Fractionation and Determination of Aluminum and Iron in Soil Water

Department of Chemistry, University of Oslo, Box 1033,. N-0315 Oslo, Norway, Norwegian Forest Research Institute. (SKOGFORSK), Høgskoleveien 12, N-14...
0 downloads 0 Views 74KB Size
Environ. Sci. Technol. 2002, 36, 5421-5425

Fractionation and Determination of Aluminum and Iron in Soil Water Samples Using SPE Cartridges and ICP-AES GAUTE TANGEN,† TORILD WICKSTRØM,‡ SYVERIN LIERHAGEN,§ ROLF VOGT,† AND W A L T E R L U N D * ,† Department of Chemistry, University of Oslo, Box 1033, N-0315 Oslo, Norway, Norwegian Forest Research Institute (SKOGFORSK), Høgskoleveien 12, N-1432 Ås, Norway, and Norwegian Institute for Nature Research, Tungasletta 2, N-7485 Trondheim, Norway

The use of commercially available solid phase extraction (SPE) cartridges for the fractionation of Al and Fe in soil water is described. The quantitative determination was done by inductively coupled plasma atomic emission spectrometry (ICP-AES). Different types of SPE cartridges, based on cation exchange, anion exchange, and chelation were studied. To avoid pH changes, the SPE cartridge should be conditioned with a buffer that has a pH close to that of the sample. Both strong cation exchange (SCX) and chelation were found to work well, whereas low recovery was observed for Al when anion exchange was used. For Fe, the sum of the anionic and cationic fractions that passed through the cartridges was nearly 100%. The results obtained for Al for 23 soil water samples using a SPE/ SCX cartridge and ICP-AES were compared with equilibrium calculations using the program ALCHEMI and also with a fractionation method that was based on separation on a manually prepared SCX column and detection by molecular spectrophotometry, after complexation with pyrocatechol violet (SCX-PCV method). The SPE/SCX-ICPAES results for the labile Al fraction (Al bound to the SCX cartridge) showed an acceptable correlation with the results obtained by the equilibrium calculations, except for the samples with the highest DOC concentrations, whereas the values obtained for labile Al by the more traditional SCXPCV method were much lower. We recommend that the SPE/SCX-ICP-AES procedure described in this work be selected for the fractionation of Al and Fe species in soil and freshwater samples.

Introduction The atmospheric deposition of inorganic strong acids may lead to the mobilization of Al in natural freshwater and soil water in acid-sensitive regions. This has been the cause of considerable concern (1), because dissolved Al species can be harmful to fish and plants (2, 3). However, not all Al species are equally toxic; therefore, analytical methods that can * Corresponding author phone: +47 22 85 55 34; fax: +47 22 85 54 41; e-mail: [email protected]. † University of Oslo. ‡ Norwegian Forest Research Institute (SKOGFORSK). § Norwegian Institute for Nature Research. 10.1021/es020077i CCC: $22.00 Published on Web 11/13/2002

