Evaluation of soil flushing potential for clean-up of desert soil contaminated by industrial wastewater

Chemosphere 62 (2006) 17–25 www.elsevier.com/locate/chemosphere

Evaluation of soil flushing potential for clean-up of desert soil contaminated by industrial wastewater Shai Arnon

a,b,*

, Zeev Ronen a, Alexander Yakirevich a, Eilon Adar

a,b

a Department of Environmental Hydrology and Microbiology, Zuckerberg Institute for Water Research, J. Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Sede-Boqer Campus 84990, Israel b Department of Geological and Environmental Sciences, Ben-Gurion University of the Negev, Israel

Received 13 January 2005; received in revised form 28 March 2005; accepted 6 April 2005 Available online 9 June 2005

Abstract The flushing potential of a desert loess soil contaminated by the flame retardant Tetrabromobisphenol A (TBBPA), chloride (Cl
) and bromide (Br
) was studied in undisturbed laboratory column experiments (20 cm diameter, 45 cm long) and a small field plot (2 · 2 m). While the soluble inorganic ions (Cl
and Br
) were efficiently flushed from the soil profile after less than three pore volumes (PV) of water, about 50% of the initial amount of TBBPA in the soil was also flushed, despite its hydrophobic nature. TBBPA leaching was made possible due to a significant increase in the pH of the soil solution from 7.5 to 9, which increased TBBPA aqueous solubility. The remaining TBBPA mass in the soil was not mobilized from its initial location in the topsoil due to the decrease in pH at this horizon. In situ soil flushing demonstrated that this method is a feasible treatment for reducing soil contamination at this site.
2005 Elsevier Ltd. All rights reserved. Keywords: Tetrabromobisphenol A; Loess; Leaching; Ionizable organic compounds

1. Introduction In recent years, inappropriate waste disposal has become a major environmental problem that requires cost-effective remediation solutions. The ability of contaminants to reach groundwater and put water resources at risk lies in their inherent physico-chemical properties, as well as their resistance to removal by complex physi* Corresponding author. Address: Department of Civil and Environmental Engineering, Northwestern University, 2145 Sheridan Road, Evanston, IL 60208-3109, USA. Tel.: +1 847 467 4980; fax: +1 847 491 4011. E-mail addresses: [email protected] (S. Arnon), [email protected] (Z. Ronen), [email protected]. ac.il (A. Yakirevich), [email protected] (E. Adar).

cal, chemical and biological reactions in unsaturated and saturated geological media (Mackay et al., 1985; SaintFort, 1991; Schwarzenbach et al., 1993). Applying water, with or without additives, is frequently used for soil flushing in order to clean up the vadose zone (Thomsen et al., 1989; MacKay et al., 1996). In situ soil flushing can be limited due to several reasons, primarily: (1) soils that have been contaminated for a long period of time exhibit a bi-phasic pattern of desorption with an initial fast stage (min–h) and a subsequent longer slow phase (days–years) (Pavlostathis and Jaglal, 1991); (2) the heterogeneous nature of soil implies that regions with low hydraulic conductivity may exist, from which contaminants must diffuse over a long period of time to the more effectively flushed layers (MacKay et al., 1996); and (3) the slow solubility rates of hydrophobic

0045-6535/$ - see front matter
2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2005.04.050

18

S. Arnon et al. / Chemosphere 62 (2006) 17–25

compounds further increase the flushing duration. Identifying which process limits the clean-up of a particular site may assist in suggesting approaches for enhancing the clean-up. For example, MacKay et al. (1996) concluded that the flushing pattern is highly dependent on the desorption and solubility rates of the different tested compounds (benzene, toluene, ethylbenzene and xylene). They also found that the presence of a non-aqueous phase liquid would substantially increase the siteÕs estimated flushing duration. Thomsen et al. (1989) demonstrated efficient cleaning of the unsaturated zone from volatile organic compounds by using in situ soil flushing (with water), groundwater pumping, ex situ treatment and recharge. In many other cases the use of specific solutions has been suggested, such as water with surfactants or organic solvents, rather than clean water, to enhance contaminant flushing (e.g., Bettahar et al., 1999; Di Palma, 2003). Nevertheless, the use of water for flushing, without additives, reduces operation activities and costs. In this study, we assess the flushing potential of undisturbed contaminated desert soil at the laboratory and small field scales, as an optional method for soil remediation. The site, 25 acres of desert loess soil, 0.1– 2 m thick, overlying fractured chalk bedrock, was contaminated in the late 1980s as a result of industrial wastewater disposal in a forced evaporation facility (Nativ et al., 1999). The inorganic contamination was dominated by Cl
and Br
ions and the major organic contaminant was tetrabromobisphenol A (TBBPA). Although soil flushing might increase the risk of groundwater contamination, when combined with additional groundwater treatment it can provide an inexpensive method for clean-up of contaminated soils (Thomsen et al., 1989; MacKay et al., 1996). An underground drain, a pumping station and a treatment facility were already in place at the downstream edge of this site (Nativ et al., 2003); therefore soil flushing was proposed as a solution to clean up the contaminated soil. TBBPA (Fig. 1) is used as a flame retardant in electronic circuit boards and in the plastics industry (de Wit, 2002). It is of concern since it was found in stream sediments and municipal wastewater, which also makes it a potential groundwater contaminant (Sellstrom and Jansson, 1995; Ronen and Abeliovich, 2000; Oberg

