Influence of soil properties on vapor-phase sorption of trichloroethylene

Journal of Hazardous Materials 306 (2016) 34–40

Contents lists available at ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Influence of soil properties on vapor-phase sorption of trichloroethylene Dawit N. Bekele a,b , Ravi Naidu a,b,∗ , Sreenivasulu Chadalavada a,b a

Global Center for Environmental Remediation, University of Newcastle, Callaghan, NSW 2308, Australia CRC for Contamination Assessment & Remediation of the Environment, Building X (Environmental Sciences Building), University of South Australia, Mawson Lakes, SA 5095, Australia b

h i g h l i g h t s • • • • •

Vapor intrusion is a major exposure pathway for volatile hydrocarbons. Certainty in transport processes enhances vapor intrusion model precision. Detailed understanding of vadose zone vapor transport processes save resources. Vapor sorption near-steady-state conditions at sites may take months or years. Type of clay fractions equitably affects sorption of trichloroethylene vapor.

a r t i c l e

i n f o

Article history: Received 24 August 2015 Received in revised form 29 October 2015 Accepted 2 December 2015 Available online 4 December 2015 Keywords: Volatile organic hydrocarbon Trichloroethylene Sorption Vapor intrusion model

a b s t r a c t Current practices in health risk assessment from vapor intrusion (VI) using mathematical models are based on assumptions that the subsurface sorption equilibrium is attained. The time required for sorption to reach near-steady-state conditions at sites may take months or years to achieve. This study investigated the vapor phase attenuation of trichloroethylene (TCE) in five soils varying widely in clay and organic matter content using repacked columns. The primary indicators of TCE sorption were vapor retardation rate (Rt ), the time required for the TCE vapor to pass through the soil column, and specific volume of retention (VR ), and total volume of TCE retained in soil. Results show TCE vapor retardation is mainly due to the rapid partitioning of the compound to SOM. However, the specific volume of retention of clayey soils with secondary mineral particles was higher. Linear regression analyses of the SOM and clay fraction with VR show that a unit increase in clay fraction results in higher sorption of TCE (VR ) than the SOM. However, partitioning of TCE vapor was not consistent with the samples’ surface areas but was mainly a function of the type of secondary minerals present in soils. © 2015 Elsevier B.V. All rights reserved.

1. Introduction

Abbreviations: Ca , soil air permeability, cm/s; Cw , soil saturated hydraulic conductivity, cm/s; CRCCARE, cooperative research centre for contamination assessment & remediation of the environment; ECD, electronic capture detector; GC, gas-chromatograph; IGC, inverse gas chromatography; LOR, limit of reporting; Rt , retardation rate; SOM, soil organic matter; VI, vapor intrusion; VIM, vapor intrusion model; VR , specific volume of retention; VOC, volatile organic compound; w , dynamic viscosity of water, centipoise (cP); a , dynamic viscosity of air, centipoise (cP); w , density of water, kg/m3; a , density of air, kg/m3. ∗ Corresponding author at: Global Center for Environmental Remediation, University of Newcastle, Callaghan, NSW 2308, Australia Fax: +61 8 8302 3124. E-mail addresses: [email protected], [email protected] (R. Naidu). http://dx.doi.org/10.1016/j.jhazmat.2015.12.002 0304-3894/© 2015 Elsevier B.V. All rights reserved.

The distribution of volatile organic compounds (VOCs) between soils and sediments and the soil solution aqueous phase underpins modeling of the fate and transport of such substances in the subsurface environment [1]. Sorption of VOCs onto the soil matrix as they migrate upwards from subsurface contaminated sources results in strong binding to SOM, and therefore retarding the transport of vapor to the ground surface [2–4]. Ignoring attenuation of vapor whether via microbial degradation or sorption processes can lead to incorrect predictions about the vapor transport and potential risk it poses to receptor. This may well lead to the wrong clean-up technology being chosen [4]. Current vapor intrusion models (VIMs) assume that the vapor source has been in place long enough to fully develop near-steady-

