Analyzing the characterization of pore structures and permeability of diesel contaminated clays under different aging conditions

Abstract

In this study, mercury intrusion porosimetry (MIP) and X-ray micro-computed tomography (XRμCT) were used to characterize the pore structures and investigate the permeability characteristics of clay after aging and contamination with diesel. The results of the MIP tests showed that aging leads to reductions in porosity and average diameter, as well as an increase in tortuosity. The XRμCT analysis yielded consistent results; it showed that aging renders pores more spherical and isotropic and pore surfaces smoother. This weakens the pore connectivity. Micromorphological analysis revealed that aging led to the rearrangement of soil particles, tighter interparticle overlapping, and a reduction in pore space. The combination of MIP and XRμCT provided a comprehensive and reliable characterization of the soil pore structure. An increased diesel content increased the porosity and average diameter and reduced the tortuosity of the pores. Mechanistic analysis showed that aging weakens interparticle cohesion; this causes large agglomerates to break down into smaller agglomerates, resulting in a tighter arrangement and a subsequent reduction in porosity. An increase in diesel content increases the number of large agglomerates and pore spaces between agglomerates, resulting in increased porosity. Both aging and diesel content can weaken the permeation characteristics of soil.

Graphical abstract

Keywords

MIP / XRμCT / aging / diesel content / pore structure / permeability characteristics

1 Introduction

The increasing contamination of soil with petroleum hydrocarbons represents an urgent problem for humanity [1–3]. Contaminant removal represents one of the main ways to solve this problem [4]. Currently, thermal desorption is the primary technique used to this end [5]. It is a remediation method in which contaminants that have been thermally desorbed or pyrolyzed are removed via negative-pressure extraction; its success depends upon the pore structure of the soil and the volatile nature of the contaminants [3,6]. Thus, alongside the thermophysical-chemical properties of the contaminants, the porous structural characteristics of the clay are fundamental factors influencing the thermal desorption efficiency. This is because the pore structure constitutes the material basis for percolation (gas and liquid percolation) occurring in the soil [7,8] and determines the flow paths and rates of contaminants during permeation [9]. Thus, a thorough and clear understanding of the pore structures of contaminated soils is essential.

Petroleum hydrocarbons are mixtures of aliphatic and aromatic hydrocarbons; their properties considerably differ from those of water [10]. When petroleum hydrocarbons contaminate soil, they do not dissolve in water and can affect the interactions between the soil particles and water and between the soil particles themselves [3]. Some of these changes occur immediately at the macroscopic or microscopic level. However, other changes gradually develop over time and are present in both macroscopic and microscopic forms [11]. Petroleum hydrocarbons have been present in soil for decades [1]. Naturally, the pore structure changes produced by petroleum hydrocarbon contaminants are time-dependent processes; however, only a limited number of studies have reported the effects of aging upon the pore structures of soils contaminated with petroleum hydrocarbons. Researchers have reported that aging causes shrinkage of contaminated soils and the formation of microscopic cracks [12]. Furthermore, aging can cause the microscopic flocculation structure of the soil to break up; as a result, the soil particles disperse and take on a flaky form (Scanning Electron Microscope analysis) [3]. The abovementioned literature is mostly qualitative and descriptive; it provides no quantitative investigation of the effects of aging on pore structure parameters, and research regarding the effects and mechanisms of aging upon the pore structures of soils contaminated with petroleum hydrocarbons is lacking.

Changes in the levels of petroleum hydrocarbon contaminants in the soil can also affect the pore structure. These changes can alter the interactions between soil and water and between soil particles [3]. Current research regarding the variation in contamination levels is mainly focused on the macro-engineering properties of soils [11,13,14]; only limited research has been conducted regarding its effects on pore structures. Petroleum hydrocarbons affect soil agglomeration by adsorbing to the surrounding soil particles; these then form flocculated structures (coarse particles) [15], which increases the sizes of macropores in the soil (as determined by scanning electron microscopy analysis) [16–18]. However, according to other results, when petroleum hydrocarbons leach into the soil, they block the aggregation of particles and the soil takes on a dispersed structure. This reduces the number of larger pores [3]. Hence, petroleum hydrocarbons can exert two different impacts upon the pore structure of soil, and the mechanisms underlying these effects remain unclear.

The pore structures of soils can be determined via various methods. Previously, the pore structures were mostly measured using breakthrough curves, the tension infiltrometer technique, the slice method, and mercury piezometric methods [19–22]. Some of these methods are complex and highly accurate and can provide a wide range of measurements. Some methods are simpler to operate but produce erroneous results. In recent years, the development of X-ray micro-computed tomography (XRμCT), a non-destructive technology, has offered a reliable technique for rapidly and accurately analyzing pore structures. This technique offers several advantages. It is non-invasive, causes no damage, can visualize the pore structure, and can obtain a large number of pore structure parameters. It has gradually become the mainstream technique for characterizing pore structures. For example, Nakashima et al. [23] used CT techniques for the destructive analysis of core petroleum-contaminated soil samples. Al-Raoush [24] studied the effects of particle geometry upon residual non-aqueous phase liquid (NAPL) properties in porous media, using CT techniques. They visualized the effect of the particle shape upon the NAPL saturation and the area of the constitutive curved moon surface. Javanbakht et al. [25] used CT techniques to dynamically characterize the effects of surfactants and microemulsions upon the efficiency of NAPL expulsion from nonhomogeneous rocks. Guan et al. [26] used CT technology to obtain the changes in pore structure when a sample was treated with a surfactant. However, XRμCT suffers from technical shortcomings, including the problem of test accuracy and the conflict between test accuracy and test area (i.e., the smaller the test area, the more accurate the test result, but the less representative the specimen) [27–29]. Accordingly, many scholars have combined mercury intrusion porosimetry (MIP) and XRμCT techniques, to exploit their respective strengths in the complementary characterization of pore structures. Wang et al. [30] used both techniques to characterize the pore structure of coal and found that XRμCT was superior to MIP in characterizing the pore morphology; however, the pore size analysis was not as extensive as that of MIP. The results obtained using both techniques were consistent and provided a good characterization. Wu et al. [31] combined XRμCT and MIP techniques to characterize pore structures. The results showed that both methods were highly reliable and accurate, and they complemented each other in characterizing the pore structure. Hashemi et al. [32] found good agreement between the results when they applied these two techniques to characterize the pore structure. Thus, the combination of XRμCT and MIP represents a good method of characterizing the pore structures of soils, and it has been well accepted by many scholars. The two techniques can be combined to characterize the effects of aging upon the soil pore structure, as well as to analyze the causes of changes in permeability.

In this study, clay soil contaminated with diesel was prepared using clay as the background soil and diesel as the petroleum hydrocarbon contaminant. Aging time and diesel content were considered as influencing factors, and MIP and XRμCT measurements were applied to investigate the characteristics of the pore structure changes, as well as their effect mechanisms (here, pore structure refers to pore size, distribution, pore shape, and pore connectivity). In addition, the permeation characteristics of the specimens were tested for different aging times and levels of diesel doping, to determine the relevant design parameters before thermal desorption.