Br

Br OH

C CH3 Br

Br

Fig. 1. Chemical structure of TBBPA.

K d ¼ foc bðk ow Þa

ð1Þ

where a and b are coefficients peculiar to the organic compounds under consideration. For example, Schellenberg et al. (1984) found from their experiment with chlorinated phenols that the values of a and b are 0.82 and 1.05, respectively. They also stated that if the pH of the solution is not more than one unit above the acid dissociation constant (pKa) of the compound, the contribution of the phenolate ion to the overall adsorption of the phenolic compound could be neglected. For pHs higher than one unit above the pKa, the aforementioned adsorption model (1) failed to describe phenolate adsorption, as was shown by Fiore et al. (2003) and by Shimizu et al. (1992). The latter have demonstrated that the mineral constituents (mainly clay content) of the soils controlled the adsorption of the ionic form of pentachlorophenol.

2. Materials and methods

CH3 HO

et al., 2002). In some cases, increased levels of TBBPA were linked to industrial sources (Oberg et al., 2002). There is evidence that prolonged exposure of rats to TBBPA disturbs the liver heme metabolism (Szymanska et al., 2000), as well as the neural system (Eriksson et al., 1998). In vitro assays demonstrated that TBBPA was up to 25 times more potent in binding to human transthyretin (thyroid transport protein) than thyroxin (the native hormone) (Brouwer, 1998). TBBPA aqueous solubility, sorption and bioavailability are pH dependent since TBBPA can be partially ionized (Lee et al., 1991). TBBPA is not easily biodegraded and only recently was its mineralization demonstrated through dehalogenation under anaerobic conditions and further biodegradation of bisphenol A under aerobic conditions (Ronen and Abeliovich, 2000; Voordeckers et al., 2002; Arbeli and Ronen, 2003). Ionized compounds such as TBBPA exhibit complex transport behavior, as some of their physico-chemical properties are pH dependent. For example, Schellenberg et al. (1984) demonstrated that the equilibrium partition coefficient (Kd) of chlorinated phenols between the sorbent and the solution could be estimated based upon their lipophilicity, as expressed by the octanol/water partition coefficient (Kow), and on the organic carbon contents (foc) of the sorbent. Mathematical relationships between Kd, Kow and foc have been derived for various sets of compounds and natural sorbents:

2.1. Sample collection and soil properties The soil monoliths were collected from a contaminated site located in the northern Negev desert, Israel. The sampling technique was designed to collect soil cores, 10 cm in diameter and 45 cm long, in an attempt

S. Arnon et al. / Chemosphere 62 (2006) 17–25

to represent the soil structure at the field site. The columns consisted of a PVC cylinder with a removable sharpened steel edge attached to its base. As the sharpened column edge was pressed down by weight, shovels were used to remove the soil around the exterior of the steel edge to lower the soil resistance and compaction near the monolith edge. At all times, 5 cm of soil remained constantly around the column edge to prevent separation of the soil and the steel edge. When the column was filled, the steel edge was removed, and a 2 cm layer of glass fibers was positioned at both core ends to prevent solid particles from being washed out of the column during the flushing experiment. The column was then sealed with caps and transported to the laboratory. Soil samples adjacent to the sharpened edge were taken during the collection of the monoliths for assessment of the initial contaminant concentrations and soil properties. Selected soil properties appear in Table 1. The major soil constituent was quartz (>50%). Other minerals that were found in the soil, in order of abundance, are: calcite, feldspar, gypsum, kaolinite and illite, as determined by X-ray diffraction analysis (Philips XRD diffractometer). Halite was found only in the upper section of the soil. Soil particle distribution was measured using a laser diffraction system (Malvern mastersizer). Specific surface area was also measured using the Malvern mastersizer. Natural organic matter was measured using the dichromate oxidation method (Lowell, 1993). Bulk density was found by measuring the dry mass of a known structured sample with a defined volume, and porosity was estimated using the bulk density and particle density (taken as 2.65 gcm
3) (Lowell, 1993). 2.2. Batch desorption kinetics Five grams (dry weight) of contaminated sieved soil (<2 mm) were mixed with 15 ml of double distilled water inside 20 ml glass vials with Teflon screw caps (24 samples). The vials were shaken at 200 rpm at 25 C. At each time interval, two vials were removed and centrifuged

