Fractionation of oil sands-process affected water using pH-dependent extractions: A study of dissociation constants for naphthenic acids species

Chemosphere 127 (2015) 291–296

Contents lists available at ScienceDirect

Chemosphere journal homepage: www.elsevier.com/locate/chemosphere

Fractionation of oil sands-process affected water using pH-dependent extractions: A study of dissociation constants for naphthenic acids species Rongfu Huang, Nian Sun, Pamela Chelme-Ayala, Kerry N. McPhedran, Mohamed Changalov, Mohamed Gamal El-Din ⇑ Department of Civil and Environmental Engineering, 3-133 Markin/CNRL Natural Resources Engineering Facility, University of Alberta, Edmonton, Alberta T6G 2W2, Canada

h i g h l i g h t s  The dissociation constants for naphthenic acids species were estimated.  Sum of O2-, O3- and O4-NAs species accounted for 33.6% of extracted organic matter.  Accumulative masses at different pHs revealed every O atom added increases the pKa.  Electron-withdrawing groups (double bonds and aromatic groups) may lead to lower pKa.  Model pKa values were 3.5 for O2-NAs, 4.8 for O3-NAs, and 6.8 for O4-NAs.

a r t i c l e

i n f o

Article history: Received 22 August 2014 Received in revised form 8 November 2014 Accepted 16 November 2014 Handling Editor: Keith Maruya

Keywords: Oil sands process-affected water Naphthenic acids pH fractionation Ultra-performance liquid chromatography time-of-flight mass spectrometry Dissociation constant

a b s t r a c t The fractionation of oil sands process-affected water (OSPW) via pH-dependent extractions was performed to quantitatively investigate naphthenic acids (NAs, CnH2n+ZO2) and oxidized NAs (Ox-NAs) species (CnH2n+ZO3 and CnH2n+ZO4) using ultra-performance liquid chromatography time-of-flight mass spectrometry (UPLC-TOFMS). A mathematical model was also developed to estimate the dissociation constant pKa for NAs species, considering the liquid–liquid extraction process and the aqueous layer acidbase equilibrium. This model provides estimated dissociation constants for compounds in water samples based on fractionation extraction and relative quantification. Overall, the sum of O2-, O3-, and O4-NAs species accounted for 33.6% of total extracted organic matter. Accumulative extracted masses at different pHs revealed that every oxygen atom added to NAs increases the pKa (i.e., O2-NAs < O3-NAs < O4-NAs), indicating that the additional O atoms exist as –OH in O3- and O4-NAs. Molecule electron-withdrawing groups such as double bonds and aromatic groups, as indicated by higher carbon and Z number, may be responsible for the lower pKa of O2-, O3-, and O4-NAs. The model obtained estimated pKa values of 3.5 for O2-NAs, 4.8 for O3-NAs, and 6.8 for O4-NAs via nonlinear regression curve fittings. These pKa values are valuable physicochemical parameters for environmental engineering applications targeting OSPW NAs treatment. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction In past decades, the growth of the oil sands industry in Northern Alberta has generated huge amounts of oil sands process-affected water (OSPW). OSPW is comprised of suspended particles, salts, ⇑ Corresponding author at: NSERC Senior Industrial Research Chair in Oil Sands Tailings Water Treatment, Helmholtz – Alberta Initiative Lead (Theme 5), Department of Civil and Environmental Engineering, 3-093 Markin/CNRL Natural Resources Engineering Facility, University of Alberta, Edmonton, Alberta T6G 2W2, Canada. Tel.: +1 780 492 5124, cell: +1 780 231 3712; fax: +1 780 492 0249. E-mail address: [email protected] (M. Gamal El-Din). http://dx.doi.org/10.1016/j.chemosphere.2014.11.041 0045-6535/Ó 2014 Elsevier Ltd. All rights reserved.

metal ions, and organic compounds, among which naphthenic acids (NAs) are abundantly present (Kelly et al., 2009). Studies have been conducted to assess the acute and chronic toxicity of NAs species towards organisms, including goldfish, larval zebrafish, Pimephales promelas, Vibrio fischeri, and the mammalian immune system (Hagen et al., 2012; He et al., 2012; Scarlett et al., 2013; Wang et al., 2013b). The persistent toxic compounds in OSPW, such as NAs, are of concern due to their potential to adversely affect the environmental and public health (Kim et al., 2012). NAs are a class of mono-carboxylic acids with the empirical formula of CnH2n+ZO2 (O2-NAs), where n is the carbon number and Z is

