Science of the Total Environment 499 (2014) 220–227

Contents lists available at ScienceDirect

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

Leachability and desorption of PCBs from soil and their dependency on pH and dissolved organic matter Silviu-Laurentiu Badea ⁎, Majid Mustafa, Staffan Lundstedt, Mats Tysklind Department of Chemistry, Umeå University, SE-901 87 Umeå, Sweden

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• The log Kd-values of 11 selected PCBs were investigated in pH static leaching tests • The log Kd-values of all PCBs decreased with increasing pH, by up to 3 log units • The de-protonation of humics may explain why Kd-values fell as pH rose from 5 to 7

a r t i c l e

i n f o

Article history: Received 31 May 2014 Received in revised form 8 August 2014 Accepted 8 August 2014 Available online 3 Sepember 2014 Editor: Damia Barcelo Keywords: Contaminated soil PCBs Leaching Distribution coefficient DOM FTIR

a b s t r a c t pH affects both soil–water partitioning coefficient (Kd) of polychlorinated biphenyls (PCBs) and dissolved organic matter (DOM), thereby influencing PCBs' leachability from contaminated soils. To explore these incompletely understood interactions, the leachability of 11 selected PCBs in a naturally aged soil was investigated in pH static leaching tests spanning a wide pH range (2 to 9). The Kd was calculated for each of the PCBs, based on their observed concentrations in the soil and leachates obtained from each test. The concentration and composition of DOM in each leachate were also determined, the latter using FTIR spectroscopy. Correlations between the DOM's FTIR spectra and Kd values were investigated by orthogonal projections to latent structures. The log Kd-values varied among the PCB congeners and were most variable at low pH, but the values for all studied congeners decreased with increasing pH, by up to 3 log units (for PCB 187). In the pH 5–7 interval, an abrupt decrease in log Kd values with increases in pH was observed, although the total organic carbon content remained relatively stable. The FTIR data indicate that fulvic and humic acids in DOM partially deprotonate as the pH rises from 5 to 7. © 2014 Elsevier B.V. All rights reserved.

1. Introduction

⁎ Corresponding author. Tel.: +46 76 776 4336; fax: +46 90 7867655. E-mail address: [email protected] (S.-L. Badea).

http://dx.doi.org/10.1016/j.scitotenv.2014.08.031 0048-9697/© 2014 Elsevier B.V. All rights reserved.

Polychlorinated biphenyls (PCBs) are ubiquitous pollutants in the environment and cycle between sediments, water, air, and soil. Until their production was restricted in the 1970s and under the Stockholm

S.-L. Badea et al. / Science of the Total Environment 499 (2014) 220–227

Convention (Lallas et al., 2001) on Persistent Organic Pollutants (2001) (Fuerhacker, 2009; Islam et al., 2014; Muir and Sverko, 2006), they were commercially produced as complex technical mixtures (e.g. Aroclor), which were used in diverse applications (e.g. as commercial waxes, non-flammable transformer fluids, sealants, and resins) in various industries. Although PCBs are now banned in most parts of the world, they continue to be released from old equipments and waste sites (Li et al., 2009). It was estimated that approximately 99% of environmental PCBs are reportedly present in soil (Bi et al., 2002). Currently, it is estimated that of the approximated 1.3 × 106 t cumulative global production, about 440 to 92,000 t is estimated to have been emitted into the environment, mostly into soil (Breivik et al., 2002, 2007; Li et al., 2009). In soils, the PCBs, like other hydrophobic organic contaminants (HOCs), are mainly adsorbed to mineral surfaces, organic matter, clay and active oxide surfaces (Wang et al., 2008) and/or exist as a separate nonaqueous-phase liquid (NAPL) (Mackay and Cherry, 1989). The transport of PCB congeners from such contaminated materials is limited due to their low aqueous solubility, slow dissolution from NAPL phases, and slow desorption from surfaces (Gan and Berthouex, 1994; Oliver, 1985). However, in some cases PCBs may be transported in water not only as a dissolved phase, but also in suspensions, either sorbed to mobile soil colloids or in emulsions with solubilizing agents (Adeel et al., 1997). The mobility may be studied in the laboratory through leaching tests in which soil–water partitioning coefficients (K d ) can be estimated. These may subsequently be used to estimate the risk for groundwater and surface water contamination (Grathwohl and Susset, 2009). PCBs, and other HOCs, are largely associated with the soil organic matter (SOM) (Frankki et al., 2006; Karickhoff et al., 1979; Persson et al., 2008; Schwarzenbach and Westall, 1981), but their retention by SOM is strongly pH dependent. This is partly due to pH effects on protonation and de-protonation states of fulvic and humic acids (FA and HA, respectively), which influence SOM solubility, and protonation of aromatic moieties (i.e. aromatic rings with multiple carboxyl groups, aromatic amines, or heteroaromatic amines), which influence π–π electron donor–acceptor interactions and thus sorption parameters (Zhu et al., 2004). Several studies have addressed the influence of pH on soil sorption coefficients of non-ionizable organic compounds in soil. Most have focused on the sorption of polycyclic aromatic hydrocarbons (PAHs) and found at most weak Kd dependence on pH (DePaolis and Kukkonen, 1997; Pan et al., 2008; Traina et al., 1989). For example, Traina et al. (1989) found that changes in pH from 1.5 to 7.3 had no effect on naphthalene's binding to a watersoluble soil organic matter extracted from a muck soil. Furthermore, using a commercial HA, Pan et al. (2008) found that the carbonnormalized sorption coefficient of pyrene (KDOC) only slightly increased with decreasing pH, increasing ionic strength, and decreasing DOM concentration. Similarly, DePaolis and Kukkonen (1997) found that changing the pH from 5.0 to 8.0 only weakly affected the soil organic carbon–water partitioning coefficient (Koc) of benzo[a]pyrene between water and humic substances (HA and FA) extracted from water and sediments. Zhu et al. (2004) detected up to 0.8 log unit changes in partitioning coefficients of phenanthrene in diverse soils (organic carbon contents: 1.17 to 17.9%, w/w) across the pH range of 2.5 to 7, but only minor changes in the partitioning of PCBs 52 and PCB 77. More recently, Bronner and Goss (2010) found that pH has no significant effect on the sorption of nonionic compounds to Pahokee Peat and Euro soils 3–5 (organic contents: 1.55 to 9.25%, w/w), regardless of their polarity. Fourier transform infrared spectroscopy (FT-IR) has been widely used to characterize the functional and structural properties of DOM (Chen et al., 2002). Recently, Choi et al. (2012) also used FTIR to demonstrate the presence of preferential PCB sorption sites in spiked sediments, but there have been few studies of pH effects on the behavior of both SOM and organic contaminants. Thus,

