Bicyclic naphthenic acids in oil sands process water: Identification by comprehensive multidimensional gas chromatography–mass spectrometry

Journal of Chromatography A, 1378 (2015) 74–87

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Bicyclic naphthenic acids in oil sands process water: Identification by comprehensive multidimensional gas chromatography–mass spectrometry Michael J. Wilde a , Charles E. West a,1 , Alan G. Scarlett a , David Jones a , Richard A. Frank b , L. Mark Hewitt b , Steven J. Rowland a,∗ a b

Petroleum and Environmental Geochemistry Group, Biogeochemistry Research Centre, University of Plymouth, Drake Circus, Plymouth PL4 8AA, UK Water Science and Technology Directorate, Environment Canada, 867 Lakeshore Road, Burlington, ON, Canada L7R 4A6

a r t i c l e

i n f o

Article history: Received 20 October 2014 Received in revised form 30 November 2014 Accepted 3 December 2014 Available online 12 December 2014 Keywords: Naphthenic acids Bicyclics GC × GC–MS

a b s t r a c t Although bicyclic acids have been reported to be the major naphthenic acids in oil sands processaffected water (OSPW) and a well-accepted screening assay indicated that some bicyclics were the most acutely toxic acids tested, none have yet been identified. Here we show by comprehensive multidimensional gas chromatography–mass spectrometry (GC × GC–MS), that >100 C8–15 bicyclic acids are typically present in OSPW. Synthesis or purchase allowed us to establish the GC × GC retention times of methyl esters of numerous of these and the mass spectra and published spectra of some additional types, allowed us to identify bicyclo[2.2.1]heptane, bicyclo[3.2.1]octane, bicyclo[4.3.0]nonane, bicyclo[3.3.1]nonane and bicyclo[4.4.0]decane acids in OSPW and a bicyclo[2.2.2]octane acid in a commercial acid mixture. The retention positions of authentic bicyclo[3.3.0]octane and bicyclo[4.2.0]octane carboxylic acid methyl esters and published retention indices, showed these were also possibilities, as were bicyclo[3.1.1]heptane acids. Bicyclo[5.3.0]decane and cyclopentylcyclopentane carboxylic acids were ruled out in the samples analysed, on the basis that the corresponding alkanes eluted well after bicyclo[4.4.0]decane (latest eluting acids). Bicyclo[4.2.1]nonane, bicyclo[3.2.2]nonane, bicyclo[3.3.2]decane, bicyclo[4.2.2]decane and spiro[4.5]decane carboxylic acids could not be ruled out or in, as no authentic compounds or literature data were available. Mass spectra of the methyl esters of the higher bicyclic C12–15 acids suggested that many were simply analogues of the acids identified above, with longer alkanoate chains and/or alkyl substituents. Our hypothesis is that these acids represent the biotransformation products of the initially somewhat more bio-resistant bicyclanes of petroleum. Although remediation studies suggest that many bicyclic acids can be relatively quickly removed from suitably treated OSPW, examination by GC × GC–MS may show which isomers are affected most. Knowledge of the structures will allow the toxicity of any residual isomers to be calculated and measured. © 2014 Elsevier B.V. All rights reserved.

1. Introduction ‘Naphthenic acids’ occurring naturally in the oil sands of Alberta, Canada are concentrated by processing, resulting in oil sands process-affected water (OSPW) which, after much re-use, is stored in large tailings ponds or lagoons, awaiting final reclamation [1]. Undiluted OSPW has been shown to be somewhat toxic in numerous biological assays, but with time in storage the composition and

∗ Corresponding author. Tel.: +44 01752 584557; fax: +44 01752 584710. E-mail address: [email protected] (S.J. Rowland). 1 Present address: EXPEC Advanced Research Center, Saudi Aramco, Dhahran 31311, Saudi Arabia. http://dx.doi.org/10.1016/j.chroma.2014.12.008 0021-9673/© 2014 Elsevier B.V. All rights reserved.

toxicity changes, the latter usually reducing [2]. Nonetheless, residual toxicity remains and this has promoted numerous studies of treatment methods with oxidants or ozone, or by photocatalysis or bioremediation [3]. Numerous studies have shown that the major acids in different OSPW samples comprise, as a group, unknown alicyclic bicyclic compounds [2,4–7] and a well-accepted screening assay indicated that some synthetic alicyclic bicyclics were the most acutely toxic acids tested [8]. However, almost nothing is known about the identities, or even the numbers, of bicyclic acids present in OSPW. Cyr and Strausz [9] isolated a C16 bicyclic acid from oil sands deposits in Alberta which had a mass spectrum similar to that of drimane or labdane bicyclanes, but these have not yet been reported in OSPW acids (cf [10–12]). Bowman et al. [13] recently identified

M.J. Wilde et al. / J. Chromatogr. A 1378 (2015) 74–87

bicyclic monoaromatic, indane and tetralin acids in a pore water sample from a composite tailings deposit, which combines fluid fine tailings from oil sands processing with gypsum to form a nonsegregating deposit, but no alicyclic bicyclics were identified. Elucidation of acid structures also has geochemical significance, providing an insight into the microbial degradation mechanisms of petroleum [14]. Some alicyclic bicyclic acids in crude oils and commercial naphthenic acids preparations derived from refining petroleum, have been identified [14–16], but several studies have noted the differences between the latter and OSPW acids, so perhaps nothing directly can be inferred from a comparison [17]. Furthermore, the few bicyclic acids identified in commercial naphthenic acids to date represent only a small fraction of those actually present, as the >100 compounds revealed by comprehensive multidimensional gas chromatography–mass spectrometry of the methyl esters (GC × GC–MS) of two commercial naphthenic acids mixtures attests [18]. Fortunately the bicyclic acids in OSPW seem to be quite prone to removal by ozone treatment and bacterial action [3]. Nonetheless, it is important to establish the identities of these acids so that the toxicity of relevant isomers can be measured, the mechanisms of remediation treatments better understood and the products of remediation treatment predicted. In the present study we examined several methylated OSPW acidic extracts and a commercial acid mixture, by GC × GC–MS and identified several of the bicyclic acids present. Some bicyclics previously assumed to be representative of OSPW constituents, were not common.

2. Materials and methods The naming of bicyclic compounds varies considerably throughout the literature. As an attempt to keep the naming of the compounds discussed consistent, the IUPAC nomenclature rules for polycyclic compounds based on the Von Baeyer system [19] have been used, with numbering of the carbons within the bicyclic core starting at a bridgehead carbon (Fig. 1A and B). Alternative names for compounds commonly used by chemical suppliers and search engines (e.g. decalin or octahydro-pentalene) are given alongside the systematic names. Authentic bicyclo[2.2.1]heptane-2-ethanoic acid (Fig. 1A; Structure Ib), 2,6,6-trimethylbicyclo[3.1.1]heptane-3-carboxylic acid ((+)-3-pinanecarboxylic acid) (IIa), bicyclo[2.2.2]octane-2carboxylic acid (IVa), 4-pentylbicyclo[2.2.2]octane-1-carboxylic acid (IVc), bicyclo[3.3.0]octane-2-carboxylic acid (VIa), 4-methylacid (3-methyl-octahydrobicyclo[3.3.0]octane-2-carboxylic pentalene-1-carboxylic acid) (VIb) and bicyclo[3.3.1]nonane1-carboxylic acid (VIIIa) were purchased from Sigma (Poole, UK). Authentic bicyclo[2.2.1]heptane-1-carboxylic acid (Ia), bicyclo[2.2.2]octane-1-carboxylic acid (IVb) and 5-methylbicyclo[3.3.1]nonane-1-carboxylic acid (VIIIc) were purchased from Molport (Riga, Latvia). Bicyclo[3.2.1]octane-6-carboxylic acid (Va) was synthesised from 2-hydroxybicyclo[3.2.1]octane-6carboxylic acid (Sigma) by base catalysed dehydration followed by hydrogenation [20]. Bicyclo[3.3.1]nonane-3-carboxylic acid (VIIIb) was synthesised essentially by the methods of Sasaki et al. [21] as modified by Peters et al. [22]. Thus, reaction of adamantan-2-one in methanesulphonic acid in the presence of sodium azide produced the mesylate which was not isolated but heated with potassium hydroxide to give the unsaturated bicyclo[3.3.1]non-2-ene-7-carboxylic acid, obtained after extraction into acidified chloroform [21]. The corresponding saturated bicyclo[3.3.1]nonane-3-carboxylic acid (VIIIb) was obtained by hydrogenation [22] and the methyl esters by heating with BF3 /methanol. Bicyclo[4.3.0]nonane-3-carboxylic (Xa)

