Anal. Chem. 2000, 72, 3647-3652
Determination of Distribution Coefficients of Priority Polycyclic Aromatic Hydrocarbons Using Solid-Phase Microextraction Ruey-an Doong* and Sue-min Chang
Department of Nuclear Science, National Tsing Hua University, Hsinchu, 300, Taiwan
The determination of distribution coefficients is important for prediction of the chemical pathways of organic compounds in the environment. Solid-phase microextraction (SPME) is a convenient and effective method to measure the distribution of chemicals in a two-phase system. In the present study, the SPME distribution coefficient (Kspme) of 16 priority aromatic hydrocarbons (PAHs) was determined with 100-µm poly(dimethylsiloxane) (PDMS) and 85-µm polyacrylate (PA) fibers. The partition coefficients and LeBas molar volumes were used to describe the linearity of the log Kspme values of PAHs. Also, the validation of the distribution coefficient was examined using different sample volumes. The extraction time was dependent on the types of PAHs, and 20 min to 60 h was needed to reach equilibrium. The determined log Kspme values ranged from 3.02 to 5.69 and from 3.37 to 5.62 for 100-µm PDMS and 85-µm PA fibers, respectively. Higher Kspme values of low-ring PAHs were observed using 85-µm PA fiber. Good linear relationships between log Kow and log Kspme for PAHs from naphthalene to benzo[a]pyrene and from naphthalene to chrysene for 100-µm PDMS and 85-µm PA fibers, respectively, were obtained. The correlation coefficients were 0.969 and 0.967, respectively. The linear relationship between log Kspme and the LeBas molar volume was only up to benz[a]anthracene for 85-µm PA fiber and up to chrysene for 100µm PDMS fiber. Moreover, the effect of sample volume can be predicted using the partition coefficient theory and excellent agreement was obtained between the experimental and theoretical absorbed amounts of low-ring PAHs. This result shows that the determined log Kspme is more accurate than the previous method for estimating analytes with log Kow < 6 as well as for predicting the partitioning behaviors between SPME fiber and water. The physicochemical data of xenobiotic compounds are usually used to predict their environmental fate and effects and that of compounds with similar physicochemical properties. Hydorphobicity is one of the most important parameters governing the distribution behavior of xenobiotics in the environment. The octanol-water partition coefficient (Kow) is widely accepted as the * Corresponding author: (mail) 101, sec. 2, Kuang-Fu Rd., Department of Nuclear Science, National Tsing Hua University, Hsinchu, 300, Taiwan; (phone) 886-3-5726785; (fax) 886-3-5718649; (e-mail)
[email protected]. 10.1021/ac000040l CCC: $19.00 Published on Web 06/30/2000
© 2000 American Chemical Society
best two-phase system to model the partitioning of analytes between water and organic phases.1-3 Also, measuring the distribution coefficient in a two-phase system is necessary for studying chemical pathways in diverse fields such as environmental processes, food quality, and biological metabolism. Currently, two experimental systems, the slow-stirring and shakeflask methods, are used for the determination of Kow values of xenobiotics.4-7 The slow-stirring method has been shown to be applicable for the determination of Kow values of very high hydrophobic compounds. Slow stirring prevents octanol microemulsion formation in the water phase. This is in contrast to the shake-flask method in which the presence of octanol in the water phase prevents the reliable determination of compounds with log Kow values higher than 4-5. Solid-phase microextraction (SPME) is a recently developed direct, solvent-free extraction method for the determination of analytes in the environment.8-11 The mechanism of SPME is based on the equilibrium between analyte concentration in the aqueous phase and that in the polymeric phase of the fiber. Therefore, the SPME method is an equilibrium method rather than an exhaustive method, such as liquid-liquid extraction (LLE) or solid-phase extraction (SPE). Equilibrium methods are more selective because they take full advantage of differences in the extracting-phase-matrix distribution constants to separate target analytes from