Microwave-Assisted Steam Distillation for Simple Determination of

Anal. Chem. 2003, 75, 1450-1457

Microwave-Assisted Steam Distillation for Simple Determination of Polychlorinated Biphenyls and Organochlorine Pesticides in Sediments Masahiko Numata,* Takashi Yarita, Yoshie Aoyagi, and Akiko Takatsu

National Metrology Institute of Japan, AIST, Tsukuba, Ibaraki 305-8563, Japan

A novel sample extraction technique for polychlorinated biphenyls (PCBs) and organochlorine pesticides (OCPs) analysis using microwave-heating device is developed. In this study, microwave-assisted extraction (MAE) and steam distillation techniques were combined. Desorption of the anatytes from solid matrixes was accelerated with water vapor which was generated by microwave irradiation. A sample holder in a commercial microwave extraction cell kept the sample from direct contact with the organic solvent for analytes trapping during the treatment process. Therefore, relatively clean extracts were obtained with small amount of solvents. Without any cleanup steps, the obtained extract could be analyzed with gas chromatograph/mass spectrometers (GC/MS). Six PCB congeners (PCB15, 28, 70, 101, 180, 194, 209) and three OCPs (γ-HCH, 4,4′-DDE, 4,4′-DDD) in two marine sediment samples (a sediment collected from a bay of Kyusyu Island, Japan, and a certified reference material NIST SRM1944) were analyzed by using this microwave-assisted steam distillation (MASD) technique and another extraction method (exhaustive steam distillation, MAE, and Soxhlet extraction); and comparisons of the results are shown in this report. Although recovery yields of highly chlorinated biphenyls (PCB180, 194, 209) and relatively polar OCPs (γ-HCH, 4,4′-DDD) were low (3060%) compared with other analytes (PCB15, 28, 70, 101, 4,4′-DDE; recovery, 80-100%), use of isotope labeled internal standards for the MASD technique gave comparable results with the values obtained by other extraction methods and the certified values in the samples. Polychlorinated biphenyls (PCBs) and organochlorine pesticides (OCPs) are representative persistent organic pollutants (POPs). Because of their volatility, persistence and lipophility, they are ubiquitous among environmental samples, such as air, water, soil, sediments, and biological tissues.1,2 Because some of them are suspected to be carcinogens or endocrine disrupters, determination of PCB and OCP concentrations is important to evaluate their risk. However, the analytical procedures for determination of them are usually tedious and time-consuming. * Corresponding author. Fax: +81-29-861-6865. E-mail: [email protected]. (1) Tanabe, S.; Iwata, H.; Tatsukawa, R. Sci. Total Environ. 1994, 154, 163177. (2) Wania, F.; Mackay, D. Environ. Sci. Technol. 1996, 30, 390A-396A.

