J. Cent. South Univ. (2016) 23: 1924-1930

DOI: 10.1007/s11771-016-3248-0

Diesel oil infiltration in soils with selected antecedent water content and bulk density

MA Yan-fei(马艳飞)^{1, 3}, LI Yong-xia(李永霞)², S. H. Anderson³, ZHENG Xi-lai(郑西来)⁴,

FENG Xue-dong(冯雪冬)¹, GAO Pei-ling(高佩玲)¹

1. Department of Resource and Environmental Engineering, Shandong University of Technology, Zibo 255091, China;

2. Shandong Academy of Environmental Science, Ji’nan 250013, China;

3. Department of Soil, Environmental and Atmospheric Sciences, University of Missouri,Columbia, Missouri 65211, USA;

4. Key Laboratory of Ocean Environment and Ecology, Ministry of Education, Ocean University of China,Qingdao 266100, China

Central South University Press and Springer-Verlag Berlin Heidelberg 2016

Abstract: The effects of soil texture, initial water content and bulk density on diesel oil infiltration in fine sand and silty clay loam materials were evaluated. Three physical and two empirical equations express diesel oil infiltration through soils with time, with coefficients of determination greater than 0.99. Diesel oil infiltrates more quickly in the fine sand than in the silty clay loam material. Diesel oil infiltration rates are found to decrease with increasing initial water content and bulk density for the silty clay loam material. The infiltration rate of diesel oil in the fine sand material increases slightly with increasing initial water content. The diesel oil saturated conductivity (K_diesel) decreases with increasing bulk density for the silty clay loam column. Diesel oil sorptivity(S) decreases linearly with increased initial water content and bulk density of the silty clay loam material. Changes in empirical parameters relative to initial water content and bulk density are similar to the parameter S.

Key words: infiltration; diesel oil; soil; physical and empirical models; initial water content; bulk density

1 Introduction

Leaks and spills are inevitable in the process of exploring, refining, transporting and using petroleum products [1]. These products present a potential danger as a source of pollution to subsurface environments because of their toxicity and the possible migration of pollutants to groundwater systems [2]. Therefore, it is of vital importance to research the infiltration of petroleum products in soils, including the effects of soil physical properties and soil conditions on infiltration.

The infiltration of petroleum products in porous media is a complicated process which is affected by the characteristics of both soils and petroleum products [3-5]. Numerical simulation of multi-phase flow in porous media was first applied in the field of petroleum engineering, and was developed in a petroleum- contaminated field [6]. Approaches to modeling multi-phase flow have been discussed as two-phase and three-phase flow models [7-8]. Much research has been done to develop models of the migration of petroleum components in porous media [9-10]. However, many of the models have been developed under assumptions or approximate conditions; for example, non-aqueous phase liquids (NAPLs) as completely immiscible liquids with water or gas movement in porous media have not been taken into account [11-12]. Laboratory data are needed to test the reliability of these prediction models. There are some laboratory studies that have examined the infiltration of NAPLs under unsaturated-zone conditions [13-14]. Soil water content, bulk density and soil structure affect NAPLs infiltration [15-16], but most laboratory studies only account for the oil saturation distribution or the oil infiltration curve [3-4, 17-18]. Therefore, the laboratory data and models both need to be evaluated together.

There are many studies on water infiltration [19-20]. Physical models such as the Green and Ampt model (1911), the Parlange model (1982) and the Philip model (1957) have often been used for modeling water infiltration in soils and soil water redistribution [21].Researchers have also recommended empirical equations to design and evaluate water infiltration and surface water transport, including the Kostiakov (1932), the Kostiakov–Lowies and the USDA-SCS equations [22]. Some researchers have used the Kostiakov model fitted to crude oil infiltration in soils [15, 23]. The application of these models could take into account the parameters of saturated conductivity (K_s) and sorptivity (S). This has vital significance for reducing and controlling the pollution caused by NAPLs if the parameters K_s and S of NAPLs can be obtained with these models.

