Outside-In Trimming of Humic Substances During Ozonation in a

This paper addresses the change of molecular size distribution of humic substances (HS) during ozonation in a membrane contactor. It focuses on the ...
0 downloads 0 Views 145KB Size
Download PDF
Environ. Sci. Technol. 2006, 40, 6460-6465

Outside-In Trimming of Humic Substances During Ozonation in a Membrane Contactor R. H. S. JANSEN, A. ZWIJNENBURG, W. G. J. VAN DER MEER, AND M. WESSLING* Faculty of Science and Technology, Membrane Technology Group, University of Twente, P.O. box 217, 7500AE Enschede, The Netherlands

This paper addresses the change of molecular size distribution of humic substances (HS) during ozonation in a membrane contactor. It focuses on the characterization and identification of some small ozonation products. The membrane contactor setup allows very precise control of ozone transfer into the solution as well as precise sampling of the products in time. The molecular size distribution was followed by gel permeation chromatography (GPC). Characterization and identification of small ozonation products was performed by membrane nanofiltration and high-performance liquid chromatography (HPLC). Measurements on molecular size distribution indicate that during the ozonation process the size of the HS molecules decreases slowly and only small highly oxidated compounds are being split off the larger molecules. Pyruvic acid, formic acid, methylglyoxal, and acetaldehyde could be identified by substantial peaks. Glyoxilic acid and glyoxal were identified to a lesser extent. This suggests that HS molecules consist of a relatively stable backbone network structure and that the HS molecule degrades according to an outside-in trimming mechanism.

Introduction Ozonation of humic substances is a subject of major interest in drinking water treatment. In some cases ozonation is followed by biodegradation to prevent biofouling of the process systems (1-6). The efficacy of a plain biofiltration process to accomplish total removal of humic substances is hampered by the low biodegradability of humic substances. This is generally explained by the size and complex composition of the humic molecules. In a combination of ozonation and subsequent biofiltration one tries to degrade the humic substances into smaller and less complex compounds which may than be biodegradable (5, 6) or removed by adsorptive processes (7). The biodegradability of ozonated humic substances depends on the products that have been formed after ozonation. In this, the effect of ozonation on the biodegradability can be different for humic substances from different origins (4, 8, 9). The enhanced biodegradability lies in the conversion of the large complex molecular structures into smaller less complex molecules. It is therefore of interest to investigate the change in molecular size distribution of the compounds in the solution during the ozonation process. Even more interesting is the identification of the ozonation products. * Corresponding author phone: +31 53 4892950; fax: +31 53 489 4611; e-mail: [email protected]. 6460

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 20, 2006

Humic substance molecules can have sizes that range from 103 to 105 Da (10, 11). For aquatic humic substances molecular weights of 103 to 104 Da are common (12). The carbon structure may consist of aliphatic chains, (poly-) aromatic groups, and furanic groups. These contain various organic functional groups. Ester and ether bonds may be found throughout the structure as well as carboxyl, ketone, aldehyde, and alcohol groups. Through this variety in humic structures it is evident that the products of oxidation also show a wide variety in size and structure. One can make some estimation of ozonation products with the help of the information that has been gathered over the years of reactions of known organic compounds and functional groups with ozone. In a review article Von Gunten (13) discusses the possible products stemming from the ozonation of drinking water for both the direct and indirect ozonation reaction. The discussion addresses organic compounds and several functional groups that can be found in drinking water. Generally, ozonation products can contain functional groups like alcoholic, carbonyl, and carboxyl groups. Also esters can be formed in the ozonation process. Splitting of bonds yields smaller compounds and complete oxidation yields water and carbon dioxide. Incomplete ozonation of the colored humic substances results in decoloration and a certain DOC loss that depends on the amount of ozone that has been utilized. Characterization of reaction products from ozonation of humic substances shows the formation of low molecular weight compounds consisting of aldehydes, ketones, ketoacids, and carboxylic acids (14-17). Some products identified so far are pyruvic acid, glyoxylic acid, mesoxalic acid, propanal, glyoxal, methylglyoxal, and acetaldehyde (2, 14). Quantity and type of potentially reactive sites, each with its own reactivity, vary within a HS. This has consequences for the overall reactivity of the humic substance as well as for the products evolving in the course of the reaction. In batch-wise experiments the reaction kinetics, and hence the products, may change during the ozonation period making a controlled, precise evaluation of the reaction products a difficult task. For this purpose an experimental setup which can be operated stationary is highly desirable. Membrane contactors have advanced in recent years as an effective means for mass transport across the gas-liquid interface. The membrane does not have a separation function: it is macroporous and allows rapid diffusion of gases and vapors through the porosity. In aqueous application, the membrane must be hydrophobic to prevent wetting of the pores. The advantages of a membrane contactor lie in the (potentially) high interfacial area and the hydrodynamic decoupling of the phases. Also in bubble-free transport of a gas to a liquid phase membrane contactors can be of much use. Due to the low solubility of ozone in water a high contact area would be very beneficial to enhance the ozone transfer from the gas to the liquid phase. Jansen et al. (18) presented the methodology to use a membrane contactor to quantify the reaction kinetics of HS during ozonation. In the work presented here, we use the well-controlled ozonation conditions of a membrane contactor to identify the products of the ozonation process. Gel permeation chromatography (GPC) is a useful technique to follow the molecular size distribution. The separation is based on size exclusion and preferably as little as possible on affinity with the stationary phase. However, one can never avoid any chemical interaction with either phase. GPC has been practiced on humic substances in several studies (19-22). 10.1021/es060533t CCC: $33.50