 2002 American Chemical Society

discriminate between the different forms of the metal have been developed. A large number of such methods are described in the literature; most of them are based on the fractionation of Al according to charge and/or reactivity. The methods published up to 1994 have been reviewed by Clarke et al. (4), and more recent articles on the same subject are listed by Wickstrøm et al. (5). In the fractionation methods that are based on a separation according to charge, the “free/labile” Al species which are positively charged, are usually trapped on a cation exchange column, while Al-species that are neutral or negatively charged, such as Al bound to humic substances, pass through the column. A SCX resin such as Amberlite IR-120 has often been used for this fractionation, in accordance with the method originally developed by Driscoll (6). The quantitative determination of Al has often been done by molecular spectrophotometry, e.g. after complexation with pyrocatechol violet (PCV) (7), but atomic absorption spectrometry (8) and inductively coupled plasma atomic emission spectrometry (ICP-AES) (5, 9) have also been used. In most of the papers where ion exchange has been used for the fractionation of Al, the column was packed manually with a strong cation-exchange resin of a suitable particle size (often 0.4-0.6 mm), using rather large bed volumes. Today, the use of a commercially available solid phase extraction (SPE) cartridge seems to be a promising alternative, in particular for laboratories primarily concerned with routine analysis. These cartridges have so far mainly been used for chemical isolation and purification of organic compounds, but the SPE approach has several advantages also for metal speciation studies: It is simple, the SPE cartridge is only used once, it facilitates fractionation in the field with short delay times from sampling to conservation, and filtration is easily combined with the SPE separation. Results obtained for Al using SPE cartridges were included in a recent comparison study by Wickstrøm et al. (5). In the present work, the use of different types of SPE cartridges for the fractionation of Al and Fe species was studied, focusing on the type of sorbent material and the conditioning procedure required. The metal fraction that passed through the cartridge and the total metal concentrations were determined by ICP-AES. The free/labile metal fraction, which is probably the most toxic part, was calculated as the difference between the total metal in the sample and the Al concentration after passage through the cartridge. The sample was filtered through a 0.45 µm filter before fractionation, and also before measurement of total metal, hence the particulate metal fraction was not determined in this work. The results obtained for the Al fractionation by the SPE/ SCX-ICP-AES method for a number of soil water samples were compared with those obtained by an alternative fractionation method which utilized strong cation exchange on a manually prepared column and detection by molecular spectrophotometry after complexation with pyrochatecol violet. The results were also compared with equilibrium calculations using the program ALCHEMI, which was based on the stability constants of metal complexes with welldefined ligands. The program also included metal complexes with natural organic matter.

Experimental Section Instrumentation and Equipment. Three ICP-AES instruments were used: a Perkin-Elmer 5500B, a Thermo Jarrell Ash IRIS/AP, and a Spectro Spectroflame SND-07 with a microconcentric nebulizer. The instrument settings are given VOL. 36, NO. 24, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5421

TABLE 1. Operating Conditions for the ICP-AES Instruments Thermo Perkin-Elmer Jarrel Ash Spectro Rf power, W plasma gas flow rate, L min-1 auxiliary gas flow rate, mL min-1 nebulizer gas flow rate, mL min-1 wavelength for Al, nm wavelength for Fe, nm

