Prediction of naphthenic acid species degradation by kinetic and surrogate models during the ozonation of oil sands process-affected water

Science of the Total Environment 493 (2014) 282–290

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Prediction of naphthenic acid species degradation by kinetic and surrogate models during the ozonation of oil sands process-affected water Md. Shahinoor Islam, Jesús Moreira, Pamela Chelme-Ayala, Mohamed Gamal El-Din ⁎ Department of Civil and Environmental Engineering, University of Alberta, Edmonton, Alberta T6G 2W2, Canada,

H I G H L I G H T S • • • • •

A kinetic model was developed to predict the removal of NAs during ozonation. Surrogate parameters were used to determine the degradation of NAs during ozonation. Speciation and distribution of classical and oxidized NAs were examined. The structure–reactivity of individual NA species was obtained. High correlations between the AEF and COD of OSPW and NA species were found.

a r t i c l e

i n f o

Article history: Received 28 February 2014 Received in revised form 20 May 2014 Accepted 26 May 2014 Available online xxxx Editor: Simon Pollard Keywords: Kinetic modeling Naphthenic acids species Ozonation Speciation Correlations Structure–reactivity

a b s t r a c t Oil sands process-affected water (OSPW) is a complex mixture of organic and inorganic contaminants, and suspended solids, generated by the oil sands industry during the bitumen extraction process. OSPW contains a large number of structurally diverse organic compounds, and due to variability of the water quality of different OSPW matrices, there is a need to select a group of easily measured surrogate parameters for monitoring and treatment process control. In this study, kinetic and surrogate correlation models were developed to predict the degradation of naphthenic acids (NAs) species during the ozonation of OSPW. Additionally, the speciation and distribution of classical and oxidized NA species in raw and ozonated OSPW were also examined. The structure–reactivity of NA species indicated that the reactivity of individual NA species increased as the carbon and hydrogen deficiency numbers increased. The kinetic parameters obtained in this study allowed calculating the evolution of the concentrations of the acid-extractable fraction (AEF), chemical oxygen demand (COD), and NA distributions for a given ozonation process. High correlations between the AEF and COD and NA species were found, suggesting that AEF and COD can be used as surrogate parameters to predict the degradation of NAs during the ozonation of OSPW. © 2014 Elsevier B.V. All rights reserved.

1. Introduction The bitumen extraction process of the oil sands in Northern Alberta, Canada, leads to the generation of oil sands process-affected water (OSPW), which is a complex mixture of organic compounds, inorganic compounds, and suspended solids (Fedorak et al., 2002; Holowenko et al., 2002). OSPW has been shown to be toxic toward aquatic organisms (Anderson et al., 2012; Gagné et al., 2012; He et al., 2012; Kavanagh et al., 2011), therefore, it needs to be extensively treated before being released into the receiving environments (Schramm et al., 2000). Of particular environmental concerns are the complex mixture of naphthenic acids (NAs) found in OSPW since they are both highly ⁎ Corresponding author at: 3-093 Markin/CNRL Natural Resources Engineering Facility, Department of Civil and Environmental Engineering, University of Alberta, Edmonton, Alberta T6G 2W2, Canada. Tel.:+1 780 492 5124; fax: +1 780 492 0249. E-mail address: [email protected] (M. Gamal El-Din).

http://dx.doi.org/10.1016/j.scitotenv.2014.05.138 0048-9697/© 2014 Elsevier B.V. All rights reserved.

recalcitrant and toxic to fish (He et al., 2012), bacteria (Gamal El-Din et al., 2011), and benthic invertebrates (Anderson et al., 2012). The NAs are described as a mixture of alicyclic and alkyl-substituted aliphatic carboxylic acids with the general formula of CnH2n + ZOx, where n is the carbon number, Z is the hydrogen deficiency due to ring or double bond formation, and x represents the number of oxygen atoms. When x = 2, these acids are referred as classical NAs, while they are referred as oxidized NAs (oxy-NAs) when x ≥ 3. In this study, the oxy-NAs with x = 3, x = 4, x = 5, and x = 6 are designated as O-NAs, O2-NAs, O3-NAs, and O4-NAs, respectively. Advanced Oxidation Processes (AOPs), particularly ozonation of OSPW at a laboratory scale, have been considered previously as OSPW remediation treatment processes (Gamal El-Din et al., 2011; Scott et al., 2008; Wang et al., 2013) and the impact of ozonation on the speciation of classical NAs has been demonstrated (Hwang et al., 2013; Wang et al., 2013). However, reports on the structure–reactivity of classical NAs in OSPW have been limited (Han et al., 2008; Perez-Estrada et al., 2011). To our knowledge,