In this wrok, diesel oil was used as an light non-aqueous phase liquid (LNAPL) contaminant, and two selected uncontaminated materials (fine sand and silty clay loam) were collected as porous media. The purpose of the experiments is to study the effects of initial water content and bulk density of soils on LNAPL infiltration in soils and to ascertain the reliability of physical and empirical models for LNAPL infiltration in soils.

2 Methods and materials

2.1 Materials

Porous media samples include sediment and soil samples. Non-contaminated soil samples were collected from the petrochemical area of Linzi District, Zibo City, China, where Dawu Water, a source of drinking water, was located. Soils at this site were classified as Meadow Cinnamon soils in the Chinese soil classification system. In addition, soil was removed from the 40 to 50 cm depth horizon of a cultivated soil not contaminated with petroleum products. The sandy soil sediment was gathered from the 0 to 20 cm depth from an alluvial riverbank not contaminated with petroleum products. The soil samples were air dried, ground, passed through a 2 mm sieve, and stored in sealed vessels at room temperature prior to the experiments. According to the soil particle size classification of the Chinese Geology

and Mineral Ministry, the sandy alluvial soil was classified as fine sand (fine sand) with the cultivated soil classified as silty clay loam texture class (silty clay loam) [24].

The initial gravimetric water contents of the air- dried fine sand and silty clay loam soils were 0.6% and 1.1%, respectively. The air-dried soils were dried in a oven under 105-110 °C until they achieved constant mass. The water contents of the materials were calculated by measuring data of the material mass before and after stoving. The particle size distributions and physicochemical properties of the selected porous media are listed in Table 1. The sieving method and hydrometer method were used to analyze the grain size distribution of the materials. The specific surface area of the materials was measured by the BET/N₂-adsorption method with an F-Sorb2400 surface analyzer. Determination of the organic matter was conducted by the potassium dichromate method.

The diesel oil, 0^#, for the experiments was supplied by Qilu Petrochemical Corporation. The composition of diesel oil was analyzed with gas chromatography-mass spectrometry (Agilent Technologies 7890A-5975C), and it was composed of alkanes (67.69%), cyclanes (15.22%) and aromatic hydrocarbons (17.09%). The density and viscosity of the diesel were measured by pycnometer method and capillary viscosimetry, and they were 0.848 g/cm³ and 3.96 MPa·s at 20.0 °C, respectively.

2.2 Experimental setup and procedures

An experimental setup is shown in Fig. 1 which consisted of a transparent glass cylinder and a Mariotte bottle which kept a constant pressure head of diesel oil (3 cm). The cylinders were 4.5 cm in diameter and 40 cm in height. Steel meshes were placed on the bottom of the columns to contain the soil.

The water contents of air-dried fine sand and silty clay loam were 0.6% and 1.1%, respectively. All the soil samples were packed in cylinders according to the bulk density values listed in Table 1. The top of the soil sample was 10 cm below the top of the cylinder to allow the addition of liquid.

Table 1 Grain size distribution and physicochemical properties of selected soil materials.

Fig. 1 Schematic of diesel oil infiltration experiments

For the soil samples with different initial water contents, the soil samples were mixed with water by mass ratio according to the selected initial water contents. The soil sample was packed by layers within the cylinder with each 2.5 cm increment. For one layer，the mixed soil sample was weighted by elevator balance, and the mixed soil sample mass was determined by packing volume (h=2.5 cm , D=4.5 cm), the bulk density value (oven-dried soil sample) and initial water content. The final length of each sample within the cylinder was 30 cm. For the silty clay loam cylinders with different bulk density, the initial water contents were all 1.1%. Therefore, the soil sample for each layer of one soil cylinder was weighed according to the selected bulk density. The packing method and layer increment are the same as the soil samples with different water contents.

