Comparison of Accelerated Solvent Extraction and Soxhlet Extraction

Nov 1, 2000 - The two major input routes of PCBs to the Baltic Sea are through .... The air-dried sediment samples were then ground in an automatic gr...
0 downloads 0 Views 73KB Size
Environ. Sci. Technol. 2000, 34, 4995-5000

Comparison of Accelerated Solvent Extraction and Soxhlet Extraction for the Determination of PCBs in Baltic Sea Sediments CECILIA BANDH* Institute of Applied Environmental Research and Department of Analytical Chemistry, Stockholm University, SE-106 91 Stockholm, Sweden ERLAND BJO ¨ RKLUND AND LENNART MATHIASSON Department of Analytical Chemistry, Lund University, P.O. Box 124, SE-221 00 Lund, Sweden C A R I N A N A¨ F A N D Y N G V E Z E B U ¨ HR Institute of Applied Environmental Research, Stockholm University, SE-106 91 Stockholm, Sweden

The effectiveness of the extraction conditions proposed in ASE U.S. EPA Method 3545 was evaluated for the extraction of polychlorinated biphenyls (PCBs) in Baltic Sea sediments with different sample characteristics such as organic carbon, soot carbon, sulfur, water content, and PCB concentration range. PCB concentrations determined with accelerated solvent extraction (ASE) using n-hexane/ acetone (1:1, v/v) as organic solvents were compared to Soxhlet extraction using toluene as organic solvent. The results indicated that sediments containing relatively low amounts of organic and soot carbon can be quantitatively extracted by ASE in 5 min with n-hexane/acetone (proposed in EPA Method 3545). However, for sediments with larger amounts of carbon, the total amount of PCBs present in the sample might be underestimated when using n-hexane/ acetone. For these latter sediments, toluene seems to increase the extraction efficiency in ASE.

Introduction Over the past two decades, polychlorinated biphenyls (PCBs) have been of environmental concern due to their persistence, toxicity, carcinogenic potential, and ability to interfere with reproductive systems (1, 2). The Baltic Sea has experienced a considerable load of various organic pollutants such as PCBs, 1,1,1-trichloro-2,2-bis(4-chlorophenyl)ethane (DDT), and polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs) (3, 4). The two major input routes of PCBs to the Baltic Sea are through atmospheric input and water runoff (5, 6). One important sink for these pollutants is considered to be bottom sediments, which play an important role for the distribution and dynamics of organic pollutants in the aquatic environment. Therefore, it is of great interest to determine the concentrations of key pollutants in different sediments. Sample preparation in environmental analysis is usually time-consuming and involves expensive processes that aim * Corresponding author fax: +46-8-6747638; phone: +46-86747333; e-mail: [email protected]. 10.1021/es991064g CCC: $19.00 Published on Web 11/01/2000

 2000 American Chemical Society

to isolate target analytes before their final determination. The conventional extraction methods generally require several hours or days to perform and demand large amounts of hazardous organic solvents. Therefore, substantial efforts have been made to develop and evaluate new sample preparation techniques (7-9). Considerable interest has developed, for example, in supercritical fluid extraction (SFE) (10, 11), microwave-assisted extraction (MAE) (12), and accelerated solvent extraction (ASE) (13) for the preparation of solid matrixes. These techniques all offer reduced extraction times, increased sample throughput, and decreased consumption of organic solvents. Accelerated solvent extraction (ASE) is one of the latest techniques developed for extraction of solid samples and was introduced in 1995 (14, 15). The principle of ASE is simple. The sample is placed in a stainless steel cell that is heated to temperatures in the range of 50-200 °C. Normally the same organic solvent or combination of solvents as used in Soxhlet extraction are pumped into the cell (16), and by pressurizing the extraction cell, the solvents are kept in their liquid state. A more detailed discussion on performing extraction of organic pollutants can be found in a recent review by Bjo¨rklund et al. (17). While literature describing the use of ASE is increasing, little research has been devoted to using ASE for PCB extraction (13, 15, 18). A few papers have recently been published, evaluating the effectiveness of ASE for PCB analysis in various matrixes (19-22), but only two of these deal with PCBs in sediments (19, 20). Our aim in the work presented here was to evaluate the effectiveness of the extraction conditions proposed in ASE U.S. EPA Method 3545 (23) as compared to traditional Soxhlet extraction for the determination of PCBs in Baltic Sea sediments. Since the extraction efficiency of organic pollutants in many studies are known to be matrix dependent (24, 25), Baltic Sea sediments with varying content of carbon, sulfur, and PCBs were chosen for the present study.