M.S. Islam et al. / Science of the Total Environment 493 (2014) 282–290

the distribution of the oxy-NAs after ozone treatment considered in the present study has not been previously examined. The use of kinetic modeling of structure–reactivity relationships may be useful in broadening the understanding of organic compound degradation (or lack thereof) following the ozonation of OSPW. Kinetic modeling of this relationship for binary mixtures of model NA compounds has been considered previously by Perez-Estrada et al. (2011), where OSPW and commercial NAs were assumed to be of first- and second-order reactions, respectively. Additionally, it was found that model NA compounds with multiple rings and alkyl branching were preferably degraded during the ozonation process (Perez-Estrada et al., 2011). It has been reported that classical NAs in OSPW are more recalcitrant than commercial NAs during biodegradation (Han et al., 2008). Moreover, it has been reported that n had little effect on the biodegradation rate, whereas a general structure–persistence relationship was observed with the Z number, which indicates that an increase in cyclization (i.e., increase in Z) resulted in a decrease in the biodegradation rate of both commercial and OSPW classical NAs. The OSPW contains thousands of unidentified organic compounds (in addition to NAs) in a highly complex OSPW mixture; many of these compounds do not have established analytical methods for their identification and quantification (Grewer et al., 2010; Rowland et al., 2011). Given this inability to directly identify/quantify these organic compounds, the use of simply measured ‘surrogate’ parameters (such as chemical oxygen demand) may be considered to determine the degradation of these compounds in OSPW, following a treatment process without the need for their direct measurement. Although surrogate parameters have been investigated extensively for municipal wastewaters, the use of surrogates for OSPW has not been extensively considered. Mohamed et al. (2008) used UV absorption and fluorescent emission spectroscopy as surrogate parameters to determine chromophoric surrogate compounds that serve as an internal standard for the indirect analysis of NAs in OSPW. Surrogate parameters used for municipal wastewaters have included total organic carbon (TOC), chemical oxygen demand (COD), UV absorbance at 254 nm (UV254), and UV fluorescence (Chang et al., 1998; Gerrity et al., 2012; Lee and Ahn, 2004; Lee et al., 2013). The selection of specific surrogate parameters depends on the water treatment step being monitored. As illustration, COD has been commonly used in kinetic studies during the ozonation of wastewater because it provides insight of the magnitude of the oxidation process (Beltrán, 2004). TOC has been used as “true” parameter for the determination of organic pollution (Bourgeois et al., 2001). UV254 has been found to be an accurate precursor surrogate parameter to predict the trihalomethane formation in raw and alum-coagulated waters (Pifer and Fairey, 2014). In contrast to the typical surrogate parameters used for municipal wastewater, a more suitable surrogate typically measured during OSPW remediation is the acid-extractable fraction (AEF) which is commonly employed to monitor treatment effectiveness (Gamal El-Din et al., 2011; Zubot et al., 2012). The benefit of using the AEF measurements is that they determine the total organic acid fraction concentrations, which includes all NAs species. Currently, surrogate parameter(s) used to predict the oxidation of the organic fraction in OSPW have not been fully developed despite the usefulness of these easily measured parameters for onsite assessment of treatment efficiency via non-specialized, user-friendly instrumentation. The main objective of this study was to develop, for the first time, kinetic and surrogate correlation models to predict the degradation of NA species during the ozonation of OSPW. The surrogate parameters investigated in the present study included COD, TOC, UV254, and fluorescence, in addition to a typical OSPW remediation AEF parameter. Correlation analysis was evaluated using the Pearson product–moment correlation coefficient with statistically significant correlations determined based on the p-value and Student's t-test on slope and intercept during the linear regression. In addition to the main objective, this study also examined, for the first time, the effects of ozonation on the speciation of

283

oxy-NAs, while the structure–reactivity of both classical and oxy-NAs species were estimated using a pseudo-first-order reaction model. 2. Materials and methods 2.1. Chemicals and reagents The following chemicals were used, as received, without any further treatment: Fluka NA mixture (Sigma-Aldrich, ON, Canada); lucine enkaphenlin and sodium formate standard solutions (Waters Corporation, Milford, MA, USA) used as standards for NA analysis; as well as potassium hydrogen phthalate (Fisher scientific, Fair Lawn, New Jersey, USA) as a COD standard. Phenol bioextra (N99.5% Sigma, Aldrich, CA) was used in Microtox analysis. The OSPW was provided by Syncrude Canada Ltd., and collected from West In-pit pond, Fort McMurray, Alberta, Canada, on September 27th 2010. OSPW samples were stored at 4 °C in sealed polyethylene containers and used without additional pre-treatment prior to the ozonation experiments. 2.2. Ozonation experiments The OSPW was ozonated following the procedure described elsewhere (Wang et al., 2013). Ozone gas (O3) was produced by an ozone generator (AGSO 30 Effizon, WEDECO AG Water Technology, Herford) from extra dry high purity oxygen. Ozonation experiments were performed in a 4-L vacuum flask reactor equipped with a cylindrical shape rock fine bubble gas diffuser. Throughout the experiment, the ozone concentrations in feed and off-gas lines were continuously monitored by two identical ozone monitors (HC-500, PCI-WEDECO). The experiments were performed at natural OSPW pH (8.4 ± 0.1) and at room temperature (20 ± 1 °C). For sample analysis, triplicate samples (n = 3) were taken periodically over the 20 min experiment duration. 2.3. Sample analyses The AEF was measured using Fourier Transform Infrared Spectroscopy (FT-IR) as described elsewhere (Gamal El-Din et al., 2011). A calibration curve for the quantification of AEF was done using a Fluka NA mixture. The concentrations of classical and oxy-NAs were measured using a high performance liquid chromatography-high resolution mass spectrometry (UPLC-HRMS) Waters Acquity UPLC® System, Milford, MA, USA equipment, according to a method described elsewhere (Wang et al., 2013) and included in the Supplementary data. Ion-mobility spectrometry (IMS) was conducted in a Tri-Wave® ion-mobility cell of 15 cm long, using nitrogen (purity N 99%) as the drift gas. The IMS consisted of a transfer cell that collected a certain amount of ions and a helium gate that released the ions into the ion-mobility cell. The COD was measured according to standard methods (American Public Health Association, 2005). An Apollo 9000 TOC Combustion Analyzer (FOLIO Instruments Inc.) was used to measure the TOC levels. A Varian Cary Eclipse fluorescence spectrophotometer (Ontario, Canada) was used to measure the fluorescence using a 1-cm cuvette with excitation–emission spectra collected over a range of excitation wavelengths from 220 nm to 400 nm and emission wavelengths from 260 nm to 500 nm. The excitation and emission slits were maintained at 10 and 5 nm, respectively, with a scanning speed of 600 nm min−1. A medium photomultiplier voltage (600 mV) was applied for the fluorescence signal. Both raw and ozonated OSPW samples were filtered through a 0.45 micron nylon filter (Whatman) before performing the fluorescence spectroscopy measurements. A Microtox analyzer (AZUR Environmental, Carlsbad, USA) was used to determine the toxicity of the untreated and ozonated OSPW samples (Chelme-Ayala et al., 2011). The results are reported for an incubation time of 5 min by using the 81.95% screening test protocol. The percentage of inhibition caused by the untreated and ozonated samples after

