1 Introduction

The aging phenomenon, which reflects a material state change as function of the time, can produce considerable changes in the mechanical and physical soil properties. The long term change in these properties should be considered during geotechnical design. Previous investigators (Mitchell 1960, 1986; Leonards and Ramiah 1960; Leonards and Altschaefl 1964; Zeevaert 1949, 1983; Schmertmann 1983, 1987, 1993) developed a variety of reliable ideal models which account for soil aging. The aging process itself, however, remains a complex and uncontrolled phenomenon.

Generally, aging produces an improvement in soil properties. In select cases however (Schmertmann 1991), degradation of soil properties was observed. All soils age with time (Mitchell 1986). “Pure” aging is often characterized as the phenomenon which involves time-dependent changes only. Chemical weathering, freezing–thawing, swelling–desiccating and changes in ground-water level are the main in-situ processes observed during aging. These processes occur simultaneously and their individual influence is difficult to quantify during “pure” aging (Schmertmann 1993). Also, many references have focused on changes caused by later stage processes such as creep and/or secondary consolidation aging.

Thixotropy is widely linked to aging mechanisms (Mitchell 1960). Most geotechnical engineers associate thixotropy with an increase in compressive strength remolding. Mitchell (1960) has refined his definition of aging to include isothermal conditions and reversible behavior at constant composition and volume. Mesri (1993) similarly defines thixotropic hardening as a reversible process that can occur under conditions of constant composition and volume.

Schmertmann (1991), presents examples of improvement in preconsolidation, modulus, strength, and bearing capacity based on lab and field studies. He also describes soils where little aging effect was observed. Schmertmann (1991) suggests that in those solid displaying a lack of observable improvement with age, other processes might be at work. Schmertmann (1991) also argues that the effects of aging can be explained by mechanics and, that strength gain is due to an increase in frictional resistance and not from any increase in cohesion.

Mitchell and Soga (2005) also address the effect of time on the strength and deformation of soils. Again, their focus was on how time changes the structural, deformation, and strength properties of soils under stress. They present suggestions and methods to quantify and mathematically predict the effects of time on soil properties. In recognizing that macroscopic stress on a soil results in tangential and normal interparticle stresses which can be modeled by discrete particle simulation. These models “show that changes in creep rate with time can be explained by changes in the tangential and normal force ratio at inter-particle contacts that result from particle rearrangement during deformation.” They conclude that changes in the micro-fabric lead to a non-homogeneous strong particle network with locally weak clusters which explain the mechanical aging process.

Dunn and Mitchell (1984) relate the time effect of aging to the permeability (hydraulic conductivity) of fine-grained soils. Hydraulic conductivity was found to increase with the delay in time between sample preparation and test initiation. This was attributed to thixotropic changes in sample fabric. They argue that flocculation of soil particles increases with time after compaction. This change in fabric increases the effective pore size, which explains the higher hydraulic conductivity. Mitchell et al. (1965) also document an “increased degree of flocculation” when they find a high correlation between the increase in hydraulic conductivity with time and the increase in strength and modulus of soil.

Previous paragraphs document the complexity and practical importance of the aging phenomenon. This paper reports the results of experimental studies performed with observational periods sufficiently long to document the aging process. In particular, the main focus of this study is the influence on the relationship between storage time and constitutive soil properties as measured in laboratory. Storage time is defined as the time between sample compaction and the initiation shear strength. The effect of aging (soil storage) on shear strength was assessed by performing undrained unconsolidated triaxial compression tests after the soil had been compacted and stored for various periods of time. The laboratory measured influence of the storage time on soil permeability provided a secondary focus for the present study.

2 Materials and Methods

The experimental study was carried out on a silt originating from the Xeuilley area, located at Nancy North–West (France). The silt had a Liquid Limit of 56, Plastic Limit of 31, and a Plasticity Index of 25 (as determined in accordance with ASTM D4318). The specific gravity (Gs) of the soil was 2.64 (in accordance with ASTM D854-10). The silt is classified as MH to the unified soil classification system (USCS). The mineralogical composition of the silt used in this study is shown in Table 1.