Experimental Section Sediment Samples: Collection and Preparation. Seven sediment samples (A-G) from different areas in the Baltic Sea were used in this study. All sediments were collected using gravitational corers, and the top section of the cores was sliced into 1 cm thick slices immediately after sampling. These sediment slices were placed in new prewashed polypropylene boxes and stored at -18 °C. Sampling positions (coordinates), layers (cm) used for extraction, content of carbon (organic and soot in % of dry weight (dw)), sulfur and water contents (% of dw), approximate concentration ranges of investigated PCBs (ng/g dw), and amount of sample extracted (g) in ASE and Soxhlet are presented in Table 1 for each sample. Prior to extraction (ASE and Soxhlet), the sediments were homogenized (Ultra Turrax Janke and Kunkel, Staufen, Germany) and then centrifuged at approximately 1000g for 15 min. After the supernatant water was removed, aliquots of each sediment sample were collected and stored at -18 °C for later Soxhlet extraction. For the ASE extractions, the centrifuged sediment residues were then dried at room temperature. The approximate mean time for the samples to dry was 48 h. The air-dried sediment samples were then ground in an automatic grinding machine to homogeneous powders, which were stored in dark glass jars (+4 °C) for later ASE extraction. Accelerated Solvent Extraction. An ASE 200 Accelerated Solvent Extraction system (Dionex Corporation, Sunnyvale, CA) was used to perform the accelerated solvent extractions. VOL. 34, NO. 23, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4995

TABLE 1. Sample Characteristics for the Seven Bottom Sediments from the Baltic Sea position (coord)

A (Southern Baltic Proper) B (Southern Baltic Proper) C (Bothnian Bay) D (Bothnian Sea) E (Gotland Deep) F (urban waters of Stockholm) G (Baltic Proper)

lat

long

55°47′27′′ 55°38′27′′ 65°22′92′′ 61°05′00′′ 57°25′63′′ 59°19′25′′ 58°10′14′′

14°38′90′′ 20°28′58′′ 23°06′93′′ 19°33′67′′ 19°50′51′′ 18°05′28′′ 18°14′38′′

(% of dw) layers organic soot water PCB concn (cm) carbon carbon sulfur (%) range (ng/g dw) 1-5 1-5 1-6 2-5 0-10 0-3 1-5

0.21 1.2 2.2 2.8 8.5 8.9 9.2

This extraction unit consists of 24 extraction cells that can be processed sequentially. Maximum pressure and temperature are 20 MPa and 200 °C, respectively. Extracts are transferred to the collection unit where 26 collection vials can be positioned. ASE extractions were performed on the air-dried sediment samples A-G according to the ASE U.S. EPA Method 3545 (23). Dionex standard stainless steel cells with a volume of 11 mL were loaded with air-dried sediment and Hydromatrix (Hydromatrix IST Hengoed, Mid Glamorgan, UK) mixed in equal proportions. Prior to extraction, four 13C12-labeled PCB standards (IUPAC Nos. 52, 101, 138, and 180; Cambridge Isotope Laboratories, Woburn, MA) were added as internal standards to the samples. The extraction sequence began with a filling step when n-hexane/acetone (1:1, v/v) (HPLC grade, Merck, Darmstadt, Germany) was pumped into the cell. The second step was 5 min of preheating to 100 °C ensuring complete sample heating prior to the 5-min static extraction step. During the static extraction, the pressure and temperature applied were 14 MPa and 100 °C, respectively. After the extraction was completed, the pressure was released and the extract collected in 25-mL glass vials. Fresh solvent was pumped through the cell (60% of the extraction cell volume as set by the software), and finally, as a last step in the sequence, purging for 1 min with an inert gas (Nitrogen, 4.8 grade AGA gas AB, Sweden) assured complete transfer of the solvent to the collection vial. The final extract volume was approximately 15 mL. Four ASE extractions using toluene (Burdick and Jackson, Fluka Chemie AG, Buchs, Switzerland) as solvent were also performed on the air-dried sediment sample E. For these extractions, the pressure and the temperature were held at 20 MPa and 160 °C, respectively. All other parameters were the same as for the ASE extractions using n-hexane/acetone. Soxhlet Extraction. Triplicates of each centrifuged sample A-G were placed in pre-extracted cellulose thimbles and extracted (wet) for 24 h with toluene using a Soxhlet apparatus. Toluene has previously been demonstrated superior to n-hexane/acetone in Soxhlet extractions of aromatic compounds from samples containing black carbon (soot) (26). To remove the water from the samples, a DeanStark trap was attached to the Soxhlet extractor (27). Prior to extraction, four 13C12-labeled PCBs (IUPAC Nos. 52, 101, 138, and 180) were added as internal standards. After the Soxhlet extractions, the samples were heated to 105 °C for 24 h in their preweighed cellulose thimbles. The thimbles containing the dried samples were then kept in a desiccator for later dry weight determinations of the extracted samples. Triplicates of the air-dried sediment sample E were also Soxhlet extracted for 24 h in n-hexane/acetone (1:1, v/v). The sediment was mixed with Na2SO4 in equal proportions, and four 13C12-labeled PCBs (IUPAC Nos. 52, 101, 138, and 180) were added as internal standards prior to extraction. Cleanup of Extracts. Prior to cleanup, a 13C12-labeled PCB standard (IUPAC No. 118 or 153; Cambridge Isotope Laboratories) was added to the samples in addition to the internal standards added prior to the extractions for recovery estima4996

