Sampling of soil and groundwater for microcosms were collected and preserved as described by Lesser (2013) [26]. Microcosm experiments were done under aerobic and anaerobic conditions with sterile controls. The anaerobic microcosm experiments were performed without amendment, and with sulphate and nitrate amendments. The purpose of adding sulphate and nitrate was to promote anaerobic degradation of naphthalene by adding possible electron acceptors which may have become depleted in site groundwater. All the experiments were done in triplicate. These microcosms received 20 g of aquifer material and approximately 55 ml of groundwater in 60 ml hypovials with no remaining headspace. Initial site groundwater did not contain enough naphthalene to perform the experiments, so the addition of naphthalene was necessary (approximately 12 mg/l). The appropriate additions of sulphate and nitrate were made via syringe adding 1 ml from stock solutions (73.96 mg/l of sodium sulphate and 180.43 mg/l of potassium nitrate respectively). Anaerobic microcosms were assembled in an anaerobic chamber. A total of 144 microcosms were performed as shown in Table 1, which includes the characteristics and time of sampling for each microcosm.

The microcosms were periodically sacrificed and the aqueous phase was analyzed for the concentration of organic constituents, dissolved oxygen, and nitrate or sulphate where appropriate. The samples from the microcosm experiments to be analyzed for stable carbon isotopes were collected in 4.8 ml glass vials with a Teflon coated septum and leaving no headspace. They were preserved with sodium azide at a rate of 1 μL per 1 ml of sample, and stored at 4˚C until isotopic analyses were performed.

The concentrations of organic and inorganic compounds in water samples were analyzed by the Organic Geochemistry Laboratory at the University of Waterloo. Organic compounds were analyzed by gas chromatography and inorganic compounds by ion chromatography (anions) and by a plasma spectrometer (cations).

For the stable carbon isotope ratio analysis, an isotope ratio mass spectrometers (IRMS) was employed. These apparatuses are designed to measure mass differences in compounds such as CO₂. For isotope analyses of organic compounds contained in aqueous samples, the organic compounds must first be extracted from the sample using headspace and/or SPME techniques. Each compound must then be isolated using a gas chromatograph (GC). Then the carbon contained in the compound is converted into CO₂ in a combustion interface. Water from the combustion is trapped and the CO₂ is swept into an IRMS.

For this study a solid-phase microextraction technique (SPME) was used for extraction of naphthalene from the

d = days; NA = Not analyzed.

aqueous samples. Chai et al. [27] and Potter et al. [28] outline the SPME technique to extract PAHs from aqueous samples. Hunkeler and Aravena [29] compare the SPME extraction technique with other methods and point out its accuracy for δ¹³C determination.

The SPME fiber used in this study is coated with Polydimethylsiloxane (Supelco, Bellefonte, PA, USA). Different film coating thicknesses were employed: 7 μm, 30 μm and 100 μm. The amount of analyte that can be sorbed is related to coating thickness. The thicker the coating, the greater the sensitivity and the lower concentrations can be analyzed. The IRMS system has a limited range of concentrations that can be analyzed, therefore when samples are highly concentrated a thinner coating film must be used or the samples must be diluted.

The fiber is contained in a fiber-holder which resembles a syringe. The needle is used to penetrate the septa of the sample vials and then the fiber is lowered into the water. Prior to lowering the fiber, a magnetic stir bar had been deposited inside the vial, and 0.5 ml of water had been removed in order to leave a headspace and avoid the needle touching the sample. All magnetic stir bars were washed with pentane, dried at 100˚C for 30 min and deposited into the vials using a sterilized Teflon coated magnet. Then, the sample is stirred for 20 minutes, until naphthalene reaches equilibrium between the water sample and the silica fiber.

Then the fiber is withdrawn into the needle and the needle is removed from the vial. The needle is then used to penetrate the GC septum, and the fiber is lowered into the GC inlet. The Gas Chromatograph (GC) model HP 6890 has an injector (split/splitless) equipped with a SPME liner from Supelco (Bellefonte, PA, USA). The column used (Rtx5) was 60 m long × 0.25 mm I.D. and 1.0 μm coating film thickness (5% diphenyl-95% dimethyl polysiloxane) from Restek Corporation (Bellefonte, PA, USA).