850 14.0 700 1050 396.15 238.20

1150 1200 14.0 14.0 500 500 700 700 396.15 167.08 238.20 259.94

in Table 1. A peristaltic pump was used for pumping the samples through the SPE cartridges. For the flow injection analysis, an Alpkem Superflow 3590 was used. A Dohrman Carbon Analyzer DC-190 was used for the determination of the dissolved organic carbon (DOC) concentration in the soil water samples. Cartridges. The types of SPE ion exchange cartridges that were used are shown in Table 2. Strong cation exchange, strong and weak anion exchange and chelating resin were studied. The SPE cartridges were obtained from IST, Varian, Supelco, and Alltech. All the cartridges contained 500 mg of adsorbent and were silica-based, except the chelating resin from Alltech, which was based on styrene-divinylbenzene. The counterion provided by the manufacturer is shown in Table 2. Conditioning of the cartridge, i.e., adjustment of the counterion, was found to be essential, to avoid a pH change in the sample solution. Based on the results obtained in the present work, the following procedure was used: 14 mL of water were pumped at a flow rate of 5 mL min-1 through the cartridge, followed by 14 mL of 0.4 M ammonium acetate buffer of pH 4.5, and finally 4 mL of sample to rinse out the buffer. Reagents and Samples. All reagents were of analyticalreagent grade, and distilled, deionized water was used throughout. The standard metal solutions were 1000 mg L-1 Al3+ (in 2.5% HCl + 0.3% HNO3) and Fe3+ (in 2.5% HNO3) from Spectrascan (Norway). Twenty-three soil water samples (taken at 5, 15, and 40 cm depths) were collected from monitoring plots in Norway. The samples were stored in polypropylene flasks in a refrigerator (4 °C). The pH of the samples was in the region 3.9-5.5 (mean value 4.8). SPE-ICP-AES Procedure. The sample was filtered through a 0.45 µm (cellulose acetate, 13 mm) prefilter (Advantec MFS) before it passed through the SPE cartridge. The sample was pumped through the prefilter and cartridge at a flow rate of 5 mL min-1; 10 mL was collected and analyzed by ICP-AES, using the operating conditions specified in Table 1. The total metal concentrations in the soil water samples were also determined by ICP-AES, after filtration through a 0.45 µm filter. The labile metal fraction was calculated as the difference between total metal and the metal concentration after passage through the cartridge. SCX-PCV Procedure. The Al concentrations were determined by molecular spectrophotometry in a flow injection system (Alpkem Superflow 3590), based on the complexation of Al with pyrocatechol violet (PCV) and detection at 580 nm (7, 10). The carrier solution was 3.2 ‚ 10-5 M HCl (pH 4.5), the concentration of PCV was 1.57 ‚ 10-4 M, and 0.44 M hexamethylenetetramine served as buffer. 1.25 ‚ 10-3 M 1,10phenanthroline and 0.36 M hydroxylammonium chloride were used as masking agents for Fe. Total “monomeric” Al was determined by direct analysis of the sample, and organically bound “monomeric” aluminum was determined by analysis of the sample after passage through a manually prepared strong cation exchange column. The labile Al fraction was calculated as the difference between these two values. The pH adjusted (pH 5.5) strong cation exchange column (Amberlite IR-120; 0.3-1.1 mm (52-14 5422

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 24, 2002

mesh)) was prepared in a 150 mm long, 3 mm diameter glass tube (Omnifit 416322). Equilibrium Calculations using ALCHEMI. The concentration of labile Al in the soil water samples was estimated by equilibrium calculations using the program ALCHEMI (14). The program calculates the sum of the labile species, i.e., Al3+ and aluminum complexes with OH-, SO42-, Si(OH)4, and fluoride, based on the total ionic concentrations and the stability constants of the individual complexes. ALCHEMI also takes into consideration the presence of natural organic matter. The program includes two models for the organic matter: the organic matter is considered to resemble either a diprotic or a triprotic acid (both assumptions may represent an oversimplification).

Results and Discussion Conditioning of the Cartridge. The manufacturers normally recommend that the cartridge be conditioned with an organic solvent such as methanol (1 mL/100 mg sorbent), followed by a 0.02-0.05 M buffer solution with a pH close to that of the sample (1 mL buffer/100 mg sorbent). However, this conditioning procedure was found to cause a change in the pH of the soil water sample when it passed through the cartridge and also a change in the Al-speciation. Therefore, different conditioning procedures were tried out, by varying the amount of buffer, water, and methanol used. The concentration of Al and pH of a soil water sample after passage through the SPE/SCX cartridge are given in Figure 1 for different conditioning procedures. In addition to the conditioning which is shown in the figure, the cartridge was rinsed with 4 mL of the sample. It can be seen from Figure 1 that the concentration of Al may change from 1.3 mg L-1 to 0.6 mg L-1, depending on the conditioning procedure, and pH changes from 4.6 to 3.3. A low pH value is accompanied by a decrease in the amount of Al that passes the cartridge, indicating that organically bound Al dissociates when the pH is lowered. However, it can be seen from the figure that there is no simple relationship between the concentration of Al species and pH; the reagents used for the conditioning are also important. Conditioning the cartridge with only water results in a marked decrease in both pH and the amount of organically bound Al that passes the column; hence the inorganic Al fraction will be overestimated in this case. The treatment with 14 mL of water followed by 14 mL of 0.4 M ammonium acetate buffer (pH 4.5) and finally 4 mL of sample to rinse out the buffer was selected as the best approach in the present work. With this treatment, the change in pH was minimized, the dissociation of organically bound Al was avoided, and the concentration of Al was determined with good precision. This conditioning procedure also limits the amount of buffer used and avoids the use of methanol. In the present work, the flow rate through the cartridge was 5 mL min-1. The ability of the cartridge to bind Al from soil water samples was found to be unaffected by the flow rate in the range 1-5 mL min-1, so the highest practical flow rate (5 mL min-1) was used. The absence of a flow-rate effect indicates that the metal which is bound to natural organic matter does not dissociate during passage through the cartridge (hence the term “nonlabile” metal fraction seems to be justified). Control Samples. The performance of the SPE cartridges was checked using control samples which contained Al or Fe in 0.02 M acetate buffer (pH 4.5). In addition, the samples contained EDTA or citrate. The results for the SCX and chelate cartridges are shown in Table 3. It can be seen from the table that the cation exchange and chelate cartridges behaved as expected; the Al and Fe complexes with EDTA and citrate, which are negatively charged at pH 4.5, pass through the cartridge, whereas the positive “free” metal ions are bound to the cartridge.