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 23, 2000

0.043 0.20 0.17 0.47 0.89 1.0 1.1

0.032 0.25 0.062 0.099 0.77 2.4 1.2

0.4 2 2 4 3 14 8

0.01-0.1 0.02-0.4 0.1-1 0.1-1 0.3-3 3-70 0.3-4

sample amount ASE; Soxhlet (g dw) 1.5-2.2; 3.6-5.5 1.3-1.4; 3.1-4.2 0.93-1.1; 2.6-2.9 0.97-1.3; 1.7-2.2 0.93-1.3; 1.3-1.5 0.12-0.20; 0.24-0.41 0.27-0.38; 1.1-1.3

tions of the cleanup steps. All extracted volumes were reduced by rotary evaporation to approximately 1 mL and then eluted with 50 mL of n-hexane (Burdick and Jackson) through a column packed with SiO2 (deactivated with 10% water w/w, Merck, 63-200 µm). The eluate volumes were once again reduced to approximately 15 mL, and 5 g of activated copper powder (Merck) was added to the extracts to remove elemental sulfur. The sediment extracts together with the copper powder were sonicated for 30 min, left overnight, and then sonicated for another 30 min before the copper powder was removed by elution through glass wool in a Pasteur pipet. The extracts were then fractionated by high-performance liquid chromatography (HPLC) using a nitropropylphenylsilica (nitro) column (Nucleosil, 5-µm particles, 250 × 4.6 mm, Macherey-Nagel, Germany) with n-hexane as the mobile phase. When the nitro column is operated in straight phase mode, hydrocarbons can be separated according to the size of the conjugated π-electron system (28). In short, three fractions were collected: (i) aliphatic/monocyclic aromatic compounds (∼4.5 mL), (ii) dicyclic aromatic compounds (including the PCBs, ∼9 mL), and (iii) polycyclic aromatic hydrocarbons (four-ring PAHs and higher, ∼5 mL). The HPLC fraction 2 containing the PCBs was further cleaned up prior to GC/MS analysis. After volume reduction, the extracts were eluted with 7 mL of n-hexane through columns containing three layers of modified silica. The silica gels were packed in a Pasteur pipet and consisted of a bottom layer of SiO2/ H2O (5 mm), a middle layer of SiO2/KOH (10 mm), and a top layer of SiO2/H2SO4 (5 mm). On top of the silica layers 5 mm of anhydrous sodium sulfate (Na2SO4) was added (29). This method was modified from Smith et al. 1989 (30). The silica used in the cleanup steps was activated at 500 °C for 24 h and then modified with water (10% w/w), 7.7 M potassium hydroxide (KOH) in methanol (35% w/w) or concentrated sulfuric acid (H2SO4, 40% w/w). PCB Analysis. The sediments were analyzed for their content of the following 15 PCB congeners denoted by their IUPAC Nos. (52, 49, 44, 95, 101, 110, 149, 118, 153, 105, 138, 180, 170, 199, and 196). All samples were analyzed on a Fison 8060 gas chromatograph (GC) with a Fison MD800 lowresolution mass selective detector (MS). A 13C12-labeled recovery standard (IUPAC No. 28 or 153; Cambridge Isotope Laboratories) was added to all samples before injection on the GC/MS. The samples were introduced by on-column injection (1 µl) into a PTE-5 fused silica column (Supelco, Bellefonte, PA; 30 m × 0.25 mm i.d., 0.25 µm film thickness). Helium was used as the carrier gas at a head pressure of 15 psi. The column temperature was held at 100 °C for 2 min, then increased at a rate of 10 °C/min to 200 °C, then increased at 3.5 °C/min to a temperature of 295 °C, and finally held constant for 20 min. The analyses were performed in the electron impact (EI) mode at an ionization energy of 70 eV, and the detection was carried out in selected ion monitoring (SIM) mode, recording two isotopic molecular ions for each analyte. The ion source and interface temperatures were 200

