Extraction, Separation, and Intramolecular Carbon Isotope

Nov 5, 2012 - Geological Survey of Canada, Natural Resources Canada, Québec City, QC, ... elemental analysis isotope ratio mass spectrometry (TC/EA-...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/ac

Extraction, Separation, and Intramolecular Carbon Isotope Characterization of Athabasca Oil Sands Acids in Environmental Samples Jason M. E. Ahad,*,† Hooshang Pakdel,‡ Martine M. Savard,† Marie-Christine Simard,† and Anna Smirnoff† †

Geological Survey of Canada, Natural Resources Canada, Québec City, QC, Canada, INRS Eau Terre Environnement, Québec City, QC, Canada



ABSTRACT: Here we report a novel approach to extract, isolate, and characterize high molecular weight organic acids found in the Athabasca oil sands region using preparative capillary gas chromatography (PCGC) followed by thermal conversion/elemental analysis−isotope ratio mass spectrometry (TC/EA-IRMS). A number of different “naphthenic acids” surrogate standards were analyzed as were samples from the bitumen-rich unprocessed McMurray Formation, oil sands process water, groundwater from monitoring wells, and surface water from the Athabasca River. The intramolecular carbon isotope signature generated by online pyrolysis (δ13Cpyr) showed little variation (±0.6‰) within any given sample across a large range of mass fractions separated by PCGC. Oil sand, tailings ponds, and deep McMurray Formation groundwater were significantly heavier (up to ∼9‰) compared to surface water and shallow groundwater samples, demonstrating the potential use of this technique in source apportionment studies.



considerable overlap in the range of δ13C signatures found for fossil fuels and other types of organic matter can significantly inhibit its applicability in source apportionment studies.12 As a result, initial attempts to use δ13C signatures to trace oil sands mining-related contamination have thus far been limited to indirect approaches with minimal success, such as the isotopic characterization of different carbon pools from various types of oil sands reclamation sites.13,14 In contrast to total or “bulk” δ13C values that are measured on the CO2 generated by oxidation, however, intramolecular carbon isotope analysis of polar organic compounds that measures the δ13C signature of the CO2 generated by pyrolysis of carboxyl groups (−COOH) may offer a more direct approach to source discrimination of NAs. As has been observed in low molecular weight (MW) organic acids in the C2−C6 range, the carboxyl group can undergo exchange with typically more 13C-enriched dissolved inorganic carbon (DIC) under hydrous pyrolysis conditions which can significantly influence the carbon isotope composition of the entire molecule.15,16 The rate of this exchange reaction is thought to increase as pH becomes more alkaline.16 Consequently, as a result of the alkaline hot water process used by commercial oil sands operations to extract bitumen or due to isotopic variability associated with the original geological depositional

INTRODUCTION The organic acids found naturally in bitumen that become concentrated in oil sands process water pose a serious threat to surface water and shallow groundwater.1−3 These compounds are collectively described as naphthenic acids (NAs), a complex mixture of alkyl-substituted acyclic and cycloaliphatic carboxylic acids that follow the general chemical formula CnH2n+ZO2, where n indicates the carbon number and Z is zero or a negative, even integer that specifies the hydrogen deficiency resulting from ring formation.4 New developments in high- and ultrahigh-resolution mass spectrometry (MS) and multidimensional comprehensive gas chromatography mass spectrometry (GC × GC-MS), however, have shown that this “classical” definition of NAs cannot be used to describe the majority of components comprising the polar organics in fresh surface waters and oil sands process waters.5−7 The difficulty in separating “NAs” from interfering coextractives such as humic and fulvic acids thus contributes to significant discrepancies in concentrations between different analytical methods used for quantification.5,8−11 And while high and ultrahigh resolution MS and multidimensional comprehensive GC-MS provide valuable insight into the molecular species and individual compounds comprising NAs, the complex mass spectra generated by these techniques do not generally lend themselves to an easily interpretable format suitable for quantification and hence source apportionment. Natural abundance stable carbon isotope ratios (δ13C) have the potential to discriminate sources of NAs, though the © 2012 American Chemical Society

