Analysis of naphthenic acids in aqueous solution using HPLC-MS/MS

Abstract

During the oil sand hot water extraction process, naphthenic acids are extracted from the bitumen into the oil sands process water, which leaves the water acutely and chronically toxic to aquatic biota. Because these acids can have toxic effects even at relatively low concentrations, depending on their composition and age, it is important to have analytical techniques sensitive enough to quantify the amounts of naphthenic acids in the aqueous environment. In this study, a new HPLC-MS/MS method was developed to separate and analyze mixtures of naphthenic acids. Using this method, inorganic salts are separated from the acids without any sample pre-treatment, such as an extraction step. Different naphthenic acids give characteristic retention times due to their structures, thus enabling measurements of the individual concentrations. Adsorption characteristics of certain naphthenic acids were explored using the new analytical method. It is expected that this analytical method could be employed to determine the fate and stability of individual naphthenic acids in the environment and thus provides insight to potential environmental impacts of oil sands processing.

---

Introduction

The oil sands of Alberta, Canada constitute the world's largest bitumen reserve, containing initial in-place reserves of 1.73 trillion barrels with an ultimate potential of 314.5 billion barrels.¹ During the processing of mined oil sand about three barrels of river water have to be imported in order to produce one barrel of oil. During this process organic acids (e.g., naphthenic acids) are leached from the bitumen to the water, which makes this water acutely and chronically toxic to aquatic organisms.^2,3 Under a zero-discharge policy, all process water and tailings must be stored on-site, which has resulted in an increasing number of tailings ponds.⁴

There has been a lot of research in the past few years on naphthenic acids (NAs) because of their toxicity. Naphthenic acids are a group of naturally occurring aliphatic and polycyclic carboxylic acids with a general formula of C_nH_2n+ZO₂, where n is the carbon number and Z is zero or a negative even number defining the hydrogen deficiency due to cyclization.⁵ In tailings ponds, concentrations of NAs usually range from 40 mg L⁻¹ to as high as 130 mg L⁻¹, whereas the concentration in the Athabasca River or regional lakes is typically less than 1 mg L⁻¹.^4,6 In addition, NAs with differing structures may have different degrees of toxicity and different fates in the environment.⁷ Therefore, analytical techniques that can quantitate individual naphthenic acids would aid in determining the potential environmental impacts of oil sand process water due to NAs.

Fourier transform infrared (FTIR) spectroscopy and gas chromatography (GC) are commonly used to detect NAs in tailings pond water due to the ease of adapting the existing methods used in oil-phase detection of NAs. However, it has been found that FTIR usually overestimates the concentration of NAs.^8,9 This overestimation may be due to the contribution of the carboxylic groups of naturally occurring non-naphthenic acids in the aqueous sample and which also have a carboxylic group and therefore contribute to the measured IR absorbance at 1743 and 1706 cm⁻¹.¹⁰ In the GC method, NAs are usually derivatized to form their methyl-, methyl-tert-butylsilyl-, or trimethylsilyl-esters, which are then separated by GC and detected by flame ionization detection or mass spectrometry (MS).¹¹ However, the esters are usually eluted from the GC column shown as an unresolved hump in the chromatogram and, therefore, the total area of the hump has to be integrated and then compared to the area of an internal standard to calculate the total NA concentration in an unknown sample.² Consequently, characterization of individual NAs is not possible.

Recently, soft ionization MS has been used very successfully in the direct characterization of NAs since the derivatization step can be eliminated. Negative ion mode atmospheric pressure chemical ionization (APCI) and electrospray ionization (ESI) have both been employed to analyze NAs. These ionization methods produce relatively little fragmentation of the deprotonated molecular ions.¹² It has been found that APCI-MS has a larger linear response range but a lower signal-to-noise ratio than ESI-MS.⁷ Since the concentration of NAs in the environment is relatively low, negative-ion ESI-MS has emerged as the method most widely used for analysis of NAs, although care is needed to minimize dimer formation at high ion concentrations.^13,14

Fragmentation of NAs using the tandem MS function may be possible. However, as it is often necessary to identify hundreds of NAs in tailings pond water by accurately measuring their individual mass-to-charge ratios, high-resolution and ultra-high-resolution MS are the most appropriate tools for this task, whereas unit-resolution MS is commonly employed to characterize NAs in aqueous samples, mainly due to the availability and ease of maintenance of the instrument.