During the experiment, diesel oil was added to the top of the soil column continuously with a Mariotte bottle with a constant pressure head of 3 cm. The experiments were conducted in a constant temperature incubator at 20.0 °C. The amount of diesel oil infiltrating the column was monitored as a function of time after experiment initiation. The infiltration experiment was discontinued after a few minutes when diesel oil exited the bottom of the column. A plastic wrap with small holes for aeration covered the top of the soil column to minimize diesel oil volatilization.

2.3 Infiltration equations

Three physically-based infiltration equations and two empirically-based equations were used to fit the diesel oil infiltration data [22]. The first physically-based diesel oil infiltration equation (Eq. (1)) from the Green and Ampt equation is as follows:

               (1)

where t is the time after infiltration initiation, h; I is the cumulative infiltration, mm; K_diesel is the diesel oil saturated conductivity, mm/h; and S is the diesel oil sorptivity,

The second physically-based diesel oil infiltration equation from the Parlange equation is as follows:

             (2)

The third physically-based diesel oil infiltration equation from the Philip equation is as follows:

                            (3)

The first empirically-based diesel oil infiltration equation from the Kostiakov equation is as follows:

                                    (4)

where I is the cumulative infiltration, mm; t is the time after infiltration initiation, h; and A, mm/hⁿ, and n is the coefficient of the equation.

The second empirically-based diesel oil infiltration equation from the Kostiakov-Lewies equation is as follows:

                                (5)

where A and n are defined in Eq.(4); and f₀ is the final infiltration rate, mm/h.

For Eqs. (1) and (2), the Solver routine of Excel was used for estimating K_diesel and S parameters based on cumulative diesel oil infiltration measurements. The initial S parameter was estimated from initial diesel infiltration divided by the time (t^1/2), and the initial K_diesel value was estimated from the final/steady state infiltration rate, mm/h. For the other three equations, the non-linear fitting routine of Origin software was adapted to obtain the parameters for fitting measured diesel oil infiltration. Comparisons were made by fitting equations with both procedures with similar results.

3 Results and discussion

3.1 Effect of antecedent water content on diesel oil infiltration

Infiltration curves as a function of time for the fine sand and silty clay loam materials for selected antecedent water content values are shown in Fig. 2.

The curves show diesel oil infiltrating rapidly during the early period, with the infiltration rate decreasing as time progresses. Diesel oil requires less time to infiltrate into the fine sand material than the silty clay loam material due to differences in pore size. The hydraulic conductivity of crude oil is proportional to the particle size of soils [25], and similar results are found for crude oil, kerosene and PCE, which infiltrates faster in sandy textures soils than in loam soils [4, 26]. Gravity and capillary forces control the fluid infiltration [27] and the effect of capillary forces on fluid sorption is dominant early in the infiltration process, while gravity forces are dominant later in the infiltration process. The mean grain size of the silty clay loam material is much finer than that of the fine sand; these finer pores offer greater resistance to flow.

Fig. 2 Physical and empirical models fitted to measured diesel oil infiltration data for fine sand and silty clay loam:

The fine pores in the silty clay loam material allow sorption of fluid through capillary action. The organic matter and the surface area are higher for the silty clay loam material than these properties for the fine sand material; these properties directly affect the sorption of diesel oil. Organic matter plays an important role in the sorption (partitioning) of hydrophobic organic compounds into soil and in the retardation of hydrophobic organic compounds in these systems. Soils with higher organic matter content are observed to have higher residual gasoline saturation [28].

Antecedent water content plays a significant role in the initial infiltration rate. The lower the water content, the higher the initial infiltration rate for the silty clay loam material (Figs. 2(d)-(f)). The effect of antecedent water content on initial infiltration rates is observed for the silty clay loam material, since during the first 2.0 h, faster diesel infiltration occurred for the 1.1% initial water content treatment as compared to the 10% initial water content (Figs. 2(d)-(f)). As initial water content increases, the initial pressure gradient decreases; therefore, the initial infiltration rate decreases. Once the material becomes saturated with diesel oil, the gravity gradient is more dominant, and the infiltration rates are more similar.

