Earlier research in soil mechanics have mostly been focused on saturated soil mechanics. There are two main reasons for this. Firstly, saturated soil systems which consist of only two phases of soil and water are relatively easier to study than multiphase unsaturated soil systems (Barbour, 1998). Secondly, most soil mechanics were developed in regions where saturated soils were problematic (Barbour, 1998). The work of Terzaghi has led to the development of practical technologies, test methods, and better understanding of soil behaviour.
However, one third of the earth's land consists of arid and semi-arid regions where the rate of evaporation exceeds the rate of precipitation (Dregne, 1976). In recent times, the changes in the global economy have resulted in many of these regions being centres of rapid economic development. In these regions, the behaviour of unsaturated soils has caused numerous problems that prove to be costly. For example, Krohn and Slosson estimated in 1980 that $7 billion are spent each year in the United States for the damage caused to all types of structures built on expansive soils.
Get your grade
or your money back
using our Essay Writing Service!
This work aims to study the behaviour of unsaturated soils; the development of aggregate structures. Soils containing aggregate structures present major engineering problems by either swelling (expansive soils) or collapsing (collapsing soils). In both cases, ground movements are the result of negative pore-water pressure (Fredlund and Rahardjo, 1993). Increase in pore-water pressure results in volume changes and therefore ground movements can inflict major damage to structures or cause slope failures.
The significance of this work is to assess the development of aggregate structures through a series of drying and wetting tests on previously saturated soils. The soils used in this experiment are: commercially available powdered Kaolin, natural London clay, and natural Virginia clay. A series of suction values are applied to the saturated soil samples. Identification of aggregate structures is done by scanning electron microscopy (SEM). This can provide further information to understanding the mechanical and hydraulic behaviour of unsaturated soils.
CHAPTER 2 LITERATURE REVIEW
2.1 Unsaturated soils
Unsaturated soils are multiphase soils that can be commonly found in arid and semi-arid regions where rates of evaporation and evapotranspiration exceed precipitation. Whether a soil is saturated or unsaturated is dependent on the climate of the region. Such regions tend to have deep water tables and therefore unsaturated soils have negative pore water pressure as they are well above the water table. Besides being found in dry environments, unsaturated soils can also be the result of excavation and compaction of soils.
In saturated soil mechanics, the stress state variable can be commonly expressed by:
Where σ' = effective normal stress, σ = total normal stress, and uw = pore-water pressure.
This equation is also referred to as the effective stress equation. Past work has shown that the value of effective stress or the one stress state variable is sufficient to describe the mechanical behaviour of a saturated soil. Its validity has been well accepted and verified (Rendulic, 1936; Bishop and Eldin, 1950; Laughton, 1955; Skempton, 1961).
The behaviour of unsaturated soils is more complex due to the fact that it is a multiphase soil. Past work has attempted to translate the concept of effective stress for saturated soils to unsaturated soils. However, difficulties arose as soil properties measured do not yield a single-valued relationship to the proposed effective stress. Different magnitudes are obtained for different problems, different stress paths, and different types of soil (Jennings and Burland, 1962; Coleman, 1962; Bishop and Blight, 1963; Burland, 1964; Burland, 1965; Blight, 1965). To deal with the matter, stress state variables in unsaturated soil were used in an independent manner. The effective stress equation has been separated into two independent stress state variables. It is justified by the consideration of the air-water interface as the fourth phase (Fredlund and Morgenstern, 1977). Hence, it was recognised that the effects of total stress and pore-water pressure need to be separated.
One of the distinct properties of an unsaturated soil is that it is a multiphase soil unlike saturated soils, which only have two phases of soil and water. It is important to establish the number of phases a soil comprises as it can influence the stress state of the soil mixture. The work of Fredlund and Morgenstern (1977) has established that an unsaturated soil has four phases, namely, 1) solids, 2) water, 3) air, and 4) air-water interface. The fourth phase is also known as the contractile skin or the meniscus and it has a distinctive property to exert a tensile force. It behaves like a concave membrane pulling the soil particles together under tension. The properties of the air-water interface are different from that of the water phase in a soil (Davies and Rideal, 1963). For example, it has reduced density and increased heat conductivity. There is a distinct transition from liquid water to the contractile skin (Derjaguin, 1965).