It has been shown that unit-resolution MS can generate a trend for a given change in distribution of NAs similar to that produced by high or ultra-high mass resolution MS, although unit-resolution MS methods are prone to giving false-positive results when quantifying some NAs.¹¹ It was thought that combining on-line high-performance liquid chromatography (HPLC) with MS detection would minimize the occurrence of false positives associated with unit-resolution MS and shed more light on the structural differentiation of NAs based on their characteristic HPLC retention times.

In this study, an HPLC-MS/MS method was developed to separate and quantify a mixture of NA compounds. One advantage of this method is that sample pre-treatment is eliminated, thus avoiding any artifacts from sample preparation. In addition, each NA can be eluted at the characteristic retention time associated with its structure, thus enabling quantitative detection of certain NAs in the mixture. This newly developed method was used to investigate adsorption of NAs in an environment simulating a tailings pond. Factors that affect the adsorption of NAs are discussed.

Experimental

Chemicals and reagents

Six different NAs were purchased from Sigma Aldrich, Canada for use as model compounds in this study: 1-methyl-cyclohexane carboxylic acid, 4-tert-butylcyclohexane carboxylic acid, trans-4-pentylcyclohexane carboxylic acid, lauric acid, 2,2-dicyclohexylacetic acid, and 2-hexyldecanoic acid. The structures of these model compounds are given in Table 1. These six compounds were used to create a mixture of known molecular weights and structures. A commercial mixture of NAs ordered from Sigma Aldrich, Canada was also used in this study (Fluka brand, technical grade).

Table 1 Model NAs used in this study

1-Methyl-cyclohexane carboxylic acid (142.1 amu, n = 8, Z = −2)	trans-4-Pentylcyclohexane carboxylic acid (198.2 amu, n = 12, Z = −2)
4-tert-Butylcyclohexane carboxylic acid (184.1 amu, n = 11, Z = −2)	Lauric acid (200.2 amu, n = 12, Z = 0)
2,2-Dicyclohexylacetic acid (224.2 amu, n = 14, Z = −4)	2-Hexyldecanoic acid (256.2 amu, n = 16, Z = 0)

---

HPLC-grade ammonium acetate (NH₄Ac), water, and methanol were all obtained from Fisher Scientific and were used for the HPLC mobile phase. The mobile phase was filtered through a 0.45 µm filter before being used in any HPLC experiments. The two mixtures of NAs were each dissolved in methanol and directly injected into an HPLC column.

Well-crystallized kaolinite was obtained from Washington County, GA and bitumen was extracted from an Athabasca oil sand. To make up a mixture of 2.5 wt% bitumen on clay, a known amount of bitumen was fully dissolved in toluene (Fisher Scientific) and thoroughly mixed with a known amount of kaolinite. After mixing, the toluene was evaporated by a gentle air stream in a fume hood until the mixture was completely dry. The resulting bitumen–clay mixture was used in adsorption studies as a sorbent modeling clay from oil sand tailings.

HPLC-MS/MS

Chromatographic separation was conducted using an Agilent 1200 HPLC instrument. The mobile phase consisted of 10 mM NH₄Ac in HPLC water (A) and methanol (B), and was pumped using a binary pump at a flow rate of 700 µL min⁻¹. Injections were made by autosampler and volumes were typically 5 µL. Separation was performed on a reverse-phase column, Zorbax® Eclipse Plus C8 (4.6 × 150 mm, 3.5 µm, Agilent). A gradient elution was utilized: 30% B 0–6 min, ramped to 100% B 6–50 min, and held for 10 min before returning to initial conditions and being left to equilibrate for 10 min before injection of the next sample.