1450 Analytical Chemistry, Vol. 75, No. 6, March 15, 2003

Improving extraction techniques for analysis of POPs in solid matrixes is widely investigated to minimize waste solvents and to shorten analytical procedures and time. Conventional Soxhlet extraction is one of the standard methods, but it is time-consuming and requires a large amount of solvent. Recently, many extraction techniques for environmental analysis have been developed. Although supercritical fluid extraction 3-6 and subcritical water extraction 5,7-9 are selective and basically organic solvent-free techniques, their sample throughput is low, and a relatively complex system is required. Pressurized fluid extraction (PFE),4-6,10,11 and microwave-assisted extraction (MAE)4,6,11-15 require a relatively short extraction time and small amount of solvent. Sample throughput by MAE is especially high, because multiple samples can be treated simultaneously. However, tedious cleanup procedures are necessary for determination of PCBs and OCPs in the extracts obtained by PFE or MAE in general because selectivity of the extraction methods is low. Although filtered extract is automatically obtained by Soxhlet extraction or PFE methods, the solid residue should be removed after the batch MAE process. One of the classical techniques of separating substances, steam distillation is useful for recovery of hydrophobic and volatilesemivolatile substances. The technique is also applicable to PCBs and OCPs analysis in water, soils, sediments, and foods.16-23 (3) Schantz, M. M.; Bøwaldt, S. B.; Benner, B. A., Jr.; Wise, S. A.; Hawthorne, S. B. J. Chromatogr., A 1998, 816, 213-220. (4) Dean, J. R.; Xiong, G. Trends Anal. Chem. 2000, 19, 553-564. (5) Hawthorne, S. B.; Grabanski, C. B.; Martin, E.; Miller, D. J. J. Chromatogr., A 2000, 892, 421-433. (6) Camel, V. Analyst 2001, 126, 1182-1193. (7) Yang, Y.; Bøwaldt, S. B.; Hawthorne, S. B.; Miller, D. J. Anal. Chem. 1995, 67, 4571-4576. (8) Hartonen, K.; Inkala, K.; Kangas, M.; Riekkola, M.-L. J. Chromatogr., A 1997, 785, 219-226. (9) Hawthorne, S. B.; Grabanski, C. B.; Martin, E.; Miller, D. J. J. Chromatogr., A 1998, 814, 151-160. (10) Schantz, M. M.; Nichols, J. J.; Wise, S. A. Anal. Chem. 1997, 69, 42104219. (11) Martens, D.; Gfrerer, M.; Wenzl, T.; Zhang, A.; Gawlik, B. M.; Schramm, K.-W.; Lankmayr, E.; Kettrup, A. Anal. Bioanal. Chem. 2002, 372, 562568. (12) Lopez-Avila, V.; Young, R.; Benedicto, J. Anal. Chem. 1994, 66, 1097-1106. (13) Eskilsson, C. S.; Bjo ¨rklund, E. J. Chromatogr., A 2000, 902, 227-250. (14) Carro, N.; Saavedra, Y.; Garcıá, I.; Llompart, M. J. Microcolumn Sep. 1999, 11, 544-549. (15) Carro, N.; Garcıá, I.; Llompart, M. Analusis 2000, 28, 720-724. (16) Veith, G.D.; Kiwus, L. M. Bull. Environ. Contam. Toxicol. 1977, 17, 631636. (17) Peter, T. L. Anal. Chem. 1980, 52, 211-213. (18) Richter, E.; Renner, G.; Bayerl, J.; Wick, M. Chemosphere 1981, 10, 779784. 10.1021/ac0262513 CCC: $25.00

© 2003 American Chemical Society Published on Web 02/13/2003

Table 1. Compositions of the Surrogate Solution and the GC/MS Calibration Solutionsa

PCB15 PCB28 PCB70 PCB101 PCB180 PCB194 PCB209 PCB170 γ-HCH 4,4′-DDE 4,4′-DDD

surrogate solution isotope labeled

calibration solution A native

calibration solution B native

calibration solution C native

calibration solutionb A,B,C isotope labeled

16.6 320 540 267 63.3 12.1 6.98

7.29 149 254 120 30.8 5.60 3.24

15.2 310 530 251 64.1 11.7 6.76

22.9 467 800 378 96.7 17.6 10.2

45.4 37.7 109

22.9 18.0 53.1

46.0 37.5 111

69.0 56.5 167

16.6 320 540 267 63.3 12.1 6.98 136 45.4 37.7 109

a Concentrations, ng/g the shown values.

b

Actual concentrations of the isotope labeled compounds in the three calibration solutions were 1-3% different from