284

M.S. Islam et al. / Science of the Total Environment 493 (2014) 282–290

incubation was calculated from the change in luminescence intensity. Phenol was used as quality control to verify the sensitivity of the luminescent bacteria prior to the analyses of OSPW. 3. Results and discussions 3.1. Speciation and distribution of NA species after ozonation As the utilized O3 dose increases, there is a resultant decrease in the concentrations of all NAs, including classical NAs, O-NAs, O2-NAs, O3-NAs, oxy-NAs (sum of all oxidized NAs from x = 3 to 5), and (classical + oxy)-NAs (Fig. 1a,b). After a 50 mg L−1 utilized O3 dose, removals of 65.8, 47.4 and 54.1% were achieved for classical NAs, oxy-NAs and (classical + oxy)-NAs, respectively. As the utilized O3 dose reached 170 mg L− 1, the overall removals increased to 98.8, 85.0 and 90.3% for classical NAs, oxy-NAs and (classical + oxy)-NAs, respectively. In agreement with previous work by Wang et al. (2013), two distinct reaction zones were observed during the ozonation of OSPW, with oxy-NAs and classical NAs showing a sharp decrease in the degradation rate up to 50 mg L− 1 utilized O3 dose, followed by a more gradual decrease as the utilized ozone dose increased to 170 mg L−1. Focusing on a utilized O3 dose less than 50 mg L−1, oxidative removals (per mg L− 1 utilized O3 dose) were 0.26, 0.21, and 0.47 mg L− 1 for classical NAs, oxy-NAs and (classical + oxy)-NAs, respectively. In contrast, utilized O3 doses between 100 and 170 mg L−1 resulted in oxidative removals

a

Reaction time, minutes 0

2

4

6

8

10

12

14

16

18

1.2

classical NAs O-NAs O2-NAs O3-NAs

NAs/NAs0

1.0 0.8 0.6 0.4 0.2 0.0 0

20

40

60

80

100

120

140

160

180

14

16

18

Utilized O3 dose, mg L-1

b

Reaction time, minutes 0

2

4

6

8

10

12

1.2

(classical + oxy)-NAs oxy-NAs

NAs/NAs0

1.0 0.8 0.6 0.4 0.2 0.0

0

20

40

60

80

100

120

140

160

180

Utilized O3 dose, mg L-1 Fig. 1. Degradation profile for different NA species present in OSPW, (a) [classical NAs]0 = 17.4 mg L−1, [O-NAs]0 = 11.4 mg L−1, [O2-NA]0 = 11.8 mg L−1, [O3-NA]0 = 4.7 mg L−1, and (b) [(classical + oxy)-NAs]0 = 45.3 mg L−1, [oxy-NAs]0 = 27.9 mg L−1.

(per mg L−1 utilized O3 dose) of 0.14, 0.15 and 0.29 mg L−1 of classical NAs, oxy-NAs and (classical + oxy)-NAs, respectively. As suggested by Wang et al. (2013), these dose ratios may be useful in the optimization of the ozonation process in addition to the scale-up of bench-scale experiments to full-scale processing. Fig. 1 also shows that for utilized O3 doses smaller than 30 mg L−1, oxy-NAs exhibited a slight increase in concentration due to the formation of oxy-NAs during the ozonation process (Martin et al., 2010). Qualitative analyses of the OSPW samples before and after ozone treatment were conducted using IMS (Fig. S1 in the Supplementary data). Raw OSPW showed three characteristic regions (Fig. S1a), including classical NAs, oxy-NAs and sulfur-NAs. The retention times for classical NAs and sulfur-NAs were similar, with differing drift times only, therefore, it can be concluded that these two classes of compounds have similar acid and aliphatic moieties, with similar ring numbers in their molecular structures (Wang et al., 2013). Overall, the IMS spectrum signals decreased after ozonation for classical, oxy-, and sulfur-NAs (completely removed after a 50 mg L−1 utilized O3 dose) which indicates (qualitatively) their overall degradation (Figs. S1 a,b). The oxy-NAs showed a slight increase in IMS signal at b30 mg L− 1 utilized O3 doses (data not shown) which is in agreement with their increase in concentration as presented in Fig. 1a. At the increased utilized O3 doses N30 mg L−1, there is a reduction in the IMS signal for oxy-NAs (Fig. S1). Although oxy-NAs can be formed during the ozonation of classical NAs, it is suspected that at higher utilized O3 doses (N50 mg L−1) their rate of degradation exceeds the rate of formation. An added benefit of IMS is that it may be further used for the identification of heteroatom-containing hydrocarbons, which will allow comparing the chemical fingerprinting of different OSPWs (Wang et al., 2013). IMS combined with mass spectrometry can also be used as a tool to separate complex mixtures and to resolve ions that may be indistinguishable by mass spectrometry alone (Lanucara et al., 2014). Due to the complexity of OSPW mixture, having this additional analysis capability may help elucidate the structural distribution of specific NAs isomers present in OSPW and warrants further investigation. The UPLC-HRMS analyses revealed that raw OSPW contained 45.3 mg L−1 of (classical + oxy)-NAs, with a distribution of classical NAs (39%), O-NAs (25%), O2-NAs (26%) and O3-NAs (10%) (Fig. S2). Classical NAs with n = 14–18 constituted about 80% of all NAs present, mostly distributed in |Z| = 6 (32%), |Z| = 4 (26%), and |Z| = 12 (19%) (Fig. S2a). Additionally, classical NAs with |Z| = 6 (tricyclic) and n = 15 had the highest concentrations in OSPW. These patterns were similar to those reported in previous studies (Hwang et al., 2013; Sohrabi et al., 2013; Wang et al., 2013). For O-NAs and O2-NAs, it was found that the |Z| number was mostly distributed from 4–10, with n = 13–17 (Fig. S2b,c), with the highest concentration found for |Z| = 6 and n = 14. Overall, the O-NA profiles indicated that these acids were mostly distributed in |Z| = 6 (35%), |Z| = 8 (25%), and |Z| = 4 (18%), whereas the majority of the O2-NA distribution was in |Z| = 6 (37%), |Z| = 8 (27%), and |Z| = 10 (13%). Previously, Martin et al. (2010) reported similar trend for O-NAs in untreated OSPW. The O3-NAs were distributed from |Z| = 2–12, with |Z| = 8 (26%) and |Z| = 6 (22%) being the most abundant compounds for n = 11–19 (Fig. S2d). The maximum concentration for O3-NAs was found for |Z| = 8 and n = 15. The ozonation process initially had a higher degradation rate for classical NAs and oxy-NAs with higher Z and n numbers. For example, with a 50 mg L−1 utilized O3 dose, 75, 70, 34 and 22% of O-NAs with |Z| = 12, 10, 8 and 6, were removed, respectively, whereas O-NAs with |Z| = 2–4, were removed by less than 20% (Fig. S3). For utilized O3 doses lower than 50 mg L−1, a slight increase in the concentrations of NAs with lower Z numbers including classical NAs (|Z| = 2–4), O-NAs (|Z| = 2–6), O2-NAs (|Z| = 4) and O3-NAs (|Z| = 4) was observed, which may be due to the ring opening of more cyclic (i.e., higher Z number) NAs during ozonation (Gamal El-Din et al., 2011). In all cases, ozone doses higher than 100 mg L−1 promoted the removal of 95, 64,