2 Materials and methods

2.1 Introduction of soil site and treatment of experimental soil

The borrow site is located in the Nanning fault depression basin in China, on the second terrace of the Yongjiang River; the terrain is relatively flat and open. The site includes an abandoned chemical factory area and a surrounding old residential area that is to be used for commercial housing development. The geological structure of the site is simple; that is, no Holocene active faults pass through it, and its regional geological structure is stable. According to the drilling results, the strata of the site (from top to bottom) are as follows: Quaternary artificial fill, alluvial clay layer, silt layer, pebble layer, and underlying Paleogene mudstone. No stable groundwater level was identified during the survey, and it is speculated that the buried depth exceeds 11 m. No adverse geological phenomena such as debris flow, collapse, or landslides were noted at the site. The investigation and assessment of the pollution situation at the site found that the production area in the chemical plant area was seriously polluted by hydrocarbon compounds, with a maximum pollution level of 69.8 g/kg and a maximum pollution depth of 10 m. The other areas were either less polluted or exhibited no pollution.

The experimental soil was collected from an uncontaminated alluvial clay layer at the site. The collected soil samples were sealed, packaged in plastic bags, and transported to the laboratory for drying, crushing, and impurity removal. The key physical properties of the soil samples are listed in Tab.1.

2.2 Diesel

The diesel oil used in this experiment was 0# diesel oil, as sold on the Chinese market. It has a light green color, a relative density of 0.856, a viscosity coefficient of 3.95 mPa·s at 20 °C, and an oil−water surface tension coefficient of 1.8 × 10⁻² N·m⁻¹.

2.3 Experimental design

To investigate the effects of aging and diesel content upon the pore structure, the aging time and diesel content were set as influencing variables (Tab.2). MIP and XRμCT were used to characterize the pore structures. The mechanism was investigated by measuring the particle size distribution and exchangeable cation content (i.e., the cation exchange capacity (CEC)). Constant-head permeability experiments were conducted upon specimens with different aging times and diesel contents.

2.4 Experimental methods

2.4.1 Non-contaminated background soil particle gradation test

The Malvern particle size analysis method was used to test the particle size distributions of the non-contaminated background soil. This analysis was performed using a laser particle size analyzer (MAZ3000, Malvern Panalytical, UK), which measures particle sizes in the range of 0–3800 µm and allows for data processing (e.g., particle classification and numerical homogenization) [34]. The analysis was repeated for each soil sample, to ensure the accuracy of the data; the average value was calculated (n = 5). Based on the international system of soil particle size fractions, the soil particle composition was described according to the clay content: clay (< 0.002 mm), powder (0.002–0.02 mm), and sand (> 0.02 mm) [35].

2.4.2 Soil samples contaminated with diesel

The layering method was used to configure the samples of contaminated soil [36]. The procedure was as follows: First, a specified mass of dry soil was laid and then sprayed with a specified mass of water and diesel fuel, using a spray can. This was repeated until the desired contaminated soil content was achieved. After configuration, the sample was sealed in a dark environment for 7 d, to allow the diesel to be evenly distributed in the pores of the soil [2,37]. The diesel-contaminated soil was conditioned for 7 d in a dark environment and then made into soil bread (ϕ100 mm × 63.7 mm) with a dry density of 1.4 g∙cm⁻³ (accuracy controlled to ±0.02 g∙cm⁻³) via static pressing at a specified rate [22]. The specimens were sealed and aged in a dark environment [38,39].

2.5 Pore structure testing

2.5.1 Mercury intrusion porosimetry

The instrument was a fully automatic mercury piezometer (AutoPore IV 9500, Micromeritics, USA), with a pore size range of 0.003–1000 µm and a maximum pressure of 60000 psi (Fig.1). The pore size of the specimens was determined according to Washburn [40], using

(1)P= \frac{-2\mathit{\text{σ}}\text{cos}\mathit{\text{θ}}}{r},

where P is the additional pressure of the applied mercury (MPa); r is the pore radius (µm); \mathit{\text{σ}} is the surface tension of the mercury, σ = 485.0 mN/m; and \mathit{\text{θ}} is the contact angle between the introduced liquid and solid material, θ = 140°.

Preparation of MIP test samples: A cut soil block (0.5 cm × 1.5 cm × 0.5 cm) was quickly placed in liquid nitrogen (boiling point −196 °C), frozen for 1–2 h, and then dried via the freeze-drying method [41,42]. The brief process of the freeze-drying method is as follows: the cut specimen was placed into a container with liquid nitrogen and the soil sample was completely immersed and cooled rapidly to –196 °C such that the pore water in the soil sample froze and became amorphous ice; the freezing time was 1–2 h. The vacuum freeze dryer was precooled for ~1 h, and the frozen specimen was quickly placed into the drying tray of the vacuum freeze-dryer. The quick filling valve was removed, the vacuum pump was turned on, and the specimen was dried at –56 °C under a continuous vacuum for 24–30 h. The dried specimens were quickly placed in a sealed bag with a silica gel desiccant and labeled with a number for the MIP test [43].

2.5.2 X-ray micro-computed tomography

The test instrument was a Zeiss Xradia 510Versa 3D X-ray microscope (Carl Zeiss AG, Germany) with a maximum resolution of 0.7 µm, X-ray tube high voltage range of 30–160 keV, and a maximum power of 10 W. The test accuracy was set at 10 µm, based on several previous tests and instrumentation conditions. The test was conducted under a voltage of 140 V, a current of 72 µA, and an exposure time of 3.000 s. Each scan yielded ~600 two-dimensional (2D) images in .tiff format with a size of 1000 × 1024 dpi and 16-bit depth. The laminar imaging cross-section is shown in Fig.2.

Preparation of the XRμCT sample: The large specimen sizes can affect the mounting and testing accuracy during testing. Therefore, the specimens were cut under the required working conditions. Their sizes were ~ϕ39 mm × 20 mm. The CT test specimens and field tests are shown in Fig.3.

2.5.3 XRμCT image processing program

The three-dimensional (3D) visualization software Avizo was used to process the CT images [44,45]. Image processing was divided into five steps, as shown in Fig.4.

Image import: The main aim was to import the measured CT images into Avizo 3D visualization software via the Open data module. In this step, the brightness and contrast of the images were adjusted using an Orth slice. The volume-rendering module was also used to reconstruct the soil media in 3D.

Denoising: This consisted of two steps: The first substep was beam-hardening correction, which was performed using the beam-hardening correction module. The second step removed noise from the graph. The median filtering method was the most effective way to remove the noise; it was performed using a filter sandbox module.

Threshold segmentation: This step focused on the identification of intermediate pores using correlation functions, as well as segmentation of the pores. Avizo contains many thresholding methods, including interactive thresholding, watershed segmentation, interactive top-hat segmentation, mathematical morphology, texture-based segmentation, and deep-learning-based segmentation. From experiments conducted using various methods, the interactive thresholding segmentation method was found to achieve the optimal results. The threshold source used for threshold segmentation was a commonly used algorithm [46].

Determination of ROI_REV (selected representative elemental volume (REV) ROI in the image) representative elementary volume: owing to the heterogeneity of the soil pore structure, it is not possible to accurately characterize soil pore structures [47]. To address this problem, the concept of a representative elementary volume has been proposed in Ref. [48]; they concluded that the obtained pore structure parameters fluctuated when the computational unit was smaller than the elementary volume. When the size of the calculated unit exceeded that of the elementary volume, the obtained pore structure parameters remained constant and could be used as the basic calculation unit for the pore structure parameters of the response soil [49,50]. In this study, volumetric porosity was used as an indicator to characterize the ROI_REV size [47]. The main steps in the determination of the ROI_REV representative elementary volume were as follows: for operational purposes, the largest cube was cut out of the threshold split soil core, and this cube was used as the basis for ROI_REV area size determination. Any point within this cube was chosen as the center, a cube of a certain side length was cut, and its porosity was measured. Then, cubes of different side lengths were cut off at the center of this point, and the porosity was calculated. Finally, the variation curve of the porosity with respect to the side length of the cube was calculated, and the length of the cube corresponding to a stable porosity was determined. These steps were repeated by identifying another central point in the largest cube [51]. Fig.5 presents a graph of the variation in the volumetric porosity with respect to cube size for the sample corresponding to an aging time of 7 d and a diesel doping content of 5.00 wt.%. The graph shows that the volumetric porosity tends to exhibit a constant value at a cube size of ~450 voxels. Therefore, the size of the ROI_REV representative elementary volume for this condition was determined as 450 voxels. The size of the volumetric porosity corresponding to a constant voxel for the other conditions is listed in Tab.3.