Table 1 Soil properties Parameter Sand % Silt % Clay % Specific surface Natural organic carbon content Porosity Bulk density

Value (50–2000 lm) (2–50 lm) (<2 lm) (m2 g
1) (%)

13 ± 3 63 ± 4 25 ± 6 3.9 ± 1.3 0.14 ± 0.03

(g cm
3)

0.4 ± 0.02 1.54

The averaged results were calculated from soil samples collected at depths of 0–50 cm.

19

for 15 min at 3000 rpm to separate the supernatant and the soil. TBBPA concentrations were quantified at 14 time intervals, not equally distributed, over 100 h (0.016, 0.083, 0.25, 0.5, 0.75, 1, 1.5, 2, 3, 5, 13, 24, 50 and 100 h). Biodegradation of TBBPA was not prevented in this experiment, either by soil sterilization or by biocide addition, as it was previously shown that TBBPA is not degraded in this soil under aerobic conditions (Ronen and Abeliovich, 2000). 2.3. Soil extracts Soil extractions were performed for measurement of the major ions, TOC (total organic carbon) and pH by mixing air-dried sieved soil (<2 mm) and double distilled water at a 1:1 weight ratio (40 g) inside 125 ml flasks (triplicates). The soil–water mixtures were shaken at 200 rpm at 25 C for 24 h and filtered through GF/C filters (Whatman) prior to the analysis. TBBPA was extracted from contaminated sieved soil (<2 mm) according to method no. 3550 (EPA, 1997). Ten grams of contaminated soil and 50 ml of ethyl-acetate were added to 250 ml flasks equipped with Teflon screw caps (triplicates). The mixtures were sonicated for 15 min and shaken at 200 rpm at 25 C for 12 h. The solution was separated from the soil by filtration through GF/C filters (Whatman) and concentrated via evaporation to 1 ml. O-hydroxybiphenyl (97%, Aldrich) was added to the soil before extraction, as an internal standard for the assessment of the TBBPA recovery. 2.4. Batch solubility test TBBPA solubility was examined under a pH range of 7–9. Five different buffer solutions were prepared from Tris-HCl 0.1 M with pH values of 7, 7.5, 8, 8.5 and 9. TBBPA in acetone was added to 20 ml glass vials equipped with Teflon screw caps in quantities which, after acetone evaporation and addition of 10 ml of buffer solutions and upon complete dissolution, yielded maximum concentrations of 100, 200, 300, 400 and 500 mg l
1 (i.e. each concentration was examined at five pH values). Duplicate samples were shaken at 200 rpm at 25 C for 10 h. The 10 ml solutions were filtered (25 mm; 0.45 lm; Gelman) and the last 1 ml was taken for analysis by high-performance liquid chromatography (HPLC), after preliminary experiments verified that losses of TBBPA by sorption onto the filter in this procedure are negligible. 2.5. Column experiments Soil flushing with tap water was investigated during column experiments using three undisturbed soil cores. The composition of the tap water used for flushing was as follows (mg l
1): 68.1 (Ca2+), 40.5 (Mg2+), 120