292

R. Huang et al. / Chemosphere 127 (2015) 291–296

a zero or a negative even integer representing the formation of rings or double bonds. In addition, NAs derivatives, such as oxidized NAs with empirical formula CnH2n+ZOx (Ox-NAs; 3 6 x 6 6), sulfur-containing NAs, and nitrogen-containing NAs have also been identified in OSPW (Headley et al., 2009; Grewer et al., 2010; Pereira et al., 2013). Recently, tandem mass spectrometry examinations of NAs species have been used to investigate the possible structural information of Ox-NAs and sulfur-containing NAs (Wang et al., 2013a; West et al., 2014). Use of structure-related information, including physicochemical properties, such as the dissociation constant pKa, is essentially important for better understanding of NAs to assess their potential toxicity and for the investigation of prospective environmental engineering treatment applications (Hoigne and Bader, 1983; Faust and Aly, 1998; Kojima et al., 2005; Babic et al., 2007). The existence of O3- and O4-NAs species, which result from the addition of oxygen (O) atoms to O2-NAs, has been reported previously (Headley et al., 2009; Grewer et al., 2010; Rowland et al., 2011; Lengger et al., 2013; Wang et al., 2013a). Rowland et al. (2011) reported hydroxyl NAs found as O3-NAs based on multidimensional comprehensive gas chromatography mass spectrometry (GC  GC–MS) of methyl esters of acids in an OSPW sample. Wang et al. (2013a) found that, O3-NAs and O4-NAs were composed of OH-NAs and (OH)2-NAs, respectively, based on the dansyl derivation of Ox-NAs followed by MS/MS characterization. Lengger et al. (2013) deduced dicarboxylic acids may be present as O4-NAs in OSPW based on results obtained via Fourier transform ion cyclotron resonance high-resolution mass spectrometry (FTICR-HRMS). It is clear from the literature that the O3-NAs are present as OH-NAs. However, of current interest is whether O4-NAs are present as (OH)2-NAs and/or di-carboxylic acids. The position of the additional O atom in Ox-NAs will have a significant influence on the dissociation constant pKa, which, in turn, will influence the extraction process of NAs using solutions of various pH values. The pH dependent extraction of O2-NAs species was previously investigated using refined O2-NAs (Headley et al., 2002), revealing that they are preferentially extracted in acidic conditions (pH 6 3); there are no known reports for Ox-NAs where x P 3. The objective of the current study is to investigate the influence of pH on the extraction of O2-, O3-, and O4-NAs from OSPW. To accomplish this objective, extracted fractions were analyzed by using ultra-performance liquid chromatography time-of-flight mass spectrometry (UPLC-TOFMS). Based on the pH-dependent experimental data, a novel mathematical model was developed to estimate pKa for different NAs species, considering experimental parameters from the liquid–liquid extraction process and the acid-base equilibrium of the water layer. The pKa values for O2-, O3-, and O4-NAs species were estimated via the model and the possible structures for O3-, and O4-NAs species were indicated based on their corresponding pKa values. The influence of n number and Z number to the pKa values of NAs were also investigated. Knowledge of the pKa values of different NAs is useful for further understanding of their fate and transport in various wastewater treatment processes.

2. Material and methods 2.1. Chemicals and reagents Cyclohexane (Acros Organic, NJ) and dichloromethane (Fisher Scientific, ON) were purchased and used as extraction solvents. NaOH and H2SO4 (Sigma–Aldrich, ON) were used to adjust solution pH. The anhydrous Na2SO4 (Sigma–Aldrich, ON) was used to remove trace water in the separated organic layers. Optima-grade water, methanol, and acetonitrile (Fisher Scientific, ON) were used for UPLC-TOFMS analysis. The myristic acid-1-13C (Sigma–Aldrich, ON) was used as the internal standard in quantification analysis. 2.2. Fractionation extraction from OSPW at different pH conditions Syncrude Canada Ltd. provided OSPW that was collected from the West In-pit pond on September 27, 2010 and was stored at 4 °C prior to use. Before sampling, OSPW was mixed uniformly using a motor driven paddle mixer. A 2-L sample was collected and filtered through a 0.45 lm nylon filter (Millipore, ON). After filtration, the OSPW sample (pH 8.5) was alkalized to pH 12.4 by adding 2 M NaOH dropwise, while stirring on an ice bath. The OSPW sample was then extracted with cyclohexane (900 mL). The cyclohexane layers were separated and dried with anhydrous Na2SO4 that was subsequently filtered off. The organic solvent was evaporated to dryness under vacuum. The remaining organic matter was precisely weighed and recorded (1). The remaining aqueous layer was acidified to pH 10.0 with H2SO4 on a continuous stirring ice bath. The aqueous layer was extracted using dichloromethane (DCM) 900 mL. As previously, the organic layers were separated, trace water removed, filtered, solvent evaporated, and the remaining organic matter precisely weighed and recorded (Table 1). The pH of the remaining aqueous layer was then decreased sequentially by one unit until pH 2.0, using the same extraction process as used at pH 10.0 (Table 1). 2.3. Quantification analysis of fractions For UPLC-TOFMS analysis, each fraction was prepared in acetonitrile with a concentration of 2 000 lg L1 total organic matter. A 1 mL of each fraction was centrifuged at 10 000 RPM for 5 min. The injection solution was prepared with 500 lL of sample supernatant, 100 lL of 4.0 mg L1 internal standard (myristic acid-1-13C) in methanol, and 400 lL methanol to reach a final sample volume of 1 mL. Chromatographic separations were performed using a Waters UPLC Phenyl BEH column (1.7 lm, 150 mm  1 mm), with mobile phases of 10 mM ammonium acetate in water (A) and 10 mM ammonium acetate in 50/50 methanol/acetonitrile (B). The elution gradient was 0–2 min, 1% B; 2–3 min, increased from 1% to 60% B; 3–7 min, increased to 70% B; 7–13 min, 95% B; 13– 14 min, 1% B and hold 1% B until 20 min to equilibrate column with a flow rate of 100 lL min1. The column temperature was set at 50 °C and the sample temperature at 4 °C. Samples were analyzed