Pore structure analysis: all pore structure parameters were obtained based on the ROI_REV basic unit cell; the calculation method is described in the next section.

2.5.4 XRμCT pore structure parameters

Porosity: the porosity of the ROI_REV area was calculated according to Ref. [47].

(2)\phi =\frac{V_{\text{pore}}}{V_{\text{total}}}\times 100\mathrm{\%},

where \phi is the porosity, V_{\text{pore}} is the volume of all pore voxels in the ROI_REV characterization unit, and V_{\text{total}} is the total volume of the ROI_REV characterization unit.

Equivalent and average diameters: in avizo, the standard sphere is equal to the volume of irregular pores, and the diameter of the pores is replaced by the standard sphere; thus, we have

(3)EqD=\sqrt[3]{\frac{\text{6×}{V}_{\text{pore}}}{\text{π}}},

(4)AvD=\frac{\sum _1^n{EqD}_i}{\sum _{\text{1}}^{n}i},

where EqD is the equivalent diameter of the pores in the soil porous medium, AvD is the average pore diameter, and {EqD}_i is the equivalent diameter of numbered pores i, where i = 1,2,3,...,n.

Shape factor: the shape factor refers to the shape characteristic of the pore relative to the sphere; it is defined as [52]

(5)SF=\frac{A_{\text{pore}}^{\text{3}}}{36\text{π}{V}_{\text{pore}}^{\text{2}}},

where SF is the shape factor, and A_pore is the surface area of the pores.

Degree of anisotropy: anisotropy indicates whether a pore has a directional dependence. It is calculated as [27]

(6)DA=\text{1}-\frac{{MIL}_{\text{min}}}{{MIL}_{\text{max}}},

where DA is the anisotropy value, {MIL}_{\text{min}} is the minimum mean intercept length, and MI{L}_{\text{max}} is the maximum mean intercept length, respectively.

Fractal dimension: in the analysis software, the fractal dimension describes the roughness of the pore surface. It is calculated as [27]

(7)\text{log}{N}_r\left(A\right) =d \text{log}\left(\text{1}/r\right)+\text{log}k,

where A represents the cover object of the box ROI_REV, r is the edge length of the box, N_{\text{r}}\left(A\right) is the number of boxes of edge length r required to cover the object ROI_REV area, d is the box dimension, and k is a positive number.

Pore−throat ratio: the pore−throat ratio reflects the homogeneity of the pore structure network model. It is defined as [53]

(8)PTR=\frac{R_{\text{pore}}}{R_{\text{throat}}},

where PTR refers to the pore−throat ratio, R_{\text{pore}} is the radius of the pore, and R_{\text{throat}} is the radius of a pore−throat adjacent to the pore.

Euler number: the Euler number is a measure of the strength of the pore connectivity; it is defined as

(9)E_{\text{v}}=\frac{I-C}{V_{\text{total}}},

where I is the volume of the unconnected pores in the ROI_REV representative elementary volume, C is the volume of the connected pores in the ROI_REV representative elementary volume, and V_{\text{total}} is the total volume of the ROI_REV representative elementary volume.

Tortuosity: tortuosity is a measure of the degree of curvature of the pore channel. It is defined as [54,55]

(10)\mathrm{\Gamma }=\frac{L_{\text{t}}}{L_{\text{o}}} ,

where Γ is the tortuosity, L_{\text{t}} is the actual length of the permeability channel, and L_{\text{o}} is the macroscopic straight length.

2.6 Permeability test

The instrument used was a fully automatic solidification permeameter from GDS (GDS-Instruments-Ltd. 2002), UK [56]. The permeation head pressure was maintained at a constant value. The operation was divided into two steps. 1) Backpressure saturation: for this, the specimens were aged to the set time, cut to the specified size of the permeameter (ϕ76.8 mm × 20 mm), and mounted on the permeameter for back pressure saturation. Backpressure saturation and permeability tests were performed by coordinating the control of axial, back, and base pressures. In the case of backpressure saturation, it was necessary to close the base pressure valve and strictly fix the pressure difference between the axial and back pressures; in general, the axial pressure was 10 kPa higher than the back pressure. When the B value exceeded 0.98, the back-pressure saturation was completed. 2) Permeability testing: during the permeability test, it was necessary to ensure that the base and back pressures were less than the axial pressure (axial pressure > back pressure > base pressure) and that the back pressure was generally 10 kPa less than the axial pressure. The pressure difference between the base and back pressures depended on the type and characteristics of the specimen. The end of the permeation test was indicated by an equal water flow between the back and base pressure controllers. After several attempts, the testing effect was found to be optimal when the difference between the back and base pressures was 5 kPa. Three parallel experimental groups were used for each condition.

The permeability coefficient was calculated as

(11)k= \frac{{\mathit{\text{γ}}}_{\mathrm{w}}QH}{\mathrm{\Delta }pAt},

where {\gamma }_{\mathrm{w}} is the gravity of water, Q is the flow rate through the specimen water, A is the cross-sectional area of the specimen, and H is the height of the specimen.

2.7 Mechanistic investigation

Agglomerate size distribution testing: for this, the particle size distribution was measured using a sedimentation method [57]. The soil particle size was classified according to the National Standard of the People’s Republic of China for Work Test Methods (GB/T50123-2019), as follows: clay (< 0.005 mm), silt (0.005–0.075 mm), and sand (> 0.075 mm). In this study, the particle size distribution testing method used for the diesel-contaminated soils was chosen via a comparison of several methods. Initially, three methods were used to test the particle size distribution of the contaminated soils: Malvern particle size analysis [58,59], sieving [60] and sedimentation [57]. Malvern particle size analysis uses high-speed rotation and ultrasonication to disperse soil particles and then measures the particle sizes using a laser [61]. Although this method provides comprehensive and detailed results regarding particle size gradation, the high-speed rotation and ultrasonication can disrupt the diesel-soil structure and affect the accuracy of the results. The sieving method of particle gradation testing presupposes that the soil sample is dry and therefore requires soil baking. However, when baking diesel-contaminated soils, diesel fuel is partially removed along with the water; this has a significant effect on the diesel-soil structure and the accuracy of the test results. For the sedimentation method, although the soil was agitated during pre-dispersion, an effect was observed on the diesel-soil structure; however, this was much less intense than the agitation of the soil in the Malvern particle size method. In the sedimentation experiments, the water-in-diesel-soil structure was stable, because more water than oil was present, and the oil was more viscous than water [62]; this affected the test results. Therefore, the diesel-soil structure was more stable. Comparing the three particle gradation testing methods, it was found that the sedimentation method was the least affected. Therefore, in this study, the sedimentation method was used for diesel-contaminated soil particle gradation testing.

