Can the toxicity of naphthenic acids in oil sands process-affected water be mitigated by a green photocatalytic method?

The source of the OSPW and its chemical characterization used in this experiment were described previously by Gutierrez-Villagomez et al. (2019). Briefly, a 600-L sample of OSPW was collected from an active settling pond near Fort McMurray, Alberta, Canada. This was then acid-extracted in a similar manner to that reported by Rogers et al. (2002). In the laboratory of JVH, OSPW was left to settle and then acidified to pH 2.5 using sulfuric acid. A 1-L portion of OSPW was placed in a funnel, 500 mL of dichloromethane (DCM) was added, and the mixture was manually shaken for several minutes. The organic phase was then collected and concentrated with a rotary evaporation (at 40 °C). A 150-mL portion of 0.1 N NaOH was added to the acid extractable organic (AEO) fraction containing NAs. Sulfuric acid was then added to reduce the pH to 10 before the solution was filtered using 1000 MW cutoff membranes (Millipore). This volume of acid-extracted NAFC was then based-neutralized to pH 10 via 0.1 N NaOH, shipped to Queen’s University, and stored in the dark at 4 °C prior to further dilution for use in embryo bioassay exposures. Here, we used the term NAFC to broadly include the AEO fraction of OSPW containing all oil sands derived NAs. Though NAFC are made up predominantly of classical NAs (e.g., C_nH_2n + z O₂) (Brown and Ulrich 2015), this mixture also contains other complex structures including heteroatomic species, diamondoid compounds, resins, and sulfur- and nitrogen-containing compounds (Rogers et al. 2002; Hughes et al. 2017). The extracted NAFC stock (5300 mg/L, pH 10), was then nominally diluted in 10-L high-density polyethylene containers to a target concentration of 30 mg/L prior to photocatalytic treatment for solutions used in embryotoxicity bioassays. These solutions were adjusted to a pH of 8.2 ± 0.1 prior to photocatalytic treatment (described in Leshuk et al. 2018a, 2018b). The photocatalyst was added to NAFC solutions and placed under UVA fluorescence (350 nm, ∼30–40 W/m²) until the desired level of organic degradation was achieved for “low” and “high” intensity treatments. Approximately 25% of the initial concentration of NAFC were removed by low-intensity treatment and more than 80% by high-intensity treatment; the NAFC concentration of untreated solutions was 25.8 ± 2.3 mg/L (by Fourier-transformed infrared (FTIR)), which was reduced to 19.1 and 20.9 mg/L after low-intensity treatment, and to 2.5 and 3.9 mg/L after high-intensity treatment for laboratory and semi-natural experiments, respectively (Table 1).

The concentration of NAs present in OSPW–NAFC was assessed using two methods. The first used FTIR (PerkinElmer, Waltham, MA, USA) spectroscopy to assess the total level of carboxylic organics within solution, generally representing the estimated concentration of NAFC from the extracted OSPW sample. This was done according to standard industry methods (Jivraj et al. 1995), with modifications described in Leshuk et al. (2018a). Briefly, the acidified samples were extracted thrice with DCM in a 1:12.5 solvent-to-sample volumetric ratio, with >80% total recovery. A commercial NA mixture was then used to prepare a reference calibration curve (e.g., Sigma-Aldrich, Oakville, ON; No. 70340). Samples were filtered prior to the analysis (Whatman 934-AH glass fiber filter, Little Chalfont, Buckinghamshire, UK). FTIR analysis of exposure solutions was used to estimate the overall changes in total NA concentration before and after the photocatalytic treatment, as well as over the course of the experiments to monitor exposure concentrations during embryo bioassays. Water samples of the exposure stocks solutions for each treatment were taken: (i) following initial stock dilutions into 10-L HDPE containers prior to the advanced oxidation treatment, (ii) upon completion of photocatalytic treatment to the desired intensity and related concentration, and (iii) to monitor changes in the stock solution used to replenish bioassay exposure test solutions.