Table 1 Fractionation extraction masses at different pH conditions and UPLC-TOFMS quantification results of O2-, O3-, and O4-NAs species in different fractions. Extraction pH

Total organic matter (mg)

O2-NAs (lg)

O3-NAs (lg)

O4-NAs (lg)

Sum of Ox-NAs (lg)

Mass percentage of Ox-NAs (%)

12.4 10.0 9.0 8.0 7.0 5.9 5.0 4.0 3.0 2.0

3.9 14.3 10.1 8.2 11.4 9.0 27.2 16.7 14.8 8.5

25.0 1939.7 241.9 388.3 557.3 1582.8 6392.7 8228.5 7556.3 2309.3

1.2 182.4 126.6 235.0 1315.5 1923.1 2140.8 1164.4 842.2 223.9

0.8 609.5 775.0 651.1 1191.6 589.8 257.9 140.0 96.1 49.1

27.0 2731.6 1143.5 1274.4 3064.4 4095.7 8791.4 9532.9 8494.6 2582.3

0.7 19.1 11.3 15.5 26.9 45.5 32.3 57.1 57.4 30.4

R. Huang et al. / Chemosphere 127 (2015) 291–296

with a high-resolution time-of-flight mass spectrometer (Synapt G2, Waters) with the electrospray ionization (ESI) source operating in negative ion mode and TOF analyzer in high-resolution mode. The negative ESI has been commonly used to determine naphthenic acids species based on previous comprehensive studies of OSPW (Pereira et al., 2013). The data acquisition process was controlled using MassLynx (Waters) and data extraction from spectra was performed using TargetLynx (Waters). The UPLC-TOFMS quantification method was developed previously and has been verified in a number of publications (Hwang et al., 2013; Wang et al., 2013b).

and 3.5% for O4-NAs) of the total extracted organic matter. It was observed that various NAs species were preferentially extracted in different pH ranges with O2-NAs, O3-NAs, and O4-NAs in ranges of pH 3–5, pH 5–7, and pH 6–8, respectively. These overall results suggest that the various NAs species have different dissociation constants (pKa) which determine the favorable extraction pH for each NAs species. For NAs as organic acids, the extraction masses are reliant on both the sample pH and the pKa value of NAs (Grbovic et al., 2012). In classic acid-base equilibrium theory, for a monoacid compound HA, the dissociation constant is calculated as follows (Skoog et al., 2013):