The ammonium acetate method was used to measure the CECs of the specimens [63].

Pore diesel content tests: A certain quantity of diesel-contaminated soil was soaked in deionized water, tumbled, shaken [64], and centrifuged to obtain the supernatant. Infrared spectrophotometry was then used to measure its diesel content [65]. The infrared spectrophotometer model was TANGO-R, and the wavelength range was 11500–4000 cm⁻¹.

2.8 Data analysis

The porosity, average pore diameter, pore size distribution, and pore connectivity indicators (tortuosity) measured by MIP were analyzed using a two-factor ANOVA method [66]. The XRμCT results for the porosity, average diameter, pore size distribution, morphological parameters (anisotropy, shape factor, and fractal dimension), and pore connectivity indicators (Euler number, tortuosity, and pore−throat ratio) were analyzed using one-way ANOVA [58,67]. Furthermore, the correlation between the pore structure parameters and soil permeability characteristic parameters was determined using Pearson’s coefficient [59,68]. The analysis software used was SPSS (version 26.0) [69]. Origin 8.0 and Microsoft Office PowerPoint were used for image creation and manipulation.

The SPSS 26.0 General Linear Model univariate was used to perform two-factor analysis of the obtained data. To calculate the correlation parameters [36], we used the following expressions.

Variation rate:

(12)T=({TSS}_{\text{Aging}}+{TSS}_{\text{diesel}}+{TSS}_{\text{Aging}-\text{diesel}})/{TSS}_{\text{CT}}.

Pore structure changes attributable to the proportion of aging effects:

(13)F_1={TSS}_{\text{Aging}}/{TSS}_{\text{CT}}.

Pore structure changes attributable to the proportion of diesel content:

(14)F_2={TSS}_{\text{diesel}}/{TSS}_{\text{CT}}.

Pore structure changes attributable to the interactions between aging and diesel content:

(15)F_{\text{3}}={TSS}_{\text{Aging}-\text{diesel}}/{TSS}_{\text{CT}}.

In the above, {TSS}_{\text{Aging}} is the third type of the sum of square values corresponding to high-temperature action, {TSS}_{\text{diesel}} is the third type of the sum of square values corresponding to diesel content, {TSS}_{\text{Aging}-\text{diesel}} is the third type of the sum of squares corresponding to the interaction between heating temperature and diesel fuel content, and {TSS}_{\text{CT}} is the third sum of squares corresponding to the corrected total.

3 Results and discussion

3.1 Results of MIP

In our study, the apparent porosity, average diameter, pore size distribution, and tortuosity measured by MIP differed for different aging times and diesel contents (Fig.6). The apparent porosity of the soil was reduced by aging under the same diesel content, and the apparent porosity was reduced from 48.81% (S.D. = 0.84, 7 d) to 44.35% (S.D. = 0.56, 90 d) at 5 wt.% diesel. The apparent porosity change was divided into two stages: a greater change in apparent porosity below 30 d of aging and a slower change in apparent porosity at 30–90 d of aging. A reduction in apparent porosity means that fewer pores are connected to the outside world, and their pore connectivity is simultaneously reduced. A similar trend in the average diameter variation was observed. Again, using 5 wt.% diesel content as an example, the average pore size measured by MIP decreased from 60.35 nm (S.D. = 2.25, 7 d) to 32.09 nm (S.D. = 2.72, 90 d). An aging time of 30 d represented the cutoff point between the two phases. The tortuosity exhibited the opposite trend: it increased with increasing aging time. The tortuosity increased from 1.60 (S.D. = 0.016, 7 d, 5.0 wt.%) to 1.81 (S.D. = 0.014, 90 d, 5.0 wt.%). Pore size distribution analysis showed that the total pore size consisted primarily of pore size classes of 20–100 nm and > 5000 nm, followed by 5–20 nm and 100–5000 nm (Fig.7 shows the pore size distribution for a diesel content of 5.0 wt.% and an aging time of 30 d; the pore size percentage distributions for all working conditions are shown in Fig.8). The proportions of pore size classes of 5–20 nm and > 5000 nm in the soil decreased, and the proportions of pore size classes of 20–100 nm and 100–5000 nm increased with aging. Aging time led to a more homogeneous pore size distribution in the soil. Aging reduced the pore space within the soil and increased the extent of pore curvature. At a constant aging time, an increase in the diesel content led to an increase in the apparent porosity and average diameter, as well as a decrease in tortuosity. When the aging time was fixed at 30 d, the apparent porosity increased from 45.23% (S.D. = 0.23, 0.0 wt.%) to 47.01% (S.D. = 0.73, 7.0 wt.%). The average diameter increased from 28.17 nm (S.D. = 0.19, 0.0 wt.%, 30 d) to 55.14 nm (S.D. = 4.33, 7.0 wt.%, 30 d); the tortuosity value increased from 1.84 (S.D. = 0.016, 0.0 wt.%, 30 d) to 1.71 (S.D. = 0.019, 7.0 wt.%, 30 d). The pore size distribution analysis revealed that the increase in diesel content reduced the percentages of pore size classes 5–20 nm and 20–100 nm and increased the percentages of pore size classes 100–5000 nm and > 5000 nm.

A two-factor analysis of the experimental results revealed that the effects of aging time, diesel content, and the interaction there between on the MIP pore structure (the apparent porosity, mean pore size, and tortuosity measured by MIP are collectively referred to as the MIP pore structure) were all statistically significant (p value < 0.05, Tab.4–Tab.6), and that the statistical model reflected the observed rate of variation. The aging time accounted for F₁ = 59.48% of the effect degree for causing the variation in apparent porosity, diesel content accounted for F₂ = 23.65%, and the interaction of the two accounted for F₃ = 8.66%. The statistical model reflected the observed variation rate of t = 91.67%. In terms of average pore diameter, the effect of aging upon the variation in average diameter was reduced, and the value decreased to F₁ = 45.07%. The effect of diesel content was enhanced to F₂ = 48.8%, and the effect degree of the interaction between the two accounted for F₃ = 2.85%. The statistical model accounted for t = 96.72% of the observed variance. The effect of aging upon the variation in tortuosity was F₁ = 53.35%, the effect of diesel content was F₂ = 31.55%, and the interaction between the two was F₃ = 8.33%. The statistical model showed t = 93.23% for the observed variability. In terms of pore size distribution, the effect of aging time upon the variation in pore size classes 5–20 nm, 20–100 nm, 100–5000 nm, and > 5000 nm were F₁ = 16.33% and 0.86% (not significant), 6.54%, and 25.17%, respectively, and the effects of diesel content were F₂ = 79.11%, 84.09%, 79.16%, and 65.28%, respectively. The effect levels for the interaction between the two were F₃ = 1.48%, 12.03%, 6.68%, and 2.50%, respectively. The statistical model reflected the observed variability at t = 96.92%, 96.98%, 92.39%, and 92.95%, respectively. The results of the above analyses show that the aging time, diesel content, and interaction there between have significant effects on the variation in the measured apparent porosity, average pore diameter, and tortuosity measured by MIP. The aging time and diesel content had a greater individual effect than the interaction between aging time and diesel content.