Because mainly only volatile and hydrophobic compounds are recovered, in some cases, cleanup is unnecessary to analyze the obtained extract using gas chromatography. However, the application of the technique for environmental analysis is limited because of its low sample throughput. The purpose of this study is evaluation of a novel microwaveassisted steam distillation (MASD) technique, a combination of MAE and steam distillation methods. We have developed an extraction system for this purpose. Wet sediment samples were heated by microwave irradiation, and the analytes (PCBs and OCPs) were desorbed from the matrix with water vapor. The condensed analytes were trapped in a small amount of nonpolar organic solvent, and then the relatively clean extracts were analyzed with GC/MS without any clean up procedure. The MASD technique was applied for the determination of PCBs and OCPs in sediments, and the effects of the extraction conditions on efficiency and a comparison between this method and other techniques, such as exhaustive steam distillation, MAE, and Soxhlet extraction are presented in this report. EXPERIMENTAL SECTION Samples and Reagents. Two marine sediment samples were used in this study. Sediment D was collected from a bay of Kyusyu Island as a candidate for Japanese reference material for environmental analysis. The sediment was air-dried, pulverized, sieved (99%, Cambridge Isotope Laboratories; γ-HCH, >99%, Wako Pure Chemical Industries, Japan; 4,4′-DDE: neat, Supelco; 4,4′-DDD, >99%,;Dr. Ehrenstorfer, Germany) and the PCB-[13C12] solutions described above. Compositions of the calibration solutions (three concentration levels of native compounds) and the surrogates solution are shown in Table 1. Pesticide analysis grade solvents (acetone, toluene, and hexane, Kanto Chemical, Japan) and reagent grade solvents (2,2,4-trimethylpentane, Kanto Chemical; octane and heptane, Wako Pure Chemical Industries, Japan) were used for the extraction procedures. Microwave-Assisted Steam Distillation. The apparatus used in this study is shown in Figure 1. A sintered glass filter (thickness, 3 mm) attached on the bottom of a glass tube (50 mm × 20 mm i.d.) was covered with a piece of filter paper (diameter, 20 mm). The sediment (1.25 g of sediment D or 0.5 g of SRM1944) was weighted in the glass tube. The tube was put into a Teflon PFA extraction cell (∼140 mm × 32 mm i.d., GreenChem Plus, CEM) containing 3.0 mL of water. After the sample soaked up the water, the surrogates solution (135 mg of 2,2,4-trimethylpentane solution) and 10 mL of a nonpolar solvent, such as hexane, heptane, octane, 2,2,4-trimethylpentane, and toluene, was added to the glass tube. The cells were covered by a pressure-resistant holder and were heated using a microwave at 110-170 °C for 10-90 min using a microwave extraction system MARSX (CEM). Three to six vessels were used for each extraction experiment in this study. The temperature was monitored and controlled with a fiber-optic probe inside one of the cells. Upon the termination of the microwave irradiation, the cells were aircooled. After cooling, the extraction cells were opened, and the glass tubes placed inside were removed. A small amount of aqueous layer that remained in the PFA extraction cell was removed with a pipet. To dry the organic layer, 1.0 g of anhydrous sodium sulfate was added to the cell, then the organic solvent layer was recovered with a pipet, and the inner wall of the cell was rinsed with a small amount of hexane to recover the whole Analytical Chemistry, Vol. 75, No. 6, March 15, 2003

1451

Figure 1. Principle of the microwave assisted steam distillation.

extract. The combined extract was analyzed by GC/MS without any cleanup procedure. Exhaustive Steam Distillation. The apparatus used in this study was modified from the method of Veith et al.16 The sediment (1.25 g of sediment D or 0.5 g of SRM1944) was weighed into a flask. After addition of the surrogates solution (135 mg of 2,2,4trimethylpenane solution), the flask was evacuated to remove the solvent. Water (25 mL) and octane (2.0 mL) were added to the flask. Then a “J-shape” glass tube (solvent reservoir, ∼55 mm height × 12 mm i.d.; lower curved stem, 2 mm i.d.) was supported inside the neck of the flask with a stainless steal wire, and a Liebig condenser was installed on the top of the flask. The flask was heated in an oil bath (115-120 °C) for 5 h to maintain reflux. The condensed water, octane, and analytes were collected in the inner glass tube, and only water dropped through the thinner tube. After cooling, the organic solvent layer was recovered and then dried with 0.1 g of anhydrous sodium sulfate. Microwave-Assisted Extraction with Water and Nonpolar Solvents. After addition of the surrogates solution (270 mg), the sediment (2.5 g of sediment D or 0.8 g of SRM1944) was extracted with the MAE system with 20 mL of octane and 3.0 mL of water for 10 min at 150 °C. The extract from MAE was centrifuged at 3000 rpm for 3 min to remove residues. The octane layer was recovered and dried with 1.0 g of anhydrous sodium sulfate. The extract was passed through a solid-phase extraction (SPE) cartridge (500 mg of silica, International Solvent Technology, U.K.; precleaned with 10 mL of hexane) to remove polar constituents, and then the cartridge was washed with 7 mL of hexane to recover the PCBs. Microwave-Assisted Extraction with Polar Solvent. After addition of the surrogates solution (270 mg), the sediment (2.5 g of sediment D or 0.8 g of SRM1944) was extracted with 20 mL of hexane/acetone (1:1) for 10 min at 130 °C using the MAE system. The extract was centrifuged at 3000 rpm for 3 min to remove the solid residues. The obtained supernatant was cleaned up by the method of Schantz et al.10 with some modifications. The supernatant was transferred to a glass vial, and ∼10 g of activated copper powder and 2.0 g of anhydrous sodium sulfate were added. The mixture was shaken for 10 min in order to remove elemental sulfur and water. After filtration with a PTFE membrane filter (pore size, 1452