Detection was performed in the negative ion mode by either selective ion monitoring scan or enhanced MS scan mode (m/z 50–800) at a rate of 1000 Da s⁻¹ on a QTrap 3200 mass spectrometer equipped with an ESI interface (Applied Biosystems). Typical mass spectrometer conditions were as follows: source temperature 500 °C, curtain gas 1.5 L min⁻¹, nebulizer gas (GS1) 6 L min⁻¹, auxiliary gas (GS2) 6 L min⁻¹, interface heater (ihe) ON, collision-activated dissociation (CAD) gas high, ion spray (IS) voltage −4500 V, declustering potential (DP) −20 V, entrance potential (EP) −10 V, collision energy (CE) −10 V, and collar 2 barrier (C2B) 100 V. The electron multiplier was set to 2000 V. Compound optimization was performed using model NA compounds. The complete HPLC-MS/MS system was controlled by Analyst 1.5 software (Applied Biosystems).

Adsorption studies

The exchange of NAs between the aqueous phase and the solid phase (i.e., clay with adsorbed organics in this case) is a function of many variables including pH, ionic strength, clay type, and adsorbed organics on the clay surface.^3,15,16 In this study, the adsorption of 25 ppm trans-4-pentylcyclohexane carboxylic acid and 25 ppm lauric acid onto the sorbent (clay with 2.5 wt% adsorbed bitumen) was determined using the newly developed HPLC-MS/MS method. These two model compounds were chosen because they have similar molecular weights but quite different molecular structures. Adsorption studies on other model compounds and Athabasca oil sands NAs will be provided in a future publication. The acids were dissolved in phosphate buffers at pH 8.1. The ionic strength of buffer was controlled at 1 M using NaCl. The weight of the solid sorbent was 0.2000 g ± 0.0002 g. During the adsorption studies, 20 mL NA solution were added to the solid sorbent in a capped 20 mL glass vial. The vial was stirred vigorously for 4 h then left to sit overnight. Aliquots from the water phase after centrifugation were taken for analysis by HPLC-MS/MS. Control runs were made using 20 mL plain buffer with the same amount of solid sorbent (about 0.2 g) and 20 mL NA solution with no added sorbent.

Results and discussion

Method development on Eclipse Plus C8 column

In this study, model NA compounds representing a range of n and Z values were selected to develop a method that would optimize chromatographic separation and mass spectrometric response. Various reverse-phase HPLC columns, including Hydro-RP C18 (Phenomenex), Aquasil C18 (Thermo), Zorbax® Eclipse XDB C18 (Agilent), and Zorbax® Eclipse Plus C8 (Agilent), were tested using the model NA compounds.

Initial runs showed that Aquasil C18 and Zorbax® Eclipse Plus C8 produced the best separation and gave minimal peak tailing. However, in later experiments, it was found that cationic process chemicals, such as benzalkonium chloride (BAC),‡ were difficult to elute from the Aquasil C18 column but could be eluted from a Zorbax® Eclipse Plus C8 column within a reasonable time. This is possibly due to the strong interaction between positive ions formed from BAC and the hydroxyl (–OH) group in the stationary phase of the Aquasil column. Since oil sand tailings pond water may contain these cationic process chemicals, the Aquasil C18 column may only be suitable for separating NA compounds in model solutions, whereas the Eclipse Plus C8 column should be useful for real-world sample analysis. Consequently, the Eclipse Plus C8 column was used in this study to develop the HPLC method.

Oil sand process water can contain high concentrations of inorganic salts, which can overload the MS system with charged ions, block the electrospray probe and thus harm the MS instrument. There can also be hundreds of NAs in the water. A gradient elution method was developed to remove the problem of salts entering the MS as well as allow faster analysis of the NAs.

