A SIMPLE TENAX EXTRACTION METHOD TO DETERMINE THE AVAILABILITY OF SEDIMENT-SORBED ORGANIC COMPOUNDS

Transcription

1 Environmental Toxicology and Chemistry, Vol. 20, No. 4, pp , SETAC Printed in the USA /01 $ A SIMPLE TENAX EXTRACTION METHOD TO DETERMINE THE AVAILABILITY OF SEDIMENT-SORBED ORGANIC COMPOUNDS GERARD CORNELISSEN, HENK RIGTERINK, DORIEN E.M. TEN HULSCHER, BEA A. VRIND, and PAUL C.M. VAN NOORT* Institute for Inland Water Management and Waste Water Treatment, P.O. Box 17, 8200 AA Lelystad, The Netherlands Veluws College, Postbus 307, 7300 AH, Apeldoorn, The Netherlands (Received 27 March 2000; Accepted 10 August 2000) Abstract A simple method to determine the availability of sediment-sorbed organic contaminants was developed and validated. For 10 polycyclic aromatic hydrocarbons, 4 polychlorinated biphenyls, and 9 chlorobenzenes in 6 sediments, we measured the fraction extracted by Tenax in 6 and 30 h. These fractions were compared with the rapidly desorbing fractions determined by consecutive Tenax extraction. Extraction by Tenax for 30 h completely removed the rapidly desorbing fraction plus some part of the slowly desorbing fraction. The fraction removed after 30 h was about 1.4 times the rapidly desorbing fraction. The fraction extracted by Tenax after 6 h is about 0.5 times the rapidly desorbing fraction for chlorobenzenes (CBs)/polychlorinated biphenyls (PCBs) and polycyclic aromatic hydrocarbons (PAHs). The rapidly desorbing fraction probably represents the fraction of sorbed organic compound that poses actual risks for transport to (ground) water and determines the uptake by organisms and that can be microbially degraded. Extraction by Tenax for 6 h provides an easy way to address these issues more accurately than does the measurement of total concentrations. Keywords Slow desorption Organic compounds Actual risks Chemical availability Tenax extraction INTRODUCTION Organic contaminants in sediments can pose a substantial risk to animals and humans. Up till now, risk estimations have mostly been done by means of total contaminant concentrations in sediments. However, there is increasing awareness that total contaminant concentrations (measured by vigorous extractions with organic solvents) are not representative for actual risks for ecotoxicological effects and contaminant transport to groundwater and surface water. These actual risks should be based on the concentrations of contaminants in sediment that are available for transport and for uptake [1]. An important process that influences this availability is the sequestration of organic compounds in soils and sediments [2 4]. Sequestered compounds desorb only slowly and are mostly unavailable for transport and uptake by earthworms [5] and also by microorganisms [2,4]. Compounds at slowly and very slowly desorbing sites show Langmuir-like sorption whereas compounds at rapidly desorbing sites show linear sorption [6]. In other words, concentrations of solute freely dissolved in the pore water, which are sometimes taken to determine concentrations in organisms, are linearly related to concentrations of rapidly desorbing sorbate in sediments. The fractions of polycyclic aromatic hydrocarbon (PAH) extracted from marine sediment by both semipermeable membranes, Tenax TA (Chrompack, Middelburg, The Netherlands), and dialysis for an arbitrary contact time of 24 h have been shown to be related to the degradable fraction of PAH [7]. The amounts of DDT, dichlorodiphenyldichloroethylene (DDE), and dichlorodiphenyldichloroethane (DDD) taken up by C18 membrane disks from soils on a extraction time of 24 h have been found to be highly correlated with the amounts * To whom correspondence may be addressed (p.vnoort@riza.rws.minvenw.nl). taken up from these soils by Eisenia foetida [8]. The accumulation of polychlorinated biphenyls (PCBs), PAHs, and linear alkylbenzenes from sediment to a deposit-feeding bivalve has been found to correlate with the fraction desorbed to XAD- 4 resin in 48 h [9]. It may well be that desorption to an infinite sink for 24 or 48 h removes a fraction that is closely related to the rapidly desorbing, linearly sorbed fraction. The rapidly desorbing fraction can be determined