TABLE 2. Mean Concentrations (ng/g dw) and Standard Deviations (RSD in %) for PCBs in Sediments with Low Levels of Carbon, Sulfur, and PCBsa sediment A ASEb PCB no. 52 49 44 95 101/90 110 118 105 149 153 138 180 170 199 196 av

sediment B

Soxhletc

ASEb

sediment C

Soxhletc

ASEb

sediment D

Soxhletc

ASEb

Soxhletc

concn RSD concn RSD recoveryd concn RSD concn RSD recoveryd concn RSD concn RSD recoveryd concn RSD concn RSD recoveryd (ng/g) (%) (ng/g) (%) (%) (ng/g) (%) (ng/g) (%) (%) (ng/g) (%) (ng/g) (%) (%) (ng/g) (%) (ng/g) (%) (%) 0.042 0.020 0.021 0.036 0.066 0.060 0.056 0.023 0.080 0.118 0.140 0.097 0.060 0.016 0.015

6.5 4.3 13.3 10 3.2 3.7 4.8 7.4 5.5 5.6 7.4 6.6 4.8 9.9 6.2

0.036 0.019 0.019 0.038 0.060 0.061 0.052 0.020 0.077 0.119 0.129 0.116 0.071 0.014 0.018

8.3 3.2 6.6 7.3 7.6 5.2 3.5 3.4 7.2 6 7.5 8.7 9.5 11.7 8.2

116 104 109 94 111 100 107 111 105 99 108 83 84 114 85 102

0.052 0.021 0.029 0.075 0.126 0.132 0.113 0.048 0.178 0.298 0.375 0.257 0.150 0.036 0.042

9.6 11.2 8.6 6.6 5.7 5.3 5.0 13.0 2.2 5.8 2.7 8.1 7.6 9.6 11.0

0.043 0.019 0.025 0.075 0.134 0.144 0.111 0.046 0.227 0.325 0.410 0.280 0.174 0.034 0.037

7.9 12.2 12.0 7.2 2.6 3.4 4.7 5.3 8.3 5.7 5.2 3.0 3.5 12.2 9.6

120 109 118 101 94 92 101 103 79 92 91 92 87 105 113 99

0.191 0.068 0.090 0.192 0.371 0.417 0.368 0.146 0.557 0.843 1.021 0.729 0.458 0.118 0.114

10.1 8.8 12.6 5.0 4.2 6.2 6.5 8.7 4.8 8.2 6.8 9.3 8.7 6.9 4.2

0.209 0.066 0.080 0.187 0.405 0.470 0.322 0.129 0.562 0.793 1.146 0.739 0.482 0.128 0.123

7.2 7.2 10.7 13.2 7.4 10.1 12.7 11.0 8.0 10.8 10.6 7.8 13.3 13.3 25.2

91 103 113 103 92 89 114 113 99 106 89 99 95 92 93 99

0.216 0.093 0.105 0.177 0.355 0.392 0.476 0.187 0.433 0.761 0.890 0.661 0.375 0.129 0.129

5.0 3.4 4.9 7.1 4.6 4.2 4.1 6.0 4.8 3.7 2.5 5.3 6.5 9.7 7.7

0.276 0.093 0.126 0.224 0.419 0.439 0.498 0.201 0.518 0.868 1.053 0.845 0.489 0.163 0.156

5.0 8.4 3.1 8.0 8.3 8.1 3.4 4.3 7.7 3.2 3.5 6.1 8.8 6.2 13.0

78 100 83 79 85 89 96 93 84 88 85 78 77 80 83 85

a The PCBs were recovered using ASE and Soxhlet extractions. (Two significant figures are valid.) b ASE using n-hexane/acetone as solvent (n ) 4). c Soxhlet using toluene as solvent (n ) 3). d Recovery of ASE as compared to Soxhlet concentrations.