M.S. Islam et al. / Science of the Total Environment 493 (2014) 282–290

72, and 65% of NAs, O-NAs, O2-NAs, and O3-NAs respectively. The distribution of classical NAs, O2-NAs and O3-NAs is provided in the Figs. S4–S6. The distribution of NAs and oxy-NAs before and after ozonation of OSPW is an important consideration in the implementation of efficient treatment process. For instance, it has been reported that biodegradation alone did not decrease the concentration of classical NAs with |Z| N 8, but caused a large decrease in classical NAs with |Z| b 6. Whereas when ozonation was used alone, classical NAs with |Z| N 8 were rapidly reduced and NAs with |Z| = 0–2 were almost unchanged (Wang et al., 2013). Thus, a combination of ozonation followed by biodegradation will certainly increase the overall degradation of NAs during OSPW treatment. In a similar fashion, Gamal El-Din et al. (2011) concluded that adsorption followed by ozonation and biological treatment as part of a treatment train appears to be promising option for the management of OSPW. In this study, the distribution of oxy-NAs during the ozonation of OSPW, is presented for the first time. Of the oxy-NAs, the O-NAs and O2-NAs are the most abundant (40 and 42%, respectively); therefore, a summary of the effect of ozone on their degradation is given in Fig. 2. Fig. 2 clearly shows the increase of oxy-NAs for different n numbers at low utilized ozone doses. Similar trend was found at different Z numbers. Synchronous fluorescence spectroscopy (SFS) showed three characteristic peaks located at 270–274 nm, 308–310 nm, and 324–326 nm respectively (Fig. S7). These peaks were related to the structure of OSPW aromatic organic compounds with one-member ring (peak a) and two-member ring (peaks b and c) (Kannel and Gan, 2012; Kavanagh et al., 2009). The decreasing peak intensity with increasing utilized ozone dose may be attributed to the weakening of aromatic structures with different rings and by the enhancement of electron withdrawing groups in aromatic compounds, as shown by the SFS profile (Uyguner and Bekbolet, 2005). The decrease in the overall fluorescence intensity by ozonation may occur because ozonation breaks down chromophoric groups within the structure, which decreases the fluorescence (Henderson et al., 2009). The peak intensity decrease

a

3.2. Development of kinetic modeling Global or ‘surrogate’ parameters may be used to follow the degree of organic compound degradation through ozonation of wastewaters that contain multiple compounds at unknown concentrations (Beltrán, 2004), which is the case for the OSPW in the current study. The ozonation of wastewaters is a multiple series–parallel system of ozone reactions in which it is common to portray the kinetics of ozone reactions by using surrogate parameters which are representative of contaminant concentrations of the wastewater (Beltrán, 2004). Thus, the degradation profiles for the potential surrogate parameters including AEF, COD, classical NAs, O-NAs, O2-NAs, O3-NAs, oxy-NAs, and (classical + oxy)-NAs were used in this study to determine the pseudo-first-order reaction rate for bulk parameters. Kinetic modeling also allows the prediction of bulk parameters as a function of time and their initial concentrations. The knowledge of these concentrations over time throughout the ozonation process will allow for (1) assessing the treatment of OSPW with respect to water regulations and (2) determining the ozonation time required to achieve a minimum level of contaminant concentration. The observed degradation of organic contaminants in OSPW during ozonation is a result of a direct reaction of molecular O3 with compounds in OSPW, and indirect reactions involving hydroxyl radicals (•OH), formed as a consequence of ozone decomposition in water (Valdes and Zaror, 2006), as described by the following equation (Lee et al., 2013):
− ln

½C  ½C 0



Z ¼ kO3

Z ½O3 dt þ k• OH

•

 OH dt

ð1Þ

where [C], [O3], and [•OH] are the compound, ozone, and •OH concentrations, respectively; [C]0 is the initial compound concentration in mg L−1;

1.2

n = 12 n = 13 n = 14 n = 15 n = 16 n = 18 n = 19 n = 20

1.0

O-NAs/O-NAs0

O-NAs/O-NAs0

may be useful as a surrogate parameter indicating the removal of aromatic species contained in OSPW.

b

1.2 1.0 0.8 0.6 0.4 0.2

0.8 0.6 0.4 0.2

0.0

0.0 0

20

40

60

80

0

100 120 140 160 180

20

-1

c

40

60

80

100 120 140 160 180

Utilized O3 dose, mg L-1

Utilized O3 dose, mg L

d

1.2

1.2

1.0

n = 12 n = 13 n = 14 n = 15 n = 16 n = 18 n = 19 n = 20

1.0

O2-NAs/O2-NAs0

O2-NAs/O2-NAs0

285

0.8 0.6 0.4 0.2

0.8 0.6 0.4 0.2

0.0

0.0 0

20

40

60

80

100 120 140 160 180 -1

Utilized O3 dose, mg L

0

20

40

60

80

100 120 140 160 180

Utilized O3 dose, mg L-1

Fig. 2. Distribution of oxy-NAs in OSPW based on Z and n after ozonation for: O-NAs (a) |Z| = 6, (b) |Z| = 10; O2-NAs (c) |Z| = 6, and (d) |Z| = 10.