For the fine sand material, the effect of antecedent water content on the initial infiltration rate is less significant (Figs. 2(a)-(c)). However, it appears that increasing initial water content slightly increases the initial infiltration rate. Possible reasons for these observations may be that the water in the fine sand material covers the sand particles [26], thus reducing interaction between diesel oil and sand. It is also possible that water may fill the smaller soil pore spaces, leaving these locations unavailable for diesel oil sorption with increasing initial water content. Therefore, diesel oil may infiltrate through the larger pore spaces and require less time to pass through the soil. The role of water films was also found to reduce the resistance to the flow of kerosene in sand, and the different effects of initial moisture content on kerosene conductivity in medium sand soil have been attributed to this reduced resistance [29].

3.2 Effect of bulk density on diesel oil infiltration

The air-dried silty clay loam materials are packed in the columns with three bulk densities: 1.35, 1.45 and 1.55 g/cm³. The infiltration curves of 1.35 and 1.55 g/cm³ silty clay loam are shown in Fig. 3.

Significant effects of soil bulk densities on the infiltration of diesel oil in the soils are shown. The diesel oil infiltration rate in silty clay loam material decreases with increasing bulk density. The infiltrations of tetrachlorethylene and kerosene in soils also accord with this rule [30]. The pore sizes become finer with the silty clay loam compacted heavily, and the residual porosity becomes smaller for diesel oil infiltration. It takes diesel oil more time to infiltrate the silty clay loam with higher bulk density for the same depth because of its finer pore system. From this standpoint, intentionally increasing the soil compaction can extend the migration time of petroleum pollutants into soils, which can allow more time to clean up polluted fields and reduce the risk of underground water contamination.

3.3 Fitting equations

The parameters used to fit equations for diesel oil infiltration in fine sand and silty clay loam materials under different initial water contents are listed in Table 2. The fitting results for silty clay loam material with 1.35 and 1.55g/cm³ are also listed in Table 2.

The coefficients of determination (R²) are greater than 0.99. It is apparent from the coefficients of determination that all the physical and empirical equations fit the data well and can be used to quantify diesel oil infiltration in soils. Equation (5) is optimized by comparing the proximity between the fitting results and experimental data as shown in Figs. 2 and 3. The diesel oil conductivity (K_diesel) of fine sand material is higher than that of silty clay loam material. As can be seen from the fitting of the parameters of the physical equations, there are no regular changes for K_diesel of fine sand and silty clay loam materials under different initial water content values.

According to the parameters K, S and A in Table 2, the fitting results of the parameters changing with initial water content and bulk density are listed in Table 3.

The K_diesel decreases with increasing bulk density of the silty clay loam column, and the determination coefficient of the fitting equation is greater than 0.92. Two linear correlations can partially describe the diesel oil sorptivity parameter (S) which decreases with increasing bulk density and initial water content for the silty clay loam material (determination coefficients are 1 and 0.99). S is lower with increasing initial water content because of lower pressure gradients. The values of A from the two empirical equations decrease with increasing initial water content and bulk density of the silty clay loam material; these relationships can be described with linear correlations (correlation coefficients are greater than 0.99). The experimental data of LNAPL infiltration in soils can be described with clay or silt content [14]. As for fine sand and silty clay loam materials, the fitting value of S is related to the clay content of the soil, initial water content and bulk density, and a multiple linear regression can show the relationship of S and the factors with a determination coefficient greater than 0.94.

Fig. 3 Physical and empirical models fitted to measured diesel oil infiltration data for silty clay loam with different bulk densities:

Table 2 Equation parameters for diesel oil infiltration in soil columns under selected initial water content and bulk density values estimated with three physical (Eqs. (1)-(3)) and two empirical equations (Eqs. (4)-(5))

Table 3 Relationship of K and S with selected initial water content and bulk density values

4 Conclusions

1) Soil texture, initial water content and bulk density play important roles in diesel oil infiltration in fine sand and silty clay loam materials. Diesel oil infiltrates faster in fine sand than in silty clay loam materials.