Received: September 14, 2012 Accepted: November 5, 2012 Published: November 5, 2012 10419

dx.doi.org/10.1021/ac302680y | Anal. Chem. 2012, 84, 10419−10425

Analytical Chemistry

Article

environment, it may be possible to differentiate between oil sands process water and natural background organic acids. Carboxyl group δ13C signatures of semivolatile compounds such as benzoic acid and 2- and 4-ring aromatic acid standards have been accurately determined by both conventional off-line in vacuo preparation17,18 and online thermal conversion/ elemental analysis isotope ratio mass spectrometry (TC/EAIRMS) techniques.17 Online TC/EA-IRMS analysis has also been used to examine the intramolecular carbon isotope signature of insoluble organic matter from meteorites.19 To the best of our knowledge, however, this approach has yet to be tried on complex mixtures of high MW organic acids extracted from water samples. Here we report a novel protocol for the extraction, separation, and intramolecular isotopic characterization of acid extractable organics (AEO) containing “naphthenic acids” for use in source apportionment studies. The techniques used here were tested at the local scale, in the context of an Athabasca oil sands operation, where the host of the bitumen reserves is the Cretaceous McMurray Formation (sand and sandstone), which is both mined to extract oil and naturally eroded by the Athabasca River. We specifically applied the method to samples collected from several oil sands tailings ponds, groundwater monitoring wells, unprocessed McMurray Formation oil sand, and surface water from the Athabasca River. The successful development of this technique and large range (up to ∼9‰) in intramolecular carbon isotope signatures points to a highly promising approach to discriminating different sources of polar organic compounds in the Athabasca oil sands region.



Figure 1. Simplified schematic outlining the protocols used for extraction, purification, and analysis of acid extractable organics (AEO).

EXPERIMENTAL SECTION Standards. Since classically defined naphthenic acids are a complex mixture of cyclic and acyclic carboxylic acids, acquiring adequate standard material is not straightforward. In lieu of precise standards, the following compounds (Sigma-Aldrich, Oakville, ON) were analyzed to evaluate the range of intramolecular carbon isotope values generated by pyrolytic decarboxylation: benzoic acid, BA (99.99%), 4-hydroxybenzoic acid, HBA (≥99%), cyclohexanecarboxylic acid, CHCA (≥98%), 1-methylcyclohexanecarboxylic acid, MCHCA (99%), trans-1,2-cyclohexanedicarboxylic acid, CHDCA (95%), 1-adamantanecarboxylic acid, ACA (99%), 1,3-adamantanedicarboxylic acid, ADCA (98%), and 3-oxo-1-indanecarboxylic acid, ICA (98%). To assess potential isotopic fractionation associated with each step in the methodology outlined below, several standards from the above list were dissolved in 4 L of 18.2 MΩ water and extracted and processed in the same manner as described for samples. The intramolecular carbon isotope signatures of these “procedural standards” were compared to those analyzed from aliquots taken directly from the container. Extraction and Purification of AEO from Water Samples. A simplified schematic outlining the protocols used for extraction, purification, and analysis of AEO is shown in Figure 1. Between approximately 1 and 25 L of water was filtered under vacuum using precombusted (450 °C for 4 h) glass fiber filters (∼1 μm pore size diameter, VWR). All samples were acidified to pH 4.5 using 10 N HCl prior to AEO extraction. Strata-X-A (Phenomenex, Torrance, CA, USA) solid phase extraction (SPE) sorbent was used to recover sufficient amounts of AEO required for isotopic analysis. Due to the difficulty in treating very large volume water samples, AEO was

extracted using loose SPE sorbent preconditioned with 5 mL of methanol followed by 5 mL 18.2 MΩ water. Approximately 100 mg loose SPE sorbent per liter of filtered water were vortexed inside a 4 L amber glass bottle for a period of 4 h using a magnetic stir bar. The sorbent was then filtered from the aqueous phase under vacuum using precombusted glass fiber filters as described above and subsequently eluted with 10 mL methanol followed by another 10 mL of 10% formic acid in methanol. Each water subsample was extracted twice to maximize recovery. The extracts containing total AEO were combined and then evaporated to dryness under ultrahigh purity N2. A recovery test carried out using an Environment Canada “oil sands naphthenic acid fraction” standard obtained from the acid extraction of oil sands process water yielded an AEO recovery of >90%. Extraction blanks carried out using 4 L of 18.2 MΩ water and put through the same procedure as described above yielded no detectable background AEO. Extraction and Purification of AEO from Oil Sand. A 250 g sample of bitumen-rich (∼18 wt %) sand of the McMurray Formation was extracted in a Soxhlet apparatus using dichloromethane (DCM). Bitumen was then dissolved in 300 mL of cold hexane while stirring. Asphaltenes, the hexane insoluble fraction, were filtered off to recover the maltene fraction. The extraction was repeated twice to remove most of the asphaltenes. The maltene fraction was then dissolved in 100 mL hexane and extracted with 4 × 100 mL 1 N NaOH in water. The aqueous phase was further extracted with 2 × 100 mL hexane. The hexane fraction was discarded. The aqueous fraction was acidified to pH 2 using a 10 N HCl solution followed by extraction with 5 × 100 mL of 10% (vol) methanol 10420