D.N. Bekele et al. / Journal of Hazardous Materials 306 (2016) 34–40

state conditions and sorption is assumed to occur instantaneously for modeling purposes. Therefore models consider no attenuation of vapor due to partitioning [5]. However, vapor phase partitioning could be significant at the contaminated subsurface with a finite source. VIMs such as that of Johnson and Ettinger [6] and BioVapor [7] are used by environmental agencies and practitioners to evaluate the health risks arising from vapor intrusion (VI). These models, however, assume steady-state conditions and do not consider vapor phase attenuation [4] and as a consequence prediction of risks when in reality vapor emanating from the subsurface environment may be significantly less than that at the source. Despite these limitations and given the lack of existing models that consider vapor attenuation, Cooperative Research Centre for Contamination Assessment and Remediation of the Environment (CRC CARE) in consultation with State EPAs in Australia and industry sector representatives have recommended the Johnson and Ettinger (J&E) model for the assessment of risks emanating from vapor intrusion [8]. In the absence of realistic information on subsurface fate and behavior of VOC, limited attenuation of vapor was considered a safe option because it minimizes liability resulting from exposure to VOC. As a consequence sites posing limited risks could end up being remediated costing owners of contaminated sites more than otherwise needed to manage such sites. Sorption isotherms for VOCs on saturated soils or sediments have been demonstrated in a number of studies. However, vapor phase sorption isotherms have not been dealt with in sufficient depth [9] especially in light of subsurface soil heterogeneity. Vapor phase sorption has been studied using a range of soil organic matter (SOM) sourced from activated carbon and humic and fulvic acids by grafting these amendments onto clay minerals [10,11]. Rhue et al. [12] and Pennell et al. [13] have also investigated sorption on pure clay minerals such as smectite and kaolinite. However, in the present study, vapor attenuation was investigated using natural surface and subsurface soils varying in SOM content and mineralogy. This study addresses the research gap created by the use of grafted organic amendments on surfaces of mineral soil for experimental purposes. The coating of organic substance on mineral soil has its own disadvantage since it is controlled by the chemical nature of the organic materials (presence of functional groups), nature of clay minerals and physical properties such as surface area and mesopores [14]. Thus sorption studies using natural soil have direct application to the generation of generic sorption factors for inclusion in risk assessment tools, as well as decisions regarding remediation action plans for managing hydrocarbon contaminated sites.

2. Experimental 2.1. Material Trichloroethylene (TCE) (>99% purity) was purchased from Sigma–Aldrich (Germany) and used as supplied. The laboratory study was conducted using clean sand as a control and six soil types representing topsoil formations in various parts of Australia. Soil samples were selected to cover a range of factors such as soil pH, SOM, surface area and clay mineralogy that contribute to vapor phase sorption processes. Pertinent properties of the soils are summarized in Table 1. Soil physico-chemical tests were conducted as per the standard methods documented by Rayment [15]. The glass columns were made of 5 cm internal diameter and 15 cm length with three pieces, a bottom compartment separated from the middle column of repacked soil with a built-in fine perforated glass disc and capped with a glass lid on the top (Fig. 1). The glass column top and bottom compartment has the side port

35

Fig. 1. Packed soil glass column set-up.

with a screw cap sealed with PTFE-silicon septum for injecting TCE solution. Septum with PTFE face is selected to prevent reaction with TCE vapor and silicon body minimizes vapor loss from syringe injections, resists coring and is recommended when multiple injections are required. 2.2. Apparatus and procedure Surface area and total mesopore was determined by N2 adsorption at −196 ◦ C using a Micromeritics Gemini-V Surface Area and Pore Size Analyzer, assuming a value of 0.164 nm2 for the cross-section of the N2 molecule. The specific surface area was calculated by the BET method [16]. The clay size fractions (<2 ␮m) were extracted by treating the soils (<2 mm) with sodium hexametaphosphate (1 M). Ultrasound (50 W and 24 kHz for 15 min) treatment was provided to ascertain complete dispersion of the soil particles [17]. The clay size fractions were then pipetted out using Stokes’ law. Soil organic matter was removed by sodium hypochlorite treatment (1 M NaOCl and pH adjusted to 8 by adding 1 M HCl) [18]. After drying the samples at 60 ◦ C, soil mineralogy was characterized by X-ray diffraction (XRD) of the powdered samples using CuK␣ radiation ( = 1.5418 Å) on a PANalytical, Empyrean Xray Diffractometer (PANalytical, Australia) operating at 40 kV and 40 mA between 2.0 and 80◦ (2) at a step size of 0.016◦ . Trichloroethylene concentrations were analyzed using a conventional gas chromatograph (Agilent 7980) equipped with electron capture detector (ECD) and J&W 122–1334, DB-624 GCcolumn (0.25 mm i.d., 30 m length). The gas chromatograph was operated with initial isothermal oven temperature (35 ◦ C) for 5 min and then raising the temperature to 95 ◦ C at 10 ◦ C/min ramp-1 followed by an increase to 200 ◦ C at 50 ◦ C/min ramp-2. Injection port and detector temperatures were set at 250 ◦ C and 300 ◦ C, respectively. Helium carrier gas flow 30 ml/min and Nitrogen makeup gas flow 10 ml/min were used. Nitrogen and helium gas (extra pure >99.99%) was obtained from BOS Inc., Adelaide, South Australia. A chromatography data-handling system served to collect the data. Trichloroethylene vapor that passed through the repacked soil was monitored in the top headspace using gas-tight syringes. A