interferences. Although the SPME method has been applied to the determination of organic chemicals in diverse environmental media, an additional benefit of the SPME approach (1) Mackay, D.; Shiu, W. Y.; Ma, K. C. Illustrated Handbook of Physical-Chemical Properties and Environmental Fate of Organic Chemicals. V. 2: Polynuclear Aromatic Hydrocarbons, Polychlorinated Dioxins, and Dibenzofurans; Lewis Publishers: Chelsea, MI, 1992. (2) Schwarzenbach, P. R.; Gschwend, P. M.; Imboden, D. M. Environmental Organic Chemistry; John Wiley & Sons: Toronto, ON, Canada, 1993; pp 124156. (3) Lyman, W. J.; Reehl, W. F.; Rosenblatt, D. H. Handbook of Chemical Property Estimation Methods: Environmental Behaviors of Organic Compounds; McGraw-Hill: Washington, DC, 1996; Chapter 1. (4) De Bruijn, J.; Busser, F.; Seinen, W.; Hermens, J. Environ. Toxicol. Chem. 1989, 8, 499. (5) Van Haelst, A. G.; Heesen, P. F.; Van der Wielen, F. W. N.; Govers, H. A. J. Chemosphere 1994, 29, 1651. (6) Sijm, D. T. H. M.; Sinnige, T. L. Chemosphere 1995, 31, 4427. (7) De Maagd, P. G. J.; Hulscher, D. E. M.; Van der Heuvel, H.; Opperhuizen, A.; Sijm, D. T. H. M. Environ. Toxicol. Chem. 1998, 17, 251. (8) Potter, D. W.; Pawliszyn, J. Environ. Sci. Technol. 1994, 28, 298. (9) Magdic, S.; Pawliszyn, J. J. Chromatogr., A 1996, 723, 111. (10) Zhang, Z.; Pawliszyn, J. Anal. Chem. 1993, 65, 1843-1852. (11) Gorecki, T.; Martos, P.; Pawliszyn, J. Anal. Chem. 1998, 70, 19.
Analytical Chemistry, Vol. 72, No. 15, August 1, 2000 3647
is to obtain distribution coefficients of analytes. Several studies have attempted to estimate the SPME distribution coefficient (Kspme) and establish the relationship between partition coefficient (log Kow) and log Kspme.12-16 However, predictions of the amount of compounds extracted are different and no straightforward relationship has been shown between partition coefficients and the characteristics of the analytes.17,18 Although the relationship has been established for analytes having low molecular weights, the correlation between the apparent SPME distribution coefficient (Kspme) and Kow values appears to fail as the molecular weights of the analytes increase. Previous work has demonstrated that SPME is only valid for estimating the compounds with log Kow < 3.5.13,14 Yang et al.15 demonstrated that surface adsorption rather than absorption is the primary mechanism controlling high molecular weight analytes partitioning from water into SPME sorbents. More recently, Gorecki et al.18 showed that the determination of the partition coefficient of semivolatile organic compounds between water and polymeric-coating fiber is hampered by long equilibrium time, microbial degradation, the adsorption of vial wall and the existence of organic solvent from standard solution. This means that the distribution of analytes using SPME techniques is difficult to be precisely predicted by Kow values. However, a number of experimental data on Kow were obtained using less accurate methods.7 Also, the SPME distribution coefficient of analytes between water and polymeric-coating fiber is an important parameter for understanding the sorption mechanism of SPME as well as for quantitatively determining the extracted amounts of analytes. Therefore, the determination of the Kspme values and the establishment of the relationship between Kow and Kspme for better understanding of the distribution behaviors of analytes using the SPME method are still in demand. In this study, we present the linear relationship between the distribution coefficients (Kspme) and Kow using the SPME method. To ensure broad applicability of the method, 16 priority polycyclic aromatic hydrocarbons (PAHs), which fall into moderate to extreme hydrophobicity classes with molecular weights of >128 and log Kow > 3.3, were selected as the target compounds. A 100µm poly(dimethylsiloxane) (PDMS) and an 85-µm polyacrylate (PA) fiber were used to determine the distribution coefficient and estimate the partition coefficient of 16 PAHs. The LeBas molar volume was used to describe the linearity of determined Kspme. Moreover, the Kspme was validated using the effect of sample volume by comparing the extracted amounts of PAHs between predicted and experimental data. EXPERIMENTAL SECTION