Methanol was used as the organic component of the mobile phase of the gradient elution. Although it was found that high concentrations of NAs (e.g., 500 ppm) could not be fully dissolved in less than 30% (v/v) methanol, the mobile phase initially has to be more polar in order to get rid of salts in the sample before any organic chemicals are eluted. Therefore, 30% methanol was chosen as the initial organic solvent content in the mobile phase. Additionally, in our experiments, it was found that NH₄Ac can deprotonate acids and thus minimize peak tailing. Therefore, the aqueous component of the mobile phase was composed of 10 mM NH₄Ac in water. That is to say, 70% 10 mM NH₄Ac in water (A) plus 30% methanol (B) (mixed by a binary pump on HPLC) was chosen as the initial composition of the mobile phase for the gradient elution of NAs from Eclipse Plus C8 column.

Using the above mobile phase composition, it was found that most inorganic salts were eluted from the HPLC column within the first 5 min and the retention times of model NAs on the column were well over 5 min. Consequently, sample cleanup was accomplished by means of a gradient elution method using the above mobile phase (70% A and 30% B) during at least the first 5 min of the run. In this way, inorganic salts were separated from the NAs.

In order to design the final method, a short chromatographic run (30% B 0–10 min, ramped to 70% B 10–10.01 min, ramped to 100% B 10.01–25 min, held for 7 min before returning to initial conditions, and held for 3 min; total 35 min) was tested on the Eclipse Plus C8 column. Under these conditions, most of the peaks representing NAs came out together as a hump, shown as the red curve in Fig. 1. Even when the information dependent acquisition (IDA) function of Analyst 1.5 (which can discriminate individual peaks for all the masses) was used, the retention times of different NAs were similar, especially at the higher-molecular-weight end (retention time differences were less than 0.5 min). Therefore, it is difficult to identify individual NAs in an unknown sample in a short chromatographic run using a unit-resolution mass spectrometer, even when gradient elution is employed.

Fig. 1 Red: commercial NAs (1000 ppm) eluted by a short chromatographic run. Blue: commercial NAs (1000 ppm) eluted by a long run. Both runs on the Eclipse Plus C8 column using gradient elution under selective ion monitoring mode at 20 °C.

Although the longer the gradient elution time, the better the peak resolution, if the retention time is too long, peak broadening may occur by the end of the chromatographic run, which reduces the separation efficiency. A method was needed with a run time sufficiently long to resolve peaks but not so long as to reduce separation efficiency. After a series of tests, a 1 h chromatographic run using gradient elution (30% B 0–6 min, ramped to 100% B 6–50 min, and held for 10 min) appeared to be the optimal choice. The separation based on this method is shown as the blue curve in Fig. 1. It can be seen that the peaks are more resolved than in the short run shown as the red curve in Fig. 1, even though there is still a broad hump. With the aid of MS as a detector, it was possible to find the retention times for most of the NAs.

In order to identify an individual NA in an NA mixture, one has to know its retention time. With the MS detector, the retention time of different NAs can be found based on their molecular weights. Although NA isomers will not be differentiated by the detector as they have the same molecular weight, differing NA isomer classes (for each n and Z combination) can be easily distinguished using the developed method. Knowing the general formula of NAs, C_nH_2n+ZO₂, one can calculate their molecular weights (see ESI†). As a result, based on the molecular weights and experimental results shown as the blue curve in Fig. 1 obtained using the selective ion monitoring mode and IDA, the retention times of different NA isomer classes eluted from Eclipse Plus C8 column were found and are shown in Fig. 2.

Fig. 2 Retention times of detected NAs in the commercial mixture as a function of n and −Z based on the experiment shown as the blue curve in Fig. 1. Different colours represent different n values.

The results show that the retention times of NAs decrease with increasing |Z| and increase with increasing n. This finding is consistent with previously reported results obtained using a high resolution MS detector.⁵ Although different isomers of NAs may have slightly different retention times, the difference in retention time among most isomers may be on the order of just a few minutes. Consequently, one can expect that, for a given n and Z, the retention time of an unknown NA will closely follow the curves shown in Fig. 2, but isomers may not be fully separated. Further improvement may be achieved using ultra-high-performance liquid chromatography (UHPLC).