75

and 2-methylbicyclo[4.3.0]nonane-3-carboxylic acids (Xb) were obtained by catalytic hydrogenation (cf [15]) of the corresponding indane acids (Sigma). Bicyclo[4.4.0]decane-2-carboxylic (XIVa), 3-carboxylic acid (XIVb), 2-ethanoic (XIVc), 3-ethanoic (XIVd), and 2-propanoic acids (XIVe; numbers refer to position of alkanoate substituents on bicyclic core) were synthesised as described [15]. 7-methylbicyclo[4.2.0]octane-7-carboxylic previously acid (VIIa) was prepared by hydrogenating 1-methyl-1,2dihydrocyclobutabenzene-1-carboxylic acid methyl ester over a Raney Nickel catalyst at 100 ◦ C and 100 bar using a H-Cube® (ThalesNano Nanotechnology Inc., Budapest). Four different OSPW samples and a commercial naphthenic acids mixture were analysed (Table 1). The OSPW included two samples (#1 and #2) from industry A described in two previous studies [23,24]. Briefly, sodium salt concentrates of #1 and #2 were acidified to pH < 2 and the acids extracted with ethyl acetate before derivatisation with BF3 /methanol [24]. Another OSPW (#3) was provided from industry B (Table 1) at a site with a high concentration of particulate matter. This water sample was filtered, acidified and then eluted through a 200 mg ENV+ SPE cartridge with acetonitrile before being dried under N2 and derivatised with BF3 /methanol. A fourth OSPW acid extract (#4) from industry A was obtained by extracting a sample of raw OSPW, collected from a different tailings pond using the methods described previously [24]. The latter sample had undergone no pre-treatment/clean-up prior to extraction and derivatisation. In addition to the above samples, a commercial naphthenic acids mixture (#5) was obtained from Merichem Co. for comparison (Table 1) and fractionated based on a method previously used [24–26]. Derivatisation of the acids with BF3 /methanol was followed by silver ion solid phase extraction (Ag+ SPE). Analysis herein focused on fraction 3 obtained by elution through the argentation solid phase extraction column with hexane, since this contained the bicyclic acids (methyl esters). Accurate mass measurements were made using a Thermofisher LTQ Orbitrap XL high resolution mass spectrometer with electrospray ionisation. The mass range was m/z 120–2000; mass accuracy <3 ppm RMS with external calibration. For negative ionisation the instrument was externally calibrated using the above, sodium dodecyl sulphate and sodium taurocholate. For loop-injections a Thermo Scientific Surveyor MicroLC was used to provide solvent flow at 20 ␮L min−1 ., through a 2 ␮L sample loop. Solvents used were H2 O:MeOH (1:1). For nano-electrospray an Advion Triversa NanoMate was used to deliver samples diluted into MeOH ± 10% NH4 OAc at a flow of approximately 0.25 ␮L min−1 . API source settings: Infusion NanoMate source temperature 275 ◦ C or 200 ◦ C, sheath gas flow 3–7 (arb. units) 2 (arb. units), aux gas flow was not used capillary (ionising) voltage positive ionisation: +3.2 to 3.7 kV negative ionisation: −3.5 to −4.0 kV. Mass spectra were acquired at a minimum resolution of 30,000 (at m/z 400). Theoretical masses and mass accuracies were calculated using an online calculator tool [27]. Comprehensive multidimensional gas chromatography–mass spectrometry (GC × GC–MS) analyses were conducted as described previously [23,28], using an Agilent 7890A gas chromatograph (Agilent Technologies, Wilmington, DE) fitted with a Zoex ZX2 GC × GC cryogenic modulator (Houston, TX, USA) interfaced with an Almsco BenchTOFdxTM time-of-flight mass spectrometer (Almsco International, Llantrisant, Wales, UK). The first-dimension column was a 100% dimethyl polysiloxane 60 m × 0.25 mm × 0.25 ␮m Rxi® -1ms (Restek, Bellefonte, USA), and the second-dimension column was a 50% phenyl polysilphenylene siloxane 2.5 m × 0.1 mm × 0.1 ␮m BPX50 (SGE, Melbourne, Australia). Helium was used as carrier gas and the flow was kept constant at 1.0 mL min−1 . Samples (1 ␮L) were injected at 300 ◦ C splitless. The oven was programmed from 40 ◦ C (hold for 1 min), then heated to 130 ◦ C at 10 ◦ C min−1 then at

76

M.J. Wilde et al. / J. Chromatogr. A 1378 (2015) 74–87

Fig. 1. (A) Structures of synthesised or purchased authentic bicyclic acids and (B) examples of generalised structures and names of possible C11 bicyclic acids.

M.J. Wilde et al. / J. Chromatogr. A 1378 (2015) 74–87

77

Table 1 Details of samples studied with assigned sample numbers. Sample No. #

Sample type

Originator

Details

1 2 3 4 5

OSPW OSPW OSPW OSPW Commercial Naphthenic acids

Industry A Industry A Industry B Industry A Merichem

Collected 2004 Collected 2009, isolated concentrated naphthenates 2011 Collected 2011 Collected 2013 Batch no. CN/138 CAS# 1338-24-5

2 ◦ C min−1 to 320 ◦ C (held for 15 min). The modulation period was 6 s. The MS transfer line temperature was 290 ◦ C and ion source 300 ◦ C. Data processing was conducted using GC ImageTM v2.1 (Zoex, Houston, TX, USA).

3. Results and discussion When examined by ESI Orbitrap high-resolution mass spectrometry with negative ion electrospray ionisation, the bicyclic acids of the OSPW extracts produced ions which we attributed to the [M−H]− ions, mainly of C12–16 acids (m/z 195–251; ESI Orbitrap), although some samples had small amounts of lower carbon number species. Although OSPW is known to be a heterogeneous substrate, this is consistent with the data presented in numerous studies of different OSPW samples [2,4–7]. The accurate masses of some of the more abundant acids in one of the OSPW samples (#1), for example were: 209.1544 (C13 H21 O2 requires 209.1547, mass accuracy 1.4 ppm; C13 H21 S requires 209.1369 and C11 H13 SO2 requires 209.0642), 223.1699 (C14 H23 O2 requires 223.1704, mass accuracy 2.2 ppm; C14 H23 S requires 223.1526 and C12 H15 SO2 requires 223.0798) and 237.1855 (C15 H25 O2 requires 237.1860, mass accuracy 2.1 ppm; C15 H25 S requires 237.1682 and C13 H17 SO2 requires 237.0955), indicating that the major ionised bicyclic species were acids fitting the formula Cn H2n−4 O2 and not, for example, nominally isobaric keto bicyclics or tricyclic hydroxy acids or sulphur compounds (at least in this OSPW sample). When examined by GC × GC–MS as the methyl esters, selected ion mass chromatography of the molecular ions produced by electron ionisation confirmed the presence of C8–15 bicyclic acids in samples #3 and #4, with at least C11–15 bicyclics in all the samples (#1–#5; e.g. Fig. 2). Moreover, GC × GC–MS revealed the true complexity of the OSPW mixtures. The exact ranges varied between samples. For example, one OSPW acid extract (#3) contained at least nineteen C9 , twenty seven C10 , forty C11 and numerous C11 + peaks within the chromatogram (Fig. 2B), whereas the extract from another OSPW tailings pond (#4) appeared even more complex (Fig. 2C), possibly due to the lack of pre-treatment or clean-up, which may have removed some lower molecular weight compounds in the more treated samples (e.g. #1 and #2, Figs. 2A and S1). Recently Damasceno et al. [18] analysed two commercial acid mixtures by GC × GC–MS, characterising groups of naphthenic acids by their ‘z’ value (i.e. hydrogen deficiency attributed to the number of rings). They detected 124 and 132 individual bicyclic acids (z = -4) in two samples (Sigma Aldrich and Miracema-Nuodex naphthenic acids) with carbon numbers ranging from C9–16 [18]. Similar numbers were detected in a sample of Merichem commercial acids herein (#5, Fig. 2D), so similar numbers of bicyclics appear to be present in OSPW and commercial naphthenic acids. Such large numbers must represent many different structural types of bicyclic acids, not just those routinely cited as examples [29–31]. We attempted to calculate the likely maximum possible number present for the simplest (C8–11 ) acids. We assumed that at least one (necessarily), and sometimes two, carbon atoms would be associated with the carboxylate/alkanoate chain. The latter is