Reagents and Materials. The standard mixtures of 16 PAHs at a concentration of 2000 µg/mL in methylene/benzene (1/1, v/v) were purchased from Supelco Co., Inc. (Bellefonte, PA). (12) Paschke, A.; Popp, P.; Schuurmann, G. Fresenius J. Anal. Chem. 1998, 360, 52. (13) Langenfeld, J. J.; Hawthorne, S. B.; Miller, D. J. Anal. Chem. 1996, 68, 144-155. (14) Dean, J. R.; Tomlinson, W. R.; Makovskaya, V.; Cumming, R.; Hetheridge, M.; Comber, M. Anal. Chem. 1996, 68, 130. (15) Yang, Y.; Hawrhorne, S. B.; Miller, D. J.; Liu, Y.; Lee, M. L. Anal. Chem. 1998, 70, 1866. (16) Poerschmann, J.; Kopinke, F. D.; Pawliszyn, J. J. Chromatogr., A. 1998, 816, 159. (17) Gorecki, T.; Pawliszyn, J. Analyst 1997, 122, 1079. (18) Gorecki, T.; Khaled, A.; Pawliszyn, J. Analyst 1998, 123, 2819.
3648
Analytical Chemistry, Vol. 72, No. 15, August 1, 2000
These standards were stored at 4 °C and used for the preparation of working standard solutions. A working standard (20 µg/mL in acetone) was prepared every week. Deionized water was obtained from a Milli-Q water purification system (Millipore, Bedford, MA). Methanol, dichloromethane, acetone, and n-hexane were obtained from the Mallinckrodt Co. (Phillipsburg, NJ). Toluene was purchased from Redial de Haen Co. (Seelze, Germany). The SPME holders for manual use were obtained from Supelco Co., Inc. Two different fibers, 100-µm PDMS and 85-µm PA, were also obtained from Supelco. All fibers were conditioned in the hot injector part of the gas chromatograph for 0.5-2 h and at 250280 °C, according to the instructions provided by the manufacturer. The glassware used in this study was first washed with detergent and then with deionized water followed by placement in a cleaning solution overnight to remove the trace amounts of organics on the surface of the vials. The glassware was then rinsed with deionized water, methanol, acetone, and hexane in sequence, dried in an oven at 105 °C, and wrapped with aluminum foil before use. For the SPME procedure, the vials were further silanized by soaking the glassware overnight in a 10% (v/v) mixture of dichlorodimethylsilane (Supelco) in toluene. Finally, the vials were rinsed with toluene and methanol and oven-dried at 105 °C. Apparatus and Analysis. A Hewlett-Packard 6890 gas chromatograph, with a split/splitless injection port and a flame ionization detector (FID), was used for the experiments to determine the optimized SPME conditions. The carrier gas was nitrogen with a flow rate of 3 mL/min. The detector flow rates were 350 mL/min for air, 35 mL/min for hydrogen, and 30 mL/ min for nitrogen (makeup gas). The injector was maintained between 270 and 300 °C, depending on the fiber used. The temperature of the detector was 350 °C. A 30-m Ultra Alloy-5 stainless steel capillary column (0.5-mm inner diameter, 0.5-µm film thickness, Quadrex Co., New Haven, CT) was used for separating PAHs. The column was held at 40 °C for 5 min, increased to 180 °C at a rate of 20 °C/min, and again ramped at 4 °C/min to 250 °C; the temperature was increased to 270 °C at a rate of 2 °C/min, held for 2 min, finally ramped to 320 °C at a rate of 10 °C/min, and then held for 10 min. Solid-Phase Microextraction Procedures. The SPME extractions were performed by placing 10 mL of deionized water in 15-mL amber vials capped with PTFE-coated septa. Aqueous standards were prepared by spiking an appropriate amount of the working standard to give a final concentration of 10 ng/mL. Magnetic stirring with a 1-cm-long Teflon-coated stir bar was used to agitate the solution at ∼1000 rpm. The SPME equilibrium was conducted by immersing the fiber into the aqueous phase of the sample with stirring at room temperature for an appropriate time period, during which analytes were absorbed onto the stationary phase of the fibers. After extraction, the fiber was thermally desorbed for 5 min into the glass liner of the GC injector port at 270 (100-µm PDMS fiber) or 300 °C (85-µm PA fiber). Possible carryover was removed by keeping the fiber in the injector for an additional time with the injector in the split mode. Reinserting the SPME fiber after the run did not show obvious carryover. Moreover, blanks were run periodically during the analysis to confirm the absence of contaminants.