and 250 °C, respectively. The PCB congeners were quantified using relative response factors (RRFs) against the 13C12-labeled standard with the same or nearest number of chlorine atoms. A five-point calibration curve ranging from 2.5 pg to 1 ng/ congener was used to check the linearity of the MS. Detection limits for the GC/MS were determined to a range of 0.1 pg/ triCB congener - 0.8 pg/octaCB congener. Sample Characterization. The water content (Table 1) and the dry weight of the air-dried, ground samples were determined by heating overnight in an oven at 105 °C. Heating air-dried samples to temperatures just above 100 °C is normal procedure for determination of water content, and values generated are very close to certified values as demonstrated for air-dried reference materials (31, 32). The content of organic carbon was quantified according to standard procedures with a Carlo Erba elemental analyzer. The soot carbon was differentiated from the organic carbon by thermal oxidation at 375 °C in air for 24 h. The samples were then analyzed for soot carbon using elemental analysis (Carlo Erba) (33). The content of sulfur (oxidized, reduced, and elemental) was analyzed according to standard procedures using an element analyzer. QA/QC. The recoveries were calculated for the internal standards added prior to extraction (13C12-labeled PCBs Nos. 52, 101, 138, and 180; see Accelerated Solvent Extraction and Soxhlet Extraction) as well as the internal standard added prior to cleanup (13C12-labeled PCBs No. 118 or 153; see Cleanup section). These were compared to the recovery standard added just before the injection into the GC/MS (13C12-labeled PCBs No. 28 or 153; see PCB Analysis). The recoveries for the standards added prior to the extractions were generally good and were in the range of 67-97%. The standard added prior to cleanup also showed good recoveries and were in the range of 79-105%. Procedure blanks (n ) 8) for the ASE extractions were run every fifth sample, and three procedure blanks were also run to determine background levels for the Soxhlet extractions. The procedure blanks for both ASE and Soxhlet included all cleanup steps. Blank levels were typically e3% of the total levels for all samples except sediments A and B. For a few of the dominating congeners in these sediments (mainly PCB Nos. 52, 118, and 180), blank levels ranged between 4 and 10% of total levels. The PCB levels for these congeners were therefore corrected accordingly.

Results and Discussion Sediments with Low Levels of Carbon, Sulfur, and PCBs. Sediments A-D all contained low amounts of organic carbon (0.7%), and PCBs (Table 1). The concentrations of the 15 native PCBs varied between 0.3 and 70 ng/g (Table 3). Two of the sediments (E and F) showed significantly lower recoveries from the ASE extractions as compared to the Soxhlet extractions, with recoveries in the range of 61-83% and 47-92% for sediments E and F, respectively. In contrast, sediment G was quantitatively extracted with congener recoveries in the range of 82-118%. Comparison of Sediments with Low or High Levels of Carbon, Sulfur, and PCBs. The results obtained for the lowlevel sediments (Table 2) and the high-level sediments (Table 3) indicate that the extraction with ASE, to some extent, is correlated to the carbon content of the sample (Figure 1). These findings are in agreement with previous investigations performed on fly ash, where an increasing amount of carbon in the matrix made PAHs less accessible to the extraction solvent in ASE (24, 25). Note, however, that sediment G showed good PCB recoveries using ASE despite the high carbon content. This means that, even though the carbon content seems to be of importance, other sample characVOL. 34, NO. 23, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4997

TABLE 3. Mean Concentrations (ng/g dw) and Standard Deviations (RSD in %) for PCBs in Sediments with High Levels of Carbon, Sulfur, and PCBsa sediment E ASEb

sediment F

Soxhletc

ASEb

PCB no.

concn (ng/g)

RSD (%)

concn (ng/g)

RSD (%)

recoveryd

52 49 44 95 101/90 110 118 105 149 153 138 180 170 199 196 av

0.74 0.32 0.40 0.60 1.13 1.03 1.35 0.53 1.22 2.17 2.40 1.87 0.92 0.33 0.34

5.8 3.9 7.3 2.3 2.1 4.0 3.8 4.4 2.4 1.4 2.1 5.1 4.4 4.5 5.6

0.93 0.45 0.54 0.88 1.61 1.39 1.77 0.67 2.00 3.31 3.34 2.33 1.14 0.40 0.45

2.3 5.0 4.2 9.9 7.0 4.0 1.7 6.4 13.5 6.0 5.1 1.9 1.3 2.9 1.8

79 71 74 68 70 74 77 80 61 66 72 80 81 83 75 74

(%)

sediment G

Soxhletc

ASEb

concn (ng/g)