The mixture of model NAs was also separated by the Eclipse Plus C8 column using the developed analytical method. As shown in Fig. 3, 1-methyl-cyclohexane carboxylic acid eluted at 20.01 min; 4-tert-butylcyclohexane carboxylic acid eluted at 32.72 min and 35.24 min (cis and trans); trans-4-pentylcyclohexane carboxylic acid eluted at 37.34 min; lauric acid eluted at 40.00 min; 2,2-dicyclohexylacetic acid eluted at 40.38 min; and 2-hexyldecanoic acid eluted at 46.15 min. The peaks are separated from each other, except those of lauric acid and 2,2-dicyclohexylacetic acid. However, since the detector is MS, they could easily be differentiated based on their molecular weight. The retention times of these model compounds fall closely onto the curves shown in Fig. 2, which demonstrates the reproducibility of the developed analytical method.

Fig. 3 NA model compounds (50 ppm each in methanol) eluted from Eclipse Plus C8 column, detected by negative-ion-enhanced MS scan at 20 °C. A comparison of the chromatograms of model NAs and commercial NAs is shown in ESI†.

Previously, Lo et al. have used low resolution MS to analyze and characterize model NAs and NAs from the tailings ponds of the Athabasca oil sands.⁷ However, there was no chromatographic separation before MS analysis. As a result, false positive errors are inevitable. Smith and Rowland synthesized amide derivatives of NAs and successfully used HPLC-MSⁿ to characterize the structures of the molecules.¹⁷ However, the derivatization procedure can be time-consuming and, more importantly, there is sometimes incomplete conversion of NAs to their derivative, which means that it is difficult to obtain quantitative data using this method. Headley et al. compared the MS profile of NAs among low-, high-, and ultra-high-resolution MS.¹⁸ Because of the false positive results and misclassification given by the low-resolution MS, they claimed that high- or ultra-high-resolution mass spectrometry would be suitable for NA analysis. But, in their work, there is no HPLC analysis before the low-resolution MS detection whereas it is the combination of retention time and the low-resolution MS that gives this method its usefulness, analogous to the previously mentioned method using HPLC-high resolution MS.⁵ However, the method reported here does not require the extensive sample preparation used for the HPLC-high resolution MS.

The current research work emphasizes the utility of combining low-resolution MS detection with HPLC. The advantages of the developed method are that there is no sample pre-treatment and it will reduce the false positive results of NA analysis given by low-resolution MS alone, by checking the retention time of the NAs of interest. That is to say, the retention time of the wide range of classes observed will add an important level of detail that is lost using low-resolution MS alone.

Calibration and validation of quantitative analysis

Standard solutions of NAs for the mixture of model compounds were made by dissolving 0, 1, 25, and 100 ppm of each NA in methanol. These four standard solutions were separately injected into the column and eluted utilizing the developed HPLC-MS/MS method. The chromatograms were similar to the one shown in Fig. 3 and the model compounds gave reproducible retention times between each HPLC run. The only differences were in the heights and widths of peaks due to concentration differences among different standard solutions. Calibration curves of these model compounds all showed good linearity. For example, correlation coefficients of the calibration curves of 4-tert-butylcyclohexane carboxylic acid, trans-4-pentylcyclohexane carboxylic acid, and lauric acid were 0.9998, 0.9988, and 0.9970, respectively (as shown in ESI†).

The Fluka NA mixture was dissolved in methanol to make a 500 ppm solution. 5 µL of this solution were injected into the column in order to detect whether it contained any of the model compounds. The chromatogram acquired using the selective ion monitoring mode is shown in Fig. 4, which is similar to the blue curve shown in Fig. 1 despite the intensity difference. In order to identify whether there are any NA model compounds in the commercial mixture, the extracted ion chromatograms of the molecular weights of the unknown compounds in the commercial mixture were compared with the extracted ion chromatograms of the molecular weights of the model compounds. It was found that in the commercial mixture there were possibly three of the NA model compounds: 4-tert-butylcyclohexane carboxylic acid, trans-4-pentylcyclohexane carboxylic acid, and lauric acid.

