The isotopic compositions of selected crude oil PAHs during biodegradation

The isotopic compositions of selected crude oil PAHsduring biodegradation

Peter J. Yanika, Thomas H. O’Donnella,1, Stephen A. Mackoa,*,Yaorong Qianb, Mahlon C. Kennicutt IIb

aDepartment of Environmental Sciences, University of Virginia, Charlottesville, VA 22903, USAbGeochemical and Environmental Research Group, Texas A&M University, College Station, TX 77845, USA

Abstract

Compound specific isotope analysis (CSIA) has been used to measure the d13C of selected individual polycyclicaromatic hydrocarbons (PAHs) from a crude oil that was exposed to enhanced microbial biodegradation. Twenty-one

liters of a crude oil were introduced to series of separate 5 m�5 m wetland plots along the San Jacinto River, Texas.Smaller grids were established within each area to serve as control plots, plots for addition with nitrogen fertilizer, or asplots for addition with petroleum degrading microbes. Samples of randomly selected, oil impacted soils were collectedat 4, 28, and 56 days after oil inoculation. These samples initially were analyzed using GC/MS for identification of

PAH compounds. The initial results on the soils indicated a tendency for samples to shift from low and high molecularweight PAH compounds to eventually predominantly high molecular weight compounds. Fifteen specific PAH com-pounds, including two containing sulfur, were selected for compound specific isotope analysis based on chromato-

graphy and expected abundance in the soils. Five alkylated components, a dimethylnaphthalene, twotrimethylnaphthalenes, a methyldibenzothiophene, and a dimethylphenanthrene, were seen to be best resolved and aresuggested to be the most reliable for observations of isotopic changes during biodegradation. Isotopic variability dur-

ing the treatment period follows patterns with an overall trend toward enrichment in the residual PAH pool of 2–8%.All treatment grids, including the control, underwent isotopic alteration throughout the duration of the experiments.ANOVA testing of the isotopic results indicated that the control grids showed the smallest amount of changes over

time, whereas the nutrient and microbe amended grids were not significantly different from each other. Although thetrends on isotopic enrichments were predictable, the magnitude of changes was not. One additional finding indicatedthat time trials beyond 2-month periods were needed to determine if nutrient or microbial additions are advantageousfor PAH biodegradation in wetland soils. The results of these tests highlight the importance of understanding the

magnitude and direction of isotopic variations in PAH compounds as part of investigations aimed at tracing thesecompounds from potential sources.# 2002 Elsevier Science Ltd. All rights reserved.

1. Introduction

Polycyclic aromatic hydrocarbons (PAHs) are ubiqui-tous compounds in the environment (Laflamme and Hites,

1978). Formed from thermal alteration of buried organic

matter (petrogenic) or through incomplete combustion of

fossil fuels or biomass (pyrogenic), PAHs are ubiquitous inecosystems. Because fossil fuels are used extensively,anthropogenic inputs are significant contributors of petro-

genic and pyrogenic PAHs into the environment. Further-more, biomass burning of forests and grasslands andhuman-induced fires are also significant source of PAHs inthe environment. In recent decades there has been

increased attention focused on these compounds as poten-tial negative health effects associated mainly with themetabolites of PAHs have been identified.

0146-6380/03/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.

PI I : S0146-6380(02 )00164-X

Organic Geochemistry 34 (2003) 291–304

www.elsevier.com/locate/orggeochem

* Corresponding author.1 Present address: Academy of Natural Sciences, Patrick

Center for Environmental Research, Philadelphia, PN 19103,

USA.

http://www.sciencedirect.com



http://www.elsevier.com/locate/orggeochem/a4.3d

Petrogenic PAHs, such as those found in crude oil,are characterized by considerable amounts of alkyl-substitution. Pyrogenic PAHs can be unalkylated, oralkylated, depending on the temperatures achieved dur-

ing combustion. Generally, the higher the temperature,the more parent (unalkylated) forms of PAHs that areproduced, with the degree of alkylation increasing with

decreasing temperature (Hites et al., 1980; Harvey,1997). Like a number of other hydrocarbons, PAHs arevery hydrophobic, which leads to their strong partition-

ing to sediments, and when present, biological tissue(Wilson and Jones, 1993; Roper et al., 1997 and refs.therein). PAHs are considered to be semi-volatile with

volatility decreasing with increasing molecular weight.This characteristic, coupled with the hydrophobicity ofPAHs and the tendency to sorb to particulate matterleads to the persistence of these compounds in the

environment. Once in sediments, the PAHs tend toremain there, although there can be redistribution,remobilization and subsequent partitioning between

other phases such as biological tissue. The lipophilicPAHs can bioaccumulate in the tissue, and in somecases, be metabolized by the organism (Roper et al.,

1997; Hellou et al., 1995; Escartı́n and Porte, 1999).Certain PAHs, such as benzo(a)pyrene, are the pri-

mary cancer-causing agents in some fossil fuel products

(Dipple, 1985 and refs. therein). However, it is not theparent molecule, but rather the oxidized metabolites ofPAHs that exhibit mutagenicity and carcinogenicity. Inparticular, the dihydrodiol, the dihydrodiol epoxide and

the diol epoxide derivatives of certain PAHs [such asbenzo(a)pyrene] have been identified as the problematiccompound classes (Yang et al., 1985). Additional work

has shown that methyl substitution also enhances thecarcinogenic character in PAHs (Hecht et al., 1985).Much of the recent research on PAHs has focused on

characterizing emissions from primarily anthropogenicsources and developing techniques for tracing the sour-ces of these contaminants. Many of the efforts forapportioning sources of PAHs have utilized compound

ratios or compound class ratios (Maldonado et al.,1999). A special emphasis has been placed on alkylatedhomologues of naphthalene, phenanthrene, diben-

zothiophene, fluorine and chrysene. It has been pro-posed that by taking the abundance ratio of the sum of‘‘other’’ three to six ring PAHs to the sum of the above

series of PAHs, one could arrive at whether the PAHswere combustion derived or petrogenic in nature (Wanget al., 1999). In differentiating sources of purely anthro-

pogenically-derived combustion products, certain PAHshave been used as indices of each type of combustion.The predominance of a suite of PAHs over all otherscould narrow the list or be used identify source of PAHs

affecting an area (Khalili et al., 1995).Neglected in many of these approaches is the poten-

tial problem of alteration of the signal of one compound

or compound class relative to another. Chemical or bio-logical alteration of the abundances and distribution ofcertain PAHs would add considerable confusion to thesource apportionment investigation. There is evidence of