reasonable based on the identifications of ethanoate side chains of the co-occurring tricyclic and pentacyclic acids [20,23,32] and what is known of the biodegradation processes from which the acids originate [33–35]. Thus, for a C11 acid, for example, at most ten carbons are left for formation of the bicyclic ‘core’ of the acid. If any alkyl substituents were present, the number of carbon atoms in the ‘core’ would be less than ten and more alkylation would be present. Since alkyl groups identified or tentatively established in OSPW acids to date have not exceeded those comprising four carbon atoms in total (e.g. a combination of ethyl and methyl groups), it is reasonable to assume that the smallest number of atoms in the bicyclic ‘core’ would probably be six. We calculated that three structural types exist for acids with a C6 core. These have cyclopropyl- or cyclobutyl rings; the former are present in carane- and thujane-type compounds and the latter present in bicyclo[3.1.1]heptanes (pinanes), bicyclo[4.2.0]octanes and bicyclo[2.2.0]hexanes, which are known in the ladderane acids [36]. For the acids with a C7 core and the requisite substituents, there are four structural types (examples given in Fig. 1B; I–III), for the C8 core acids, six (e.g. Fig. 1B; IV–VII), for the acids with a C9 core, seven (e.g. Fig. 1B; VIII–XI) and for the acids with a C10 core, nine possible structures (where the rings are fused at two carbon atoms, e.g. Fig. 1B; XII–XVI). Spiro- and nonfused structures were also considered, such as spiro[4.5]decane carboxylic acid (Fig. 1B; XVII), fused at one carbon atom and the nonfused cyclopentylcyclopentane carboxylic acid (Fig. 1B; XVIII). Thus, our calculations suggest that even the simplest acids in the OSPW sample might comprise over 30 structural types and for each of these, many stereoisomers exist. Examples of some of these bicyclic structural types are given in Fig. 1B; most ring types have been identified within natural products. Thus, in theory, it is easy to account for the >100 bicyclic acids we observed in the OSPW and commercial acids. The remaining analytical challenge is to identify what at least some of these actually are. Examination by GC × GC–MS, of methyl esters of authentic purchased acids or those synthesised herein, allowed us to establish GC retention regions of nine structural types (Fig. S1). It was clear from the GC1 retention positions of pseudohomologues within a given structural type (e.g. the bicyclo[4.4.0]decane (decalin) carboxylic, ethanoic and propanoic acids; Fig. S1) that increasing molecular weight within a structural class increased the GC1 retention times in an approximately linear fashion, as expected. Also, as expected, similar homologues (e.g. C10 ) from different structural types, were generally quite well separated both in the GC1 and GC2 dimensions. For example, the methyl ester of 4-methylbicyclo[3.3.0]octane-2-carboxylic acid (Fig. S1; C10 acid VIb) was well separated (GC1) from those of the bicyclo[3.3.1]nonane-1- and 3-carboxylic acids (Fig. S1; C10 acids VIIIa and b). Within a group of related stereoisomers (e.g. cis/trans, or positional isomers) of a particular acid (e.g. isomers of bicyclo[4.4.0]decane acids) the relative retention positions produced a so-called grouping or ‘tiling’ effect, with both GC1 and GC2 retention positions differing (by up to about 50 retention index units in GC1) between isomers (Fig. S1). The combined effects produced chromatograms in which the profiles of the complex distributions of individual OSPW bicyclic acids of carbon numbers C8–15 were

78

M.J. Wilde et al. / J. Chromatogr. A 1378 (2015) 74–87

Fig. 2. Extracted ion chromatograms for molecular ions of C8–15 bicyclic acid methyl esters (m/z 154, 168, 182, 196, 210, 224, 238 and 252) in (A) an OSPW acid extract from industry A collected 2009, #2, (B) an OSPW acid extract from industry B, #3, (C) an OSPW acid extract from industry A collected from a different tailings pond in 2013, #4 and (D) a fraction of Merichem acid extract, #5.

apparent (Figs. 2 and S1). In general, the GC2 retention position seems to give a good separation of the different structural classes of acids (Fig. S1) but is also clear that differences in positional substitution (e.g. of the bicyclo[3.3.1]nonane-3- and 1-carboxylic acids and decalin-1- and 2-carboxylic acids) can produce a difference of about 0.25 s in the GC2 dimension. The quaternary-substituted acid (VIIIa) eluted earlier in the GC2 dimension, as did the 3-substituted decalin acids (Fig. S1). We next undertook a systematic examination of the retention positions and mass spectra of the >100 individual GC × GC peaks and compared these with those of reference compounds. Since interpretation of the data for the lower homologues was likely to be simplest and might give clues to the identities of the presumably more alkylated higher homologues, we began with the C8–10 acids. 3.1. C8 bicyclic acids C8 acids were present in the OSPW acid extracts #3 and #4. Fig. 3 shows the GC × GC retention positions of a series of peaks within #3 and #4, with the expected retention positions and molecular ions (m/z 154) of C8 bicyclic acid methyl esters indicated. One peak which was present within both OSPW acid extracts, #3 and #4 (Fig. 3A and B, peak 1a) was identified as bicyclo[2.2.1]heptane2-carboxylic acid methyl ester after comparison with a NIST library spectrum (exo-bicyclo[2.2.1]heptane-2-carboxylic acid when compared with the mass spectrum reported by Curcuruto et al. [37]) (Fig. 3C and D). Interpretation of the mass spectrum of a second peak (Fig. 3A and B, peak 1b) resulted in the identification of

bicyclo[2.2.1]heptane-1-carboxylic acid methyl ester, confirmed by matching the GC × GC retention time and mass spectrum with that of an authentic standard (Fig. 3E and F). The other peaks within this series (Fig. 3A and B, peaks 1c and d) were believed to be isomers possessing the same bicyclo[2.2.1]heptane core; either endo/exo isomers, or isomers with the methyl carboxylate group substituted elsewhere on the ring (mass spectra detailed in supplementary information Fig. S2). Examination of the retention behaviour of purchased bicyclo[2.2.1]heptane ethanoic acid (Fig. 1A; Ib) compared with an OSPW acid extract which did not contain C8 acids (e.g. #1 and #2), also indicated that more alkylated bicyclo[2.2.1]heptanes were present (Fig. S1) and literature data were also available for the retention indices of some C8–11 isomers of the latter on apolar and polar phases [38]. These also suggested that numerous bicycloheptane acids were possibilities for the unknowns (Fig. S1). Compounds with the bicyclo[2.2.1]heptane skeletons (Fig. 1B; I, e.g. norbornane and bornane), are well-known in nature and are most often encountered as derivatives of camphor. Thus, there is precedence for the biosynthesis of compounds with this skeleton and numerous analogues have been studied. Seifert and Teeter [39] suggested that naphthenic acids from a Californian petroleum might include such structural types. GC retention indices on apolar and polar phases and mass spectra or partial spectra of the methyl esters of isomers of C8–11 acids have been published [37,38,40] and we obtained the mass spectrum of the methyl ester of the C9 bicyclo[2.2.1]heptane ethanoic acid (Figs. 1A; Ib and S3). Common spectral features seem to be small molecular ions (<10% abundant) and abundant (often base peak) ions at m/z 95 (Fig. S2C and D). The

M.J. Wilde et al. / J. Chromatogr. A 1378 (2015) 74–87

79

Fig. 3. Extracted ion chromatograms (m/z 154, 125, 95 and 87) of (A and B) two OSPW acid extracts from industries A and B (#3 and #4), analysed by GC × GC–MS and mass spectra of (C and E) unknown peaks 1a and 1b, identified by comparison with (D) a NIST library spectrum of bicyclo[2.2.1]heptane-2-carboxylic acid methyl ester and (F) a purchased reference standard of bicyclo[2.2.1]heptane-1-carboxylic acid methyl ester. Unknown peaks labelled 1c and 1d were speculated to be bicyclo[2.2.1]heptane carboxylic acid methyl ester isomers based on mass spectral interpretation (mass spectra are given in supplementary information, Fig. S2).