3. Results and discussion 3.1. Fractionation extraction from OSPW and quantification analysis of fractions The UPLC-TOFMS quantification was performed to investigate the relative abundances and concentration distributions of O2-, O3-, and O4-NAs in fractions (see Fig. 1 and Table 1; concentration distributions are presented in Figs. S1–S3 in the Supplementary Material; SM). The concentration of Ox-NAs (2 6 x 6 4) species was the sum of concentrations of all detected compounds with the formula CnO2n+ZOx (7 6 n 6 26, 0 6 Z 6 18). Fig. 1 shows the composition of organic matter for fractions extracted at pH 12.4–2.0. The relative percentages were calculated based on masses in Table 1. To remove nonpolar organic compounds from OSPW, the sample was initially alkalized to pH 12.4, at this pH all NAs species are completely ionized in water (>99.9% assuming pKa 4.9 (Havre et al., 2003). The majority of the nonpolar organic compounds would be extracted by cyclohexane at pH 12.4 with an extracted mass of 3.9 mg, which accounts for 3.1% of the total extracted organic matter (Fig. 1). The sum of O2-, O3-, and O4-NAs in this fraction was negligible (0.7% of the fraction mass) as shown in Table 1. The total organic mass was calculated using the gravimetric method and this percentage was calculated using the TOFMS determined semi-quantitative amounts of Ox-NAs (2 6 x 6 4) in the corresponding mass fractions. The total extracted masses in the pH range 5.0–3.0 were 47.3% of the total organic matter in OSPW, which suggests that the most abundant organic constituents in OSPW are weak acidic compounds (Table 1). For the NAs specifically, the total mass of O2-, O3-, and O4-NAs species accounted for 11.3–57.4% of organic matter in fractions of pH 10.0–2.0. Overall, the sum of O2-, O3-, and O4-NAs accounted for 33.6% (23.5% for O2-NAs, 6.6% for O3-NAs,

Fig. 1. The relative percentage of organic matter extracted at different pH conditions in total extracted organic matter. The O2-, O3-, and O4-NAs species were identified quantitatively using UPLC-TOFMS.

293

Ka ¼

½Hþ   ½A  ½HA

ð1Þ

where Ka is the dissociation constant of acid HA, and [H+], [HA], and [A] are concentrations of proton, HA molecule, and A ion in solution, respectively. Thus provided [HA] = [A] = 0.5  [Total] in solution, where [Total] is the total concentration of the acid HA including both molecule and ion forms, the pKa value is equal to the solution pH. In this work, the accumulative extracted mass at a specific pH was defined as the sum of extracted quantities from pH 10 to that specific pH (e.g., accumulative extracted mass at pH 7.0 is the sum of extracted masses from pH 10.0 to 7.0). The [HA] in Eq. (1) can be approximately represented by the accumulative extracted mass based on the NAs molecule form [HA], being preferentially extracted versus the NAs ion A (Mitra, 2003). In Fig. 2, the accumulative extracted mass versus sample pH was used to reveal the pH response of different NAs species in the pH-dependent fractionation extraction process. For direct comparison of species, the intensities of each NAs species were normalized against their final accumulative extracted masses. To investigate the trends of pKa values for different NAs species in OSPW, it was assumed that the NAs species were completely extracted through multiple extractions at pH 10.0–2.0. At the relative intensity of the accumulative extracted mass equal to 0.5 (Fig. 2), indicating [HA] = 0.5  [Total], the corresponding pKa for different NAs species (pKa = pH) were taken directly from the curves as pH 4.2 for O2-NAs, 6.2 for O3-NAs, and 7.8 for O4-NAs, respectively. These results show that with each additional O atom added to the O2-NAs (i.e., O3- and O4-NAs) led to an increase in pKa.

Fig. 2. Accumulative extracted mass of O2-, O3-, and O4-NAs species at different pH conditions. The accumulative extracted mass at a specific pH was the sum of extracted amounts from pH 10 to that specific pH. The intensities of each NAs species were normalized to allow for direct comparisons.

294

R. Huang et al. / Chemosphere 127 (2015) 291–296

As mentioned previously, Wang et al. (2013a) reported that O3-NAs and O4-NAs were mainly composed of OH-NAs and (OH)2NAs respectively, while Lengger et al. (2013) deduced the O4-NAs may present as a dicarboxylic acid in OSPW. Although the issue whether O4-NAs exist as (OH)2-NAs or dicarboxylic acid still remains controversial, the current examination in terms of trends of pKa for different NAs species may be useful to help clarify which of these potential NAs species are currently present. As indicated in Fig. 2, each addition of an extra oxygen atom to O2-NAs species has shifted its pKa towards being more basic, supporting that the additional oxygen atoms are more likely present as hydroxyl groups in O3- and O4-NAs species because electron-donating groups, such as hydroxyl group, decrease the acidity of a carboxylic acid (Bruice, 2004). Alternatively, the di-carboxyl group addition to create O4-NAs species would lead to an increase in acidity and result in lower pKa values compared with the O2-NAs species (Bruice, 2004). 3.2. Influence of n number and Z number to pKa of NAs compounds The accumulative extracted mass versus extraction pH was applied to investigate the impact of n number and Z number on the pKa of the NAs compounds. Fig. 3(a–c) shows the normalized accumulative extracted mass of Ox-NAs [CnH2n+ZOx, x = 2 (a), x = 3 (b), x = 4 (c)] species at different pH conditions in terms of n number, ranging from 11 to 22. The relative intensity values taken at equal intensities of 0.5 (Fig. 3a–c) showed similar trends for O2-, O3-, and O4-NAs with higher n number, resulting in lower pKa values, with the exceptions of n = 16, 18, 22 in Fig. 3a. The exceptions were due to three major interferential peaks (C16H32O2, C18H36O2, and C22H42O2; Fig. S1), as they were not observed abundantly in natural OSPW as previously reported (Hwang et al., 2013; Wang et al., 2013b). Given this lack of observation, the interference could