The correlation matrix (Tab.7) shows a strong positive correlation between the apparent porosity and average pore diameter obtained by MIP measurements (r = 0.95, p = 0.01). This is because a reduction in apparent porosity implies a reduction in the pore space and an overall contraction of pores within the soil. The apparent porosity was strongly negatively correlated with the tortuosity obtained from MIP measurements (r = 0.82, p = 0.01), and the average diameter was correlated with tortuosity via a coefficient of r = 0.86 (p = 0.01). This indicates a reduction in the space within the soil, altering the arrangement of the soil particles and increasing the degree of curvature of the pores. The correlation coefficients between each pore size class and the apparent porosity measured by MIP were 0.02 (5–20 nm), 0.26 (20–100 µm), 0.26 (100–5000 µm), and 0.88 (> 5000 µm); the correlation coefficients between the pore size classes and the average diameters measured by MIP were 0.23 (5–20 nm), 0.45 (20–100 nm), 0.45 (100–5000 nm), and 0.94 (> 5000 nm). This suggests that the reduction in apparent porosity measured by MIP was mainly due to a reduction in the proportion of pores corresponding to pore size class > 5000 nm; the reduction in average diameter measured by MIP was attributable to a reduction in the proportion of pores corresponding to pore size classes 100–5000 nm and > 5000 nm.

3.2 Results of XRμCT

In this experiment, the XRμCT pore structure was evaluated after three aging periods (7, 30, and 90 d; 5.0 wt.%) using the XRμCT technique (the true porosity, mean pore size, pore size distribution, shape factor, anisotropy, fractal dimension, Euler number, tortuosity, and pore−throat ratio are collectively referred to as the XRμCT pore structure). The CT slice images were processed using Avizo visualization software, and the results are shown in Tab.8 and Fig.9. According to XRμCT, the true porosities were 19.13% (S.D. = 0.26, 7 d), 15.42% (S.D. = 0.35, 30 d), and 13.25% (S.D. = 0.47, 90 d); the average pore size decreased from 16.76 µm (S.D. = 0.21, 7 d) to 12.66 µm (S.D. = 0.25, 90 d); and the tortuosity increased from 1.72 (S.D. = 0.03, 7 d) to 2.75 (S.D. = 0.01, 90 d). In terms of pore size distribution, because of the difference in testing accuracy between the two pore structure testing methods (MIP test accuracy: 3 nm; XRμCT test accuracy: 10 µm), the pore size distribution results covered different pore size intervals and could not be fully compared and analyzed. Therefore, a comparative analysis was performed by intercepting the aperture interval covered by both test methods. The analysis revealed (Fig.10(a)) that the difference in porosity values between MIP and XRμCT was not significant in the pore size range 10–500 µm; the trends were identical in this class; therefore, the test results of MIP and XRμCT were mutually validated. As described in the study by Wang et al. [30], the pore volume contribution was used as a metric to measure the change in MIP and CT with respect to the pore size distribution under an increasing aging time (Fig.10). As can be seen from the graph, in the pore size range of 10–100 µm, the differences between the pore volume contributions of MIP and XRμCT were large but exhibited the same (decreasing) trend under increasing aging time. The difference in the pore volume contribution between MIP and XRμCT decreased in the pore size interval of 100–300 µm; both showed an increasing trend under increasing aging time. No corresponding pore size pores were observed in the pore size interval 300–500 µm in the MIP measurements; meanwhile, the pore volume contribution in the pore size interval 300–500 µm in the XRμCT measurements decreased under increasing aging time. Therefore, a large difference was observed in the magnitude of the pore size distribution values for MIP and XRμCT as the aging time increased, though they followed a similar trend. Compared with the MIP results, the trends of the pore structure parameters with respect to aging time were similar, and the test results were mutually verified.

In this study, shape factor, anisotropy, and fractal dimensions were used to characterize the pore morphology of the soil. The shape factor value of the pores decreased from 1.473 (S.D. = 0.031, 7 d) to 1.014 (S.D. = 0.009, 90 d) under the effects of aging time. Aging rendered the pore morphologies more spherical. The orientation of the pores (anisotropy) became increasingly homogeneous (i.e., the anisotropy values decreased) under aging. The correlation coefficient (R²) between the logarithm of the number of boxes (N(\epsilon)) and box size (\epsilon) always exceeded 0.985 (data not shown) when calculating the fractal dimension. This indicates that the analysis yields acceptable values for the fractal dimension, and that the values are all in the range of 2–3. The change in fractal dimension values indicates that aging resulted in a smoother pore surface (lower fractal dimension). The fractal dimension showed a significant positive correlation (R² = 0.9815) with the total true porosity identified via CT (Fig.11). The Euler number was used to characterize the connectivity of the XRμCT pore structure. The tests showed that the Euler number of the soil XRμCT pore structure gradually increased with aging time. This indicates that the pore connectivity was weakened by the aging effect, which broke the connections between the pores. The pore−throat ratio of the soil decreased as a result of aging. Based on the definition of the pore−throat ratio, a decrease in the value indicates that the pore diameter becomes more similar to the pore−throat diameter. This indicates that the pore size of the soil became more homogeneous through aging. To determine the measurement accuracy of XRμCT, the pores were classified into four pore size classes: 10–100, 100–300, 300–500, and 500–700 µm (Fig.9). The pore size classes that varied significantly during the three aging periods were 10–100 µm (p≤ 0.01), 100–300 µm (p = 0.02 < 0.05), and 300–500 µm (p≤ 0.01). The percentage of pore size classes 10–100 and 100–300 µm gradually increased under increasing aging time, whereas the percentage of pores in the pore size classes 300–500 µm decreased with increasing aging time. The results identified by XRμCT can be compared and cross-validated with the test results obtained by MIP, in addition to yielding numerous other pore structure parameters. As reported in previous studies [31,66], MIP and microscale CT techniques can be combined to analyze a variety of soil pores.

3.3 Micromorphological evolution

Fig.12 shows the microscopic morphologies of the specimens aged for 7, 30, and 90 d at 5.0 wt.% diesel content. A comparative analysis showed that the microstructure changed as the aging time increased. At 7 d, the photograph shows a clear boundary between the particles and a loose lap between them. The coarse particles (yellow circles) were separated by large pore spaces (red boxes). When the aging time increased to 30 d, the gaps decreased, and the particle laps became more compact; however, larger pores were present (red box). The areas of the pores (black) in the photographs decreased and the true porosity decreased. After 90 d of aging, the gaps between the particles in the soil became more compact, and the coarser particles (yellow circles) appeared to be submerged or buried, significantly reducing the pore size (red box). The areas occupied by the pores (black) were again reduced in the photograph, and the true porosity continued to decrease.

4 Discussion

4.1 Mechanisms affecting changes in pore structure

The free diesel content in the pores of diesel-contaminated soils and the CEC were tested for a 5.0 wt.% diesel admixture at different aging times, to characterize the changes in diesel and soil particles during aging. The results are shown in Fig.13 and Fig.14. As can be seen in Fig.13, the free diesel content in the pores decreased as the aging time increased (from 45740.01 µg/g (7 d) to 29635.70 µg/g (9 d)), indicating that the soil particles are constantly “adsorbing” the free diesel from the pores during aging. The longer the aging time, the smaller the cation exchange; this is consistent with the results of Ref. [2]. A smaller CEC means that the hydration of the soil particles is less intense, and the amount of water bound to the surface of the soil particles is reduced. Owing to the increased adsorption of diesel by soil particles, an increasing number of soil particles are exposed to diesel and increase the number of adsorption sites on the surface of the soil particles through displacement and mass transfer [70,71]; this reduces the amount of bound water on the surfaces of the soil particles, ultimately leading to a reduction in the cohesion between soil particles [72]. This causes the agglomerates formed by diesel adsorption prior to the preparation of the diesel-contaminated remodeling samples to break down [10,15,73,74] into smaller agglomerates. This can be seen in the variations in agglomerate size and number for the diesel-contaminated soil measured using the sedimentation method. The smaller agglomerates formed by the reduced combined water content and fragmentation resulted in a tighter arrangement, which ultimately led to a reduction in the porosity of the diesel-contaminated soil [75,76].