abundance of the molecular ion can vary dramatically, however, in different stereoisomers of the same acid type (vide infra). The retention position of 2,6,6-trimethylbicyclo[3.1.1]heptane3-carboxylic acid methyl ester meant C8 , as well as higher bicyclo[3.1.1]heptane acids (Fig. S1), were also a possibility for the identities of some of the unknowns, but no exact match was found in the spectra of the OSPW acids (methyl esters). The most common compounds found with bicyclo[3.1.1]heptane (Fig. 1B; II) skeletons are pinenes; trimethylmonterpenes produced by plants, particularly abundant in resin from pine trees (i.e. turpentine oil). The mass spectrum of

2,6,6-trimethylbicyclo[3.1.1]heptane-3-carboxylic acid methyl ester (3-pinanecarboxylic acid methyl ester) is complex (Fig. S4), perhaps as a result of the bridged structure containing a cyclobutane ring within the core. The cyclobutane ring makes the bicyclic acid liable to ring-opening and subsequent rearrangement. This is supported by an extremely low molecular ion abundance (<2%) at m/z 196 (Fig. S4). Distinguishable features of the mass spectrum included a strong (95%) M-60 ion (m/z 136) corresponding to loss of the methyl carboxylate moiety with a hydrogen transfer and a base peak at m/z 81, as well as an intense ion at m/z 83 consistent with cyclic C6 H9 + and C6 H11 + ions respectively (Fig. S4).

80

M.J. Wilde et al. / J. Chromatogr. A 1378 (2015) 74–87

3.2. C9 bicyclic acids Next, we examined the GC retention positions of commercially available C9 (and C14 ) bicyclo[2.2.2]octane acids (Fig. 1A; IVa and c). When compared with the retention positions of unknown bicyclic acids within OSPW acid extracts #1 and #2 (Figs. S1 and S2) both reference acids had relatively long GC2 retention times but despite not being identical to those of some unknowns, bicyclo[2.2.2]octane acids were considered as possible identities. Furthermore, bicyclo[2.2.2]octane-1-carboxylic acid methyl ester (Fig. 1A; IVb) was identified within the Merichem acid extract (#5); an unknown peak had a matching retention position and mass spectrum to that of the authentic reference compound (Fig. 4B and C, peak 2a). Mass spectral interpretation of other peaks within the series indicated other isomers were also present. Compounds with the bicyclo[2.2.2]octane (Fig. 1B; IV) skeleton are found as stable, cage-like skeletons in natural products such as eremolactone [41,42] isolated from Eremophila fraseri and (−)-seychellene [43] found in patchouli oil, extracted from Pogostemon cablin. In addition to bicyclo[2.2.2]octane-1-carboxylic acid (Fig. 1A; IVb) we were able to purchase bicyclo[2.2.2]octane-2carboxylic acid (Fig. 1A; IVa) and 4-pentylbicyclo[2.2.2]octane1-carboxylic acid (Fig. 1A; IVc) and obtain the spectrum of the methyl esters (Fig. S5A and B). Whilst the mass spectrum of bicyclo[2.2.2]octane-2-carboxylic acid methyl ester (Fig. 1A; IVa) was characterised by a small molecular ion (m/z 168) and base peak ion (m/z 136) due to loss of methanol from the latter (Fig. S5A), the mass spectrum of bicyclo[2.2.2]octane-1-carboxylic acid methyl ester (Fig. 4C) contained pronounced molecular and M-29 ions, similar to those of some of the unknowns, as did the mass spectrum of the 4-methyl-1-carboxylic acid isomer (free acid, NIST library) perhaps due to the loss of C2 H5 . The mass spectrum of the C14 4-pentylbicyclo[2.2.2]octane-1-carboxylic acid methyl ester (Fig. S5B) also showed a fairly abundant molecular ion (m/z 238) and the loss of M-29 and M-28 (m/z 209 and 210). Denisov et al. [44] reported the mass spectra of a range of substituted bicyclo[2.2.2]octanes with many showing loss of an ethyl group (C2 H5 , 29 Da) from the molecular ion. They proposed a mechanism for the loss of ethyl from a monocyclic intermediate brought about by the rupture of a bond at a bridgehead carbon coupled with a hydrogen transfer [44]. When examined by GC × GC–MS, bicyclo[3.2.1]octane-6carboxylic acid methyl ester was identified within #3 and #4. The retention position and mass spectrum of the authentic reference compound (Fig. 1A; Va) matched those of an unknown peak present in both samples (Fig. 5C–E, peak 3a). Compounds with the bicyclo[3.2.1]octane-type skeleton are common in several natural products [45]. However the hydrocarbon, bicyclo[3.2.1]octane and alkyl substituted homologues have also long been known in petroleum [46,47]. The mass spectrum of bicyclo[3.2.1]octane-6-carboxylic acid (Fig. 5E) contained a small molecular ion (m/z 168), ion attributed to methanol loss (m/z 136) and a base peak ion typical of methyl esters (m/z 87). Similarly, cis-bicyclo[3.3.0]octane was identified in petroleum over 50 years ago [47]. Previously we identified 4-methylbicyclo [3.3.0]octane-2-carboxylic acid in a commercial sample of naphthenic acids [15] by comparison of the mass spectrum with that of a purchased reference sample. Since then we were able to purchase the C9 parent acid, bicyclo[3.3.0]octane-2-carboxylic acid (two isomers) and comparison of the mass spectra and GC × GC retention times has now led to the identification of the corresponding methyl esters within a fraction of Merichem acid extract (#5, Fig. S6B–E). Another unknown compound within the Merichem acid extract, also in the OSPW acid extract from industry B (#3) and a different tailings from industry A (#4), had a very similar mass spectrum to that of the minor bicyclo[3.3.0]octane-2-carboxylic acid methyl

ester isomer (Fig. S7, peak 7c). However, the retention time of the unknown was different from the ester of the authentic acid and thus the unknown was postulated to be a different isomer. The mass spectrum contained an ion at m/z 150, attributed to the loss of water (M+ -18; Fig. S7). The loss of water is often observed in the mass spectra of non-derivatised acids, keto- or hydroxy acids, but uncommon in spectra of methyl esters. However, loss of water (M+ 18) is observed in the mass spectra of some bicyclo[4.4.0]decane acid methyl esters and again appears to be specific to certain isomers [15]. The mass spectrum of the unknown displayed an ion at m/z 74 (Fig. S7), also a characteristic ion of methyl esters suggesting, it was not a non-methylated C10 acid. The molecular ion did not show multiple isotopic peaks suggesting the compound did not contain sulphur and the lack of tailing in the chromatogram often observed for non-derivatised or more polar compounds indicated it was not a keto- or hydroxy acid and was most likely a different isomer of bicyclo[3.3.0]octane-2-carboxylic acid. 3.3. C10 bicyclic acids The retention positions of the synthetic bicyclo[3.3.1]nonane1- and 3-carboxylic acid methyl esters (Fig. 1A; VIIIa and b) showed that these acids were absent from some OSPW samples (Fig. S1). However two unknown acids in #5 (Fig. 6; peaks 4a and c) had matching retention positions and mass spectra with those of bicyclo[3.3.1]nonane-3-carboxylic acid methyl ester and bicyclo[3.3.1]nonane-1-carboxylic acid methyl ester (Fig. 6B, C, F and G). One unknown (Fig. 6; peak 4b) had the same retention position and a mass spectrum containing similar ions but with different intensities to that of a C11 homologue, 5-methylbicyclo[3.3.1]nonane-1-carboxylic acid methyl ester (Fig. 6D and E). The GC × GC retention position and mass spectrum of authentic 5-methylbicyclo[3.3.1]nonane-1carboxylic acid methyl ester was similar to that of an unknown present within the OSPW acid extract from industry B, #3 (Fig. S8). Previously it has been speculated that biodegradation of adamantanes might produce ring-opened acids with the bicyclo[3.3.1]nonane skeleton, since this occurs in the biodegradation of adamantan-2-one [23,48]. Indeed this seems to be possible, at least for some of the present samples (viz.: #3 and #5). The data suggest the acids in the OSPW extracts (and some commercial acids) sometimes included bicyclo[3.3.1]nonane carboxylic acids. Bicyclo[4.3.0]nonane carboxylic acids (e.g. Fig. 1B; X) were also identified previously in a commercial naphthenic acids mixture by comparison of the mass spectra with literature mass spectra of the methyl esters of synthetic 2-carboxylic acid isomers [15]. In the present study, we were able to synthesise the corresponding 3carboxylic acids (Figs. 1A; Xa and S9) and a 2-methyl-3-carboxylic acid isomer (Figs. 1A; Xb and S10). The GC2 retention times of the bicyclo[4.3.0]nonane acid standards were generally greater than those of most of the unknowns (Fig. S1). However, a few of the unknowns within an OSPW acid extract (#3) possessed mass spectra very similar to those previously identified as C9 bicyclo[4.3.0]nonane carboxylic acids (methyl esters) in a commercial acid mixture [15,37] (Fig. 7C and D) as well as the synthesised standards (Figs. 7(A, B and E, F) and S9). It is even more likely that members of the C12–15 acids include this structural type, since there are several more late-eluting peaks in these classes (Figs. 2 and S1). These data suggest the acids in the OSPW extracts sometimes include bicyclo[4.3.0]nonane carboxylic acids. The spectra of the bicyclo[4.3.0]nonane acids within the commercial acid mixture were characterised by medium abundance molecular ions (ca 20%) and in the C10 parent acid, bicyclo[4.3.0]nonane-2-carboxylic acid methyl ester, by ions due to loss of methanol (m/z 150) and m/z 87 [15,49]. The mass spectra of the isomers of the synthesised bicyclo[4.3.0]nonane-3-carboxylic