PAH degradation even under anaerobic conditions(Coates et al., 1996, 1997). Once valid source–con-tamination site relationships may be made to appear

invalid owing to these altered concentration relation-ships. One approach to surmount this problem is to useconcentration relationships of only the most refractory

compounds when trying to establish sources of con-taminants (Hostettler et al., 1999). This technique hasshown some promise in early work, although it relies on

the fact that the refractory compounds of interest wereoriginally present at the site(s) in question.In recent years, the technique of compound specific

isotope analysis (CSIA) has found applications for a

variety of biogeochemical questions. Because the stableisotopic signature of individual compounds is deter-mined, added resolution to traditional molecular ana-

lyses is gained. The stable isotopic signature of eachcompound of interest gives an indication of the historyof the compound. In most cases, the details that

researchers are interested in are centered on the reac-tants that went into producing a particular compoundor what reservoir it came from. Owing to the utilization

of different reactants and production processes, twocompounds with identical elemental compositions andatomic arrangements would have unique isotopic sig-natures. For contaminants that were derived from a

certain company, combustion process or natural reser-voir, unique stable isotopic signatures would serve as afingerprint for tracing that compound in the environ-

ment. This approach to studying isotopic signatures ofcontaminants has aided in the apportioning of sourcesof these compounds. Fundamental to this research is the

theory that CSIA approaches are less subject to chemi-cal and biological alterations than traditional moleculartechniques. Because each compound is investigatedseparately, the technique does not rely on having

numerous compounds or compound classes remainunaltered for their entire history. Compound specificisotope analysis thus has the potential for being very

robust in light of the effects of chemical or biologicalalteration because each compound can be treated andinterpreted independently rather than being part of an

interpretation based on the interdependence of severalcompounds.With respect to PAHs, compound specific isotope

analysis first was used to apportion sources of anthro-pogenic compounds, such as those found in varioussoots and oils transported by urban runoff (O’Malley etal., 1994). Distinct isotopic signatures were observed for

many PAHs and were utilized in apportioning sourcesof the contaminants extracted from harbor sediments.Owing to different isotope fractionations in C3 and C4

292 P.J. Yanik et al. / Organic Geochemistry 34 (2003) 291–304

vegetation, insight has also been gained on determining thesources of PAHs derived from biomass burning (O’Malleyet al., 1997) and between anthropogenic and biomass-derived compounds (Ballentine et al., 1996). CSIA has also

been used to differentiate between coal-derived PAHs thatwere produced under varying processing conditions(McRae et al., 1999). In most of the above applications of

CSIA, PAH isotopic compositions were determined forcompounds that had not been exposed to possible bio-degradation. However, chemical or biological altera-

tions may affect the isotopic compositions of thecompounds of interest. If there were relatively smallchanges in concentration of a few compounds because

of biodegradation, the concentration ratios wouldpotentially be more affected than the isotopic composi-tions. However, it has not been established whether theisotopic compositions of some, none, or many PAHs

would be affected by extensive biodegradation. Regard-less of the concentration changes, if the isotopic com-positions of only some of the PAHs were affected,

source apportionment could still be achieved by CSIA.Furthermore, fractionation due to biodegradationshould be predictable, as the microbes would tend to

attack the isotopically depleted (containing less 13C)molecules. Knowledge of fractionations associated withbiodegradation would aid greatly in source apportion-

ment studies when systems have experienced complexhistories.The biodegradation of PAHs is mediated by a range of

microorganisms, the most often studied being bacteria.

Some of the most common bacteria genera encounteredor used for biodegradation are Alicaligenes, Pseudomo-nas, Bacillus, Cyanobacter, Vibrio and Mycobacterium

(Wilson and Jones, 1993 and refs. therein). There havebeen a number of studies focused on monitoring thebiodegradation of PAHs. Phenanthrene, fluorene and

fluoranthene have been shown to be biodegraded by amixed culture of bacteria (Weissenfels et al., 1990).Others have shown numerous PAH species were biode-graded, but the smaller molecules were preferentially

attacked (Mahaffey et al., 1991). Enhanced biodegrada-tion has been observed under denitrifying conditions,with nitrate apparently serving as a complementary or

replacement electron acceptor for oxygen (McRae andHall, 1998). Others have reported no observable biode-gradation of PAHs at contaminated sites even when

microbes capable of degrading PAHs were present. Thelikely cause of this was that the PAHs were stronglybound to the soil and unavailable for biodegradation

(Erickson et al., 1993). It has never been conclusivelyshown whether biodegradation alters the isotopic sig-nature of PAHs, and if so, the magnitude of this effectand the trend with the extent of degradation. There is

some evidence that for low molecular weight PAHs, theeffect is minimal (O’Malley et al., 1994; Meckenstock etal., 2001; Lollar, et al., 1999; Ahad et al., 2000) but the

effects for higher molecular weight molecules has notbeen investigated. If CSIA is to be used effectively forsource apportionment of PAHs, knowledge of the pos-sible effects of biodegradation is vital, especially if the

PAHs have resided in areas, such as sediment, thatcontain populations of PAH-degrading microbes. Theextent and general trends of any isotope fractionations

will aid in not only assessing the general applicability ofCSIA for sites of possible biodegradation, but will alsoaid future studies that monitor the fates of PAH in the

environment. The overall purpose of this study, there-fore, was to monitor in a controlled experiment, theisotopic composition of a suite of PAHs as they were

exposed to varying conditions of degradation over time.