286

M.S. Islam et al. / Science of the Total Environment 493 (2014) 282–290

k03 and k•OH are the second-order rate constants in L mg−1 min−1; and t is the time in min. A kinetic model of NAs present in OSPW has been reported previously which assumes a first-order reaction rate since the concentrations of ozone and •OH are present in great excess throughout the ozonation experiments (Perez-Estrada et al., 2011). Following the same logic in the current study, a kinetic model was developed according to an apparent overall reaction constant (Valdes and Zaror, 2006). The solution of Eq. 1 yields: ½C  ¼ ½C 0 expð−koverall t Þ

ð2Þ

•  where koverall ¼ kO3 ½O3  þ k• OH OH is the overall apparent first-order −1 rate constant in min , which considers both the removal of pollutants by ozone and •OH. Eq. (2) can be used to model the kinetics of global parameters, including AEF, COD, NAs, oxy-NAs, and (classical + oxy)-NAs, as well as being used to determine individual structure–reactivity through kinetic constants of different classical and oxy-NAs based on Z and n numbers. There is generally good agreement between experimental and model fitting profiles for all ozonation times for AEF (Fig. 3a) and for COD (Fig. 3b) for the various NAs groups. The overall kinetic constants for these results are presented in Table S1, with a good fitting observed for the model results with the experimental data with R2 N 0.883 obtained in all cases with appropriate 95% confidence intervals. Using a combination of the measured distributions of AEF, COD and different NAs species together with the kinetic parameters (Table S1) will allow for calculating the evolution of the concentrations of global

AEF, classical NAs, Oxy-NAs, mg L-1

a

80

AEF classical NAs oxy-NAs Kinetic model

60

40

20

0 0

2

4

6

8

10

12

14

16

18

Time, minutes

b

210

COD Kinetic model

200

COD, mg L-1

190

parameters and NAs distributions for a given ozonation process as indicate previously by Lee et al. (2013). Knowledge of the evolution of NA species degradation during ozonation of OSPW will permit assessing treatment efficiency for future application of AOPs used in remediation of OSPW. Kinetics modeling used to predict the degradation profiles of micropollutants during ozonation has been developed by determining the second-order rate constants for the reaction of the micropollutant with ozone and •OH radicals (Lee et al., 2013). The development of this kinetic modeling is only possible if the ozone and •OH radical concentrations are known during the ozonation process, which would be a limitation of application of this approach in the present study. Instead, contributions of direct ozone and •OH radical reactions were grouped currently into an apparent kinetic constant (koverall) which will allow the prediction of the degradation of bulk parameters during the ozonation of OSPW by using a single kinetic parameter. 3.2.1. Reactivity of individual NA species The overall first-order kinetic rate constant for classical NAs and oxy-NAs (x = 3–5) during the ozonation of OSPW was calculated based on initial and time-based concentrations (Fig. 4). These results indicated that the reactivity of each NAs species generally increased as n increases, which is expected due to increasing overall number of H atoms available for •OH or ozone attack of NAs structures (Perez-Estrada et al., 2011). The explanations for the high structure– reactivity shown by smaller NAs in O2-NAs (|Z| = 2 and n = 13–14) and O3-NAs (|Z| = 4 and n = 13–14) are not clear, but have been previously attributed to possible branching patterns in isomers in the smallest n classes (Perez-Estrada et al., 2011). The structure–reactivity of all NA species based on their overall first-order kinetic rate constants showed a linear relationship with increasing n number (Fig. 5). Alternatively, there was no distinctive trend in reactivity with respect of Z number (Fig. 6). For classical NAs, the structure–reactivity for n = 12–15, reached maximum reactivity at |Z| = 4, 6, and 8, respectively, increasing again at higher Z numbers (Fig. 6a). Whereas for all n = 16–17 and all Z values, a steady increase in structure–reactivity was observed. In other words, for both classical NAs and oxy-NAs present in OSPW, increasing number of rings did not show any clear reactivity trend. A possible explanation for this behavior is that energy bounds of tertiary carbon in compounds with higher Z numbers differ from those found in lower Z numbers as shown for the degradation of classical NAs in OSPW after UV/H2O2 treatment previously (Acero et al., 2000). Previous studies have also indicated that during the oxidation of OSPW with persulfate and permanganate, the reactivity of classical and oxidized NAs generally increased with increasing Z number (Sohrabi et al., 2013) which may be explained by the fact that NAs with higher Z numbers containing more C_C double bonding, thus making them more susceptible to oxidation attack. 3.3. Correlations between different surrogate parameters to predict NA oxidation during the ozonation of OSPW for process control

180 170 160 150 140 130 0

2

4

6

8

10

12

14

16

18

Time, minutes Fig. 3. Experimental vs. model results in the pseudo-first-order kinetic modeling of ozonated OSPW. Figure (a) shows the experimental degradation profiles for AEF, classical NAs, oxy-NAs, and model results. Figure (b) shows degradation profiles for COD and model results.

Research on the mechanism of removal of organic compounds during the application of AOPs for municipal wastewater indicates that it should be possible to predict their removal by utilizing surrogate measurements (Acero et al., 2000). The use of surrogate parameters for the prediction of the behavior of difficult-to-measure contaminants has been extensively studied for municipal wastewater treatment (Acero et al., 2000); however, it has not been adopted widely by researchers studying the removal of both classical and oxidized NAs during the remediation of OSPW by AOPs (Mohamed et al., 2008). To address the need for simplifying the analysis of contaminants during the remediation of OSPW, measurable surrogate parameters were considered in this study to assess the removal of classical and oxy-NAs during ozonation treatment.