by consecutive Tenax extraction in either a direct [10] or indirect way [11]. However, that is a tedious and time-consuming method. A more practical method, based on the difference in desorption rate constants between linearly and nonlinearly sorbed fractions, is needed to easily determine this rapidly desorbing, linearly sorbed fraction. Therefore, in this study, we determined the fractions removed from sediment by a single Tenax extraction using two extraction times (6 and 30 h) for 19 compounds (PAHs, PCBs, and CBs) in six different contaminated sediments from various locations in The Netherlands. We compared these fractions with the linearly sorbed, rapidly desorbing fractions as determined by a consecutive Tenax extraction method described earlier [10]. Experimental methods Chemicals. The PAHs tested include fluorene, phenanthrene, anthracene, fluoranthene, pyrene, benzo[a]anthracene, chrysene, benzo[b]fluoranthene, benzo[k]fluoranthene, and benzo[a]pyrene. The PCBs tested include trichloro-pcb-28, pentachloro-pcb-101, hexachloro-pcb-138, and hexachloro- PCB-153. The CBs tested include 1,2-dichlorobenzene, 1,3- dichlorobenzene, 1,4-dichlorobenzene, 1,3,5-trichlorobenzene, 1,2,4-trichlorobenzene, 1,2,3-trichlorobenzene, 1,2,3,4-tetrachlorobenzene, pentachlorobenzene, and hexachlorobenzene. Tenax TA (60 80 mesh; m), a porous polymer 706

2 Tenax to extract available sorbed organic compounds Environ. Toxicol. Chem. 20, Table 1. Sediments used in the experiments Sediment Locations in The Netherlands Compounds a Concentrations ( g/kg dry matter) b f OC Details Lake Ketelmeer PCBs, CBs 300 (5 100) Large lake in sedimentation area of the Ijssel River (a tributary of the River Rhine; sediment core ( cm) WD Wemeldinge PAHs 2400 (20 600) Small inland harbor Hollands Diep PAHs 400 (8 79) Lake in sedimentation area of River Rhine SH Sassenheim PAHs 2000 (70 500) Small inland harbor AV Amstelveen PAHs 600 (14 118) Small inland harbor Tolkamer PCBs, CBs 250 (1 60) Sediment from River Rhine a Total concentrations of all compounds tested; in parentheses are the ranges of concentrations for the individual compounds. b PCB polychlorinated biphenyl; CB chlorobenzenes; PAH polycyclic aromatic hydrocarbon. based on 2,6-diphenyl-p-phenylene oxide, was purchased from Chrompack, Middelburg, The Netherlands. Before use, the Tenax TA beads were rinsed with water, acetone, and hexane (each 3 10 ml/g Tenax) and dried overnight at 75 C. Sediments. Six sediments were used for the validation of the single-step Tenax extractions. These sediments are described in Table 1. They were sieved over 250 m to remove coarse material. Total content determination. Sediment solvent extractions were performed in triplicate by refluxing wet sediment (1 3 g) with a mixture of water (50 ml), hexane (50 ml), and acetone (20 ml) for 6 h. This extraction method has been shown to recover 85 to 100% of contaminants from aged sediment reference materials [12]. After separation of the hexane and acetone/water, the hexane was evaporated to 1 ml. For the PCBs and CBs, the hexane was directly analyzed by gas chromatography-electron capture detector (Hewlett Packard 5890 [Avondale, PA, USA] with 63 Ni electron-capture detector and HP 7673 autosampler; column: Chrompack, fused silica CP sil 8cb length 50 m, diameter 0.25 mm; carrier gas: He [Groenband high purity], 1 ml/min; make-up gas: N 2 [Groenband high purity], 60 ml/min). For the PAHs, the hexane supernatant was evaporated to 1 ml and dissolved in 10 ml of acetonitrile. The acetonitrile was subsequently evaporated to 1 ml. Analysis was carried out by high-performance liquid chromatography (Hewlett Packard 1050 with fluorescence detector [HP 1046]; column: reversed phase C 18, Vydac 201TP54 [Hesperia, CA, USA]; gradient elution with acetonitrile and water). Desorption technique. Desorption was determined at 20 C by means of a Tenax solid-phase extraction method described previously [10]. For each sediment, two types of desorption experiments were performed in triplicate (two single extractions, for 6 and 30 h, and an extraction in 11 consecutive steps). For consecutive desorption experiments, a mixture of Tenax TA (0.6 g), sediment (1 g dry