Anal. Chem. 1995,67, 307-311
Mediated, Thin-Layer Cell, Coulometric Determination of Redox-Active Iron on the Surface of Asbestos Fibers Zhihua Shen, Vemon D. Parker, and Ann E. Aust* Department of Chemistry and Biochemistry, Utah State University, Logan, Utah 84322-0300
Redox-active iron on the surface of asbestos fibers was detected and quanaed using a thin-layer cell, coulometric method with soluble mediators to shuttle electrons between the mineral fibers and the solid electrode. The working and counter electrodes consisted of gold films on a glass slide with reference electrodes of silver. Asbestos fibers were entrapped in a thin-layer cell of 25 pm thickness. Hexaa"ineruthenium(1I) or o-dianisidine (dication) was used as the reducing or oxidizing mediator, respectively. Hexaa"ineruthenium(II1) undergoes a one-electron reduction, and protonated odianisidine undergoes a sequential two-electron oxidation. The measurement involved determination of the total charge for the oxidation or reduction of surface-immobilized Fe(1I) or Fe(1II) on the asbestos fibers. Analysis of the results showed that crocidolite and amosite have 4.3 f 0.7 and 3.3 f 0.7 nmoVmg of total redox-active iron that is accessible to the mediators, respectively. This corresponded to a surface coverage of accessible redoxactive iron of approximately 4.3 x lo-" mol/cm2 for crocidolite and 9.5 x 10" mol/cm2 for amosite. Furthermore, Fe@) constituted 76%or 25%of the accessible redox-activeiron on the surface of crocidolite or amosite, respectively. The method may be applied to other types of solid materials with redox-active species on their surfaces. Exposure to asbestos results in an increased risk of bronchogenic carcinoma and mesothelioma.' The mechanism by which asbestos causes cancer is not well understood. However, the most carcinogenic forms of asbestos, crocidolite and amosite, contain 27% iron by weight.2 It has been proposed that iron associated with the carcinogenic fibers may catalyze sequential reduction of oxygen, leading to the formation of the highly reactive *OH, which can damage DNA, resulting in ~ a n c e r . ~In the presence of a physiologically important reductant, such as ascorbate, Feu10 may also be reduced to F e O , producing a high level of *OH in a continuous manner. Iron can be mobilized from crocidolite or amosite asbestos by chelators, such as citrate, in a * ?his author may be reached at (801) 797-1629,FAX (801)797-3390,or email
[email protected]. (1) Khan, G.; Mahmood, N.; Arif, J. M.; Rahman, Q. .] Sci. Indus. Res. 1992, 51,507-514. (2)Campbell, W.J.; Huggins, C. W.; Wylie, A G. Chemical and Physical Characterization ofAmosite, Chrysotile, Crocidolite, and Nonfibww Tremolite for Oral Ingestion Studies @ the National Institute of Environmental Health Sciences; Bureau of Mines Report of Investigations 8452; United States Department of the Interior: Washington, DC, 1980. (3) Lund, L. G.; Aust, A E.Biofactors 1991,3, 83-89. 0003-2700/95/0367-0307$9.00/0 0 1995 American Chemical Society
redox-active form which can catalyze the formation of DNA singlestrand breaks by *OH+ However, iron on the surface of the fibers may also be redox-active and may play an important role in the biological activity of asbestos. Traditional electrochemical methods cannot be used to determine the amount of redox-active iron on a solid surface, since these techniques only measure the redox-active species in solution. A mediated, thin-layer cell, coulometric technique developed by Widrig and Majda5was used to determine the amount of redoxactive viologen silane derivatives immobilized on Au, Si02, and A 1 2 0 3 surfaces.6-* We have now applied this method with moditications to determine the amount of redox-active iron on asbestos fibers. The method described here made it possible to determine the surface redox-active iron on asbestos fibers and may be applied to other solids with redox-active species on their surfaces. EXPERIMENTAL SECTION