The analytes are thermally desorbed from the fiber and swept onto the GC column. The inlet temperature was 270˚C with a purge time of 1 minute (i.e. it runs in splitless mode for 1 minute before turning into split mode). The desorption time was 20 minutes (i.e. the total time the fibre is in the injector). The column is temperature programmed as follows: 80˚C for 2 min, then the temperature is increased 20˚C per min, up to 200˚C. Then the temperature remains at 200˚C for 12 min. Each sample was run twice to measure the reproducibility of the results. The GC separation of naphthalene from the rest of the organics found in the sample was confirmed by using neat compounds.

The GC column is connected to a quartz tube filled with copper oxide [combustion interface (Micromass UK Limited, Manchester, UK)] with an operating temperature of 850˚C. The combustion interface converts the organic compounds to CO₂ and H₂O. A liquid nitrogen cold trap set at –100˚C is used to retain the water while the CO₂ gas is swept with helium by capillary tubing into the isotope ratio mass spectrometer (IRMS) model Isochrom (Micromass UK Limited, Manchester, UK).

To measure the reproducibility and accuracy of δ¹³C measurements, a naphthalene standard was prepared. Seven subsamples from the standard were run. Each subsample was run twice and the average was calculated as it was done for the real samples. The standard deviation obtained was 0.083‰, then the 95% confidence interval was calculated to be twice the standard deviation (0.166‰). This is the reported precision for each sample [30].

Because the microcosm experiments were performed by analyzing bottles at different times each bottle is, on its own, a single experiment. Each experiment has then a different initial concentration (C_o) depending on the sorptive characteristics of the soil particles on each bottle. It is not practical to attempt to measure this variability due to the nature of the experiment, but this is assumed to have minimum effects on the final results of the experiment (biodegradation of naphthalene with time). Hence, the first set of triplicates was assumed to represent the initial concentration of all samples (C_o) in each of the conditions of the microcosm experiments for each condition shown in Table 1. Control microcosms showed no decrease in the concentration of naphthalene.

The aerobic microcosm experiments show that naphthalene was consumed aerobically in less than 60 hours (Figure 1). These results are in agreement with previous studies that have shown aerobic biodegradation of naphthalene [6,7]. These data show that after 48 hours there were significant differences in the amount of naphthalene consumed (from 50% to 95%). The best estimate of naphthalene aerobic biodegradation was taken as the average of biodegradation in the triplicate microcosms.

Apparent biodegradation rates were calculated using zero-and first-order kinetics as follows:

where:

t = time C = concentration at time t C_o = initial concentration

λ_o = zero order kinetics apparent biodegradation rate [t^‒1]

λ₁ = first order kinetics biodegradation rate [t^‒1].

The apparent biodegradation rates were calculated by fitting the naphthalene data. The first-order decay rate (λ₁) was 2.4 days⁻¹, and the zero-order apparent decay rate (λ_o) was 0.72 days⁻¹.

Taking into account the results of the biodegradation of naphthalene, 13 microcosms samples were analyzed for δ¹³C. The results show a slight δ¹³C enrichment dur-

ing aerobic biodegradation of naphthalene. The δ¹³C increased from approximately −25.5‰ to −25.1‰ (enrichment of 0.4‰ ± 0.08‰), only in one sample where 95% of the naphthalene had been consumed. The data clearly shows that this enrichment is only apparent after most of naphthalene has been degraded (Figure 2).

The statistical significance of the δ¹³C enrichments observed during biodegradation of naphthalene was calculated using the statistic test T [31]. This test determines if there is any difference between two random variables from approximate normal distributions. The statistic test T is calculated as follows [31]:

where:

T = T statistic value X_a = Mean value a X_b = Mean value b n_a = number of repetitions for value a n_b = number of repetitions for value b and is a pooled variance calculated as:

where:

= variance of population a

= variance of population b with (n_a + n_b ‒ 2) degrees of freedom.

The δ¹³C for each microcosm were compared using this statistical test. For the sterile controls, except in one case, the δ¹³C values were statistically equal. The sample from the aerobic microcosm where most of naphthalene had been biodegraded was statistically different from the other 3 samples with a lower degree of biodegradation. Therefore, the isotopic enrichment observed during the aerobic biodegradation, though small, was statistically significant.

Other researchers have also observed small isotopic enrichments on naphthalene in the final stages of their aerobic biodegradation [32]. Comparing with the results obtained for other compounds biodegraded aerobically, benzene was found to have an isotopic enrichment of 1.1‰ [21], and aerobic biodegradation of toluene did not exhibit any significant fractionation [20].