M.J. Wilde et al. / J. Chromatogr. A 1378 (2015) 74–87

81

Fig. 4. Extracted ion chromatogram (m/z 168) of (A) a fraction of Merichem acid extract (#5) analysed by GC × GC–MS and mass spectrum of (B) unknown peak 2a, identified by comparison with (C) that of authentic bicyclo[2.2.2]octane-1-carboxylic acid methyl ester with the same GC × GC retention position.

acid methyl esters varied considerably (Fig. S9). Thus in the major isomer (69% of total resolved peaks) the molecular ion was abundant (80%; Fig. S9), whereas in more minor isomers the molecular ion was only <5% abundant (Fig. S9). The mass spectra were easily distinguished from those of the 2-isomers and we can now assign the isomers present in commercial acids studied herein (#5) and previously [15], to both 3-isomers and almost certainly 2-isomers, given the mass spectra (Fig. 7). The GC × GC retention position of 4methylbicyclo[3.3.0]octane-2-carboxylic acid (Fig. S1 and Fig. 1A; VIb) was close to those of some of the unknowns within the OSPW acid extracts (#1; Fig. S1). However, there was no exact retention time or mass spectral match. The mass spectrum of the C10 4-methylbicyclo[3.3.0]octane-2-carboxylic acid methyl ester (Fig. S11) was characterised by a quite strong (30%) molecular ion and characteristic fragment ions, particularly at m/z 140 (70%) assumed to be due to loss of a propene moiety, likely via a cycloreversion/retro-Diels–Alder rearrangement, typical of cyclic hydrocarbons [50,51]. The GC2 retention position of the C10 authentic acid methyl ester (Fig. S1) and the tentative identification of the C9 parent acid (Fig. S7) suggested that some of the C11 + acids present in the OSPW might have this skeleton. 7-methylbicyclo[4.2.0]octane-7-carboxylic acid methyl ester eluted very closely to unknown bicyclic acids in an OSPW acid extract (#1, Fig. S1) but none matched exactly the particular isomers present. Bicyclo[4.2.0]octane carboxylic acids contain a fused cyclobutane ring (Fig. 1B; VII), similar in structure to short-chain ladderane fatty acids previously identified as degradation products

of ladderane lipids [36]. Ladderane lipids are specific for bacteria capable of anaerobic ammonium oxidation (anammox) and therefore the acids can be used as biomarkers for anammox bacteria [52]. The mass spectra of both 7-methylbicyclo[4.2.0]octane-7carboxylic acid methyl ester isomers (Fig. S12) displayed weak molecular ions (m/z 182) as expected for alicyclic acids containing a highly strained, fused cyclobutane ring. The base peak at m/z 101 was attributed to the fragmentation across the cyclobutane ring.

3.4. C11 + bicyclic acids Bicyclic naphthenic acids, believed to be products of biodegradation, have frequently been assumed to possess bicyclo[4.4.0]decane structures (e.g. [29–31]) and this has been supported by the occurrence of such acids identified within at least one commercial acid mixture [15]. The retention positions of the synthetic bicyclo[4.4.0]decane (decalin) carboxylic, ethanoic and propanoic acid methyl esters substituted in either the 2- or 3-positions on the decalin core showed that these acids were absent or had a very low abundance in some of the samples of OSPW acids which we examined (#1 and #2), as demonstrated by the elution of these acids late in the GC2 retention window (#1, Fig. S1). A small number of bicyclo[4.4.0]decane acids were tentatively identified within another OSPW (#3), based on mass spectral comparison with those previously reported in commercial acid mixtures [15], such as an isomer of bicyclo[4.4.0]decane-3-carboxylic acid methyl ester, as

82

M.J. Wilde et al. / J. Chromatogr. A 1378 (2015) 74–87

Fig. 5. Extracted ion mass chromatograms (m/z 168 and 87) of (A and B) OSPW acid extracts from Industries A and B (#3 and #4) analysed by GC × GC–MS and mass spectra (C and D) of the same unknown (3a) present within both samples identified by comparison with (E) the mass spectrum and retention position of authentic bicyclo[3.2.1]octane6-carboxylic acid methyl ester.

well as bicyclo[4.4.0]decane-1-carboxylic acid methyl ester which was compared with a NIST library mass spectrum (Fig. 8). Petroleum hydrocarbons and related compounds possessing bicyclo[4.4.0]decane cores such as drimanes, cadinanes and eudesmanes have been well studied [12,53,54]. Fused cyclohexyl rings are common in biologically derived compounds, e.g. hopanes. Therefore, bicyclic sesquiterpenes can be reasonably postulated to be biodegradation products of higher terpenes [53]. Although we could obtain no samples of bicyclo[3.2.2]nonane (Fig. 1A; IX), bicyclo[4.2.1]nonane (XI), bicyclo[4.2.2]decane (XII), bicyclo[5.3.0]decane (XIII), bicyclo[5.2.1]decane (XV), bicyclo[3.3.2]decane (XVI), spiro[4.5]decane (XVII) or cyclopentylcyclopentane (XVIII) carboxylic acids, when we examined the reported NIST GC retention indices of the hydrocarbons

bicyclo[5.3.0]decane and cyclopentylcyclopentane, it was clear that these eluted well after decalin (bicyclo[4.4.0]decane). Since the acid methyl esters would be expected to have the same relative retention orders and bicyclo[4.4.0]decane acids, when present, were the latest eluting acids; we can fairly confidently rule out these acids in these samples of OSPW. Since we could find no sources of bicyclo[4.2.1]nonane, bicyclo[3.2.2]nonane, bicyclo[4.2.2]decane, bicyclo[3.3.2]decane or spiro[4.5]decane carboxylic acids to allow us to study the mass spectra or GC retention behaviour, and the retention indices of the alkanes appear not to have been published, we cannot rule these out as possibilities. Examination of the mass spectral features observed for the authentic reference compounds (Fig. 1A) were used to postulate

M.J. Wilde et al. / J. Chromatogr. A 1378 (2015) 74–87

83

Fig. 6. Extracted ion chromatogram (m/z 196, 182, 151, 137 and 123) of (A) a fraction of Merichem acid extract (#5) analysed by GC × GC–MS and mass spectra of (B, D and F) two unknown peaks (4a and c) identified and one C11 unknown peaks (4b) tentatively identified by comparison with the mass spectra and retention positions of (C) purchased bicyclo[3.3.1]nonane-1-carboxylic acid methyl ester, (E) 5-methylbicyclo[3.3.1]nonane-1-carboxylic acid methyl ester and (G) bicyclo[3.3.1]nonane-3-carboxylic acid methyl ester.