 2006 American Chemical Society Published on Web 09/12/2006

Two main problems exist in the determination of the molecular size distribution of humic substances. The first is lack of reference compounds. There are no reference compounds available with a known molecular weight that can act as model compounds for humic substances concerning the chemical interactions (12). Instead, calibration of the GPC column is usually done by polymer standards. By choosing a number of different compounds as reference a good estimation of the separation curve may be possible. The second problem is the effect of the water conditions on the conformation of the molecules (size and shape). Ionic strength and pH can affect the conformation of the molecules severely. The result is that the molecular size distribution is never an absolute value but merely an apparent molecular size distribution (19). Despite these problems GPC may be used to follow at least qualitatively the (apparent) molecular size distribution of humic substances during a treatment. Nanofiltration is a pressure-driven membrane filtration process in which the membrane allows the passage of small organic molecules and monovalent ions. Larger organic molecules are rejected mainly based on molecular sieving, but sometimes also based on charge and hydrophobicity. Bivalent ions are frequently rejected based on charge (Donnan) exclusion. We will use nanofiltration to fractionate ozonation products. The research focuses on investigation of the products that result from ozonation of humic substances under wellcontrolled ozonation and sampling conditions. With GPC and fractionation by membrane nanofiltration the molecular size (distribution) of the humic substances and ozonation products during the ozonation process was studied. HPLC analysis was used to identify some of the small ozonation products. From the results of the molecular size measurements, product identification, and measurements of some water quality parameters like DOC, color, and COD, we suggest the mechanism of ozonation to proceed according to an outside-in trimming process.

Materials and Methods Humic substances were obtained from tap water from pumping station Spannenburg (The Netherlands). The HS molecules were concentrated by an anion exchange adsorption process, where Purolite A806S was used as ion-exchange resin on which the HS molecules adsorb. According to the manufacturer’s instruction the column was regenerated by a 10% NaCl solution in water. The HS concentration obtained in the regenerate reached up to 30 g DOC/L. The concentrated HS solution was diluted to a solution with 20 g DOC/L to form the stock solution. The NaCl concentration in the stock solution was kept below 1 M. According to Hoigne et al. (23) chloride ions do not react significantly with ozone and it is assumed that chloride ions do not take part in the reactions. HS solutions for the experiments were prepared out of the stock solution. The amount of HS in the experimental solution was 3.3 mol DOC/m3 (40 mg DOC/L). In these dilutions the NaCl concentration in the HS solutions is 0.002 M. At these low concentrations the ionic strength of the reaction solutions is considered to have no significant effect on the ozone solubility in water. Experimental Setup. The ozonation experiments were performed in a setup as presented in the Supporting Information (Figure S5). A membrane module with hollow fiber membranes was used as gas-liquid contactor for the ozonation experiments. The module is constructed with PVDF capillary membranes in a glass housing. The membranes were supplied by EMI University of Twente, The Netherlands. For detailed information on the membrane module refer to ref 18. Air was led through an ozone generator (Ozone Generator Peripheral Com, Anseros Klaus Nonnenbacher GmbH). The