In a second subset a limited number of exposure solutions were analyzed by electrospray ionization–high-resolution mass spectrometry (ESI-HRMS) for the concentrations of specific NAs, including analysis of the O₂ components (i.e., classical NAs) from the bitumen-influenced source water. Briefly, 5 μL—samples were introduced into the mass spectrometer by loop injection (flow injection analysis) using a Surveyor MS pump (Thermo Fisher Scientific Inc., San Jose, California) and a mobile phase of 50:50 acetonitrile–water containing 0.1% NH₄OH. Mass spectrometry analysis was carried out using a dual pressure linear ion trap-orbitrap mass spectrometer (LTQ Orbitrap Elite, Thermo Fisher Scientific Inc) equipped with an ESI (HRMS) interface operated in negative ion mode; referred to here as ESI-HRMS. Data were acquired in full-scan mode from m/z 100 to 600 at a setting of 240 000 resolution (average mass resolving power (m/Δm 50%) was 242 000 at m/z 400). For scan-to-scan mass calibration correction n-butyl benzenesulfonamide was used as lock mass compound. Xcalibur version 2.2 software (Thermo Fisher Scientific Inc.) was used for data acquisition, instrument operation, and semiquantitative data analysis. Class distributions were established with acquired accurate mass data and Composer version 1.5.3 (Sierra Analytics, Inc., Modesto, California) with mass accuracies of <2 ppm. A previously characterized Athabasca oils sands OSPW large volume extract was used as a standard for the quantification of NA in the AEO samples (Rogers et al. 2002). While there is currently no standard method for the quantification of NAs in AEO samples from OSPW, so results presented here are considered semi-quantitative.

FHM eggs were reared from a minimum of 10 breeding pairs at the Aquatox Testing and Consulting Inc. (Guelph, Ontario, Canada). Eggs were transported in 23 ± 1 °C water within 24 h of fertilization and upon arrival at Queen’s University (Kingston, Ontario, Canada) were immediately placed in replicate 240-mL glass jars containing 200 mL of treatment solutions: negative control (water only), untreated NAFC (∼30 mg/L), low-intensity treated NAFC (∼20 mg/L), and high-intensity treated NAFC (<5 mg/L). Dechlorinated tap water from Lake Ontario (pH 7.9 ± 0.4, Table S1) was used for both experiments. Animal procurement and experimental procedures were carried out in accord with the approved consent of the Queen’s University Animal Care Committee (Orihel 2018-1829).

An embryo bioassay experiment was conducted on 10–16 July 2018 under controlled laboratory conditions based on the Organization for Economic Cooperation and Development Guideline 210 (OECD 2013). Four replicate jars, each with a starting minimum of 45 embryos, were used for each treatment. Exposure jars underwent daily static renewal with mortality and hatching success recorded daily. Replicate jars were aerated and maintained above 7 mg/L dissolved oxygen (Table S1) via an aeration pump (Hailea HAP60, Guangdong, China), at an individual tank flow rate of 5 mL/min. Conditions of the exposure were maintained at 23 ± 2 °C, with a 16L:8D photoperiod (Table S2). As viability of embryos were indistinguishable from initial rapid mortality, each replicate was inspected 48-h postfertilization for unfertilized embryos which were then removed from the treatments. Once embryos hatched, they were inspected using a fluorescence microscope (Leica DMLB, Germany, 10×/22 mm FOV eyepiece, 10× objective, 12 V 100 W Halogen light source) for signs of developmental malformations (e.g., pericardial edemas, tube heart, cranial or spinal deformities). All hatched fish were photographed and then euthanized in buffered of 0.1 g/L tricaine methanesulfonate solution (Acros Organics, Thermo Fisher Scientific, New Jersey, USA).

A second embryo bioassay experiment following similar parameters was conducted outdoors, under semi-natural conditions at the Queen’s University Biological Station (44.5675°N, 76.3245°W) from 13 to 21 August 2018. The experimental was set-up in an open-sided tent subject to natural light fluctuations and ambient weather conditions. Water temperature and light intensity were measured continuously using HOBO data-loggers (Onset, Bourn, Maine, USA; Table S2). A subset of the hatched embryos (n = 162) from this embryotoxic bioassay were raised for an additional 7 d post hatch in clean water (Norberg and Mount 1985). Larvae were fed brine shrimp (Artemia franciscana) twice daily and were assessed for physical indices, mortality, and prevalence of malformations.

Passive photocatalytic treatment using TiO₂ microparticles reduced the acute toxicity of NAFC extracted from OSPW to developing FHM embryos (Fig. 1A). In untreated NAFC solutions, we observed a three-fold (i.e., ∼74% and 83% for laboratory and semi-natural experiment, respectively) increase in acute mortality relative to negative controls. In contrast, high-intensity photocatalysis of NAFC solutions resulted in almost complete elimination of NAFC-induced acute mortality. Low-intensity treatments approximately halved the acute mortality in developing FHM observed relative to untreated NAFC solutions, but notably, mortality in treated fish was overall higher than that observed in earlier experiments that assessed 48-h LC50 (8–21 mg/L) and EC50 (∼22–24 mg/L) for FHM with similar concentrations of NAFC (Marentette et al. 2015a, 2015b). However, differences in NAFC quantification methods between these studies may account for the observed differences in FHM mortality. Mortality within respective treatments was comparable between laboratory and semi-natural experiments, bolstering confidence in the reproducibility of these findings outside of controlled settings.