2. Methods

2.1. Experimental site and treatment

The biodegradation experiment was carried out in thefield utilizing indigenous microbes for PAH and generalhydrocarbon degradation. For this experiment, a series

of 5 m�5 m plots were established in an uncontami-nated wetland adjacent to the San Jacinto River,approximately 30 km east of Houston, Texas. The soils

were generally uniform across the entire experimentalplot, and were not anaerobic. The plots were in directsunlight over the entire term of the experiment, whichran from May through July, being exposed to daytime

temperatures commonly in excess of 30 �C. The experi-mental plots were further broken down into 26�26 cellgrids. To each of the plots, 21 l of Arabian medium

crude oil were added. The oil remained on the surface ofthe plot for the duration of the experiment, penetratingto a depth at most of a few centimeters. Three of the

plots were further amended with a commercial inorganicfertilizer (the nutrient amendment), three of the plotswere untreated (the control), and three of the plots wereamended with a commercial product containing petro-

leum-degrading microbes. For the first 5 weeks of thisexperiment, nutrients were added weekly to achieverelatively constant nitrogen concentrations. The target

level was 40 mg nitrogen per kilogram of dry soil (sumof ammonia plus nitrate) for the nutrient-amended soil.After the first 5 weeks, treatments were carried out at

2-week intervals. For this study, samples were obtainedof each treatment type after 4, 28 and 56 days. Samplingwas done randomly from each plot (using a random

number generator) by obtaining a sample from one ofthe 26 cells for all of the nine plots. After each sampling,the adjacent cells were removed from consideration,such that each future sample could be obtained without

concern about effects from nearby sampling. The sam-ples were taken with a coring device to a depth of 25 cmand homogenized prior to extraction and analysis.

P.J. Yanik et al. / Organic Geochemistry 34 (2003) 291–304 293

2.2. Extraction of PAHs and analysis

Approximately 20 g of freeze-dried sediment wasSoxhlet extracted with 100 ml of distilled dichloromethane

for 12 h. The initial sets of compound classes were firstseparated using column chromatography with an alu-mina/silica stationary phase. Aliphatic hydrocarbons were

eluted using 50 ml of pentane. Aromatic hydrocarbonswere then eluted using 150 ml of pentane/dichloro-methane (1:1, v:v). The aromatic hydrocarbon fraction

was then further purified with high-performance liquidchromatography (HPLC) using a cyano/amino bondedphase column as per Killops and Readman (1985). The

aromatic hydrocarbons were separated based on thenumber of double bonds in the molecule (Killops andReadman, 1985). A pentane and dichloromethane gra-dient was used for the elution. The aromatic fractions of

interest were then combined and taken up in dichloro-methane to a volume of approximately 0.5 ml.Each of the nine samples were first analyzed by GC/

MS to establish a temperature program that providedthe best resolution for the PAHs, to determine whatcompounds were present, and the relative contribution

of each. A Hewlett-Packard 5890 Series II GC with a 30m DB-5 column (0.25 mm i.d., 0.25 mm film thickness;J&W Scientific, Folsom, CA) interfaced to a Hewlett-

Packard 5971A mass selective detector (MSD) was usedfor these initial analyses. The temperature program was70 �C for 4.5 min, ramp at 2�/min to 230 �C, hold for 10min, ramp at 3�/min to 275 �C and hold for 15 min The

GC was operated in splitless mode, the column headpressure was 10 psi and the carrier gas was helium. TheMSD was operated in full scan mode (50–450 amu), and

peaks were identified using the factory-installed operat-ing software.

2.3. GC/C/IRMS

For compound specific isotope analyses, a Hewlett-Packard 5890 Series II gas chromatograph (GC) was

interfaced to a Micromass Isoprime (Manchester, UK)isotope ratio mass spectrometer (IRMS) through acombustion (C) furnace. The GC had a flame ionization

detector (FID) as a secondary detector and it was heldat 280 �C. The combustion furnace was operated at850 �C and the IRMS was operated in continuous flow

mode and tuned to analyze stable carbon isotopes. TheGC column was a 60 m DB-5 (0.32 mm i.d., 0.25 mmfilm thickness), and was connected between the GC

injection port and a VSOS fitting (SGE, Austin, TX). Apiece of capillary column ran from the VSOS fitting to aheart split valve and the combustion furnace. The heartsplit valve was pneumatically actuated to divert the car-

rier gas flow to and from the FID and IRMS. High purityhelium (99.9999%) was used as the carrier gas and theGC was operated in splitless mode. The temperature

program was 70 �C for 4.5 min, ramp at 2�/min to 230�,hold for 10 min, ramp at 3�/min. to 275� and hold for 20min Water was removed from the effluent of the com-bustion furnace by an ethanol/CO2 cryogenic trap. The

performance of the GC/C/IRMS system was regularlytested by injecting a standard of known isotopic com-position. Each sample was analyzed multiple times to

test the precision of the measurements.All stable carbon isotopic data is reported below as

delta (d) values in the ‘‘per mil’’ (%) notation, and all

values reported are relative to the international stan-dard, Pee Dee Belemnite (PDB). Although the repro-ducibilities of isotopic compositions of well-resolved

components of the original oil nor of a single compoundwere focii of this study, previous work from ourlaboratory and others has reported typical standarddeviations of 0.3% for analyses of well-resolved com-

ponents (Ballentine et al., 1996; O’Malley et al., 1997).

3. Results

3.1. GC/MS data

Representative GC/MS chromatograms showingcompound presence and change over time for the one of

the treatments (nutrient amendment) are shown (Figs. 1–3). It should be noted that although the same volume ofeach sample was injected (1.5 ml), absolute concentra-tion values were not determined for these samples owing

to inconsistencies in the volumes of the samples afterextraction and cleanup procedures. A wide variety ofPAHs were found in the samples, with a considerable

number of the alkylated homologues. For each timestep, there were peak to peak variations and after 56days, there was a shift toward higher molecular weight

PAHs for the nutrient amended sample. For a giventreatment, the trend was for a transition from a mixtureof low and high molecular weight compounds to a pre-dominance of the higher molecular weight species.