2) Three physical and two empirical equations fit the diesel oil infiltration curve in soils as a function of time with coefficients of determination greater than 0.99, and the Kostiakov-Lewies equation is the optimized one.

3) The K_diesel parameter decreases with increasing bulk density for the silty clay loam material. The S parameter decreases linearly with the increasing initial water content and bulk density of silty clay loam; and the change in the A parameter with initial water content and bulk density of silty clay loam could also be expressed as a linear function. As for fine sand and silty clay loam materials with different clay contents, initial water contents and bulk densities, a multiple linear regression can show the relationship of S and the factors with a correlation coefficient greater than 0.97.

References

[1] SCHIRME M, BUTLER B J. Transport behaviour and natural attenuation of organic contaminants at spill sites [J]. Toxicology, 2004, 205(3): 173-179.

[2] LEE K Y. Phase partitioning modeling of ethanol, isopropanol, and methanol with BTEX compounds in water [J]. Environmental Pollution, 2008, 154(2): 320-329.

[3] MA Y. Studies on migration, transportation and leaching technique of oil pollutants in unsaturated aquifer media [D]. Qingdao: Ocean University of China, 2011. (in Chinese)

[4] WANG Y, SHAO M. Infiltration characteristics of non-aqueous phase liquids in undisturbed loessal soil cores [J]. Journal of Environmental Sciences, 2009, 21(10): 1424-1431.

[5] LUCIANOA A, VIOTTI P, PAPIPI M P. Laboratory investigation of DNAPL migration in porous media [J]. Journal of Hazardous Materials, 2010, 176 (1/2/3): 1006-1017.

[6] MCLAREN R G,SUDICKY E A,PARK Y J,ILLMAN W A. Numerical simulation of DNAPL emissions and remediation in a fractured dolomitic aquifer [J]. Journal of Contaminant Hydrology, 2012, 136-137: 56-71.

[7] LUCIANO A, VIOTTI P, PAPINI M P. On morphometric properties of DNAPL sources: Relating architecture to mass reduction [J]. Water Air Soil Pollution, 2012, 223: 2849-2864.

[8] NIESSNER J, HELMIG R. Multi-scale modeling of three-phase– three-component processesin heterogeneous porous media [J]. Advances in Water Resources, 2007, 30: 2309-2325.

[9] ATTEIA O, HOHENER P. Semianalytical model predicting transfer of volatile pollutants from groundwater to the soil surface [J]. Environmental Science and Technology, 2010, 44(16): 6228-6232.

[10] CHRIST J A, PENNEL K D, ABRIOLA L M. Quantification of experimental subsurface fluid saturations from high-resolution source zone images [J]. Water Resources Research, 2012, 48(1): W01517.

[11] ERVIN R E, BOROUMAND A, ABRIOLA L M, RAMSBURG C A. Kinetic limitations on tracer partitioning in ganglia dominated source zones [J]. Journal of Contaminant Hydrology, 2011, 126 (3/4): 195-207.

[12] OOSTROM M, ROCKHOLD M L, THORNE P D, TRUEX M J, LAST G V, ROHAY V J. Carbon tetrachloride flow and transport in the subsurface of the 216-Z-9 trench at the Hanford Site [J]. Vadose Zone Journal, 2007, 6(4): 971-984.

[13] ALLEN-KING R M, BARKER J F, GILLHAM R W, JENSEN B K. Substrate-limited and nutrient-limited toluene biotransformation in sandy soil [J]. Environmental Toxicology and Chemistry, 1994, 13(5): 693-705.

[14] SCHRAMM M, WARRICK A W, FULLER W H. Permeability of soils to four organic liquids and water [J]. Hazardous Waste and Hazardous Materials, 1986, 3(1): 21-27.

