Document kmGBx5rY4yZ2K8xa282N2m580

. .'i/i| . < r> 1 f-f. r c< 77 t^'-' | ** V/*3 AfrV April 1970 J:oiccr 72Jiff*! 72, G, 44 lomlwin, J. T., "Bc o of Water-Oriented : preliminary Texas exes Water Develop- 8J, 2S (10GS). ;:i;- bakes Can be a like. Woria, ) its, A. W., "Cooling Southwest." Paper lnu;riean'Society of pr. 1901), United Na vis on thonrdl over tin: is are tabure available S TRACE METAL CHARACTERIZATION IH AQUATIC ENVIRONMENTS BY ANODIC STRIPPING VOLTAMMETRY H. E. Allen, W. R. Matson, and K. H. Money Generally it is accepted that trace metals such as Cu, Zn, Cd, Co, Cr, Ni, and Mo play a significant role in bio logically mediated reactions in the aquatic environment. In trace quanti ties many of these metal ions are es sential micronutrients for enzymatic transformations, but in high concen trations they may be inhibitory or toxic to biological systems. The con trolling role of metal ions on the eco logical balanee and productivity of fresh waters has been emphasized by several authors (1) (2) (3) (4). For example, molybdenum, considered an essential micronutrient for nitrogen fixation and the formation of the plant enzyme nitrate reductase, controls the primary productivity of Castle Lake, Calif. (1). The traee metals in the aquatic en vironment also may serve as '`'mark ers" in the identification of water masses. The distribution and relative abundance of metal ions may offer valuable information on the geochemi cal history and pollutional character istics of natural waters. The exchange of metal ions between the hydrosphere and the biosphere, lithosphere, and atmosphere must he considered in studying metals in natural waters. S. E. Allen, W. B. Matson, and K. E. Money are, respectively, Chemist, Bureau of Commercial Fisheries, V, S, Department of the Interior, Assistant Professor, and Assodate Professor, Department of Environmental Health, School of Public Health, University of Michigan, Ann Arbor, Mich, The paper was presented at the 41sl An nual Conference of the Water Pollution Con trol Federation, Chicago, III., Sept. 22-27, ises. Therefore, tracc-metal analysis is im portant in most water pollution charac terization programs. The availability of a trdee metal to a biological system depends on the oxi dation state, degree of hydration, and organic eomplexation of the metal ion. Shapiro (4) emphasized this point in his studies on the availability of iron as an algal nutrient, when he showed that the availability of iron to the biological system depends on the de gree of eomplexation of iron with yel low organic acids extracted from lake water. He concluded that in biologi cal systems it is important to differen tiate between the analytical concentra tion of metal ions in solution and the metal ions available to the biota. Trace metals in the aqnatie environ ment arc frequently in the part-perbillion range or lower (lO^-lO'10!^), which places them below the sensitivity limit of most analytical proeednres. (A summary of the sensitivity limits of various analytical procedures is given in Table 1.) Except by activa tion analysis and anodic stripping voltammetry, direct determinations of metal ions cannot be made without concentrating the water samples. Sam ple pretreatment, i.e., separation or concentration by various procedures, usually modifies the physicochemical characteristics of the species under investigation. Although activation analysis offers the high sensitivity required for tracemetal analysis, its use is limited by the availability of reactor facilities; fur thermore, it is an elemental analysis procedure that offers no information N36884 DUP050311841 574 JOURNAL WPC5* April 1970 TABLE I.--Sensitivity of Some Analytical Methods for the Determination of Trace Metals Method of Analysis Sensitivity Molecular absorption spectro photometry Molecular fluorescence spec trophotometry Atomic absorption spectro photometry Atomic fluorescence spectro photometry Optical anU emission spectroscopy Neutron activation analysis Potentiometry with metal specific glass or membrane electrodes Classical polaroEraphv Derivative polarography Square-wave and linear-sweep voltammetry Anodic stripping voltammetry with hanging mercury drop electrodes Anodic stripping voltammetry with thin film or solid electrodes io -mo -m lO-'-lO-W io-'-icrw io-t-io~w 10-*-10-5jW 10-M0-W 10-'-10-sM lOr-'-lQ-m about oxidation state or degree and type of complexation. Perhaps the best method for tracemetal characterization is anodic strip ping voltammetry, which potentially is capable of performing in situ analysis of free and complexed metal ions in natural waters. Stripping voltam metry, in general, consists of a deposi tion step in which the desired compo nent is deposited eathodieally or anodieally as a solid or an amalgam, and a stripping step (reverse electrolysis) in which the components are deter mined. Theory Ordinarily, the deposition step is performed at controlled potential in a uniformly stirred solution. The depo sition under reproducible stirring con ditions follows the first order kinetic expression: ' Qc = Qo(l - crkt)..............1 where Qt = number of coulombs in the stripping peak at any time, Q,, = num ber of coulombs equivalent to the total metal ion in the sample solution, t -- deposition time (see), and h = rate constant. The rate constant, k, is equal to: where: D == diffusion constant of eleetroactive species in the test solution, sq em/see, A = electrode surface area, sq cm, V = volume of the solution, eu em, and S = effective thickness of dif fusion layer (em). Ordinarily partial deposition of a sample is sufficient for analysis. Un der conditions where the rate constant is not reproducible for different sam ples, extension of the plating increases the aeeuracy of the determination. With a long deposition time, the ex pression. e~M approaches zero; there fore slight changes in Tc which may occur between analyses under adverse working conditions, e.g., on board re search vessels, have little effect on the precision of the results. The depo sition step can be arranged to segre gate the species present as well as deposit a reproducible fraction of the sample. Quantitative determination of the deposited materials is made in the stripping step, most commonly by a chronoamperometrie technique with a linear potential sweep. Commercial polarographic equipment employing operational amplifiers may be used for slow potential sweep rates, in the order, of 2 to 100 mv/sec. A characteristic stripping curve is shown in Figure 1. "When the proper electrode is used under controlled conditions, the con centration of metals in solution is linearly proportional to the peak cur rent, and qualitative identification of the metal ions in solution is obtained from peak potential values (5). These relationships for a hanging mercury drop electrode (IIMBS) are approxi- Vol. 42, No. 4 mated by the I (6): ip -- /' where: iv -- pp; geometry and j number of eler electrochemical age scan, v/sec tion of metal in The conceutr sample also ma; metrically. Tin related to the p potential, E,n, Ep = B, where F = Farn Ef will oeeur i volt more catlioi Elect Three main tj been used in sti DE's of the K< idler (8) types, (9), and mereni (10) (11). The HMDB's for hydrogen e insure a eonstan deposited in the their thickness, 1 metal must diffu during the strij fusion results Solid electrodes a greater anodic but their surfaei produce and the activity deposits complex systems, terns of diffusion negligible in Ihi trodes, sensitivit; increased, and tl age is comparable DE. However, t sag on metal am: the formation of si DUP050311842 April 1970 cr of coulombs in tho any time, Q.0 -- numqnivalent to the total sample solution, t -- (sec), and k = rate fc constant, k, is equal ^A~ ........... 2 ision constant of elecin the test solution, jlectrode surface area, me of the solution, cu etive thickness of dif- i)irtial deposition of a ent for analysis. Un.licre the rate constant ible for different sam> the plating increases of the determination, ^position time, the expproaches zero; thcrenges in k which may analyses under adverse ions, e.g., on board rehave little effect on the results. The depobe arranged to segrees present as well as jducible fraction of the determination of the erials is made in the , most commonly by a aetrie technique with a al sweep. Commercial equipment employing ipliflers may be used for sweep rates, in the order iv/sec. A characteristic e is shown in Figure 1. oper electrode is used led conditions, the con- metals in solution is irtional to the peak curilitative identification of s in solution is obtained ential values (5). These for a hanging mercury e (HMDS) arc approxi- Vol. 42, No. 4 ANODIC STRIPPING VORTAMMETUV mated by the Bandlcs-Sevcik equation (.6) : IOC 575 i,, = IKz^-A.D^'^Co........... 3 where: peak current for a given geometry and plating time, amp, z = number of electrons involved in the electrochemical step, v -- rate of volt age scan, v/sec, and C,, ~ concentra tion of metal in the sample. The concentration of metal in the sample also may he determined eoulo- metrieally. The peak potential, EP) is related to the polarographie half-wave potential, i/2, by JSv = Ew - l.IRT/zF......... 4 where F - Faraday. This means that Bl: will oecur at a potential, 0.028/a volt more cathodie than 2?i/2 at 25 C. Electrode Types *0-65 -0.55 VOLTS vs, SC.E. 0.35 FIGURE 1.--Anodic stripping voltam- Three main types of electrodes have been used in stripping analysis: HM DS!'s of the Kemula (7) and G-eris- cher (8) types, solid metal electrodes magram shovring the sensitivity and range of the method. Full-scale ranges were 100 iia (Zn'=), 2 jua (CcT), 5 pa <Pb), and 20 ma (Cu"). (9), and mercury thin-film electrodes (10) (11). The HMDS's offer high overvoltage for hydrogen evolution, and usually compounds, e.g., Ni-Zn or Co-Zn, is enhanced. insure a constant activity in the metal The working electrode used in deposited in the amalgam. Because of the present study was a composite their thickness, however, the deposited mercury-graphite electrode (GMGE), metal must diffuse out of the mercury which is composed of small droplets during the stripping step; this dif of mercury electroplated on a graphite fusion results in broadened peaks. surface impregnated with paraffin. Solid electrodes are rugged and have The graphite serves as an inert sup a greater anodic range than HMDXS's, port and the dispersed mereury is but their surfaces arc difficult to re more durable than a mercury film on produce and the formation of mixed metal. This electrode system offers activity deposits limits their use in high sensitivity and repuodueibil ity, complex systems. Because the prob lems of diffusion iu the mercury arc and a ruggedness that, makes it very suitable for field work. negligible in thin-film mercury elec Roe and Toni (10) derived and trodes, sensitivity and resolution are tested the anodic stripping equation increased, and the hydrogen overvolt for a mercury-film electrode in a age is comparable to that for the HM stirred solution. Within the limits of DS. However, the mercury film may their approximation the authors sag on metal, amalgam electrodes, and proved that at constant stirring rate the formation of bimetallic association with a known deposition time (tj) the peak current is directly proportional DUP050311843 570 JOURNAL WPCF April 197Q to the metal ion concentration in solu tion. zFA l H Ca............ 5 where l is the thickness of the mer cury film, t/j is equal to zW/RT, and e is the base of Napierian logarithms. The peak current is directly propor tional to the potential scan rate; in this characteristic the electrode differs from the HMDE, in which the peak current depends on v1/z, as shown in Equation 3. The peak potential equa tion for the mercury-film electrode was given as follows (10): EP + j log^........... 6 <p JJ where E is the standard electrode po tential (which is equal to E-^ in many systems) and S is the thickness of a boundary diffusion layer at the solution-electrode interface. In con trast to that in Equation 4, the peak potential depends on the logarithm of the scan rate. -Experimental variation in S and l will have little effect on Ev, which generally is cathodic to E. Qualitative and quantitative infor mation on free and combined metal ions and metal-binding ligands and their distribution can be gained by varying the sample pretreatment and the anodic stripping procedures. A flow chart showing various analytical procedures is given in Figure 2. Total metal content can be determined by digestion of the sample before anodic stripping analysis, and filtration sepa rates dissolved from particulate frac tions on the basis of variations in par ticle size. Complexation Metal ions in natural waters may be eomplexed with simple inorganic ligands such as water-, halides, carbo nate, and sulfate, or they may be tied up with complex organic ligands such as amino acids, organic acids, vitamins, porphyrins, liumie acids, and tannins. It is possible to exchange the metal in many complexes by additions of acid (H+) or metal cations; the freed metal ions subsequently may be determined by anodic stripping voltammetry (Fig ure 2). The metal-eomplexing ability of natural waters, which is a measurable Sample digestion total filtration digestion particulate dissolved addition of matoriai variation of ASV parameters H+ electroaetive metals non-cleetroaetive metals sweep rate plating time stripping medium plating potential mixing FIGURE 2.