ozone-enriched air entered the membrane contactor module on the shell side of the membranes. The ozone concentration both in the gas feed and outlet was measured spectrophotometrically at a wavelength (λ) of 253 nm with a Pharmacia LKB Ultrospec Plus double beam spectrophotometer. During the experiments the applied ozone concentration was 0.4 mol/m3 at a flow of 20 L/h. The liquid flow was 1 L/h. The mass transfer processes were performed countercurrently with the liquid flowing through the fibers. The ozone consumption was calculated from the difference in ozone concentration between the gas feed and the gas outlet. The benefit of using a membrane contactor in this process is that the operating conditions can be controlled precisely. The phases can flow through the reactor independent of each other and the gas/liquid adsorption process can be operated stationary so that the chosen experimental conditions remain constant throughout the experiments. This is essential for accurately measuring the ozone consumption during the oxidation process. HS Characterization. The HS solutions were characterized by DOC (mol DOC/m3 or mg DOC/L), by GPC analysis and HPLC analysis. The DOC was measured on a Shimadzu TOC-V CPH total organic carbon analyzer. The GPC analysis was performed on a Metrohm HPLC 761 compac IC setup with a PSS MCX-1000 Å GPC column. A Bisschoff Lambda 1010 was used as UV-vis detector. The eluent consisted of an aqueous solution of 0.1 M Na3PO4 running at 1 mL/min. The organic acids were also measured on a Metrohm HPLC 761 compac IC setup with a Metrosep Organic acids 6.1005.200 column, a suppressor with 50 mM LiCl, and a Metrohm 1006 conductivity detector. The eluent was 0.5 mM H2SO4 running at 0.7 mL/min. GPC and HPLC Calibration. Several organic compounds were used as reference standards and for calibration of the GPC system. These compounds were several types of polymers with different molecular weights. Further, low molecular weight acidic and nonacidic organic compounds were used, as summarized in the Supporting Information (Table S1 and Figure S2). Also for reference purposes, several small ketones and carboxylic (keto-) acids were analyzed on the HPLC IC column (Figure S3). Detection Wavelength of the UV-vis Detector for GPC. In research for water treatment usually the UV absorption is measured at a wavelength of 253.7 nm (254 nm rounded off). However, during the ozonation process products are formed that have an absorption over a smaller spectrum ranging from 200 to 240 nm. All observed peaks during the analyses of humic substances and ozonated humic substance solutions have their maximum at a wavelength of 223 nm. For that reason this wavelength was chosen for detection in the GPC analysis procedure. Details are available in the Supporting Information (Figure S1). Experiments for Sequential Ozonation of a Humic Substance Solution. A HS solution with a concentration of 3.3 mol DOC/m3 (40 mg DOC/L) was subjected to sequential ozonation, which means that the solution was run through the ozonation module several times. The liquid flow rate was 1.0 L/h (0.28 mL/s) and the gas flow rate was 14.9 L/h (4.14 mL/s) at an ozone concentration of 0.378 mol/m3. To investigate the molecular size of the ozonation products a part of the ozonated HS solution was fractionated by nanofiltration after each run. The fractionation was performed in a stirred dead-end cell with a SpectraPor type C cellulose ester membrane with a MWCO of 500 Da (given by manufacturer). Solutions were filtered until half the feed volume had permeated. For reference purposes a solution of glucose and sucrose was also filtered by the membrane. The filtration test with lactose (Mw ) 342.3 Da) showed that 93% of the lactose was retained. The same test with a glucose (Mw )180) solution showed that 50% of the feed concentration VOL. 40, NO. 20, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6461

FIGURE 1. GP chromatogram of humic substances in several dilutions. According to the GPC calibration line (Figure S2) the molecular weight ranges from approximately 1 × 104 to 0.5 × 102 Da. The figures at the curves indicate the DOC content of the solutions in mol DOC/m3. was found in the permeate. GPC and HPLC (IC) analyses were performed to characterize the samples or identify specific compounds.