wt), and Milli-Q water (70 ml, Bedford, MA, USA) was constantly shaken in a 100-ml separation funnel. HgCl 2 (5 mg) and NaN 3 (32 mg) were added to inhibit microbial activity. The Tenax was refreshed 10 times at periodic intervals and extracted with hexane (20 ml). The hexane was analyzed as described above. After termination of desorption, the remaining sediment and supernatant water were refluxed as described above with hexane (50 ml) and acetone (20 ml) for 17 h to extract and analyze all residual organic compounds. For the single-extraction experiments, the Tenax (1.5 g) and sediment (1 g dry wt) were separated after 6 or 30 h of shaking. Instead of introducing fresh Tenax, the sediment was directly refluxed as described above. The Tenax was also extracted and analyzed. Tenax resin and solvents were periodically checked for background signals. In some cases, the background signals prevented the quantification of extracted compounds present on Tenax. This was sometimes the case for anthracene due to interferences from material present on Tenax. Some concentrations of chlorobenzenes in the sediments and were too small relative to background signals from Tenax and solvents to be accurately determined on Tenax after desorption. They were not used for this study. Consecutive desorption data interpretation. Desorption from sediment has been found to be triphasic [6,11,13], with the disappearance of each fraction (rapidly, slowly, or very slowly desorbing) following distinct first-order kinetics. So, during desorption, the change in the concentrations of these fractions is described by df rapid rapid kdes F rapid (1) dt df slow slow kdes F slow (2) dt df very slow kvery slow des F very slow (3) dt where F rap, F slow, and F very slow are the rapidly, slowly, and very slowly desorbing fractions, respectively. The first-order rate constants for rapid, slow, and very slow desorption are designated k rap, k slow, and k very slow (h 1 ), respectively. At the start of the desorption (t 0), the sum of the three fractions is, by definition, Finitial Finitial Finitial rapid slow very slow 1 (4) The balance of fractions at any time t is given by Frapid Fslow Fvery slow Fdes 1 (5) in which F des (g/g) is the fraction cumulatively desorbed. Values of k rap, k slow, k very slow, and initial values for F rap, F slow, and F very slow were determined by fitting experimental data for F des to Equations 1 through 5 using the Scientist software from Micromath (Salt Lake City, UT, USA). RESULTS AND DISCUSSION Comparison of rapidly desorbing fractions with fractions from single Tenax extractions The consecutive desorption data were fitted to Equations 1 through 5. As an example, the cumulative desorption of fluor-

3 708 Environ. Toxicol. Chem. 20, 2001 G. Cornelissen et al. Table 3. Rapidly desorbing fractions of chlorobenzenes and polychlorinated biphenyls (PCBs, with standard deviations from triplicates) in the sediments and as determined by consecutive desorption to Tenax Compound 1,3-Dichlorobenzene 1,4-Dichlorobenzene 1,2-Dichlorobenzene 1,3,5-Trichlorobenzene 1,2,4-Trichlorobenzene 1,2,3-Trichlorobenzene 1,2,3,4-Tetrachlorobenzene Pentachlorobenzene Hexachlorobenzene Polychlorinated biphenyl-28 Polychlorinated biphenyl-101 Polychlorinated biphenyl-153 Polychlorinated biphenyl Fig. 1. The cumulative desorption of fluoranthene from sediment. The line through the datapoints results from the fitting. The dashed lines indicate the times (6 and 30 h) used for the single extractions. anthene from sediment is given in Figure 1. The fitting resulted in rate constants for rapid, slow, and very slow desorption of about 1, 10 2, and 10 4 /h (irrespective of sediment and compound), respectively, as observed earlier for other sorbent sorbate combinations [10,11,13]. In Figure 1, the first 10 h of desorption are dominated by desorption from the rapidly desorbing pool. From 10 to 100 h and 100 h of desorption, the desorption is dominated by the slowly and very slowly desorbing pools, respectively. The fitting of the data in Figure 1 to Equations 1 through 5 resulted in a sum of squared deviations of Fitting the data to a twophase model instead afforded a sum of squared deviations