S. Arnon et al. / Chemosphere 62 (2006) 17–25

(Na+), 5.35 (K+), 220 (Cl
), 41.4 (SO2
4 ) and pH of 7.5. The columns were placed upside down and flow was in an upward direction (from the surface soil into the soil profile), to diminish air trapping. The soil was saturated by introducing water in the direction of flow under relatively small head differences (2–4 cm). Water level was increased based on the predicted movement of the water inside the column (based on the hydraulic conductivity of this soil), until water exited the column. During the flushing experiments, a constant head controlled the flow, and DarcyÕs law was used to calculate the hydraulic conductivity based on timed collections of fixed volumes of water under three different hydraulic gradients. The effluent was analyzed for Cl
, Br
and TBBPA concentrations until TBBPA concentration fell below the detection limit. The pH, EC (electrical conductivity) and TOC were also measured. 2.6. In situ soil flushing A ‘‘double ring-like’’ infiltration pond was constructed at the contaminated site. The dimension of the outer pond was 2 · 2 m, and the inner pond was 1.2 · 1.2 m. Both ponds were filled with tap water gradually over 4 h, maintaining the same levels at both sections, until the water level reached 45 cm above the surface. This level was kept constant for two more hours and then the experiment was terminated (i.e. water was drained). EC and TOC profiles in the soil were measured by analyzing soil solution extracts using the same method discussed earlier. 2.7. Chemical analyses The following measurements were carried out in the water samples: EC, pH and TOC with a Dohrmann DC-190 (Rosemount). Cl
and Br
were measured with an ion chromatograph (Dionex, 4500i). TBBPA concentrations in the soil (extracted by ethyl-acetate) were analyzed with a gas chromatograph (GC). The GC in use was an HP 5490 equipped with a 15 m capillary column (SPB5), 0.53 mm inner diameter and 1.5 lm film thickness (Supelco). The temperature program began at 100 C for 4 min, followed by a ramp at 10 C min
1 until 280 C, where the temperature was held for 8 min. Helium was used as a carrier gas at 12 ml min
1. The compounds were detected with a flame ionization detector at a temperature of 250 C. TBBPA concentrations were corrected according to the internal standard recoveries. Error in analysis was estimated at ±5% and a concentration of 1 mg l
1 was considered the low detection limit based on the signal to noise ratio. TBBPA concentrations from the desorption and solubility experiments were analyzed with HPLC (Kontron), equipped with a 4.6 mm by 25 cm C18 column (Supelcosil), and with UV detection (Diode array 440)

at a wavelength of 290 nm. The mobile phase consisted of two solutions: (A) methanol and acetic acid (1%) and (B) ammonium acetate 0.018 M-H2O with acetic acid (1%). The initial A:B ratio was 60%:40%. The A:B ratio was changed at a constant rate during 13 min until 100% A was reached. The final stage was maintained for 1 min. Error in analysis was estimated at ±2% and a concentration of 0.2 mg l
1 was considered the low detection limit based on the signal to noise ratio. All values in the chemical analysis are given as mean ± standard deviation (SD).

3. Results and Discussion 3.1. TBBPA solubility The effect of the solution pH on TBBPA solubility was investigated at a variety of pHs and initial TBBPA masses (Fig. 2). Below pH 7.5 TBBPA was not detected (concentrations <0.2 mg l
1, i.e. below the detection limit), whilst above pH 8 there was a sharp increase in its solubility, reaching complete solubility of the compound at pH 9 (up to 500 mg l
1). In addition, the soluble fraction of TBBPA increased with higher initial TBBPA input under the entire examined pH range (7–9). When the pH is increased by one unit above the pKa, 90% of the hydrophobic ionizable organic compounds, such as TBBPA, is in its ionized form and its solubility in water is expected to rise by several orders of magnitude compared to its solubility at pH values below the pKa (Schwarzenbach et al., 1993). Recently, Eriksson et al. (2004) reported that although TBBPA has two hydroxy-groups, the differences in the pKaÕs (7.5 and 8.5, WHO, 1995) is relatively small, indicating that the two rings may function, at least in part, independently of each other. In another report that was recently submitted to the EPA by the Brominated Flame Retardant

100

Soluble fraction (%)

20

Initial concentration 100 mg/l 200 mg/l 300 mg/l 400 mg/l 500 mg/l

80 60 40 20 0 7

7.5

8

8.5

9

pH Fig. 2. The effect of pH on TBBPA solubility. The relative amount of soluble TBBPA is displayed (soluble fraction) at five different initial concentrations. The results are averages ± the SDs of duplicates.

S. Arnon et al. / Chemosphere 62 (2006) 17–25

Industry Panel (BFRIP) (BFRIP, 2004), the pKa of TBBPA was reported to be 9.4. While the pKaÕs of TBBPA in the aforementioned studies (BFRIP, 2004; Eriksson et al., 2004) differ significantly from each other, our experimental results suggest that the values reported by Eriksson et al. (2004) better describe the behavior of TBBPA, since TBBPA concentrations increase significantly above pH 8 (up to 500 mg l
1). The BFRIP report also states that the solubility of TBBPA at pHs 5, 7, and 9 was 0.148, 1.26 and 2.34 mg l
1, respectively. These results also differ from our results shown in Fig. 2. One possible reason for the difference between our results and those reported by the BFRIP might lie in the specific details of the experimental procedure (e.g., what is the maximum TBBPA concentration that might be expected from the amount of TBBPA that was added to the water). 3.2. TBBPA desorption