Results and Discussion GPC Analysis of Humic Substances. Analysis of concentrated humic substances shows chromatograms that are in good proportion with each other regarding the DOC content of the samples as shown in Figure 1. Based on the calibration plot (shown in the Supporting Information, Figure S2), the molecular weights of the humic substances vary from approximately 10 000 down to 500 Da. The weight average molecular weight Mw is calculated as 2.6 × 103 Da and the number average molecular weight Mn is calculated as 2.1 × 103 Da. With these figures the polydispersity is 1.23. These weight averages have not been corrected for the skewing and peak broadening caused by the GPC system. Accounting for skewing and peak broadening, the corrected molecular weights would be Mw ) 2.5 × 103 to 2.9 × 103 Da and Mn ) 2.2 × 103 to 2.5 × 103 Da. The polydispersity would be between 1.15 and 1.16, which would imply that the molecular weight distribution of these humic substances is relatively narrow. DOC, Color, and COD of a HS Solution during Ozonation. During the ozonation process a strong color decrease of the reaction solution can be observed visually. This is also expressed in the specific color (color/DOC) of the solution, which decreases considerably during the ozonation process resulting in a low color/DOC ratio (Figure 2). Meanwhile, the DOC content of the solution is lowering slowly. Apparently volatile compounds that may have been formed during the ozonation process are leaving the reaction solution. Together with the DOC, the chemical oxygen demand (COD) of the solution is also slowly decreasing, however, the COD/DOC ratio is reducing as the solution undergoes the ozone treatment. This is an indication that the humic substances are being oxidized. Figure 3 shows that the pH of the solution drops rapidly from 7.75 to 3.2. This indicates the formation of acidic groups during the ozonation process. These groups could either be formed on the HS molecules or be part of (acidic) compounds that are split from HS molecules as a result of the oxidation. GPC and HPLC Analysis of HS during Ozone Oxidation: Course of the Molecular Size of Large Compounds. During the ozonation of humic substances one might expect a wide range of ozonation products of different size and structure, because of the diversity of the humic substance content. 6462

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 20, 2006

FIGURE 2. (a) Course of the DOC content, color, and COD of a HS solution during sequential ozonation. (b) Normalized DOC content, specific color, and COD of the HS solution during sequential ozonation. Initial HS concentration 40 mg DOC/L (3.3 mol DOC/m3). The parameters are plotted against the specific ozone consumption. An ozone consumption of 0 mol O3/mol DOC indicates an untreated HS solution.

FIGURE 3. pH of a HS solution as function of the specific ozone consumption during sequential ozonation. HS concentration: 3.3 mol DOC/m3. Therefore it might be expected that the large peak of the untreated humic substance would become smaller and be “smeared out” toward higher elution times, indicating the gradual formation of small compounds. Figure 4 presents the GPC results of samples taken during the sequential ozonation of a HS solution. The large peak of the original humic substance at around 8 min is clearly decreasing and a shift of the peak can be observed. After an ozone

FIGURE 4. GP chromatogram of a humic substance solution after sequential ozonation in a membrane contactor. The solution is repeatedly sent through the membrane contactor. Initial HS concentration: 3.3 mol DOC/m3 (40 mg DOC/L). The figures indicate the specific ozone consumption (mol O3/mol DOC).

FIGURE 5. Course of the calculated average Mw (weight average molecular weight) of the large compounds of a HS solution (represented by the left peak in Figure 4) against the specific ozone consumption. HS concentration: 3.3 mol DOC/m3 (40 mg DOC/L). consumption of 0.65 mol O3/mol C the weight average molecular weight Mw has appeared to shift from 2600 to 1500 Da as shown in Figure 5. Simultaneously, we observed a remarkable decrease of the apparent polydispersity during the ozonation process: from 1.23 to 1.09. The spreading in molecular weight of the larger molecules seems to become smaller during ozonation. As the ozonation sequence goes on, another peak emerges around 9 min elution time indicating that small compounds are being formed during the ozonation process. The apparent Mw is calculated to be around 430 Da with a polydispersity Dp of 1.07. This low polydispersity indicates that products being formed have a narrow molecular size distribution. However, since it is likely that these ozonation products contain acidic groups, and considering the outcome of the GPC analysis of the organic compounds, the molecular sizes given above can only be taken as quantitative indication rather than an absolute number. It is furthermore remarkable to observe that the size of the ozonation products seems to be restricted to only two distinct peaks throughout the sequence of ozonation steps. This observation applies to the whole wavelength range between 200 and 300 nm of the detector. It indicates that during ozonation only small molecular fragments are split off from the large HS molecules: the main structure reduces size slowly during ozonation, but remains intact. This may