The higher the diesel content, the greater the agglomeration effect upon the soil particles (large agglomerate content increases with increasing diesel content whereas small agglomerate content decreases, see Fig.15), which is consistent with the results of Refs. [62,75], and the lower the bound water content (cation exchange decreases under increasing diesel content). Together, these two changes led to a reduction in the porosity of diesel-contaminated soils. The analysis suggests that changes in the agglomerate size and number affected the pore structure to a greater extent than bound water. Specifically, the increase in pore space attributable to significant changes in agglomerate size and number under increased diesel admixture and content surpassed the decrease in pore space attributable to the rearrangement of soil particles brought about by the reduced bound water content. The change in porosity matches the change in porosity produced by the change in the agglomerate size and number.

4.2 Permeability characteristics

In our study, both aging time and diesel content weakened the permeation characteristics (Fig.16). Correlation analysis revealed a significant correlation (p = 0.01, p = 0.05) between the MIP pore structure parameters and permeability coefficients under different aging times (Tab.9). The permeability coefficient K_sa decreased under decreasing porosity and average diameter (r = 0.97, r = 0.95, p = 0.01), which is consistent with the findings of Nishiyama and Yokoyama [77] and Zhang et al. [78]. It also increased with decreasing tortuosity (r = −0.86, p = 0.01). K_sa was significantly positively correlated with porosity for pore size classes > 5000 µm and 20–100 µm (r = 0.98, r = 0.96, p = 0.01) and negatively correlated with apparent porosity for pore size class 100–5000 µm (r = −0.67, p = 0.05). The decrease in water channel volume (apparent porosity and average diameter) and increase in pore curvature within the soil reduces the soil’s intrinsic permeability [79] and increases the retention capacity of water and oil within the soil [78]; these factors weaken the permeability properties of the soil. Under the effect of diesel content, K_sc decreased under increasing apparent porosity and average diameter (r = −0.71, p = 0.05; r = −0.83, p = 0.01), increased with increasing tortuosity (r = 0.94, p = 0.01), was positively correlated with apparent porosity for pore size classes 5–20 µm and 20–100 µm (r = 0.94, r = 0.92, p = 0.01), and was negatively correlated with apparent porosity for pore size classes 100–5000 µm and > 5000 µm (r = −0.93, r = −0.94, p = 0.01). There is uncertainty whether the pore structure parameters (e.g., apparent porosity and average diameter) correlate with K_sc. The reason for this uncertainty may be related to the diesel in the soil pores. The higher the diesel content incorporated in the soil, the greater the content of free diesel present in the soil pores. The stronger the impediment to water permeability, the lower the coefficient of permeability [80–82].

5 Conclusions

To determine the effects of aging and diesel content upon clay pore structures, measurements were conducted using MIP and XRμCT techniques. Experimental studies were performed upon the effects of aging and diesel content, as well as the permeability characteristics. The significant findings are summarized below.

1) The combination of MIP and XRμCT facilitates the measurement and comprehensive characterization of the soil pore structure. The test results are validated by their mutual agreement.

2) Both the aging and diesel content change the pore structure. At constant diesel levels, increasing the aging time reduces the apparent porosity and average diameter of the soil, and the pore morphology becomes more spherical (reduced shape factor values) and isotropic (MIP). The reduction in apparent porosity means lower connectivity with the outside world, and the increased aging time also reduces the connectivity of soil pores. The true porosity, average diameter, pore size distribution, tortuosity, and connectivity measured by XRμCT exhibit similar trends to those measured by MIP. At the same time, XRμCT tests showed that a longer aging time renders the pores more spherical and isotropic. When the aging time was stable, the increase in diesel content increased the pore space within the soil (increase in porosity and average diameter) and decreased the degree of tortuosity.

3) Aging causes the soil particles to adsorb more diesel fuel, resulting in lower cohesion between particles and a lower bound water content. This leads to the subsequent fragmentation of the agglomerates formed by the adsorption of diesel fuel prior to the preparation of the diesel-contaminated remolded samples, with the larger agglomerates breaking down into smaller ones. The smaller agglomerates produced by the reduced combined water content and agglomerate fragmentation resulted in a tighter arrangement, which ultimately reduced the porosity of the diesel-contaminated soil. The higher the diesel content, the stronger the aggregation effect. The larger the size of the agglomerates, the larger the internal voids in the soil, and the greater the porosity.

4) Both aging and diesel dosing weakened soil permeability. The permeability coefficient K_sa decreased under increasing aging time for a constant diesel fuel level. Correlation analysis revealed a strong correlation with the variation in pore structure parameters. When the aging time was stable, the permeability coefficient K_sc decreased under increasing diesel content. The correlation between the pore structure parameters and permeability coefficient K_sc was inconsistent with this fact. The analysis suggests that diesel in the pores of the soil increase the resistance to percolation and weakens the permeability properties.

References

[1] Delage P, Marcial D, Cui Y J, Ruiz X. Ageing effects in a compacted bentonite: A microstructure approach. Geotechnique, 2006, 56(5): 291–304

[2] Wei Y, Han J, Wang Y, Wang H, Sun Y, Yan B. Effects of oil pollution on water movement in soils with different textures. Water, Air, and Soil Pollution, 2020, 231(4): 1–12

[3] Rajabi H, Sharifipour M. Geotechnical properties of hydrocarbon-contaminated soils: A comprehensive review. Bulletin of Engineering Geology and the Environment, 2019, 78(5): 3685–3717

[4] Kuppusamy S, Thavamani P, Venkateswarlu K, Lee Y B, Naidu R, Megharaj M. Remediation approaches for polycyclic aromatic hydrocarbons (PAHs) contaminated soils: Technological constraints, emerging trends and future directions. Chemosphere, 2017, 168: 944–968

[5] Piña J, Merino J, Errazu A F, Bucalá V. Thermal treatment of soils contaminated with gas oil: Influence of soil composition and treatment temperature. Journal of Hazardous Materials, 2002, 94(3): 273–290

[6] Vidonish J E, Zygourakis K, Masiello C A, Sabadell G, Alvarez P J J. Thermal treatment of hydrocarbon-impacted soils: A review of technology innovation for sustainable remediation. Engineering, 2016, 2(4): 426–437

[7] Koestel J, Dathe A, Skaggs T H, Klakegg O, Ahmad M A, Babko M, Giménez D, Farkas C, Nemes A, Jarvis N. Estimating the permeability of naturally structured soil from percolation theory and pore space characteristics imaged by X-ray. Water Resources Research, 2018, 54(11): 9255–9263

[8] Tsakiroglou C D, Ioannidis M A. Dual-porosity modelling of the pore structure and transport properties of a contaminated soil. European Journal of Soil Science, 2008, 59(4): 744–761

[9] Koestel J, Larsbo M, Jarvis N. Scale and REV analyses for porosity and pore connectivity measures in undisturbed soil. Geoderma, 2020, 366: 114206

[10] Izdebska-Mucha D, Trzciński J. Clay soil behaviour due to long-term contamination by liquid petroleum fuels: Microstructure and geotechnical properties. Bulletin of Engineering Geology and the Environment, 2021, 80(4): 3193–3206