dx.doi.org/10.1021/ac302680y | Anal. Chem. 2012, 84, 10419−10425

Analytical Chemistry

Article

of methylated AEO sample in a 4 mL amber glass vial with a PTFE-lined cap and heated in an oven at 60 °C for one week following a modified protocol described by Frank et al.23 The solution was transferred to a separation flask, and 1 mL of 3 N HCl and 5 mL distilled water were added then extracted with 4 × 30 mL 10% methanol in DCM. All organic phases were mixed, evaporated to dryness, and weighed to obtain the mass of recovered AEO. Intramolecular δ13C Analysis. Analysis by TC/EA-IRMS requires that samples are transferred into metal capsules and placed onto an automated carousel. Due to the difficulty in accurately weighing out aliquots of polar organics into such small containers, demethylated standards and samples were dissolved in methanol and transferred by syringe as a 10 μL injection into a 12 μL rigid silver capsule (IVA-Analysentechnik e.K., Düsseldorf, Germany) that was placed inside a larger, more malleable silver capsule (Elemental Microanalysis Ltd., Okehampton, UK). The capsules containing samples and standards were dried in an oven at 60 °C for 20 min and then sealed with pliers. On the basis of the CO2 signal sizes obtained for mass 44 from analysis of standards of known weight, we estimated an optimal AEO sample size of ∼300−800 μg for intramolecular carbon isotope characterization. Each 10 μL aliquot of sample transferred to the silver capsules thus contained approximately 300−800 μg of AEO. The carbon isotope signature of the CO2 generated by pyrolytic decarboxylation of AEO (δ13Cpyr) was determined by TC/EA-IRMS using a Thermo-Finnigan Delta+ XL (Thermo Fisher Scientific, Bremen, Germany) IRMS system at the DeltaLab of the Geological Survey of Canada, Québec. The system was equipped with a stainless steel GC column (2 m × 1/4 in. × 5.3 mm) packed with HayeSep Q (80/100 mesh). A cold trap with a temperate between −78 and −85 °C was placed between the pyrolysis furnace and GC oven to help purify CO2 and to remove potentially interfering compounds prior to IRMS analysis. The pyrolysis reactor (Al2O3 tube) was filled with ceramic granules approximately 3 mm in diameter on top of which ∼30 cm of crumpled platinum wire was placed. The GC column was kept at 70 °C, and the flow of the helium carrier gas (>99.9999% purity) was 100 mL/min. Tests carried out using a range of temperatures for the pyrolysis of carboxylic acids standards confirmed an optimal temperature of 750 °C as previously reported.17 The δ13Cpyr values determined by TC/EA-IRMS were analyzed using CO2 calibrated against international carbonate standards (NBS 18 and NBS 19). The reference gas was analyzed by conventional off-line techniques with a PRISM-III dual inlet IRMS system following protocols adapted from Santrock and Hayes18 and Oba and Naraoka.17 In order to evaluate the accuracy of the online continuous flow technique for naphthenic acids, the following standards were also prepared off-line: BA, HBA, CHCA, CHDCA, and ACA. Precombusted (450 °C for 8 h) quartz tubes containing 3.0− 3.5 mg standard material were evacuated under vacuum for several minutes and then heated at 550 °C for 1 h. Subsequent CO2 generated by pyrolytic decarboxylation was purified by standard procedures. The “bulk” or nonintramolecular carbon isotope signatures of AEO were analyzed by EA-IRMS using standard protocols (uncertainty of ±0.5‰).

