Environ. Sci. Technol. 2005, 39, 6770-6776
Polycyclic Aromatic Hydrocarbon Purification Procedures for Compound Specific Isotope Analysis MOONKOO KIM,* MAHLON C. KENNICUTT II, AND YAORONG QIAN Geochemical and Environmental Research Group, Texas A&M University, 833 Graham Road, College Station, Texas 77845
Purification and isotope analysis methods were developed to accurately measure the stable carbon isotope ratio of individual PAHs in environmental samples. Sample extracts were separated into several fractions by column chromatography and were purified by high-performance liquid chromatographic and thin-layer chromatographic techniques. The mean recovery of the purification method for all compounds was approximately 80%. The purity of isolates was verified by GC/MS in the full scan mode. The accuracy and precision of the method were verified using authentic standards of known isotopic compositions. The standard deviation for multiple analyses ranged between 0.1 and 0.4‰ for standard material and between 0.1 and 0.7‰ for environmental extracts after purification. The purification and isotope analysis methods were used to discern the sources of PAHs in environmental samples using variations in stable carbon isotopic compositions of individual compounds. It was confirmed that the purification method effectively purifies environmental samples while retaining original isotopic signatures so that PAH content and stable isotopic compositions can be used to infer contaminant sources in complex environmental settings.
Introduction Many industrial and combustion-derived chemicals, including polycyclic aromatic hydrocarbons (PAHs), are released to and persist in the environment. PAHs are detectable almost everywhere in the environment, including water, sediment/ soil, air, and organismal tissues (1). PAHs are formed by either thermal alteration of buried organic matter or the incomplete combustion of organic matter (2, 3). The widespread occurrence of PAHs in the environment, along with their persistence and toxic properties, has prompted extensive research on the fate and effects of PAHs in the environment (4-6). Integrated assessments of the environmental effects of pollutants document the chemical and toxic nature of the contaminants, their sources, and the pathways of exposure. This information can then be used to remediate contaminated sites and prevent future contamination. One traditional method is the use of compositional information to identify contaminant sources. However, chemical and biological alterations of contaminants often change the original composition of contaminants preventing identification of con* Corresponding author present address: Department of Chemistry, Western Michigan University, Kalamazoo, MI 49008; phone: (269)387-2876; fax: (269)387-2909; e-mail:
[email protected]. 6770
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 17, 2005
taminant sources (7). Alternatively, identification of sources using stable isotopic compositions can be effective if isotopic compositions are resistant to chemical and biological alteration processes (3). PAHs from different processes or source materials often exhibit unique stable isotopic compositions that may be useful as a method of source identification (3, 7-11). Environmental samples, including sediments and soils, can contain varying amounts of many types of organic matter. To accurately determine the stable carbon isotope ratio of specific compounds, it is necessary to extract, isolate, and purify the components to avoid coelution, peak overlap, and unresolved complex mixture (UCM) interferences during stable isotope analysis. A minor component coeluting with a target compound can have a significant effect on the isotope ratio if their carbon isotope ratios are significantly different (12, 13). Therefore, producing a high-purity extract, that has not been modified in isotopic composition, is important for the accurate and precise analysis of isotopic compositions. Various purification methods, including silica gel column (14), alumina/silica gel column (8), and sephadex and silica gel column chromatography (7, 9, 10), have been used for the purification of environmental samples for compound specific isotope analysis of PAHs. Yanik et al. (3) used alumina/silica column and cyano/amino bonded phase column to isolate crude oil PAHs in wetland soils. Okuda et al. (15-18) used silica gel column and chemically bonded aminopropyl-silica gel column chromatography to purify PAHs in automotive exhausts, atmospheric particles, and sediments. In the current study, a combination of alumina/ silica gel column chromatography, gel permeation chromatography, and thin-layer chromatography is tested as a purification method to prepare extracts for compoundspecific isotope analysis of PAHs. The objective of this study was to develop a purification and isotope analysis method to accurately measure the stable carbon isotope ratios of PAHs and to use these intrinsic characteristics to define the sources of pollutants in environmental matrixes. The developed method would minimize interferences from coeluting compounds by producing highly pure extracts. The purification and isotope analysis method was then used to analyze environmental samples to confirm the utility of the method as a source identification tool. The methods were tested on soils and sediments from marine, lacustrine, and land environments.