Materials and Reagents. Asbestos was obtained from Richard Griesemer, NIEHS/NTP (Research Triangle Park, NC). HexaamminerutheniumUII) chloride, o-dianisidine, 2-mercaptoethanol, potassium hexachloroiridateo, and lithium perchlorate were obtained from Aldrich Chemical Co. (Milwaukee, WI). Acetonitrile W was obtained in high purity from Burdick & Jackson (Muskegon, MI). The carcinogenic substances,asbestos and dianisidine, were handled in a sealed glovebox or with the investigator wearing gloves and a mask. Ferrous sulfate was obtained from Mallinckrodt, Inc. Paris, KY). Water for all purposes was prepared by passing distilled water through a Barnstead E-pure water puritication system. All other chemicals used were analytical grade. Electrode Preparation. Before preparation of electrodes, glass slides were cleaned by submersion in fresh dichromic acid solution for 12 h, followed by thorough rinsing with water. The three-electrode pattern, shown in Figure 1,was prepared by vapor deposition of the following metals onto a 6.56 cm2glass slide. First, an adhesion layer of chromium, 5 nm thick, was deposited, followed by approximately 100 nm of gold as the working and counter electrodesand 100 nm of silver as the reference electrode. The surface area of the working electrode was 0.13 cm2. The working electrodes were cleaned in freshly prepared, cold dichro(4) Lund, L. G.;Aust, A E. Arch. Biochem. Biophys. 1991,287,91-96. (5)Widrig, C.A;Majda, M. Anal. Chem. 1987.59,754-760. (6)Miller, C. J.; Widrig, C. A; Charych, D. H.; Majda, M.]. Phys. Chem. 1988, 92,1928-1936. (7) Widrig, C. A; Miller, C. J.; Majda, M.]. Am. Chem. SOC.1988,110,20092011. (8)Widrig, C. A; Majda, M. Langmuir 1989,5,689-695. Analytical Chemistry, Vol. 67, No. 2, January 15, 1995 307
Working-electrode with asbestos on it
Reference-electrode
0
Counter-electrode
1 - 1
Contacts of electrodes Figure 1. The three-electrode pattem for the thin-layer cells. Gold was deposited as the working and counter electrodes and silver as the reference electrode. The dotted line indicates the level of the electrolyte solution on the electrode.
mic acid solution for several minutes and then washed thoroughly with water. To prevent adsorption of o-dianisidine, the working gold surface was submerged in 20%2-mercaptoethanol/ethanol for 30 min to form a monolayer on the gold surface? The reference electrode used to measure E112 for various compounds was a Ag/AgCl/saturated KCI electrode. This was different from the one used for the measurements of redox-active iron on the surface of the fibers in the thin cell. 'Ihin-Layer Cell. Approximately 100 pg of asbestos was weighed using a microbalance (M5 SA, Mettler Instrument Corp., Hightstown, NJ). The asbestos sample was placed directly on the working electrode. The thin-layer cell was made by entrapping the asbestos between the glass slide containing the electrodes and an acidcleaned glass backplate separated by 25 pm aluminum foil spacers. The backplate covered only the lower portion of the electrodes, corresponding to the area below the dashed line in Figure 1. The electrodes and the backplate were held firmly together with metal clips. Electrical contact was made by attaching electrical connectors at the points where the film electrodes protrude from the cell above the glass backplate. The thin-layer cell was filled with solution by touching its lower edge to the surface of the mediator solution. The experiments were performed anaerobically by purging the mediator solution with argon in a closed Teflon beaker for at least 30 min prior to filling the thin-layer cell. This prevented oxygen from reacting with iron and affecting the results. The thickness of the thin-layer cells used in these experiments was well reproduced by assembling the cells in the manner described in the previous paragraph. Results from cyclic voltammetric experiments with aqueous FeSO4 solutions (0.5 mM or 1.0 mM) showed that the thickness was 27 f 1pm (n = 4). Cyclic Voltammetric Experiments. The cyclic voltammetric experiments were performed using an EG & G PAR Model 273A potentiostat/galvanostat (Princeton, NY) and a Yokogawa Model 3023 X- Y recorder (Newnan, GA) . The scan range was from 0.7 to -0.5 V at a scan rate of 10 mV/s. The reference electrode used in the measurements was Ag, which was deposited on the glass slide adjacent to the working electrode (Figure 1). The potential of this pseudo-reference electrode varied vs a Ag/AgCI/ saturated KCl electrode relative to the C1- concentration of the mediator solution. The concentration of the hexaammineruthe nium(I1I) was 1.2 mM and that of the dianisidine was 0.8 mM (9) Porter, M. D.; Bright, T. B.; Allara, D. SOC.1987,109.3559-3568.
L:Chidsey, C. E. D.J.
Am. Chem.
308 Analytical Chemistry, Vol. 67, No. 2, January 15, 1995
unless stated otherwise. The supporting electrolyte was 0.5 M KCl, pH 2.6. Integration of the cyclic voltammograms was done manually and was used to calculate the amount of accessible, surface redox-active total iron and Fe(II) on crocidolite and amosite asbestos. Spectrophotometric Experiments. Potassium hexachloro i r i d a t e o (0.1-0.3 mM) was prepared in CH3CN containing 0.1 M LiC04. Crocidolite asbestos was incubated in this solution at 5.0 mg/mL, anaerobically, by continuous purging with Ar. Aliquots were removed at specified times, and the fibers were removed by centrifugation (Model 59 A microcentrifuge, Fisher Scienac, Pittsburgh, PA). The absorbance of the supernatant and the IrCl+!- solution before incubation was measured at 490 nm using a Shimadzu W 160 spectrophotometer (Kyoto,Japan). Th6 concentrations of IrCk2- were calculated from the absorbance measurements using the molar absorptivity (4085 cm-l M-').lo The amount of accessible, surface redox-active ferrous iron was calculated kom the concentration difference between the solutions before and after incubation and is expressed as nanomoles of ferrous iron per milligram of asbestos. RESULTS AND DISCUSSION
Mediators and Electrolyte Conditions. Mediators were used in the experiments to shuttle electrons between the solid electrode and the solid sample. The experimental design required two mediators which were selected on the basis of the following criteria. One mediator must have a redox potential sufficiently more positive than that of Fe(III)/Fe(II) to ensure oxidation and the other sufficiently more negative than that of Fe(III)/Fe(II) to ensure reduction. The mediators must not chelate the iron from the surface of the asbestos fibers and mobilize it into solution. In addition, the two mediators must not form stable complexes with each other, nor with the electrodes or the asbestos. The redox potential of Fe(III)/Fe(II) on the surface of asbestos fibers is not known, but the redox potential of aqueous FeSO4 solution was determined to be E112 = 0.51 V. This was used as a starting point for selection of the appropriate mediators. Hexaammineruthenium [RU(NH~)~~+] was chosen as the reducing mediator. It undergoes a one-electron oxidation with E112 = -0.17 V. The dication form of o-dianisidine @IA2+)was selected as the oxidizing mediator, and it undergoes an apparent twoelectron reduction with E112 = 0.53 V at pH 2.6. The potential differencesbetween the mediators and the redoxactive species to be determined are critical for complete mediation. A potential difference of 100 mV is considered enough to assure a complete oneelectron mediati~n.~It is apparent that the mediation by Ru (NH3)62+is thermodynamicallyfavorable because of the large potential difference. For the mediation of ferrous iron by DIA2+, a 20 mV potential difference, as described in the previous paragraph, might appear insufficient, even though this is a twoelectron mediation. However, the redox potential of the Fe(III)/Fe(II) couple on the surface of the fibers appeared to be greater than 200 mV more negative (Eo = 0.33 V-0.52 V)ll than that of the Fe3+/Fe2+ couple in aqueous solution (Eo = 0.77 V).ll Therefore, these two mediators should meet the potential requirement. The detailed redox reactions of dianisidine are shown in Figure 2. When the pH of the solution is less than 3, the amino groups (10) Wong, K J. Am. Chem. SOC.1979. 101,5593-5603. (11) Write, k F.; Yee, k Ceochim. Cosmochim. Acta 1985.49, 1263-1275.