in DCM. All organic phases were mixed and evaporated to dryness under ultrahigh purity N2 to recover the AEO. The amount of AEO contained within the oil sand sample was determined to be approximately 0.1% by weight of bitumen. Separation of AEO Using Preparative Capillary Gas Chromatography (PCGC). The difficulty in separating and purifying different classes of process water-derived organic acids prior to their analysis is one of the most pressing issues concerning their quantification.20 To address this, we have used preparative capillary gas chromatography (PCGC), a technique that is becoming increasingly popular in geochemical research for collecting the large amounts (i.e., ≥40 μg carbon) of sample required for compound-specific natural abundance radiocarbon analysis (e.g. ref 21 and 22). All AEO samples were methylated prior to PCGC to produce acid methyl esters. Approximately 2 mL BF3-methanol reagent was added per 10 mg of sample in an amber vial with a PTFE-lined cap and heated at 60 °C for 2 h. After the samples were cooled, 5 mL of 18.2 MΩ water was added and then extracted with 3 × 50 mL DCM. All DCM phases were combined and evaporated to dryness to recover the methylated acids. Concentrated methylated AEO samples redissolved in 1 mL DCM were separated and collected by PCGC to obtain a sufficient amount required for intramolecular δ13C characterization of six distinct AEO fractions. PCGC was carried out using an Agilent (Santa Clara, CA, USA) 7890A GC equipped with two 30 m × 0.5 mm i.d. DB-5MS columns (0.5 μm film thickness) and a flame ionization detector (FID) coupled to a Gerstel (Mülheim an der Ruhr, Germany) preparative fraction collection (PFC) system. The injection volume was 5 μL, and each sample was injected ∼100 times. The following GC oven temperature program was used: 40 °C (1 min), 30 °C/min to 130 °C, and 10 °C/min to 290 °C (18 min). The Gerstel cold injection system (CIS) was kept at 10 °C for 0.2 min at the start of the run. Approximately 5% of effluent was passed to the FID, and the remaining 95% was passed to the PFC system. The first AEO fraction collected on the PCGC covered a time span of 3.5 min, whereas the other five fractions each covered a time span of 1.5 min. Each of the six AEO fractions within a given sample is directly comparable to its corresponding fraction from another sample. The application of PCGC to separate AEO into six different fractions that are comparable between samples also works to eliminate potentially interfering lower or higher MW compounds that can bias the assessment of AEO. The transfer line between the GC and PFC was set at 300 °C, and the PFC interface was set at 310 °C. Samples were collected cryogenically in glass U-traps at −20 °C using chilled methanol to condense the methylated AEO fractions. U-traps were eluted with DCM and transferred to amber 2 mL GC vials. All samples were analyzed by an Agilent 5975C GC-MS equipped with a 30 m × 0.25 mm i.d DB-5 column (0.25 μm film thickness) to confirm comparable mass fractions and to check for potential impurities, which were insignificant in all cases. Demethylation. A parallel study to this will report the compound species classes comprising AEO in these samples as determined using Orbitrap high resolution mass spectrometry. Orbitrap MS characterization is optimized for the detection of polar acids as opposed to their analogous methyl esters; thus, all samples collected using PCGC were converted back to their original organic acid form. Demethylation also acts to avoid any potential carbon isotope interferences that the methyl groups added to AEO may cause during pyrolysis. A 1 mL portion of 1.67 N NaOH and 1 mL methanol were added to every 50 mg



RESULTS AND DISCUSSION Figure 2 shows a representative series of GC-MS total ion current chromatograms illustrating the six different mass 10421

