Pore-Scale Formation Characteristics of Impermeable Frozen Walls for Shallow Groundwater Contamination Remediation

Abstract

Impermeability and water blocking are crucial for remediating shallow groundwater contamination. Traditional methods often employ curtain-grouting technology to create impermeable layers. However, cement slurry curing is irreversible, leading to permanent closure of underground aquifers and secondary pollution. This study employs an innovative approach by fabricating cylindrical models that simulate actual strata and utilizing a high-temperature and high-pressure displacement device. It systematically analyzes the variations in soil pore structure, distribution, porosity, and permeability under different temperatures, pressures, and freezing durations. The microscopic characteristics of the freezing process in water-bearing soils were studied. Results demonstrate that longer freezing time improves the effectiveness of soil freezing, reaching complete freezing at temperatures as low as −4 °C for samples with low water content. For water-saturated samples, freezing below −6 °C results in nearly zero porosity. Increased pressure at a certain freezing temperature significantly reduces permeability. When freezing temperature falls below −4 °C, water permeability in saturated samples after freezing reaches near-zero levels, while unsaturated samples experience complete freezing. These findings provide a theoretical foundation for constructing freezing curtains in remediating shallow groundwater pollution.

1. Introduction

Groundwater as a water resource is an important safeguard for maintaining the virtuous cycle of water resource systems [1]. The groundwater environment has suffered serious damage, and the shallow groundwater is seriously polluted because protection measures, awareness of protection, and the improvement of pollution treatment technology are not harmonized with economic development [2]. Shallow groundwater, especially surface soil pore water, is the most affected by human activities, the most sensitive to pollution, and the most obviously controlled by hydrological, meteorological, and hydrogeological conditions [3]. Surface aquifers constitute the main aquifers for various industrial and agricultural water withdrawals and domestic water use, and a variety of human activities are closely related to this type of water-bearing medium [4]. Pollution remediation of quaternary pore aquifers is important for national economic development.

One of the main ideas of water environment remediation projects is to block the source of pollution by setting up water-blocking barriers and then remediate the contaminated groundwater [5]. Existing plugging and containment technologies are mainly used to form an impermeable barrier around the contaminated site to prevent the further spread of contaminated groundwater for remediation works [6,7,8]. At present, scholars’ research on water seepage prevention and barriers focuses on groundwater-level control [9,10,11], seepage barriers [12,13], filler materials [14,15], leakage plugging and repair [16,17,18], isolation measures [19,20], and other aspects. Zhang et al. [21] introduced the safety problems caused by groundwater seepage during coal mine construction. By analyzing the influencing factors, such as coal reserves, stability of coal pillar dams, and the evolution of water quality, groundwater extraction was carried out on the enclosing layer of the mine, and groundwater content around the engineering operation was ensured to be reduced by the method of lowering the water table, which prevented water seepage and thus ensured the safety of the construction. Ramin Vali et al. [22], through the construction process of the Isfahan metro tunnel project, found that the internal pressure of the tunnel and changes in the water table have a great impact on the surface displacement, so the impact of the water pressure on the project was avoided by installing the corresponding structures through the lining of the water pressure balance method. Alicja et al. [23] describe the pollution of groundwater channels by mine water and post-mining effluents from mining operations and propose to utilize the effects of mining of chemical raw materials such as hard coal, lignite, metal ores, iron, etc. from the mining area on the conditions of the native water. By adding relevant chemicals and fillers to the groundwater around the operation area to change the chemical properties of the groundwater, the viscosity of and dip in the groundwater will be increased, which in turn will achieve the purpose of seepage prevention and water blockage to ensure that the normal groundwater will not be contaminated. Barbu et al. [24] prepared viscosity-rich, strong, flexible, low-permeability cutoff walls composed of water–cement–bentonite mixtures, and by studying the relationship between the three ratios and using experiments, it was found that the cutoff walls obtained with a bentonite content of less than 17% were the most effective in blocking groundwater. Yuan et al. [25], taking the safe and efficient exploitation of mineral resources and the protection of regional groundwater resources from pollution as the main purpose, researched the direction of grouting curtain construction conditions, drilling structure, grouting materials, etc., and evaluated the consolidation and metamorphosis mechanism of grouting slurry under long-term high pressure and the effect of grouting curtain, and finally obtained the optimal water-blocking method. M. Vitel et al. [26] proposed a coupled use of hydrothermal and frozen tube models to solve the seepage problem of the Cigar Lake underground mine in Saskatchewan, Canada, by injecting higher-temperature and -viscosity liquids around the mine formation and then utilizing freezing technology to freeze the high-temperature liquids, resulting in the formation of a less permeable frozen wall, which prevented the seepage of groundwater.

However, while solving problems by utilizing different impermeability and water-blocking methods, other corresponding problems are also brought about. For example, although the water table-lowering method reduces the infiltration pressure of groundwater, the pressure of the protected soil relative to the surrounding soil will be higher, and as a result, it may lead to the occurrence of soil collapse and loosening problems [27]. The groundwater pressure balance method requires careful and precise calculations prior to operation, and there is a high probability that the method will fail if the support components fail or the ground stress field changes during operation [28]. The chemical substances added to the groundwater chemical seepage control method firstly increase the cost of the project operation, and secondly, the chemical substances left in the ground will also cause a certain amount of pollution of the underground environment in the future [29]. The groundwater isolation method represented by the grouting curtain is difficult to construct, the grouting parameters are demanding, and the results do not always achieve the effect of blocking water seepage after the completion of grouting [30]. Therefore, appropriate subsurface space confinement technologies need to be developed for the remediation of shallow groundwater contamination.