be attributed to potential issues with the extraction process. Fig. 3(d–f) shows the normalized accumulative extracted mass of Ox-NAs (CnH2n+ZOx, x = 2 (d), x = 3 (e), x = 4 (f)) species at different pH conditions in terms of -Z numbers, ranging from 4 to 16. The relative intensity values taken at equal intensities of 0.5 (Fig. 3d– f) showed similar trends for O2-, O3-, and O4-NAs with higher Z number resulting in lower pKa values. Fig. 3a–f indicated that higher n number and Z number leads to a decrease in pKa values for OSPW NAs. The higher n number and Z number are representative of more saturated rings, double bonds, or aromatic groups in a molecule. The presence of electronwithdrawing groups, such as double bonds and aromatic groups other than saturated rings, could increase the acidity of carboxylic acids (Bruice, 2004). The trends indicated that NAs molecules with double bonds or aromatic groups are likely to be present in the entire n and Z range observed. Thus, the influence of different n and Z number to NAs pKa can be further utilized to extract specific compounds that may be of interest in subsequent toxicity or treatment studies. The pKa values were observed to distribute in a relatively wide range according to n and Z numbers. The pKa range covered, but was not centered on, the corresponding values acquired in Fig. 2. This difference is due to the pKa for Ox-NAs is not only being decided by the pKa of individual compounds but is also dependent on their individual abundances. Thus, engineered OSPW treatment processes must still take into account a range of pKa values to fully remove entire NAs species groups. 3.3. Modeling estimation of pKa values for O2-, O3-, and O4-NAs species Based on the quantification results obtained in this study, a mathematical model was developed to estimate the pKa values for O2-, O3-, and O4-NAs species. In this model, it is assumed each

Fig. 3. Normalized accumulative extracted mass of Ox-NAs [CnH2n+ZOx, x = 2 (a), x = 3 (b), x = 4 (c)] species at different pH conditions in terms of carbon number (n = 11–22). Normalized accumulative extracted mass of Ox-NAs [CnH2n+ZOx, x = 2 (d), x = 3 (e), x = 4 (f)] species at different pH conditions in terms of Z number (Z = 4–16).

295

R. Huang et al. / Chemosphere 127 (2015) 291–296

NAs species has a representative pKa value since they comprise of a class of compounds with the same functional groups (e.g., carboxyl group) which primarily determined the acid dissociation constant (Skoog et al., 2013). Therefore, the pKa values are an essentially important physicochemical parameter to accurately determine for better understanding of the speciation of different NAs species. The pH-dependent fractionation using liquid–liquid extraction together with acid-base equilibrium in the aqueous layer determined the extracted quantities of NAs from OSPW. Thus, the total mass of compound HA in the extraction process comprised of the molecule form [HA] in the organic and water layers plus the ion form [A] in the water layer, based on the molecule form [HA] being preferably extracted by organic solvent. This relationship can be shown as follows:

mT ¼ m½HAo þ m½HAw þ m½A w

ð2Þ

where the total mass (mT) of a specific NAs species can be represented using its accumulative extraction mass at pH 2.0 because NAs are almost entirely extracted in multiple extractions from pH 12.4 to pH 2.0 (>99.5% assuming pKa  4 and extraction efficiency is 90% at each extractions). In the water layer, the dissociation constant Ka can be calculated using Eq. (3) based on an approximation that the molecular weight of HA is equal to that of A.

Ka ¼

þ ½Hþ   ½A  ½H   m½A w ¼ ½HA m½HAw

ð3Þ

For liquid–liquid extraction, the distribution constant Kd was defined as: Kd = [HA]o/[HA]w, (Mitra, 2003) provided HA is the compound to be extracted. Thus, in a liquid–liquid extraction with given volumes of organic solvent and water, a derived distribution constant Kw in this work was defined in Eq. (4) (note Kw – Kd).