Experimental Methods PAH Quantification. An aliquot (∼15 g) of partially dried sediment was mixed with anhydrous sodium sulfate and was extracted with an accelerated solvent extractor (ASE, Dionex) with dichloromethane. The concentrated extracts were purified and fractioned with an alumina/silica gel (Al/Si, 10 g/20 g) chromatographic column (300 × 13 mm i.d.). Alumina oxide (∼150 mesh) was activated by heating at 400 °C for 4 h and then was deactivated with HPLC grade water (1%, w/w). Silica gel (100∼200 mesh) was activated by heating at 170 °C for 12 h and was deactivated with HPLC grade water (5%, w/w). The aromatic hydrocarbons were eluted from the column with 200 mL of pentane/dichloromethane (1/1, v/v). The collected fraction was concentrated and exchanged to 1 mL of hexane for instrumental analysis. Elemental sulfur was removed from the extracts by adding activated copper during ASE extraction and Al/Si column chromatography. The PAHs were analyzed on a Hewlett-Packard 6890 GC coupled with a Hewlett-Packard 5973 mass selective detector. Separation of PAHs was accomplished with a 30 m × 0.25 10.1021/es050577m CCC: $30.25
2005 American Chemical Society Published on Web 08/05/2005
temperature of 60 °C was held for 1 min. The oven was programmed to increase from 60 °C to 150 °C at 12 °C/min and then at 5 °C/min to a final temperature of 300 °C with a final hold time of 20 min. The GC effluent passed through a combustion interface at 980 °C, quantitatively combusting the PAHs to carbon dioxide. The stable carbon isotope ratios of individual PAHs were determined by comparison to a working standard of CO2 (99.996%, δ13CPDB ) -11.57‰) introduced at the beginning of each run.
Results and Discussion
FIGURE 1. Analytical procedure for stable carbon isotope ratios of PAHs. mm (i.d.), 0.25-µm film thickness, DB-5MS fused silica capillary column (J&W Scientific). The oven temperature was programmed to increase from an initial temperature of 60 °C to 150 °C at 15 °C/min, then 5 °C/min to 220 °C, and finally at 10 °C/min to a final temperature of 300 °C with a final holding time of 10 min. The PAHs were identified on the basis of the comparison of the retention times and mass spectrum of selected ions with the calibration standards. Compound Specific Isotope Analysis (CSIA). The procedure to separate and purify PAHs from extracts for stable carbon isotope ratio analysis is shown in Figure 1. Approximately 15 g of sediments, dried chemically by mixing with anhydrous sodium sulfate, was extracted on ASE system with dichloromethane. For sediments with low PAH concentrations, 50∼100 g of chemically dried samples were Soxhlet extracted with dichloromethane. Concentrated extracts were separated into aliphatic and aromatic fractions using Al/Si column chromatography. Aliphatic hydrocarbons were eluted from the column with 100 mL of pentane. Aromatic hydrocarbons were then eluted with 200 mL of pentane/dichloromethane (1/1, v/v). The aromatic hydrocarbon fraction was further purified by an automated gel permeation chromatography (GPC) (19) using a highperformance liquid chromatography (HPLC) equipped with two size exclusion columns (22.5 × 250 mm, Phenomenex Phenogel 100 Å). Dichloromethane was used as the mobile phase at a flow rate of 7 mL/min. Collection start and end times were monitored using an UV detector. The PAH fraction collected from the GPC was further purified by thin-layer chromatography (TLC). TLC plate (silica gel 60 Å, 500-µm thickness, EM Science) was prewashed by developing with dichloromethane/methanol (1/1, v/v) and then was activated by heating at 120 °C for 1 h. After cooling, about 100 µL of concentrated extract was applied as a thin band onto the TLC plate. The plate was developed with about 50 mL of cyclohexane/toluene (3/2, v/v), and the PAH band (RF ∼ 0.81) was located by comparison with standard materials under short wavelength (254 nm) UV light. The silica gel containing the PAHs was scraped off the TLC plate and the PAHs were extracted with dichloromethane by sonication. The purified samples were analyzed for compoundspecific isotopic composition with a Varian 3400 GC coupled to a Finnigan MAT 252 IRMS. A DB-5MS fused silica capillary column (30 m × 0.32 mm i.d., 0.5-µm film thickness, J&W Scientific) was used to separate the PAHs. The initial oven