Artificial ground-freezing technology [31,32] is the use of artificial refrigeration technology to arrange freezing pipes in the strata to be reinforced. By circulating chemicals such as low-temperature brine or liquid nitrogen in the pipes, the heat in the surrounding soil layer is absorbed, and the loose water-bearing stratum is turned into a strong stratum with high strength and low permeability, which is used to block groundwater and bear a certain amount of stress, and plays the role of support and seepage control. Tunnel excavation, village masonry, or protection of groundwater is carried out under the protection of the frozen wall, thus ensuring construction safety and water quality. Qi et al. [33] investigated the thermal mechanism and permafrost development mechanism of gravel soil under seepage in the process of artificial freezing, and the experimental results showed that the greater the seepage rate, the longer the closure time of the freezing wall and the better the freezing effect. Han et al. [34] analyzed the case of coal mine water surge and considered the artificial freezing method instead of the grouting method. The law of groundwater migration in the ring-freezing channel of the mine was obtained through experiments, and the shaft-freezing device was designed to effectively inhibit the migration of water outside the project area. This paper only aims to study the final operational effect of the freezing wall, and does not discuss in depth the characteristics that affect the water-blocking mechanism of the groundwater freezing wall. Yang et al. [35] established a thermal–water–mechanical equation to simulate the freezing process of soil, and the reasonableness of the coupled model was verified by comparing the calculation results with the model and field tests. The results showed that the migration of water can be represented by the combined effect of matrix and fracture permeability, and with the increase in freezing time, the permeability of frozen clay decreased greatly, as did the water gathered around the freezing tube, but the study did not analyze the permeability of the frozen wall under temperature and pressure conditions. Qin et al. [36] studied and analyzed the freezing engineering of salt-containing groundwater and drew conclusions by establishing a seepage–freezing hydrothermal coupling calculation model and combining it with a small seepage–freezing model test. The results showed that the freezing wall is in an inhomogeneous state under the action of horizontal seepage. Freezing walls perpendicular to the direction of groundwater flow are formed slowly, while those parallel to the direction of water flow are formed quickly. The salt content affects the formation of the freezing wall, and the thickness of the freezing wall decreases as the salt content increases. Although the effect of temperature is considered in the text, the effect of ground pressure on the frozen wall after freezing is not mentioned.

Compared to other seepage control and water-blocking methods, a groundwater-freezing wall not only does not need to lag the blocking material into the ground but also has low cost and simple construction, and utilizes the characteristics of groundwater heating and melting, and cooling and freezing, to prevent ground contamination brought about by the completion of the blocking project. Therefore, based on the seepage characteristics of pore water and migrating water in the process of groundwater freezing, the real situation of the stratum is taken into account. We analyze the changes in pore type, pore distribution, effective porosity, and permeability of saturated and unsaturated samples under the action of freezing time, freezing temperature, and pressure. The conclusion is used to elucidate the seepage prevention and water-blocking mechanism after freezing of the soil body to guide the practical engineering of seepage prevention and water blocking of a groundwater freezing wall and to provide favorable theoretical support for the actual engineering operation.

2. Models and Experiments

In order to better reveal the actual operation effect of the underground freezing project, the experimental model was simplified based on the actual engineering conditions. Utilizing the nuclear magnetic high-temperature and high-pressure replacement device, the seepage experiment and freezing experiment were designed, and the temperature change and pressure change in the experimental process were regulated. The changes in porosity and permeability after the freezing of water-saturated samples and unsaturated samples was observed, and then the reliability of the freezing water resistance of the freezing wall was judged.

2.1. Experimental Model

In order to prepare the specimens for the experiment, the soil was first sampled, and after the sampling was completed, it was put in the dryer for drying. The soil particles were then sieved using a sieve mesh to retain the soil below 2 mm. Specimens with a target dry density of 1.71 g/cm³, a target moisture content of 16%, and a target compaction of 90% were prepared by querying the maximum dry density and optimum moisture content of different soils [37,38,39]. The specimen had a diameter of 25 mm, a height of 60 mm, a mass of about 47 g of soil, and a mass of about 9 g of water. The preparation process is shown in Figure 1. The prepared samples were vacuum-saturated, and unsaturated samples based on 25%, 50%, and 75% water content of the saturated samples were prepared by the method of conservation of mass to be used to simulate water seepage through the ground soil.

For the experiment to be easily calculated and smoothly analyzed, the prepared samples must comprise homogeneous and fully saturated soil with uniform and interconnected pores; seepage should be one-dimensional and laminar, following Darcy’s law, without accounting for temperature effects on water’s physical properties; the experimental seepage rate is considered the initial average flow rate across the flow section, disregarding any changes during the test; the temperature of both the specimen and the seepage water is assumed to be uniform, meaning the sensor measures a combined temperature for both; fluorinated liquids in the freezing tube are assumed to flow without altering their properties, keeping the coolant temperature constant; and the influences of the stress field on seepage and temperature fields are neglected [40,41,42].

2.2. Experimental Procedure

To measure the changes in porosity and permeability of soil samples at different temperatures and different pressures, first, the manufactured sample was placed in the gripper, and after the placement was completed, the indications of the temperature sensor T₁ and the pressure sensor P were observed, which were used to determine whether the sensors were working properly. The sample was frozen to the target temperature, valve 5 was opened, fluorinated liquid was added continuously to the refueling container, and valve 5 was closed when refueling was complete. The hot box was opened, circulating the pump and ring pressure-tracking pump, and the temperature was frozen to the T₂ temperature to observe whether there was leakage in the oil cup. If there was no leakage, the valve seal was intact; if there was a leakage of liquid, the experiment was stopped to check whether the valve was tightened or whether there was a sealing problem. After freezing to a certain temperature, the porosity of the sample was measured. When the sample in the holder was maintained at a certain temperature, valve 4 was opened, and distilled water in the storage tank was injected into the sample in the holder at the pressure required for the experiment, the change in the number shown on the meter was observed, and the data were processed to analyze the magnitude of the permeability. Among them, the kerosene tank and the constant pressure and constant flow pump were used to balance the pressure in the water storage tank so that the distilled water was injected into the gripper according to the pressure required for the experiment. Part of the experimental procedure is shown in Figure 2, and the schematic diagram of the experimental setup is shown in Figure 3.