Fig. 4 Chromatogram of commercial NAs (500 ppm) eluted from Eclipse Plus C8 column at 20°C under selective ion monitoring mode.

To prove that these unknown compounds in the commercial mixture were the three model compounds, the chromatograms of the identified NA compounds found in the commercial mixture were extracted from Fig. 4 and are shown in Fig. 5–7. It can be seen that the retention times of these three compounds were similar to those of the model compounds, as shown in Fig. 3, but their peak shapes were slightly different. The difference may be due to the coexistence of isomers in the commercial NA mixture. As a result, using the developed analytical method, individual NA isomer classes could be easily detected but individual NAs will be indistinguishable from their isomers.

Fig. 5 Extracted ion chromatograms of a compound (deprotonated M_w of 183.1) in a commercial NA mixture eluted from the column at 33.42 min.

Fig. 6 Extracted ion chromatograms of a compound (deprotonated M_w of 197.2) in a commercial NA mixture eluted from the column at 37.04 min.

Fig. 7 Extracted ion chromatograms of a compound (deprotonated M_w of 199.2) in a commercial NA mixture eluted from the column at around 40 min.

Quantitation using the developed method was tested on the commercial NA mixture in order to validate the method. Commercial NAs dissolved in methanol, with total concentrations ranging from 0 to 500 ppm, were injected into the Eclipse Plus C8 column and detected using the developed HPLC-MS/MS method. To quantify individual NA isomer classes in this complex NA mixture, the selective ion monitoring mode and IDA were employed.

Certain unknown NAs found in the commercial mixture were randomly selected to show the linearity of their calibration curves. In these cases, the calibration curves of individual NA isomer classes in the commercial mixture were arbitrarily established based on the concentration of total commercial NAs, since the concentrations of these individual NAs in the commercial mixture were unknown. Considering that there are thousands of NAs in the mixture, the concentration of an individual NA will be much lower than that of the total NAs. As shown in Fig. 8, C₈H₁₆O₂ with a deprotonated molecular weight of 171.1 amu was eluted at 35.92 min, and C₁₅H₂₆O₂ with a deprotonated molecular weight of 237.2 amu was eluted at 41.90 min. The measured retention time corresponds well to the curves shown in Fig. 2, which verifies the reproducibility of the developed method and identifies the unknown molecules as NAs. Additionally, as shown in Fig. 8, the calibration curve for C₈H₁₆O₂ gives a correlation coefficient of 0.9891 and the calibration curve for C₁₅H₂₆O₂ gives a correlation coefficient of 0.9998. This shows that the developed HPLC-MS/MS method is quantitative.

Fig. 8 Extracted ion chromatograms from Fig. 4 at 171.1 amu (top) and 237.2 amu (middle); calibration curves for 171.1 amu (bottom left) and 237.2 amu (bottom right).

While a method is still needed to calculate absolute concentrations of individual molecules in a mixture, this method will be able to quantify changes in concentrations or compare concentrations between experimental systems containing those components in a mixture. In this paper, adsorption studies on NAs in an environment simulating an oil sand tailings pond were explored with the aid of the developed HPLC-MS/MS method.

Applications of the developed method in adsorption studies

It has been shown that the acutely toxic fraction of NAs degrades naturally over time.^19–22 However, the lengthy residence time required for degradation may not be practical for reclamation purposes.⁴ During this period, the fate of NAs in a tailings pond is mainly controlled by adsorption onto sorbents (such as tailings solids) and by other physical, chemical, and biological processes.^15,16 Adsorption is normally characterized by an adsorption isotherm, which represents, at constant temperature, the equilibrium distribution between the concentration of species on the sorbents and the concentration in the aqueous phase. The isotherm is related to the free energy and entropy of the adsorption reaction, as well as to both the surface morphologies and the adsorption properties of the sorbents.^16,23,24