221

the objective of the study presented here was to characterize the effects of pH on the leachability of structurally diverse PCBs in a contaminated soil (PCBs 28, 52, 66, 101, 105 118, 138, 153, 156 180 and 187, among them 7 indicator PCBs which are the most environmentally relevant PCBs) and elucidate its correlations with the concentration and composition of dissolved organic carbon (DOC) in the leachate. To our best knowledge, the multivariate methods have never been used to correlate the leachability of PCBs with the composition of dissolved organic matter at different pH values. The leachability studies were conducted using a pH static batch leaching test with initial acid/base addition, within a wide range of pH values (2 to 9), and the DOM characterization was performed using FTIR spectroscopy. 2. Materials and methods 2.1. pH-static tests Batch leaching tests were performed with 50 g portions of a PCBcontaminated soil collected from the site of a former paper mill in Västervik, Sweden. The SOM content of the soil, determined as losson-ignition (LOI) at 550 °C, was 37.5% (see Supplementary Data, S.D.). Kaolinite (Al2Si2O5(OH)4) was identified as the main mineral phase of the soil (55%), using X-ray diffraction (see S.M.). Before the leaching tests the soil was homogenized manually. The tests were performed in 0.5 L Schott bottles at a liquid to solid ratio (L/S) of 10 L/kg, set using 0.5 L of an aqueous solution containing 0.001 M CaCl2. The pH was measured with a Mettler Delta 320 pH meter equipped with a Thermo Fischer electrode and adjusted to pre-selected set points in the range of pH 2–9 using 1, 0.1 and 0.01 M solutions of HCl and NaOH. Each leaching test was performed in duplicate at pH 3, 5, 7, and 8, and in triplicate at pH 2, 4, 6, and 9. To counter the buffering of the soil, the pH was measured again after the soil was added and adjusted to the initial values when needed. A blank containing 50 g of clean OECD soil (OECD, 1984) was also prepared, at pH 7, and subjected to the same treatments as the test samples. The bottles (21 in total) were placed in a horizontal shaker and agitated at 120 rpm, for 48 h, to equilibrate the contaminants in the solution and the soil. The pH of each leachate was measured again at the end of each test. The leachates were then separated from the soil by filtration, using a cellulose filter followed by vacuum filtration through a glass fiber filter with 0.7 μm diameter pores (Sartorius Stedim Biotech GmbH, Gottingen, Germany). 2.2. Extraction of soil leachates After separation, the leachates were spiked with 40 μL of Internal Standard (IS) solution containing 13C-labeled PCBs 28, 52, 101, 118, 138, 153, and 180 (so-called indicator PCBs), each at a concentration of about 0.05 ng/μL, purchased from LGC Standards (Borås, Sweden). A 400 mL portion of each leachate was then extracted by sequential liquid–liquid extraction with 3 × 50 mL dichloromethane using separatory funnels. The organic extracts (approx. 150 mL each) were evaporated using a rotary evaporator and dried by passage through anhydrous sodium sulfate (Na 2SO 4) columns. The rest of the leachates (approx. 100 mL of each) were saved for further DOC measurements. 2.3. Extraction of soil samples Samples (5 g dry weight) of the original PCB-contaminated soil were transferred in triplicate to pre-washed Soxhlet thimbles, spiked with 100 μL13C-labeled PCB 80, then extracted with toluene in Soxhlet Dean-Stark extractors for 24 h. Due to the high levels of PCBs in the soil, only 500 μL (approx. 0.25% by weight) of each organic extract was used for clean-up and analysis of PCBs. The extraction procedure

222

S.-L. Badea et al. / Science of the Total Environment 499 (2014) 220–227

for the soil particles trapped on the glass fiber filters was identical to that for the soil samples.

2.4. Clean-up and analysis of PCBs in soil and leachates Portions of the soil extracts intended for PCB analysis were spiked with 40 μL of the IS-solution added to the leachates. Tetradecane (50 μL) was also added and the extracts (about 150 mL each) were evaporated to near dryness (the residual volume of tetradecane) using a rotavapor. The residual extracts (including soil extracts) were dissolved in 2–3 mL n-hexane and transferred to multi-layer silica columns for clean-up (Liljelind et al., 2003). The columns consisted of 16 mm (i.d.) glass columns packed with the following components from top to bottom: 1 cm layer of Na2SO4, 4 cm of H2SO4 acidified silica, 4 cm of neutral silica, 4 cm of KOH alkaline silica, and 2 cm of glass wool. The PCBs were eluted with 50 mL of n-hexane, collected in round flasks and evaporated again until to dryness (the residual volume of 50 μL tetradecane). The bulk PCB fractions were spiked with 40 μL of a recovery standard (RS) solution, containing 13C-labeled PCBs 97 and 188 (~0.25 ng/μL of each). The PCB fractions were evaporated to the residual volume of tetradecane and transferred to 200 μL GC-vials for analysis. The PCBs were analyzed using an Agilent 6890 N gas chromatograph equipped with a non-polar DB-5 MS column (60 m × 0.25 mm × 0.25 μm from J&W Scientific, CA, USA) coupled to a high-resolution Waters Autospec Ultima NT 2000D mass spectrometer (mass resolution, 10,000), operated in electron impact ionization (EI)/selected ion monitoring (SIM) mode. The GC oven temperature program consisted of 190 °C for 2 min, followed by a 3 °C/min gradient to 260 °C, then 10 °C/min to 325 °C, held for 0.4 min. In all GC–HRMS analyses a quantification standard was used in every sequence to check the accuracy of the response factors used for the calculations of PCB concentrations. The recoveries of 13C-labeled PCBs were always above 70%.