of The statistical F test at 90% confidence level on these sums of squared deviations revealed that the use of the threephase model gave a statistically significant improvement over the two-phase model. The values for F rap for the PAHs in the various sediments are given in Table 2. Those for the CBs and PCBs are in Table 3. For a few compound sediment combinations, fitted values for F very slow were found to be very small (usually in cases of F rap 0.4 for some PAHs in sediments and AM). In those cases, the precision in rate constants for very slow desorption was very low. In other words, the three-compartment model was not very sensitive to changes in either k very slow or F very slow in those cases. Then a two-compartment model could have been used equally well. However, we still applied the three-compartment model in those cases since in the other cases the three-compartment model was needed, and earlier we observed the desorption to be triphasic, with occasionally dominant sorption in the very slow domain [10,11,13]. The standard variations from triplicates were about equal to or higher than standard deviations from the individual fits. The rapidly desorbing fractions for chlorobenzenes are in the (low) range of to Earlier we found similarly low rapidly desorbing fractions of chlorobenzenes in sediment from the same area from which sediment was taken [11]. In the sediments and, rapidly desorbing fractions for PCBs were about one order of magnitude higher than those for the chlorobenzenes. For the PAHs in the various sediments, F rap shows considerable variation. In general, the magnitude of F rap may vary considerably due to variation in nonlinear sorption capacity, possibly leading to competition for sorption at these sites by other compounds. Competition for resistant sorption sites in soil organic matter has been found for aromatic acids [14] and lipids [15]. By and large, values for F rap show considerable variation and are much less than one. This is needed to determine to what extent rapidly desorbing fractions contribute to the fractions desorbed in 6 and 30 h. The ratios of F rap and the fractions desorbed to Tenax in a single step after 30 h (F 30h )and6h(f 6h ) for the CBs and PCBs are given in Table 4. Those for the PAHs are in Table 5. These ratios show some variation and are given as averages Table 2. Rapidly desorbing fractions of polycyclic aromatic hydrocarbons (PAHs, with standard deviations from triplicates) in the sediments,,, and s determined by consecutive desorption to Tenax Compound Fluorene Phenanthrene Anthracene Fluoranthene Pyrene Benzo[a]anthracene Chrysene Benzo[b]fluoranthene Benzo[k]fluoranthene Benzo[a]pyrene a Duplicates.

4 Tenax to extract available sorbed organic compounds Environ. Toxicol. Chem. 20, Table 4. Ratios (with standard deviations) of the rapidly desorbing fractions (F rap ) and the fractions desorbed after 30 h (F 30h )and6h(f 6h ) for chlorobenzenes and polychlorinated biphenyls (PCBs) Sediment F rap /F 30h F rap /F 6h 1,3-Dichlorobenzene 1,4-Dichlorobenzene 1,2-Dichlorobenzene 1,3,5-Trichlorobenzene 1,2,4-Trichlorobenzene 1,2,3-Trichlorobenzene 1,2,3,4-Tetrachlorobenzene Pentachlorobenzene Hexachlorobenzene Polychlorinated biphenyl-28 Polychlorinated biphenyl-101 Polychlorinated biphenyl-153 Polychlorinated biphenyl with standard deviations for the various compound classes in the various sediments in Table 6. The data in Table 6 show that, for all compound classes in all sediments, the fraction extracted in 30 h overestimates the linearly sorbed, rapidly desorbing fraction except for the PAHs in the sediment. We have no explanation for this. On average, the fraction extracted in 30 h is, within a factor of two, equal to 1.4 times F rap. This can be understood from the kinetics of the desorption in the following way. For many compound sediment combinations, F rap was found to be comparable to F slow (individual data not shown). The desorption plot in Figure 1, in which the desorption between 10 and 100 h is dominated by desorption from the slowly desorbing pool, shows that, after desorption for 30 h, about 50% of the slowly desorbing fraction is removed. So extraction for 30 h can be expected to