3.2. GC/C/IRMS data

In order to track the effects of biodegradation on theisotopic composition of PAHs, compounds had to bechosen that showed measurable concentrations in most,

if not all, of the samples. In addition, the process ofdetermining the isotopic compositions of all compoundswithin complex mixtures, such as these PAH samples,

and may not be necessary. Because of this, a subset of thecompounds were chosen for this study, with an emphasison compounds found in most or all of the samples. Fif-teen peaks in all were chosen, with representation from

early to late elution times. Thirteen of the compoundswere ‘‘traditional’’ PAHs, while two were sulfur con-taining analogs (Table 1). Exact identifications of the


peaks were not performed for this study because therewere often a large number of possible library matchesfor a given alkylated PAH, and exact peak identification

was not vital to the success of the study. Instead, generalclassifications were determined and were based on theparent PAH molecule, and each of the 15 peaks was

identified in each sample by the mass spectrum coupledwith the retention times (Table 1).For each of the nine samples, stable isotopic compo-

sitions were determined for the 15 peaks of interest. Allsamples were analyzed multiple times (n55) in order toobtain a measure of the precision of the analyses in the

Fig. 1. GC/MS chromatogram of PAHs extracted after four days from sediment that was amended with nutrients. The 15 peaks of

interest for this study are labeled. Peaks labeled with numbers represent ‘‘true’’ PAHs while those labeled with letters represent sulfur

containing analogs of PAHs. Compound identifications can be found in Table 1.

Fig. 2. GC/MS chromatogram of PAHs extracted after 28 days from sediment that was amended with nutrients.


complex mixtures. The mean of the isotopic composi-

tions was determined for each compound in each sam-ple, and standard deviations were computed. Owing tothe complexity of the chromatograms, and the fact that

some peaks dropped below detection limits for somesamples, isotope values were not able to be determined

for every peak in all samples. When the number of iso-tope values computed for a compound in a sample wasbelow three, standard deviations are not reported.The 15 peaks of interest were originally chosen

because they showed appreciable abundance in the GC/MS chromatograms and they could be resolved. Uponanalysis on the GC/C/IRMS system, however, some of

these peaks lost some of their resolution. Additionalpeaks from degradation products of the PAHs, andperhaps unavoidable ‘‘pooling’’ zones in the GC/C/

IRMS system may have served to complicate matters orworsen the resolution. With this is mind, isotopic com-positions were determined for all fifteen peaks whenextreme coelution and/or low abundances did not pre-

clude this (Table 2). As is typical of fossil fuel-derivedcompounds, the isotopic composition of most of thepeaks were in the range of �20 to �30 per mil (%). The

precision for replicate analyses was usually less than 1%,although there were some exceptions. Attention was paidduring the processing of the raw isotopic data as to what

peaks from each sample were the best resolved, and there-fore most likely represented the most precise data. Becauseof the necessity for reliable isotopic values when monitor-

ing any possible fractionations associated with biode-gradation, those compounds that were regularly the bestresolved were chosen as a focus for the remainder of thestudy. The peaks chosen were numbers 2, 5, 8 and 13 as

well as the peak labeled B (Fig. 1, Table 1). Of this subset,peak 13 was regularly the best resolved, as it showed onlyvery minor coelution for some of the analyses.

Fig. 3. GC/MS chromatogram of PAHs extracted after 56 days from sediment that was amended with nutrients.

Table 1

The 15 compounds of interest, along with their peak labels and

approximate retention times for the GC/MS chromatograms

Compound Peak no. Approximate

retention

time (min)a

Methylnaphthalene 1 22.99

Dimethylnaphthalene 2 29.94



Trimethylnaphthalene 5 36.11





Dimethyl-9H-fluorene 10 54.2

Methyldibenzothiophene A 55.32

Methyldibenzothiophene B 56.4

Methylanthracene 11 58.68

Methylphenanthrene 12 58.9

Dimethylphenanthrene 13 64.36

a The retention times vary by as much as 0.1 min between

analyses. The retention times shown are those from Fig. 1. The

compound identification are general, as the library on the soft-

ware package could not differentiate between the different

alkylated homologues.


The data from these five compounds are summarizedin graphical form in Figs. 4 and 5. In all cases the errorbars represent �1 SD. Fig. 4 shows the trend in carbonisotopic data over time for the three treatments. For

many of the peaks for the three treatments, the mostdepleted isotopic values were for those compoundsextracted after 4 days. A noticeable exception was for

the microbe treatment in which compounds 5,8 and Bwere the most depleted after 28 days (Fig. 4c). In gen-eral, the compounds extracted from the sediments were

the most enriched after 56 days, although there was someoverlap and reversal of this trend with those compoundsextracted after 28 days (Fig. 4a–c). Overall, the isotopiccomposition of the compounds became enriched by

2–8%. For all three treatments, compounds 2 and 13were the only two that showed the consistent trend ofincreasing enrichment in 13C over time (Fig. 4a–c).

Fig. 5 shows the difference between the three treat-ments at each time step. Overall, the isotopic values forthe three treatments showed similar compound to com-

pound trends for each time step (Fig. 5a–c); again, theexception was for peaks 5, 8 and B after 28 days for themicrobe treatment (Fig. 5b). Fig. 5a reinforces the fact

that after only 4 days, most compounds showed isotopicdifferences between the three treatments. Early in theexperiment, the compounds extracted from the sedimenttreated with nutrients were the most depleted (Fig. 5a)

and this trend continued late in the experiment (Fig. 5c).Compounds 5 and 8 (excluding the very depleted con-trol treatment values after 28 days) were affected the

most by a change in treatment. In general, compound 13showed the least but most consistent isotopic changewith treatment (Fig. 5a–c).Although there were isotopic differences for each

compound over time and between treatments, someanalyses did overlap within the precision of themeasurement. For further analysis, the data were parti-

tioned into two broad classifications: changes with timeand changes with treatment. For each of the five peaksof interest, either time or treatment was held as the

independent variable while the other was treated as thedependent variable. A single factor analysis of variance(ANOVA) test was run for the data from the dependent

variable to find out if there were any isotopic differ-ences. If statistically significant differences wereobserved at the 0.05 confidence level, two-tailed t-tests(assuming equal variances) were run between each of the

possible combinations to determine where the isotopicdifferences were originating. If the ANOVA test showedthat there were no statistically significant differences, no

t-tests were run. These data are summarized inTables 3–12. The first five tables (3–7) represent thestatistical analyses for the compounds of interest with

treatment acting as the independent variable. The latterfive tables (8–12) represent the statistical analyses withtime acing as the independent variable.