M.S. Islam et al. / Science of the Total Environment 493 (2014) 282–290

b

0.5 0.4 0.3 0.2 0.1 10

0.0 8

12

8 10

6

12

14

Car

bon

16

18

Num

0.5 0.4 0.3 0.2 0.1 10

0.0 8

20

12

8 10

2

6

12

14

Car

0

bon

22

-Z

4 16

18

Num

2 20

ber

c

0 22

d 0.6

-1

-1

0.6

-Z

4

ber

koverall for O2-NAs, min

koverall for O-NAs, min

-1

0.6

0.5 0.4 0.3 0.2 0.1 10

0.0 8

12

8 10

6

12 Car

14 bon

4 16

18

Num

ber

0.6 0.5 0.4 0.3 0.2 0.1 10

0.0 8

-Z

2 20

koverall for O3-NAs, min

koverall for classical NAs, min

-1

a

287

10

6

12

14

Car

0

bon

22

12

8 4 16

18

Num

ber

-Z

2 20

0 22

Fig. 4. Reactivity of individual NAs species based on Z and n numbers by using koverall kinetic rate constant for (a) classical NAs, (b) O-NAs, (c) O2-NAs, and (d) O3-NAs.

Surrogate correlations with NA degradations were obtained by using the Pearson coefficients and the following linear relationship (Gerrity et al., 2012):
1−

 
  ½NAs ½Surrogate 100 ð % Þ ¼ Slope 1− 100 ð % Þ þ Intercept ∗ ½NAs0 ∗ ½Surrogate0 

ð3Þ where [NAs] and [Surrogate] represent the concentrations of different NAs species and surrogate parameters in mg L−1 during ozonation process. Slope and intercept are the linear parameters obtained after linear regression is performed. A liner regression through the origin (i.e., zero) was not used in order to determine whether the elimination of NAs started at the same time as the elimination of the surrogate parameter. As for municipal wastewater, the potential for fluorescence (EEM), TOC, UV254 and COD, as surrogates is considered in addition to AEF which is commonly measured during various treatment processes for OSPW. As fluorescence (EEM), TOC and UV254 have been shown to be useful as municipal wastewater surrogate parameters, their usefulness as surrogates for NAs species was considered herein. The contour plots of fluorescence EEM spectroscopy for raw and ozonated OSPW with 10 mg L−1 utilized ozone dose are shown in Fig. S8. Four peaks at different excitation and emission wavelengths were present at the Ex/Em wavelengths of 230–235/340–350 nm, 280–290/340–355 nm, 265–270/300–310 nm and 225–230/290–300 nm for peaks 1 to 4, respectively (Fig. S8). The degradation profiles of the four peaks (Fig. S9b) indicate that there was a high removal of all fluorescence peaks even at low ozone doses. The TOC and UV254 marginally decreased

with the increasing utilized ozone dose (Fig. S9a), showing removal levels lower than 25%. A representative set of correlation based on percent removal between TOC and classical NAs is presented in Fig. S10. This figure shows the low linear correlation between the percent reduction of different NAs species and the reduction of TOC. Overall, the correlations between NAs species and surrogates including TOC, UV254, and EEM peak intensities (Peak 2) were all weak, characterized by high slopes with high vertical intercepts (Gerrity et al., 2012) with R2 = 0.54–0.68 (p N 0.07 and pmax = 0.22) for TOC (Fig. S9), R2 = 0.67–0.72 (p N 0.07 and pmax = 0.09) for UV254, and very low R2 = 0.35–0.59 for (p N 0.05 and pmax = 0.29) EEM. Based on these findings, it was concluded that TOC, UV254 and EEM could not be used as surrogate parameters to describe the removal rate of any NAs during the ozonation of OSPW. However, it may also be argued that if linear models are not obtained, then other non-linear models may be considered to relate removal of these surrogate parameters. In this respect, the fit of logarithmic, power equations, and exponential models were investigated to relate the percentage of NAs degradation with different surrogate parameters (Fig. S11). Although a good fit was obtained using the exponential models, these approximations are not correct since when these models are applied they result in large reading errors due to the abrupt increase of the exponential curves as observed. Essentially, small changes in percent reduction of surrogate parameters will lead to large readings in percent reduction in NAs. In contrast to the previous parameters, useful correlations were obtained by using AEF and COD as surrogate variables for all NAs species. Table S2 summarizes the degree of correlations of these two surrogate variables with all NAs species and Microtox measurements. Since

288

M.S. Islam et al. / Science of the Total Environment 493 (2014) 282–290

b

NAs reactivity based on koverall, min-1

0.6 0.5 0.4 0.3 0.2 0.1 0.0

10

12

14

16

18

20

O-NAs reactivity based on koverall, min-1

a

0.6

Z = -2 Z = -4 Z = -6 Z = -8 Z = -10 Z = -12

0.5 0.4 0.3 0.2 0.1 0.0

10

12

14

Carbon number

18

20

d

0.6 0.5 0.4 0.3 0.2 0.1 0.0

10

12

14

16

18

20

O3-NAs reactivity based on koverall, min-1

c O2-NAs reactivity based on koverall, min-1

16

Carbon number

0.6

Z = -2 Z = -4 Z = -6 Z = -8 Z = -10 Z = -12

0.5 0.4 0.3 0.2 0.1 0.0

10

12

14

Carbon number

16

18

20

Carbon number

Fig. 5. Structural reactivity of individual classical NAs based on carbon number for (a) NAs, (b) O-NAs, (c) O2-NAs, and (d) O3-NAs.

toxicity is also an important variable that determines the effectiveness of any water treatment, surrogate parameters to determine reduction of toxicity toward Vibrio fischeri were considered herein. This approach

b

O-NAs reactivity based on koverall, min-1

NAs reactivity based on koverall, min-1

a

can be extended to other types of toxicity assessments, such as in vitro studies. Currently, there is a high correlation for all NAs species degradation and toxicity toward V. fischeri using AEF measurements (R2 ≥ 0.92,

0.6 0.5 0.4 0.3 0.2 0.1 0.0 0

2

4

6

8

10

0.6

n = 10 n = 11 n = 12 n = 13 n = 14 n = 15 n = 16 n = 17 n = 18 n = 19 n = 20

0.5 0.4 0.3 0.2 0.1 0.0

12

0

2

4

-Z number

8

10

12

d

0.6 0.5 0.4 0.3 0.2 0.1 0.0 0

2

4

6

8

-Z number

10

12

O3-NAs reactivity based on koverall, min-1

c

O2-NAs reactivity based on koverall, min-1

6

-Z number

0.6

n = 10 n = 11 n = 12 n = 13 n = 14 n = 15 n = 16 n = 17 n = 18 n = 19 n = 20

0.5 0.4 0.3 0.2 0.1 0.0 0

2

4

6

8

10

12

-Z number

Fig. 6. Structural reactivity of individual (a) classical NAs, (b) O-NAs, (c) O2-NAs, and (d) O3-NAs based on Z number.