Analytical Chemistry, Vol. 75, No. 6, March 15, 2003

0.2 µm), the solution was passed through the SPE silica cartridge, and then the PCBs and OCPs were recovered with 15 mL of a dichloromethane/hexane mixture (1:9). The obtained fraction was concentrated and then loaded onto a normal phase liquid chromatograph (Gulliver system, Jasco, Japan; equipped with YMCPack NH2 column, 150 mm × 10 mm i.d., YMC, Japan). The PCBs and OCPs were eluted with n-hexane and dichloromethane (012 min, 100% of hexane, 12-30 min, a linear gradient of 0-9% dichlorometane in hexane) at a flow rate of 4.0 mL/min. For GC/ HRMS analysis, the collected two fractions (3-13 min fraction, PCBs and 4,4′-DDE; 20-25 min fraction, more polar OCPs [4,4′DDD and γ-HCH]) were mixed. Soxhlet Extraction. The mixture of the sediment (2.5 g of sediment D or 0.8 g of SRM1944) and 10 g of anhydrous sodium sulfate was placed into a filter paper thimble, and 270 mg of the surrogates solution was added. The sample was loaded onto an automated Soxhlet extraction system B-811 (BU ¨ CHI, Switzerland) and then extracted with 250 mL of hexane/acetone (1:1) for 12 h. The extract was cleaned up by SPE and HPLC as described above. Determination of PCBs and OCPs by GC/MS. Analyses of PCB congeners and OCPs in the extracts were performed using a gas chromatograph-high-resolution mass spectrometer (GC/ HRMS) system (AutoSpec, Micromass, U.K.). After 50 µL of the syringe spike solution (including 20 ng of PCB170[13C12]) was added to the extract (or cleaned fraction) obtained by each method, the solution volume was reduced to 0.2 mL by means of a rotary evaporator and nitrogen gas stream. The analytes were separated using a gas chromatograph (Agilent 6890, Agilent Technologies) equipped with an HT-8 capillary column (50 m × 0.22 mm i.d., film thickness 0.25 µm, SGE, Australia). A portion of the solution (1.0 µL) was injected in splitless mode. The injection port was at 200 °C, and the column was held at 60 °C for 2 min, programmed at 30 °C/min to 170 °C, then ramped at 3.5 °C/min to 300 °C and held for 6 min. Helium gas was used as carrier gas at a pressure of 276 kPa. The resolution of the mass spectrometer was 10 000, and the PCB signals were monitored in SIM mode (lock mass, PFK). Analyses under lower mass resolution (3000 and 1000) and by using a gas chromatograph/ quadropole mass spectrometer (GC/QMS) system (Agilent 6890/

Table 2. Quantification Limits and Experimental Blank of the Analytesa quantification limit (ng/cell)

PCB15 PCB28 PCB70 PCB101 PCB180 PCB194 PCB209 γ-HCH 4,4′-DDE 4,4′-DDD a

GC/HRMS (R ) 10000)

GC/HRMS (R ) 3000)

GC/HRMS (R ) 1000)

GC/QMS

exptl blank (ng/cell)

0.02 0.03 0.02 0.02 0.04 0.09 0.06 0.02 0.02 0.13

0.01 0.05 0.01 0.01 0.02 0.04 0.02 0.01 0.05 0.04

0.9 0.2 0.04 0.2 0.05 0.05 0.07 5 0.2 0.2

0.6 1.5 0.7 0.7 0.8 0.7 1.2 1.8 0.1 0.8

ndb 0.1 0.3 0.1 nd nd nd nd nd nd

Quantification limit, S/N ) 10

b

nd, lower than detection limits (S/N ) 3, R ) 10 000)

Figure 2. Comparative studies on PCB concentrations (A) and recoveries (B) using different analyte-trapping solvents for the MASD process: isooctane, 2,2,4-trimethylpentane.

5973N MSD, Agilent Technologies were also performed to compare analytical results. Quantification limits of the analytes and the MASD experimental blank (analysis was performed without sample) are summarized in Table 2. Analytical results are represented as dry-mass base in this report. Moisture contents of the samples were determined gravimetrically. The samples were dried at 105 °C for >24 h, and the moisture contents were calculated from weighting before and after drying (sediment D, 4.7%, SRM1944, 1.8%). Safety Consideration. Because the treatment temperature is monitored and controlled with a temperature probe inside the control cell, the content of the control cell should be the same as contents of other extraction cells. If the contents are different, the temperature in the extraction cells would not be controlled. Moreover, the temperature and pressure in the extraction cells could exceed their safety limits. RESULTS AND DISCUSSION The Extraction Device. Figure 1 shows the principle of the MASD system. When a nonpolar organic solvent was used for MASD, the moist sediment inside the glass tube was directly heated by microwave, because microwave penetrates the nonpolar solvent. It was supposed that water vapor accelerated vaporization of the solvent and volatile compounds, including PCBs and OCPs adsorbed in the sediment. Then the vapor was condensed on the