36

Soils tested CEC(meq/g) pH

EC (␮s/cm) % SOM Surface area (BET), m2 /g

Cumulative mesopore volume, cm3 /g

XRD clay structure type

USDA texture classification

Soil texture

Soil hydraulic conductivity (cm/s)

Soil air permeability (cm/s)

Neat packed soil weight (g)

Bulk density (g/cm3 )

Sandy clay loam Loamy sand

Sand = 61.9 Silt = 21.2Clay = 16.8 Sand = 79.3 Silt = 8.7Clay = 11.8 Sand = 85.2 Silt = 4.9Clay = 9.9

3.33

2.19E-01

121.8

0.414

0.243

1.60E-02

170.7

0.056

0.348

2.29E-02

267.4

0.091

MTA

10.20

5.05 86.95

13.3

34.71

9.81E-02

MTB

7.18

4.85 28.60

6.69

58.70

1.37E-01

STB

0.91

5.27 6.35

0.40

1.31

1.96E-03

IWB

14.48

8.35 102.10

1.89

9.32

1.04E-02

AT

4.59

5.34 47.50

3.44

1.33

2.15E-03

BI

2.24

4.41 51.70

3.56

0.11

1.30E-04

Kaolinite = 91% Magnetite = 9% Kaolinite = 30% Magnetite = 58% Anorthite = 39% Albite = 24% Quartz = 11% Quartz = 55% Halite = 23% Muscovite = 15% Dolomite = 31% Quartz = 25% Halite 23% Quartz = 94%

SAND

1E-03

7.63 31.20

0.00

NA

NA

NA

NA: not applicable.

Loamy sand

Loamy sand

Sand = 79.3 Silt = 11.9Clay = 8.8

0.302

1.99E-02

138.7

0.047

Loam

Sand = 41.6 Silt = 38.9Clay = 19.3

0.026

1.72E-03

182.8

0.062

Sand

Sand = 94.0 Silt = 2.8Clay = 3.0 Sand = 100

2.52

1.66E-01

200.1

0.068

13.77

9.03E-01

237.5

0.081

Sand

D.N. Bekele et al. / Journal of Hazardous Materials 306 (2016) 34–40

Table 1 Physical characteristics of the soils used to study TCE vapor sorption.

D.N. Bekele et al. / Journal of Hazardous Materials 306 (2016) 34–40

37

200-␮l aliquot of pure analytical-grade liquid TCE (0.292 g, density 1.46 g/cm3 ) was injected using a gas-tight syringe through the septum of the bottom compartment and allowed to volatilize slowly and migrate by diffusion at room temperature (Fig. 1). The volume of TCE injected in these experiments was such that the vapor pressure was sufficiently low to preclude vapor condensation. Prior to their use, the glass column and glass wool were deactivated using dimethyldicholorosilane (DMCS) to minimize sorption. Sodium azide (2 ml) was added to the soil to inhibit microbial activity. Furthermore the soil samples were oven-dried at 105 ◦ C for 24 h and directly packed into heated glass columns with glass wool pads placed at the bottom. Repacked soil was placed in the glass column at every 2 cm and tapped vertically by hand 10 times for compaction. The soil column was wrapped with cotton cloth and aluminum foil to create a uniform temperature and placed in a temperature-regulated room at 22± 2 ◦ C. Then it was equilibrated at this temperature for 24 h. Soil air permeability was estimated from soil characteristics such as grain size distribution or hydraulic conductivity by assuming a linear correlation [19]. The relationship is expressed by Eq. (1): Ca = Cw