Always on Time
Marked to Standard
With the presence of air in the soil, aggregation will occur as pore water pressure becomes increasingly negative. Soil particles group together under the tensile force of the air-water interface to form aggregate structures. These aggregate structures have voids within resulting from air entering the aggregates when suction is applied. These aggregated soils exhibit a bimodal pore size distribution.
2.2 Matric suction
Surface tension is a property exerted by the air-water interface which results from the intermolecular forces acting on the molecules in the interface. These forces are different from the molecular forces in the interior of the liquid water (Fredlund and Rahardjo, 1993). Unlike the molecules in the interior water, which experience equal forces in all directions, a molecule in the air-water interface experiences an unbalanced force towards the interior of the water. Essentially, this is because one side of the molecule is exposed to air, and thus lacks opposite forces. For the air-water interface to be in equilibrium, a tensile force is generated along the air-water interface. This property is known as the surface tension, Ts. Surface tension is tangential to the surface of the air-water interface and is measured as the tensile force per unit length of the interface (Fredlund and Rahardjo, 1993).
In an unsaturated soil, the air-water interface will be subjected to air pressure, ua, which is greater than the water pressure, uw. The pressure difference, (ua - uw), is known as the matric suction. Kelvin's capillary model equation shows that:
Where (ua - uw) = matric suction, Ts = surface tension, and Rs = radius of curvature.
The matric suction of the soil is inversely proportional to the radius of curvature of the air-water interface. When matric suction is zero, the radius of curvature is infinity, hence resulting in a flat air-water interface. Matric suction is a hydrostatic pressure because it has equal magnitude in all directions.
The factors affecting the magnitude of matric suction are: 1) ground surface conditions, 2) environmental conditions, 3) vegetation, 4) water table, and 5) permeability of the soil profile. Matric suction is soil increases during dry seasons and decreases during wet seasons. Maximum changes in suction are observed near the ground surface. Accumulation of moisture below covered surfaces such as pavements can cause a reduction in soil suction. Vegetation on the ground surface can exert a tensile pull to the pore-water through evapotranspiration. Increased rates of evapotranspiration result in increased matric suction. The depth of the water table significantly affects the magnitude of matric suction. The deeper the water table is, the higher the matric suction, and hence the magnitude of matric suction is much higher near the ground surface (Blight, 1980). The permeability of a soil is its ability to drain water, which in turn, affects the magnitude of the matric suction. Matric suction increases as water is drained from the soil.
2.3 Development of aggregate structure
The development of unsaturated soil was demonstrated at different stages of matric suction by Childs (1969). As a soil becomes unsaturated, air begins to replace the water in the larger pores, resulting in water flow in smaller pores with an uneven flow path. With increasing suction, the pore volume occupied by water decreases and the air-water interface is drawn closer to the soil particles. Available space for water flow is significantly reduced, resulting in decreased permeability with respect to the water phase.
In the present work, different values of suction are applied to simulate the development of unsaturated soil. As suction increases, the surface tension exerted by the air-water interface pulls the soil particles together, forming aggregate structures. For air to replace the space occupied by water, the matric suction applied must exceed a parameter known as the air entry value of the soil, (ua - uw)b. The air entry value is a measure of the maximum pore size in a soil.
The air entry value can be determined from the curve of effective degree of saturation versus matric suction (Brooks and Corey, 1964). The effective degree of saturation, Se (Corey, 1954), can be expressed as:
Where Se is the effective degree of saturation and Sr is the residual degree of saturation, which can be defined as the degree of saturation at which an increase in matric suction does not produce a significant change in the degree of saturation. The following equation describes the curve of effective degree of saturation versus matric suction (Brooks and Corey, 1954):
This Essay is
a Student's Work
This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.Examples of our work
Where λ is the pore size distribution index, which can be defined as the negative slope of the curve of effective degree of saturation versus matric suction. A small value of λ means that the soil has a wide range of pore sizes whereas a large value means that the soil has a more uniform distribution.