2.3. Control Equations

During the freezing process, the coolant circulates and transfers the cold to the holder and to the sample by means of convective heat transfer. The heat transfer process refers to the transfer of heat between a moving fluid and a solid surface, and includes both heat conduction and convection. The heat flux is expressed by the Newtonian cooling equation [43]:

$$\mathsf{\Phi }=hA\left(T_a-T_b\right)$$

(1)

where h is the convective heat transfer coefficient, W/m²∙°C; T_a is the wall temperature of the contact, °C; and T_b is the temperature of the coolant. The convective heat transfer coefficient is related to the cause of the coolant movement, the surface factors of the pipe wall, the physical properties of the coolant, the presence or absence of phase change, and the flow state.

Water flows uniformly through the pores during sample percolation, satisfying Darcy’s law [44]:

$$V=-k{\displaystyle \frac{dH}{dL}}=-{\displaystyle \frac{k\rho g}{\mu }}\cdot {\displaystyle \frac{dH}{dL}}$$

(2)

where V is the seepage velocity, m/s; k is the permeability, m²; μ is the hydrodynamic viscous coefficient, kg/m∙s; H is the infiltration head, m²; L is the infiltration path, m; and ρ is the density of water, kg/m³.

In addition to this, the factors affecting the freezing temperature field under seepage include the thickness between the gripper and the sample, the initial temperature of the sample, the specific heat capacity of the soil, the thermal conductivity of the soil sample after freezing, the thermal conductivity of the unfrozen soil, the thermal conductivity of the water, the thermal conductivity of the soil, and so on.

3. Results and Discussion

3.1. The Change Law of Freezing Time and Pores of Water-Saturated Samples

To investigate the changes in pore type, pore distribution, and effective porosity of aquifer sediments under freezing, using the nuclear magnetic high-temperature and high-pressure torso substitution device, the water-saturated samples were first frozen to 10 °C, 0 °C, −2 °C, −4 °C, and −6 °C; continuously frozen at each temperature for 60 min; and measured in gradient changes every 10 min. The results are shown in Figure 4.

Samples were frozen from room temperature to 10 °C to simulate the state of groundwater when it is unfrozen. As can be seen from Figure 4a, the trends exhibited by the measured pore size and pore distribution of the samples during the freezing process did not differ much, and when the samples were frozen for 10 min, the largest percentage of the pore distribution with a pore size of 0.021 μm was achieved, which amounted to 0.1527%. The largest peak in the graph is considered the small pore, the middle peak the medium pore, and the smallest peak the large pore [45]. When the freezing time reached 50 min, the percentage of the small pore gradually decreased, and the percentage of the medium pore also gradually decreased, but then the large pore relatively increased. When the freezing time reached 60 min, the pore distribution of the small pores increased again, but the medium pores showed a decreasing trend. Since the samples were still at an unfrozen temperature, the change in pore size was not significant and did not yet show a clear pattern.

As can be seen from Figure 4b, for the samples frozen to 0 °C, the large, medium, and small peaks were shifted to the left with respect to those at 10 °C and the pore distributions were all decreased, but the reduction trend was not obvious. Taking the sample frozen at 0 °C for 10 min as an example, the pore distribution of 0.021 μm was also the largest, reaching 0.1421%, which is 0.0106% less than the same period of the previous year. Samples frozen to 0 °C have already begun to freeze, the pore water and migrating water content is decreased, and the water freezes into ice to form crystals, lenses, and ice interlayer, which makes the volume increase [46], which in turn compresses the pores in the samples, decreases the pore size, and reduces the pore distribution of the different pore types. As can be seen in Figure 4c, the pore distribution of the sample frozen to −2 °C decreased sharply, but the large, medium, and small peaks were still obvious, and the changes in pore size and pore distribution were obvious with the increase in the freezing time, which also indicates that the phase state of water in the sample was undergoing a continuous change. When the samples were frozen at −2 °C for 10 min, the pore distribution with a pore size of 0.018 μm had the largest percentage, of 0.0471%. As can be seen from Figure 4d, the continuous decrease in freezing temperature decreased the pore size and pore distribution of the samples significantly, the change in pore type was gradually insignificant, and during the freezing process, when the freezing time was more than 50 min, the large, medium, and small peaks did not have any obvious characteristics and were almost indistinguishable from each other. At this point, the pore water content of the sample decreased further and the total volume occupied by ice increased further. As can be seen in Figure 4e, the sample was frozen to −6 °C, at which point the sample was almost completely frozen. At a freezing time of less than 20 min, the pore distribution of the largest pore size of the sample was no more than 0.02%, and the small peaks of the larger pore sizes disappeared directly. At a freezing time of more than 30 min, the sample was completely frozen and the free water content could be measured to be zero. During the freezing process, water within the pores forms ice crystals, which may lead to alterations in the size and distribution of the pores. The formation of ice crystals has the potential to block some of the smaller pores, consequently resulting in a decrease in the proportion of small pores [47]. Conversely, the larger pores may experience a relative increase due to the growth of ice crystals. Throughout the freezing process, water might migrate from larger to smaller pores, causing a redistribution of pore sizes. This migration of water could give rise to dynamic changes in the pore distribution [48].