RSD (%)

concn (ng/g)

RSD (%)

recoveryd

16.0 11.3 10.5 12.7 29.0 24.9 22.5 8.0 26.1 31.8 48.5 30.3 17.0 2.7 3.7

3.2 14.5 2.5 2.8 7.5 2.8 2.1 8.0 2.6 2.5 2.0 1.3 7.7 8.3 6.3

22.5 12.3 15.8 22.9 33.8 42.7 35.5 14.4 38.9 45.7 67.7 49.4 36.2 4.8 5.1

1.3 9.1 13.6 12.2 8.0 10.3 2.7 13.7 4.4 11.6 4.2 10.5 11.4 9.2 10.4

71 92 66 55 86 58 63 56 67 69 72 61 47 57 72 66

(%)

Soxhletc

concn (ng/g)

RSD (%)

concn (ng/g)

RSD (%)

recoveryd (%)

1.17 0.48 0.58 1.12 1.89 1.89 2.40 0.87 1.96 3.44 3.61 1.86 0.91 0.34 0.39

4.1 2.8 8.3 4.3 1.4 2.8 1.4 4.3 2.7 2.2 3.8 0.5 4.0 1.3 5.3

1.12 0.50 0.57 1.05 1.79 1.68 2.03 0.76 1.98 3.57 3.57 2.25 1.11 0.39 0.44

3.2 4.0 1.4 2.2 3.9 2.6 6.6 6.6 1.1 10.0 7.5 7.5 10.8 9.1 7.8

104 96 102 107 106 113 118 116 99 96 101 83 82 88 88 100

a The PCBs were recovered using ASE and Soxhlet extractions. (Two significant figures are valid.) b ASE using n-hexane/acetone as solvent (n ) 4). c Soxhlet using toluene as solvent (n ) 3). d Recovery of ASE as compared to Soxhlet concentrations.

FIGURE 1. Relationships between PCB recoveries and the content of (a) organic carbon and (b) soot carbon in seven Baltic Sea sediments (A-G). The sediments are presented in order of increasing organic carbon and soot carbon content. The PCB recoveries are calculated as the percentage of total PCB obtained by using ASE (n-hexane/acetone) as compared to Soxhlet (toluene). teristics might be govern the release of target analytes during the course of extraction. For the low-level sediments, sediment D showed the lowest recoveries using ASE with an average recovery of 85% (Table 2). This sediment also contained the largest amount of water and second highest level of sulfur for the low-level sediments (Table 1). Similarly the high-level sediment F, which showed the lowest recoveries using ASE (Table 3), contained the highest levels of both water and sulfur (Table 4998

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 23, 2000

1). This contrasts from sediments E and G, where the former yielded an average recovery of 74%, while the latter was completely extracted by ASE (100%), (Table 3). Yet sediment G contained about twice the amount of sulfur and water as compared to sediment E (Table 1). Accordingly, it seems that water and sulfur have only a limited influence on the extraction efficiency, which is further clarified in Figure 2, where no strong correlation between sulfur, water, and PCB recoveries can be observed.

FIGURE 2. Relationships between PCB recoveries and the content of (a) sulfur and (b) water in seven Baltic Sea sediments (A-G). The sediments are presented in order of increasing sulfur and water content. The PCB recoveries are calculated as the percentage of total PCB obtained by using ASE (n-hexane/acetone) as compared to Soxhlet (toluene). Toluene as Extraction Solvent in ASE. Since according to existing literature toluene might be a better choice than n-hexane/acetone, toluene was investigated as a solvent in ASE extractions of air-dried sediment E (n ) 4). The effects of using n-hexane/acetone as solvents in Soxhlet extractions of this sediment (n ) 3) were also determined. The use of toluene instead of n-hexane/acetone for ASE extractions of sediment E as proposed in ASE U.S. EPA Method 3545 increased the average recovery from 74 (Table 3) to 85% (Table 4). The use of n-hexane/acetone for Soxhlet extraction rather than toluene reduced the average recovery to 82%. Consequently, if the results obtained for sediment E with ASE U.S. EPA Method 3545 are compared to the Soxhlet data obtained with n-hexane/acetone, the average recovery for all investigated congeners is 90%. Therefore, for this “hardto-extract” sediment, quantitative recoveries would have been achieved if the results were compared to Soxhlet extraction using n-hexane/acetone as solvents. It is likely that the same would have occurred for all other samples investigated. In this case, an erroneous conclusion would have been that the conditions proposed in ASE U.S. EPA Method 3545 are sufficient to accomplish quantitative recoveries for all types of samples (15, 16). It is likely that the conditions proposed in ASE U.S. EPA Method 3545 are valid for most matrixes. However, there is a risk that the concentrations are underestimated in some cases due to insufficient interaction between the solvent and the matrix. Even after replacing the n-hexane/acetone mixture in the ASE method with toluene, not more than an average recovery of 85% could be achieved for sediment E (Table 4). One reason for this might be that the extraction time (5 min) used in the ASE method is too short. Literature data support this idea; for example, Popp et al. (34) suggested that a static step of 2 × 5 min should be used for these types