--Possible variations in procedures for anodic stripping voltammetric det ruination of metal ions in natural waters. i 'tty DUP050311844 April 1970 Vol. 43, No. 4 ANODIC STRIPPING VOLTAMMETRY 577 .lautilfttivc infiorcombiued metal iug ligands and m be gained by pretreatment and ; procedures. A various analytical i Figure 2. Total ie determined by pie before anodic id filtration sepa- partieulate fracvariatious in par- ration tural waters may simple inorganic er, halides, cai-bothey may be tied janic ligands such lie aeids, vitamins, cids, and tannins, mnge the metal, in additions of acid is; the freed metal iay be determined voltammetry (Fig- cxing ability of ;h is a measurable Situation dissolved iV parameters E voltammctric deter- v I'1;/' characteristic, depends on the type .V !SV and eoncentuation of simple or com plex ligands. Quantitative and quali- v: ji,. tative information on metal-binding ligands can be obtained by a complex- ometrle titration procedure in which an elcetroactive metal ion is used as -rliJi-:- v.,; the titrant. In this technique the ' 1 j '. stepwise metal-ligand complex forma, tion can be followed and the stoiehio- jj'::; metric end points can be detected by []); ;I: anodic stripping. Similarly, the rates `V ! of metal-ligand dissociation on titra- ':Vv' *in with a strong acid can be deter, : mined. The analytical feasibility of anodic ;iy. {?.;.!..... stripping voltammetry can be extended />$,(>'{ H.;:,;//:greatly by varying the deposition time, > V the potential sweep rate, the stripping j{ medium, and the mixing regime. These [][" variations can be used to advantage to characterize free and eomplexed metal ions in the aqueous environment. Experimental Design >: ' 1 !: : i ,; I; ; The anodic stripping voltammetric cell and electrodes used in the present study are shown in Figure 3. The CMGE was prepared by the eleetrodeposition of mercury (with vigorous stirring) at -- 400 mv vs. Ag/AgCl, onto speetrographie grade graphite electrodes impregnated with paraffin (11). A platinum counter electrode and a Ag/AgCl reference electrode were in elcetrolytie contact with the test solution through leached Yyeor plugs. The cell was a 3- X 1-in. (7.62X 2.54-cm) quartz vial with a Teflon plug- for support of the eleetrodes. A Teflon tube with a platinum wire in sert for rigidity was used to bubble nitrogen for deaeration and mixing of the test solution. A previously described multiple anodic stripping unit (12), with Heath operational amplifier modules, was used to provide the necessary volfammetric deposition and stripping modes * Mention of products and manufacturers is for identification only and docs not imply endorsement by tlic Bureau of Commercial Fisheries. PI Counter Electrode FIGURE 3.--Anodic stripping voltammet ric cell. (In. X 2.54 = cm.) for simultaneous use of eight anodic stripping cells. Each CMGE had an effective area of about 1.3 sq cm and a half-time for deposition of 15 min for a 10-ml sample. By using similar electrode and cell designs a sensitivity of KP10 M (0.02 ppb) for lead can be achieved with a linear anodic sweep of 20 mv/sec (5) (11). A typical stripping curve for Zn, Gd, Pb, and Cu in a lake-water sample based on the use of a CMGE is shown in Figure 1.. Analysis for Trace Metals The analytical feasibility of anodic stripping voltamlhetry has been illus trated by the analysis of a number of lake- and river-water samples for thenfree and eomplexed (aeid-cxchange able) metal, content. Samples were collected at the surface from the lower Rouge and Detroit Rivers, at the sur face and bottom (15 m) from the cen tral basin of Lake Eric 10 miles (16 Ion) northeast of Sandusky, Ohio; and at 10' m from Lake Michigan near Waukegan, 111., and Imdington, Midi. The samples wore refrigerated and brought to the laboratory for analysis. The water samples were analyzed before and 30 min after aeidi- DUP050311845 578 JOURNAL WPCF April 1970 TABLE II.--Trace Meta! Concentrations in Various Natural and WasteWaters Determined by Anodic Stripping Voltammetry Sample i 2 3 4 5 6 7 8 9 Location Rouge River 0.6 mile (0.8 km) above Ford Motor Go. turning basin Rouge River turning basin Rouge River mouth Detroit River, Trenton Channel Detroit River, Livingston Channel Lake Erie, central basin, surface water Lake Erie, central basin, bottom water Lake Michigan, Waukegan, III. Lake Michigan, Ludington, Mich. Cd <0.1 7.4 1.5 0.8 1.9 -- 0.3 1.0 -- Concentration C^g/1) Free Pb i .i 2.2 0.4 0.4 0.4 -- 0.07 0.2 -- A.cid Exchangeable Cu Pb 2 15 Cu 27 14 8 18 11 -- 0.5 0.8 -- 37 11 5 6 1.8 1.6 3.3 0.6 108 19 * 28 29 6.8 1.8 19 0.4 fieation to pH 2 with, perehlorie acid, at a plating time of 15 min. Al though detailed analysis of so few data is not possible, some inferences may be drawn from the results (Table FIGURE 4.