2.7. Calculation of distribution coefficients The distribution of the contaminant between the solid and the liquid phase was described by the soil–water distribution coefficient (Kd) (mL/g) (Badea et al., 2013), which is defined as follows: Kd ¼

Cs Cw

where Cs is the concentration of a compound in the soil (μg/g), and Cw is the concentration of the same compound in the aqueous phase (μg/mL). Since the concentration of a compound in the soil after the leaching test (Cs) can be calculated by subtracting the amount of compound that migrated in the leachate phase, the Kd-values were determined solely from the concentrations in the original soil (measured in triplicate) and leachates. A demonstration of accuracy of this assumption was shown in a mass balance described previously (Badea et al., 2013). 2.8. Calculation of the ortho-PCB fraction To evaluate the influence of ortho-PCBs' molecular structure on their leachability, the concentration fraction of ortho-substituted PCBs (F(o)PCB) in each sample was calculated using the following equation: FðoÞPCB ¼

CðoÞPCB CðoÞPCB þ Cðoþ1ÞPCB

where, C(o)PCB and C(o + 1)PCB are the concentrations in aqueous phase of the least and most ortho-substituted PCBs, at the same degree of chlorination, respectively. A major advantage of using concentration fractions is that they can only vary between 0 and 1, which enables clearer graphical representation than the use of concentration ratios (Harner et al., 1999). 2.9. Multivariate data analysis

2.5. Fourier-transform infrared (FT-IR) spectroscopic measurements To characterize the DOM, leachates obtained from each experiment were placed in a freezer for 24 h at −20 °C. The frozen leachates were then freeze-dried (Kalbitz et al., 2003) and powdered using a HETOSICC instrument (Heto-Holten A/S, Allerød, Denmark), maintaining pressure and temperature at 0.2–0.1 mPa and − 50 °C to −40 °C, respectively. After freeze-drying, 4 to 10 mg of each dry sample was mixed with potassium bromide (Kalbitz et al., 2003) (infrared spectroscopy grade, Fisher Scientific, Loughborough, UK) to a total weight of about 400 mg. The aliquots were ground and homogenized using an agate pestle and mortar. FT-IR spectra were recorded over the 4000– 400 cm−1 range, with 4 cm−1 resolution, under mild vacuum conditions (3 mbar) using a BRUKER IFS 66v/S instrument, equipped with a DTGS detector. 128 interferograms were co-added to obtain a high signal-to-noise ratio. The spectra were baseline-corrected (64 point rubber band) and vector (area square sum)-normalized using OPUS software (version 5.5, Bruker Optics GmbH, Ettlingen, Germany) prior to multivariate analysis.

2.6. Total organic carbon analysis Total organic carbon (TOC) contents in the leachates were determined, after sparging and acidification, using a Shimadzu TOC-5000 high temperature catalytic oxidation instrument with non-dispersive infrared (NDIR) detection (Alvarez-Salgado and Miller, 1998). Calculation of carbon concentrations was made with potassium hydrogen phthalate as standard substance.

To evaluate overall trends in the experimentally obtained Kd values, two Principal Component Analysis (PCA) models were established using the SIMCA-P + 12.0 multivariate statistical software package (Umetrics, Umeå, Sweden). To evaluate the relationships between the wavelengths of the FT-IR spectroscopic data obtained for DOM and the leachability of the target PCBs (log Kd values) three orthogonal projections to latent structure (OPLS) models were established using the same SIMCA-P + 12.0. The three OPLS models were calculated in order to correlate the log Kd of specific PCB congeners, (i.e. PCB 28, PCB 52 and PCB 156) with the FT-IR spectra. The quality of the PCA and OPLS models was evaluated by calculating R2X and R2Y values, defined as the proportions of variance in the X- and Y-data matrices explained by the models, respectively, which also indicate goodness of fit (Holmes et al., 2008). 3. Results and discussions 3.1. Assessment of pH-dependent variation of the distribution coefficients of PCBs The concentrations of target PCBs in the contaminated soil (μg/g) determined in triplicates are shown in Table 1, indicating a homogenous distribution of PCBs in the soil samples. The equilibrium concentration of target PCBs in leachates after leaching were ranged between 3.7 ± 0.1 pg/L for PCB 156 at pH 2 and 2.483 ± 0.543 μg/L for PCB 28 at pH 9. In Table 1, the log Kd-values for all target PCBs, recorded in the experiments with the pH between 2 and 9, are also summarized. Generally, the Kd-values of all PCBs decreased with increasing pH and the highest compound-specific variation of log Kd with pH was recorded

S.-L. Badea et al. / Science of the Total Environment 499 (2014) 220–227

223

Table 1 Average log Kd (mL/g) values for target PCBs obtained from the tests at each pH. Compound

pH 2

PCB 28 PCB 52 PCB 66 PCB 101 PCB 105 PCB 118 PCB 138 PCB 153 PCB 156 PCB 180 PCB 187 Average initial pH Average final pH TOC (mg/L)