2.4 Bimodal pore size distribution
Bimodal pore size distribution is commonly observed in compacted clays at dry side of optimum water (Diamond, 1970; Sridharan et al., 1971; Collins and McGown, 1974). An extension of the original Barcelona Basic Model (BBM) (Alonso et al., 1987, 1990), the BExM considers two structures: 1) the microstructure, which accounts for the particles and 2) the macrostructure, which accounts for the larger structure of the soil. The model provides better understanding of soil behaviour and the mechanisms involved (Lloret et al., 2003).
The double porosity feature of aggregated soils is made up of two different pores: 1) inter-aggregate pores, which are the voids between the aggregates, and 2) intra-aggregate pores, which are the voids within the aggregate itself. Aggregation may be regular or irregular (Collins and McGown, 1974). In regular aggregation, unlike irregular aggregation, aggregates have a distinct physical boundary and therefore particular hydraulic and mechanical properties can be attributed to the aggregates themselves. Sizes of aggregates depend on the amount of suction applied and the amount of air present in the soil.
Volume of voids present in soils are represented by one of the following parameters: porosity, n, defined as the ratio of void volume over the total volume of the soil, void ratio, e, defined as the ratio of void volume over the volume of soil particles, or specific volume, v, which is the total volume of soil which contains unit volume of solid particles.
Where Vv = volume of voids, V = total volume, and Vs = volume of soil solids. From the two equations, the following can be obtained:
In saturated soil mechanics, the coefficient of permeability is a function of the void ratio (Lambe and Whitman, 1979). However in unsaturated soil mechanics, the coefficient of permeability is highly dependant on the combined changes in the void ratio and the water content of the soil, where water content, w, is defined as the ratio of the mass of water to the mass of soil solids.
Where Mw = mass of water, and Ms = mass of solids. The pore size distribution is used to give reasonable estimates of the permeability of the soil.
2.4 Problems caused by aggregated soils
Aggregated soils present major engineering problems when they are saturated, i.e. suction in the soil decreases. When suction in the aggregates fall below a critical value, the aggregates may swell and occupy pore spaces due to the change in stresses. Water will also be taken up by the air-water interface at the aggregate contacts, which can reduce the inter-aggregate stresses and shear strength at the contacts leading to collapse settlement (Sivakumar et al., 2006).
CHAPTER 3 EXPERIMENTAL PROCEDURES
3.1 Sample preparation
The clays used have to be sieved to a powder form prior to mixing with the exception of Kaolin, which is readily available in powder form. The clays are crushed in a mortar and dry sieved with a 425μm sieve. For this work, a minimum of 2kgs of clay is needed for each clay type.
To make two samples, 300g of clay is required. The clay is saturated with de-aired water at different moisture content according to clay types. Kaolin is saturated at 1.5 X 70%, London clay at 1.5 X 50% and Virginia clay at 85%. The clay and water are mixed thoroughly with a high speed mixer and later pressed repeatedly with a palette knife on a glass plate to ensure that there are no lumps or aggregates. It is important that the mixture is smooth and contains no structure at all. This mixture should be cured for 24 hours before compression.
A cylindrical compression chamber (Fig. 3) is used to form the sample. The chamber consists of a piston that moved upwards when pressure is applied. The inner diameter of the chamber is 50mm, hence samples formed are of this size. A saturated stone disc is placed on top of the piston followed by a piece of filter paper. The mixture is then scooped into the chamber and onto the piston. It must be ensured that a sufficient amount of mixture is placed into the chamber as the resulted sample formed will be at a shorter height compared to the mixture prior to compression. The mixture is repeatedly disturbed with a palette knife to ensure that there are no air pockets in the mixture. Once the mixture has been placed inside the chamber, another piece of filter paper is placed on top of the mixture followed by another saturated stone disc. The top of the chamber is then closed with a lid that is tightly screwed onto the chamber. A drainage line is attached to the lid which drains water out from the top whereas the piston itself will drain water out from the bottom of the chamber. A pressure line is attached to the bottom of the chamber to apply pressure to the piston which will cause it to compress the mixture by moving upwards.