   a w w a

(1)

where Ca = soil air permeability, cm/s; Cw = soil saturated hydraulic conductivity; cm/s; w = dynamic viscosity of water; centipoise (cP) (=1.84E-02); a = dynamic viscosity of air, centipoise (cP) (=1.002); w = density of water, kg/m3 (=998.2); and a = density of air, kg/m3 (=1.204), measurements at 20 ◦ C. Trichloroethylene vapor samples were drawn through PTFE/silicon septa using gas-tight syringes from the top headspace of soil columns with a stepping time of 72 h (Fig. 1). Syringes were washed with methanol and heat treated between sampling and checked for contamination before use. 200-␮l vapor samples were transferred to pre-treated 21-ml headspace gas-chromatograph vials in duplicate. Preparation of the 21-ml sampling vials included purging with pure lab grade nitrogen gas for one minute; the vials were immediately crimped capped with aluminum caps with a PTFE/silicon septum. The vials were placed in a 60 ◦ C oven and allowed to equilibrate for 15 min. Calibration gas standard was prepared using static bottle techniques as per US EPA [20]. Pure analytical grade TCE was injected into the pre-treated 121-ml crimped-top bottles and the weight was measured with a high-precision analytical scale. A range of gaseous concentrations (e.g., 24.1, 60.3, 120.7, 241.3, 603.3, 905.0 ␮g/cm3 ) were prepared as a calibration standard and transferred to pre-treated 21-ml crimped vials according to the calibration ranges required considering the dilution factor. The gaseous concentration in the crimped bottle was calculated using Eq. (2): Sconc =

TCEinj V121

(2)

where TCEinj = weight of TCE injected (g); V121 = volume of crimped vial (standard bottle, volume of 121 ml) (cm3 ); and Sconc = standard concentration in standard bottle (g/cm3 ). The linear calibration curve was prepared for GC-ECD response signal area versus concentration with linear regression, showing R2 ranging from between 0.95 to 0.998. The GC response signal was implemented for interpreting the result to demonstrate breakthrough of TCE in soil column before equilibrium is established. Equilibrium between the soil and TCE vapor was considered when the GC response peak area for the vapor sample from the bottom compartment and the top headspace were identical. The glass column was opened after 72 h and the soil core was tested to quantify the amount partitioned to soil. Methanol extractions were conducted in both sand control and soils according to Method 8265

Fig. 2. Trichloroethylene vapor emission rate at top headspace of soil column.

[20]. The amount of TCE sorbed on soils was used to compare the contributions of SOM and clay fractions to sorption properties. 3. Results and discussion 3.1. Soil properties Pertinent properties of the soils are listed in Table 1. The soil pH in water ranged from strongly acidic (4.4) to alkaline (8.4) with the texture varying from: sand 41.6% to 100%, silt 2.8% to 38.9% and clay 3.0% to 19.8%. The SOM content ranged from 13.3% in the study soils to below detection as in sand while the soils’ surface area ranged from 58.7 m2 /g to 0.11 m2 /g. The linear regression analysis of key soil parameters shows a strongly significant correlation (R2 = 0.99*** ) between surface area and mesopores. 3.2. Vapor attenuation The concentration of TCE vapor which is estimated using the GC response peak area versus time in the upper headspace of repacked columns are plotted in Fig. 2. Comparison of TCE vapor between soils shows that the time required for the vapor to reach the top headspace of soil column (Rt), varied considerably among the soils and increased in the order MTA > MTB > IWB > AT. Linear regression analyses indicate a significant positive relationship between the SOM and Rt, r2 = 0.91, P < 0.05. Although the SOM (3.44%) of soil AT exceeds that of IWB (SOM- 1.89%) (Table 1), Rt in soil AT is lower than soil IWB. This could be attributed to large surface area (i.e., more sorption spaces for the vapor to bind with SOM) of soil IWB which is 7 times higher than that of soil AT resulting from the presence of muscovite. The step wise multiple linear regression with 95% confidence interval used to evaluate soil properties with total TCE vapor adsorbed to soil. It exhibits no significant linear correlation having P value less than 5% due to interference of independent variable factors. One of the major limitations of the vapor flow experiment’s using repacked columns includes edge flow or macroflow channels should the soils be unevenly packed. However, the experimental result in the present study depicts consistent Rt values which are significantly correlated with soil–air permeability. Moreover,