inner surface of the extraction cell. The condensed water permeated into the sintered glass filter and the sample by capillary action. As a result, the water in the sample was heated by microwave and took part in the extraction of the volatile compounds repeatedly. Through this process, PCB and OCP molecules carried by water vapor were trapped in the condensed organic solvent without direct contact between the solvent and the sediment. Investigations of the Extraction Conditions. Effects of solvents, water amount in the extraction cell, extraction temperature, and time on the PCB analytical results were investigated in this study. Effects of the extraction solvents on the PCB analysis in the sediment D are shown in Figure 2. Six kinds of solvents (aliphatic hydrocarbon: hexane, heptane, octane, nonane, and 2,2,4-trimethylpentane [isooctane]; aromatic hydrocarbon: toluene) were used as the analyte trapping solvent of the MASD (150 °C, 30 min), and the analytical results of the sediment D obtained by GC/HRMS measurement (resolution, 10 000) were compared. Observed concentration values in Figure 2A normalized to the values obtained by octane extraction are shown as the relative values, and values in Figure 2 are shown as mean values of triplicate measurements (error bars, SD). Although the calculated concentration values were not so fractionated depending on solvent (Figure 2A), the recovery yields were sensitive to the Analytical Chemistry, Vol. 75, No. 6, March 15, 2003

1453

Figure 3. PCB concentrations (A) and recoveries (B) as a function of the MASD process temperature. Extraction time, 30 min.

Figure 4. PCB concentrations (A) and recoveries (B) as a function of the MASD process time. Extraction temperature, 150 °C.

solvent used (Figure 2B). The low recovery with toluene might be caused by the microwave absorption in toluene. Although the reason for the difference in the yields was not clear in other cases, octane gave the highest recovery yield among the tested solvents. Water and solvent amounts were mainly constrained with the structure of the extraction vessel. The level of the solvent should not exceed the upper rim of the glass tube. To prevent direct contact between the sample and the organic solvent, water should saturate the opening in the sample and filter. On the other hand, an excess amount of water (>5 mL) made the filter paper lift during microwave irradiation. As a result, fine fraction of the samples passed through the glass filter. A combination of much water (10 mL) and octane (3 mL) gave a yellowish extract and much lower PCB analytical values (data not shown). It was supposed that water outside of the glass tube would block microwave irradiation of the sample and prevent water circulation. Practically, the range of the amount of water was restricted to 3-4 mL in the case of the extraction vessel and sample amount (0.5-1.2 g) used in this study. Analytical results of PCBs were not significantly different when 3 and 4 mL of water were used. Consequently, 3 mL of water and 10 mL of solvents were used for the other experiments. The effects of process temperature and time on the analytical results and recovery yields of the surrogate in sediment D are shown in Figures 3 and 4. Octane was used for MASD, and the analytical results of sediment D obtained by GC/HRMS measure1454


ment (resolution, 10 000) depending on treatment temperature (Figure 3) and time (Figure 4) were compared. In Figures 3 and 4, the observed concentrations are shown as mean values of triplicate extraction results (error bars, SD). When extraction temperature was 110 °C or lower, octane still remained inside the glass tube after heating, and the recovery yields of PCBs were very low. Although a high temperature (>170 °C) would be necessary to recover highly chlorine-substituted PCB congeners completely, the observed values of concentration of most PCB congeners reached a plateau under 150 °C. The observed values of concentration and the recovery yield of most PCB congeners reached a plateau within 60 min. The required treatment time was longer compared with the treatment time of MAE for quantitative recoveries (see, e.g., ref 12). This result would indicate that recoveries of analytes were not driven by direct solvent extraction but by steam distillation. As following these experimental results, all subsequent sediment extractions were performed at 150 °C for 60 min using octane. Because the dielectric constant (i.e., polarity) of water decreases as the temperature is increased, the solubility of nonpolar organic compounds in the subcritical water is higher than in the normal water. This phenomenon has been also applied as a selective extraction technique of PCBs and other organic pollutants.5,7-9 However, the extraction temperature used in this study was not high enough to extract PCB congeners with liquid water quantitatively (typically 200-250 °C). Therefore, the PCBs