dx.doi.org/10.1021/ac302680y | Anal. Chem. 2012, 84, 10419−10425

Analytical Chemistry

Article

isotopic fractionation caused by PCGC separation of individual compounds was previously reported and was mainly attributed to incomplete collection at the front or tail ends of eluting peaks.24,25 It should be noted however that those studies involved the separation of individual compounds, whereas AEO can contain thousands of different components.26 Thus partial peak collection would not have affected the vast majority of AEO peaks contained within the relatively large windows of separation employed here. Furthermore, the MW of CHDCA (172.2 g/mol) is such that it eluted in the PCGC separation window that corresponded to the very early part of fraction 1. We therefore do not consider isotopic fractionation observed in relatively low MW naphthenic acid surrogates to be representative of the much larger components comprising AEO in fractions 2−6. A more accurate assessment of potential isotopic fractionation associated with PCGC separation of AEO is provided by examining the results obtained for samples. As shown in Table 2, the δ13Cpyr ratios showed little variation across fractions 2−6 within a particular sample, within a precision of ±0.6‰. It is highly unlikely such precision would be obtained for each AEO sample if there were significant isotopic fractionation associated with this methodology. In some instances (e.g., samples C2−B and C3), however, the δ13Cpyr signatures of fraction 1 showed a slight 13C-enrichment of around 1‰ compared to the other 5 fractions (Table 2). This result suggests a possible isotope fractionation effect associated with PCGC separation in compounds comprising the very low MW range of AEO as observed for the CHDCA standard. It is also possible that the slight 13C-enrichment observed in fraction 1 was not an artifact of PCGC separation but is in fact a real observation. For instance, the 13C-enrichment in carboxyl group carbon in C2− C6 organic acids due to exchange with isotopically heavier DIC observed by Dias et al.16 was in many cases more pronounced with decreasing chain length. Until further investigation provides more insight into this phenomenon, the δ13Cpyr data obtained for fraction 1 of AEO is not included in the discussion or figures reported here. Perhaps the most compelling argument for the lack of significant isotopic fractionation associated with PCGC separation of AEO can be made through comparison of the average δ13Cpyr values obtained for fractions 2−6 of any particular sample to those obtained by analysis of the same sample not separated by PCGC (i.e., non-PCGC). As illustrated in Figure 3, the non-PCGC δ13Cpyr values were within the ±0.6‰ precision obtained for the fraction 2−6 averages for the five samples in which this comparison was made. Additionally, this comparison was carried out for samples covering almost the entire range of average δ13Cpyr values reported here (−29.2 to −20.0‰). This observation would be

Figure 2. Superimposed GC-MS total ion current chromatograms illustrating the six different mass fractions of methylated acid extractable organics separated by PCGC (sample C1).

fractions of methylated AEO separated by PCGC (a different GC column and oven temperature program were used for GCMS analysis; hence the retention times illustrated on Figure 2 do not correspond to those obtained by PCGC). The approximate MW ranges (m/z) for the principal components comprising each of the six fractions separated by PCGC as determined by Orbitrap high resolution MS (manuscript in preparation) were 180−240 (fraction 1), 210−260 (fraction 2), 230−280 (fraction 3), 250−300 (fraction 4), 260−320 (fraction 5), and 280−340 (fraction 6). The higher MWs for the majority of components in actual AEO samples compared to naphthenic acids surrogate compounds (Table 1) highlights the difficulty in obtaining suitable standard material for Athabasca oil sands polar organics. The δ13Cpyr ratios for the standard compounds analyzed here ranged between −47.6 and −22.6‰ (Table 1). No significant isotopic difference was observed between standards that were analyzed in solid form directly from the container and those that were prepared in the dissolved phase; thus results from both are shown together in the total number of analyses (n). The precision for replicate analyses of naphthenic acids surrogate standards varied between 0.1 and 0.9‰ (Table 1). The δ13Cpyr signatures of standard compounds (CHCA, CHDCA, MCHCA, ACA, and ADCA) extracted using the same protocol as that for large water samples and then subjected to demethylation prior to isotopic analysis were comparable within ±0.7‰ to the δ13Cpyr signatures shown in Table 1. A procedural blank (CHDCA) that was water extracted, demethylated, and subsequently put through the PCGC showed a 13C-enrichment of approximately 4‰. Significant

Table 1. Intramolecular Carbon Isotope Ratios (δ13Cpyr) of “Naphthenic Acids” Surrogate Compounds standard

acronym

MW (g/mol)

δ13Cpyr

σ

n

cyclohexanecarboxylic acid trans-1,2-cyclohexanedicarboxylic acid 1-methylcyclohexanecarboxylic acid 1-adamantanecarboxylic acid 1,3-adamantanedicarboxylic acid benzoic acid 3-oxo-1-indancarboxylic acid 4-hydroxybenzoic acid

CHCA CHDCA MCHCA ACA ADCA BA ICA HBA

128.2 172.2 142.2 180.2 224.3 122.1 176.2 138.1

−25.7 −25.9 −47.6 −22.9 −35.8 −31.4 −27.5 −22.6

0.6 0.9 0.8 0.7 0.3 0.4 0.2 0.1

15 63 41 7 6 32 3 3

10422

dx.doi.org/10.1021/ac302680y | Anal. Chem. 2012, 84, 10419−10425

Analytical Chemistry

Article

Table 2. Intramolecular Carbon Isotope Ratios (δ13Cpyr) of AEO in Samples for Fractions 1−6 Separated by PCGCa sample oil sand PW1 PW2 PW3d C1 C2−B C3 R1c R4c