Significant isotopic differences between treatmentswere observed for all five peaks (Tables 8–12). Most ofthe differences between treatments were observed after 4or 28 days, with only one difference observed after 56

days (for compound 2). All forms of treatment to treat-ment variation were observed (nutrient-control, nutri-ent-microbes, microbes-control) with no clear

predominance of one form of inter-treatment variation(Tables 8–12). With respect to changes over time, com-pounds 8 and 13 showed the most variation (Tables 5

and 7) while compound B showed no change in time forany of the treatments (Table 6). None of the treatmentsshowed a predominance of changes over time for thefive compounds, although the untreated sediment

showed the fewest changes over time (Tables 3–7).Compounds 2, 8, B and 13 appear to show isotopic dif-ferences over time for the nutrient amended sediment.

Tables 3 and 6 show that the differences are not statis-tically significant for compounds 2 and B, but are so forcompounds 8 and 13 (Tables 5 and 7). For compound

13, there were significant isotopic changes between days4 and 56 and between days 28 and 56 (Fig. 4a andTable 7). In the untreated (control) sediment, the iso-

topic changes appear to be less pronounced, with onlycompounds 8 and 13 potentially showing isotopic differ-ences (Fig. 4b). The statistical analyses show that the iso-topic differences only occurred for compound 13, between

days 4 and 28 and between days 4 and 56 (Table 7). For themicrobe-amendment, compounds 5, 8 and B, and perhapscompound 2 showed potential isotopic differences

Table 2

Representative stable carbon isotopic data from this studya

Compound Peak

no.

d13C(vs. PDB)

Precision

(SD)

Methylnaphthalene 1 �27.5 0.54

Dimethylnaphthalene 2 �27.1 0.90



Trimethylnaphthalene 5 �25.7 1.00





Dimethyl-9H-fluorene 10 �25.3

Methyldibenzothiophene A �22.7 0.95

Methyldibenzothiophene B �26.6 0.42

Methylanthracene 11 �20.5 1.30

Methylphenanthrene 12 �25.8 0.84

Dimethylphenanthrene 13 �25.8 1.11

a These data are for the compounds extracted from the

nutrient amended sediment after 4 days. Isotopic values are

reported relative to PDB and the precision values represent one

standard deviation. The lack of precision value for compound

10 is because less than three separate isotopic values were

determined for that compound.


(Fig. 4c). There were isotopic differences between days 4and 56 for compound 2, between days 4 and 28 and days28 and 56 for compounds 5 and 8, but no significant iso-

topic differences over time for compound B (Tables 3–6).Similar to the temporal variations, the graphical

appearance of isotopic variations between treatments

was also verified with statistical analyses. Fig. 5a showsthat compounds 5, 8 and B appear to show changes inisotopic composition after 4 days because of the differ-

ent treatments. The differences between the microbeamended and both the nutrient amended and controlsediment are statistically significant for compound 5

Fig. 4. Plots showing the isotopic composition of the five compounds of interest for this study over time for the three treatments. Plot

(a) represents the time series for the nutrient treatment, plot (b) represents the time series for the control treatment, and plot (c)

represents the time series for microbe amended treatment. The isotopic values are reported relative to PDB (Pee Dee Belemnite). Error

bars represent �1 SD from the average value.


(Table 9). There are also statistically significant differ-ences between the nutrient and control treatments andthe between the microbe and control treatments forcompound 8 (Table 10). Compound B has significant

isotopic differences between the nutrient and controltreatments and between the nutrient and microbe treat-ments (Table 11). Although not graphically obvious,

there is also a significant isotopic difference between thecontrol and microbe treatment for compound 13(Table 12). After 28 days, only compound 8 appears tohave isotopic differences (Fig. 5b) (excluding the

microbe treatment data for compound 5,8 and B). Thisis verified in Table 10 that shows that there were sig-nificant isotopic differences between the nutrient and

Fig. 5. Plots showing the change in isotopic composition of the five compounds of interest with treatment for the three time steps of

this study. Plot (a) represents the isotopic composition of the compounds for the three treatments after 4 days, plot (b) after 28 days

and plot (c) after 56 days.


control treatments. After 56 days, much of the isotopicdata between treatments appears to be similar (Fig. 5c),

and the statistical analyses show that only a differencebetween the microbe and nutrient treatments for com-pound 2 existed (Table 8).

4. Discussion

The results of this investigation showed that biode-

gradation has an effect on the isotopic composition ofindividual PAH molecules. Where biodegradationoccurs, there would be an expected change not only in

the concentration of some PAHs but also a trendtoward isotopic enrichment of the residual PAH mate-rial. This is compatible with theory where a kinetic iso-tope effect occurs and isotopically lighter molecules

react (are utilized by the microbes) preferentially com-pared to the heavier molecules. The GC/MS chromato-grams showed a general trend toward a predominance

Table 3

Results of the statistical analysis of the isotopic data for com-

pound 2 for each treatment with varying timea

Peak 2 ANOVA Different

than Day 4?

Different

than Day 28?

Different

than Day 56?

Nutrients

Day 4

Day 28 0.312

Day 56

Control

Day 4

Day 28 0.818

Day 56

Microbes

Day 4 – N Y (0.014)

Day 28 0.034 N – N

Day 56 Y (0.014) N –

a Analysis of variance (ANOVA) was performed on the time

data for each treatment. When ANOVA results showed sig-

nificant differences at the 0.05 confidence level, two-tailed

t-tests assuming equal variance were performed between the

different time step combinations. The t-test results are shown in

matrix form. When significant results were found at the 0.05

confidence level, the P-values are displayed.

Table 4


pound 5 for each treatment with varying time. Descriptions of

the statistical tests are the same as for Table 3


than Day 4?

Different

than Day 28?

Different

than Day 56?

Nutrients

Day 4

Day 28 0.146

Day 56

Control

Day 4

Day 28 0.322

Day 56

Microbes

Day 4 – Y (0.011) N

Day 28 0.004 Y (0.011) – Y (0.003)

Day 56 N Y (0.003) –

Table 5





than Day 4?

Different

than Day 28?

Different

than Day 56?

Nutrients

Day 4 – Y (0.000009) N

Day 28 0.001 Y (0.000009) – Y (0.013)

Day 56 N Y (0.013) –

Control

Day 4

Day 28 0.14

Day 56

Microbes

Day 4 – N Y (0.0000002)

Day 28 0.0005 N – Y (0.005)

Day 56 Y (0.0000002) Y (0.005) –

Table 6


pound B for each treatment with varying time. Descriptions of


Peak B ANOVA Different

than Day 4?