--Decrease of cupric ion added to Rouge River water (0.13 f<g Cu'" added to a 10-ml sample). II). Free and acid-exchangeable metals were much higher in the Rouge River below the point of introduction of industrial wastes (sample 2) than above (sample 1). Concentrations of aeid-exckangeable metals were lower at the mouth of the Rouge River (sam ple 3) than above most of the pollutional sources. This low concentra tion may be due to the presence in the river of large amounts of ferric hy droxide, which aet as scavengers for trace metal. ions. It was to be ex pected that large quantities of trace metals would be lost quieldy to bottom deposits by such precipitation mecha nisms. The difference in acid-cxchangcable copper in Lake Erie between the sur face (sample 6, 6.8 ppb) and bottom (sample 7, 1.8 ppb) may be due to the concentration of copper by phyto plankton in the surface water or the removal of copper from the bottom of the analyses of sample 9, from the relatively pure water in northern Lake Michigan, and sample 8, collected near 'Waukegan, 111., were dramatically different. Acid-exchangeable copper and lead were about 50- and 5-fold waters by the sediments. The results higher, respectively, in the sample col lected from the southern portion of tire lake. 1 i t ! j ;s i i ( DUP050311846 . April 1970 nnd Wastetry ration O.-./l) Acid Exoliaogcablo On 2 it 8 18 11 -- 0.5 0.8 --* Vb Cu 15 27 37 . 11 5 6 l.S 1.6 3.3 0.6 108 19 28 29 6.8 1.8 19 0.4 acid-exchangeable uglier in the Rouge int of introduction s (sample 2) than Concentrations of metals were lower Rouge River (sammost of the polluiis low concentrathe presence in the mite of ferric hy- as scavengers for It was to be exiinantities of trace t quickly to bottom rceipitation mccha- t acid-cxchangcablc e between the sur! ppb) and bottom j ) may be due to f copper by phytorfaee water or tbe from the bottom sample 9, from the ;r in northern Lake >le 8, collected near were dramatically changeable copper iut 50- and 5-fold cents. The results , in the sample eolaitlicm portion of Vol. 42, No. 4 ANODIC STRIPPING VOLTAMMETRY 579 FIGURE 5.--Release of lead in water from Lake Erie and the Rouge River after the addition of acid. DUP050311847 580 JOURNAL WPCR Aprii 1970 Characterization oE Complexes The release of metal from the Lake Anodic stripping voltammetry is ex tremely .useful for the study of kinetics of complexation reactions. Figure 4 shows the reduction of free copper in a sample after 0.13 p.g of Cu'2 was added to a 10-uil sample of Rouge River water containing' excess ligand. This sample was collected near an out fall of a Detroit wastewater treatment plant. Analyses were performed at 2-min intervals at plating times of 1 min. The complexation of cop per followed first-order kinetics with a rate cqnstant of 0.031/min. The rate constant for the complexation of a metal with naturally present ligands or the ratio of two rate constants would be extremely useful in the char acterization of ligand types present in natural waters. The aeid exchange rates for lead in water samples from the Rouge River and from Lake Brie are shown in Figure 5. The Rouge River sample Brie surface sample was much, slower and was characteristic of exchange of metal from a more stable complex. Only an extremely small amount of lead was released by acidification when the pH of Lake Brie water was changed from 2.40 to 1.60. This in crease was due almost exclusively to lead iu the aeid--most of the lead in complexes would have been exchanged by pH 2.40. This amount of aeid was used to acidify the water samples from Rouge River, Lake Brie, and other' samples listed in Table II. Titration of a ligand excess of natu ral or pollutional origin also can be done by additions of spikes of various metal ions such as Zn, Cu, In, Bi, Pb, and T1 (13). The rates of complex formation or complex exchange with specific metal ions ean be used as diag nostic criteria for the identification of labile and nonlabile complex com pounds. had an initial pH of 7.30 and con Conclusion tained 1.3 mg Fe/1. Upon acidifica tion of the sample to a pH of 6.30, the lead signal rapidly increased and then slowly decreased. The changes in the lead concentration were attrib uted to dissolution of a portion of the ferric hydroxide followed by its reformation, which exerted a slow' scavenging action on the lead. The newly