Assuming that the rate of change in the concentration is mainly due to adsorption and that the flow rate is low enough to establish adsorption equilibrium, several equilibrium adsorption models could be applied in adsorption studies.^15,25 Generally, the adsorption of organic compounds at low concentrations onto a sorbent can also be described by a linear isotherm, a special case of Freundlich model (a nonlinear isotherm), as shown below:^26,27

C_s = K_dC

where C_s is the concentration of the solute on the sorbent (µg g⁻¹), C is the concentration of the solute in the aqueous phase (mg L⁻¹), and K_d is the equilibrium distribution coefficient or adsorption coefficient (mL g⁻¹).

In these adsorption studies, lauric acid and trans-4 pentylcyclohexane carboxylic acid were used as model compounds. Using the developed HPLC-MS/MS method, the concentrations of NAs in the solution were measured before and after the partitioning experiments and the concentrations of NAs on the sorbents (clay with 2.5 wt% adsorbed bitumen) were calculated based on the difference between the two measured concentration values. It was found that when NA concentrations were below 25 mg L⁻¹, their adsorptions onto the sorbents could be described by a linear isotherm (details to be given in a future publication). Our experiments showed that lauric acid has a larger adsorption coefficient (25 mL g⁻¹) than trans-4-pentylcyclohexane carboxylic acid (19 mL g⁻¹) at the same pH value. In other words, lauric acid is more likely to partition to the solid phase than trans-4-pentylcyclohexane carboxylic acid.

The fact that trans-4-pentylcyclohexane carboxylic acid has a cyclic ring while lauric acid is a straight-chain molecule may explain their different adsorption coefficients given the same pH value, with straight-chain NAs possibly adsorbing onto the sorbents more readily. This is consistent with the observation that more cyclic naphthenic acids are present in the water found in oil sand tailings ponds than straight-chain ones.²⁸ However, it could be that the greater abundance of cyclic naphthenic acids in the tailings water is due to their resistance to biodegradation by microorganisms.²¹ Consequently, the true cause of the NA composition in tailings water is still under debate. Meanwhile, it has been claimed that the acutely toxic fraction of NAs are straight-chain molecules.⁷ If this is the case, NAs in the water column may be less toxic than would be expected after fully partitioning with the sorbents in the tailings pond. However, further studies using other NAs with and without cyclic structures are needed in order to validate these claims.

Conclusions

This paper describes a useful HPLC-MS/MS method for analyzing naphthenic acids. It separates inorganic salts from the sample matrix directly by use of a single HPLC column, thus eliminating the extraction step of sample preparation. In addition, this new HPLC-MS/MS method can help differentiate NA molecules based on their unique structures and quantify their concentrations in aqueous samples. The method could therefore be used to measure the fate of NAs in the environment.

Acknowledgements

Funding for this work was provided by Natural Resources Canada through the Canadian Program of Energy Research and Development (PERD) and the ecoEnergy Technology Initiative (ecoETI). X.W. thanks Natural Resources Canada and the Natural Sciences and Engineering Research Council of Canada for the visiting fellowship.