90 min, was much slower. The fast increase in TBBPA also implies that if dissolution is taking place, its kinetics do not have a major impact on the release of TBBPA into the solution. The second sample (after 2 min) was exceptional since it did not follow the general smooth increase as expected from a desorption experiment and as seen for the rest of the data. Given that we used 24 replicates, in which duplicates were sacrificed at each time measurement, it might be that some heterogeneity was apparent despite the initial soil mixing, causing the non-smooth increase in the first 20 min. 3.3. Soil flushing The breakthrough curves (BTCs) of Cl
, Br
and TBBPA during the soil flushing are presented in Fig. 4. Since similar observations were documented for all tested cores (Arnon, 1996), the results presented in Figs. 4–7 were obtained only from column CL2, to avoid redundancy. High Cl
and Br
concentrations (accounting for more than 95% of the water EC) were observed soon after the water started to leach out from the column, while maximum concentrations appeared simultaneously after 0.42 pore volumes (PV). The saturation procedure of the initially dried soil caused high solute concentrations soon after water started to flow out (Fig. 4). Since the water for saturation was applied in the direction of the flow, the first PV that was introduced in order to saturate the column leached considerable amounts of the contaminants toward the outlet. Nevertheless, the volume of water that was used for saturation is not shown in Fig. 4, since the PV was calculated based on the accumulated effluent volume (and soil porosity data). The transport of TBBPA was slow compared with that of Cl
and Br
, with calculation of a retardation factor based on the relative timings of Cl
, Br
and TBBPA maximum concentrations. The averaged retardation factor was 3.2 (±0.6), as calculated using data from all three tested columns. After approximately

1 300

0.8 TBBPA

0.4 0.2 0

250

TBBPA (mg/l)

0.6

0

20

40

60

80

Time (min)

16000 TBBPA ClBr-

200 150

12000 8000

100 4000

50 0

0

1

2

3

4

5

Cl- and Br- (mg/l)

Relative concentration

The kinetics of TBBPA emergence in the solution, integrating both the release from the sorbent (desorption) and the dissolution (if TBBPA is in solid form) as a single process, are illustrated in Fig. 3. During the first five min of the experiment the concentrations reached 70–80% of the maximum released concentration, observed 90 min after the start of the experiment. Over the next 4 d, the concentrations did not differ from the value observed after 90 min (data not shown). The solution pH increased after 5 min from 7.5 to 8.5 and remained constant until the end of the experiment. When desorption controls the appearance of a compound in a soil–water solution, we usually observe bi-phasic behavior involving an initial fast rise and a second slower rise (e.g., Pavlostathis and Jaglal, 1991; Opdyke and Loehr, 1999). This behavior can be seen clearly in Fig. 3, where the initial fast increase was in the order of min and a further increase, up to the maximum concentration after

21

0

Pore volume Fig. 3. TBBPA desorption kinetics from the contaminated soil. The concentrations are normalized to the maximum observed concentration (15 mg l
1, after 90 min). The results are averages ± the SDs of duplicates.







Fig. 4. BTCs of Cl , Br and TBBPA during the continuous flushing of initially dry, undisturbed contaminated soil columns using tap water.

22

S. Arnon et al. / Chemosphere 62 (2006) 17–25 0

a

b

Depth (cm)

10 20 Cl- - before flushing

30

Cl- - after flushing

Before flushing After flushing

Br- - before flushing Br- - after flushing

40 0

1000

2000

3000

4000

5000

0

50

100

Concentration (mg/kgdry soil)

150

200

250

300

350

400

TBBPA (mg/kgdry soil)

100

Table 2 Hydraulic conditions during the soil flushing

9

80 8.5

60 pH TBBPA

40

pH

TBBPA relative mass removal (%)

Fig. 5. Soil concentration profiles of Cl
, Br
(a) and TBBPA (b) before and after the continuous flushing of an undisturbed contaminated soil column (Fig. 3). Each profile was constructed by five measurements evenly distributed along the profile. The results are averages ± the SDs of triplicates.

8

20 0 0

0.5

1

1.5

2

2.5

3

Pore volume

Hydraulic Flux gradient (ml h
1)

0–0.92 0.92–1.69 1.69–6

0.22 0.44 0.66

Hydraulic conductivity (cm s
1)

23.1–100 1.15 · 10
3– 4.3 · 10
3 19.1– 43.8 4.78 · 10
4– 1.09 · 10
3 31.8– 46.9 5.29 · 10
4–7.81 · 10
4

7.5 3.5

Pore volume Fig. 6. The change in effluent pH along with the TBBPA mass that was flushed out of the column. TBBPA mass is calculated relative to its initial amount in the soil. The results are averages ± the SDs of triplicates.