afford about 1.5 times F rap, which is qualitatively in agreement with our observed overall average F rap /F 30h ratio. The average ratios F rap /F 6h for the compound class of PCBs suggests them to be slightly smaller than one. Nevertheless, the data for the PCBs are statistically not significantly different from most other average F rap /F 6h ratios. In general, average F rap /F 6h ratios are higher than one. The average F rap /F 6h ratio for all compounds and sediments is SoF rap is about two times F 6h within a factor of two. Consequences for observed relations between uptake by organisms and desorbed fractions In the following, we will attempt to apply the influence of desorption contact time found in this study to the relations between uptake by organisms versus uptake by artificial sorbents obtained by others. The isotherms for the sorption of phenanthrene by 27 different soils and sediments after a contact time of 14 d have been analyzed in terms of a combination of a linear and a Langmuir isotherm [16]. On the assumption that, due to the relatively short contact time of 14 d, the Langmuir isotherms represent sorbates at slowly desorbing sites Table 5. Ratios (with standard deviations) of the rapidly desorbing fractions (F rap ) and the fractions desorbed after 30 h (F 30h )and6h(f 6h ) for polycyclic aromatic hydrocarbons (PAHs) Compound Sediment F rap /F 30h F rap /F 6h Sediment F rap /F 30h F rap /F 6h Fluorene Phenanthrene Anthracene Fluoranthene Pyrene Benzo[a]anthracene Chrysene Benzo[b]fluoranthene Benzo[k]fluoranthene Benzo[a]pyrene Phenanthrene Fluoranthene Pyrene Benzo[a]anthracene Chrysene Benzo[b]fluoranthene Benzo[k]fluoranthene Benzo[a]pyrene a Duplicates.

5 710 Environ. Toxicol. Chem. 20, 2001 G. Cornelissen et al. Table 6. Ratios of the rapidly desorbing fraction (F rap ) and the fractions desorbed to Tenax in 30 and 6 h for the compound classes polycyclic aromatic hydrocarbon (PAH), chlorobenzenes (CB), and polychlorinated biphenyls (PCB) in the various sediments Sediment F rap /F 30h PAH CB PCB All F rap /F 6h PAH CB PCB All AM All only, we calculate from the data in Huang et al. [16] that, at low concentrations of phenanthrene, F slow /F rap for these 27 soils and sediments has been In the present study, we observed F rap to be about equal to F slow for all sediments. So this seems to be a general phenomenon. Desorption to Tenax in 30 h removes the rapidly desorbing fraction plus about half of the slowly desorbing fraction. Therefore, the fraction removed by an infinite sink in 24 or 48 h, as employed in studies by others, can now be estimated to be 1 to 2 or 1.5 to 2.5 times the rapidly desorbing fraction, respectively. The biodegradable fractions of PAHs in sediment correlate with the fraction desorbed in 24 h [7]. Our present results now suggest that biodegradable fractions correlate with either the rapidly desorbing fraction or the sum of the rapidly desorbing and slowly desorbing fractions. Earlier we demonstrated that biodegradable fractions correlate with rapidly desorbing fractions [17]. The correlation of (i) the uptake of DDT, DDE, and DDD by Eisenia foetida with fractions desorbed in 24 h [8] and (ii) sediment to biota accumulation factors for a deposit feeding bivalve with the fraction desorbed in 48 h [9] suggests that at least very slowly desorbing fractions are not taken up by these organisms. General discussion This study shows that the concentration in sediment of rapidly desorbing, linearly sorbed fractions can be easily determined by the measurement of the amount desorbed to Tenax. The use of a contact time of 6 h is to be preferred over a contact time of 30 h as desorption for 30 h will also remove a substantial part of the slowly desorbing, nonlinearly sorbed fraction and F slow /F rap is not constant. Our combined data for chlorobenzenes, PCBs, and PAHs show that the amount linearly sorbed is about two times the amount desorbed to Tenax on a contact time of 6 h. These linearly sorbed concentrations can be used to assess contamination of sediments (and soils) by organic compounds with respect to topics that are related to concentrations in pore water, such as, e.g., the transport to (ground) water, the