References

1. A. Burrows, R. Marsh, C. Evans, M. Teare, and S. Ramos, Alberta's Energy Reserves 2008 and Supply/Demand Outlook 2009–2018, Report ST98-2009, Energy Resources Conservation Board, Government of Alberta, 2009, http://www.ercb.ca/docs/products/STs/st98_current.pdf, accessed 12 May, 2010.
2. J. S. Clemente and P. M. Fedorak, Chemosphere, 2005, 60, 585.
3. J. V. Headley and D. W. McMartin, J. Environ. Sci. Health, Part A: Toxic/Hazard. Subst. Environ. Eng., 2004, 39, 1989.
4. E. W. Allen, J. Environ. Eng. Sci., 2008, 7, 123.
5. M. Bataineh, A. C. Scott, P. M. Fedorak and J. W. Martin, Anal. Chem., 2006, 78, 8354.
6. F. M. Holowenko, M. D. MacKinnon and P. M. Fedorak, Water Res., 2002, 36, 2843.
7. C. C. Lo, B. G. Brownlee and N. J. Bunce, Water Res., 2006, 40, 655.
8. A. C. Scott, R. F. Young and P. M. Fedorak, Chemosphere, 2008, 73, 1258.
9. T. W. Yen, W. P. Marsh, M. D. MacKinnon and P. M. Fedorak, J. Chromatogr., A, 2004, 1033, 83.
10. F. M. Holowenko, M. D. MacKinnon and P. M. Fedorak, Water Res., 2001, 35, 2595.
11. J. V. Headley, K. M. Peru and M. P. Barrow, Mass Spectrom. Rev., 2009, 28, 121.
12. C. S. Hsu, G. J. Dechert, W. K. Robbins and E. K. Fukuda, Energy Fuels, 2000, 14, 217.
13. J. V. Headley, K. M. Peru, D. W. McMartin and M. Winkler, J. AOAC Int., 2002, 85, 182.
14. K. Qian, K. E. Edwards, G. J. Dechert, S. B. Jaffe, L. A. Green and W. N. Olmstead, Anal. Chem., 2008, 80, 849.
15. A. Janfada, J. V. Headley, K. M. Peru and S. L. Barbour, J. Environ. Sci. Health, Part A: Toxic/Hazard. Subst. Environ. Eng., 2006, 41, 985.
16. J. Peng, J. V. Headley and S. L. Barbour, Can. Geotech. J., 2002, 39, 1419.
17. B. E. Smith and S. J. Rowland, Rapid Commun. Mass Spectrom., 2008, 22, 3909.
18. J. V. Headley, K. M. Peru, S. A. Armstrong, X. M. Han, J. W. Martin, M. M. Mapolelo, D. F. Smith, R. P. Rogers and A. G. Marshall, Rapid Commun. Mass Spectrom., 2009, 23, 515.
19. A. C. Scott, M. D. MacKinnon and P. M. Fedorak, Environ. Sci. Technol., 2005, 39, 8388.
20. R. A. Frank, R. Kavanagh, B. K. Burnison, G. Arsenault, J. V. Headley, K. M. Peru, G. Van Der Kraak and K. R. Solomon, Chemosphere, 2008, 72, 1309.
21. X. M. Han, A. C. Scott, P. M. Fedorak, M. Bataineh and J. W. Martin, Environ. Sci. Technol., 2008, 42, 1290.
22. V. Nero, A. Farwell, L. E. J. Lee, T. Van Meer, M. D. MacKinnon and D. G. Dixon, Ecotoxicol. Environ. Saf., 2006, 65, 252.
23. E. K. Quagraine, H. G. Peterson and J. V. Headley, J. Environ. Sci. Health, Part A: Toxic/Hazard. Subst. Environ. Eng., 2005, 40, 685.
24. C. H. Giles. Interpretation and Use of Sorption Isotherms, in Sorption and Transport Processes in Soils. S.C.I. Monograph 37, Society of Chemical Industry, London, 1970, pp. 14–32.
25. C. C. Travis and E. L. Etnier, J. Environ. Qual., 1981, 10, 8.
26. R. M. Bayard, L. Barna, B. Mahjoub and R. Gourdon, Environ. Toxicol. Chem., 1998, 17, 2383.
27. I. Sabbah and M. Rebhun, Water Environ. Res., 1997, 69, 1032.
28. J. V. Headley, K. M. Peru, M. P. Barrow and P. J. Derrick, Anal. Chem., 2007, 79, 6222.

---

Footnotes

† Electronic supplementary information (ESI) available: The ESI includes a table of NA molecular weights, comparison of the chromatography between the model and commercial NAs, and calibration curves for some model NA compounds. See DOI: 10.1039/c0ay00204f

‡ An important representative of quaternary amine compounds (QACs), which were commonly used as corrosion inhibitors in oil sands operations.

---