Me4
pMe
8.0
r------
6.0 Me0
pMe
3
Me0
,OMe
MeO,
4.0
-
2.0
-
I-
2
,OMe
w
E3
0 -
0
-2.0 -
Me0
E
-4.0
-
-6.0
-
I
OMe
I
I
I
I
I
H 2 & = b = d h H 2
Figure 2. Reaction pathways for oxidation of dianisidine. These represent two pathways for oxidation of dianisidine. Pathway I is A B C D E, which predominates when the pH is below 3.0. Pathway II is A 6 F D E, which predominates when the pH is greater than 3.0.
- - - -- - - -
on dianisidine are protonated.12J3 The protonated molecule [A, DIA(2H2+)]must lose a proton (B) before oxidation can take place. This oxidized form of dianisidine [C, DW+(H+)I,a radical cation, can lose a second proton (D) to be oxidized again to form a dication (E, DIA2+,pathway I). Protonation is greatly decreased when the pH is greater than 3. Under these circumstances, the neutral compound (F, DIA) becomes predominant in solution. It will undergo two oneelectron oxidations, from F to D and from D to E, pathway 11. The cyclic voltammogram for oxidation of dianisidine showed one redox peak when the pH was below 3,13 because the potential needed for the reaction from B to C is approximately the same as that from D to E (Figure 2). However, the potential for the oxidation from F to D is smaller than that from B to C, since the charged amino compound B makes the oxidation more difficult. As a result of this, the cyclic voltammogram of dianisidine, when the pH was above 3, showed two redox peaks.'3 The reduction of dianisidine would follow the reverse of pathway I (Figure 2) when the pH is below 3, and the reverse of pathway I1 when the pH is greater than 3. A solution pH less than 3 was selected for these experiments because the voltammograms were difficult to interpret when the pH was greater than 3, due to the complexity of the reactions occurring. Iron mobilization from asbestos fibers by chelators has been shown to be inversely proportional to the pH of the solution in which the fibers are suspended.14 In solutions below pH 2.5, iron was leached from the fibers even in the absence of a chelator, which made interpretation of the cyclic voltammogram impossible. ~~
~
(12) Barek J.; Berka, A Collect. Czech. Chem. Commun. 1976.41, 1334-1342. (13) Barek, J.; Berka, A; Tocksteinovi, Z.; Zma, J. Talanta 1986, 33, 811815. (14) Lund, L. G.; Ausf A E. Arch. Biochem. Biophys. 1990,278, 60-64.
Since the goal of these experiments was to study surface redoxactive iron, the pH was maintained above 2.5 to prevent leaching. The final pH of the electrolyte solution to be used also depended upon the conditions necessary for the oxidation/reduction reactions with the mediators selected. An initial pH of 2.6 was selected as the lowest pH that could be used for the time periods of these experiments which would not damage the asbestos but would simplify the oxidation/reduction chemistry of dianisidine. Scans conducted with crocidolite or amosite in the absence of mediators at pH 2.6 showed only background current, with no peaks indicative of iron leaching into the solution (data not shown). Surface Adsorption on the Working Electrode. The cyclic voltammogram of dianisidine for the Au electrode is shown in Figure 3. The reversible peak at -0.35 V may be due to adsorption of dianisidine on the surface of the Au electrode, resulting in a change in the potential of dianisidine. From the calculated approximate amount of dianisidine (6.5 x lo-" mol) that would be responsible for the shoulder peak observed, divided by the surface area of the working electrode (0.13 cmz), a surface coverage of 5 x 10-lo mol/cm2 for the dianisidine adsorption was calculated. This was very close to the surface coverage of silylviologen (5.5 x 10-lo mol/cm2> of similar molecular size, which was immobilized as a monolayer on a gold electrode.5 In order to eliminate any surface adsorption of dianisidine, the Au electrode surfaces were treated with Zmercaptoethanol, which formed a monolayer on the Au electrode.9J5J6 The modified Au electrodes showed no shoulder peaks, indicating that dianisidine ~~~
(15) Sabatani, E.; Rubinstein,I.; Maoz, R;Sagiv,J.]. Electroanal. Chem. 1987, 219, 365-371. (16) Bain, C. D.; Whitesides, G. M.]. Am. Chem. SOC.1988, 110, 3665-3666.