Figure 1
RESULTS AND DISCUSSION Determination of Extraction Time. Since the mechanism of SPME is based on the equilibrium between analyte concentration in the aqueous phase and that in the polymeric phase of the fiber, time to equilibrium is needed for the determination of partition coefficients. Figure 1 illustrates the extraction profiles of 10 ng/ mL PAHs using 100-µm PDMS and 85-µm PA fibers at room temperature (25 ( 2 °C). The equilibrium times for the PAHs increased with molecular weights. The equilibrium time was 20 min for two-ring PAHs, 30-90 min for three-ring PAHs, 2-36 h for four-ring PAHs, and 60 h for five- and six-ring PAHs. Longer equilibrium time of PAH was observed when 85-µm PA fiber was used. The equilibrium time was 40-240 min for low-ring PAHs and 720-2400 min for four- and five-ring PAHs. No equilibrium
was observed for the six-ring PAHs (benzo[ghi]perlyene and indeno[1,2,3-cd]pyrene) in the extraction period of 80 h. PAHs are typically hydrophobic organic compounds and are expected to partition readily into a more nonpolar fiber coating rather than a polar one. The extraction efficiency increased with increasing molecular weight from naphthalene to pyrene and then decreased from benz[a]anthracene to benzo[ghi]perlyene. Although PA fiber is designed for the purpose of extracting polar compounds, similar absorption behaviors were observed. However, the PDMS fiber showed good absorption efficiency for PAHs rather than PA fiber. The high molecular weight PAHs (higher than four rings) were less effectively extracted by using PA fiber because PA fiber has a more polar coating and has low affinity to high-ring PAHs. Analytical Chemistry, Vol. 72, No. 15, August 1, 2000
3649
Table 1. Comparison of the Determined Distribution Coefficients (kspme) of PAHs Using 100-µm PDMS and 85-µm PA Fibers determined log Kspme value compounds naphthalene acenaphthylene acenaphthene fluorene phenanthrene anthracene fluoranthene pyrene benz[a]anthracene chrysene benzo[b]fluoranthene benzo[k]fluoranthene benzo[a]pyrene dibenzo[a,h]anthracene benzo[ghi]perylene indeno[1,2,3-cd]pyrene a
mol formula C10H8 C12H8 C12H10 C13H10 C14H10 C14H10 C16H10 C16H10 C18H12 C18H12 C20H12 C20H12 C20H12 C22H12 C20H14 C20H14
MW 128 152 154 166 178 178 202 202 228 228 252 252 252 276 278 278
1,3,7
log Kow
3.24-3.40 4.33 4.07 4.18 4.46-4.64 4.55-4.79 5.12-5.31 5.0-5.18 5.74-6.04 5.63-5.94 5.78 5.86-6.28 5.91-6.28 6.5 5.95-6.38 na
100-µm PDMS 3.02 3.40 3.63 3.71 3.96 3.98 4.71 4.86 5.26 5.69 5.17 5.33 5.39 4.86 4.28 4.43
published log Kspme value
85-µm PA
7-µm PDMS
100-µm PDMS
3.37 4.01 4.09 4.32 4.39 4.66 4.87 4.84 5.34 4.95 4.34 4.39 5.62 4.91 4.03 4.16
2.7313
3.01,8 2.8513 na na na 3.4113 3.14,13 4.108 4.1113 4.0713 3.83,13 4.958 3.9713 na na 3.47,13 4.868 na na na
naa na na 4.4213 3.9713 4.3813 4.4413 4.4613 4.7213 na na 4.2613 na na na
na, not available.