The immediate toxicity observed with NAFC exposure can likely be attributed to nonspecific membrane disruption (i.e., acute narcosis), which arises from the surfactant properties of NAs (Marentette et al. 2015b; Morandi et al. 2015; Yue et al. 2015). Here, the toxicity of NAFC appeared to be related to the relative proportions of lower numbered oxygen class (O_x), including classical (O₂) NAs and O₃ to O₄ compounds (Fig. 2A). These have long been characterized as dominant components of fresh OSPW (Han et al. 2008; Headley et al. 2011, 2013; Leshuk et al. 2018b), and toxicity in general decreases with increased oxygen content (Frank et al. 2008, 2009). This coincides with our finding that acute mortality of FHM embryos decreased with increasing photocatalytic treatment intensity and increased abundance of O₄, O₅, and O₆ compounds, possibly related to the degradation of higher DBE components (de Oliveira Livera et al. 2018). Leshuk et al. (2018b) also reported a clear relationship between increased photocatalytic treatment intensity and increased oxygen content with high-intensity treatment of raw OSPW increasing average oxygen numbers from 2.5 to 4.3. We also observed that photocatalytic treatment of NAFC extracts from OSPW greatly reduced the proportions of similar sized O_x classes, growing the proportion of oxygen numbers above 3.5 (Fig. 2B) and reducing the proportion of O₂ compounds with 13 to 18 carbons (Fig. 2C). More specific detail of the chemical changes to particular NA classes resultant of this TiO₂-mediated photocatalytic treatment can be found in related work by de Oliveira Livera et al. (2018).

The relationship between the intensity of photocatalytic treatment of NAFC solutions and chronic toxicity experienced by developing fish is more nuanced than its outright lethality. Though the photocatalytic treatment intensity was proportional to hatch success, almost all treatments, irrespective of intensity, had more cardiovascular abnormalities than negative controls at the time of hatch (Fig. 1B). The cardiovascular abnormalities we observed in FHM exposed to NAFC were predominantly (∼95%) pericardial edema with tube heart. Rearing environment appeared to influence the toxic responses of fish in high-intensity treatments, although the final concentration of NAFC following treatment was not identical between laboratory and field experiments, i.e., 2.5 and 3.9 mg/L, respectively. This effect is, perhaps not unexpected, as NAFC measured in our solutions following high-intensity treatment were comparable to previous estimates of LOEC (∼4 mg/L) for similar NAFC extracts (Marentette et al. 2015a, McQueen et al. 2016). These findings highlight the need for assessments of treatment methods for oil sands waste waters to consider both acute and chronic toxicity endpoints, as well as the test environment when considering safe release threshold concentrations for OSPW into surrounding receiving waters.

The proportion of FHM hatched with cardiovascular abnormalities that were exposed to NAFC solutions in our study are consistent with previous studies examining the toxicity of OSPW (He et al. 2012) and NAFC (Marentette et al. 2015a, 2015b). General cardiovascular abnormalities may be attributed to DNA damage caused by oxidative stress. Oxidative stress and the activation of the aryl hydrocarbon receptor (AhR) contribute to cellular damage in fish embryos, including cardiovascular abnormalities resulting from DNA damage (Toomey et al. 2001; Barron et al. 2004; Marentette et al. 2017). OSPW contains other petroleum-related compounds, such as PACs, and that could contribute to toxicity of oil sands extracts (Headley et al. 2013). Recent analyses of diluted bitumen toxicity to fish links PACs to oxidative stress initiated through action on central AhR ligands and downstream biotransformation and detoxification processes (Madison et al. 2015, 2017; Alsaadi et al. 2018). Though information on the PACs derived from OSPW is limited, levels of PACs present in OSPW have been shown to cause carcinogenic, mutagenic, immunotoxic, and endocrine disruptive effects in a variety of organisms (Li et al. 2017). For example, Japanese medaka (Oryzias latipes) exposed to OSPW water fractions had an inhibited ATP-binding cassette (ABC transport proteins) (Alharbi et al. 2016). This inhibition of membrane protein activity could impede natural biotransformation and detoxification of PACs, thereby increasing severity of the toxic effects of OSPW on developing fish from NAFC alone.