38

D.N. Bekele et al. / Journal of Hazardous Materials 306 (2016) 34–40

the trends in GC response signal at the top head space at different sampling time interval, till near equilibration at 72 h shows slow diffusive vapor movement (see Fig. 2). These results demonstrate either minimum or no edge or macrochannel effects as vapor migrates upwards through the repacked columns. The diffusion coefficient with the limited-source, in which a fixed quantity of TCE vapor is deposited at the bottom of repacked soil column, is described by Fick’s second law. The solution to Fick’s second law of diffusion with time-dependent diffusion coefficient [21], is described in Eq. (3) assuming the diffusion is unidirectional: Q C (x, t) = √ exp Dt

 −

x2 4Dt



 = C (0, t) exp

−

x2 4Dt

 (3)

where: C(x,t) = concentration of TCE at the headspace of soil column (g/cm3 ); C(0,t) = concentration of TCE at the bottom compartment of soil column (g/cm3 ); Q = dose (g/cm2 ); D = diffusion coefficient (cm2 /s); x = soil column height (15 cm); and t = time in seconds (sampling time). √ There is a direct correlation between diffusion distance ‘x’ i.e., 2 Dt) and diffusion coefficient ‘D’. Consequently the diffusion coefficient of repacked soil column at near equilibrium (i.e., 72 h) can be described in the order DSAND > DBI > DSTB > DAT > DIWB > DMTB > DMTA . Given the higher air permeability of soil BI (0.166 cm/s), the GC response signal for TCE vapor at the top of the headspace of the column can be expected to be detected earlier than soils AT, IWB and STB (Table 1). However, the results show that the TCE vapor breakthrough was detected earlier for soils AT, IWB and STB. Similarly, despite soil MTA’s higher air permeability compared to soil MTB, TCE vapor was detected earlier for soil MTB. This may be attributed to attenuations of vapor in the soil. Although the vapor transport in soil is directly correlated to soil–air permeability, the delay in TCE vapor appearance at the top headspace is attributed to its sorption to soils. The results for TCE vapor monitoring from the bottom compartment and the top headspace suggest a gradual approach toward breakthrough after 60–72 h (Fig. 2). At breakthrough the percentage difference of GC response signal between the bottom compartment and top headspace ranged from 15 to 30 after 72 h. Although SOM has been closely correlated to vapor attenuation [10], it is apparent from Fig. 2 that it may not be the sole factor influencing vapor attenuation given that despite greater SOM content of BI relative to IWB, TCE vapor concentration at the top headspace after 72 h was higher in soil BI. In parallel, the VR by soil IWB is greater than BI. This may be attributed to the higher surface area (IWB = 9.32 m2 /g and BI = 0.11 m2 /g) and the larger total clay fraction in the soil IWB (i.e., percentage of clay multiplied by the net soil weight) (IWB = 12 g and BI = 6 g). More importantly, IWB consists of 15% muscovite which has higher external reactive edges (Table 1). Shih and Wu [22] studied organic vapor sorption kinetics and described two distinctive sorption mechanisms with SOM and mineral components in soils, and concluded that the amount adsorbed on the mineral soil surface can be correlated with the surface area of soils. However, our research findings demonstrated that sorption on the mineral surface greatly depends on the mineralogical composition of the soil rather than surface area per se (Fig. 3a and c). For instance, soil STB sorbed twice as much vapor relative to IWB despite the surface area of IWB being almost 7 times larger than STB. This may be attributed to differences in clay mineral composition (STB comprised of quartz, anorthite and albite which are primary minerals in the clay fraction of the soils compared to IWB which contains significant amounts of secondary mineral muscovite). Partitioning of TCE vapor was consistent with theoretical observations reported earlier by Sposito et al. [23], including previous experimental work by Cheng and Huang [24] who studied the