However, the pore size and pore distribution change graph cannot intuitively indicate the freezing effect of the sample. Therefore, the distribution of pore size during the freezing process at different temperatures was summarized and the changes in the effective porosity of the samples were analyzed, and the results of the summarization and analysis are shown in Figure 5 and Figure 6.

It is obvious from Figure 5 that the 10 °C sample exhibited a much larger pore distribution in the range of 0–0.1 μm than that of 0.1–1 μm, 1–10 μm, and 10–100 μm during the freezing process. It was assumed that the pore sizes in the range of 0–0.1 μm were small, those in the range of 0.1–1 μm were medium, and those in the range of 1–200 μm were large, i.e., the small pores were the most numerous, followed by the medium pores and finally the large pores. Changes in large, medium, and small pores are in dynamic equilibrium, as evidenced by a decrease in the number of small pores and an increase in the number of medium or large pores; the temperature at which water freezes was already reached at 0 °C, and the size of the pores in the samples was obviously much diminished compared to that of the samples at 10 °C, but at this temperature it was still a mixing state of ice and water. Upon subjecting the samples to a freezing temperature of −2 °C, a significant decrease was observed in the population of large, medium, and small pores. This phenomenon suggests a substantial reduction in the pore water content within the samples, indicative of a profound transformation in the phase state [49]. During this process, the distribution of pore sizes for small and medium pores exhibited a decreasing trend over time. Conversely, the relative increase in the number of large pores can be attributed to the freezing of free water within the small pores, which induces volumetric expansion [50]. This expansion exerts pressure on the soil matrix, leading to the formation of a novel spatial structure. Consequently, this rearrangement results in an increase in the number of large pores. When the samples were frozen to −4 °C, the pore sizes decreased again, and the small pores accounted for less than 1%. When the freezing time was 30–40 min, only small pores existed in the samples, and the medium and large pores had disappeared. When the freezing time was 60 min, the distribution of small pore apertures was only 0.2009%; when the samples were frozen to −6 °C, the samples were almost completely frozen; and when the freezing time was longer than 30 min, the presence of sample apertures could not be measured experimentally. As can be seen in Figure 6, the porosity of the samples continued to decrease with both the decrease in freezing temperature and the increase in freezing time. The porosity of a 10 °C sample freezing for 10 min was 5.58%; the porosity when freezing 60 min porosity was 5.10%, at which time the change in porosity is not obvious; the porosity of a 0 °C sample freezing for 10 min was 4.94%, and the porosity of a sample freezing for 60 min was 4.66%, due to the pore water belonging to the solid–liquid mixture of the state existing at this time. Therefore, compared to the 10 °C sample, the porosity saw a decrease of only less than 1%. Nevertheless, upon subjecting the samples to freezing temperatures ranging from −2 °C to −4 °C, a significant decrease in porosity was observed. After 10 min of freezing at −2 °C, the porosity dropped to 1.75%, and after 60 min of freezing, only 0.76% of porosity was retained. At −4 °C, a 10-min freeze reduced the porosity to 0.67%, and extending the freezing duration to 60 min resulted in a porosity of merely 0.35%. This decrease in porosity indicates that a considerable volume of free water within the sample had transformed into ice, thereby occupying the void spaces within the sample [51]. The reason for this phenomenon is that as the temperature drops, the kinetic energy of water molecules decreases, leading to the formation of ice. Since ice has a greater volume than liquid water, its formation within the sample’s pores results in a physical occupation of space, compressing the original pore structure and thus decreasing the porosity. The lower the temperature and the longer the duration of freezing, the more water is converted into ice, exacerbating the decrease in porosity [52]. And when the sample was frozen to −6 °C, the porosity was 0, as the freezing time exceeded 30 min, at which point the sample was completely frozen.

From the above conclusions, it can be seen that the saturated samples simulated the freezing process of groundwater completely infiltrating into the soil layer, and the underground freezing wall is formed more effectively at lower temperatures. When the freezing temperature was lower than −6 °C, the porosity of the frozen wall produced by the freezing of groundwater was almost 0, the frozen wall achieved a sealing effect and acted as a barrier, and there was no seepage.

3.2. The Change Law of Pressure and Permeability of Saturated Samples

The change in volume of groundwater after freezing also inevitably results in the freezing wall being subjected to pressure from the ground, so it is particularly important to consider the effect of ground pressure on the seepage characteristics of the freezing wall. To investigate the change in permeability of water saturated samples under freezing at different pressures, aided by the observation of the magnitude of porosity during the process and using the nuclear magnetic high-temperature and high-pressure torso replacement device, the samples were frozen to 10 °C, 0 °C, −2 °C, and −4 °C; the pressure difference between the two ends of the gripper was adjusted to 0.3 MPa, 0.5 MPa, 1 MPa, and 2 MPa, respectively; and an incremental increase in the metrological values within 100 s was observed. The results are shown in Figure 7 and Figure 8. Due to the way the data were recorded by the experimental equipment, the mass added to the meter in 100 s is used in this paper to define the magnitude of the permeability [53,54] in g/s.