The Al/Si column separates organic compounds on the basis of their polarity and separates aromatic compounds from aliphatic hydrocarbons and other polar components on the basis of the eluting solvent. Because of the high concentrations of aliphatic hydrocarbons in environmental samples and their coelution with PAHs during GC separations, it is necessary to remove aliphatic hydrocarbons prior to isotope ratio analysis. To determine the necessary volume of pentane for the complete elution of aliphatic hydrocarbons, multiple Al/Si columns were tested using PAH standards of different concentrations (120-6000 ng) of individual PAH compounds and varying volumes of pentane elution. Except for naphthalene, all compounds had losses of less than 1% when the column was eluted with 100 mL of pentane (Table 1). The naphthalene detected in pentane fraction was most likely derived from reagents and solvents. When background contamination was eliminated, loss of naphthalene from the PAH fraction was negligible. Even though 50-80 mL of pentane was sufficient to remove the aliphatics for PAH quantification, it is preferable to elute with excess pentane to ensure complete removal of all aliphatic interferences if there is no loss of aromatic compounds. Therefore, 100 mL of pentane was adopted as the method to remove aliphatic hydrocarbons from sample extracts. During the second elution of the Al/Si column with dichloromethane/pentane, all the PAH compounds tested had recoveries of greater than 90%, except for dibenzothiophene which had a recovery of 82% (Table 1). It is likely that dibenzothiophene was not completely eluted from the column because the sulfur atom in the aromatic ring forms a temporary hydrogen bond with the O-H functional groups on the surface of Al/Si adsorbent or other polar functional groups on the surface of the adsorbents. Polar compounds are preferentially retained by the Al/Si adsorbents. After the Al/Si cleanup, the sample extracts were further purified by a GPC which separates compounds on the basis of molecular size. Because of the small and compact structural nature of PAHs, they enter the pores of the packing material and are retained on the column longer than compounds of larger molecular size. Mostly linear-structured impurities, such as aliphatic hydrocarbons and long-chain fatty acids, were eluted first while the PAH compounds were eluted between 13.9 and 21.0 min after the injection. The recoveries of PAHs were greater than 90% after GPC purification for most compounds (Table 1). No PAHs were detected in the early eluting fractions (less than 0.1% recoveries). In addition to the GPC, TLC was tested as a further purification technique. Various TLC solid phases and development solvents were tested. Among the techniques tested, the method using a silica gel plate and cyclohexane/ toluene for development, modified after Levins’ (20), was selected for the purification of samples for isotope analysis. This method produced the narrowest PAH band and greatest separations from interferences providing the best purification scheme. The full scan GC/MS analyses of seven TLC bands recovered from the plate confirmed that all PAH compounds were recovered in the seventh band and that there were no PAHs in other bands (bands 1-6). In an effort to further improve the efficiency of the purifications, two different VOL. 39, NO. 17, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6771
TABLE 1. Recoveries (%) of PAH Compounds after Each Purification Stepa compound names
abbr.
A
B
C
D
E
F
G
H
naphthalene 2-methylnaphthalene 1-methylnaphthalene biphenyl 2,6-dimethylnaphthalene acenaphthylene acenaphthene 1,6,7-trimethylnaphthalene fluorene phenanthrene anthracene 1-methylphenanthrene dibenzothiophene fluoranthene pyrene benz[a]anthracene chrysene benzo[b]fluoranthene benzo[k]fluoranthene benzo[e]pyrene benzo[a]pyrene perylene indeno[1,2,3-cd]pyrene dibenz[a,h]anthracene benzo[ghi]perylene average
NAP 2MN 1MN BIP DMN ACL ACN TMN FLU PHE ANT 1MP DBT FLT PYR BaA CHR BbF BkF BeP BaP PER INP DBA BPE
4.70 0.10 0.40 0.10 0.00 0.10 0.00 0.00 0.10 0.30 0.20 0.00 0.00 0.70 0.60 0.40 0.50 0.50 0.20 0.30 0.50 0.10 0.30 0.10 0.40 0.4
93.3 92.2 91.6 96.0 94.6 94.7 95.4 96.8 94.4 94.0 91.1 95.2 81.5 96.6 94.6 99.7 98.5 94.4 108.0 100.4 98.9 98.2 103.9 104.7 103.0 96.5
93.7 93.2 93.5 91.7 91.8 92.3 91.6 92.7 92.3 91.7 88.8 92.1 91.0 94.2 93.0 96.1 95.3 91.4 101.8 94.9 93.7 94.4 100.5 99.5 95.4 93.9
56.5 69.7 70.3 87.6 67.0 81.9 76.7 79.2 94.5 95.8 87.7 97.1 94.3 97.5 96.8 94.8 95.7 93.9 97.1 95.9 89.0 91.3 93.3 93.7 93.9 87.7
31.7 43.2 44.2 65.1 39.6 64.6 49.4 54.3 79.5 86.5 78.3 90.7 70.4 93.8 93.8 91.4 95.8 92.3 94.9 93.2 86.0 91.4 88.6 89.6 90.6 76.0
5.7 28.7 29.4 58.9 36.4 39.1 45.9 53.7 81.9 88.9 72.5 90.9 72.8 93.6 91.7 91.2 93.6 93.7 93.5 94.1 80.0 79.8 91.3 92.1 91.1 71.6
19.8 34.9 35.6 60.1 35.5 62.5 45.0 52.6 83.6 90.1 81.5 93.5 78.4 95.2 95.3 93.5 94.1 93.5 93.7 94.2 86.9 83.1 92.9 93.3 92.6 75.3
50.0 58.8 59.4 68.0 60.6 67.5 60.2 63.3 78.4 86.3 78.2 87.0 69.9 89.7 88.6 87.2 87.8 91.7 85.1 88.4 83.9 80.3 89.3 89.2 86.1 77.4
a A: Amount recovered during 100 mL pentane elution (n ) 5). B: Recoveries after Al/Si column chromatography (n ) 5). C: Recoveries after GPC (n ) 5). D: Recoveries after TLC (unsaturated, n ) 3). E: Recoveries after TLC (saturated, n ) 3). F: Recoveries after TLC with 30 min air exposure (n ) 3). G: Recoveries after TLC with N2 dry (n ) 3). H: Recoveries after all purification steps (n ) 5).
development environments, saturated and unsaturated chambers, were tested. The difference in recoveries between the two development environments is shown in Table 1. Higher recovery rates were observed in unsaturated chamber developments, especially for low molecular weight PAHs. The difference in recoveries between saturated and unsaturated chambers is in the amount of PAHs recovered in the area above the PAH band. The solvent-saturated environment accelerated the movement of the low molecular weight PAHs as compared to the high molecular weight PAHs. The recovery of the PAHs in the unsaturated TLC separations was about 88% on average (Table 1). The high molecular weight PAHs had higher recoveries (more than 90%), while the low molecular weight PAHs, especially naphthalene, had lower recoveries (about 57%). The low recoveries of naphthalene appear to be due to evaporative losses during the application of the sample. The evaporative loss of low molecular weight PAHs was confirmed by exposing the TLC plates to the air for extended periods of time after sample application. After the exposure of 30 min, the recoveries of low molecular weight PAHs decreased markedly while no significant change was observed for high molecular weight PAHs (Table 1). To reduce the evaporative losses, application time was minimized and minimal amounts of solvent were used to apply the sample. This minimized the time needed to remove the application solvent. To accelerate the evaporation of solvent, nitrogen gas was blown across the plate surface. Although nitrogen gas dramatically reduces the drying time, it also increases evaporative loss of low molecular weight PAHs (Table 1). The integrity of the samples during the extraction, purification, and instrumental analysis was verified by quantitative mass balance of all the analytes of interest. The recoveries of all the PAH compounds after all the purification steps are shown in Table 1. Although the recoveries of low molecular weight PAHs are low, especially for naphthalene (∼50%), no isotopic alteration during the purification was observed (-23.5 ( 0.3‰ before and -23.3 ( 0.4‰ after purification; n ) 5). The mean recovery for the purification method for all the compounds was about 80%. The purity 6772
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 17, 2005
of the isolates was confirmed by the full scan GC/MS analysis. The mass fragmentograms of samples after each purification step are shown in Figure 2. It is clear that peak resolution is improved and that the raised baseline, which is mostly UCM, is decreased after each purification step. The mass spectra of the PAH compounds in the samples were compared with the reference spectra from the National Institute of Standards and Technology (NIST) mass spectral database for the pure compounds. The mass spectra of sample components were, in general, very similar to the reference spectra (e.g., 93% and 95% match qualities for phenanthrene and fluoranthene, respectively), indicating minimal interference from coeluting material. A Finnigan MAT 252 IRMS coupled to a Varian 3400 GC via a combustion interface was used for the stable carbon isotope analysis. The initial GC conditions were similar to that used for PAH quantification. The GC method was modified by changing the carrier gas flow, column head pressure, purge time, oven temperature program, and so forth to provide an optimal separation of PAHs from other interfering compounds on the GC/IRMS system. One of the changes in the GC conditions was the modification of the purge delay time. To increase analyte response and remove solvent tailing, samples were injected in splitless mode and the inlet was initially purged at 2 min after injection. However, under these GC conditions, a long solvent tail was produced that interfered with the analysis of PAHs. To minimize solvent tailing in the GC system, the response of selected peaks in terms of peak height and area according to their purge time was tested. More than 90% of the detector response occurred within the initial 50 s and no further increase was observed after 1 min. Therefore, the analyte can be introduced onto the column within 50 s after injection with about 10% of the analyte being purged from the injection port. Thus, the solvent tail can be removed, and the analyte responses remain unchanged because solvent tail usually accounts for less than 5% of total detector response (21). The peak resolution was further improved by narrowing the peak width by adjusting the injection port head pressure.
FIGURE 2. Full scan GC/MS chromatograms of sediment sample after each purification step (A: after Al/Si column chromatography, B: after GPC, C: after TLC).
FIGURE 3. GC/IRMS chromatographic traces (m/z 44) of standard material along with their m/z 45/44 ratio traces. Higher column head pressure caused faster column gas flow which resulted in narrower peak shapes (13). A head pressure of 30 psi was enough to narrow the peak shape. GC/IRMS chromatographic traces of standard materials under these conditions are shown in Figure 3. The precision of the stable carbon isotope analysis was verified by repeated processing and injection of authentic standards and samples. Standard deviations of stable carbon isotope values for multiple injections of standard materials
are shown in Table 2. For all the measured compounds, the standard deviation for multiple analyses ranged between 0.1 and 0.4‰. The relationship between peak area and the standard deviation for multiple isotope measurements is shown in Figure 4. It is clear that the precision of the measurement increases with peak size mainly because the smaller peaks have greater contribution from background or interfering materials. For the precise analysis of stable carbon isotope ratio, peaks greater than 10 V s (volt‚second) in peak area were used. The standard deviations of isotope ratios for multiple injections of environmental samples after purification fall in the range of 0.1-0.7‰, which is similar or better than the results reported in the literature which used different methods for purification (Table 2). A similar range of precision has been documented in the literature probably reflective of the instrumental precision of GC/IRMS analysis. The analytical protocol developed in this study can provide similar or more precise isotope ratio data than those analytical methods reported previously. Accurate measurements can be achieved by introducing standard materials of known isotopic composition and determining the isotopic ratio of individual compounds in the sample on the basis of the isotope ratio of a standard. Standard CO2 gas was introduced in triplicate at the beginning of each run by means of an open slit and was subjected to the same analytical conditions as the samples. The analytical protocols were also evaluated to ensure that the compositional and stable isotopic integrity is preserved throughout the purification scheme by processing standards of known stable isotopic compositions. The results of stable carbon isotopic measurements of standard materials after purification are compared with those of unprocessed standard materials in Figure 5. A diagonal line represents a 1:1 ratio between the isotope values of processed and unprocessed standard materials. The ratio for most compounds fall on the diagonal line demonstrating the general stable isotopic conservation during the purification process. Although the method shows general isotopic conservation, some analytes show a slight isotopic enrichment during purification. There could be several reasons for the slight enrichment. One effect is an isotopic fractionation during the concentration steps which might cause a preferential loss of isotopically light PAH compounds during evaporation. Physicochemical processes such as evaporation involve an isotope exchange effect which results in preferential enrichment of lighter molecular species in the vapor phase and enrichment of heavier molecules in the residue (22). Additional sources or processes may also cause minor changes in carbon isotope compositions for some compounds. O’Malley et al. (7) demonstrated that evaporation had no significant effect on the measured stable carbon isotope ratios of PAHs. In the current study, even though there were slight enrichments for some compounds, these differences are minimal, and all the isotope ratios of individual PAHs for the purified standard material fell within two standard deviations (2σ) of the mean isotope ratios of unprocessed standard materials. Application to Lacustrine Environmental Samples. The purification and isotope analysis methods were applied to extracts from environmental samples to test the applicability of the method. PAHs were extracted from the sediments of an urban lake in the northwestern United States. The isotopic compositions of PAHs were compared with those of sediment samples from other sites, including a shipping waterway, harbor, and relatively undisturbed remote lake in the same watershed. The stable carbon isotope ratios of PAHs from the study sites are shown in Figure 6. Compound-specific isotope analysis suggested that PAHs in the study areas were derived from multiple sources. The error bar shows (1 standard VOL. 39, NO. 17, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6773
TABLE 2. Standard Deviation of Multiple Measurements of Stable Isotope Ratios
a
references
year
matrix
Ballentine et al. (8) O’Malley et al. (10)
1996 1997
Okuda et al. (15) Yanik et al. (3) this study (n ) 5)
2002 2003 2005
aerosol aerosol standard standard sediment standard standarda sediment
purification Al/Si Sephadex, Si Si, aminopropyl-silica gel Al/Si, cyano/amino bonded column Al/Si, GPC, TLC
standard deviation 0.2-0.8 0.15-0.65 0.15-0.4 0.4-0.7 0.43-1.21 0.1-0.4