5.71 5.89 6.00 5.97 6.08 5.93 5.68 5.71 6.56 5.68 5.87 2.02

pH 3

± ± ± ± ± ± ± ± ± ± ± ±

0.09 0.09 0.10 0.07 0.08 0.08 0.08 0.09 0.13 0.10 0.17 0.04

5.47 5.71 5.79 5.92 5.92 5.78 5.78 5.84 6.62 5.91 6.05 3.32

pH 4

± ± ± ± ± ± ± ± ± ± ± ±

0.01 0.01 0.03 0.01 0.02 0.01 0.02 0.04 0.01 0.08 0.02 0.01

5.72 6.03 5.98 6.19 6.18 6.01 6.01 6.00 6.76 6.03 6.25 4.17

pH 5

± ± ± ± ± ± ± ± ± ± ± ±

0.03 0.05 0.05 0.06 0.06 0.06 0.07 0.09 0.07 0.09 0.09 0.06

5.57 5.92 5.88 6.08 6.13 5.92 5.92 5.94 6.70 5.98 6.14 5.19

pH 6

± ± ± ± ± ± ± ± ± ± ± ±

0.13 0.02 0.02 0.02 0.06 0.03 0.02 0.06 0.01 0.01 0.03 0.08

5.63 4.53 4.54 4.55 4.59 4.39 4.39 4.45 5.10 4.52 4.49 6.11

pH 7

± ± ± ± ± ± ± ± ± ± ± ±

0.15 0.14 0.13 0.13 0.12 0.12 0.12 0.12 0.13 0.12 0.12 0.03

4.25 4.20 4.19 4.19 4.17 4.16 4.16 4.16 4.73 4.15 4.14 7.28

pH 8

± ± ± ± ± ± ± ± ± ± ± ±

0.06 0.11 0.09 0.10 0.09 0.09 0.10 0.11 0.08 0.10 0.10 0.12

3.83 3.76 3.80 3.75 3.73 3.71 3.71 3.72 4.43 3.71 3.68 8.22

pH 9

± ± ± ± ± ± ± ± ± ± ± ±

0.01 0.01 0.01 0.01 0.00 0.00 0.02 0.00 0.01 0.02 0.02 0.05

3.34 3.22 3.25 3.24 3.25 3.21 3.21 3.22 3.93 3.20 3.20 9.31

± ± ± ± ± ± ± ± ± ± ± ±

0.11 0.12 0.10 0.11 0.11 0.10 0.08 0.07 0.07 0.07 0.07 0.06

2.32 ± 0.04

3.95 ± 0.01

4.16 ± 0.03

4.79 ± 0.04

5.28 ± 0.03

6.33 ± 0.01

7.20 ± 0.01

8.05 ± 0.35

62.4 ± 19.3

57.7

67.2 ± 33.7

65.0 ± 19.4

92.8 ± 51.5

80.9 ± 23.4

133.0 ± 19.8

465.0 ± 115.3

for PCB 187, as its log Kd decreased from 6.25 of 3.20 with more than 3.00 log Kd units, while the pH increased from 5 to 9 (see Table 1). To obtain a complete overview of the Kd values for all congeners, the data were also explored by PCA. A first PCA model was designed with N = 11 observations (all PCB congeners) and K = 8 variables (log Kd values). This two component model explained 95.6% of the variance in the data (R2X-value 0.956), of which PC1 contributed with 75.6% and PC2 with 19.0%, while the cumulated Q2-value was 0.380. The observations in the model were divided into five classes, corresponding to the target PCBs' degree of chlorination (tri, tetra, penta, hexa, and hepta PCBs). The resulting score plot (Fig. 1A) showed that two congeners (PCB 28 and PCB 156) had deviating properties, while all the others appeared to have similar distribution patterns. To further explore differences among PCB congeners that clustered in the first model, a second PCA model was designed with N = 9 observations and K = 8 variables (X = 8), excluding Kd values for PCBs 28 and PCB 156. The resulting two-component model explained 86.1% of the variation in data (PC1 44.8% and PC2 41.3%). In the resulting score plot (Fig. 1B), the remaining congeners can be divided into four groups that corresponds to their degree of chlorination (tri, tetra, penta, hexa and hepta PCBs). The variability in Kd values was clearly related to the degree of chlorination, as tetra congeners PCB 52 and PCB 66, the penta congeners PCB 101 and PCB 105, the hexa congeners PCB 138 and PCB 153 and the hepta congeners PCB 180 and PCB 187 were clustered together, respectively. However, the distribution of PCB 118 deviated from that of the other two penta CBs. In order to further explore how the variation of pH influences the Kd values of the PCBs, the log Kd values of relevant PCBs were both plotted vs. pH and TOC (see Fig. 2). In Fig. 2, the variations of log Kd values vs. pH and TOC are shown for PCB 28, PCB 52, and PCB 156 (Fig. 2A) as well as for PCB 66, PCB 105, PCB 118 and PCB 187 (Fig. 2B). For the selected PCB congeners, three pH intervals can be observed: I, the pH interval of 2–5, where the changes in both the TOC-content of the leachate and in log Kd-values were negligible or very small for all studied PCB congeners. II, the pH interval of 5–7, where there was an abrupt decrease of log Kd values, even though the TOC remained relatively stable. III, the pH interval of 7–9; where a continuing decrease in log Kd was accompanied with a significant increase in TOC. For example within the pH interval of 2–5, the log Kd of PCB 52 and PCB 156 were between 5.89 and 5.92 for PCB 52 and between 6.56 and 6.71 for PCB 156 respectively, showing no clearl trend (Fig. 2A), while the TOC of the leachates did not changed. Similarly within the same pH interval, the log Kd of PCB 66, PCB 105, PCB 118 and PCB 187 varied without a clear trend (Fig. 2B). For PCB 28, the Kd-value was relatively unaffected all the way to pH 6, which possibly

Concentration in soil (μg/g)

Kow (Walters et al., 2011)

Solubility (μg/mL) (Mackay et al., 2006)