The findings indicate a demonstrably significant influence of hydrostatic pressure on the permeability characteristics of the sample, with a pronounced enhancement in permeability observed in direct proportion to applied pressure increments, such as the increase from 0.3 MPa to 2 MPa at a constant temperature of 10 °C, where hydraulic conductivity increased from 0.0068 g/s to 0.0139 g/s, corroborating the positive correlation. However, at −4 °C, freezing induced a considerable decrease in permeability due to the phase transition of water to ice, forming a cohesive “freezing wall” that impeded water migration, leading to a significant decrease to 0.0026 g/s at 2 MPa and near imperceptible permeability at 0.3 MPa. These observations substantiate the hypothesis that increased formation pressure enhances seepage propensity, aligning with air-driven water-stopping technologies, and highlight the marked difference in permeability between the liquid phase of water and its solid or mixed solid–liquid phases [55], which can be attributed to the reduction in available pore volume and increased resistance to fluid flow during the transition to a solid state [56].

To further characterize the change in permeability, the porosity of the samples was similarly measured under pressure and temperature variations. The graph results show that with the increase in pressure, the sample porosity showed a slightly decreasing trend. The reason for this may be that the axial load on the sample increased under pressure, resulting in an increase in the compaction of the sample, a decrease in the volume of the sample, a squeezing of the original space, and a decrease in the porosity of the sample. When the water in the sample was in liquid form, the pressure effect did not decrease the porosity much, with the porosity of the sample at 0.3 MPa, 10 °C being 3.40%, and that of the sample at 2 MPa, 10 °C being only 3.25%. Conversely, when the water within the sample existed in a mixed solid–liquid or solid phase, the application of increased pressure resulted in a marked decrease in porosity. Nonetheless, when the sample was maintained at a constant temperature within the frozen state, the increase in pressure led to a decrease in porosity while concurrently inducing a slight increase in permeability, indicative of an improved pore connectivity attributed to the applied pressure. Despite the retention of porosity in the samples at −4 °C, the pore channels became entirely obstructed in the presence of a freezing wall, thereby imparting an impervious character to the sample at this specific temperature [57].

From the above conclusion, it can be seen that the saturated samples simulated the process of groundwater freezing, and considering the role of stratum pressure, the occurrence of seepage in the underground freezing wall was greater under the condition of higher underground pressure. In order to ensure the freezing effect of the freezing wall, the freezing thickness or freezing temperature can be increased. When the freezing temperature was lower than −4 °C, the permeability of the frozen wall produced by the freezing of groundwater was almost 0. If the freezing temperature is lowered to −6 °C, the frozen wall will easily achieve a sealing effect and act as a barrier within a certain pressure range, and there will be no seepage.

3.3. The Change Law of Freezing Time and Pores of Unsaturated Samples

A preliminary understanding of the pore types, pore distribution, changes in effective porosity, and changes in permeability under different pressures of aquifer sediments under freezing in water-saturated samples has been developed. However, during groundwater freezing, there are some areas in the soil layer that are not water saturated. Therefore, in order to further investigate the changes in pore type, pore distribution, and effective porosity of aquifer sediments under freezing in unsaturated conditions, specimens with 25%, 50%, and 75% water saturation were subjected to experiments.

In the same way as the water-saturated sample experiment, the 25%, 50%, and 75% water-saturated samples were frozen to 10 °C, 0 °C, −2 °C, −4 °C, and −6 °C using the nuclear magnetic high-temperature and high-pressure carcass substitution device and continuously frozen at each temperature for 60 min, with a gradient change of 10 min every 10 min for measurements. The results were as shown in Figure 9, Figure 10, Figure 11, Figure 12, Figure 13, Figure 14, Figure 15 and Figure 16.

As can be seen from the figure, the increase in water saturation resulted in an increase in the number of small pores distributed in the samples, which roughly demonstrates that an increase in the water content of the samples by 25% increased the distribution of small pores in the samples by 0.02%. After the water content was reduced from 75% to 50%, the number of large pores decreased and the number of medium pores increased, but the number of small pores also decreased, and when the water content was reduced to 25%, the number of large and medium pores in the samples almost completely disappeared and there were a lot of small pores, but the overall porosity decreased. Since the temperature of the samples was continuously maintained at about 10 °C, the pore distribution in the samples changed with the increase in freezing time, but the changes were not regular. Unsaturated samples had a change in the distribution of the 0–0.1 μm pore size ranging from 3.47% to 4.79%, the distribution of the 0.1–1 μm pore size ranging from 0 to 0.54%, the distribution of the 1–10 μm pore size ranging from 0.031% to 0.19%, and the distribution of the 10–100 μm pore size ranging from 0 to 0.074% as the freezing process took place. The variation range of the distribution of 100–200 μm pore size was 0~0.0037%.

The small pores of the 75% saturated sample showed stronger signals, which means that when the sample had more water content, the free water was mainly distributed in the small pores of the sample, the pores of 0.003–0.09 μm occupied most of the pores, and the percentage of pores with a size of 0.014 μm reached 0.1679%, of which the number of medium pores and large pores was almost the same. But compared with the freezing process at 10 °C, the percentage and distribution of the individual pores were significantly reduced. However, compared with the freezing process at 10 °C, the percentage and distribution of each pore were significantly reduced, which also confirms that the free water had changed its phase state during the freezing process, and the measured pore size and pore distribution were reduced due to the phase change. The water sample at 50% saturation exhibited the most pronounced signal within the small pore fraction, with a subsequent decrease observed in the mesoporous and macroporous regions [58]. This observation suggests that a significant portion of the water was predominantly retained within the small pores. Additionally, a considerable amount of free water was detected within certain mesopores, indicating a heterogeneous distribution of water across the pore sizes. Among them, small pores with a pore size of 0.0014–0.1 μm accounted for the largest proportion, and the pore size of 0.014 μm reached 0.1432%. The proportion of medium pores with a pore size of 0.145–1.45 μm was the next largest, and the proportion of pores with a pore size of 1.1 μm reached 0.029%. Furthermore, as the duration of freezing increased, the overall pattern of pore size and distribution alterations consistently demonstrated a similar trajectory, exhibiting a gradual decrease in the relative proportion of pore distribution as a function of extended freezing periods [59]. In the case of the 25% water-saturated sample, a notable fraction of mesopores was observed, which can be attributed to the diminished water content [60]. Despite this, the preponderance of water was still primarily localized within the small pores. Among them, small pores with a pore size of 0.0014–0.12 μm accounted for the largest percentage, and the percentage of pores with a pore size of 0.013 μm reached 0.059%. Medium pores with a pore size of 0.143–1.2 μm accounted for the second largest percentage, and the percentage of pores with a pore size of 0.92 μm reached 0.021%. Samples frozen to 0 °C began to freeze, and as the freezing time continued to increase, the trend was toward a decrease in pore water content, with water freezing to ice causing an increase in volume, which in turn compressed the pores in the samples, reducing pore size and decreasing the distribution of pore sizes across different pore types. The 0–0.1 μm pore size distribution of the unsaturated samples as the freezing process occurred ranged from 0.942% to 4.64%, the 0.1–1 μm pore size distribution of the samples ranged from 0 to 0.336%, the 1–10 μm pore size distribution of the samples ranged from 0.0262% to 0.103%, the 10–100 μm pore size distribution of the samples ranged from 0 to 0.0851%, and the 100–200 μm pore size distribution was 0.