84

M.J. Wilde et al. / J. Chromatogr. A 1378 (2015) 74–87

Fig. 7. Electron ionisation mass spectra of (A, C and E) unknown peaks (5a–c) within an OSPW acid extract (#3), tentatively identified by comparison with the mass spectra of (B and F) two isomers of synthesised bicyclo[4.3.0]nonane-3-carboxylic acid methyl ester and (D) an unknown within a fraction of Merichem acid extract (#5) previously identified as an isomer of bicyclo[4.3.0]nonane carboxylic acid methyl ester [15] (most likely 2-isomer, similar to mass spectrum reported by Curcuruto et al. [37]).

structural features of the unknown acids. For example, methyl esters of acids in which the methylated carboxylic group is substituted onto the ring, creating a tertiary carbon atom, e.g. in the mass spectra [15] of the esters of bicyclo[4.4.0]decane-2- or 3-carboxylic acid (Figs. 1A (XIVa and b) and 8) or bicyclo[4.3.0]nonane-2 [15] or 3-carboxylic acids (Figs. 1A (Xa), 7 and S9), commonly lose a neutral methanol molecule, or methoxy radical (M-31/32) (though the spectra of stereoisomers vary; Fig. S9). In contrast, methyl esters of acids in which the methylated carboxylic group is substituted onto the ring via a longer alkanoate chain (e.g. in the mass spectra [15] of the esters of bicyclo[4.4.0]decane-2- or 3-ethanoic or propanoic acids (Fig. 1A; XIVc-e) or bicyclo[2.2.1]heptane-2-ethanoic acid (Figs. 1A; Ia and S3)), commonly lose a ·CH2 CO2 CH3 radical (mass 73 and mass 74 with occurrence of hydrogen transfer).

Methyl esters of acids in which the methylated carboxylic group is substituted onto the bridgehead carbon, creating a quaternary carbon atom (e.g. in the NIST mass spectrum of the esters of bicyclo[4.4.0]decane-4a-carboxylic acid or mass spectrum of bicyclo[3.3.1]nonane-1-carboxylic acid (Fig. 1A; VIIIa, and 6C)), commonly lose a methylated carboxy radical ·CO2 CH3 radical (mass 59 and sometimes mass 60 with the occurrence of hydrogen transfer). Abundant lower mass fragment ions such as m/z 55, 67, 79 and 81 present in many of the reference compound mass spectra are common ions observed in the mass spectra of cycloalkanes/polycycloalkanes, particularly those containing substituted cyclohexyl and cyclopentyl rings [55]. Therefore these ions were postulated to originate from fragmentations within the bicyclic core via more complex mechanisms and rearrangements i.e.

M.J. Wilde et al. / J. Chromatogr. A 1378 (2015) 74–87

85

Fig. 8. Electron ionisation mass spectra of (A, C, D and E) unknown peaks within an OSPW acid extract (#3), tentatively identified by comparison with the mass spectra of previously identified bicyclo[4.4.0]decane acid methyl esters [15] e.g. (B) bicyclo[4.4.0]decane-3-carboxylic acid methyl ester and (F) NIST library spectrum of bicyclo[4.4.0]decane-1-carboxylic acid methyl ester.

corresponding to C4 H7 + , C5 H7 + , C6 H7 + and C6 H9 + fragment ions respectively. Fragmentation within a bicyclic core requires fission of at least two bonds. Mass spectral studies of cycloalkanes, specifically bicyclic hydrocarbons, suggest that electron ionisation results in the fission of one of the bonds at a tertiary carbon bridgehead, followed by subsequent rearrangement and fragmentation [44,56,57]. The mass spectra of the C11 unknowns in the OSPW samples exhibited some of the above features. In general they were also characterised by abundant molecular ions (m/z 196) and ions at m/z 81, 95 and 107 were often predominant (Fig. S13). In some spectra, ions which may indicate losses of ethyl (M-29) and other alkyl (e.g. M-57, butyl) substituents, were present. To contain alkyl substituent groups of this size (e.g. C4 ), a C11 acid would require a bicyclic core to contain only six carbons (e.g.

C4 -bicyclo[2.2.0]hexane carboxylic acids). Spectra of the methyl esters of such acids are distinctive and do not match those observed here [36]. Thus, we conclude that the apparent C3 /C4 losses from the unknown represent losses from the rings, as observed in the spectrum of the methyl ester of authentic 4methylbicyclo[3.3.0]octane-2-carboxylic acid, which shows an ion due to loss of propene (Figs. 1A; VIb and S11). Although a number of structural features can be observed from the mass spectra of the methyl esters of the C12–16 acids, including molecular ions, ions due to losses of methanol (M-32) and to losses of alkyl groups or alkene moieties (e.g. M-15, M-28 and 29) from the molecular ion and ions due to losses of ethanoate (M-73) and propanoate (M-87) side chains, no more rigorous assignments of the structural types could be made than for the

86

M.J. Wilde et al. / J. Chromatogr. A 1378 (2015) 74–87

C11 acids. Thus we assume these are mostly higher homologues of the bicyclo[2.2.1]heptane, bicyclo[2.2.2], [3.2.1], [3.3.0]octane, bicyclo[3.3.1] and [4.3.0]nonane and some bicyclo[4.4.0]decane skeletal, with possibly bicyclo[4.2.1]nonane, bicyclo[3.2.2]nonane, bicyclo[3.3.2]decane or spiro[4.5]decane carboxylic acids represented also. We include the mass spectra of a few unknown bicyclic acid methyl esters (Fig. S13) in order that they may in future be compared with those of synthesised acids as these become available. 4. Conclusions Consideration of the GC retention behaviour, numbers of structural types and interpretation of the electron ionisation mass spectra of the methyl esters of a number of synthetic and purchased bicyclic carboxylic acids allowed identification of various bicyclic acids in OSPW and commercial acids. More than one hundred C8–15 bicyclic acids were typically present in OSPW. Synthesis or purchase allowed us to identify bicyclo[2.2.1]heptane, bicyclo[3.2.1]octane, bicyclo[4.3.0]octane and bicyclo[3.3.1]octane acids in OSPW and a bicyclo[2.2.2]octane acid in a commercial acid mixture. The retention positions of authentic bicyclo[3.3.0]octane and bicyclo[4.2.0]octane carboxylic acid methyl esters and published retention indices, showed these were also possibilities, as were bicyclo[3.1.1]heptane acids. In most OSPW acid extracts analysed the bicyclo[4.4.0]decane carboxylic (decalin) acids which have always been assumed to be present in OSPW, were relatively minor components. Bicyclo[5.3.0]decane and cyclopentylcyclopentane carboxylic acids were ruled out on the basis that the corresponding alkanes eluted well after bicyclo[4.4.0]decane (latest eluting acids). Bicyclo[4.2.1]nonane, bicyclo[3.2.2]nonane, bicyclo[3.3.2]decane, bicyclo[4.2.2]decane and spiro[4.5]decane carboxylic acids could not be ruled out or in, as no authentic compounds or literature data were available. Mass spectra of the methyl esters of the higher bicyclic C12–15 acids suggested that many were simply analogues of the above, with longer alkanoate chains and/or alkyl substituents. Our hypothesis is that these acids represent the biotransformation products of the initially somewhat more bio-resistant bicyclanes of petroleum. Remediation studies suggest at least some bicyclic acids can be relatively quickly removed from suitably treated OSPW [3], but a closer examination of which isomers are degraded will now be possible using the methods demonstrated here. This may be deemed important as some bicyclic acids are more acutely toxic than others [8]. Clearly many bicyclic acids remain to be identified. Since a wider literature of mass spectra of bicyclic hydrocarbons (e.g. [44,57–61]) is available than is extant for the acids, a useful approach may be to convert the acids to the hydrocarbons (cf [62,63]). Combining this older approach with the modern chromatography methods (viz.: GC × GC–MS) used here, may prove particularly valuable. Acknowledgments We are grateful for a Plymouth University PhD scholarship to DJ. Funding of this research was provided by an Advanced Investigators Grant (no. 228149) awarded to SJR for project OUTREACH, by the European Research Council, to whom we are also extremely grateful. We thank Dr C. Anthony Lewis for his contributions, particularly in delimiting the number of possible isomers as well as all his advice and input. We would like to acknowledge the EPSRC National Mass Spectrometry Service Centre at Swansea University, UK for obtaining the accurate mass data. We thank Professor O. Baudoin, Université Claude Bernard Lyon, France, for providing a sample of 1-methyl-1,2-dihydrocyclobutabenzene-1-carboxylic acid methyl ester.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chroma. 2014.12.008.