11 4.1 4.3 0.88 0.55 0.72 0.333 0.253 0.160 0.189 0.084

5.67 5.84 6.20 6.38 6.65 6.74 6.83 6.72 7.18 7.21 7.17

0.085 0.046 0.058 0.031 0.000982 0.00107 0.00173 0.00095 0.000397 0.00031 0.00047

± ± ± ± ± ± ± ± ± ± ±

2 0.3 0.3 0.06 0.04 0.04 0.005 0.008 0.007 0.002 0.002

can be explained by its relatively high water solubility (0.085 mg/L) in comparison with the other studied PCBs (Fig. 2A). With increases in pH within the 5–7 interval the log Kd values of all the abovementioned PCBs decreased (by up to 2.00 units, for PCB 187), while the TOC content remained relatively stable. This is an interesting finding, showing that a slight change in pH can affect the leachability of some PCBs significantly, without affecting the TOC content in the leachate. This might be attributed to the de-protonation of carboxylic groups in fulvic and humic matters, a process that is soil-specific. With further increases in pH within the 7–9 interval, the log Kd values decreased, accompanied by a significant increase in TOC (from 80.9 ± 23.4 mg/L to 465.0 ± 115.3 mg/L) (Fig. 2A and B). The highest and lowest congener-specific variations in this pH interval were recorded for PCBs 52 and 156 (about 1.00 and 0.8 log Kd units, respectively) (Fig. 2A), while the variations in log Kd values of PCBs 66, 105, 118 and 187 with pH were virtually identical (Fig. 2B). 3.2. Correlation between leachability of ortho-PCBs and pH As ortho-chlorine atoms reportedly influence pH effects on the sorption of selected PCBs (PCB 77 versus PCB 52), albeit weakly (Zhu et al., 2004), we also investigated the influence of chlorine position on the leachability of the target PCBs by calculating concentration fractions of selected pairs of mono-ortho to di-ortho PCBs — PCB 66 (2,3′,4,4′) versus PCB 52 (2,2′,5,5′), PCB 105 (2,3,3′,4,4′) versus PCB 101 (2,2′,4,5,5′), and PCB 156 (2,3,3′,4,4′,5) versus PCB 153 (2,2′,4,4′,5,5′) — and di-ortho to tri-ortho PCBs — PCB 180 (2,2′,3,4,4′,5,5′) versus PCB 187 (2,2′,3,4′,5,5′,6). In contrast to the cited finding, we detected no clear trend, except in the PCB 66/PCB 52 fraction. Ortho-chlorine atoms also have no reported effects on the sorption of hexa-PCB pairs 138/156 and 156/169 to charcoal (Koelmans et al., 2009), possibly because compounds in this homologue group might be too bulky to show preferential adsorption of the planar congeners (PCB 156 and PCB 169, respectively). However, the PCB 66/PCB 52 fraction decreased from 0.541 ± 0.001 at pH ~ 4 to 0.472 ± 0.013 at pH ~ 3, and further to 0.451 ± 0.012 at pH ~2, the fractions recorded at pH 2 and 3 being significantly lower than the average across the full pH range for this pair (0.503 ± 0.031, n = 20). The lower concentration fractions at pH 2 and 3, relative to the fraction at pH 4, indicate that orthochlorine atoms may have pH-related effects on the leachability of this pair, possibly due to steric effects hindering PCB 52's adoption of a coplanar ring conformation. However, further studies are clearly needed to elucidate the correlation between leachability of orthoPCBs and pH.

224

S.-L. Badea et al. / Science of the Total Environment 499 (2014) 220–227

A

7.0

A

600

6.5

500 400

5.5

300

5.0 4.5

200

TOC (mg/L)

Log Kd

6.0

4.0 100

3.5

0

3.0 1

2

3

4

5

6

7

8

9

10

pH 7.0

B

600

6.5

500 400

5.5

300

5.0 4.5

200

TOC (mg/L)

B

Log Kd

6.0

4.0 100

3.5

0

3.0 1

2

3

4

5

6

7

8

9

10

pH Fig. 2. The variation of log Kd of PCB 28 (black), PCB 52 (red) and PCB 156 (green) and the change in total organic carbon ( ) vs. pH (A). The variation of log Kd of PCB 66 (blue), PCB 105 (cyan), PCB 118 (magenta) and PCB 187 (dark yellow) and the change in total organic carbon ( ) vs. pH.

Fig. 1. Score scatter plot of log Kd values for (A) all 11 PCBs and (B) nine PCBs (excluding PCBs 28 and 156). The five classes, corresponding to the degree of chlorination of target PCBs, are: 1, tri-; 2, tetra-; 3, penta-; 4, hexa- and 5, hepta-PCBs.

3.3. Qualitative characterization of DOM by FTIR spectroscopy The composition of DOM in the leachates obtained from the tests at pH 3–9 was explored by FTIR spectroscopy, particularly in the pH 5–7 interval where the PCBs' Kd values substantially changed but the leachates' TOC content did not significantly vary (Fig. 3). As shown by the spectra in Fig. 3, the FTIR spectra of DOM from the pH 5 leachates include a broad band at approximately 3300–3650 cm−1, which can be attributed to free hydroxyl groups in alcohols and phenols and/or N\H bonds in amines and amides (Xue et al., 2009). This band is much smaller in the pH 7 spectra. A shoulder peak at approximately 3270 cm−1, and two peaks at approximately 1623 cm−1 and 1685 cm−1 indicative of C_O stretching in amide groups, in the pH 5 spectra confirm the presence of amide functional groups (Xue et al., 2009) (Silverstein et al., 1991). The pH 7 spectra show bands at 3359 cm−1 and 1620 cm−1 attributable to N\H stretching and C_O stretching in amide groups, but they are substantially weaker than in the pH 5 spectra. In addition, the pH 7 spectra show signals attributable to C\O stretching, but not C_O stretching in carboxylic acids, at about 1150 cm−1 and 1690 cm−1, respectively. These features indicate that partial de-protonation of fulvic and humic acids in DOM may occur as the pH rises from 5 to 7, potentially explaining the abrupt associated decrease in log Kd values for all