0

The fluxes, hydraulic gradients and hydraulic conductivity during the flushing experiment are summarized in Table 2. The flow rate through the columns was not constant as the hydraulic conductivity decreased by a factor of 4, from 4 · 10
3 to 1 · 10
3 cm s
1 during the first PV (Table 2), probably due to swelling of clay minerals. 3.4. Soil flushing efficiency

Before flushing After flushing

Depth (cm)

10 20 30 40 7

7.5

8

8.5

9

9.5

10

pH Fig. 7. Soil extract pH before and after the continuous flushing of an undisturbed contaminated soil column (Fig. 3). Each profile was constructed from five measurements evenly distributed along the profile. The results are averages ± the SDs of triplicates.

5 PV, the TBBPA concentration declined to below its detection limit (0.2 mg l
1) and the experiment was terminated.

The concentrations of Cl
, Br
and TBBPA in the soil before and after the flushing are illustrated in Fig. 5. Most of the initial contaminant mass was concentrated in the upper part of the soil. This phenomenon is typical for arid climates where evaporation is dominant, rather than infiltration and deep percolation (Allison and Hughes, 1983; Scanlon, 1991). Fig. 5a shows that Cl
and Br
were completely flushed from the soil profile. The average Cl
concentration in the soil profile after the flushing was 56 mg kg
1, since the injected water contained 220 mg l
1 of Cl
. Conversely, TBBPA was only partly flushed from the soil (Fig. 5b), with the remaining mass being retained in the topsoil. The mass recoveries of TBBPA, Cl
and Br
for the three tested columns are summarized in Table 3. Mass balance calculations were made according to:   F l
In Mass balance ð%Þ ¼  100 ð2Þ Bf
Af

S. Arnon et al. / Chemosphere 62 (2006) 17–25

23

Table 3 Mass balance for Cl
, Br
and TBBPA during the column experiments Column number

CL2

CL3

Contaminant

Cl







Br

Mass balance (%) Retained TBBPA in the soil (% of the initial mass)

103

84

CR3

TBBPA




Cl

Br

TBBPA

Cl


Br


TBBPA

110 12

105

88

116 17

109

81

114 42

where Fl is the mass that was leached out of the column (M), In is the mass of the solute that was introduced by the water (relevant only for chloride) (M), Bf is the mass in the soil before flushing (M) and Af is the mass in the soil after flushing (M). None of the above mass balance calculations exceeds ±20% from the ideal mass recovery. Error in mass balance was probably due to the non-dense sampling in the upper part of the soil where most of the mass was concentrated. Imperfect soil extraction may also have contributed to the error. Nevertheless, the results presented in Fig. 5 and Table 3 clearly illustrates the efficient salt flushing and the incomplete TBBPA flushing. 3.5. The effect of pH on the flushing of TBBPA The effluent pH rose above 8 after approximately 1.2 PV; at the same time a sharp increase in TBBPA mass removal was observed (Fig. 6). Despite the fact that the effluent pH reached values between 8.5 and 9, a significant fraction of TBBPA was retained in the topsoil (Fig. 5). The profound effect of pH on TBBPA solubility (Fig. 2) and desorption (Fig. 3) was discussed earlier in the text. This is also the reason for the incomplete flushing of TBBPA during the column experiments, which is further illustrated in Fig. 7. Fig. 7 shows the pH of the soil extract at the beginning and end of the column experiments. The pH was slightly reduced in the upper section of the soil and significantly increased in the regions closer to the outlet at the end of the experi-




ments. The reduced potential for pH increase in the topsoil caused incomplete dissolution of TBBPA and was the reason why significant mass of TBBPA was not mobilized. Another major factor that may affect TBBPA transport is adsorption to the soil and natural organic carbon. The Kd of TBBPA on the loess soil (at pH 8.5) was found to be 0.09 ml g
1 (Arnon, 1996). This value alone could not account for the delayed BTC of TBBPA as compared with that of Cl
and Br
(Fig. 4). The Kd of TBBPA is affected by its ionic and non-ionic species as discussed previously with respect to other ionized organic compounds (Lee et al., 1991; Shimizu et al., 1992; Fiore et al., 2003). As the pH increases we expect the Kd to be less dependent on the natural organic content of the soil (Schellenberg et al., 1984; Lee et al., 1991) and to be more influenced by the mineralogy of the soil, mainly by the clay content (Shimizu et al., 1992). Thus, we suggest that pH changes in the soil columns with time affected the Kd and sorption of TBBPA and thus played a major role in controlling its transport. However, the lack of information on the spatial and temporal pH values prevents a quantitative assessment of the relative role of Kd on TBBPA transport. 3.6. Soil flushing at the field site The flushing potential was also examined at a contaminated field site from which the columns for the laboratory experiments were taken. In this experiment,

0

Depth (cm)

b

a

Before flushing

20

After flushing

40 60 80 Before flushing

100

After flushing

120 0

5

10

EC (mS/cm)

15

10

20

30

40

50

60

70

Concentration (mg/Kgdry soil)

Fig. 8. Soil EC (a) and TOC (b) profiles before and after the continuous flushing of a pilot scale (2 · 2 m), undisturbed contaminated soil. Each profile is generated from two separate soil cores taken at the outer ring (before flushing) and the inner ring (after flushing). The results are averages ± the SDs of duplicates.