[11] DrummondCIsraelachvili J. Fundamental studies of crude oil-surface water interactions and its relationship to reservoir wettability. Journal of Petroleum Science Engineering, 2004, 45(1−2): 61−81

[12] Nazir A K. Effect of motor oil contamination on geotechnical properties of over consolidated clay. Alexandria Engineering Journal, 2011, 50(4): 331–335

[13] Ijimdiya T S. Effect of oil contamination on particle size distribution and plasticity characteristics of lateritic soil. Advanced Materials Research, 2011, 367: 19–25

[14] Khosravi E, Ghasemzadeh H, Sabour M R, Yazdani H. Geotechnical properties of gas oil-contaminated kaolinite. Engineering Geology, 2013, 166: 11–16

[15] Kaya A, Fang H Y. The effects of organic fluids on physicochemical parameters of fine-grained soils. Canadian Geotechnical Journal, 2000, 37(5): 943–950

[16] Ghadyani M, Hamidi A, Hatambeigi M. Triaxial shear behaviour of oil contaminated clays. European Journal of Environmental and Civil Engineering, 2019, 23(1): 112–135

[17] Olgun M, Yıldız M. Effect of organic fluids on the geotechnical behavior of a highly plastic clayey soil. Applied Clay Science, 2010, 48(4): 615–621

[18] Nasehi S A, Uromeihy A, Nikudel M R, Morsali A. Influence of gas oil contamination on geotechnical properties of fine and coarse-grained soils. Geotechnical and Geological Engineering, 2016, 34(1): 333–345

[19] Vogel H J. Morphological determination of pore connectivity as a function of pore size using serial sections. European Journal of Soil Science, 1997, 48(3): 365–377

[20] Radulovich R, Solorzano E, Sollins P. Soil macropore size distribution from water breakthrough curves. Soil Science Society of America Journal, 1989, 53(2): 556–559

[21] Watson K W, Luxmoore R J. Estimating macroporosity in a forest watershed by use of a tension infiltrometer1. Soil Science Society of America Journal, 1986, 50(3): 578–582

[22] Li X, Zhang L M. Characterization of dual-structure pore-size distribution of soil. Canadian Geotechnical Journal, 2009, 46(2): 129–141

[23] NakashimaYMitsuhata YNishiwakiJKawabeYUtsuzawaS JinguujiM. Non-destructive analysis of oil-contaminated soil core samples by X-ray computed tomography and low-field nuclear magnetic resonance relaxometry: A case study. Water, Air, and Soil Pollution, 2011, 214(1−4): 681−698

[24] Al-Raoush R I. Experimental investigation of the influence of grain geometry on residual NAPL using synchrotron microtomography. Journal of Contaminant Hydrology, 2014, 159: 1–10

[25] Javanbakht G, Arshadi M, Qin T, Goual L. Micro-scale displacement of NAPL by surfactant and microemulsion in heterogeneous porous media. Advances in Water Resources, 2017, 105: 173–187

[26] Guan Z, Tang X, Nishimura T, Katou H, Liu H, Qing J. Surfactant-enhanced flushing enhances colloid transport and alters macroporosity in diesel-contaminated soil. Journal of Environmental Sciences (China), 2018, 64: 197–206

[27] Dal FerroNCharrier PMorariF. Dual-scale micro-CT assessment of soil structure in a long-term fertilization experiment. Geoderma, 2013, 204–205: 84−93

[28] Wei T, Fan W, Yuan W, Wei Y, Yu B. Three-dimensional pore network characterization of loess and paleosol stratigraphy from South Jingyang Plateau, China. Environmental Earth Sciences, 2019, 78(11): 1–15

[29] LiPShaoS. Can X-ray computed tomography (CT) be used to determine the pore-size distribution of intact loess? Environmental Earth Sciences, 2020, 79(1): 1–12

[30] Wang X, Pan J, Wang K, Ge T, Wei J, Wu W. Characterizing the shape, size, and distribution heterogeneity of pore-fractures in high rank coal based on X-ray CT image analysis and mercury intrusion porosimetry. Fuel, 2020, 282: 118754

[31] Wu S, Yang J, Yang R, Zhu J, Liu S, Wang C. Investigation of microscopic air void structure of anti-freezing asphalt pavement with X-ray CT and MIP. Construction & Building Materials, 2018, 178: 473–483

[32] Hashemi M A, Massart T J, Salager S, Herrier G, François B. Pore scale characterization of lime-treated sand-bentonite mixtures. Applied Clay Science, 2015, 111: 50–60

[33] Supriyo H, Matsue N, Yoshinaga N. Chemical and mineralogical properties of volcanic ash soils from Java. Soil Science and Plant Nutrition, 1992, 38(3): 443–457

[34] MonteroE. Rényi dimensions analysis of soil particle-size distributions. Ecological Modelling, 2005, 182(3−4): 305−315

[35] CambardellaC AElliottE T. Methods for physical separation and characterization of soil organic matter fractions. Geoderma, 1993, 56(1−4): 449−457

[36] Chun Y, Du J, Zhou D, Liu Z, Liang T, Qin J, Su J. Effect of thermal treatment on the expansion characteristics of the diesel-contaminated soil and its mechanism. Bulletin of Engineering Geology and the Environment, 2022, 81(4): 1–16

[37] LeeJ KPark DKimBDongJ ILeeS. Remediation of petroleum-contaminated soils by fluidized thermal desorption. Waste Management, 1998, 18(6−8): 503−507

[38] Estabragh A R, Ghayamara F, Soltanian M R, Babalar M. Effect of ageing on the properties of a clay soil contaminated with glycerol. Geomechanics and Geoengineering, 2022, 17(2): 586–597

[39] Estabragh A R, Amini M, Soltanian M R. Aging effects on the swelling behavior of compacted bentonite. Bulletin of Engineering Geology and the Environment, 2020, 79(5): 2341–2352

[40] Washburn E W. Note on a method of determining the distribution of pore sizes in a porous material. Proceedings of the National Academy of Sciences of the United States of America, 1921, 7(4): 115–116

[41] Burton G J, Pineda J A, Sheng D, Airey D. Microstructural changes of an undisturbed, reconstituted and compacted high plasticity clay subjected to wetting and drying. Engineering Geology, 2015, 193: 363–373

[42] Riley R G, Thompson C J, Huesemann M H, Wang Z, Peyton B, Fortman T, Truex M J, Parker K E. Artificial aging of phenanthrene in porous silicas using supercritical carbon dioxide. Environmental Science & Technology, 2001, 35(18): 3707–3712

[43] Klimenko A, Molinier V, Bourrel M. Mechanisms underlying the adhesion of crude oil to mineral surfaces: Relevance of oil-brine interactions. Journal of Petroleum Science Engineering, 2020, 190: 107036

[44] Gan M, Zhang L, Wang Y, Mei K, Fu X, Cheng X, Bai M, Liu H, Li X. 3D micro-structural changes of an artificial flow channel in wellbore cement under geologic CO₂ storage conditions: Combined effect of effective stress and flow. Construction & Building Materials, 2022, 325: 126761

[45] Feng Y, Wang J, Bai Z, Reading L, Jing Z. Three-dimensional quantification of macropore networks of different compacted soils from opencast coal mine area using X-ray computed tomography. Soil & Tillage Research, 2020, 198: 104567

[46] Ostu N, Nobuyuki O, Otsu N. A thresholding selection method from gray level histogram. IEEE Transactions on Systems, Man, and Cybernetics, 1979, 9(1): 62–66