Analylical Chemistry, Vol. 67,No. 2,January 15, 1995
309
sz
8.0
-
6.0
-
4.0
-
2.0
-
-
I-
W
E3
0 -
0 -2.0 -
-4.0
mL, the effect of crocidolite on the electrolyte pH was tested. With only 10 mg of crocidolite/mL, the pH increased from 2.6 to 5.4. Clearly, under the experimental conditions (300 mg of fibers/ mL), a pH change is expected. This would favor the formation of the neutral form of dianisidine (F, Figure 2), and the oxidation would proceed F D E (pathway 11, Figure 2). The oxidation F D requires less positive potential than D E, resulting in the observation of the two oxidation peaks at -0.35 and 0.53 V (Figure 4). This explanation would appear to require that the reduction process of dianisidine should also appear as two peaks. This was sometimes observed but not reproducibly. We conclude that whether or not two peaks are observed is very dependent on the pH change that takes place in the presence of the fibers and that this change may be dependent upon the amount and type of sample being used. After reduction of dianisidine (DIAz+) by Fe(II) on the fiber surface, DIA(2HZf)diffused back to the working electrode and was reoxidized to DIAz+. This process continued until all of the Fe(II) was converted to Fe(I1I). The charge determined as a result of this scan (Qz, Figure 4) was due to the oxidation of DIA(2H2+)and Fe(II). When the potential was scanned from 0.7 to 0 V, the charge determined (91) was due only to the reduction of DIAz+ and had no contribution from iron, since all of the iron was Fe(I1I) at this point. Thus, the difference in the areas of Qz and 91was due to the charge resulting from oxidation of the Fe(ID. This was converted to moles of Fe(II) using the Faraday constant (fl as shown in eq 1.
-
0.60
0.40
0.20
0
-0.20
-0.40
POTENTIAL (V vs. Ag/AgCI, sat'd KCI) Figure 4. Mediated, thin-layer cell cyclic voltammogram of surface redox-active iron on 83 pg of amosite asbestos. eDianisidine (0.6 mM) and hexaammineruthenium(1.3 mM) were used as mediators in 0.5 M KCI, pH 2.6. The scan rate was 10 mV/s. The dotted line represents the estimated background current, which was subtracted to calculate the charge. The areas represented by QI, and Q4 are the total charges resulting from the electrochemical reactions as the potential is scanned.