Distribution Constants (Kspme) with Different Fiber Coatings. The distribution coefficients of PAHs were determined using 100-µm PDMS and 85-µm PA fibers with an extraction time of 60 h. Table 1 compares the experimentally determined Kspme values of PAHs for 100-µm PDMS and 85-µm PA fibers. The log Kspme values increased from 3.02 for naphthalene to 5.69 for chrysene and then decreased to 4.28 for benzo[ghi]perylene when measured with 100-µm PDMS fiber. Higher Kspme values for low-ring PAHs than those determined with the 100-µm PDMS fiber were obtained when an 85-µm PA fiber was used. The Kspme values of low-ring PAHs observed using 85-µm fiber were 2-5 times higher than those obtained using the 100-µm PDMS fiber. However, the Kspme for five- and six-ring PAHs was much lower. Also, these determined Kspme values were compared with octanol-water partition coefficients (Kow) and previously published values of Kspme for the same compounds.7,8,13 Although no previous log Kspme have been reported for acenaphthylene, acenaphthene, fluorene, benzo[b]fluoranthene, benzo[k]fluoranthene, dibenzo[a,h]anthranthene, benzo[ghi]perylene, and indeno[1,2,3-cd]pyrene, more reasonable agreement was obtained between the Kspme values determined in this study and those determined previously and Kow values. Although the experimental Kspme value may be dependent on the fiber coating thickness and sorbent preparation method, results obtained in this study showed excellent linear relationships between log Kow and log Kspme for the analytes from naphthalene to chrysene (log Kow < 6). Figure 2 shows the relationship between log Kow and distribution coefficient determined in this study (log Kspme). The correlation coefficients (r) for 100-µm PDMS and 85-µm PA fiber were 0.969 and 0.967, respectively. Moreover, the linearity slope for a 100-µm PDMS system was close to unity (0.98), showing that the 100-µm PDMS fiber is more suitable for the determination of Kow values for PAHs. However, a negative correlation for five- and six-ring PAHs was also found. Yang et al. reported that high molecular weight compounds may adsorb onto the surface of polymeric fiber rather than absorb into the polymeric phase.15 Low diffusion coefficients of high molecular weight compounds may be the primary possibility for this discrepancy. The diffusivities of high molecular weight PAHs in 3650
Analytical Chemistry, Vol. 72, No. 15, August 1, 2000
Figure 2
an aqueous phase are in the range from 10-5 to 10-6 cm2/s, while the apparent diffusivities could be decreased to 10-8-10-9 cm2/s in polymeric material.19 Therefore, an extraction time of 60 h may be not enough for high-ring PAHs to reach the equilibrium of six-ring PAHs, subsequently decreasing the determined Kspme values. Molar volume is one of the often used indicators for serving as a descriptor for hydrophobicity.1,2 In the case of PAHs, the LeBas molar volume is regarded as a good indicator because of the compact nature of the multiring compounds. Figure 3 illustrates the relationship between log Kspme and the LeBas molar volume of the PAHs. The relationship between log Kspme and the LeBas molar volume of the 16 PAHs is not adequately described using a linear fit. The linear relationship was only up to benz[a]anthracene for 85-µm PA fiber and up to chrysene for 100-µm PDMS fiber. An increasing incompatibility and activity coefficient of the high-ring PAHs in the organic phase with increasing molar volume may be responsible for this deviation.7,20 Moreover, a (19) Chang, M. L.; Wu, S. C.; Chen, C. Y. Environ. Sci. Technol. 1997, 31, 2307. (20) Chou, C. T. Environ. Sci. Technol. 1982, 16, 4.