TABLE 2. Types of Solid Phase Extraction (SPE) Cartridges Studied sorbent description

mode

type/manufacturer

SCX, benzenesulfonic acid (H+) SCX, benzenesulfonic acid (H+) PRS, propylsulfonic acid (H+) SAX, quaternary amine (Cl-) NH2, aminopropyl SAX, quaternary amine (Cl-) Chelate, iminodiacetate (Na+)

strong cation exchange strong cation exchange strong cation exchange strong anion exchange weak anion exchange strong anion exchange chelating resin

Bond Elut/Varian ISOLUTE/IST ISOLUTE/IST ISOLUTE/IST ISOLUTE/IST Supelco Maxi-Clean IC/Alltech

FIGURE 1. Effect of conditioning of a SCX cartridge for the fractionation of Al in soil water. For each conditioning procedure, the pH and the concentration of Al were measured after the sample had passed through the cartridge. The soil water had a pH of 4.55 and a total Al concentration of 1.57 mg L-1. Each result given in the figure is the mean of three measurements.

TABLE 3. Fractionation Using SCX and Chelate Cartridges; Al and Fe in Control Samples complexing agent type

concn (µM)

none 0.0 EDTA 9.27 EDTA 100 citric acid 4.63 none 0.0 EDTA 4.48 EDTA 100

metal concn not bound to cartridge (µM)

metal Al Al Al Al Fe Fe Fe

total concn experiment experiment (µM) SCX chelate expected 18.5 18.5 18.5 9.27 8.95 8.95 8.95

0.0 9.56 17.7 5.01 0.0 4.4 9.0

0.0 9.41 18.2 0.0 4.4 9.0

0.0 9.27 18.5 4.63 0.0 4.48 8.95

TABLE 4. Fractionation Using SAX and NH2 Cartridges; Al and Fe in Control Samples metal concentration not bound to cartridge (µM) EDTA total metal experiment experiment concn (µM) metal concn (µM) SAX (IST) NH2 (IST) expected 0 100 0 4.48 100

Al Al Fe Fe Fe

18.5 18.5 8.95 8.95 8.95

11.9 0.0 8.92 4.64 0.0

11.5 0.0 8.90 4.17 0.0

18.5 0.0 8.95 4.48 0.0

The results for two anion exchange cartridges are given in Table 4. It can be seen from the table that the negatively charged EDTA complexes of Al and Fe are bound to the cartridge, as expected. In the absence of EDTA, all the Fe pass through the cartridges, but 36-38% of the Al are retained. An anion-exchange resin may have secondary cation exchange properties (11), but in the present case this appears only to affect the results for Al. It can be seen from Tables 3 and 4 that for Fe in the presence of EDTA, the sum of the