24

S. Arnon et al. / Chemosphere 62 (2006) 17–25

TOC was monitored as a marker of organic contaminant (mainly TBBPA) concentrations, while the EC represented the soluble salt (Cl
and Br
) concentrations. Both tracers, TOC and EC, exhibited linear correlations to TBBPA and to Cl
+ Br
(R2 = 0.8 and 0.94, respectively), as measured during the laboratory experiments. Fig. 8 shows the EC and TOC profiles in the soil before and after flushing. No comprehensive effort was made to deal with the spatial distribution of the soil contaminants. Therefore, the results in Fig. 8 should be treated as qualitative, expressing mainly the major patterns of the contaminantsÕ distribution in the soil. The amount of water that percolated into the soil was equal to 1 PV of a soil section up to a depth of 0.9 m. This was calculated based on the volume of water that percolated through the inner pond (1.2 · 1.2 m) and a porosity of 0.4 (Table 1). The salts front (EC) leached into the soil profile up to 90 cm, leaving the topsoil almost completely leached (Fig. 8a). The calculated PV and the depth of the front verify that the salts were efficiently flushed from the soil, as observed in the laboratory studies. The TOC profile suggests that a major fraction of the organic contaminants still remained in the topsoil (Fig. 8b), similar to the laboratory experiments (Fig. 5). Nevertheless, a significant amount of TOC was still abundant in the entire soil profile as a result of the relatively small volume of water that was used (1 PV).

4. Summary and conclusions The results of this study suggest that soil flushing can be efficiently applied to remediate contaminated desert soils. The laboratory experiments show that complete removal of Cl
and Br
from the soil required about 3 PV, while at the same time more than 50% of TBBPA initial mass was also removed. Nevertheless, after 5 PV, TBBPA leaching became inefficient due to inability to increase the pH to the value needed for dissociation of TBBPA in the topsoil. Thus, the remaining TBBPA mass was sustained in the uppermost part of the soil profile rather than being leached to the lower parts of the profile. An artificial increase in the flushing solution pH has the potential to increase the efficiency of this treatment. A pilot scale field experiment demonstrated that in situ soil flushing is a feasible treatment for reducing contamination levels in desert loess soils, when accompanied by groundwater extraction and treatment (‘‘pump & treat’’). Since saline soils are ubiquitous in arid zones it is expected that high pH will be developed upon wetting. (Hillel, 1998), which is favorable for the flushing of ionogenic compounds. Nevertheless, further study is needed, mainly on the two-dimensional flow patterns

that are expected to develop when irrigating with water on a highly saline soil such as loess.

Acknowledgements We thank Zoe¨ Grabinar for editorial assistance and Dr. Helen Graber for constructive comments.

References Allison, G.B., Hughes, M.W., 1983. The use of natural tracers as indicators of soil–water movement in a temperate semiarid region. J. Hydrol. 60, 157–173. Arbeli, Z., Ronen, Z., 2003. Enrichment of a microbial culture capable of reductive debromination of the flame retardant tetrabromobisphenol-A, and identification of the intermediate metabolites produced in the process. Biodegradation 14, 385–395. Arnon, S., 1996. Transport of organic and inorganic contaminants in desert soil—evaluation of contaminants flushing potential from a contaminated soil near Ramat–Hovav industrial park (in Hebrew). M.Sc. Thesis, Ben Gurion University of the Negev, Beer–Sheva. Bettahar, M., Ducreux, J., Schafer, G., Van Dorpe, F., 1999. Surfactant enhanced in situ remediation of LNAPL contaminated aquifers: large scale studies on a controlled experimental site. Transp. Porous Media 37, 255–276. BFRIP, 2004. HPV data summary and test plan for phenol, 4,4 0 -isopropylidenbis[2,6-dibromo-(TETRABROMOBISPHENOL A, TBBPA). EPA report number: 201-15678. Available from: . Brouwer, A., 1998. Structure-dependent multiple interactions of polyhalogenated aromatic hydrocarbons with the thyroid hormone system. Organohalogen Compd. 37, 225–228. de Wit, C.A., 2002. An overview of brominated flame retardants in the environment. Chemosphere 46, 583– 624. Di Palma, L., 2003. Experimental assessment of a process for the remediation of organophosphorous pesticides contaminated soils through in situ soil flushing and hydrolysis. Water Air Soil Pollut. 143, 301–314. EPA, 1997. Test methods for evaluating solid waste, physical/ chemical methods. Method no. 3550 (CD-ROM version 2), Springfield, VA. Eriksson, J., Rahm, S., Green, N., Bergman, A., Jakobsson, E., 2004. Photochemical transformations of tetrabromobisphenol A and related phenols in water. Chemosphere 54, 117– 126. Eriksson, P., Jakobsson, E., Fredriksson, A., 1998. Developmental neurotoxicity of brominated flame-retardants, polybrominated diphenyl ethers, and tetrabromo-bis-phenol A. Organohalogen Compd. 35, 375–377. Fiore, S., Zanetti, M.C., Genon, G., 2003. Experimental study of the pH influence on the transport mechanisms of phenols in soil. Ann. Chim. 93, 595–605. Hillel, D., 1998. Environmental Soil Physics. Academic Press, San Diego, Ca.