[15] ZHANG B, SHAO M. Effect of initial soil water content on crude oil infiltration into soils [J]. Transactions of the CSAE, 2010, 26(3): 9-13. (in Chinese)

[16] PANTAZIDOU M, LIU K. DNAPL distribution in the source zone: Effect of soil structure and uncertainty reduction with increased sampling density [J]. Journal of Contaminant Hydrology, 2008, 96(1/2/3/4): 169-186.

[17] CHEVALIERA L R, FONTE J M. Correlation model to predict residual immiscible organic contaminants in sandy soils [J]. Journal of Hazardous Materials, 2000, B72: 39-52

[18] ZHAO Y, SHAO M, ZHANG X. Impact of localized compaction and ridge fertilization on field nitrate transport and nitrate use efficiency [J]. Journal of Applied Ecology, 2004, 15(1): 68-72. (in Chinese)

[19] HE Z, ZHOU G, ZHANG L, ZHANG J. Seepage characteristics of high grained soil embankment under rainfall condition [J]. Journal of Central South University: Science and Technology, 2014, 45(1): 262-268. (in Chinese)

[20] ANDERSON S H, UDAWATTA R P, SEOBI T, GARRETT H E. Soil water content and infiltration in agroforestry buffer strips [J]. Agroforestry Systems, 2009, 75(1): 5-16.

[21] ZHANG J, HAN T, DOU H, MA S. Analysis slope safety based on infiltration model based on stratified assumption [J]. Journal of Central South University: Science and Technology, 2013, 45(9): 3211-3218. (in Chinese)

[22] RAHIMI A, BYZEDI M. The Evaluation and determining of soil infiltration models coefficients [J]. Australian Journal of Basic and Applied Sciences, 2012, 6(6): 94-98.

[23] MOLINO B, ALBERGO G, BISCIONE A. Infiltration of petroleum products through the unsaturated zone [J]. Journal of Hydraulic Research, 2007, 45(1): 126-130.

[24] YI Y. Methods of soil physical properties [M]. Beijing: Peking University Press, 2009: 40-42. (in Chinese)

[25] KHAMEHCHIYAN M, CHARKHABI A H, TAJIK M. Effects of crude oil contamination on geotechnical properties of clayey and sandy soils [J]. Engineering Geology, 2006, 89(3/4): 220-229.

[26] WANG R, ZHANG F, WANG G. Infiltration and redistribution of dense non-aqueous phase liquid in Loess soil [J]. Journal of Irrigation and Drainage, 2010, 29(4): 89-92. (in Chinese)

[27] WIPFLER E L，NESS M, BREEDVELD G D，MARSMAN A, van der ZEE S E A T M. Infiltration and redistribution of LNAPL into unsaturated layered porous media [J]. Journal of Contaminant Hydrology, 2004, 71: 47-66.

[28] HAYDEN N J, VOICE T C, WALLACE R B. Residual gasoline saturation in unsaturated soil with and without organic matter [J]. Journal of Contaminant Hydrology, 1997, 25(3/4): 271-281.

[29] GERSTL Z, GALIN T S, YARON B. Mass flow of a volatile organic liquid mixture in soils [J]. Journal of Environmental Quality, 1994, 23(3): 487-493.

[30] SHI X, ZHANG F, WANG G, JIA G, WANG L. Experimental study of infiltration of non-aqueous phase liquids in loess soil [J]. Agricultural Research in the Arid Areas, 2005, 23(4): 49-52. (in Chinese)

(Edited by FANG Jing-hua)

Foundation item: Projects(40272108, 41402208) supported by the National Natural Science Foundation of China; Projects(ZR2012DL05, ZR2015EL044) supported by Shandong Provincial Natural Science Foundation, China; Project(4072-114017) supported by Young Teachers’ Development of Shandong University of Technology, China; Project(J12LC51) supported by Shandong Province Higher Educational Science and Technology Program, China

Received date: 2015-01-21; Accepted date: 2016-03-22

Corresponding author: ZHENG Xi-lai, Professor, PhD; Tel: +86-13012505297; E-mail: zhxilai@ouc.edu.cn