Table 3. Comparative Study of PCB and OCP Concentrations and Recoveries Using Different Extraction Methods (Sample, Sediment D) MASD (150 °C, 60 min, n ) 5) observed values (ng/g)

recovery

MAE (130 °C, 10 min, n ) 3)

yieldsa

(%)

observed values (ng/g)

recovery

Soxhlet (12 h, n ) 3)

yieldsa

(%)


analytes

mean

SD

mean

SD

mean

SD

mean

SD

mean

SD

PCB15 PCB28 PCB70 PCB101 PCB180 PCB194 PCB209 γ-HCH4,4′-DDE 4,4′-DDD

2.38 31.1 55.8 27.9 7.44 1.67 1.05 4.15 6.14 13.1

0.23 1.9 2.4 1.1 1.12 0.27 0.24 0.51 0.74 1.5

73 88 81 84 63 37 32 41 70 43

11 5 8 11 19 9 6 9 11 21

2.29 34.2 61.2 30.4 8.27 1.82 1.18 5.02 5.85 12.4

0.05 0.41 0.94 0.5 0.57 0.19 0.07 0.34 0.01 0.63

75 71 70 76 82 83 85 65 79 57

1 1 2 2 5 9 11 13 3 11

2.19 35.4 61.5 33.4 9.69 1.85 1.15 4.48 6.14 12.7

0.06 1.1 1.4 1.4 0.19 0.05 0.13 0.29 0.49 0.6

a

Recovery yields of the surrogates through extraction and clean up processes.

Table 4. Comparative Study of PCB and OCP Concentrations and Recoveries Using Different Extraction Methods (Sample, NIST SRM1944) MASD (150 °C, 60 min, n ) 5) observed values (ng/g)

recovery

MAE (130 °C, 10 min, n ) 3)

yieldsa (%)

recovery yieldsa (%)


analytes

mean

SD

mean

SD

mean

SD

mean

SD

PCB15 PCB28 PCB70 PCB101 PCB180 PCB194 PCB209 γ-HCH 4,4′-DDE 4,4′-DDD

29.3 83.8 74.6 62.2 41.3 10.5 7.50 0.17 85.2 140

1.2 4.1 2.9 2.5 1.8 0.5 0.56 0.04 4.0 6

79 99 99 103 59 34 26 45 99 38

4 11 8 9 7 5 4 8 23 9

28.2 84.1 71.6 61.6 40.2 10.4 7.45 0.21 78.9 123

0.8 3.4 1.8 2.2 3.4 0.9 0.18 0.21 2.2 6

63 66 75 79 80 71 63 43 83 54

7 10 6 5 6 11 12 10 3 21

certified valuesb (ng/g) mean of means

expanded uncertainties

80.8

2.7

73.4c 44.3 11.2 6.81

2.5 1.2 1.4 0.33

86 119

12 11

a Recovery yields of the surrogates through extraction and clean up processes. b From the Certificate of Analysis.24 Extraction methods, Soxhlet extraction and pressurized fluid extraction. c The certified value reported for PCB101 included the chromatographic interference PCB90.

Table 5. Comparative Study on PCB Concentrations and Recoveries Using Different Extraction Methods with Water and Octane (Sample, marine sediment D) MASD (150°C, 60 min, n ) 5) observed values (ng/g)

recovery (%)

analytes

mean

SD

mean

PCB15 PCB28 PCB70 PCB101 PCB180 PCB194 PCB209

2.38 31.1 55.8 27.9 7.44 1.67 1.05

0.23 1.9 2.4 1.1 1.12 0.27 0.24

73 88 81 84 63 37 32

a

ESD (∼100°C, 5 h, n ) 3)

yieldsa

MAE (150°C, 30 min, n ) 3)

yieldsa


recovery (%)