FIGURE 6. (a) GP chromatograms of a HS solution (3.3 mol DOC/m3) after ozonation as function of the specific ozone consumption. (b) GP chromatograms of the permeate of the same samples after membrane fractionation by a NF membrane with MWCO 500 Da. The figures near the graphs indicate the specific ozone consumption (mol O3/mol DOC). indicate that the HS molecules consist of a relatively stable network structure instead of a chain-like structure. The DOC content of the ozonated solutions decreases considerably as is shown in Figure 2. This indicates that compounds that may have been formed during the ozonation process leave the reaction solution. These could be volatile low molecular weight organic compounds among which might be carbon dioxide. Batch experiments where the carbon dioxide concentration was measured in the offgas of the reactor revealed that a substantial amount of carbon dioxide is indeed formed during the ozonation process. The amount of carbon dioxide that was measured appears to be comparable with the measured DOC loss. Unfortunately, it is impossible to identify to what extent CO2 is formed by further oxidation of small compounds that were already a product of oxidation of HS or by direct oxidation of the large HS molecules. Molecular Size Characterization and Identification of Small Ozonation Products. Nanofiltration was used to fractionate the ozonated HS solutions to investigate the molecular size of the small ozonation products that are eluted around 9 min. With nanofiltration the small compounds could be separated from the larger compounds into the filtrate (permeate). The chromatogram of the permeate (Figure 6a) shows a predominant peak at around 9 min indicating the presence of small compounds and the absence of larger compounds. The areas of the peaks in all the permeate samples have the same size as the areas of the corresponding peaks in the unfractionated solutions. This indicates that all VOL. 40, NO. 20, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6463

humic substances. Cleavage occurs mainly on the periphery of the molecules. One may argue that steric obstruction related to the coiled structure of the humic substance molecules prevents ozone (or the OH radical) from cleaving bonds in the core section of the molecules. However, the color decrease during ozonation and assuming that the colorcausing sites are spread throughout a HS molecule indicates that the oxidation reactions occur generally all over the molecule. We hypothesize that the backbone structure in the core is relatively stable due to a large number of double bonds keeping the overall molecular structure together even though some bonds are broken randomly in the molecule. Bonds in the peripheral shell have a higher chance of being trimmed off a HS molecule. This process of eroding of the molecules on the periphery can be considered as an “outsidein trimming mechanism”. FIGURE 7. Identification of ozonation products of a HS solution (3.3 mol DOC/m3) by HPLC. The ozone consumption was 0.61 mol O3/mol DOC.

Supporting Information Available Figures, tables, and text providing additional experimental details. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited

FIGURE 8. Represenation of the outside-in trimming process during the ozonation of humic substances. compounds associated with the peak at 9.0 min have permeated. Since the molecular weight cut off of the membrane appears to be approximately around the molecular weight of lactose (338 Da) it can be assumed that the compounds are much smaller in size than lactose. However, as glucose was also partly retained by the membrane it is plausible that the molecular sizes of these compounds are even smaller than that of glucose (180 Da). With the help of the reference diagram (Figure S3) a number of peaks in HPLC diagram could be identified in Figure 7. Mesoxalic acid, pyruvic acid, methylglyoxal, and acetaldehyde could be identified by substantial peaks. Glyoxilic acid and formic acid were identified by less distinct peaks. The peak at 14.4 min was suspected to correspond to carbonate. This peak is found in most water samples and is due to dissolved carbon dioxide present in the water. The largest molecular size of these compounds is 88 Da (pyruvic acid). It is most likely that not all small compounds that result from ozonation have been identified. However, considering the membrane filtration tests it may be justified to assume that all small compounds have a molecular size smaller than 180 Da as stated previously. Regarding the course of the molecular weight and polydispersity of the large compounds during ozonation and the fact that size of the small ozonation compounds is assumed to be smaller than that of glucose (MW ) 180) it appears that ozonation of humic substances yields products in two size regions: small compounds (1000 Da) as residue from the original HS structure that became smaller only gradually. This indicates that no random splitting of HS molecules occurs, since that would result in a very broad range of molecular sizes. Figure 8 visualizes our understanding of the ozonation process. Only small molecules are cleaved from the large 6464