Our finding that the prevalence of cardiovascular abnormalities was reduced with increased photocatalytic treatment is significant considering that Leshuk et al. (2018b) observed that O₂⁻, OS⁺, and NO⁺ ions were rapidly transformed in early stages of photocatalytic treatment; further details of this treatment process on the chemical properties of NAs have been described previously (de Oliveira Livera et al. 2018). As such, degradation of NAFC in OSPW may lower chronic toxicity by transforming the smaller toxic O₂ species, the predominant class of “classical” NAs oxygenated compounds in OSPW (Barrow et al. 2014). Our data show that photocatalytic degradation of OSPW-derived NAs reduced the relative proportions of these lower molecular weight O₂ classes, but also decreased the proportion of higher molecular weight O₄–O₆ compounds while also increasing the diversity of these compounds (Fig. 2). Marentette et al. (2017) found that walleye embryos exposed to NAFC caused down-regulation of antioxidant defenses, which may also exacerbate the effects of oxidative stress. These O₂ classes appear to be a driver of the cardiovascular abnormalities as they may inhibit membrane transport proteins and damage DNA through oxidative stress.

Incomplete degradation of NAFC extracts by high-intensity treatment, e.g., >3 mg/L, elevated the incidence of cardiac abnormalities relative to untreated NAFC in the semi-natural experiment. This may be due to potential oxidized intermediates produced during the photocatalytic treatment, which could be more toxic than the result (Leshuk et al. 2018b). This threshold is important when considering the practical application of buoyant TiO₂ microparticles to the tailings ponds in the oil sands region. Not only was more than 90% degradation via high-intensity treatment needed to reduce chronic toxicity to developing FHM in the laboratory, but incomplete degradation in lower treatment intensities may have the potential to enhance toxicity through these intermediates. Degradation of these specific compounds using similar photocatalytic conditions to those used in this experiment may best serve as a viable approach in tandem with more common biodegradative techniques. In related work, de Oliveira Livera et al. (2018) demonstrated significant reduction in larger DBE compounds linked to acute narcosis in biological organisms using TiO₂-mediated photocatalysis of commercial mixtures used as analogs for OSPW-sourced NAs. However, its target effectiveness appears to be related more to the ability to degrade toxic intermediates made up of larger DBE compounds into smaller O_x species. Here, the reduction of extracted NAFC toxicity indicated by the reduction of NA-O₂ concentrations under 3 mg/L using this method of photocatalysis appears necessary to meet intended targets for both acute and chronic toxicity. However, should any additional toxic intermediates remain in these fractions they may require additional biotransformation for full remediation of those compounds with suspected longer-term effects on aquatic organisms. For example, earlier studies have also demonstrated the longer-term impact of OSPW toxicity on endocrine function in FHM at similar and environmentally relevant concentrations of oilsands NAs from OSPW (Kavanagh et al. 2012, 2013). Whether these types of effects are alleviated by the photocatalytic treatment remains unexamined at this time.

Dissimilarity in larval mortalities and malformations at hatch were also observed between treatment groups (Supplementary Fig. S1). However, when these fish were moved to clean water following the exposure, no difference in growth was noted 7 d after the exposure to NAFC treatment solutions. Post-hatch mortality in embryos exposed to low-intensity treated NAFC may be related to the cardiac abnormalities observed with aforementioned change in proportions of oxygen-containing compounds in NAFC extracts following oxidative treatment. The cardiac abnormalities in larval FHM may impair available energy for important developmental processes or osmoregulatory abilities, resulting in higher post-hatch mortality. This finding is consistent with a previous study that found zebrafish with pericardial edemas caused by embryonic exposure to crude oil generally failed to feed as larvae and died before the onset of free feeding (Hicken et al. 2011). Though embryos were able to survive to hatch in the embryotoxicity bioassay, exposure to NAFC extracts may have lasting effects on FHM performance and survival later in life. These findings together emphasize the need to further explore the effectiveness of photocatalysis to treat both immediate and long-term toxic effects stemming from NAs and related compounds extracted from oil sands effluents.

The photocatalytic treatment of NAFC extracts was by and large successful at reducing the overall toxicity of OSPW-derived NAs to developing fish, but our study demonstrates the harmful consequences of incomplete degradation of NAs to aquatic vertebrates. The power of “green” (i.e., sustainable, recyclable, and reusable) buoyant microparticles over TiO₂, along with an unlimited solar energy source, makes implementing this type of passive degradation method an attractive technology for large-scale OSPW treatment in the oil sands industry. Future work in this area would benefit from more specific chemical characterization of NAFC from OSPW (e.g., atmospheric-pressure photo-ionization-HRMS, liquid chromatography or gas chromatography, or tandem MS) to streamline advanced oxidative processes and target catalytic degradation treatments to desired NAs that drive toxicity in aquatic animals. More broadly, we recommend that careful consideration should be paid not only to the feasibility and energy demands of any advanced oxidation treatments (McQueen et al. 2016), but as we demonstrate here, rigorous testing and monitoring must be performed to prevent the unintended ecological consequences caused by incomplete degradation of NAs and other fraction components in discharged waters.