Different

than Day 28?

Different

than Day 56?

Nutrients

Day 4

Day 28 0.069

Day 56

Control

Day 4

Day 28 0.396

Day 56

Microbes

Day 4

Day 28 0.091

Day 56


of larger molecular weight PAHs over time for eachtreatment (Figs. 1–3). This trend matches that found byothers in which the smaller PAH molecules are more

prone to microbial attack and breakdown (Mahaffey etal., 1991). Although most of the molecules shown in theGC/MS chromatograms were lower to mediummolecular

weight (2–4 rings in size), one would expect the trend of

decreasing lability with number of rings to be fairlyconsistent along the entire molecular size continuum.Similarly, if the concentrations of the PAH mixtures

would have been known at the start of the experiment, itwould also have been expected that the absolute con-centrations of PAHs would decrease as a whole over time,with the small molecules being eliminated more rapidly.

The trends associated with changes in treatment were notas apparent when compared with the temporal changes.In general, there were not appreciable differences in PAH

Table 7





than Day 4?

Different

than Day 28?

Different

than Day 56?

Nutrients

Day 4 – N Y (0.003)

Day 28 0.009 N – Y (0.042)

Day 56 Y (0.003) Y (0.042) –

Control

Day 4 – Y (0.002) Y (0.0006)

Day 28 0.0004 Y (0.002) – N

Day 56 Y (0.0006) N –

Microbes

Day 4

Day 28 0.157

Day 56

Table 8


pound 2 for the various treatments at each time stepa


than

Nutrients?

Different

than

Control?

Different

than

Microbes?

Day 4

Nutrients

Control 0.56

Microbes

Day 28

Nutrients

Control 0.46

Microbes

Day 56

Nutrients – N Y (0.029)

Control 0.038 N – N

Microbes Y (0.029) N –

a Analysis of variance (ANOVA) was performed on the

treatment data for each time step. When the ANOVA results

showed significant differences at the 0.05 confidence level, two-

tailed t-tests assuming equal variance were performed between

the different treatment combinations. The t-test results are

shown in matrix form. When significant results were found at

the 0.05 confidence level, the P-values are displayed.

Table 9


pound 5 for the various treatments at each time step. Descrip-

tions of the statistical tests are the same as for Table 8


than

Nutrients?

Different

than

Control?

Different

than

Microbes?

Day 4


Control 0.037 N – Y (0.041)

Microbes Y (0.042) Y (0.041) –

Day 28


Control 0.001 N – Y (0.003)

Microbes Y (0.003) Y (0.003) –

Day 56

Nutrients

Control 0.187

Microbes

Table 10


pound 8 for the various treatments at each time step. Descrip-



than

Nutrients?

Different

than

Control?

Different

than

Microbes?

Day 4

Nutrients – Y (0.019) N

Control 0.011 Y (0.019) – Y (0.016)

Microbes N Y (0.016) –

Day 28

Nutrients – Y (0.002) Y (0.003)

Control 0.001 Y (0.002) – N


20.52

Day 56

Nutrients

Control 0.096

Microbes


distributions after 4 or 28. After 56 days, the chroma-

tograms did tend to show a slight skewing toward thelarger molecules, especially for the nutrient amendedtreatment. It was expected that the addition of nutrients

or microbes known to attack petroleum would allow formore rapid biodegradation of the PAHs in the addedpetroleum. From a molecular concentration viewpoint,this was observed only to a slight extent over the short

time period of the experiments.The isotopic data from this study generally agrees with

the expected enrichment as the extent of biodegradation

increases. For many of the compounds, the most deple-ted values were early in the experiment, and the mostenriched isotopic values after 28 and 56 days (Fig. 4a–c).As noted above, the isotopic composition of the com-

pounds changed by 2–8%, indicating a substantialenrichment caused by biodegradation. In particular,compounds 8 and 13 (a trimethylnaphthalene and a

dimethylphenanthrene, respectively) showed statisticallysignificant changes with time (Tables 5 and 7), withcompounds 2 (a dimethylnaphthalene) and 5 (a tri-

methylnaphthalene) showing some statistically sig-nificant changes over time (Tables 3 and 4). CompoundB (a methyldibenzothiophene) did not show any sig-

nificant isotopic changes over time (Table 6), which mayindicate a microbial inhibition perhaps as a result of thepresence of sulfur. Overall, although there were sig-nificant isotopic changes over time, there were also some

inconsistencies. In particular, there were occasions whenthe isotopic values for some compounds were moreenriched after 28 days than after 56 days (Fig. 4a–c).

Although some of these differences were not statisticallysignificant, one possible general explanation is that therewas perhaps spatial heterogeneity in the overall intensity

of biodegradation and the impact on certain com-pounds. A sample collected for extraction after 28 daysmay have come from a location where there was locally

more intense biodegradation than for the sampleextracted after 56 days. Another inconsistency was theisotopic values for compounds 5, 8 and B for themicrobe treatment show a very depleted signature after

28 days (Fig. 4c). The overall abundances of all com-pounds in this sample were barely above detection lim-its. Because of this, determining isotopic values for some

of the compounds was difficult, and not achievable insome of the replicate analyses. Thus, for compounds 5,8and B from this sample, isotopic values were able to be

determined for an appropriate number of analyses to geta representative isotopic composition or precision(standard deviation) values.In this investigation, it was also expected that there

would be differences in the extent of biodegradation forsamples amended with nutrients, known petroleum-degrading microbes, or not treated at all. In particular,

it was expected that the nutrient and microbe amendedtreatments would induce more extensive biodegradationthan for the control sediment. Along with more exten-

sive biodegradation, more pronounced isotopic fractio-nations were expected. As noted above, only after 56days were noticeable differences observed in concentra-

tions of PAHs between treatments. Isotopically, thecontrol-treated sediment was expected to show the mostdepleted values for all compounds at all times. Thenutrient amended treatment did tend to yield the most

isotopically depleted values, however. As noted earlier,there were statistically significant differences in the iso-topic compositions of some of the compounds between

Table 11


pound B for the various treatments at each time step. Descrip-


Peak B ANOVA Different

than

Nutrients?

Different

than

Control?

Different

than

Microbes?