Kw ¼

m½HAo m½HAw

ð4Þ

Derived from Eqs. (2)–(4), the Eq. (5) was used as the fitting equation in the modeling estimation of the parameters pKa and Kw using the extracted mass versus pH condition.

m½HAo ¼

K w  mT

ð5Þ

1 þ K w þ 10pHpK a

Table 2 presents the formulae for the modeling estimation of pKa and Kw for all NAs species using experimental data from the fractionation extractions. The symbol s in Table 2 denotes the extracted mass (i.e., m½HAo in Eq. (5)) assuming an extraction at each individual pH. The s value can be calculated based on the fractionation extraction mass a, extraction pH, and the constants pKa and Kw (as described below). The si (2 6 i 6 9) values were located within

ai < si < ei = ai + bi, since the extracted mass from a single extraction is less than that from multiple extractions at different pH conditions. Provided the bi mass of compound HA was re-dissolved in water, the mass of HA to be extracted was calculated to be Kwbi/ (1 + Kw + 10pHpKa) according to Eq. (5). Hence, si was determined using the following equation:

si ¼ ai þ

K w  bi

ð6Þ

1 þ K w þ 10pHpK a

The model was applied via multiple curve fittings to approach a constant pKa value with the initial pK0a and Kw0 estimated from curve fitting using the accumulative extracted mass e versus extraction pH (Table 2). In this manner, the revised calculation of si becomes 0 s0i ¼ ai þ K 0w  bi =ð1 þ K 0w þ 10pHpK a Þ. The second curve fitting was performed based on s0i versus pH to determine the values of pK00a and Kw00 . The pK00a and Kw00 were subsequently used to calculate the si00 . In turn, the next fitting was based on the new si00 versus pH with sequential curve fittings repeated until pK(aj1) = pK(aj) was reached, where number j is a positive integer larger than two. A representative example of this modeling approach for O2-NAs including the calculation of pKa and Kw is shown in Table S1. Nonlinear regression curve fitting was performed using the statistical software Origin. The first fitting was performed based on the accumulative extraction mass e versus pH to give an initial pK0a of 3.7 and Kw0 of 12.2, which were then used in the calculation of s0i . The second curve fitting was based on s0i versus pH resulting in the pK00a of 3.5 and Kw00 of 8.4, which were then used in the calculation of si00 . The third curve fitting was based on si00 versus pH with pK0a 00 of 3.5 and Kw000 of 6.7. At this point, the pK00a = pK0a 00 = 3.5, which was subsequently considered as the estimated dissociation constant for the O2-NAs species. The calculated extraction percentages for the O2-NAs species were estimated as 87.0% for Kw = 6.7 and to 89.3% for Kw = 8.4, respectively. Fig. S4(a) presents the final curve fitting of O2-NAs with an adjusted R-Square = 0.981 and standard error = 0.21. An analogous process was used for the O3- and O4-NAs species with Tables S2 and S3 showing the recorded process data and final curve fittings shown in Fig. S4(b) and S4(c), respectively. Overall, the modeling results for O2-NAs were pKa = 3.5 (3.0– 4.0, 95% confidence interval (CI)) with Kw = 6.7–8.4 (R2 = 0.981). The modeling results for O3-NAs were pKa = 4.8 (4.3–5.4, 95% CI) with Kw = 4.8–6.8 (R2 = 0.963). In addition, the modeling results for O4-NAs were pKa = 6.8 (6.3–7.2, 95% CI) with Kw = 5.4–7.6 (R2 = 0.937). Given the good fit between model and experimental data as indicated by high regression values (R2 P 0.93), this model provided a novel method to estimate pKa for compounds in water samples via pH-dependent fractionation extraction and quantification.

Table 2 Demonstration of modeling calculations for NAs species via multiple fittings used to approach a constant pKa value. Extraction pH

Extracted mass at pH

Extracted mass before pH

Accumulative extracted mass at pH

Calculated extracted mass using pK0a and Kw0 at pH assuming separate extractions

10.0 9.0

a10 a9

– b9 = a10

e10 = a10 e9 = a9 + b9

–

8.0

a8

b8 = a10 + a9

e8 = a8 + b8

s08 ¼ a8 þ K w  b8 =ð1 þ K w þ 10pHpK a Þ

0

s09 ¼ a9 þ K w  b9 =ð1 þ K w þ 10pHpK a Þ 0

pHpK 0a

7.0

a7

b7 = a10 + a9 + a8

e7 = a7 + b7

s07 ¼ a7 þ K w  b7 =ð1 þ K w þ 10

5.9

a6

b6 = a10 + a9 + a8 + a7

e6 = a6 + b6

s06 ¼ a6 þ K w  b6 =ð1 þ K w þ 10pHpK a Þ

5.0

a5

b5 = a10 + a9 + a8 + a7 + a6

e5 = a5 + b5

s05 ¼ a5 þ K w  b5 =ð1 þ K w þ 10pHpK a Þ

4.0

a4

b4 = a10 + a9 + a8 + a7 + a6 + a5

e4 = a4 + b4

s04 ¼ a4 þ K w  b4 =ð1 þ K w þ 10pHpK a Þ

3.0

a3

b3 = a10 + a9 + a8 + a7 + a6 + a5 + a4

e3 = a3 + b3

2.0

a2

b2 = a10 + a9 + a8 + a7 + a6 + a5 + a4 + a3

e2 = a2 + b2

s03 s02

pK0a and Kw0 calculated from first fitting using e versus pH. pK00a and Kw00 calculated from second fitting using s0 versus pH.