Figure 3
nearly constant slope of 0.022 ( 0.001 log-unit per cm3/mol for different SPME fibers was observed. This result is similar to the reported data (0.025),1 showing that the distribution coefficient determined in this study can properly predict the log Kow values of PAHs from naphthalene to chrysene (log Kow < 6). Effect of Sample Volume. Sample volume is an important parameter affecting the quantitative results. If the distribution coefficient (Kspme) is known, the effect of the sample volume in the partitioning of an analyte between the sample and the polymeric film on the fiber in equilibrium can be predicted using the following equation: Figure 4
n ) CoVfVsK/(KVf + Vs)
(1)
where n is the mass absorbed by the coating, Vf and Vs are the volumes of the coating and aqueous solution, respectively, and Co is the initial concentration of the analyte in the aqueous phase. To demonstrate the effectiveness of distribution coefficients on the determination of extracted amounts of PAHs by SPME, the effect of different water volumes on the mass extracted was determined using 15-ml vials. Also, two sets of experiments with different added amounts of PAHs were performed. The first set of experiments was conducted by delivering a constant initial concentration (10 ng/mL) into the vials (constant concentration system). The water volumes were 7.5, 10, 12.5, and 15 mL, respectively. In the second set of experiments, all of the procedures were the same as those of the constant concentration system, but the added amounts of PAHs were held constant to 100 ng instead of constant concentration (constant amount system). An extraction time of 90 min was selected because the low-ring PAHs can reach equilibrium and the analytical sensitivity of high-ring PAHs was sufficient. Also, this sampling time was applicable to the actual sample analysis. Figure 4 illustrates the comparison of the predicted and experimentally determined amounts of low-ring PAHs absorbed on the polymeric coating of a 100-µm PDMS fiber. All of the different volumes were analyzed in triplicate. The predicted values were calculated according to eq 1, and the Kspme values was taken from Table 1. It can be seen that a small increase in the water
volume produces a large change in the amount of analytes absorbed by the fiber, and threshold volumes existed for analytes to reach the constant absorbed mass. Excellent agreements were obtained between the experimental and theoretical values of the absorbed amounts of low molecular weight PAHs in both systems, showing that the distribution coefficient determined in this study is applicable to the quantitative determination of naphthalene, acenaphthylene, acenaphthene, fluorene, and phenanthrene. Because the system is far from equilibrium, however, the absorbed amounts obtained experimentally for high molecular weight PAHs were lower than the predicted values.
CONCLUSION The results obtained in this study show the effectiveness of the SPME method for determination of the distribution coefficients of PAHs in the two-phase system. The determined distribution coefficients of PAHs using the SPME method correlated well with octanol-water partition coefficient data from naphthalene to benzo[a]pyrene and from naphthalene to chrysene for 100-µm PDMS and 85-µm PA fibers, respectively. Good correlation was also demonstrated between determined distribution coefficients and LeBas molar volumes. A nearly constant slope (0.022 ( 0.001 log-unit per cm3/mol) for different SPME fibers was observed, showing that the distribution coefficient determined in this study can properly predict the log Kow values of PAHs from naphthalene Analytical Chemistry, Vol. 72, No. 15, August 1, 2000
3651
to chrysene. Moreover, the application of Kspme was validated by comparing the extracted amounts of PAHs with different sample volumes. Excellent agreements were obtained between the experimental and theoretical values of the absorbed amounts of low molecular weight PAHs in both systems, revealing that the distribution coefficient determined in this study is useful for better understanding of the sorption mechanism of SPME and is applicable to the prediction of extracted amounts of PAHs using the SPME method.
3652
Analytical Chemistry, Vol. 72, No. 15, August 1, 2000
ACKNOWLEDGMENT The authors thank Dr. Yun-chang Sun for helpful suggestions concerning preparation of the manuscript. This work is supported by the National Science Council, R.O.C. under Contract NSC 892113-M-007-040. Received for review January 10, 2000. Accepted May 9, 2000. AC000040L