TABLE 4. Mean Concentrations (ng/g dw) and Standard Deviations (RSD in %) for PCBs in Sediment Sample Ea

PCB no. 52 49 44 95 101/90 110 118 105 149 153 138 180 170 199 196 av

Soxhlet ASE toluene n-hexane/acetone (n ) 4) (n ) 3) recovery recovery concn RSD vs Soxhlet concn RSD vs Soxhlet (ng/g) (%) tolueneb (%) (ng/g) (%) tolueneb (%) 0.88 0.40 0.34 0.62 1.38 1.16 1.57 0.61 1.38 2.56 2.67 1.99 1.00 0.41 0.47

9.2 14.6 9.3 5.9 10.6 8.4 6.6 4.0 6.2 6.9 8.2 13.2 8.3 22.3 23.2

94 90 64 70 86 83 89 91 69 77 80 85 88 103 104 85

0.68 0.29 0.33 0.76 1.37 1.27 1.61 0.61 1.49 2.51 2.66 2.05 0.99 0.37 0.41

8.5 8.1 4.6 3.1 8.2 2.8 1.4 1.4 7.0 3.4 3.4 5.5 7.1 7.2 9.6

72 64 61 86 85 92 91 92 74 76 80 88 87 91 92 82

a The PCBs were recovered using ASE and Soxhlet extractions. (Two significant figures are valid.) b Soxhlet toluene data from Table 3.

of matrixes. Likewise it has been demonstrated that a second static step of 5 min increases recoveries of some PCB congeners by as much as 14% (19). The use of toluene in combination with a longer extraction time would likely result in an improved quantitative extraction of PCBs independent of the solid matrix investigated. Accordingly, before using ASE U.S. EPA Method 3545 on unknown sediments, extractions should be carried out to determine optimal solvent and extraction times for maximum recoveries. VOL. 34, NO. 23, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4999

Acknowledgments Peter Andersson at Metric Analys AB is gratefully acknowledged for the loan of the Dionex ASE 200 instrument. The authors also thank Carolyn Oldham for editing the manuscript.

Literature Cited (1) Erickson, M. D. Analytical Chemistry of PCBs, 2nd ed.; Lewis/ CRC Press: Boca Raton, FL, 1997. (2) Ahlborg, U. G.; Hanberg, A.; Kenne, K. Risk Assessment of Polychlorinated Biphenyls (PCBs); Report No. Nord 1992:26; Nordic Council of Ministers: 1992. (3) Andersson, O ¨ .; Linder, C.; Olsson, M.; Reutergårdh, L.; Uvemo, U.; Wideqvist, U. Arch. Environ. Contam. Toxicol. 1988, 17, 755765. (4) Kannan, K.; Falandysz, J.; Yamashita, N.; Tanabe, S.; Tatsukawa, R. Mar. Pollut. Bull. 1992, 24, 358-363. (5) Axelman, J.; Na¨f, C.; Bandh, C.; Ishaq, R.; Pettersen, H.; Zebu ¨ hr, Y.; Broman, D. Dynamics and distribution of hydrophobic organic compounds in the Baltic Sea. In A Systems Analysis of the Changing Baltic Sea; Springer-Verlag: Berlin, in press. (6) Agrell, C.; Larsson, P.; Okla, L.; Bremle, G.; Johansson, N.; Zelechowska, A. Atmospheric and riverine input of hydrophobic chlorinated organic pollutants to the Baltic Sea. In A Systems Analysis of the Changing Baltic Sea; Springer-Verlag: Berlin, in press. (7) Dean, J. R.; Barnabas, I. J.; Fowlis, I. A. Anal. Proc. 1995, 32, 305-308. (8) Poole, C. F.; Poole, S. K. Anal. Commun. 1996, 33, 11H-14H. (9) Wan, H. B.; Wong, M. K. J. Chromatogr. A 1996, 754, 43-47. (10) Hawthorne, S. B. Anal. Chem. 1990, 62, 633A-642A. (11) Bøwadt, S.; Hawthorne, S. B. J. Chromatogr. A 1995, 703, 549571. (12) Onuska, F. I.; Terry, K. A. Chromatographia 1993, 36, 191-194. (13) Richter, B. E.; Jones, B. A.; Ezzell, J. L.; Porter, N. L.; Avdalovic, N.; Pohl, C. Anal. Chem. 1996, 68, 1033-1039. (14) Ezzell, J. L.; Richter, B. E.; Felix, W. D.; Black, S. R.; Meilke, J. E. LC-GC 1995, 9, 233-240. (15) Richter, B. E.; Ezzell, J. L.; Felix, W. D.; Roberts, K. A.; Later, D. W. Am. Lab. 1995, 27, 24-28. (16) Jensen, D.; Ho¨fler, F.; Ezzell, J.; Richter, B. Polycyclic Aromatic Compd. 1996, 9, 233-240.