PCBs. Furthermore, a weak band present in the pH 5 spectra, but not pH 7 spectra, at approximately 1782 cm− 1 can be related to nonconjugated C_O stretching in carboxyl groups of humic and fulvic acids, or to a lesser extent aldehyde or ketone groups (Stevenso and Goh, 1971). The high peaks in both curves of Fig. 3 spectra at approximately 1380 cm− 1 respectively, can be attributed to C\H deformation of methyl groups and for the spectra of the pH 7 leachate, a peak recorded at 2960 cm−1 might indicate an aliphatic stretching. Furthermore, absorption bands at approximately 1400–1470 cm−1 in both spectra may appear because of aliphatic C\H deformation/O\H bending vibration of carboxylic groups/C\O stretching of phenolic OH (Guo et al., 2012). C\O stretching and/or O\H deformation of \COOH and/or C\O stretching of esters (Xue et al., 2009) and/or saturated C\N stretching is probably responsible for the strong and sharp bands at approximately 1150 cm−1 in both spectra. The normal peaks in the range of 600–800 cm−1 recorded in both spectra (but slightly higher at pH 7) may be the characteristics of C\H stretching, C\Cl stretching and/or by the substitution at benzene ring (Smith, 1999) and relatively high intensity of these peaks might indicate the strong aromatic nature of the DOM obtained both at pH 5 and 7. Although the de-protonation of FA and HA is dominant, since in general KDOC is increasing with a greater aromaticity of the DOM and hydrophobicity of the HOCs (Burkhard, 2000; Chin et al., 1997), the slightly higher absorbance given by aromatic moieties in the DOM from pH 7 (Fig. 3) (comparing with their absorbance at pH 5) might indicate that aromatic moieties contributed also to the decrease of log Kd values for all PCBs within pH interval of 5–7. For an overview of the pH influence on DOM, the FTIR spectra of solid DOM extracted from the leachates samples at pH 3 and 9 are presented in S.D. (in Figs. S1 and S2, respectively).

S.-L. Badea et al. / Science of the Total Environment 499 (2014) 220–227

C-O stretching of carboxylic acids (1150 cm-1)

225

0.12

0.08 0.06 0.04

Absorbance

0.10

OH and/or N-H (3300-3650 cm-1) C=O stretching in amide (1623 cm-1) C=O stretching of carboxylic acids (1690 cm-1)

0.02 0.00

N-H stretching bands (3359 cm-1)

Aliphatic C–H deformations (1380 cm-1)

Aromatic C-H stretching (600-800 cm-1) 3500

3000

2500

2000

1500

1000

500

Wavelength (cm-1) Fig. 3. FTIR spectra of solid DOM extracted from the leachates sampled at pH 5 (blue) and 7 (red).

3.4. Correlations between the characteristics of dissolved organic matter and leachability of PCBs explored by FTIR spectroscopy and multivariate analysis The correlation between the log K d of specific PCB congeners, (i.e. PCB 28, PCB 52 and PCB 156) and the FTIR spectra was explored by orthogonal projections to latent structures (OPLS) resulting in three congener specific one-dimensional models. These models were designed with N = 17 observations and K = 1661 variables (X = 1660, Y = 1). Within these models, the predictive components are capturing variation found in both X and Y. The PCB 28 model explained 96% (R2Y) of the variation in log Kd-values using 42% (R2X) of the variation in FTIR-spectra, while 52% of the variation in the FTIR-data was defined as orthogonal to the variation in log Kd. The PCB 52 model explained 98% (R2Y) of the variation in log Kd-values using 48% (R2X) of the

variation in FTIR-spectra, while 50% of the variation in the FTIR-data was defined as orthogonal to the variation in log Kd. The PCB 156 model explained 98% (R2Y) of the variation in log Kd-values using 50% (R2X) of the variation in FTIR-spectra, while 48% of the variation in the FTIR-data was defined as orthogonal to the variation in log Kd. The above mentioned quality parameters shown that these models could explain most of the variations within the data set (viz. the variation in log Kd). In order to correlate the log Kd of specific PCB congeners, i.e. PCB 28, PCB 52 and PCB 156, with the FT-IR spectra, all the three loadings of the OPLS models for all these three congeners are shown in one single figure (Fig. 4) as the predictive components vs. spectra. Responses of the absorption bands at approximately 1690 cm−1 (which might originate from the C_O stretching of carboxylic acids) are similar for PCBs 52 and 156, but somewhat stronger for PCB 28. This, combined with a higher predictive component for PCB 28 at

C-H deformation of aliphatic carbon

-C-H vibration of aliphatic carbon C=O stretching of carboxylic acids

C-O stretching of carboxylic acids

Fig. 4. OPLS loading plot for predictive component 1 vs. wavelength for the responses of PCB 28 (back), PCB 52 (red) and PCB 156 (green).

226

S.-L. Badea et al. / Science of the Total Environment 499 (2014) 220–227

about 1150 cm−1 (which may originate from C\O stretching of carboxylic acids) indicates that carboxylic groups in the DOM influenced the log-Kd values of PCB 28 more strongly than those of PCBs 52 and 156. A peak at 1370 cm−1 that may arise from C\H deformation of aliphatic carbon is also strong and sharp in the PCB 28 spectrum, but very weak or negligible in the spectra of PCBs 52 and 156. Similarly, absorption bands in the region 2933–2856 cm−1 (which may indicate \C\H vibration of aliphatic carbon) are similar for PCBs 52 and 156 but somewhat higher for PCB 28. Recently, (Choi et al., 2012) found that the sorption of 2chlorobiphenyl (PCB 1) in spiked sediment correlated with a shift in \C\H vibrational frequencies. Therefore, we suggest that desorption of low chlorinated PCB 28 from the soil matrix during our leaching tests might be associated with a shift in \C\H vibrational frequencies recorded during FTIR spectroscopic analysis of solid DOM. This is an interesting finding suggesting that alkyl-C moieties, in addition to aromatic moieties, can contribute to KDOC for the least chlorinated PCB congeners (e.g. PCB 28).