molecular attraction of adsorbed methane on variable particle surfaces to decipher fluctuations in methane adsorption with respect to changes in the organic and/or clay mineral source regions. Investigations by Cheng and Huang [24] reveal that coal sorbed gases more than oil shale, kaolinite sorbed less total gases but more methane than montmorillonite, whereas activated carbon showed expectedly much stronger sorption capacity than the other studied samples. The magnitude of these findings agreed with those reported in a Japanese study which showed a higher sorption capacity of kaolinite for methane than for montmorillonite on two out of three samples tested [25]. Moreover, the smectites revealed a large range of methane sorption capacities. However, these studies did not provide mechanistic concepts to explain the differences in the extent of methane adsorption, while the general ability of clays as sorbents was not doubted. It was also observed that soils MTB and MTA have approximately similar total clay contents (20.1 g and 20.5 g, respectively) and SOM content (∼3%). However, the total amount of TCE vapor sorbed varied considerably (Fig. 2). This difference may be attributed to the mineralogy of the clay fraction and its influence on TCE sorption capacity [13,26]. Similarly, despite larger clay content in soil AT, its ability to attenuate vapor was much lower than other soils due to its mineral content quartz, anorthite and dolomite, which have low sorption capacity and surfaces that are not conducive to retaining volatile compounds [27] (Table 1). These results demonstrate that total TCE vapor adsorbed onto soil particles is a function of both SOM content and clay mineralogy. Consistent with our findings, Ong and Lion [28] also suggest the importance of mineral fraction on sorption of TCE vapor on oven-dried soil. The linear best-fit curve of VR versus soil properties (SOM and clay content, and specific surface area and cumulative mesopore volume) are presented in Fig. 3a–d. While there is considerable scatter in the data points, the linear correlation between the total TCE vapor sorbed onto soil colloid particles versus clay content shows R2 of 0.87 excluding the data for soil BI as an outlier with P-value slightly above 0.05 (Fig. 3a). As shown in Table 1, soil BI is largely made up of quartz (94%) that has hardly any ability to sorb resulting in the outlier data [29]. Linear regression studies demonstrate that each unit increase in SOM (%) results in a 7-fold increase in total TCE vapor sorption (mg) having significant correlation with P-value <0.05. In contrast, a similar increase in clay content reveals an 11-fold increase in vapor sorption (Fig. 3a). Fig. 2 indicates that SOM plays a much significant role in Rt (rate of vapor sorption) (Fig. 3b) compared to clay content, however, as indicated in Fig. 3a clay content makes a major contribution to the total TCE vapor sorption. The mass balances (or recovery efficiencies) were used to check for leakage in the process. The result shows the recoveries for each soil test ranged from 66.9% to 109%. The cause of loss in recovery could be attributed to transfer of TCE vapor manually with gastight syringe. For the purpose of mass balance calculations, total TCE concentrations were measured at 72 h in vapor in the top headspace and bottom compartment and VR (determined from the methanol extracts multiplied by the dry weight of soil) (Table 2). There are a number of factors that could contribute to desorption and slow release of TCE vapor, but it can in any case be exceedingly slow in natural soils. Generally sorption coefficient decreases with increasing temperature i.e., less partition between vapor phase and soil particles due to the change fluid flow near the soil particles. The primary objective of this study was to investigate the contribution of SOM and clay fractions in delaying and permanently retaining volatile hydrocarbons from being released into the ambient air, or to reduce the rate of accumulation of vapor reaching the sub-slab foundation. The importance of vapor attenuation via sorption processes is critical in VI modeling especially with subsurface contamination below a layer of clean soil bridging the

D.N. Bekele et al. / Journal of Hazardous Materials 306 (2016) 34–40

39

Fig. 3. Linear regression analysis of TCE vapor sorbed onto soils versus clay (3a) and SOM contents (3b).

Table 2 TCE vapor sorbed to soil surface and remaining in the vapor phase in the top and bottom compartments after 72 h. Soils sample

TCE sorbed to soil (␮g/kg-soil)

Total TCE sorbed to soil (g)

TCE vapor in top/bottom of soil column after 72 h (g/cm3 -air)

Total TCE vapor in soil column after 72 h (g)

Percentage of recoveries (%)

MTA MTB STB IWB AT BI SAND

6.37E + 06 8.53E + 05 5.16E + 05 4.48E + 05 5.68E + 05 4.33E + 05
1.78E-01 1.46E-01 1.16E-01 7.87E-02 8.19E-02 1.03E-01
6.61E-05 6.37E-04 5.89E-04 6.15E-04 4.44E-04 6.04E-04 7.41E-04