a,a,
was not adsorbed on the surface (Figure 3). Because 2-mercap toethanol formed a monolayer on the Au electrode which did not hinder the electron transfer processes, the quantitation showed no significant difference between the bare and modified electrodes (Figure 3). Determination of SurEace Redox-Active Iron with the Mediated, Thin-LayerCell, CoulometricMethod. To measure the quantity of surface redox-active iron on the asbestos fibers, the working electrode in the thin-layer cell was cycled between f0.7 and -0.5 V. The applied potential was first scanned from 0 to 0.7 V. During this scan, DIA(2H2+)was oxidized to DIAz+at the working electrode via pathway I shown in Figure 2. DUz+ then diffused from the working electrode to asbestos, where it oxidized Fe(II) to Fe(I1D on the surface of the fibers, resulting in reduction of DIA2+. The oxidation process for dianisidine (Figure 4) appears as a double peak. We considered two explanations for this observation. The oxidation/reduction of dianisidine involved binding/release of protons which would change the pH in the cell (Figure a), but this does not appear to be responsible for the extra peak, which was not observed in the absence of asbestos (Figure 3). A more likely explanation concerns the ability of amosite (or crocidolite, where the same phenomenon was observed) to bind protons.17 The addition of asbestos fibers to acidic solutions can raise the pH. Since the quantity of asbestos in the thin-layer cell experiments was at least 300 mg of fibers/ (17) Guthrie, G. D., Jr.; Mossman, B. T. Health Efects of Mineral Dusts; BookCrafters, Inc.: Michigan, 1993; Chapter 8. 310 Analytical Chemistry, Vol. 67, No. 2, January 15, 1995
--
FeOD = (Qz
-
-
Q1)/F
(1)
Total iron was determined by scanning the potential from 0 to -0.5 V (Figure 4). In this case, R u ( N H ~ ) was ~ ~ +reduced at the working electrode to Ru(NH3)s2+.It then diffused to the surface of the fibers, where it reduced the surface iron [Fe(III) form] to Fe(II), itself b e i i oxidized to Ru(NH3)e3+.This process continued until all of the Fe(I1I) was reduced to Fe(II). The total charge determined (Q3) after this scan was due to reduction of RU(N&)~~+ and Fe(I1I). When the potential was scanned from -0.5 to 0 V, the total charge determined (94)was due to the oxidation of Ru(NH3),j2+,with no contribution from iron. Thus, the difference between the areas of Q3 and Q4 was equal to the charge resulting from the reduction of the total iron which was in the oxidized form. This was converted to total iron as shown in eq 2. After
Fe(tot) = (Q3 - QJ/F
(2)
the h s t complete scan cycle, all of the redox-active iron was in the same oxidation state. Total iron could then be determined using either mediator. For a single experiment, the standard deviation of the average of total iron for both mediators was 0.9 nmol/mg, which represents 20%error. F e O was calculated using eq 1 only, where 91and QZwere obtained from the first scan cycle. Total iron was calculated by averaging results from both mediators after the h s t scan cycle, since the results from both mediators were statistically identical. The precision of the method depended primarily on the weighing of the asbestos fibers, since weighing was the major source of error. The same crocidolite sample,weighed 20 times, showed an average f standard deviation of 105 f 16 pg. The
crocidolite. The potential of IrCP-13- could not be measured with the thin-layer cell system, because it was more positive than the gold stripping potential. Using a planar Pt electrode (0.8 mm diameter) as the working electrode, cyclic voltammetry of crocidolite amosite hexachloroiridate (IrC&2-/3-)in 0.5 M KCl showed that Ell2 was Fe(tot) Fe OD Fe (tot) FeO 0.82 V, which was -300 mV more positive than that of dianisidine 4.3 f 0.7 3.3 & 0.7 (76%)b 3.6 & 0.7 0.9 f 0.3 (25%)b in the same system. This would mean that IrCb2- should have sufficient positive potential to react with the surface ferrous iron Data are re resented as the mean f 95%coniidence interval (n = 10) determined) from Student's t test. Numbers in parentheses on the fibers. The relatively high molar absorptivity of IrCb2represent FeOD as percent of total iron. at 490 nm in CH&N (4085 cm-l M-l) compared with that of IrCb3- (10 cm-l M-l)l0 made it possible to determine the amount of IrCb2- reduced by Fear) on the surface of crocidolite. precision of the determination was also related to the signal to The experiments were done in acetonitrile solution since noise ratio of the cyclic voltammograms and to the charge IrCl&underwent hydrolysis in aqueous solution under the condiintegration. The signal to noise ratio presented no problem, as tions required for the assay.'