4. Conclusions The clearest general results from this study are that the leachability of PCBs increased with increasing pH, and that this correlated with increases in TOC contents in the leachates. Furthermore, leachability varied among the PCBs, especially at low pH, where it was correlated to their Kow (Velzeboer et al., 2014; Walters et al., 2011) (Table 1). At higher pH, the differences were much smaller, which can be explained by more similar increases in their apparent solubilities as aqueous TOC contents rise (Lara and Ernst, 1989). However, more detailed analysis of the data provides further information on the leachability of individual PCB congeners. The highest compound-specific variation with pH of the log Kd-values was recorded for PCB 187, while overall the highest Kd-value among the PCBs was recorded always for PCB 156. Thus we can conclude that the PCBs with the highest Kow values are showing the highest variation of the log Kd-values with pH. Soil-specific partial de-protonation of fulvic and humic acids, as indicated by the FTIR analysis, may explain why log Kd values substantially fell as pH rose from 5 to 7, although there were minor changes in TOC content. Differences between concentration fractions recorded at pH 2 and 3 compared to those recorded at pH 4, indicate that ortho-chlorine atoms influence the leachability of PCB 52 via pH effects. The FTIR analyses of DOM show that the least chlorinated and hydrophobic PCB congeners (represented here by PCB 28) might be more strongly associated with the hydrophilic fraction (i.e. carboxylic groups) of the DOM than the most hydrophobic PCB congeners (e.g. PCB 156). Our study demonstrates that complex interactions between both the pH of leachates and composition of dissolved organic carbon (DOC) in the leachates influence the leachability of PCBs in a compound-specific manner. These findings may contribute to the understanding and assessment of organic contaminants' mobility at contaminated soil sites.

Acknowledgments The research of Silviu-Laurentiu Badea was funded by Umeå University, Environmental- and Biogeochemistry ref no. 313-1780-08. The work was performed within the Northern Sweden Soil Remediation Center (MCN) framework. The authors are grateful to Dr. Per Liljelind at the Trace Analyses Platform for his support with GC–HRMS measurements, and Dr. Rui Climaco Pinto from BILS (Bioinformatics Infrastructure for Life Sciences) for his support with multivariate analysis. Dr. András Gorzsás and the Vibrational Spectroscopy Platform of Umeå University are gratefully acknowledged for assisting with the Fourier transform infrared (FTIR) spectroscopic measurements.

Appendix A. Supplementary data Details concerning the quantification of minerals and crystalline phases by X-ray diffraction, soil characterization, preprocessing of FTIR spectra and qualitative characterization of DOM by FTIR. Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j. scitotenv.2014.08.031. References Adeel Z, Luthy RG, Dzombak DA, Roy SB, Smith JR. Leaching of PCB compounds from untreated and biotreated sludge-soil mixtures. J Contam Hydrol 1997;28:289–309. Alvarez-Salgado XA, Miller AEJ. Simultaneous determination of dissolved organic carbon and total dissolved nitrogen in seawater by high temperature catalytic oxidation: conditions for precise shipboard measurements. Mar Chem 1998;62:325–33. Badea S-L, Lundstedt S, Liljelind P, Tysklind M. The influence of soil composition on the leachability of selected hydrophobic organic compounds (HOCs) from soils using a batch leaching test. J Hazard Mater 2013;254–255:26–35. Bi XH, Chu SG, Meng QY, Xu XB. Movement and retention of polychlorinated biphenyls in a paddy field of WenTai area in China. Agric Ecosyst Environ 2002;89:241–52. Breivik K, Sweetman A, Pacyna JM, Jones KC. Towards a global historical emission inventory for selected PCB congeners — a mass balance approach 2. Emissions. Sci Total Environ 2002;290:199–224. Breivik K, Sweetman A, Pacyna JM, Jones KC. Towards a global historical emission inventory for selected PCB congeners — a mass balance approach-3. An update. Sci Total Environ 2007;377:296–307. Bronner G, Goss K-U. Sorption of organic chemicals to soil organic matter: influence of soil variability and pH dependence. Environ Sci Technol 2010;45:1307–12. Burkhard LP. Estimating dissolved organic carbon partition coefficients for nonionic organic chemicals. Environ Sci Technol 2000;34:4663–8. Chen J, Gu BH, LeBoeuf EJ, Pan HJ, Dai S. Spectroscopic characterization of the structural and functional properties of natural organic matter fractions. Chemosphere 2002; 48:59–68. Chin Y-P, Aiken GR, Danielsen KM. Binding of pyrene to aquatic and commercial humic substances: the role of molecular weight and aromaticity. Environ Sci Technol 1997;31:1630–5. Choi H, Nfodzo PA, Lawal WA, Al-Abed SR, Seo Y. Phenomenological and spectroscopic analysis on the effects of sediment ageing and organic carbon on the fate of a PCB congener spiked to sediment. J Hazard Mater 2012;239–240:325–32. DePaolis F, Kukkonen J. Binding of organic pollutants to humic and fulvic acids: Influence of pH and the structure of humic material. Chemosphere 1997;34:1693–704. Frankki S, Persson Y, Tysklind M, Skyllberg U. Partitioning of CPs, PCDEs, and PCDD/Fs between particulate and experimentally enhanced dissolved natural organic matter in a contaminated soil. Environ Sci Technol 2006;40:6668–73. Fuerhacker M. EU water framework directive and Stockholm convention can we reach the targets for priority substances and persistent organic pollutants? Environ Sci Pollut Res 2009;16:92–7. Gan DR, Berthouex PM. Disappearance and crop uptake of PCBs from sludge-amended farmland. Water Environ Res 1994;66:54–69. Grathwohl P, Susset B. Comparison of percolation to batch and sequential leaching tests: theory and data. Waste Manag 2009;29:2681–8. Guo XJ, He XS, Zhang H, Deng Y, Chen L, Jiang JY. Characterization of dissolved organic matter extracted from fermentation effluent of swine manure slurry using spectroscopic techniques and parallel factor analysis (PARAFAC). Microchem J 2012;102: 115–22. Harner T, Wiberg K, Norstrom R. Enantiomer fractions are preferred to enantiomer ratios for describing chiral signatures in environmental analysis. Environ Sci Technol 1999; 34:218–20. Holmes E, Loo RL, Stamler J, Bictash M, Yap IKS, Chan Q, et al. Human metabolic phenotype diversity and its association with diet and blood pressure. Nature 2008;453:396–400. Islam MN, Park J-H, Shin M-S, Park H-S. Decontamination of PCBs-containing soil using subcritical water extraction process. Chemosphere 2014;109:28–33. Kalbitz K, Schwesig D, Schmerwitz J, Kaiser K, Haumaier L, Glaser B, et al. Changes in properties of soil-derived dissolved organic matter induced by biodegradation. Soil Biol Biochem 2003;35:1129–42. Karickhoff SW, Brown DS, Scott TA. Sorption of hydrophobic pollutants on natural sediments. Water Res 1979;13:241–8. Koelmans AA, Meulman B, Meijer T, Jonker MTO. Attenuation of polychlorinated biphenyl sorption to charcoal by humic acids. Environ Sci Technol 2009;43:736–42. Lallas PL. The Stockholm Convention on persistent organic pollutants. Am J Int Law 2001; 95(3):692–708. Lara R, Ernst W. Interaction between polychlorinated-biphenyls and marine humic substances — determination of association coefficients. Chemosphere 1989;19: 1655–64. Li Y-F, Harner T, Liu L, Zhang Z, Ren N-Q, Jia H, et al. Polychlorinated biphenyls in global air and surface soil: distributions, air–soil exchange, and fractionation effect. Environ Sci Technol 2009;44:2784–90. Liljelind P, Soderstrom G, Hedman B, Karlsson S, Lundin L, Marklund S. Method for multiresidue determination of halogenated aromatics and PAHs in combustionrelated samples. Environ Sci Technol 2003;37:3680–6. Mackay DM, Cherry JA. Groundwater contamination: pump-and-treat remediation. Second of a five-part series. Environ Sci Technol 1989;23:630–6.