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 20, 2006

(1) Siddiqui, M. S.; Amy, G. L.; Murphy, B. D. Ozone enhanced removal of natural organic matter from drinking water sources. Water Res. 1997, 31, 3098-3106. (2) Melin, E. S.; Ødegaard, H. The effect of biofilter loading rate on the removal of organic ozonation by-products. Water Res. 2000, 34, 4464. (3) Fahmi; Nishijima, W.; Okada, M. Improvement of DOC removal by multistage AOP-biological treatment. Chemosphere 2003, 50, 1043-1048. (4) Goel, S.; Hozalski, R. M.; Bouwer, E. J. Biodegradation of NOM: effect of NOM source and ozone dose. J. AWWA 1995, 1, 90105. (5) Graham, N. J. D. Removal of humic substances by oxidation/ biofiltration processes - A review. Water Sci. Technol. 1999, 40, 141-148. (6) Kooij, D. V. d.; Heijnen, W. A. M.; Kruithof, J. C. The effects of ozonation, biological filtration and distribution of easily assimilable organic carbon (AOC) in drinking water. Ozone Sci. Eng. 1989, 11, 297-311. (7) Nissinen, T. K.; Miettinen, I. T.; Martikainen, P. J.; Vartiainen, T. Molecular size distribution of natural organic matter in raw and drinking waters. Chemosphere 2001, 45, 865. (8) Thomsen, M.; Lassen, P.; Dobel, S.; Hansen, P. E.; Carlsen, L.; Mogensen, B. B. Characterization of humic materials of different origin: A multivariate approach for quantifying the latent properties of dissolved organic matter. Chemosphere 2002, 49, 1327-1337. (9) Gilbert, E. Biodegradability of ozonation products as a function of COD and DOC elimination by the example of humic acids. Water Res. 1988, 22, 123-126. (10) Hayes, M. H. B.; Clapp, C. H. Humic substances: considerations of compositions, aspacts of structure and environmental influences. Soil Sci. 2001, 166, 723-737. (11) Beckett, R.; Zhang, J.; Giddings, J. C. Determination of molecular weight distributions of fulvic and humic acids using flow fieldflow fractionation. Environ. Sci. Technol. 1987, 21, 289-295. (12) McDonald, S.; Bishop, A. G.; Prenzler, P. D.; Robards, K. Analytical chemistry of freshwater humic substances. Anal. Chim. Acta 2004, 527 (2), 105-124. (13) von Gunten, U. Ozonation of drinking water: Part I. Oxidation kinetics and product formation. Water Res. 2003, 37, 14431467. (14) Glaze, W. H.; Koga, M.; Cancilla, D.; Wang, K.; McGuire, M. J.; Liang, S.; Davis, M. K.; Tate, C. H.; Aieta, E. M. Evaluation of ozonation by-products from two California surface waters. J. AWWA 1989, 81, 66-73. (15) Killops, S. D. Volatile ozonation products of aqueous material. Water Res. 1986, 20, 153-165. (16) Richardson, S. D. Disinfection by-products and other emerging contaminants in drinking water. Trends Anal. Chem. 2003, 22, 666-684. (17) Carlson, K.; Amy, G. The formation of filter-removable biodegradable organic matter during ozonation. Ozone Sci. Eng. 1997, 19, 179-198.

(18) Jansen, R. H. S.; Rijk, R. J. W. d.; Zwijnenburg, A.; Mulder, M. H. V.; Wessling, M. Hollow fiber membrane contactors-A means to study the reaction kinetics of humic substance ozonation. J. Membr. Sci. 2005, 257, 48-59. (19) Hongve, D.; Baann, J.; Becher, G.; Lømo, S. Characterization of humic substances by means of high-performance size exclusion chromatography. Environ. Int. 1996, 22, 489-494. (20) Conte, P.; Piccolo, A. High pressure size exclusion chromatography (HPSEC) of humic substances: molecular sizes, analytical parameters and column performance. Chemosphere 1999, 38, 517-528. (21) Chin, Y.-P.; Aiken, G.; O’Loughlin, E. Molecular weight, polydispersity, and spectroscopic poperties of aquatic humic substances. Environ. Sci. Technol. 1994, 28, 1853-1858. (22) Perminova, I. V.; Frimmel, F. H.; Kudryatsev, A. V.; Kulikova, N. A.; Abt-Braun, G.; Hesse, S.; Petrosyan, V. S. Molecular weight characteristics of humic substances from different environments

as determined by size exclusion chromatography and their statistical evaluation. Environ. Sci. Technol. 2003, 37, 24772485. (23) Hoigne, J.; Bader, H.; Haag, W. R.; Staehelin, J. Rate constants of reactions of ozone with organic and inorganic compounds in water - III. Inorganic compounds and radicals. Water Res. 1985, 19, 993-1004. (24) O’Loughlin, E.; Chin, Y.-P. Effect of detector wavelength on the determination of the molecular weight of humic substances by high-pressure size exclusion chromatography. Water Res. 2001, 35, 333-338.

Received for review March 7, 2006. Revised manuscript received July 2, 2006. Accepted July 18, 2006. ES060533T

VOL. 40, NO. 20, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6465