5000

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 23, 2000

(17) Bjo¨rklund, E.; Bøwadt, S.; Nilsson, T. TrAC, Trends Anal. Chem. 2000, 19, 434-445. (18) Heemken, O. P.; Theobald, N.; Wenclawiak, B. W. Anal. Chem. 1997, 69, 2171-2180. (19) Bjo¨rklund, E.; Bøwadt, S.; Nilsson, T.; Mathiasson, L. J. Chromatogr. A 1999, 836, 285-293. (20) Schantz, M. M.; Nichols, J. J.; Wise, S. A. Anal. Chem. 1997, 69, 4210-4219. (21) Zuloaga, O.; Etxebarria, N.; Fernandez, L. A.; Madariaga, J. M. TrAC, Trends Anal. Chem. 1998, 17, 642-647. (22) Poster, D. L.; Schantz. M. M.; Wise, S. A.; Vangel, M. G. Fresenius J. Anal. Chem. 1999, 363, 380-390. (23) U.S. EPA Method 3545, Pressurized Fluid Extraction. Test Methods for Evaluating Solid Waste, 3rd ed., update III; U.S. EPA SW-846; U.S. Government Printing Office: Washington, DC, July 1995. (24) Kenny, D. V.; Olesik, S. V. J. Chromatogr. Sci. 1998, 36, 59-65. (25) Kenny, D. V.; Olesik, S. V. J. Chromatogr. Sci. 1998, 36, 66-72. (26) Finkel, J. M.; James, R. H.; Baughman, K. W.; Pau, J. C.; Knoll, J. E.; Midgett, M. R. Chemosphere 1989, 19, 67-74. (27) Lamparski, L. L.; Nestrick, T. J. Chemosphere 1989, 19, 27-31. (28) Nilsson, U. Properties of some chlorinated polycyclic aromatic hydrocarbons with respect to chemical analysis. Thesis, Stockholm University, Stockholm, 1992; ISBN 91-7153-043-6,. (29) Zebu ¨ hr, Y.; Na¨f, C.; Bandh, C.; Broman, D.; Ishaq, R.; Pettersen, H. Chemosphere 1993, 27, 1211-1219. (30) Smith, L. M.; Stalling, D. L.; Johnson, J. L. Anal. Chem. 1989, 56, 1830-1842. (31) Bjo¨rklund, E.; Bøwadt, S.; Mathiasson, L.; Hawthorne, S. B. Environ. Sci. Technol. 1999, 33, 2193-2203. (32) Hawthorne, S. B.; Bjo¨rklund, E.; Bøwadt, S.; Mathiasson, L. Environ. Sci. Technol. 1999, 33, 3152-3159. (33) Gustafsson, O ¨ .; Haghseta, F.; Chan, C.; MacFarlane, J.; Gschwend, P. M. Environ. Sci. Technol. 1997, 31, 203-209. (34) Popp, P.; Keil, P.; Mo¨der, M.; Paschke, A.; Thuss, U. J. Chromatogr. A 1997, 774, 203-211.

Received for review September 15, 1999. Revised manuscript received September 20, 2000. Accepted September 21, 2000. ES991064G