* Three trials of experiments showed evidenced by the typical scans displayed in Figures 3 and 4. The the surfaceaccessible, redox-active ferrous iron on crocidolite to integration was completely dependent upon the background be 3.5 f 0.1 nmol/mg. The result was in very good agreement current estimates (see dotted line in Figure 4). The error due to with the result of using the thin-layer cell method (Table 1). This estimation of background current was determined by integration confirmed the thin-layer cell method. of the charge ratios of oxidation to reduction without asbestos fibers for DIA2+/DIA(2H2+)(0.98 f0.07, n = 20) or R u ( N H ~ ) ~ ~ + / ~ + The precision of the spectrophotometricmethod was greater than that of the thin-layer cell method, since the spectrophoto(0.93 f 0.07, n = 20). metric method was solely dependent upon the absorbance The amount of redox-active iron accessible to the mediators measurements. In addition, the larger amount of sample needed determined for crocidolite or amosite is shown in Table 1. (10-15 mg) could be weighed more precisely with an analytical Crocidolite or amosite contains 27.3% or 28.6% iron by weight, balance. The two disadvantages of the spectrophotometric respectively, in the bulk sample? However, more accessible method were that it measured only redox-active FeaI) and that a redox-active iron was detected on the surface of crocidolite per larger amount of the sample was required (by 2 orders of weight than on amosite (Table 1). When corrections were made magnitude) compared with the thin-layer cell method. This for the surface area of the fibers (10.1 m2/g for crocidolite or 3.8 technique cannot be routinely used for samples of limited supply. m2/g for amosite) ,2 crocidolite had 4.3 x lo-" mol of accessible Therefore, the mediated, thin-layer cell, coulometric method is redox-activeiron per square centimeter, compared with 9.5 x lo-" very useful for determining the surface redox-active iron on mol/cm2 for amosite. Ferrous iron constituted 76%or 25%of the asbestos fibers and other solid surfaces. surface iron on crocidolite or amosite, respectively. These
Table 1. Medlated, Thin-Layer Cell, Coulometric Detennlnationof Redox-Actlve Iron on the Surface of Asbestos Fibers (nmoUmg)a
percentages are consistent with the results from previously published work where Fe(II) and Fe(I1I) were quantified after mobilization from the surface of the fibers by chelators! However, they are in contrast with the percentage of ferrous iron in the bulk composition, 52%for crocidolite or 94%for amosite? These dfierences emphasize the importance of being able to measure the actual amount of redox-active iron on the surface of these fibers, since there is increasing evidence to suggest that the reactions catalyzed by iron may be important in the biological activity of asbestos. Measurement of Surface Redox-Adive Ferrous Iron on Crocidolite Using a Spectrophotometric Method. An independent method of determining the surface redox-active ferrous iron was used to compare results with the thin-layer cell method. The asbestos fiber used for this experiment was crocidolite only, since this experiment was exclusively for the confirmation of the thin-layer cell method. Hexachloroiridate (IrCb2-) was selected as the oxidizing reagent to determine the surface redox-active ferrous iron on (18) Poulsen, I. A; Gamer, C. S. J Am. Chem. Soc. 1962,84,2032-2037.
CONCLUSION
The mediated, thin-layer cell, coulometric method described in this report is very general. It may be applied to other mineral fibers or other solid materials containing iron or other transition metals. The potential window can be changed simply by using alternate mediators. ACKNOWLEWMENT We gratefully acknowledge Dr. Cindra A. Widrig for her contributions in helping design and develop this technique. We acknowledge Dr. Cindra k Widrig, Dr. Gregory M. Swain, and Jeanne k Hardy for reviewing this manuscript and appreciate many helpful discussionswith Fredrick Norrsell. This work was supported by Grant ES05782 from the National Institute of Environmental Health Sciences (ILEA). Received for review June 16, 1994. Accepted October 27, 1994.a AC9406080 @Abstractpublished in Advance ACS Absfracts, December 1, 1994.
Analytical Chemistry, Vo1. 67, No. 2, January 15, 1995
311