S. Arnon et al. / Chemosphere 62 (2006) 17–25 Lee, L.S., Rao, P.S.C., Brusseau, M.L., 1991. Nonequilibrium sorption and transport of neutral and ionized chlorophenols. Environ. Sci. Technol. 25, 722–729. Lowell, D.L., 1993. Soil Science: Methods and Applications. Longman Scientific & Technical, Singapore. MacKay, A.A., Chin, Y.-P., MacFarlane, J.K., Gschwend, P.M., 1996. Laboratory assessment of BTEX soil flushing. Environ. Sci. Technol. 30, 3223–3231. Mackay, D.M., Roberts, P.V., Cherry, J.A., 1985. Transport of organic contaminants in groundwater. Environ. Sci. Technol. 19, 384–392. Nativ, R., Adar, E., Assaf, L., Nygaard, E., 2003. Characterization of the hydraulic properties of fractures in chalk. Ground Water 41 (4), 532–543. Nativ, R., Adar, E.M., Becker, A., 1999. Designing a monitoring network for contaminated ground water in fractured chalk. Ground Water 37, 38–47. Oberg, K., Warman, K., Oberg, T., 2002. Distribution and levels of brominated flame retardants in sewage sludge. Chemosphere 48, 805–809. Opdyke, D.R., Loehr, R.C., 1999. Determination of chemical release rates from soils: experimental design. Environ. Sci. Technol. 33, 1193–1199. Pavlostathis, S.G., Jaglal, K., 1991. Desorptive behavior of trichloroethylene in contaminated soil. Environ. Sci. Technol. 25, 274–279. Ronen, Z., Abeliovich, A., 2000. Anaerobic–aerobic process for microbial degradation of tetrabromobisphenol A. Appl. Environ. Microbiol. 66, 2372–2377. Saint-Fort, R., 1991. Groundwater contamination by anthropogenic organic compounds from waste disposal sites:

25

transformations and behavior. J. Environ. Sci. Health. A 26, 13–62. Scanlon, B.R., 1991. Evaluation of moisture flux from chloride data in desert soils. J. Hydrol. 128, 137–156. Schellenberg, K., Leuenberger, C., Schwarzenbach, R.P., 1984. Sorption of chlorinated phenols by natural sediments and aquifer materials. Environ. Sci. Technol. 18, 652–657. Schwarzenbach, R.P., Gschwend, P.M., Imboden, D.M., 1993. Environmental Organic Chemistry. John Wiley & Sons, New-York. Sellstrom, U., Jansson, B., 1995. Analysis of tetrabromobisphenol a in a product and environmental-samples. Chemosphere 31, 3085–3092. Shimizu, Y., Yamazaki, S., Terashima, Y., 1992. Sorption of anionic pentachlorophenol (PCP) in aquatic environments: the effect of pH. Water Sci. Technol. 25, 41–48. Szymanska, J.A., Piotrowski, J.K., Frydrych, B., 2000. Hepatotoxicity of tetrabromobisphenol-A: effects of repeated dosage in rats. Toxicology 142, 87–95. Thomsen, K.O., Chaudhry, M.A., Dovantzis, K., Riesing, R.R., 1989. Groundwater remediation using an extraction, treatment, and recharge system. Ground Water Monitor. Rev. 9, 92–99. Voordeckers, J.W., Fennell, D.E., Jones, K., Haggblom Max, M., 2002. Anaerobic biotransformation of tetrabromobisphenol A, tetrachlorobisphenol A, and bisphenol A in estuarine sediments. Environ. Sci. Technol. 36, 696–701. WHO, 1995. Environmental Health Criteria 172. Tetrabromobisphenol A and derivatives International Program on Chemical Safety. World Health Organization, Geneva, Switzerland.