SD

mean

SD

mean

11 5 8 11 19 9 6

2.17 32.6 56.9 29.5 7.33 1.75 1.15

0.06 0.3 1.7 0.4 0.32 0.08 0.08

65 74 80 81 45 16 10



SD

mean

SD

mean

SD

5 5 2 7 32 14 9

2.45 31.7 58.2 28.7 8.34 1.93 1.06

0.08 0.4 0.5 0.4 0.24 0.09 0.03

98 103 103 114 109 71 61

10 11 7 10 2 2 5


and OCPs may be mainly released with the water vapor under the MASD experimental conditions. Method Comparison. To evaluate feasibility of the established MASD technique, the analytical results were compared with the results obtained by the other methods (Tables 3-6, measured with the GC/HRMS, mass resolution ) 10 000). As shown in Tables 3 and 4, the analytical results obtained by MASD agreed with the results obtained by MAE and Soxhlet extraction using hexane/

acetone (1:1) and the certified values.24 Although the recovery yields of highly chlorinated PCB congeners and relatively polar OCPs (γ-HCH and 4,4′-DDD) by the MASD were lower than the yields by the MAE, the analytical values obtained by isotope dilution were in good agreement with the other methods. To recover analytes quantitatively, use of a polar solvent for extraction or saponification of the sample before extraction is effective. The disadvantage of those extraction methods is their low selectivity. Analytical Chemistry, Vol. 75, No. 6, March 15, 2003

1455

Table 6. Comparative Study of PCB Concentrations and Recoveries Using Different Extraction Methods with Water and Octane (Sample, NIST SRM1944) MASD (150°C, 60 min, n ) 5) observed values (ng/g)

recovery (%)

compound

mean

SD

mean

PCB15 PCB28 PCB70 PCB101 PCB180 PCB194 PCB209

29.3 83.8 74.6 62.2 41.3 10.5 7.50

1.2 4.1 2.9 2.5 1.8 0.5 0.56

79 99 99 103 59 34 26

a

ESD (∼100°C, 5 h, n ) 3)

yieldsa

MAE (150°C, 30 min, n ) 3)

yieldsa


recovery (%)

SD

mean

SD

mean

4 11 8 9 7 5 4

24.3 79.8 69.1 60.7 39.5 11.0 7.81

0.7 0.9 2.4 2.3 1.2 0.2 0.37

116 110 100 97 42 21 21



SD

mean

SD

mean

SD

19 12 18 22 38 25 23

28.7 83.4 75.9 61.3 38.0 9.60 7.61

0.9 1.2 3.7 1.7 0.7 0.82 0.44

124 123 110 115 98 82 66

6 7 5 5 2 2 6


Use of a polar solvent or direct contact between solvent and sample causes solubilization of polar constituents with the target analytes. The MAE and Soxhlet extraction method using a polar solvent gave yellow- or brown-colored extracts, and the extract had to be treated by SPE and normal phase HPLC to remove possible interference on quantification by GC. On the contrary, an almost colorless organic solvent phase was recovered by MASD technique, and no interference was observed in the case of determination of PCBs and OCPs by GC/HRMS without any cleanup procedure. Since the sample holder (glass tube and glass filter) kept the sample from direct contact with the organic solvent and almost exclusively volatile compounds were recovered from samples, relatively clean extracts were obtained in the case of MASD. Continuous extraction with refreshed solvent was also observed in the focused microwave-assisted Soxhlet extraction technique (FMASE).25-27 Although FMASE needs relatively short periods of time for extraction, the extraction selectivity would not differ from the conventional Soxhlet extraction. In addition, the FMASE needs an electrical heater for solvent circulation in addition to the microwave irradiation for the rupture of the anlayte-matrix bonds. In the case of the MASD, the circulation of the solvent was directly driven by the microwave energy, and almost exclusively volatile compounds were collected in the nonpolar solvent. Tables 5 and 6 show a comparison of the analytical results obtained by MASD, steam distillation, and MAE using water and octane. In these cases, water (vapor) worked as the carrier of the analytes. In addition, use of nonpolar a solvent is a means for selective extraction. One of the classical extraction processes, steam distillation has been used as a method that requires no cleanup for PCB determination in water or soils.16 In this study, a modified method of exhaustive steam distillation was applied for determination of PCB congeners. As reported, an almost colorless extract was obtained by the steam distillation technique, and the extract could be measured with GC/HRMS without any cleanup procedure. However, the recovery yields of highly (25) Garcıá-Ayuso, L. E.; Sańchez, M.; Fernańdez de Alba, A.; Luque de Castro, M. D. Anal. Chem. 1998, 70, 2426-2431. (26) Garcıá-Ayuso, L. E.; Luque-Garcıá, J. L.; Luque de Castro, M. D. Anal. Chem. 2000, 72, 3627-3634. (27) Luque-Garcıá, J. L.; Luque de Castro, M. D. Anal. Chem. 2001, 73, 59035908.