Fr 1 −21.0 −21.4 −20.8 −23.0 −26.5

Fr 2

Fr 3

Fr 4

Fr 5

Fr 6

avgb

σb

−20.1 −22.0 −21.5 −21.4 −21.3 −24.0 −27.3

−20.6 −21.1

−20.1 −21.4 −20.6

−19.6 −20.7 −21.2 −21.5 −21.0 −24.3 −28.2

−19.4 −21.3 −20.3 −21.4 −20.9 −23.8 −27.8

−20.0 −21.3 −20.9 −21.6 −21.1 −24.2 −27.7 −29.2 −29.0

0.5 0.5 0.5 0.3 0.1 0.3 0.3

−22.1 −21.1 −24.7 −27.7

−21.2 −24.4 −27.6 −29.2 −29.0

a

The oil sand sample is unprocessed McMurray Formation bitumen; PW1, PW2, and PW3 are process water samples from three different tailings ponds; C1 is McMurray Formation groundwater; C2−B and C3 are from shallow groundwater monitoring wells; and R1 and R4 are from the Athabasca River. bThe average δ13Cpyr value for fractions 2−6. cCombined masses of Fractions 4 and 6. dCombined masses of Fractions 3 and 4.

Despite repeated efforts involving various permutations of lower and higher pyrolysis temperatures and shorter or longer incubation periods, pyrolytic decarboxylation of cycloaliphatic compounds resulted in a large number of interfering compounds that could not be properly separated from CO2 in vacuo prior to dual inlet IRMS analysis. In contrast, the chromatographic separation of CO2 from other pyrolysis products afforded by the HayeSep Q column employed here resulted in reliable and precise δ13Cpyr ratios for samples that were not significantly affected by background noise (Figure 4).

Figure 3. Average intramolecular carbon isotope signatures (δ13Cpyr) of AEO for fractions 2−6 separated by PCGC (triangles) plotted with δ13Cpyr signatures for aliquots of AEO that were not separated by PCGC (squares) and the “bulk” (i.e., nonintramolecular) δ13C values of AEO from fractions 3, 4, or 5 (circles). The oil sand sample is unprocessed McMurray Formation bitumen; PW1, PW2, and PW3 are process water samples from three different tailings ponds; C1 is McMurray Formation groundwater; C2−B and C3 are from shallow groundwater monitoring wells; and R1 and R4 are from the Athabasca River. The error bars represent the uncertainty for δ13Cpyr (0.6‰) and bulk δ13C analyses (0.5‰).

incredibly unlikely if PCGC separation led to significant isotopic fractionation in fractions 2−6 of AEO. The lack of a significant isotopic difference between the nonseparated sample and its corresponding five mass fractions suggests that in many cases PCGC separation may not yield substantial additional information in studies where δ13Cpyr analyses of the total AEO can be applied as a tool to delineate sources. However, the use of PCGC in samples containing very low amounts of AEO will be warranted as an effective means to eliminate impurities and other interfering compounds that can bias the δ13Cpyr measurement, particularly with regards to the very low and very high MW spectrum for organic matter. PCGC separation could also be valuable in future studies that aim to examine the environmental behavior of individual AEO components, its advantage over previously employed distillation techniques23 being its potential to obtain highly precise and accurate fractions for a narrower range of compounds. Out of the five standards chosen for off-line δ13Cpyr analyses, only the two aromatic carboxylic acids (BA, δ13Cpyr = −31.9 ± 0.2‰, n = 12; HBA, δ13Cpyr = −23.3 ± 0.2‰, n = 8) provided results that were similar to those obtained by TC/EA-IRMS.

Figure 4. Representative TC/EA-IRMS chromatograms (samples C1 and C3, fraction 6) illustrating the ratio of mass 45/44 and the relative intensities for masses 44, 45, and 46.