The pore distribution of the samples frozen to −2 °C decreased dramatically, with the percentage of small pores in the 75% water-saturated sample decreasing by a factor of about 4 compared to the 0 °C samples, the percentage of small pores in the 50% water-saturated sample decreasing by a factor of more than 3 compared to the 0 °C samples, and the percentage of small pores in the 25% water-saturated sample decreasing by a factor of more than 2 compared to the 0 °C samples. The trends of pore size and pore distribution of the 75% water-saturated and 50% water-saturated samples were almost the same, which also indicates that almost all the water in the samples was frozen. The 25% water-saturated sample did not show a significant difference between small and large pores with increasing freezing time. The pore size present in the sample gradually decreased due to the freezing effect with the increase in freezing time, and the temperature of −2 °C almost made all the water in the sample undergo the phase transition, the pore water content in the samples further decreased, the total volume occupied by the ice further increased, and the increase in the freezing time of the sample accelerated the phase transition. The 75% water-saturated sample had the largest proportion of small pores, with pore sizes ranging from 0.0015 to 0.11 μm, and the proportion of pores with a pore size of 0.014 μm reached 0.0378%. The proportion of medium pores with a pore size of 0.11~1.51 μm was the next largest, and the proportion of pores with a pore size of 0.97 μm reached 0.0167%. The 50% water-saturated sample had the largest proportion of small pores, with pore sizes ranging from 0.0015 to 0.12 μm, and the proportion of pores with a pore size of 0.014 μm reached 0.0382%. The proportion of medium-sized pores with a pore size of 0.12~1.51 μm was the second largest, and the proportion of pores with a pore size of 0.98 μm reached 0.0171%. The distribution of unsaturated samples with a 0–0.1 μm pore size as the freezing process occurred ranged from 0.133% to 1.654%, the distribution of samples with a 0.1–1 μm pore size ranged from 0–0.262%, the distribution of samples with a 1–10 μm pore size ranged from 0–0.758%, the distribution of samples with a 10–100 μm pore size ranged from 0 to 0.0123%, and the distribution of samples with a 100–200 μm pore size had a distribution of between 0 and 0.0013%.

When the freezing temperature was lowered to −4 °C, the proportion of small pores in the 75% water-saturated sample continued to decrease, and with the increase in freezing time, small, medium, and large pores were reduced. When the freezing time was 60 min, the porosity of the sample was 0, which was completely frozen at the time. The presence of pore water was still able to be detected in the 50% water-saturated sample when the freezing time was from 0 to 30 min, but after more than 30 min, all the water in the sample was completely frozen. The presence of holes was still able to be detected for the 25% water-saturated sample during the 10 min freezing process due to the low water content, but as the time continued to increase, the presence of free water was not detected in the sample. An increase in freezing time significantly improved the freezing effect. As can be seen from Figure 16, the distribution of the 0–0.1 μm pore size of the unsaturated samples with the occurrence of the freezing process ranged from 0 to 0.766%, the distribution of samples with a 0.1–1 μm pore size ranged from 0 to 0.0771%, the distribution of samples with a 1–10 μm pore size ranged from 0 to 0.0222%, the distribution of samples with a 10–100 μm pore size ranged from 0 to 0.0121%, and the distribution of samples with a 100~200 μm pore size had a distribution of 0.

For the unsaturated water samples, the effects of temperature and freezing duration on the freezing effect of the samples were analyzed, but the freezing effect of the samples could not be judged intuitively only by the change graphs of pore size and pore distribution. Therefore, the distribution of pore size during freezing at different temperatures was summarized for the unsaturated water samples, and the results are shown in Figure 17.