[47] Wu H, Yao Y, Zhou Y, Qiu F. Analyses of representative elementary volume for coal using X-ray μ-CT and FIB-SEM and its application in permeability predication model. Fuel, 2019, 254: 115563

[48] BearJ. Dynamics of Fluids in Porous Materials. New York: Courier Corporation, 1988

[49] Koestel J, Larsbo M, Jarvis N. Scale and REV analyses for porosity and pore connectivity measures in undisturbed soil. Geoderma, 2020, 366: 114206

[50] Piccoli I, Schjønning P, Lamandé M, Zanini F, Morari F. Coupling gas transport measurements and X-ray tomography scans for multiscale analysis in silty soils. Geoderma, 2019, 338: 576–584

[51] Koestel J, Larsbo M, Jarvis N. Scale and REV analyses for porosity and pore connectivity measures in undisturbed soil. Geoderma, 2020, 366: 114206

[52] Dal Ferro N, Delmas P, Duwig C, Simonetti G, Morari F. Coupling X-ray microtomography and mercury intrusion porosimetry to quantify aggregate structures of a cambisol under different fertilisation treatments. Soil & Tillage Research, 2012, 119: 13–21

[53] Cai J, Zhang Z, Wei W, Guo D, Li S, Zhao P. The critical factors for permeability-formation factor relation in reservoir rocks: Pore−throat ratio, tortuosity and connectivity. Energy, 2019, 188: 116051

[54] Jiao W, Zhou D, Wang Y. Effects of clay content on pore structure characteristics of marine soft soil. Water, 2021, 13(9): 1–19

[55] Zhang S, Yan H, Teng J, Sheng D. A mathematical model of tortuosity in soil considering particle arrangement. Vadose Zone Journal, 2020, 19(1): e20004

[56] Zou S F, Li J Z, Wang Z J, Lan L, Xie X Y. Seepage test and empirical models for soils based on GDS apparatus. Journal of Zhejiang University, 2017, 51(5): 856–862

[57] ChenHShan WHeX. Influence of diesel contamination on engineering properties of soil and its mechanism. In: Proceedings of GeoShanghai 2018 International Conference: Fundamentals of Soil Behaviours. Shanghai: Springer, 2018, 620–627

[58] Peng G, Xiang N, Lv S, Zhang G. Fractal characterization of soil particle-size distribution under different land-use patterns in the Yellow River Delta Wetland in China. Journal of Soils and Sediments, 2014, 14(6): 1116–1122

[59] Fei Y, She D, Gao L, Xin P. Micro-CT assessment on the soil structure and hydraulic characteristics of saline/sodic soils subjected to short-term amendment. Soil & Tillage Research, 2019, 193: 59–70

[60] Thomaz E L, Antoneli V, Doerr S H. Effects of fire on the physicochemical properties of soil in a slash-and-burn agriculture. Catena, 2014, 122(12): 209–215

[61] LebedevaE VFogden A. Wettability alteration of kaolinite exposed to crude oil in salt solutions. Colloids and Surfaces. A, Physicochemical and Engineering Aspects, 2011, 377(1−3): 115−122

[62] DakaM. Geotechnical properties of oil contaminated soil. Thesis for the Master’s Degree. Manchester: The University of Manchester, 2015

[63] Muñoz-Rojas M, Erickson T E, Martini D, Dixon K W, Merritt D J. Soil physicochemical and microbiological indicators of short, medium and long term post-fire recovery in semi-arid ecosystems. Ecological Indicators, 2016, 63: 14–22

[64] Hwang S, Cutright T J. The impact of contact time on pyrene sorptive behavior by a sandy-loam soil. Environmental Pollution, 2002, 117(3): 371–378

[65] Banan Khorshid Z, Mahdi Doroodmand M, Abdollahi S. UV–Vis. spectrophotometric method for oil and grease determination in water, soil and different mediates based on emulsion. Microchemical Journal, 2021, 160: 105620

[66] Dal Ferro N, Sartori L, Simonetti G, Berti A, Morari F. Soil macro- and microstructure as affected by different tillage systems and their effects on maize root growth. Soil & Tillage Research, 2014, 140: 55–65

[67] Zhao J, Chen S, Hu R, Li Y. Aggregate stability and size distribution of red soils under different land uses integrally regulated by soil organic matter, and iron and aluminum oxides. Soil & Tillage Research, 2017, 167: 73–79

[68] Pires L F, Auler A C, Roque W L, Mooney S J. X-ray microtomography analysis of soil pore structure dynamics under wetting and drying cycles. Geoderma, 2020, 362: 114103

[69] Ren J, Song X, Ding D. Sustainable remediation of diesel-contaminated soil by low temperature thermal treatment: Improved energy efficiency and soil reusability. Chemosphere, 2020, 241: 124952

[70] Pignatello J J, Xing B. Mechanisms of slow sorption of organic chemicals to natural particles. Environmental Science & Technology, 1996, 30(1): 1–11

[71] Sørgård H N, Totland C, Nerdal W, Seland J G. Investigating mobility of crude oil adsorbates on mineral surfaces by NMR. Magnetic Resonance Imaging, 2019, 56: 86–89

[72] Kermani M, Ebadi T. The effect of oil contamination on the geotechnical properties of fine-grained soils. Soil & Sediment Contamination, 2012, 21(5): 655–671

[73] Izdebska-Mucha D, Trzcińsk J, Żbik M S, Frost R L. Influence of hydrocarbon contamination on clay soil microstructure. Clay Minerals, 2011, 46(1): 47–58

[74] Kaya A, Fang H. Experimental evidence of reduction in attractive and repulsive forces between clay particles permeated with organic liquids. Canadian Geotechnical Journal, 2005, 42(2): 632–640

[75] MoavenianM HYasrobiS S. Volume change behavior of compacted clay due to organic liquids as permeant. Applied Clay Science, 2008, 39(1−2): 60−71

[76] Shi F, Zhang C, Zhang J, Zhang X, Yao J. The changing pore size distribution of swelling and shrinking soil revealed by nuclear magnetic resonance relaxometry. Journal of Soils and Sediments, 2017, 17(1): 61–69

[77] Nishiyama N, Yokoyama T. Permeability of porous media: Role of the critical pore size. Journal of Geophysical Research. Solid Earth, 2017, 122(9): 6955–6971

[78] Zhang J, Liu G, Torsaeter O, Tao S, Jiang M, Li G, Zhang S. Pore−throat structure characteristics and its effect on flow behavior in Gaotaizi tight siltstone reservoir, northern Songliao Basin. Marine and Petroleum Geology, 2020, 122: 104651

[79] Yu J R, Zhou C, Mu Q Y. Numerical investigation on light non-aqueous phase liquid flow in the vadose zone considering porosity effects on soil hydraulic properties. Vadose Zone Journal, 2022, 21(5): e20211

[80] Cabalar A F, Hasan R A. Compressional behaviour of various size/shape sand−clay mixtures with different pore fluids. Engineering Geology, 2013, 164: 36–49

[81] Zhang J, Zheng X, Chen L, Sun Y. Effect of residual oil saturation on hydrodynamic properties of porous media. Journal of Hydrology, 2014, 515: 281–291

[82] Manoharan S, Gopalakrishnan P. A review on effect of crude oil on the geotechnical properties of soil. International Journal of Research in Engineering and Technology, 2016, 3: 1234–1236

RIGHTS & PERMISSIONS

Higher Education Press