1456


chlorine-substituted PCB congeners were low, even after 5 h of the extraction process. For MAE using nonpolar solvent, addition of microwave absorbing materials ( Carboflon disk, CEM; Weflon disk, MLS, Germany) or water is necessary, because nonpolar solvents do not absorb microwave energy.6,28 Although a relatively clean extract was obtained by using nonpolar solvents, as compared with the extract obtained by MAE using a polar solvent, the yellowish extract should be cleaned with the silica SPE cartridge before GC/MS analysis. Obviously, the MASD technique surpassed the advantages of MAE and steam distillation. The obtained extract was clean like the extract obtained by the exhaustive steam extraction, whereas higher treatment temperature compared with conventional steam distillation brought about a relatively short extraction time. In addition, high sample throughput would be realized by simultaneous sample treatment with microwave irradiation in an oven and no need for solid residue removal after extraction. Effect of Mass Resolution on GC/MS Analytical Results. Most of the data in this study were obtained by using GC/HRMS (mass resolution, 10 000). Although high-mass resolution is effective for minimizing possible interference by concomitants, instruments are expensive and adjustment for instrument tuning is tough work in general. Therefore, an application of lower mass resolution or use of GC/QMS has the advantage for environmental analysis, because of lower cost and easier maintenance. To estimate feasibility of the MASD application for simpler analysis, the effect of mass resolution on PCB analytical results was investigated. For this purpose, octane was used as the analytetrapping solvent of the MASD (150 °C, 60 min), and the analytical results of sediment D and NIST SRM1944, depending on mass resolution of GC/MS, were compared (Figure 5). Observed concentration values in Figure 5 normalized to the values obtained by the mass resolution of 10 000 measurement were shown as the relative values, and the values were shown as mean values obtained by triplicate extractions (error bars, SD). The analytical values obtained by the mass resolution of 10 000, 3000 and 1000 had no significant differences. A resolution of 1000 gave lower precision because of higher noise on the chromatogram. The precision of the data obtained under lower mass resolution conditions (mass resolution, 1000 or GC/QMS) was lower. In the case of GC/QMS analysis, high background noise was observed (28) Du ¨ ring, R.-A.; Ga¨th, St. Fresenius’ J. Anal. Chem. 2000, 368, 684-688.

Figure 5. Comparative studies of PCB concentrations using different mass-resolution GC/MS for analyses of the MASD extracts. R, mass resolution. +Cu, the extract was shaken with Cu powder to remove sulfur. Table 7. Approximate Organic Solvents and Time Required for Extraction and Cleanup Steps of the Techniques Used in This Study step

MASD

steam distillation

extraction solid-liquid separation solid-phase extraction HPLC othersa total

10 mL, 1 h

2 mL, 5 h

a

5 mL, 2 h 15 mL, 3 h

2 mL, 1 h 4 mL, 6 h

MAE octane/H2O 10 mL, 0.5 h 0 mL, 0.5 h 20 mL, 1 h 10 mL, 1 h 40 mL, 4 h

MAE acetone/hexane 20 mL, 0.2 h 0 mL, 0.5 h 20 mL, 1 h 200 mL, 1 h 20 mL, 2 h 260 mL, 5 h

Soxhlet extraction 300 mL, 12 h 20 mL, 1 h 200 mL, 1 h 20 mL, 2 h 540 mL, 17 h

Rinse, concentration, etc.

on the chromatogram of trichlorinated biphenyls, including PCB28. Because the interference was removed with activated copper powder, it probably was caused by contamination of elemental sulfur. A signal noise ratio of the PCB209 peak on the chromatogram obtained with the GC-/QMS was too low to qualify. However, most of PCB congener peak patterns on the chromatogram were similar to each other, and the analytical values obtained by each mass resolution had no significant differences. Thus, MASD would be a pretreatment technique for determination of PCBs and OCPs by using low-mass-resolution GC/MS equipment, including GC/QMS. CONCLUSIONS Because of a low quantity of solvent use, relatively greater sample throughput, and the lack of necessity of solid-liquid

separation and cleanup process after extraction, this developed MASD technique has the advantage over simple MAE, conventional steam distillation, and Soxhlet extraction methods (Table 7). Possible analytical errors caused by the low recovery yields of highly chlorine substituted PCB congeners can be corrected by using isotope-labeled surrogates. Thus, the MASD method provides a simple technique for determination of PCBs and OCPs in sediment samples.

Received for review October 21, 2002. Accepted January 15, 2003. AC0262513


1457

Microwave-Assisted Steam Distillation for Simple Determination of

Recommend Documents