References [1] J.V. Headley, K.M. Peru, A. Janfada, B. Fahlman, C. Gu, S. Hassan, Characterization of oil sands acids in plant tissue using Orbitrap ultra-high resolution mass spectrometry with electrospray ionization, Rapid Commun. Mass Spectrom. 25 (2011) 459–462. [2] D.M. Grewer, R.F. Young, R.M. Whittal, P.M. Fedorak, Naphthenic acids and other acid-extractables in water samples from Alberta: what is being measured? Sci. Total Environ. 408 (2010) 5997–6010. [3] J.W. Martin, T. Barri, X. Han, P.M. Fedorak, M.G. El-Din, L. Perez, A.C. Scott, J.T. Jiang, Ozonation of oil sands process-affected water accelerates microbial bioremediation, Environ. Sci. Technol. 44 (2010) 8350–8356. [4] S. Mishra, V. Meda, A.K. Dalai, J.V. Headley, K.M. Peru, D.W. McMartin, Microwave treatment of naphthenic acids in water, J. Environ. Sci. Health A 45 (2010) 1240–1247. [5] J.V. Headley, K.M. Peru, D.W. McMartin, M. Winkler, Determination of dissolved naphthenic acids in natural waters by using negative-ion electrospray mass spectrometry, J. AOAC Int. 85 (2002) 182–187. [6] M.P. Barrow, J.V. Headley, K.M. Peru, P.J. Derrick, Fourier transform ion cyclotron resonance mass spectrometry of principal components in oilsands naphthenic acids, J. Chromatogr. A 1058 (2004) 51–59. [7] J.W. Martin, X. Han, K.M. Peru, J.V. Headley, Comparison of high- and lowresolution electrospray ionization mass spectrometry for the analysis of naphthenic acid mixtures in oil sands process water, Rapid Commun. Mass Spectrom. 22 (2008) 1919–1924. [8] D. Jones, A.G. Scarlett, C.E. West, S.J. Rowland, Toxicity of individual naphthenic acids to Vibrio fischeri, Environ. Sci. Technol. 45 (2011) 9776–9782. [9] T.D. Cyr, O.P. Strausz, Bound carboxylic acids in the Alberta oil sands, Org. Geochem. 7 (1984) 127–140. [10] A. Dimmler, T.D. Cyr, O.P. Strausz, Identification of bicyclic terpenoid hydrocarbons in the saturate fraction of Athabasca oil sand bitumen, Org. Geochem. 7 (1984) 231–238. [11] R.P. Philp, D.T. Gilbert, J. Friedrich, Bicyclic sesquiterpenoids and diterpenoids in Australian crude oils, Geochim. Cosmochim. Acta 45 (1981) 1173–1180. [12] R. Alexander, R.I. Kagi, R. Noble, J.K. Volkman, Identification of some bicyclic alkanes in petroleum, Org. Geochem. 6 (1984) 63–72. [13] D.T. Bowman, G.F. Slater, L.A. Warren, B.E. McCarry, Identification of individual thiophene-, indane-, tetralin-, cyclohexane-, and adamantane-type carboxylic acids in composite tailings pore water from Alberta oil sands, Rapid Commun. Mass Spectrom. 28 (2014) 2075–2083. [14] C.M. Aitken, D.M. Jones, S.R. Larter, Anaerobic hydrocarbon biodegradation in deep subsurface oil reservoirs, Nature 431 (2004) 291–294. [15] S.J. Rowland, C.E. West, A.G. Scarlett, D. Jones, Identification of individual acids in a commercial sample of naphthenic acids from petroleum by twodimensional comprehensive gas chromatography/mass spectrometry, Rapid Commun. Mass Spectrom. 25 (2011) 1741–1751. [16] C.E. West, J. Pureveen, A.G. Scarlett, S.K. Lengger, M.J. Wilde, F. Korndorffer, E.W. Tegelaar, S.J. Rowland, Can two-dimensional gas chromatography/mass spectrometric identification of bicyclic aromatic acids in petroleum fractions help to reveal further details of aromatic hydrocarbon biotransformation pathways? Rapid Commun. Mass Spectrom. 28 (2014) 1023–1032. [17] S.J. Rowland, C.E. West, D. Jones, A.G. Scarlett, R.A. Frank, L.M. Hewitt, Steroidal Aromatic, ‘Naphthenic acids’ in oil sands process-affected water: structural comparisons with environmental estrogens, Environ. Sci. Technol. 45 (2011) 9806–9815. [18] F.C. Damasceno, L.D. Gruber, A.M. Geller, M.C.V. de Campos, A.O. Gomes, R.C. Guimarães, V.F. Péres, R.A. Jacques, E.B. Caramão, Characterization of naphthenic acids using mass spectroscopy and chromatographic techniques: study of technical mixtures, Anal. Methods 6 (2014) 807–816. [19] G. Moss, Extension and revision of the von Baeyer system for naming polycyclic compounds (including bicyclic compounds), Pure Appl. Chem. 71 (1999) 513–529. [20] S.J. Rowland, R. Clough, C.E. West, A.G. Scarlett, D. Jones, S. Thompson, Synthesis and mass spectrometry of some tri-and tetracyclic naphthenic acids, Rapid Commun. Mass Spectrom. 25 (2011) 2573–2578. [21] T. Sasaki, S. Eguchi, T. Toru, Synthesis of adamantane derivatives. VIII. Novel substitution reaction of adamantanone. A simple synthesis of bicyclo[3.3.1]non-2-ene-7-carboxylic acid, J. Am. Chem. Soc. 91 (1969) 3390–3391. [22] J.A. Peters, J.M. Van Der Toorn, H. Van Bekkum, 3,7-disubstituted bicyclo[3.3.1]nonanes—III: synthesis and conformation of bicyclo[3.3.1] nonane-3␣,7␣-dicarboxylic acid, its dimethyl ester and some other 3,7disubstituted bicyclo[3.3.1]nonanes; adamantane as an integrated holding system, Tetrahedron 31 (1975) 2273–2281. [23] S.J. Rowland, A.G. Scarlett, D. Jones, C.E. West, R.A. Frank, Diamonds in the rough: identification of individual naphthenic acids in oil sands process water, Environ. Sci. Technol. 45 (2011) 3154–3159.

M.J. Wilde et al. / J. Chromatogr. A 1378 (2015) 74–87 [24] A.G. Scarlett, H.C. Reinardy, T.B. Henry, C.E. West, R.A. Frank, L.M. Hewitt, S.J. Rowland, Acute toxicity of aromatic and non-aromatic fractions of naphthenic acids extracted from oil sands process-affected water to larval zebrafish, Chemosphere 93 (2013) 415–420. [25] H.C. Reinardy, A.G. Scarlett, T.B. Henry, C.E. West, L.M. Hewitt, R.A. Frank, S.J. Rowland, Aromatic naphthenic acids in oil sands process-affected water, resolved by GC × GC–MS, only weakly induce the gene for vitellogenin production in zebrafish (Danio rerio) larvae, Environ. Sci. Technol. 47 (2013) 6614–6620. [26] D. Jones, C.E. West, A.G. Scarlett, R.A. Frank, S.J. Rowland, Isolation and estimation of the ‘aromatic’ naphthenic acid content of an oil sands process-affected water extract, J. Chromatogr. A 1247 (2012) 171–175. [27] Mass accuracy and theoretical m/z calculations, available at: http://homepages. warwick.ac.uk/staff/M.P.Barrow/massaccuracy.html (accessed 27/11/14). [28] C.E. West, A.G. Scarlett, J. Pureveen, E.W. Tegelaar, S.J. Rowland, Abundant naphthenic acids in oil sands process-affected water: studies by synthesis, derivatisation and two-dimensional gas chromatography/high-resolution mass spectrometry, Rapid Commun. Mass Spectrom. 27 (2013) 357–365. [29] F.M. Holowenko, M.D. MacKinnon, P.M. Fedorak, Naphthenic acids and surrogate naphthenic acids in methanogenic microcosms, Water Res. 35 (2001) 2595–2606. [30] J.S. Clemente, P.M. Fedorak, A review of the occurrence, analyses, toxicity, and biodegradation of naphthenic acids, Chemosphere 60 (2005) 585–600. [31] J.V. Headley, K.M. Peru, M.P. Barrow, Mass spectrometric characterization of naphthenic acids in environmental samples: a review, Mass Spectrom. Rev. 28 (2009) 121–134. [32] S.J. Rowland, C.E. West, A.G. Scarlett, D. Jones, R.A. Frank, Identification of individual tetra- and pentacyclic naphthenic acids in oil sands process water by comprehensive two-dimensional gas chromatography/mass spectrometry, Rapid Commun. Mass Spectrom. 25 (2011) 1198–1204. [33] B.E. Smith, C.A. Lewis, S.T. Belt, C. Whitby, S.J. Rowland, Effects of alkyl chain branching on the biotransformation of naphthenic acids, Environ. Sci. Technol. 42 (2008) 9323–9328. [34] S.J. Rowland, D. Jones, A.G. Scarlett, C.E. West, L.P. Hin, M. Boberek, A. Tonkin, B.E. Smith, C. Whitby, Synthesis and toxicity of some metabolites of the microbial degradation of synthetic naphthenic acids, Sci. Total Environ. 409 (2011) 2936–2941. [35] R.J. Johnson, B.E. Smith, P.A. Sutton, T.J. McGenity, S.J. Rowland, C. Whitby, Microbial biodegradation of aromatic alkanoic naphthenic acids is affected by the degree of alkyl side chain branching, ISME J. 5 (2011) 486–496. [36] D. Rush, A. Jaeschke, E.C. Hopmans, J.A.J. Geenevasen, S. Schouten, J.S.S. Damsté, Short chain ladderanes: oxic biodegradation products of anammox lipids, Geochim. Cosmochim. Acta 75 (2011) 1662–1671. [37] O. Curcuruto, D. Favrctto, P. Traldi, D. Ajo, C. Cativiela, J.A. Mayoral, M.P. Lorez, J.M. Fraile, J.I. Garcia, Mass spectrometry in stereochemical problems. 6—The case of mono and di-substituted norbornanes, Org. Mass Spectrom. 26 (1991) 977–984. [38] M. Heintz, M. Gruselle, A. Druilhe, D. Lefort, Relations entre structure chimique et données de rétention en chromatographie en phase gazeuse VI-Alcools et esters méthyliques de structures cycliques, Chromatographia 9 (1976) 367–372. [39] W.K. Seifert, R.M. Teeter, Preparative thin-layer chromatography and highresolution mass spectrometry of crude oil carboxylic acids, Anal. Chem. 41 (1969) 786–795. [40] S. Manabe, C. Nishino, Sex pheromonal activity of (+)-bornyl acetate and related compounds to the American cockroach, J. Chem. Ecol. 9 (1983) 433–448. [41] Y.-H. Kuo, C.-H. Chen, S.-L. Huang, A novel bicyclo [2. 2. 2] octane skeleton diterpene, obtunone, from the heartwood of Chamaecyparis obtusa var. formosana, Chem. Pharm. Bull. 46 (1998) 181–183.

87

[42] M. Asaoka, K. Ishibashi, N. Yanagida, H. Takei, Total synthesis of (±)eremolactone, Tetrahedron Lett. 24 (1983) 5127–5130. [43] A. Srikrishna, G. Ravi, A stereoselective total synthesis of (−)-seychellene, Tetrahedron 64 (2008) 2565–2571. [44] Y.V. Denisov, N.S. Vorobeva, I.M. Sokolova, A.A. Petrov, Mass spectrometric study of hydrocarbons of the bicyclo[2.2.2]octane series, Neftekhimiya 17 (1977) 186–191 (in Russian). [45] M. Presset, Y. Coquerel, J. Rodriguez, Syntheses, Applications of functionalized bicyclo[3.2.1]octanes: thirteen years of progress, Chem. Rev. 113 (2012) 525–595. [46] I.M. Sokolova, S.S. Berman, N.N. Abryutina, A.A. Petrov, Natural concentrates of bi- and tricyclic naphthenes, Chem. Technol. Fuels Oils 25 (1989) 233–235. [47] B.J. Mair, P.E. Eberly, K. Li, F.D. Rossini, Polycycloparaffin hydrocarbons in petroleum, Ind. Eng. Chem. 50 (1958) 115–116. [48] S.A. Selifonov, Microbial oxidation of adamantanone by Pseudomonas putida carrying the camphor catabolic plasmid, Biochem. Biophys. Res. Commun. 186 (1992) 1429–1436. [49] V.S. Ranade, G. Consiglio, R. Prins, Functional-group-directed diastereoselective hydrogenation of aromatic compounds. 21, J. Org. Chem. 65 (2000) 1132–1138. [50] C. Dass, Fundamentals of Contemporary Mass Spectrometry, Wiley, Hoboken, New Jersey, USA, 2007. [51] F.W. McLafferty, F. Tureˇcek, Interpretation of Mass Spectra, University Science Books, South Orange, New Jersey, USA, 1993. [52] J.S. Sinninghe Damsté, W.I.C. Rijpstra, J.A.J. Geenevasen, M. Strous, M.S.M. Jetten, Structural identification of ladderane and other membrane lipids of planctomycetes capable of anaerobic ammonium oxidation (anammox), FEBS J. 272 (2005) 4270–4283. [53] R. Alexander, R. Kagi, R. Noble, Identification of the bicyclic sesquiterpenes drimane and eudesmane in petroleum, Journal of the Chemical Society Chem. Commun. (1983) 226–228. [54] G.N. Gordadze, T.V. Okunova, M.V. Giruts, O.G. Erdnieva, V.N. Koshelev, Petroleum C15 polyalkyl substituted bicyclo[4.4.0]decanes (sesquiterpanes) as oil maturity indicators (illustrated by the example of Jurassic and Cretaceous oils of Kalmykia), Petrol. Chem. 51 (2011) 117–122. [55] F.W. McLafferty, R.F. Gould, Mass Spectral Correlations: Advances in Chemistry, Literary Licensing, LLC, Whitefish, Montana, USA, 2012. [56] Y.V. Denisov, I.M. Sokolova, A.A. Petrov, Mass spectrometric study of hydrocarbons of the bicyclo[4.3.0]octane series, Neftekhimiya 17 (1977) 491–497 (in Russian). [57] Y.V. Denisov, I.A. Matveyeva, I.M. Sokolova, A.A. Petrov, Mass-spectrometric study of hydrocarbons of bicyclo[3.2.1]octane series, Petrol. Chem. U.S.S.R. 17 (1977) 85–93. [58] Y.V. Denisov, N.S. Vorob’eva, A.A. Petrov, Mass Spectrometric study of hydrocarbons of the bicyclo[3.3.0]octane series, Neftekhimiya 17 (1977) 656–662 (in Russian). [59] Y.S. Brodskii, I.M. Lukashenko, I.A. Musayev, E.K. Kurashova, P.I. Sanin, Massspectra of dimethylbicyclo[4.4.0]decane stereoisomers (2,3-; 2,4-; 2,8- and 2,9dimethylbicyclo[4.4.0]decanes), Petrol. Chem. U.S.S.R. 17 (1977) 77–84. [60] I.M. Lukashenko, Y.S. Brodskii, I.A. Musayev, E.K. Kurashova, V.G. Lebedevskaya, P.I. Sanin, Mass spectra of methylbicyclo[4,4,0]decane stereoisomers, Petrol. Chem. U.S.S.R. 13 (1973) 38–44. [61] L.S. Golovkina, G.V. Rusinova, A.A. Petrov, Mass spectrometry of saturated hydrocarbons, Russ. Chem. Rev. 53 (1984) 870–887. [62] N. Zelinsky, Über die chemische Natur des Naphthensäuren (I.), Ber. Deutsch. Chem. Gesellsch. (A and B Ser.) 57 (1924) 42–51. [63] W.K. Seifert, R.M. Teeter, W.G. Howells, M.J.R. Cantow, Analysis of crude oil carboxylic acids after conversion to their corresponding hydrocarbons, Anal. Chem. 41 (1969) 1638–1647.