The sample is ready when water stops draining from the chamber. For Kaolin and Virginia clay, this takes two days whereas London clay will take up to three days. The sample is extracted from the chamber and is placed inside a mold. A wire knife is used to cut off the edges of the sample and to slice the sample in two. Two separate identical samples are produced from 300g of clay. The samples are wrapped in cling film and sealed in a plastic bag so that the samples do not get significantly drier with time.
3.2 Setting up
Firstly, the drainage line connected to the GDS controller (Fig. 4) is filled with water until it flows out of the chamber (Fig. 5). It must be ensured that no air is trapped within the drainage line. The chamber is shaken as water flows out to remove the air bubbles trapped inside. A saturated stone disk is placed on the pedestal of the chamber followed by a piece of filter paper. The prepared sample is then placed on top of the filter paper. Another piece of filter paper is placed on top followed by a steel disk. A thick silicone membrane is wrapped around the sample and the disks (Fig. 6). The membrane is stretched by using a membrane stretcher (Fig. 9) attached to a vacuum pump. Once the membrane is fitted onto the sample, two O-rings are fitted onto the membrane; one at the bottom of sample on the pedestal of the chamber and the other at the top, around the steel disk (Fig. 6). Once the sample has been set up, the chamber is then screwed tightly onto its pedestal. At the top of the chamber, a small opening is left opened as the chamber is filled with de-aired water. As water is filling up the chamber, it is again shaken to allow the trapped air bubbles to escape from the chamber (Fig. 8). Once the chamber is completely filled with water, the opening is closed by tightening its screw.
3.3 Applying suction
To apply suction to the sample so that aggregate structures can develop, a flexible tri-axial chamber (Fig. 7) is fabricated for this purpose. The chamber is connected to an Automatic Consolidation System (ACONS) and a GDS controller (Fig. 4). The ACONS is connected to a computer which controls the pressure applied to the system. The ACONS is filled with water and when pressure is applied, the water in the system exerts a pressure on the sample in the chamber. The drainage line of the chamber is connected to the GDS controller where the pore water pressure is set to 200kPA. This will exert a suction force on the sample, thus allowing aggregate structures to develop.
Seven values of suction are applied to the sample during the test. For every suction value applied, the volume of water discharged is noted after the designated amount of time has passed. For the first test, all seven values are applied to obtain the characteristic curve of that clay type. For the following tests, the values are applied to individual samples of clay. These samples are then extracted after one day for Kaolin and Virginia clay, and two days for London clay. Due to the limitations in the system, an exact value of suction cannot be applied. Therefore, approximate values of suction are used: 400kPa, 600kPa, 1350kPa, 2800kPa, 5900kPa, 8900kPa and 12000kPa.
3.4 Extracting samples
Before dismantling the system, it must be ensured that the suction is removed and all valves are closed. It is also necessary that the drainage line is vacuumed before dismantling so as to ensure that water does not return into the sample. The samples are extracted and cut into 1cm3 cubes and are stored in airtight plastic containers. They are then freeze-dried with liquid nitrogen and vacuum to -40°C and approximately 5.0mbar for 48 hours in which they will be ready for SEM. This is to remove any liquids that may interfere with SEM while preserving the original fabric (Delage et al., 1982).
3.5 Scanning electron microscopy (SEM)
Scanning electron microscopy (SEM) involves imaging a sample by scanning a beam of electrons to obtain the sample topography. SEM has been extensively used to study soil structure (Collins and McGown, 1974; Delage and Lefebvre, 1984; Vulliet, 1986). In this work, only the microscopic structures of the samples are needed. An electron beam of 3.0kV is used in this case to obtain the microscopic images of the samples. The samples are first coated with gold, a conductive material because clay is non-conductive. Images of x1000, x5000 and x10000 magnification are obtained.