Day 4

Nutrients – Y (0.002) Y (0.010)

Control 0.001 Y (0.002) – N


Day 28

Nutrients

Control 0.326

Microbes

Day 56

Nutrients

Control 0.902

Microbes

Table 12


pound 13 for the various treatments at each time step.

Descriptions of the statistical tests are the same as for Table 8


than

Nutrients?

Different

than

Control?

Different

than

Microbes?

Day 4

Nutrients – N N

Control 0.025 N – Y (0.011)

Microbes N Y (0.011) –

Day 28

Nutrients

Control 0.406

Microbes

Day 56

Nutrients

Control 0.326

Microbes


treatments, especially after 4 and 28 days (Tables 8–12).Overall, these results point to the fact that the amend-ments did not have the expected effect of inducing furtherbiodegradation during the 56 day experiment. Although

there were isotopic differences observed, no clear trendwas followed. Individual compounds also did not showa clear trend in isotopic enrichment by one treatment

over another. Perhaps the indigenous microbes were notnutrient limited; hence the added nutrients did notinduce enhanced biodegradation. In particular, if there

was already enough nitrogen in the sediment to sustainthe microbial populations, the added inorganic fertilizerwould have minor effects on microbial populations.

Why the microbe-amended sediment did not show con-sistently enhanced biodegradation is more difficult todecipher. One plausible explanation is that the indigen-ous microbes were already very effective at degrading

hydrocarbons. The effects of the addition of othermicrobes capable of petroleum degradation may havebeen somewhat masked. Another possible explanation is

that some chemical or physical condition in the sedi-ment inhibited the growth and activity of the addedpetroleum-degrading microbes.

The results of this experiment tend to support anexpectation that biodegradation would isotopicallyfractionate the aromatic hydrocarbons, leaving behind

residual enriched materials. One other effect that mightisotopically fractionate the PAHs would be evaporation.Owing to differences in bond energy between 12C–12Cand 13C–12C bonds, it is expected that the isotopically

light molecules would enter the vapor phase before theheavy molecules, leaving the residual material enri-ched—an effect similar in trend to biodegradation. Work

with individual volatile chlorinated hydrocarbons hasshown the carbon isotope fractionation to be less than1%. As noted earlier, the isotopic changes in this study

were generally toward enrichment and were greater than2%. On a bulk isotopic basis, past work with crude oilweathering has also shown the effect of evaporation to beminimal, although some differences seen at 4 days may

have been influenced by volatilization.Some of the questions raised by this study could be

answered by complementary investigations involving

some of the same methodology, and also someenhancements. A possible reason why some of the iso-tope values were more enriched after 28 days than after

56 days was given earlier. It would be instructive topursue a study similar to this again, but over a long timeperiod. The experiment could run months instead of

days, with a corresponding increase of the initial dosageof oil to the sediment of a factor of five or six. If similaror larger isotopic enrichments over shorter periods oftime were observed, it would reinforce the idea of het-

erogeneity influencing the isotopic results. Parallel sam-ples mixed together and compared to single isolateswould also help to answer the question of heterogeneity.

Finally we note that this study aimed at what weconsidered to be the best-resolved compounds in thesecomplex mixtures. Some of the variation may in factrepresent contributions of minor components, which

could alter the isotopic composition of even the best-resolved compounds. With increasing biodegradation,and an obvious increase in the unresolved complex

mixture (UCM) in the chromatograms, the contribu-tions of these minor components to the isotopic com-position of even the best resolved compounds needs to

be considered.

5. Conclusions

The results of this study indicate that there was acarbon isotopic enrichment at the compound level

associated with biodegradation. Although this fractio-nation was not uniform from compound to compound,a 2–8% change over approximately 2 months was

observed. Contrary to expectations, sediment amendedwith nutrients and microbes known to attack PAHsshowed very little enhancement of biodegradation com-

pared to the control treatment. Possible reasons for thisare that the sediments were not nutrient limited, andthat indigenous microbes capable of degrading aromatic

hydrocarbons were already present. These results havevery important ramifications for the source apportion-ment of PAHs using compound specific isotope ana-lyses. Without taking into account the possible effects of

biodegradation, source apportionment using CSIA mayyield erroneous results. Although the trend of isotopicenrichment found in this study was predictable, the

varying degrees of the enrichment were not. Furtherwork needs to be pursued in an effort to better quantifythe effects of microbially mediated isotopic fractiona-

tion, especially as a function of time. In addition,knowledge of the biodegrading capability of indigenousmicrobes on a site to site basis should be characterized.If the PAHs or other hydrocarbons were associated with

considerable amounts of organic matter or soot, theymay be more resistant to biodegradation and would leadto less ambiguity associated with compound specific

isotope analysis.

Acknowledgements

We are grateful to the US-EPA STAR Program for

funding a portion of this research. The US Environ-mental Protection Agency through its Office of Researchand Development Partially funded the research descri-bed here under assistance agreement R-82-5420-01-1 to

SAM and MCK. It has not been subjected to Agencyreview and therefore does not reflect the views of theAgency, and no official endorsement should be inferred.


PJY and TCO both received fellowship support fromthe University of Virginia graduate school during thecourse of this study for which we are grateful. We arealso appreciative of the insightful comments of J. Abra-

jano and two anonymous reviewers of this paper.

References

Ahad, J., Lollar, B.S., Edwards, E.A., Slater, G.F., Sleep, B.E.,

2000. Carbon isotope fractionation during anaerobic biode-

gradation of toluene: implications for intrinsic bioremedia-

tion. Environmental Science and Technology 34, 892–896.

Ballentine, D.C., Macko, S.A., Turekian, V.C., Gilhooly, W.P.,

Martincigh, B., 1996. Compound specific isotope analysis of

fatty acids and polycyclic aromatic hydrocarbons in aerosols:

implications for biomass burning. Organic Geochemistry 25,

97–104.

Coates, J.D., Anderson, R.T, Lovley, D.R., 1996. Oxidation of

polycyclic aromatic hydrocarbons under sulfate reducing

conditions. Applied and Environmental Microbiology 62,

1099–1101.

Coates, J.D., Woodward, J., Allen, J., Philp, P., Lovley, D.R.,

1997. Anaerobic degradation of polycyclic aromatic hydro-

carbons and alkanes in petroleum-contaminated marine har-

bor sediments. Applied and Environmental Microbiology 63,

3589–3593.