As can be seen in Figure 17, the porosity of the 25%, 50%, and 75% water-saturated samples continued to decrease with decreasing freezing temperature and increasing freezing time. For the 25% water-saturated sample, the porosity was 3.51% for 10 min of freezing and 2.04% for 60 min of freezing at 10 °C. Due to the low water content, the increase in freezing time exhibited a significant decrease in porosity. The porosity of the 0 °C sample was 1.75% for 10 min of freezing and 1.03% for 60 min of freezing. Since the pore water was in a solid–liquid mixture at this time, the porosity of the sample decreased by less than 1% compared to that of the 10 °C sample, but the overall trend showed a decreasing state. When the sample was frozen to −2 °C, the porosity was 0.86% for 10 min and 0.25% for 60 min. The increase in freezing time made the sample show a slow decreasing trend, which was also due to the low water content in the samples, but the sample was not completely frozen at this time, and pore water still existed inside the sample. When the sample was frozen at −4 °C for 10 min, the porosity was only 0.06%, and when the freezing time exceeded 20 min, the porosity was already 0. At this time, the sample was completely frozen, and the free water inside the sample underwent a phase transition. For the 50% water-saturated sample, the porosity was 4.76% for 10 min of freezing at 10 °C and 4.54% for 60 min of freezing. The porosity of the sample tended to decrease with the increase in freezing time, but in general, it tended to be stable, and it only decreased by 0.22% for 50 min. The porosity of the 0 °C sample was 4.54% for 10 min of freezing and 4.24% for 60 min of freezing, although the water had already reached the solidification temperature at this time, but in general, the decrease in the pore water content was not obvious, and it was only reduced by 0.30% for 50 min. At −2 °C, the porosity was 1.23% for 10 min of freezing and 0.67% for 60 min of freezing. With the increase in freezing time, the porosity of the sample decreased significantly, and it decreased by 0.56% for 50 min. The freezing effect of the sample was significant, but pore water still existed in the sample. When the sample was frozen at −4 °C for 10 min, the porosity was 0.43%, and when the freezing time exceeded 30 min, the porosity was already 0, and the sample was completely frozen at that time. For the 75% water-saturated sample, at 10 °C, the porosity was 5.03% for 10 min of freezing and 4.79% for 60 min of freezing, at which time the change in the porosity of the sample was also not obvious, but still, with the increase in freezing time the porosity decreased, and it was only reduced by 0.24% for 50 min. The porosity of the 0 °C sample was 4.78% for 10 min of freezing and 4.63% for 60 min of freezing, and it decreased by 0.15% for 50 min. The porosity of the 50% and 75% saturated samples was much higher than that of the 25% saturated sample by a factor of more than 2 at 0 °C due to the higher water content. At −2 °C, the porosity was 1.69% for 10 min of freezing, 0.89% for 60 min of freezing, and 0.80% for 50 min of freezing; at −4 °C, when the sample was frozen for 10 min, the porosity was 0.82%, and at 50 min of freezing time, the porosity was 0.11, but when the freezing time was 60 min, the sample was completely frozen at this time. The change in permeability of different water-saturated samples shows that the freezing time and freezing temperature had an effect on the freezing effect of the soil samples, the water content also had an effect on the freezing effect of the samples, and the freezing time of the samples with high water content needed to be longer in order to have a freezing effect.

From the above conclusions, it can be seen that the unsaturated samples simulated the freezing process of incomplete infiltration of groundwater into the soil layer, and the trend of the results is in overall agreement with the conclusions for the saturated samples. Due to the low water content, when the freezing temperature was lower than −4 °C, the porosity of the frozen wall produced by the freezing of groundwater was almost zero, and the freezing process acted as a barrier, at which time there was no seepage. However, considering the freezing situation of the saturated water samples, it is better to apply a freezing temperature that is as low as possible for the underground freezing wall project.

3.4. The Change Law of Pressure and Permeability of Unsaturated Water Samples

Through the change in porosity of unsaturated water samples and the change in pore distribution, we can gain a preliminary understanding of the groundwater-freezing process, but the ground pressure and the size of the infiltration rate under the action of pressure should also be considered for the realization of the effect of seepage prevention and water blocking of the underground soil body in the freezing process. Therefore, in order to investigate the permeability of the unsaturated soil samples under pressure, 25%, 50%, and 75% water-saturated samples were used as the experimental model; the samples were frozen to 10 °C, 0 °C, −2 °C, and −4 °C; and differential pressures at the two ends of the grippers were adjusted to 0.3 MPa, 0.5 MPa, 1 MPa, and 2 MPa, respectively; and the increases in the meter values within 100 s were observed. The results are shown in Figure 18.

For 25%, 50%, and 75% water-saturated samples, the permeability gradually decreased with decreasing temperature and increased with increasing pressure. For the samples with higher water content, the permeability of the samples under pressure was significantly higher than that of the samples with lower water content. For the 25% saturated water sample, when used for the differential pressure of 0.3 MPa and freezing temperature of 10 °C, the permeability of the sample was 0.0036 g/s. As the freezing temperature was lowered to −4 °C, the permeability of the sample was 0, and by that time, the sample had already reached the water-blocking effect. When the differential pressure was increased to 2 MPa, the permeability of the sample reached 0.0079 g/s. The increased pressure acting between the samples increased the compaction of the sample, the volume of the sample changed, and the water inside the sample increased its permeability due to the compression of the space, a phenomenon that also occurred in samples of other moisture contents. For the 50% saturated water sample, the sample permeability was 0.0042 g/s when the differential pressure applied to it was 0.3 MPa and the freezing temperature was 10 °C. As the freezing temperature was lowered to −4 °C, the sample permeability was 0, and when the differential pressure was increased to 2 MPa, the sample permeability reached 0.0093 g/s. However, the permeability of the sample with a pressure difference of 2 MPa and a freezing temperature of −4 °C could be measured as 0.0006 g/s, a situation that also indicates that the increase in water content increased the permeability of water. Then, in the case of frozen water blocking, a lower freezing temperature or higher freezing time needs to be considered when working in formations with higher pressures. For the 75% saturated water sample, when the pressure difference was 0.3 MPa and the freezing temperature was 10 °C, the permeability of the sample was 0.0049 g/s, and as the freezing temperature was lowered to −4 °C, the permeability of the sample was zero, which also indicates that the sample was not compressed significantly under the lower pressure and that the sample could achieve the effect of a water barrier at −4 °C. However, upon subjecting the sample to a differential pressure of 2 MPa, it was observed that the permeability of the sample reached a maximum of 0.0101 g/s at a freezing temperature of 10 °C. Furthermore, when the freezing temperature was lowered to −4 °C, the sample maintained a permeability of 0.0019 g/s. In samples with a lower water content, the decrease in temperature readily facilitated the freezing of pore water, thereby enhancing the water-blocking effect. Conversely, in samples with a higher water content, although the majority of the free water was frozen, the presence of remaining pore water resulted in a less effective freezing process compared to samples with lower water content. Consequently, for soils characterized by high water content, it is imperative to decrease the freezing temperature further and extend the freezing duration to achieve an optimal freezing effect [61].