Table 1 Mineralogical composition of soil used in this study
Full size table

Shear strength tests were performed on samples compacted with the Harvard Miniature apparatus developed by Wilson (1970). This apparatus is unique in that the soil undergoes kneading and homogenization during compaction. This differs significantly from the widely used Proctor test which relies on solely on impact force for compaction. The use of the Harvard Miniature for compaction is justified because of the need to obtain more homogeneous samples. Laboratory testing with the Harvard apparatus determined that optimum water content of the soil was 22 % which was associated with a maximum dry density of 1.6 g/cm3, as measured using the Harvard Miniature, in accordance to the procedures described in “Suggested Method of Test for Moisture–Density Relations of Soils Using Harvard Compaction Apparatus” published in 1970. Permeability testing was performed on samples compacted according to the standard Proctor method (ASTM D698) to a water content of 23 %. Compaction curves were obtained for both Harvard and Proctor compaction are shown in Fig. 1. The compaction curves were very similar, but one should recognize that this does not means that there no difference in soil structure between the two method of compaction. The similarity of both compaction curves should not be taken to indicate similarity of soil structure.

Fig. 1
figure 1

Compaction curves obtained using standard Proctor and Harvard Miniature

Full size image

2.1 Shear Strength Testing

After compaction testing, the samples were divided into four groups and undrained unconsolidated triaxial compression tests were performed on the samples in each group in order to study soil shear strength parameters, in accordance with the procedures described in (ASTM D4767-04). Table 2 presents the triaxial compression test program. The compacted samples had a water content ranging from approximately 23–24% with only a slight deviation in dry density The samples were then aged for various periods of time. Air-tight bags were used to isolate the sample from temperature and humidity variations in the ambient air. These samples were placed in a hermetic container to ensure a constant temperature and a constant relative humidity. Redundancy was provided by analyzing multiple samples for each group and aging period.

Table 2 Summary of triaxial shear tests program
Full size table

After a pre-determined amount of time, the samples were retrieved and the water content was again computed. Triaxial testing was then performed for the determination of the shear strength parameters. The strain rate for all triaxial tests was maintained at a constant 0.9 mm/min. The rate of shear was kept constant during all tests to avoid introducing another variable into the investigation.

For each period, every group of samples were sheared under various confinement pressures (50, 100, 200 and 400 kPa), see Table 2. A single sample group was sheared immediately after compaction in order to be used as reference for the other samples. Height and diameter of the samples were 7 and 3.13 cm respectively.

2.2 Permeability Tests

The first sets of samples was used to carry out two permeability tests immediately after compaction. The remaining samples were isolated from ambient at and held for testing after various storage periods. For each test, water content was determined and the stored sample was divided into two separate specimens. Table 2 shows the water content, dry unit weight, and degree of saturation before and during aging (Table 3).

Table 3 Initial and final samples characteristics for permeability testing
Full size table

Permeability tests were performed in accordance with the procedures described in ASTM D5084-90. Once the samples were prepared, they were placed in a triaxial cell to launch the saturation phase. Variation in soil volume during the test was monitored.

Sample consolidation, a major concern during permeability testing, was accounted for by reducing the effective pressure during testing. The lower effective pressure insured against additional consolidation upon application of the hydraulic gradient. The bottom and top backpressures of 305 and 295 kPa were applied yielding a hydraulic gradient of 10. Testing continued until the last four hydraulic conductivity values were within 25 % of the mean, inflow equaled outflow (within 5 %), and stabilization of the hydraulic conductivity and outflow/inflow.

3 Results and Analysis

3.1 Shear Strength Results

Figure 2 shows typical deviator stress variation versus strain for three different times of aging (immediately after compaction, 90 days after compaction, and 328 days after compaction for various confinement pressures (50, 100, 200, and 400 kPa). Either two or three replicates were used for each aging time and at each confining pressure. (Also note that only curves three aging times are presented in Fig. 2 to keep the graph clear. The same patterns were observed for all other aging times in a previous investigation (Ltifi 1998). Figure 2 shows that the soil samples acquired more strength during the storage time, which corresponds to higher deviator stress as the soil aging increased. Moreover, a very distinct increase of the in deformation modulus as function of sample age is apparent in Fig. 2.