Dipple, A., 1985. Polycyclic aromatic hydrocarbon carcino-

genesis: an introduction. In: Harvey, R.G. (Ed.), Polycyclic

Hydrocarbon Carcinogenesis. American Chemical Society,

Washington DC, pp. 1–17.

Erickson, D.C., Loehr, R.C., Neuhauser, E.F., 1993. PAH loss

during bioremediation of manufactured gas plant site soils.

Water Research 27, 911–919.

Escartı́n, E., Porte, C., 1999. Hyroxylated PAHs in bile of

deep-sea fish. Relationship with xenobiotic metabolizing

enzymes. Environmental Science and Technology 33, 2710–

2714.

Harvey, R.G., 1997. Polycyclic aromatic hydrocarbons. Wiley-

VCH, New York.

Hecht, S.S., Amin, S., Melikian, A.A., LaVoie, E.J., Hoffmann,

D., 1985. Effects of methyl and fluorine substitution on the

metabolic activation and tumorigenicity of polycyclic aro-

matic hydrocarbons. In: Harvey, R.G. (Ed.), Polycyclic

Hydrocarbon Carcinogenesis. American Chemical Society,

Washington DC, pp. 85–105.

Hellou, J., Mackay, D., Fowler, B., 1995. Bioconcentration of

polycyclic aromatic compounds from sediments to muscle of

finfish. Environmental Science and Technology 29, 2555–

2560.

Hites, R.A., Laflamme, R.E., Windsor Jr., J.G., Farringon,

J.W., Deuser, W.G., 1980. Polycyclic aromatic hydrocarbons

in an anoxic sediment core from the Pettaquamscutt River

(Rhode Island, USA). Geochimica et Cosmochimica Acta

44, 873–878.

Hostettler, F.F., Rosenbauer, R.J., Kvenvolden, K.A., 1999.

PAH refractory index as a source discriminant of hydro-

carbon input from crude oil and coal in Prince William

Sound, Alaska. Organic Geochemistry 30, 873–879.

Khalili, N.R., Scheff, P.A., Holsen, T.M., 1995. PAH source

fingerprints for coke ovens, diesel and gasoline engines,

highway tunnels, and wood combustion emissions. Atmo-

spheric Environment 29, 533–542.

Killops, S.D., Readman, J.W., 1985. HPLC fractionation and

GC-MS determination of aromatic hydrocarbons from oils

and sediments. Organic Geochemistry 8, 247–257.

Laflamme, R.E., Hites, R.A., 1978. The global distribution of

polycyclic aromatic hydrocarbons in recent sediments. Geo-

chimica et Cosmochimica Acta 42, 289–303.

Lollar, B.S., Slater, G.F., Ahad, J., Sleep, B., Spivack, J.,

Brennan, M., MacKenzie, P., 1999. Contrasting carbon iso-

tope fractionation during biodegradation of tri-

chloroethylene and toluene: implications for intrinsic

bioremediation. Organic Geochemistry 30, 813–820.

Mahaffey, W.R., Compeau, G., Nelson, M., Kinsella, J., 1991.

Developing strategies for PAH and TCE bioremediation.

Water Environment and Technology 3, 83–88.

Maldonado, C., Bayona, J.M., Bodineau, L., 1999. Sources, dis-

tribution, and water column processes of aliphatic and poly-

cyclic aromatic hydrocarbons in the northwestern Black Sea

water. Environmental Science and Technology 33, 2693–2702.

McRae, C., Sun, C., Snape, C.E., Fallick, A.E., Taylor, D.,

1999. d13C values of coal-derived PAHs from different pro-

cesses and their application to source apportionment.

Organic Geochemistry 30, 881–889.

McRae, J.D., Hall, K.J., 1998. Biodegradation of polycyclic aro-

matic hydrocarbons (PAH) in marine sediments under deni-

trifying conditions. Water Science and Technology 38, 177–185.

Meckenstock, R.U., Annweiler, E., Michaelis, W., Richnow,

H.H., Schink, B., 2000. Anaerobic naphthalene degradation

by a sulfate-reducing enrichment culture. Applied and

Environmental Microbiology 66, 2743–2747.

O’Malley, V.P., Abrajano Jr., T.A., Hellou, J., 1994. Determi-

nation of the 13C/12C ratios of individual PAH from envir-

onmental samples: can PAH sources be apportioned.

Organic Geochemistry 21, 809–822.

O’Malley, V.P., Burke, R.A., Schlotzhauer, W.S., 1997. Using

GC-MS/combustion/IRMS to determine 13C/12C ratios of

individual hydrocarbons produced from the combustion of

biomass materials-applications to biomass burning. Organic

Geochemistry 27, 567–581.

Roper, J.M., Cherry, D.S., Simmers, J.W., Tatem, H.E., 1997.

Bioaccumulation of PAHs in the zebra mussel at Times

Beach, Buffalo, New York. Environmental Monitoring and

Assessment 46, 267–277.

Wang, Z., Fingas, M., Shu, Y.Y., Sigouin, L., Landriault, M.,

Lambert, P., 1999. Quantitative characterization of PAHs in

burn residue and soot samples and differentiation of pyro-

genic PAHs from petrogenic PAHs-the 1994 Mobile burn

study. Environmental Science and Technology 33, 3100–3109.

Weissenfels, W.D., Beyer, M., Klein, J., 1990. Degradation of

phenanthrene, fluorine and fluoranthene by pure bacterial cul-

tures. Applied Microbiology and Biotechnology 32, 479–484.

Wilson, S.C., Jones, K.C., 1993. Bioremediation of soil con-

taminated with polynuclear aromatic hydrocarbons (PAHs):

a review. Environmental Pollution 81, 229–249.

Yang, S.K., Mushtaq, M., Chiu, P., 1985. Stereoselective

metabolism and activations of polycyclic aromatic hydro-

carbons. In: Harvey, R.G. (Ed.), Polycyclic Hydrocarbon

Carcinogenesis. American Chemical Society, Washington

DC, pp. 19–34.


Date post:	15-Nov-2023
Category:	Documents
Upload:	virginia
View:	0 times
Download:	0 times

The isotopic compositions of selected crude oil PAHs during biodegradation

Documents