As with the saturated sample analyses, sample porosity was similarly measured for pressure and temperature variations in order to further characterize permeability changes in the unsaturated samples. The results are shown in Figure 19.

For 25%, 50% and 75% water-saturated samples, the porosity similarly showed a decreasing trend with the decrease in freezing temperature and showed a gradual decreasing trend with the increase in pressure between the samples. This conclusion is opposite to the change in permeability. The reason is that the increase in the sample compaction under the action of pressure increased the permeability of the water, which made some of the ineffective pores, bound pores, and micropores connect or disappear and form a larger pore space. For the 25% saturated water sample, when used for the differential pressure of 0.3 MPa and freezing temperature of 10 °C, the measured porosity of the sample was 2.67%, and the porosity was 0 when the freezing temperature was lowered to −4 °C, when the free water present in the pores of the sample were frozen. When applied to the differential pressure of 2 MPa and freezing temperature of 10 °C, the porosity measured in the sample was 2.44%, the pressure effect reduced the porosity size, and when the freezing temperature was lowered to −4 °C, the porosity was likewise zero. This situation indicates that the soil water content was low and that lower temperatures achieved better freezing results. For the 50% saturated water sample, when used for the differential pressure of 0.3 MPa and freezing temperature of 10 °C, the sample measured a porosity of 2.94%, and when the freezing temperature was lowered to −4 °C, the porosity was 0.46. When applied to the differential pressure of 2 MPa and freezing temperature of 10 °C, the sample measured a porosity of 2.51%, and at a freezing temperature of −4 °C, the porosity was 0.30. The increase in the water content of the sample may have shown a permeability of 0 at certain temperatures and pressures, but there was still porosity present, which is because some of the pores present in the samples were in the form of ineffective pores, bound pores, and micropores; the pore size was too small or not connected; and therefore, water seepage did not occur. For the 75% saturated sample, when used with a differential pressure of 0.3 MPa and a freezing temperature of 10 °C, the porosity of the sample was 3.05%, and when the freezing temperature was reduced to −4 °C, the porosity was 0.58. When used with a differential pressure of 2 MPa and a freezing temperature of 10 °C, the porosity of the sample was 2.51%, and when the freezing temperature was reduced to −4 °C, the porosity was 0.37. Due to the high water content of the sample, porosity still existed even though the permeability of the sample was zero. In the application to the design of seepage control and water blocking, in order to better serve the purpose of water blocking and further decrease the amount of porosity, the freezing temperature can be reduced or the freezing time can be increased.

From the above conclusions, it can be seen that in the unsaturated sample under pressure, despite the presence of pore water and migrating water, the permeability may also have been 0. However, when the pressure reached 2 MPa and the freezing temperature was −4 °C, the sample still demonstrated a seepage phenomenon. Therefore, applied to the freezing wall seepage control and water-blocking project, the ground pressure is large, and the soil-freezing temperature needs to be decreased to −6 °C to be optimal.

4. Conclusions

This study employed a nuclear magnetic high-temperature and high-pressure driving device to experimentally analyze the variations in pore properties and permeability of saturated and unsaturated samples under the influence of temperature, pressure, and freezing time. The results indicate that as freezing time increases, the pore size of the samples continuously changes, with a decrease in large pores and a corresponding increase in the distribution of medium and small pores, leading to a decrease in sample porosity. When the freezing temperature reaches −6 °C and the freezing time exceeds 30 min, the sample porosity becomes zero, indicating complete freezing. The higher the differential pressure of the samples, the more pronounced the permeability phenomenon; however, when the freezing temperature reaches −4 °C, the measured permeability is almost zero, suggesting that although pores exist, they are nearly non-connective. The decrease in freezing temperature significantly reduces the samples’ porosity and permeability. It is recommended in practical engineering to maintain a freezing temperature of below −6 °C and to extend the freezing time as long as possible.

The study provides a comprehensive understanding of the dynamic changes in pore size distribution, porosity, and permeability during the process of freezing samples, and offers practical guidelines for adjusting freezing temperature and time in engineering applications to optimize sample properties. However, the complexity of the experimental apparatus and the potential need for specialized equipment limit the widespread application of this method. The analysis is based on a limited range of freezing temperatures and times, which may not cover all scenarios in actual engineering. The findings are specific to particular temperature and pressure conditions and may not be directly applicable to all geological and environmental settings. The size and composition of the samples may affect the results, necessitating further research to validate the findings using different materials.

To overcome these limitations, future research should expand the range of freezing temperatures and times to cover a broader range of engineering scenarios, investigate the impact of different sample compositions on pore properties and permeability during the freezing process to enhance the universality of the results, develop simplified and more accessible experimental methods to make the research more applicable to various laboratory environments, explore the long-term stability and durability of samples after freezing to better understand their significance in engineering projects, and conduct field studies to validate laboratory findings and adjust engineering application recommendations based on actual conditions.