1.78E-02 1.72E-01 1.59E-01 1.66E-01 1.20E-01 1.63E-01 2.00E-01

66.9 108.8 94.1 83.8 69.1 91.2 68.5

LOR: limit of reporting.

source and receptors. As demonstrated above, the lack of consideration of non-equilibrium conditions arising from sorption in the fate and transport of contaminant modeling will result in excessive prediction of risk arising from exposure to vapor emanating from subsurface source zone. 3.3. Vapor sorption implication for vapor intrusion and risk assessments

unit increase in SOM (%) results in a 7-fold increase in total TCE vapor sorption (mg) and a similar increase in clay content reveals an 11-fold increase in vapor sorption. The study demonstrates that testing a wide range of soil properties could lead to a deeper understanding of the sorption mechanism itself, with potential to include generic sorption coefficients for developing simplified health risk assessment tools. Acknowledgments

VIMs for health risk assessments consider an equilibrium state between vapor, soil pore-water and dry soil, however, the sorption processes may be significant. Our study shows that both the amount and rate of vapor attenuated by soil depends on SOM content as well as clay mineralogy although the amount attenuated by SOM far exceeds that by layer silicate minerals. For sites where the contaminant source is underneath uncontaminated soil with greater clay and/or SOM contents and finite source, vapor equilibrium models tend to overestimate the risk to humans due to VI. Given the present study, VIMs ought to: firstly, provide a realistic picture of the interaction of vapor with soil constituents; and secondly, include attenuation of VOC vapor by sorption reflected in available conceptual site model and subsurface soil type. 4. Conclusions The sorption of VOC vapor by soils has widespread implications both for health risk assessment and remediation perspectives, but is poorly understood and predicted. This research demonstrated that the sorption capacity of soil such that the Rt (rate of vapor sorption) significantly correlated to SOM with R Square 0.85 and significant correlation of P value <0.02. Consequently, SOM contributes in the delay of vapor accumulation to the near-surface zone. A

We thank the Cooperative Research Centre for Contamination Assessment and Remediation of the Environment (CRC CARE) for financial support. The authors are also grateful for the infrastructural support provided by the Center for Environmental Risk Assessment and Remediation (CERAR), University of South Australia. References [1] J.J. Pignatello, B. Xing, Mechanisms of slow sorption of organic chemicals to natural particles, Environ. Sci. Technol. 30 (1995) 1–11. [2] S.M. Steinberg, J.S. Schmeltzer, D.K. Kreamer, Sorption of benzene and trichloroethylene (TCE) on a desert soil: effects of moisture and organic matter, Chemosphere 33 (1996) 961–980. [3] C.T. Chiou, Partition and Adsorption of Organic Contaminants in Environmental Systems, WileyInterscience, Hoboken, NJ, 2002. [4] D.N. Bekele, R. Naidu, M. Bowman, S. Chadalavada, Vapor intrusion models for petroleum and chlorinated volatile organic compounds: opportunities for future improvements, Vadose Zone J. 12 (2013). [5] E. Friebel, P. Nadebaum, The development of HSLs for petroleum hydrocarbons—an issues paper. CRC CARE Technical Report no. 4. CRC for Contamination Assessment and Remediation of the Environment. Adelaide, Australia, 2007. [6] P.C. Johnson, R.A. Ettinger, Heuristic model for predicting the intrusion rate of contaminant vapors into buildings, Environ. Sci. Technol. 25 (1991) 1445–1452.