metal that passes through the SCX and SAX, respectively, is nearly 100%. Soil Water Samples. For a sample containing both cationic and anionic species of a given metal (and no uncharged species), the sum of the concentrations passing a cation and an anion exchange cartridge, respectively, should equal the total concentration of the metal. The following results were obtained for a soil water sample with a total concentration of 0.640 mg L-1 Fe. After passage through the SCX cartridge, 0.104 ( 0.003 mg L-1 Fe; after passage through the anion exchange (NH2) cartridge, 0.501 ( 0.002 mg L-1 Fe. The sum of these results is 0.605 mg L-1 Fe, so that 94% of the total Fe in the sample are accounted for. This indicates that both cation and anion exchange may be used for the fractionation of Fe in soil water. However, when the same experiment was carried out for Al, the sum of the concentrations passing the two cartridges was only 70% of the total concentration of Al in the sample. Some inorganic Al is probably retained by the anion-exchange resin, as was also the case for the control samples (see above). Therefore, cation exchange seems to be more suitable for fractionation of Al than anion exchange. Cation exchange and chelate cartridges from different manufacturers were compared by fractionating four soil water samples; the results for Al are given in Table 5. The results show that there is no significant difference (t-test, P)0.95) in the amount of Al that passes through the four cartridges, except for sample 3 through the chelate cartridge. The chelate and the SCX cartridges were compared more extensively by analyzing 18 soil water samples; the results for Al are given in Figure 2. It can be seen from the figure that for some of the samples, more Al passes through the chelate cartridge than through the SCX cartridge. However, the weighted (each point is given a weighting inversely proportional to the corresponding variance) linear regression line (P)0.95) is close to the y ) x line, as shown in Figure 2. Most of the samples above the y ) x line had a pH below 5.0, so the presence of positively charged polymeric Al species (12, 13) VOL. 36, NO. 24, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5423

TABLE 5. Comparison of Different Cation Exchange and Chelate SPE Cartridges, for the Fractionation of Al in Soil Watera

TABLE 6. Results for 23 Soil Water Samples, Obtained by Equilibrium Calculations using the ALCHEMI Program and Fractionation by the SPE/SCX-ICP-AES and SCX-PCV Methods

concentration of Al not bound to the cartridge (mg L-1) ((1 SD, n)3) cartridge type

sample 1

sample 2

sample 3

sample 4

SCX, IST 1.14 ( 0.08 0.60 ( 0.09 1.31 ( 0.06 PRS, IST 1.20 ( 0.06 0.60 ( 0.05 1.36 ( 0.07 SCX, Varian 1.19 ( 0.08 0.56 ( 0.05 1.27 ( 0.04 0.40 ( 0.03 Chelate, Alltech 1.3 ( 0.1 1.6 ( 0.1 0.44 ( 0.09 a

Three cartridges of each type were used.

FIGURE 2. Comparison of the SCX (Varian) and Chelate (Alltech) cartridges for the fractionation of Al. Results are given for 18 soil water samples. The solid line represents the weighted regression line, and the dotted line is y ) x.

FIGURE 3. Comparison of the SCX (Varian) and Chelate (Alltech) cartridges for the fractionation of Fe. Results are given for 16 soil water samples. The solid line represents the weighted regression line; it coincides with the line y ) x. is not a likely explanation of the difference between the chelate and the SCX results. The results for Fe for the same 18 soil water samples are shown in Figure 3. Here, no significant difference was observed for the two types of cartridges. Comparison with Equilibrium Calculations using ALCHEMI. In Table 6, the concentration of labile Al are given for 23 soil water samples analyzed by the SPE/SCX-ICP-AES method (labile Al is the difference between total Al and Al 5424

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 24, 2002

concentration of labile Al fraction (µM) DOC (mg L-1)

total Al concn (µM)