Þ

0 0 0

¼ a3 þ K w  b3 =ð1 þ K w þ 10

pHpK 0a 0

Þ

¼ a2 þ K w  b2 =ð1 þ K w þ 10pHpK a Þ

296

R. Huang et al. / Chemosphere 127 (2015) 291–296

4. Conclusions In this work, OSPW was sequentially fractionated at pH values ranging from pH 10 to 2 and fractions were analyzed quantitatively using UPLC-TOFMS. The pKa was found to generally decrease with increasing n number and Z number. As well, the pKa trends for the various NAs species indicated the addition of O atoms in O3- and O4-NAs are likely in the form of hydroxyl groups. A novel mathematical model was developed to estimate the pKa value for different NAs species with calculated values of 3.5 for O2-NAs, 4.8 for O3-NAs, and 6.8 for O4-NAs as determined via nonlinear regression curve fitting. The pKa is a physicochemical parameter of fundamental importance in a wide range of applications in the environmental engineering area. For example, knowledge of the dissociation constants of environmentally relevant emerging contaminants, such as pharmaceuticals and personal care products, is crucial in order to estimate their occurrence, fate and impact on the environment (Babic et al., 2007). For water and wastewater treatment practitioners, accurate pKa values coupled with the water–octanol partition coefficient, can help in the estimation of compounds adsorption in various adsorption processes (Faust and Aly, 1998). Additionally, the solution pH and the pKa values of contaminants are important factors to consider when designing oxidation processes (Kojima et al., 2005). As illustration, the rate of reaction of a specific compound with ozone depends on the degree of dissociation. In general, deprotonated species react with ozone over a wide pH range many orders of magnitude faster than the protonated compounds (Hoigne and Bader, 1983). It has been found that ozonation is an effective treatment process to both degrade OSPW NAs and reduce the OSPW toxicity towards various organisms (Gamal El-Din et al., 2011; Wang et al., 2013b). Therefore, knowing the pKa values for the different NAs species could help elucidate which species would be more reactive during the ozonation processes. As well, this information may also be useful to target specific classes of NAs in OSPW that have higher toxicity toward test organisms. Acknowledgements This research was supported by research Grants from the Helmholtz-Alberta Initiative and a Natural Sciences and Engineering Research Council of Canada Senior Industrial Research Chair in Oil Sands Tailings Water Treatment through the support by 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. Finally, the assistance in the laboratory by Ms. Maria Demeter was greatly appreciated. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chemosphere. 2014.11.041. References Babic, S., Horvat, A.J.M., Pavlovic, D.M., Kastelan-Macan, M., 2007. Determination of pK(a) values of active pharmaceutical ingredients. Trac. – Trend Anal. Chem. 26, 1043–1061.