40

D.N. Bekele et al. / Journal of Hazardous Materials 306 (2016) 34–40

[7] A.P.I. BioVapor A 1-D Vapor Intrusion Model with Oxygen-Limited Aerobic Biodegradation American Petroleum Institute (2009) http://www.api.org. [8] G.B. Davis, M.G., Trefry, B.M. Patterson, Petroleum vapour model comparison, CRC for Contamination Assessment and Remediation of the Environment, March 2009, Technical Report no. 9., Adelaide, Australia, (2009). [9] M.C. Tekrony, R.C. Ahlert, Adsorption of chlorinated hydrocarbon vapors onto soil in the presence of water, J. Hazard. Mater. 84 (2001) 135–146. [10] G. Cornelissen, O. Gustafsson, T.D. Bucheli, M.T. Jonker, A.A. Koelmans, P.C. Van Noort, Extensive sorption of organic compounds to black carbon, coal, and kerogen in sediments and soils: mechanisms and consequences for distribution, bioaccumulation, and biodegradation, Environ. Sci. Technol. 39 (2005) 6881–6895. [11] C. Niederer, R.P. Schwarzenbach, K.-U. Goss, Elucidating differences in the sorption properties of 10 humic and fulvic acids for polar and nonpolar organic chemicals, Environ. Sci. Technol. 41 (2007) 6711–6717. [12] R.D. Rhue, K.D. Pennell, P.S.C. Rao, W.H. Reve, Competitive adsorption of alkylbenzene and water vapors on predominantly mineral surfaces, Chemosphere 18 (1989) 1971–1986. [13] K.D. Pennell, R.D. Rhue, P.S.C. Rao, C.T. Johnston, Vapor-phase sorption of p-xylene and water on soils and clay minerals, Environ. Sci. Technol. 26 (1992) 756–763. [14] E.S. Krull, J.O., Skjemstad, J.A. Baldock, Functions of Soil Organic Matter and the Effect on Soil Properties. Report GRDC Project No: CSO 00029, 2004, p.107. [15] G.E. Rayment, Australian Laboratory Handbook of Soil and Water Chemical Methods, in: G.E. Rayment, F.R. Higginson (Eds.), Inkata Press, Melbourne, Australia, 1992. [16] S. Brunauer, P.H. Emmett, E. Teller, Adsorption of gases in multimolecular layers, J. Am. Chem. Soc. 60 (1938) 309–319. [17] L. Zhao, Y. Sun, X. Zhang, X. Yang, C.F. Drury, Soil organic carbon in clay and silt sized particles in Chinese mollisols: relationship to the predicted capacity, Geoderma 132 (2006) 315–323. [18] K. Kaiser, K. Eusterhues, C. Rumpel, G. Guggenberger, I. Kögel-Knabner, Stabilization of organic matter by soil minerals—investigations of density and

[19] [20]

[21]

[22]

[23]

[24] [25] [26]

[27] [28] [29]

particle-size fractions from two acid forest soils, J. Plant Nutr. Soil Sci. 165 (2002) 451–459. S.S. Suthersan, Remediation Engineering: Design Concept, CRC Press LLC, Florida, 1997, pp. 33431. U.S. EPA, SW-846. Method 8265, Volatile Organic Compounds in Water, Soil, Soil Gas and Air by DirectSampling Ion Trap Mass Spectrometry (DSITMS) (2002). L. Mejlbro, The Complete Solution of Fick’s Second Law of Diffusion with Time-dependent Diffusion Coefficient and Surface Concentration, Technical University of Denmark, 1996. Y.-H. Shih, S.-C. Wu, Distinctive sorption mechanisms of soil organic matter and mineral components as elucidated by organic vapor uptake kinetics, Environ. Toxicol. Chem. 24 (2005) 2827–2832. G. Sposito, N.T. Skipper, R. Sutton, A.K. Park, J.A. Greathouse, Surface geochemistry of the clay minerals, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 3358–3364. A.-L. Cheng, W.-L. Huang, Selective adsorption of hydrocarbon gases on clays and organic matter, Org. Geochem. 35 (2004) 413–423. A. Sugimoto, J. Dan, T. Kumai, J. Murase, Adsorption as a methane storage process in natural lake sediment, Geophys. Res. Lett. 30 (2003) 2080. W. Shiyong, Y. Xiaohong, X. Wenjun, Preparation and adsorption properties for fluoride of kaolinite-magnetite association, Electronics and Optoelectronics (ICEOE), 2011 International Conference on, (2011) pp. V1-410–V411-414. L. Clausen, I. Fabricius, L. Madsen, Adsorption of pesticides onto quartz, calcite, kaolinite, and ␣-alumina, J. Environ. Qual. 30 (2001) 846–857. S.K. Ong, L.W. Lion, Effects of soil properties and moisture on the sorption of trichloroethylene vapor, Water Res. 25 (1991) 29–36. R. Wienberg, E. Heinze, U. Förstner, Experiments on specific retardation of some organic contaminants by slurry trench materials, Contaminated soil, Springer (1986) pp. 849–857.