M.S. Islam et al. / Science of the Total Environment 493 (2014) 282–290

% Reduction in NAsi concentrations

a

attack. However, increasing the number of rings did not show any clear reactivity trend. When performing kinetic modeling of OSPW, a good agreement between experimental and model fitting profiles for AEF, COD, classical NAs, oxy-NAs was found with R2 N 0.883 for the kinetic constants in all cases. High correlations between AEF and COD percentage removals and different NA species degradations were found. Thus, it should be possible to predict the extent of NAs removal by utilizing AEF and COD as surrogate measurements during the ozonation of OSPW. Thus, the use of AEF and COD as surrogate parameters can be integrated into an actual treatment train to provide real-time monitoring of process performance during ozonation of OSPW. Overall, these findings suggest that the modeling approaches used in this study could serve as conservative monitoring tools for the control of the ozonation process during the treatment of OSPW.

100

80

60

40

20

0 0

20

40

60

80

100

% Reduction in AEF concentration

% Reduction in NAsi concentrations

b

289

100

Conflict of interest statement The authors of this manuscript declare to have no conflict of interests.

100 80

80

60

60

Acknowledgments

40

The authors acknowledge the financial support provided by the Helmholtz-Alberta Initiative (HAI) (RES0006289) and Natural Sciences and Engineering Research Council of Canada (NSERC) Senior Industrial Research Chair in Oil Sands Tailings Water Treatment (RES0004073), through the support from Syncrude Canada Ltd., Suncor Energy Inc., Shell Canada, Canadian Natural Resources Ltd., Total E&P Canada Ltd., EPCOR Water Services, IOWC Technologies Inc., Alberta Innovates-Energy and Environment Solution, and Alberta Environment and Sustainable Resource Development.

20

40

0

5

10

15

20

25

30

20

0 0

20

40

60

80

100

% Reduction in COD concentration Fig. 7. Relationship between (a) percent reduction in AEF concentration and (b) percent reduction in COD concentration vs. (●) classical NAs, (■) oxy-NAs, and (▲) (classical + oxy)NAs during the ozonation of OSPW. Lines represent linear models for (—) classical NAs, (----) oxy-NAs, and (•••••) (oxy + classical)-NAs. NAsi refers to the different NA groups (i.e., classical NAs, oxy-NAs, and (classical + oxy)-NAs).

p ≤ 0.002). Consequently, there was high sensitivity of AEF measurements comparable to the different NA degradations and toxicity studies by Microtox, as presented in Fig. 7a and Table S2, respectively. The level of correlation for COD and all variables was moderate with correlation coefficients R2 ranging from 0.71 to 0.96 (p ≤ 0.031) (Table S2). Overall, the COD order of correlation from highest to lowest is O-NAs (p = 0.001) N oxy-NAs (p = 0.003) N O 2-NAs (p = 0.005) N classical NAs (p = 0.029) N Microtox (p = 0.031). The results presented here indicated that COD can also be used as a surrogate variable to estimate the NAs concentrations. When AEF was used as a surrogate, a significant high correlation was obtained with R2 ranging from 0.92 to 0.99 (p ≤ 2.27 × 10−3). The order of correlation from highest to lowest is O2-NAs (p = 5.96 × 10−5) N oxy-NAs (p = 2.12 × 10 − 4 ) N classical NAs (p = 3.95 × 10− 4 ) N O 3-NAs (p = 8.28 × 10− 4) N Microtox (p = 1.00 × 10− 3) N O-NAs (p = 2.25 ×10− 3). In addition, both the slope and intercept passed the Student's t-test indicating that these linear relationships are statistically significant. Thus, it is concluded that both AEF and COD surrogate parameters presented statistically significant linear relationships with the percent reduction of different NA species and Microtox. 4. Conclusions This study demonstrated that the degradation of both classical and oxy-NAs followed similar trends during the ozonation of OSPW. The reactivity of classical and oxy-NAs generally increased as n increased due to increasing overall number of H atoms available for •OH or ozone

Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.scitotenv.2014.05.138. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.scitotenv.2014.05.138. References Acero JL, Stemmler K, von Gunten U. Degradation kinetics of atrazine and Its degradation products with ozone and OH radicals: a predictive tool for drinking water treatment. Environ Sci Technol 2000;34:591–7. American Public Health Association APHA. Standard methods for the examination of water and wastewater. 21st ed. Washington, DC, USA: American Water Works Association and Water Environment Federation; 2005. Anderson J, Wiseman SB, Moustafa A, Gamal El-Din M, Liber K, Giesy JP. Effects of exposure to oil sands process-affected water from experimental reclamation ponds on Chironomus dilutus. Water Res 2012;46:1662–72. Beltrán F. Ozone reaction kinetics for water and wastewater systems. Boca Raton: Lewis Publisher CRC Press; 2004. Bourgeois W, Burgess JE, Stuetz RM. On-line monitoring of wastewater quality: a review. J Chem Technol Biotechnol 2001;76:337–48. Chang EE, Chiang PC, Lin TF. Development of surrogate organic contaminant parameters for source water quality standards in Taiwan, ROC. Chemosphere 1998;37:593–606. Chelme-Ayala P, Gamal El-Din M, Smith DW, Adams CD. Oxidation kinetics of two pesticides in natural waters by ozonation and ozone combined with hydrogen peroxide. Water Res 2011;45:2517–26. Fedorak PM, Coy DL, Salloum MJ, Dudas MJ. Methanogenic potential of tailings samples from oil sands extraction plants. Can J Microbiol 2002;48:21–33. Gagné F, Douville M, André C, Debenest T, Talbot A, Sherry J, et al. Differential changes in gene expression in rainbow trout hepatocytes exposed to extracts of oil sands process-affected water and the Athabasca River. Comp Biochem Physiol C 2012; 155:551–9. Gamal El-Din M, Fu HJ, Wang N, Chelme-Ayala P, Pérez-Estrada L, Drzewicz P, et al. Naphthenic acids speciation and removal during petroleum-coke adsorption and ozonation of oil sands process-affected water. Sci Total Environ 2011;409:5119–25.