The enrichment factor for aerobic biodegradation of naphthalene was calculated. The concentration for each bottle was normalized with respect to the initial concentration (C_o) and this normalized value is in fact the fraction of remaining reactant (f). These normalized concentrations were plotted against the δ¹³C and a least squares regression was obtained, which is equivalent to the Rayleigh distillation curve. Using Equation (3), an enrichment factor of (ε) of −0.16 was calculated, with a correlation factor (R²) of 0.8. This correlation factor, although reasonable, is poor for a controlled microcosm experiment, and the enrichment factor (ε) is very low. Therefore, this isotopic enrichment may not be sufficient to demonstrate biodegradation of naphthalene in field situations, where many other processes such as advection, dispersion and adsorption, also help to reduce the concentrations of contaminants.

The anaerobic microcosms were also performed in triplicate. The results show that after 288 days the microcosms with no amendments, amended with nitrate and amended with sulfate, had consumed respectively 30%, 50% and 60% of naphthalene on average, compared to control microcosms (Figure 3). Essentially all the naphthalene was consumed in 4 in microcosms. The concentration of naphthalene in the sterile controls changed only slightly during the experiments; therefore much of the change in naphthalene concentration in the active micro-

cosms is attributed to biodegradation.

Apparent zero-order decay rate constants (λ_o) were calculated by subtracting the apparent mass loss in sterile controls, obtaining 0.26, 0.77 and 0.58 yr^‒1 for the unamended, sulphate-amended and nitrate-amended respectively. The first-order decay rate constants (λ₁) obtained were 0.51, 1.50 and 2.23 yr^‒1 for the unamended, sulphate-amended and nitrate-amended respectively. These decay rates are considerably lower than those obtained for aerobic degradation. Blum et al. [11] obtained a λ_o of 1.39 yr^‒1, similar to what was found in this study; however other studies report λ_o of 5 to 31 yr⁻¹ [8].

In the nitrateand sulfate amended microcosms, concentrations of nitrate and sulphate were also analyzed to monitor their activity as electron acceptors. These data show that the decrease in naphthalene concentration in the sulphate-amended microcosms correlates with a decrease in sulphate concentrations (data not shown). This suggests that biodegradation of naphthalene occurred under a sulphate-reducing conditions. Similarly a decrease in naphthalene concentration on the nitrateamended microcosms also correlated with a decrease on nitrate concentrations (data not shown), suggesting that biodegradation occurred under denitrifying conditions.

Based on these concentration results, 20 microcosms were selected to be analyzed for the δ¹³C. The results show that under denitrifying conditions the δ¹³C of naphthalene increased from −25.2‰ to −23.9‰ (enrichment of 1.3‰ ± 0.3‰) after naphthalene biodegraded in 95% (Figure 4). The enrichment factor (ε) according to the Rayleigh distillation was −0.3, with a correlation factor (R²) of 0.7. Similarly to aerobic microcosms, the correlation factor is acceptable and the enrichment. is low. This means that under field conditions, this enrichment under denitrifying conditions would not be appreciated easily. The δ¹³C of each microcosm were compared with statistic test T [31]. For sterile controls δ¹³C values were showed to be statistically equal. For the

active nitrate-amended microcosm, most δ¹³C values were statistically different.

For the unamended microcosm a slight enrichment was detected in naphthalene δ¹³C from −25.2‰ to −24.5‰ (0.7‰ ± 0.3‰ enrichment) after a biodegradation of approximately 65% (after 288 days, Figure 5). The enrichment factor (ε) according to the Rayleigh distillation, was −0.62 with a correlation factor (R²) of 0.8. As in the previous cases, the correlation factor is acceptable but the enrichment factor is low, which would prove difficult to be verified in field situations. For these unamended microcosms, just one δ¹³C results proved to be statistically different from the rest using the statistic test T.

For reductive sulphate-amended microcosms, no significant changes were seen in the δ¹³C results.

Richnow et al. [24] found an isotopic fractionation of 1.2‰ naphthalene in an anaerobic aquifer. Kelley et al. [33] analyzed the δ¹³C of several PAHs (including naphthalene) in an aquifer with anaerobic conditions. In their field data a very small enrichment in the δ¹³C of naphthalene can be appreciated, however they do not seem to pay much attention to this result, perhaps because of the uncertainty associated with such low enrichments. The results of this study are in agreement with the results of Kelley et al. [33].