The development of a suitable off-line methodology capable of accurately determining the intramolecular carbon isotope ratios of cycloaliphatic naphthenic acids surrogate compounds is currently being investigated. In addition to the O2 compound class species indicative of carboxylic acids and hence classically defined NAs, ultrahigh resolution MS analysis of Athabasca oil sands region polar organics has revealed a wide range of other species classes such as O2N, O3N, O2, O2S O3, O3S, O4, O4S, O5, O5S, O6, O6S, O7, 10423

dx.doi.org/10.1021/ac302680y | Anal. Chem. 2012, 84, 10419−10425

Analytical Chemistry

Article

O7S, O8, and O8S.6,27 While the term “pyrolytic decarboxylation” can be unambiguously used to describe δ13Cpyr ratios obtained for individual carboxylic acids analyzed by TC/EAIRMS, in light of the other On species typically found in AEO, it is possible that the CO2 generated during pyrolysis of AEO is not solely derived from carboxylic group carbon. Whether this is the case or not, the operationally defined “pyrolysis” provides an adequate description for the reliable and precise intramolecular δ13C signatures reported here. In contrast to the lack of significant intramolecular carbon isotope variability within the different mass fractions comprising AEO in any given sample, a large difference in δ13Cpyr values of up to ∼9‰ was found between different sample types. Unprocessed McMurray Formation oil sand, deep McMurray Formation groundwater, and three separate oil sands process water samples taken from three different tailings ponds were all relatively more 13C-enriched compared to the Athabasca River and other shallow groundwater samples (Figure 3). While it appears that the unprocessed oil sand (−20.0‰) may be slightly more enriched than process water (−20.9 to −21.6‰) or McMurray Formation groundwater (−21.1‰), the small number of oil sand and McMurray Formation groundwater samples analyzed thus far makes definite assessment of this slight potential difference not feasible at this time. The isotopic similarity between process water and McMurray Formation groundwater, which represents the aqueous “natural-background” bitumen-derived AEO, implies that the alkaline hot water extraction process used for bitumen recovery in mining operations does not produce a significant carboxyl group carbon isotope fractionation effect. Consequently, discrimination between different bitumen-derived polar organics in water may not be possible using intramolecular isotopic characterization. On the other hand, the large difference in δ13Cpyr values observed between process water and surface water/groundwater samples suggests that a clear distinction can be made between other types of “natural background” AEO (i.e., fulvic and humic acids) and mining-related AEO containing naphthenic acids. This distinction would not be possible using the “bulk” (i.e., not intramolecular) δ13C signatures of AEO, which showed no systematic variation between any of the samples analyzed here (Figure 3; −31.0 to −29.0‰). In conjunction with an understanding of local hydrogeological conditions, it thus becomes possible at some sites to quantify mining-related AEO in the subsurface using an intramolecular carbon isotope mass balance containing both process water and natural background end-members. The application of δ13Cpyr values to delineate sources of AEO in groundwater adjacent to a major Athabasca oil sands tailings pond is the subject of a separate paper currently in preparation. The lack of significant carbon isotopic variation between the mass fractions separated by PCGC suggests a common origin for AEO within any particular sample regardless of the heterogeneity of its constituents. For the samples containing predominantly bitumen-derived AEO (oil sand, process water, and McMurray formation groundwater), the around 10‰ enrichment relative to bulk δ13C suggests a possible exchange with isotopically heavier DIC at one point during the formation of the Athabasca oil sands deposits as has been previously reported for light MW acids under hydrous-pyrolysis conditions.16,28 Because of the similarity in δ13Cpyr values for bitumen-derived AEO samples, this exchange must have occurred at temperatures >80 °C, which is the maximum temperature historically used in the Clark hot water process for

the commercial extraction of bitumen.29 Coincidentally, ∼80 °C is considered to be the upper limit for most subsurface microbial activity,30,31 and Athabasca oil sands reservoir temperatures have remained considerably lower than this over the last 60 Ma.31 This implies that a component of high MW Athabasca oil sands acids may not have originated from microbial degradation of oil, as is generally thought.32−34 Instead, they may be a product of the thermocatalytic degradation of kerogen, as has been demonstrated experimentally for low MW acids.16,35−37 A thermochemical origin for NAs at temperatures greater than around 100 °C due to oxidation of hydrocarbons by S0 has been previously suggested,38 and iron-bearing minerals are reactive inorganic sedimentary components that can also oxidize hydrocarbons to form carboxylic acids.39 On the other hand, it may be that the relative 13C-enrichment found in bitumen-derived AEO is not related to exchange with DIC but is instead a function of reservoir biodegradation processes at lower temperatures (i.e.,