290

M.S. Islam et al. / Science of the Total Environment 493 (2014) 282–290

Gerrity D, Gamage S, Jones D, Korshin GV, Lee Y, Pisarenko A, et al. Development of surrogate correlation models to predict trace organic contaminant oxidation and microbial inactivation during ozonation. Water Res 2012;46:6257–72. Grewer DM, Young RF, Whittal RM, Fedorak PM. Naphthenic acids and other acidextractables in water samples from Alberta: what is being measured? Sci Total Environ 2010;408:5997–6010. Han XM, Scott AC, Fedorak PM, Bataineh M, Martin JW. Influence of molecular structure on the biodegradability of naphthenic acids. Environ Sci Technol 2008;42:1290–5. He YH, Patterson S, Wang N, Hecker M, Martin JW, Gamal El-Din M, et al. Toxicity of untreated and ozone-treated oil sands process-affected water (OSPW) to early life stages of the fathead minnow (Pimephales promelas). Water Res 2012;46:6359–68. Henderson RK, Baker A, Murphy KR, Hamblya A, Stuetz RM, Khan SJ. Fluorescence as a potential monitoring tool for recycled water systems: a review. Water Res 2009;43: 863–81. Holowenko FM, MacKinnon MD, Fedorak PM. Characterization of naphthenic acids in oil sands wastewaters by gas chromatography–mass spectrometry. Water Res 2002;36: 2843–55. Hwang G, Dong T, Islam MS, Sheng ZY, Pérez-Estrada LA, Liu Y, et al. The impacts of ozonation on oil sands process-affected water biodegradability and biofilm formation characteristics in bioreactors. Bioresource Technol 2013;130:269–77. Kannel PR, Gan TY. Naphthenic acids degradation and toxicity mitigation in tailings wastewater systems and aquatic environments: a review. J Environ Sci Health A 2012;47:1–21. Kavanagh RJ, Burnison BK RAF, Solomon KR, Van Der Kraak G. Detecting oil sands processaffected waters in the Alberta oil sands region using synchronous fluorescence spectroscopy. Chemosphere 2009;76(1):120–6. Kavanagh RJ, Frank RA, Oakes KD, Servos MR, Young RF, Fedorak PM, et al. Fathead minnow (Pimephales promelas) reproduction is impaired in aged oil sands processaffected waters. Aquat Toxicol 2011;101:214–20. Lanucara F, Holman SW, Gray CJ, Eyers CE. 2014. The power of ion mobility-mass spectrometry for structural characterization and the study of conformational dynamics. Nat Chem 2014;6:281–94. Lee S, Ahn KH. Monitoring of COD as an organic indicator in waste water and treated effluent by fluorescence excitation-emission (FEEM) matrix characterization. Water Sci Technol 2004;50:57–63. Lee Y, Gerrity D, Lee M, Encinas Bogeat A, Salhi E, Gamage S, et al. Prediction of micropollutant elimination during ozonation of municipal wastewater effluents: use of kinetic and water specific information. Environ Sci Technol 2013;47:5872–81.

Martin JW, Barri T, Han XM, Fedorak PM, Gamal El-Din M, Pérez-Estrada L, et al. Ozonation of oil sands process-affected water accelerates microbial bioremediation. Environ Sci Technol 2010;44:8350–6. Mohamed MH, Wilson LD, Headley JV, Peru KM. Screening of oil sands naphthenic acids by UV–vis absorption and fluorescence emission spectrophotometry. J Environ Sci Health A 2008;43:1700–5. Perez-Estrada LA, Han XM, Drzewicz P, Gamal El-Din M, Fedorak PM, Martin JW. Structure–reactivity of naphthenic acids in the ozonation process. Environ Sci Technol 2011;45:7431–7. Pifer AD, Fairey JL. Suitability of organic matter surrogates to predict trihalomethane formation in drinking water sources. Environ Eng Sci 2014;31:117–26. Rowland SJ, Scarlett AG, Jones D, West CE, Frank RA. Diamonds in the rough: identification of individual naphthenic acids in oil sands process water. Environ Sci Technol 2011;45:3154–9. Schramm LL, Stasiuk EN, MacKinnon M. Surfactants in Athabasca oil sands slurry conditioning, flotation recovery, and tailings processes. In: Schramm LL, editor. Surfactants, fundamentals and applications in the petroleum industry. Cambridge, UK: Cambridge University Press; 2000. p. 365–430. Scott AC, Zubot W, MacKinnon MD, Smith DW, Fedorak PM. Ozonation of oil sands process water removes naphthenic acids and toxicity. Chemosphere 2008;71: 156–60. Sohrabi V, Ross MS, Martin JW, Barker JF. Potential for in situ chemical oxidation of acid extractable organics in oil sands process affected groundwater. Chemosphere 2013; 93:2698–703. Uyguner CS, Bekbolet M. Evaluation of humic acid photocatalytic degradation by UV–vis and fluorescence spectroscopy. Catal Today 2005;101:267–74. Valdes H, Zaror CA. Heterogeneous and homogeneous catalytic ozonation of benzothiazole promoted by activated carbon: kinetic approach. Chemosphere 2006;65:1131–6. Wang N, Chelme-Ayala P, Perez-Estrada L, Garcia-Garcia E, Pun J, Martin JW, et al. Impact of ozonation on naphthenic acids speciation and toxicity of oil sands process-affected water to vibrio fischeri and mammalian immune system. Environ Sci Technol 2013; 47:6518–26. Zubot W, MacKinnon MD, Chelme-Ayala P, Smith DW, Gamal El-Din M. Petroleum coke adsorption as a water management option for oil sands process-affected water. Sci Total Environ 2012;427:364–72.