1.1 1.2 1.2 1.3 1.4 1.5 1.9 2.2 2.9 3.2 5.7 8.0 9.5 14 17 19 19 20 20 23 31 51 56

6.1 6.4 7.2 7.7 4.1 4.9 3.3 17 7.1 10 16 7.1 9.3 6.2 39 41 43 41 40 41 29 38 41

pH

ALCHEMI Diprotic model

ALCHEMI Triprotic model

SPE/SCXICP-AES

SCXPCV

5.0 5.2 5.2 5.1 5.3 5.3 5.4 4.6 4.8 4.7 4.7 5.1 4.9 5.5 4.5 4.5 4.5 4.4 4.4 4.4 4.6 3.9 3.9

5.5 5.5 6.0 6.5 3.5 4.0 2.6 15 5.9 8.4 12 4.0 5.3 2.1 25 26 27 27 27 27 13 31 36

5.4 5.4 6.0 6.4 3.4 4.0 2.6 14 5.4 7.6 9.9 3.8 4.7 2.2 18 18 19 19 18 17 8.7 15 17

5.7 6.0 6.6 7.3 3.8 4.6 2.6 16 6.7 8.6 16 2.6 3.9 5.9 17 18 19 17 18 16 6.7 6.1 6.8

3.3 1.7 0.11 0.89 0.44 0.07 13 7.0 14 0.11 5.9 6.5 7.0 5.7 2.0 1.6 1.6

after passage through the cartridge). The results are presented in the sequence of increasing concentration of DOC in the soil water samples. In the same table, the results are given for the equilibrium calculations which were based on the ALCHEMI program (see Experimental Section). Considering the limitations of the basic data for the organic matter in the ALCHEMI program, the agreement between the SPE/SCXICP-AES results for labile Al and those obtained by the equilibrium calculations are acceptable, particularly at low DOC values. It can be seen from Table 6 that the correlation between the SPE-ICP-AES values and the ALCHEMI values is best when the triprotic model is used. The correlation coefficient (r2) obtained for SPE-ICP-AES versus ALCHEMI was 0.71 for the triprotic model and 0.45 for the diprotic model. From Table 6 it can be seen that for the two highest DOC concentrations, the agreement between the calculations and the SPE-ICP-AES results is poor for both the diprotic and the triprotic models. This is not surprising, considering that these samples had a low pH (pH 3.9), and that the model was empirically fitted to the more common pH range around 5. Furthermore, these samples were from the forest floor organic horizon. The chemical characteristics of this organic material are probably different from the organic material found in surface waters (15), where ALCHEMI has mostly been used. Comparison with the SCX-PCV Method. Most of the samples in Table 6 were also analyzed by an alternative Alfractionation method, which was based on separation on a manually prepared SCX column and detection of Al by molecular spectrophotometry, after complexation with pyrocatechol violet (PCV) (7, 10) (see Experimental Section). The results of the analysis of 17 soil water samples by the SCX-PCV method are given in Table 6. It can be seen from the table that most of the results for labile Al obtained by the SCX-PCV method are much lower than those obtained by either the SPE-ICP-AES method or by equilibrium calculations with the ALCHEMI program. As expected, the correlation coefficient for SCX-PCV versus ALCHEMI was very low, 0.21 for the triprotic model and 0.06 for the diprotic model.