S.-L. Badea et al. / Science of the Total Environment 499 (2014) 220–227 Mackay D, Wan YS, Kuo-Ching M, Sum CL. Handbook of physical–chemical properties and environmental fate for organic chemicals. 2nd ed. CRC Press; 2006. Muir D, Sverko E. Analytical methods for PCBs and organochlorine pesticides in environmental monitoring and surveillance: a critical appraisal. Anal Bioanal Chem 2006; 386:769–89. OECD. Test 207: earthworm, acute toxicity tests. In: Organization for Economic Cooperation and Development, editor. OECD guidelines for testing of chemicals; 1984. Oliver BG. Desorption of chlorinated hydrocarbons from spiked and anthropogenically contaminated sediments. Chemosphere 1985;14:1087–106. Pan B, Ghosh S, Xing BS. Dissolved organic matter conformation and its interaction with pyrene as affected by water chemistry and concentration. Environ Sci Technol 2008;42:1594–9. Persson Y, Hemstrom K, Oberg L, Tysklind M, Enell A. Use of a column leaching test to study the mobility of chlorinated HOCs from a contaminated soil and the distribution of compounds between soluble and colloid phases. Chemosphere 2008;71:1035–42. Schwarzenbach RP, Westall J. Transport of nonpolar organic compounds from surface water to groundwater. Laboratory sorption studies. Environ Sci Technol 1981;15:1360–7. Silverstein RM, Bassler GC, Morrill TC. Spectrometric identification of organic compounds. 5th ed. USA: John Wiley and Sons, Inc; 1991.

227

Smith B. Infrared spectral interpretation. Boca Raton: CRC Press; 1999. Stevenso FJ, Goh KM. Infrared spectra of humic acids and related substances. Geochim Cosmochim Acta 1971;35:471–83. Traina SJ, Spontak DA, Logan TJ. Effects of cations on complexation of naphthalene by water-soluble organic carbon. J Environ Qual 1989;18:221–7. Velzeboer I, Kwadijk C, Koelmans AA. Strong sorption of PCBs to nanoplastics, microplastics, carbon nanotubes, and fullerenes. Environ Sci Technol 2014;48: 4869–76. Walters DM, Mills MA, Cade BS, Burkard LP. Trophic magnification of PCBs and its relationship to the octanol–water partition coefficient. Environ Sci Technol 2011;45: 3917–24. Wang X, Lu J, Xu M, Xing B. Sorption of pyrene by regular and nanoscaled metal oxide particles: influence of adsorbed organic matter. Environ Sci Technol 2008;42: 7267–72. Xue S, Zhao QL, Wei LL, Ren NQ. Behavior and characteristics of dissolved organic matter during column studies of soil aquifer treatment. Water Res 2009;43:499–507. Zhu D, Hyun S, Pignatello JJ, Lee LS. Evidence for π–π electron donor–acceptor interactions between π–donor aromatic compounds and π–acceptor sites in soil organic matter through pH effects on sorption. Environ Sci Technol 2004;38:4361–8.

Leachability and desorption of PCBs from soil and their dependency on pH and dissolved organic matter.

pH affects both soil-water partitioning coefficient (Kd) of polychlorinated biphenyls (PCBs) and dissolved organic matter (DOM), thereby influencing P...
948KB Sizes 1 Downloads 6 Views