CHAPTER 4 RESULTS AND DISCUSSION
4.1 Soil water characteristic curve
The soil water characteristic curve is an extension of the present work. According to Williams (1982), the soil water characteristic curve is defined as the relationship between water content and soil suction. Studies have shown that there is a relationship between the soil water characteristic curve and the properties of an unsaturated soil (Fredlund and Rahardjo, 1993). However in the present work, the void ratio e is plotted against the natural log of P' (i.e. ln P') to obtain the soil water characteristic curve.
Where P' is the average effective pressure, P is the cell pressure, and uw is the pore water pressure (set at 200kPa throughout the procedure).
The void ratio e is obtained from the basic volume-mass relationship:
Where e is the void ratio, S is the degree of saturation (which is assumed to be 100% throughout the procedure), w is the water content, and Gs is the specific gravity (taken as 2.65 for all clays).
The gravimetric water content w is calculated by:
Where w is the gravimetric water content, Mw is the mass of water and, Ms is the mass of soil solids.
The soil water characteristic curves are obtained for Kaolin, London clay, and Virginia clay:
As seen from the curves, the clay samples experience a reduction in void ratio as suction increases. If suction is increased to a much higher magnitude, the curve will eventually become a flat line (Fig. 10 d) because the soil has reached it shrinkage limit, which can be defined as the water content in which any further loss in water will not result in volume reduction (Atterberg, 1913).
If suction is increased beyond 12000kPa, the curve will eventually become a flat, horizontal line as the soil has reached its shrinkage limit. The figure above illustrates a predicted soil water characteristic curve for Kaolin where its shrinkage limit is reached (taken as 12%). At shrinkage limit, the sample will not experience any further volume reduction, thus no reduction in void ratio.
It can be observed from the three curves that at lower pressures, the curves exhibit an almost linear relationship. However at higher pressures, it is less linear as the shrinkage limit is ultimately reached. The linear part of the curve can be expressed as:
Where e is the void ratio, ln (P') is the natural log of the average effective pressure, N is the y-intercept of the curve, and λ is the pore size distribution index. The N and λ values for the three soils are computed and shown below:
Table Values of N and λ
The above equation is widely used in constitutive modelling, but the assumption of a straight line is invalid for the present work. It is assumed that at certain values of suction, the particles experience crushing under high pressures. This results in a slightly steeper curve and thus is not linear.
4.2 Scanning electron microscopy (SEM) images
SEM images are taken for all samples at x1000, x5000 and x10000 magnification so that the presence of aggregates can be identified. The SEM images shown in this report are all taken at x5000 magnification.
For Kaolin samples, it can be observed that signs of aggregation are present at smaller suction values. Aggregation is more prominent at suction values between 500kPa to 5000kPa. As suction reaches more than 8000kPa, there is less aggregation and aggregates are smaller in size. It can be postulated that there are not sufficient voids present for air to enter the aggregates or that there is no available space for aggregates to occupy. For London clay samples, larger aggregates can be observed. Aggregation appears to be more prominent at suction values higher than that of Kaolin samples (1000kPa to 8000kPa). Less aggregation is also observed at higher suction values. Similar to Kaolin, Virginia clay samples exhibit prominent aggregation at suction values of 500kPa to 5000kPa and less aggregation at higher values.
CHAPTER 5 CONCLUSIONS
The report describes a detailed experimental procedure and analysis carried out on previously saturated Kaolin, London clay, and Virginia clay. The experimental procedure involves applying suction to previously saturated clay samples at a range of suction values to assess the development of aggregate structure. The testing yielded data for a volume mass relationship used to plot the soil water characteristic curve, which is an extension to the present work. SEM images were taken to study the soil structure and bimodal pore size distribution of the soil at a microscopic level.
The SEM images showed that development of aggregate structures are prominent in the samples of suction values ranging from 500kPa to 5000kPa. However at higher values of suction exceeding 8000kPa, signs of aggregate structures are less distinct. This could be due to the reduced void ratio of the sample, allowing less space for the development of aggregate structures. The sizes of aggregates vary according to clay types, e.g. large and distinct aggregates observed in Virginia clay compared to London clay.
The results provide information on unsaturated soil behaviour and its bimodal pore size distribution which can provide further understanding on collapse settlement and other problems caused by unsaturated soils.