formed ferric hydroxide had a seavening efficiency greater than that of ferric hydroxide originally present, possibly because a locally high acid concentration caused dissolution of ferric hydroxide, followed by a rees tablishment of equilibrium. Further additions of aeid were ac companied by a rapid release of lead at pH 2.82 and pH 2.48. Perturbation of the system by aeid additions yielded an instantaneous release of the trace Anodic stripping voltammetry is a valuable analytical procedure highly applicable to the characterization of metals in the aquatic environment. The technique extends the detection limits far below tlie sensitivity of other analytical procedures and permits the analysis of trace metals in natural waters without prior concentration. Because of its freedom from reagent contamination and its ability to dis tinguish between complexed and free metal ions, the method is considered very useful in studying the role of trace metals in biologically mediated reactions in aquatic ecosystems.. The speed and selectivity of anodic strip ping voltammetry facilitate the study of the metal-binding ability by ligands of natural or pollutional origin. metal, followed by slow aeid exchange References rates. Such transient responses to aeid 1. Goldman, C. I?-., "Molybdenum as an Es additions arc characteristics of the metal-binding ability of a given water sample. sential Microiuitrient ami Ifseful Wa ter Mass Marker in Castle Lake, Californin.." In: "Chemical Environ ment in the Aquatic Habitat.** N.V. V i \ i i il f { DUP050311848 April 1970 1 from t)io Lake vas muck slower ; of exchange of stable complex, mall amount of idifieatiou when '.vie water was 1.60. Tins int exclusively to t of the lead in been exchanged unit of acid was or samples from 'Me, and other e II. 1 excess of natu ral also can be pikes of various Cu, In, Bi, Pb, ites of complex exchange with be used as diage identification 2 complex com- >n Itammetry is a 'ocodure highly netcrization of e environment. 5 the detection sitivity of other iiid permits the als in natural concentration. 1 from reagent ability to dislexed and free l is considered l8 the role of cnOy mediated osvstems. The f anodic stripitate the study ' lify by ligands 1 origin. "Ion ii in as an Esnrnl Useful WaGustle Lake, Calicmical EnvironHabitat." N. Y. Td. 42; No. 4 ANODIC STRIPPING VOLTAMMETRY 581 Noord-Eollandsche TJitgevers Maatschappij, Amsterdam. (1967). 2. Goldman, 6. R., `'Primary Productivity and. Micro-nutrient limiting Factors in some North American and New Zealand Lakes." Verb. Inti. Ter. Liimtcl. (Germany), 15, 365 (1964). 3. Sliapiro, J,, "Yellow Organic Acids of Lake Waters Differences in their Composition and Behavior." In: ` ` Chemical Environment in tiic Aquatic Habitat." N. Y. Noord-Hollandsche, tlitgevcra Maatachappij, Amsterdam (1967). 4. Shairiro, X, "Iron Available to Algae.'' In: "Chemical Environment in the AqaaMc Habitat." N. Y. NoordHollandsche TJitgevers Maatscliappij, Amsterdam (1967). 5. Sliain, I., "Stripping Analysis." In: "Treatise in. Analytical Chemistry." Fart I, Yol. 4, John Wiley and Sons, New York (1964). 6. Kcnmia, W., and Eiiblik, Z., ` ` Applica tion of Hanging Mercury Drop Elec trodes in Analytical Chemistry." In "Advances in Analytical Chemistry and Instrumentation." Yol. 2, Interseienco, New York, N. Y. (1963). 7. Kemula, W., and Kublik, Z., "Applica tion of the Hanging Mercury Drop to the Determination of Small Quantities of Various Ions." Aval. Ghim. Ada, 18, 104 (1958). 8. Geriseher, II., "The Discharge Mechanism of Simple and Complex Zinc Ions." Zcits. Phys. Chem. (Leipzig), 202, 302 (1953). 9. Gardiner, K. W., and Rogers, L. B., "Coulomotric Determinations of Submicrogram Amounts of Cadmium and Zine." Anal. Chem., 25, 1393 (1953). 10. Roo, D. K., and Toni, J. E. A., "An Equation for Anodic Stripping Curves of Thin Film Mercury Electrodes." Anal. Chem-., 37, 1503 (19G5). 11. Matson, W. IS., Roe, D. K., and Carritt, D. E., "Composite Graphite-Mercury Electrode for Anodic Stripping Volt ammetry." Anal. Chem., 37, 1594 (1965). 32. Matson, W. R., and Roe, D. K., "Trace Metal Analysis in Natural Media by Anodic Stripping Voltammetry. " In "Analysis Instrumentation." Vol. 7, Plenum Press, New York, N, Y. (1967). 13. Carritt, D. E., "Direct Measurement of Chemical Species of Biochemically and Geochemiaally Reactive Elements in Natural Aquatic Systems." Presented at the 155th National Meeting Amer. Chem. Soc., Division of Water, Air, and Waste Chemistry, San Francisco, Calif. (1968). DUP050311849