uptake by organisms that feed from water, and the feasibility of rapid bioremediation. Elucidation of the chemical and sorption characteristics of the slowly and very slowly desorbing sites combined with the development of knowledge of the strategies employed by organisms for the uptake from geosorbents may aid in further ascertaining whether these type of organisms do not take up from either slowly or very slowly desorbing sites. Furthermore, the determination of rapidly desorbing, linearly sorbed fractions may aid in the determination of the eventual influence of feeding behavior (e.g., avoidance) of organisms on the extent of uptake. In many cases, rapidly desorbing fractions are much less than 1.0 (in our data set, 80% of the F rap data were 0.4). Compared with the determination of total contents by direct solvent extraction of geosorbents, extraction by Tenax for 6 h will be a facile and hardly more expensive way to determine available, linearly sorbed concentrations, which are much more meaningful than total concentrations. Acknowledgement We are grateful to Angelique Belfroid and an anonymous reviewer for many valuable suggestions and comments. REFERENCES 1. Alexander M Sequestration and bioavailability of organic compounds in soil. In Linz DG, Nakles DV, eds, Environmentally Acceptable Endpoints: Risk Based Approach to Contaminated Site Management Based on Availability of Chemicals in Soil. American Academy of Environmental Engineers, Annapolis, MD, pp Alexander M How toxic are toxic chemicals in soil? Environ Sci Technol 29: Pignatello JJ, Xing B Mechanisms of slow sorption of organic chemicals to natural particles. Environ Sci Technol 30: Loehr R, Webster M Behavior of fresh vs. aged chemicals in soil. J Contam Hydrol 5: Tang J, Carroquino MJ, Robertson BK, Alexander M Combined effect of sequestration and bioremediation in reducing the bioavailability of PAHs in soil. Environ Sci Technol 32: Cornelissen G, Rigterink H, Van Noort PCM, Govers HAJ Slowly and very slowly desorbing organic compounds in sediments exhibit Langmuir-type sorption. Environ Toxicol Chem 19: MacRae JD, Hall KJ Comparison of methods used to determine the availability of polycyclic aromatic hydrocarbons in marine sediment. Environ Sci Technol 32: Tang J, Robertson BK, Alexander MA Chemical-extraction methods to estimate bioavailability of DDT, DDE, and DDD in soil. Environ Sci Technol 33: Lamoureux EM, Brownawell BJ Chemical and biological availability of sediment-sorbed hydrophobic organic chemicals. Environ Toxicol Chem 18: Cornelissen G, Van Noort PCM, Govers HAJ Desorption kinetics of chlorobenzenes, PAHs and PCBs: Sediment extraction with Tenax and effects of contact time and solute hydrophobicity. Environ Toxicol Chem 16: Ten Hulscher ThEM, Vrind BA, Van der Velde L, Van den Heuvel H, Van Noort PCM Triphasic desorption of chlorobenzenes, PCBs and PAHs from field-contaminated sediment. Environ Sci Technol 33: European Commission BCR information on reference materials. The certification of the contents mass fractions of eight

6 Tenax to extract available sorbed organic compounds Environ. Toxicol. Chem. 20, chlorobiphenyls, IUPAC Nos 101, 118, 128, 149, 153, 156, 170, 180 in industrial soil. Report EUR EN. Committee on Reference Materials 48, Ispra, Italy. 13. Cornelissen G, Van Noort PCM, Parsons JR, Govers HAJ Temperature dependence of slow adsorption and desorption kinetics of organic compounds in sediments. Environ Sci Technol 31: Xing B, Pignatello JJ Competitive sorption between 1,3- dichlorobenzene or 2,4-dichlorophenol and natural aromatic acids in soil organic matter. Environ Sci Technol 32: Kohl SD, Toscano PJ, Hou W, Rice JA Solid-state 19 F NMR investigation of hexafluorobenzene sorption to soil organic matter. Environ Sci Technol 34: Huang W, Young TM, Schlautman MA, Yu H, Weber WJ Jr A distributed reactivity model for sorption by soils and sediments. 9. General isotherm non-linearity and applicability of the dual reactive domain model. Environ Sci Technol 31: Cornelissen G, Rigterink H, Ferdinandy MMA, Van Noort PCM Rapidly desorbing fractions of PAHs in contaminated sediments as a predictor of the extent of bioremediation. Environ Sci Technol 32: