Environ. Sci. Technol. 2007, 41, 3316-3321
The Primary Aerobic Biodegradation of Gasoline Hydrocarbons ROGER C. PRINCE,* THOMAS F. PARKERTON, AND CAROLYN LEE ExxonMobil Biomedical Sciences, Inc., Annandale, New Jersey 08801
We describe the primary aerobic biodegradation of an unleaded, unoxygenated, regular gasoline by inocula from unacclimated fresh and sea water, and from a domestic sewage treatment plant. Biodegradation was rapid and complete in all inocula, with an overall median “half-life”, at ∼70 ppm gasoline and low levels of inorganic nutrients, of 5 days. The biodegradation of 131 individual hydrocarbons in the gasoline followed a relatively consistent pattern. The larger n-alkanes and iso-alkanes, and simple and alkylated aromatic compounds were the most readily degraded compounds, followed by the smaller n-alkanes and isoalkanes and the naphthenes. The last compounds to be degraded were butane, iso-butane, and 2,2-dimethylbutane, but even these disappeared with an apparent half-life of 85% degradation of the total gasoline hydrocarbons. In more degraded samples we used the average concentration of these compounds in all the less degraded samples and t ) 0 samples. To ensure that this was appropriate, later series of experiments were conducted with carbon tetrachloride added to approximately 1% in the gasoline (m/z ) 117, elution time
% Loss ) [((A0/C0) - (As/Cs))/(A0/C0)] × 100 where As and Cs are the concentrations of the target analyte and conserved compound in the sample, respectively, and A0 and C0 are the concentrations in the sterile controls. We observed consistent results using the two different internal hydrocarbon standards or carbon tetrachloride, and all the data presented here used either 2,2,4-trimethylpentane or carbon tetrachloride. Current regulatory requirements in Canada (17) and Europe (18) focus on the half-life of chemical substances in the environment. We have calculated apparent half-lives, τ, for the disappearance of the total detectable gasoline and the individual hydrocarbons from the fraction remaining ((100 - % loss)/100), A, at time t from the equation
τ ) ln2 • (-t/lnA) Further discussion of the issues relevant to such calculations is provided in the Supporting Information.
Results and Discussion Figure 1 shows total ion chromatograms of representative samples from our experiments using pond water. Within a few days the majority of the hydrocarbon had been consumed (Figures 1 and 2A), and eventually there were no detectable hydrocarbons in the vials. The median half-life of total detectable gasoline hydrocarbons was 5.0 days (Figure 2B), and the mean of these estimates was 5.9 days (n ) 102). We attribute the variability in Figure 2A to the fact that the inocula were not pre-acclimated to hydrocarbons; there were no obvious differences between samples collected in different months, even when collected from under winter ice, or between the different inocula, although we note that all laboratory incubations were maintained at summer temperatures. Although there was some diversity in the absolute rate of degradation in different samples, the process of biodegradation generally followed a clear sequence. Figure 3 presents data on the n-alkanes. Part A shows data from individual experiments, and the variation in the process of biodegradaVOL. 41, NO. 9, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3317
FIGURE 2. Disappearance of total detectable hydrocarbons with time: (A) raw data, with the line drawn to guide the eye, and with a “lag” of about 2 days; (B) frequency distribution of calculated estimates of apparent “half-life” of gasoline in A.
FIGURE 4. Loss of C8-paraffins: (A) graphical representation of when biodegradation of a particular analyte occurs, based on all inocula (see Figure 3); (B) the preferential loss of 2-methylheptane before 3-methylheptane, where the solid line would be equal loss. FIGURE 3. Loss of n-alkanes: (A) degradation of some n-alkanes from all inocula, with lines drawn to guide the eye; (B) graphical representation of when biodegradation of a particular analyte occurs, defined as when an analyte is present at between 90 and 10% of its initial concentration. The box shows the interquartile range of samples where biodegradation was occurring. The lines indicate the extent of 10-90% of the data, and any data points indicate potential outliers. tion. An overview of the process can be gained (Figure 3B) by delineating that fraction of the total degradation during which the biodegradation of a specific compound, or group of compounds, occurred. n-Dodecane and n-undecane were the most rapidly degraded compounds, followed by n-decane and n-nonane. The smaller n-alkanes disappeared more slowly. It is noteworthy that there is apparently a sequential biodegradation: hexane, for example, typically underwent significant biodegradation only after nonane biodegradation was essentially complete, although there were a few experiments where hexane was lost almost simultaneously with nonane. Note that all the n-alkanes in the gasoline (14% by weight), even butane, are degraded in the longest incubations. Calculated apparent half-lives are presented in the Supporting Information. Figure 4 shows some data on the iso-alkanes, in this case those with eight carbons. n-Octane is included for reference. Beens and Brinkman (19) point out that there are 18 possible isomers of octane, and 12 are present at high enough concentrations in this gasoline for us to assign them with a reasonable degree of certainty. Iso-Alkanes are important constituents of gasolines, representing 27% by weight in this sample. This hydrocarbon class tends to be degraded a little 3318
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 9, 2007
less rapidly than the n-alkanes, but, as can be seen in Figure 4A, the biodegradation of monomethylalkanes begins before the biodegradation of the n-paraffins is complete. There is a preference for the biodegradation of the 2-methyl form over the 3- and 4-methyl forms (Figure 4B), which was also seen with the other alkanes in this gasoline (except 2- and 3-methylpentane which were degraded at very similar rates). The biodegradation of 3-ethylhexane lags that of 3-methylheptane, but is similar to the degradation of the dimethyl forms. The trimethyl forms are among the more recalcitrant compounds we detected in this gasoline, as they were in the experiments of Solano-Serena et al. (8). In contrast to that work, however, Figure 5 shows representative data from an experiment using a seawater inoculum and gasoline spiked with carbon tetrachloride as an internal marker. Within 5 days the majority of the iso-alkanes had been degraded, and they were eventually completely consumed. Figure 6A shows representative data for cyclohexane and its simple alkylated forms as representative naphthenes (7% of this gasoline). The parent cyclohexane is among the least degradable, although it is completely biodegraded with a median apparent half-life of 8.1 days (see SI). n-Alkylsubstituted forms, such as pentylcyclohexane, are degraded preferentially (Figure 6A). Gasoline also contains equivalent cyclopentanes, which are degraded under our test conditions with half-lives ranging from 6.5 to 9.1 days (see SI). The linear and cyclic olefins (cyclic and linear hydrocarbons with a single unsaturation, 12% of this gasoline) are also completely degraded (Figure 7), and Figure 6B shows an example that they are degraded slightly preferentially to their fully saturated analogs.
FIGURE 7. Loss of aromatics and alkenes. Graphical representation of when biodegradation of particular groups of analytes occurs, based on all inocula (see Figure 3). The individual compounds of the different groups are listed in SI Table 2.
FIGURE 5. Gas chromatograms of C8-paraffin biodegradation: seawater inoculum, gasoline amended with 1% carbon tetrachloride (black, measured at m/z ) 117), and spectra normalized to equal amounts of this compound. The rest of the chromatograms are m/z ) 57. One of the Day 5 samples had lost a little 2,2,4-trimethylpentane, the other a lot, and the Day 48 sample had lost all the alkanes. The dashed lines enclose the methyl, dimethyl, and trimethyl compounds.
FIGURE 6. Loss of cyclohexanes: (A) graphical representation of when biodegradation of a particular analyte occurs, based on all inocula (see Figure 3); (B) preferential loss of 1-methylcyclopentene before methylcyclopentane, where the solid line would be equal loss. Figure 7 also shows data on the biodegradation of the aromatic components of gasoline (39% of this gasoline). They are readily degraded, lagging only slightly behind the
n-alkanes. There is a slight overall preference for the degradation of 1,3-disubstituted compounds, but the biodegradation of the different isomers of substituted benzenes do not always follow the same pattern, even with inocula from a single source, as shown by representative chromatograms of the C3-benzenes (benzenes with three additional -CH2 groups; trimethyl, methyl-ethyl, propyl, and isopropyl benzenes) in pond water (Figure 8A). The sample that has lost 30% of the total gasoline has lost all the 1-ethyl-3-methyland 1-ethyl-4-methyl-benzene, but not the 1-ethyl-2-methylor 1,3,5-trimethyl-benzene. In contrast, the sample that has lost 40% of the total hydrocarbon exhibits the reverse preference. This diversity is further demonstrated in Figure 8B, where 1,2,4-trimethylbenzene was degraded substantially before the other two isomers (dashed line) in 7 of the 23 samples where we stopped the incubation before complete degradation of these compounds. All were in freshwater incubations, but only some of the 3-5 replicates at a particular time showed the phenomenon. On the other hand, Figure 8C shows that branching of an alkyl substituent consistently slows biodegradation. Gasoline also contains some aromatic two-ring compounds, naphthalenes, indans, and tetralins, and their biodegradation is included in Figure 7. All were readily degraded, and there was subtle but fairly consistent preference for the biodegradation of naphthalene > tetralin > indan, naphthalene >2-methylnaphthalene >1-methylnaphthalene, tetralin >5-methyltetralin >6-methyltetralin, and 4-methylindan >5-methylindan, 1-methylindan, and indan. These preferences might provide important forensic evidence in understanding spill source identification as biodegradation proceeds (6). Our results extend those presented by others (8-12) both in studying far more individual hydrocarbons (131 individual analytes, see Supporting Information), and by demonstrating that even recalcitrant iso-alkanes such as 2,2,4-trimethylpentane and even butane, iso-butane, and 2,2-dimethylbutane are routinely biodegraded under aerobic conditions by the indigenous microflora of relatively “clean” sites, a freshwater pond and the New Jersey seashore (cf. 8). Microorganisms obtained from a domestic activated sludge water treatment plant exhibited broadly similar kinetics and preferences, and again biodegradation was complete. Biodegradation of the different classes of hydrocarbon molecules followed a fairly predictable pattern in all the experiments, albeit with subtle distinctions of preference in different samples collected on the same day. The larger n-alkanes and iso-alkanes, and simple and alkylated aromatic compounds were the most readily degraded compounds, followed by the smaller n-alkanes and iso-alkanes and the naphthenes. The last compounds to be degraded were butane, iso-butane, VOL. 41, NO. 9, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3319
FIGURE 8. Loss of aromatics: (A) loss of individual C3benzenes in pond water: % total loss of gasoline calculated using 2,2,4-trimethylpentane as the conserved internal marker (measured as the m/z ) 57 ion); the chromatograms are of the m/z ) 105 ion. (B) Loss of the three trimethylbenzenes, with lines drawn to guide the eye. (C) Preferential loss of n-alkyl-substituted before iso-alkyl-substituted benzenes, both propyl versus iso-propyl and butyl versus the two iso-butyl isomers, based on all inocula. The dashed lines are drawn to guide the eye, but the solid line would be equal loss. and 2,2-dimethylbutane, and as shown in the Table in the Supporting Information, their calculated half-lives were 15.0, 17.5, and 26.5 days. Interestingly they were biodegraded after all the other hydrocarbons had been degraded. Thus they are unlikely to have been degraded by simple co-metabolism with other hydrocarbons, although their co-metabolism with intermediates of other hydrocarbon degradation seems quite likely as we will discuss below. Hydrocarbons are relatively unusual substrates for microbial growth in that they provide carbon and energy, but none of the trace elements essential for growth. We therefore added low levels of biologically available nitrogen and phosphorus (125 µM ammonium, 125 µM nitrate, and 130 µM phosphate), and it is likely that some growth occurred during our incubations. The complete, rapid, primary biodegradation of all the hydrocarbons in the gasoline stands in apparent contrast to some regulatory literature, but probably reflects the fact that many organisms can metabolize a broad array of hydrocarbons, although they may not be able to grow on individual compounds. For example, only 10 of the 25 individual gasoline components run in the OECD Ready Biodegradability 301C (MITI) protocol (13), showed significant biodegradation (20), although it must be noted that these tests follow the total biodegradation of the hydrocarbon to CO2 and H2O (monitored by oxygen consumption) rather than its simple disappearance as measured here. For a hydrocarbon containing eight carbons, the initial oxygenation would only consume about 8% of the total oxygen demand of the compound, which might be below the detection limit of the OECD method. We prefer the explanation that indigenous microorganisms can degrade a broad range of hydrocarbons, but do so much more effectively when they are administered as a mixed suite of hydrocarbon substrates that allows microbes to use intermediates from different pathways to balance their overall metabolism. After all, hydrocarbons are rarely found as single compounds in nature, but rather as complex mixtures resulting from diagenesis and catagenesis during oil generation and maturation (discussed in 6). Indeed 3320
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 9, 2007
Solano-Serena et al. (8) found that oxygen consumption in experiments very similar to ours accounted for 94% of the theoretical oxygen demand of the complete mineralization of the gasoline in the system. Further support comes from attempts to isolate microorganisms growing on individual hydrocarbons. Ridgway et al. (7) isolated 244 isolates from an unleaded gasolinecontaminated aquifer, and found that they could be sorted into 111 catabolic groups on the basis of aerobic growth on 15 gasoline hydrocarbons; most could grow on only two or three of the substrates tested. Several isolates have been found, albeit after some trouble, that can grow on cyclohexane as sole carbon and energy source (21 and references therein), but they seem rare in the environment, and most strains seem to grow by co-metabolism. Yet cyclohexane was biodegraded quite rapidly in all our experiments. We also note that although some hydrocarbons have been part of the biosphere from oil seeps from the earliest days, other hydrocarbons, such as 2,2,4-trimethylpentane, are made during catalytic processing of fossil fuel streams, and are not found in any significant concentrations in unrefined hydrocarbons. Nevertheless, such compounds were metabolized quite promptly in our experiments using pond water, seawater, and sewage treatment plant inocula. SolanoSerena et al. (22) have isolated a Mycobacterium that can grow on 2,2,4-trimethylpentane as sole carbon and energy source from contaminated groundwater, but again it seems that such organisms are rare, and it is likely that most degradation occurs by organisms growing simultaneously on other substrates or their biodegradation intermediates. Although gasoline was added at ∼70 ppm in our experiments, the individual hydrocarbons were present in the aqueous phase at much lower concentrations; the aromatics were typically in the sub-ppm level, the others at a few ppb or less (Table 2). For these latter compounds, the reported degradation kinetics may be in part limited by redistribution kinetics from air to water where the microbial degradation occurs. Such low levels and potential mass transfer limitations were nevertheless sufficient to stimulate the unacclimated
consortia to metabolize all the detectable hydrocarbons in gasolinesa quite remarkable phenomenon.
(12)
Acknowledgments We are grateful to an anonymous reviewer for helpful suggestions.
(13)
Supporting Information Available
(14)
Discussion and table of the half-lives of the individual hydrocarbons. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) U.S. Department of Energy, Energy Information Administration. 2007; http://www.eia.doe.gov/oil_gas/petroleum/ data_publications/. (2) Spaulding, R. S.; Schade, G. W.; Goldstein, A. H.; Charles, M. J. Characterization of secondary atmospheric photooxidation products: Evidence for biogenic and anthropogenic sources. J. Geophys. Res. 2003, 108, 4247-4264. (3) Gary, J. H. Petroleum Refining: Technology and Economics; Marcel Dekker: New York, 2001. (4) Ritter, S. What’s that stuff? Gasoline’s ubiquity belies its complex blend of hydrocarbons and additives. C & E News 2005, 83 (8), 37. (5) Ellis, L. B.; Roe, M. D.; Wackett, L. P. The University of Minnesota Biocatalysis/Biodegradation Database: the first decade. Nucleic Acids Res. 2006, 34, D517-D521. (6) Prince, R. C.; Walters, C. C. Biodegradation of oil hydrocarbons and its implications for source identification. In Environmental Forensics: Spill Oil Fingerprinting and Source Identification; Wang, Z., Stout, S.A., Eds.; Elsevier: New York, 2006; pp 349379. (7) Ridgway, H. F.; Safarik, J.; Phipps, D.; Carl, P.; Clark, D. Identification and catabolic activity of well-derived gasolinedegrading bacteria from a contaminated aquifer. Appl. Environ. Microbiol. 1990, 56, 3565-3575. (8) Solano-Serena, F.; Marchal, R.; Ropars, M.; Lebeault, J.-M.; Vandecasteele, J.-P. Biodegradation of gasoline: kinetics, mass balance, and fate of individual hydrocarbons. J. Appl. Microbiol. 1999, 86, 1008-1016. (9) Solano-Serena, F.; Marchal, R.; Lebeault, J.-M.; Vandecasteele, J.-P. Selection of microbial populations degrading recalcitrant hydrocarbons of gasoline by monitoring of culture-headspace composition. Lett. Appl. Microbiol. 2000, 30, 19-22. (10) Marchal, R.; Penet, S.; Solana-Serena, F.; Vandecasteele, J.-P. Gasoline and diesel oil biodegradation. Oil Gas Sci. Technol. s Rev. IFP. 2003, 58, 41-448. (11) Townsend, G. T.; Prince, R. C.; Suflita, J. M. Anaerobic biodegradation of alicyclic constituents of gasoline and natural
(15)
(16)
(17) (18)
(19)
(20)
(21)
(22)
(23)
gas condensate by bacteria from an anoxic aquifer. FEMS Microbiol. Ecol. 2004, 49, 129-135. Van Hamme, J. D.; Ward, O. P. Volatile hydrocarbon biodegradation by a mixed bacterial culture during growth on crude oil. J. Ind. Microbiol. Biotechnol. 2001, 26, 356-362. Organisation for Economic Cooperation and Development. OECD Guidelines for Testing of Chemicals; First Addendum; ISBN: 92-64-12221-4; OECD: Paris, 1993; http://www.oecd.org. Bushnell, L. D.; Haas, H. F. The utilization of certain hydrocarbons by microorganisms. J. Bacteriol. 1941, 41, 653-673. Uhler, R. M.; Healey, E. M.; McCarthy, K. J.; Uhler, A. D.; Stout, S. A. Molecular fingerprinting of gasoline by a modified EPA 8260 gas chromatography-mass spectrometry method. Int. J. Environ. Anal. Chem. 2003, 83, 1-20. Prince, R. C.; Douglas, G. S. Quantification of hydrocarbon biodegradation using internal markers. In Manual of Soil Analysis - Monitoring and Assessing Soil Bioremediation; Margesin, R., Schinner, F., Eds.; Springer-Verlag: Berlin, 2005; pp 179-188. Environment Canada. Persistence and Bioaccumulation Regulations. Canada Gazette II 2000, 134, 607-612. European Commission. Technical Guidance Document in support of Commission Directive 93/67/EEC on risk assessment for new notified substances and Commission Regulation (EC) No 1488/94 on risk assessment for existing substances (1996), 2nd ed.; ISBN 92-827-8012-0; EC: Brussels, 2003; http://ecb.jrc.it/ Documents/TECHNICAL_GUIDANCE_DOCUMENT/ EDITION_2/tgdpart2_2ed.pdf. Beens, J.; Brinkman, U. A. D. The role of gas chromatography in compositional analyses in the petroleum industry. Trends. Anal. Chem. 2000, 19, 260-275. Sakuratani, Y.; Yamada, J.; Kasai, K.; Noguchi, Y.; Nishihara, T. External validation of the biodegradability prediction model CATABOL using data sets of existing and new chemicals under the Japanese Chemical Substances Control Law. SAR QSAR Environ. Res. 2005, 16, 403-431. Rouvie`re, P. E.; Chen, M. W. Isolation of Brachymonas petroleovorans CHX, a novel cyclohexane degrading β-Proteobacterium. FEMS Microbiol. Lett. 2003, 227, 101-106. Solano-Serena, F.; Marchal, R.; Heiss, S.; Vandecasteele, J.-P. Degradation of isooctane by Mycobacterium austroafricanum IFP 2173: growth and catabolic pathway. J. Appl. Microbiol. 2004, 97, 629-639. U.S. EPA. EPI-suite; 2007; www.epa.gov/oppt/exposure/docs/ episuite.htm.
Received for review December 5, 2006. Revised manuscript received March 6, 2007. Accepted March 7, 2007. ES062884D
VOL. 41, NO. 9, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3321