Bruice, P.Y., 2004. Organic Chemistry, fourth ed. Prentice Hall, Upper Saddle River, NJ. Faust, S.D., Aly, O.M., 1998. Chemistry of Water Treatment, second ed. CRC, Boca Raton, FL. Gamal El-Din, M., Fu, H., Wang, N., Chelme-Ayala, P., Perez-Estrada, L., Drzewicz, P., Martin, J.W., Zubot, W., Smith, D.W., 2011. Wastewater treatment: naphthenic acids speciation and removal during petroleum-coke adsorption and ozonation of oil sands process wastewater. Sci. Total Environ. 409, 5119–5125. Grbovic, L., Pavlovic, K., Prekodravac, B., Kuhajda, K., Kevresan, S., Popsavin, M., Milic, J., Cirin-Novta, V., 2012. Fractionation of complex mixtures of naphthenic acids, their characterization and biological activity. J. Serb. Chem. Soc. 77, 147– 157. Grewer, D.M., Young, R.F., Whittal, R.M., Fedorak, P.M., 2010. Naphthenic acids and other acid-extractables in water samples from Alberta: What is being measured? Sci. Total Environ. 408, 5997–6010. Hagen, M.O., Garcia-Garcia, E., Oladiran, A., Karpman, M., Mitchell, S., El-Din, M.G., Martin, J.W., Belosevic, M., 2012. The acute and sub-chronic exposures of goldfish to naphthenic acids induce different host defense responses. Aquat. Toxicol. 109, 143–149. Havre, T.E., Sjoeblom, J., Vindstad, J.E., 2003. Oil/water-partitioning and interfacial behavior of naphthenic acids. J. Disper. Sci. Technol. 24, 789–801. He, Y., Patterson, S., Wang, N., Hecker, M., Martin, J.W., El-Din, M.G., Giesy, J.P., Wiseman, S.B., 2012. Toxicity of untreated and ozone-treated oil sands processaffected water (OSPW) to early life stages of the fathead minnow (Pimephales promelas). Water Res. 46, 6359–6368. Headley, J.V., Peru, K.M., McMartin, D.W., Winkler, M., 2002. Determination of dissolved naphthenic acids in natural waters by using negative-ion electrospray mass spectrometry. J. AOAC Int. 85, 182–187. Headley, J.V., Peru, K.M., Barrow, M.P., 2009. Mass spectrometric characterization of naphthenic acids in environmental samples: a review. Mass Spectrom. Rev. 28, 121–134. Hoigne, J., Bader, H., 1983. Rate constants of reactions of ozone with organic and inorganic-compounds in water: dissociating organic-compounds. Water Res. 17, 185–194. Hwang, G., Dong, T., Islam, M.S., Sheng, Z., Perez-Estrada, L.A., Liu, Y., Gamal El-Din, M., 2013. The impacts of ozonation on oil sands process-affected water biodegradability and biofilm formation characteristics in bioreactors. Biores. Technol. 130, 269–277. Kelly, E.N., Short, J.W., Schindler, D.W., Hodson, P.V., Ma, M., Kwan, A.K., Fortin, B.L., 2009. Oil sands development contributes polycyclic aromatic compounds to the Athabasca River and its tributaries. Proc. Natl. Acad. Sci. 106, 22346–22351. Kim, E.-S., Liu, Y., Gamal El-Din, M., 2012. Evaluation of membrane fouling for inline filtration of oil sands process-affected water: the effects of pretreatment conditions. Environ. Sci. Technol. 46, 2877–2884. Kojima, Y., Fukuta, T., Yamada, T., Onyango, M.S., Bernardo, E.C., Matsuda, H., Yagishita, K., 2005. Catalytic wet oxidation of o-chlorophenol at mild temperatures under alkaline conditions. Water Res. 39, 29–36. Lengger, S.K., Scarlett, A.G., West, C.E., Rowland, S.J., 2013. Diamondoid diacids (‘O4’ species) in oil sands process-affected water. Rapid Comm. Mass Spec. 27, 2648– 2654. Mitra, S., 2003. Sample Preparation Techniques in Analytical Chemistry. WileyInterscience, Hoboken, NJ. Pereira, A.S., Bhattacharjee, S., Martin, J.W., 2013. Characterization of oil sands process-affected waters by liquid chromatography orbitrap mass spectrometry. Environ. Sci. Technol. 47, 5504–5513. Rowland, S.J., West, C.E., Jones, D., Scarlett, A.G., Frank, R.A., Hewitt, L.M., 2011. Steroidal aromatic ‘naphthenic acids’ in oil sands process-affected water: structural comparisons with environmental estrogens. Environ. Sci. Technol. 45, 9806–9815. Scarlett, A.G., Reinardy, H.C., Henry, T.B., West, C.E., Frank, R.A., Hewitt, L.M., Rowland, S.J., 2013. Acute toxicity of aromatic and non-aromatic fractions of naphthenic acids extracted from oil sands process-affected water to larval zebrafish. Chemosphere 93, 415–420. Skoog, D.A., West, D.M., Holler, F.J., Crouch, S.R., 2013. Fundamentals of Analytical Chemistry, ninth ed. Brooks/Cole, Belmont. Wang, B., Wan, Y., Gao, Y., Yang, M., Hu, J., 2013a. Determination and characterization of oxy-naphthenic acids in oilfield wastewater. Environ. Sci. Technol. 47, 9545–9554. Wang, N., Chelme-Ayala, P., Perez-Estrada, L., Garcia-Garcia, E., Pun, J., Martin, J.W., Belosevic, M., Gamal El-Din, M., 2013b. 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. 47, 6518–6526. West, C.E., Scarlett, A.G., Tonkin, A., O’Carroll-Fitzpatrick, D., Pureveen, J., Tegelaar, E., Gieleciak, R., Hager, D., Petersen, K., Tollefsen, K.E., Rowland, S.J., 2014. Diaromatic sulphur-containing ‘naphthenic’ acids in process waters. Water Res. 51, 206–215.