A reason for the low values obtained by the SCX-PCV method may be that the organic matter interferes with the Al-PCV reaction (16, 17). Wickstrøm et al. (5) have reported similar results; they found that the SCX-PCV method gave lower values for labile Al than the SPE-ICP-AES method, the difference increased when the DOC concentration increased, and the SPE-ICP-AES method was better correlated to ALCHEMI than to the SCX-PCV method. Although different fractionation procedures cannot be expected to give exactly the same results, because all approaches determine operationally defined fractions, deviations as large as those observed in the present work between the SPE-ICP-AES and the SCX-PCV methods are hardly acceptable. The present results indicate that the SCX-PCV method may underestimate the labile Al fraction in the sample. The SPE-ICP-AES method has several advantages compared to the more traditional methods of Al fractionation that are based on a manually prepared cation exchange column and molecular spectrophotometric detection. It uses commercial SPE cartridges that are easily available and easy to use, the cartridge is only used once, it facilitates fractionation in the field with short delay times from sampling to conservation, and filtration is easily combined with the SPE separation. ICP-AES is available in many laboratories; because it is an element-specific multielement technique, the fractionation of several elements can be studied simultaneously. The fractionation step and the determination step are carried out in two separate operations; the separation can be performed in the field, while the determination of Al and Fe by ICP-AES can be done later in the laboratory, after acid conservation of both the total sample and the metal fraction that has passed the SPE cartridge. In contrast, molecular spectrophotometry is time-consuming and prone to interferences; e.g. the PCV method involves six chemicals and three reaction coils. However, the methods based on molecular spectrophotometry may be used more easily to determine selectively the monomeric species of Al.

Acknowledgments We thank the personnel at the chemical laboratory of the Norwegian Forest Research Institute for technical assistance.

Literature Cited (1) The environmental chemistry of aluminum; Sposito, G.; Ed.; CRC Press: Boca Raton, FL, 1989. (2) Driscoll, C. T.; Schecher, W. D. In Metal ions in biological systems; vol 24, Aluminum and its role in biology; Sigel, H., Sigel, A., Eds.; Marcel Dekker: New York, 1988; pp 59-122. (3) Taylor, G. J. In Metal ions in biological systems; vol 24, Aluminum and its role in biology; Sigel, H., Sigel, A., Eds,; Marcel Dekker: New York, 1988; pp 123-163. (4) Clarke, N.; Danielsson, L. G.; Sparen, A. Pure Appl. Chem. 1996, 68, 1597. (5) Wickstrøm, T.; Clarke, N.; Derome, K.; Derome, J.; Røgeberg, E. J. Environ. Monit. 2000, 2, 171. (6) Driscoll, C. T. Int. J. Environ. Anal. Chem. 1984, 16, 267. (7) Seip, H. M.; Mu ¨ ller, L.; Naas, A. Water, Air, Soil Pollut. 1984, 23, 81. (8) Barnes, R. B. Chem. Geol. 1975, 15, 177. (9) Sullivan, T. J.; Seip, H. M.; Muniz, I. P. Int. J. Environ. Anal. Chem. 1986, 26, 61. (10) Dougan, W. K.; Wilson, A. L. Analyst 1974, 99, 413. (11) Van Horne, K. C. Sorbent Extraction Technology Handbook; Varian: Harbor City, CA, 1985. (12) Baes, C. F.; Mesmer, R. E. The hydrolysis of cations; Wiley: New York, 1976. (13) Bertsch, P. M. In The environmental chemistry of aluminum; Sposito, G., Ed.; CRC Press: Boca Raton, FL, 1989. (14) Schecher, W.; Driscoll, C. T. Water Resour. Res. 1987, 23, 525. (15) Vogt, R. D.; Andersen, D. O.; Bishop, K.; Clarke, N.; Gadmar, T.; Gjessing, E.; Kitunen, V.; Laudon, H.; Lund, W.; Lundstrøm, U.; Mulder, J.; Starr, M.; Tervahauta, A.; Van Hees, P.; Østerhus, B. Natural organic matter in the Nordic Countries (NOMiNiC). NT Techn. Report 479; NordTest: Espo, Finland, 2001; available at http://www.nordtest.org. (16) Røyset, O. Anal. Chim. Acta 1986, 185, 75. (17) Røyset, O.; Sullivan, T. J. Intern. J. Environ. Anal. Chem. 1986, 27, 305.

Received for review April 15, 2002. Revised manuscript received September 30, 2002. Accepted October 7, 2002. ES020077I

VOL. 36, NO. 24, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5425