Fig. 2
figure 2

Deviator stress versus strain for three aging times (Black immediately after compaction, Blank 90 days after compaction, and Gray 328 days after compaction) for four various confinement pressures. Note the water content of these samples were all around 22 %

Full size image

Figure 3a shows the maximum deviator stress versus aging time for all tested specimens. For all confining pressures, the maximum deviator stress increases with aging time. These results are in agreement with the finding of Yashura and Ue (1983) Regression analysis of the data used to generate Fig. 3a provides R2 values of 0.85 or higher.

Fig. 3
figure 3

Maximum deviator stress versus aging time

Full size image

Figure 3b shows the soil deformation modulus versus aging time for the same soil specimens. For all confining pressure, the soil modulus increases as the aging time increases. Again, regression analysis on the data used to generate Fig. 3b yields R2 values of 0.72 or greater.

Figure 4 examines the relationship between undrained shear strength of the soil and sample age. An important increase in the undrained shear strength as function of sample ages is evident in Fig. 4. Regression of undrained shear strength on soil aging times shows a very high correlation between the two variables (R2 = 0.92). The obtained results indicate that aging time explains much of the increase in undrained shear strength, as well as the increase in deformation modulus of the soil. The results show that, for increasing aging time, the undrained shear strength as well as the deformation modulus of the soil will increase in value.

Fig. 4
figure 4

Undrained shear strength versus aging time

Full size image

These results lead one to question the physical processes responsible for this distinct improvement. Similar soil strengthening was reported by Zeevaert (1949), Shmertmann (1991), and Mesri (1975). These authors attribute the strength gained during aging to the increased rigidity which results from thixotropic hardening during the sample storage.

Several investigators attribute the increased resistance of soil during aging to chemical pore water alteration and/or changes in solid matter material constituents (Mitchell 1986; La Rochelle et al. 1986; Graham et al. 1990). In this study, according to the nature of the materials mineralogy (namely quartz), this increase is not expected to have a chemical origin, i.e. it must have been initiated by the transformations that the material structure could have undergone. Mesri (1993) defined thixotropic hardening as the purest form of soil aging. It does not require volume change or chemical alteration. Only the influence of aging is expected on the samples prepared for this study.

These conclusions agree with the findings of Zeevaert (1949) who worked with Mexico City clay. The conditions under which the test in this study were carried out, and the verifications made by measurements, in particular water content and dry density, show that the temperature and moisture have an insignificant influence on material. In addition, quality control was used to eliminate sample groups whose preparation water content variations were different. Thus, the results obtained during this study seem to be caused by a pure aging mechanism defined by Mesri (1993).

3.2 Permeability Results

After application of the hydraulic gradient, inflow and outflow quantities through the test apparatus were carefully monitored. Figure 5 shows the permeability versus sample aging time. Due to experimental constraints, permeability testing was performed on a maximum aging time of 232 days. Sample permeability decreases with increasing aging time. This reduction is very important during the first 5 months. Beyond the 5 month aging duration, the permeability becomes stable. This is probably due to the structural changes that occur during the storage period. These results are inconsistent with the findings of Mitchell et al. (1965), where he reported an increase in permeability due to the change in soil structure with. He also argued that the increase in permeability is due to same change in structure responsible for the increase in strength. It is difficult to relate the permeability results obtained during this study to the strength results because each set of samples were compacted using two distinct compaction methods, as described in the previous section.

Fig. 5
figure 5

Permeability versus aging time

Full size image

Dunn and Mitchell (1984) ascribed the reduction in permeability during the permeability testing period (while the sample is saturating and permeating as being due to several causes such as the micro-organisms growth), secondary sample consolidation and progressive filling by the migration of very fine particles. The effects of these phenomena could negate the expected increase in permeability with aging time.

Sample permeability converged as aging time increased. The difference in permeability between the two initial specimens (aging time is zero) is around one half order of magnitude. The difference decreased as the aging time reached 50 days and becomes insignificant for aging time higher than 150 days. This might be further evidence that that soil structure changes with aging until a steady state is reached.

4 Summary and Conclusions

This paper describes an experimental study on the effect of storage time on the aging of a silty soil on undrained shear strength and soil permeability (hydraulic conductivity). Several unconsolidated undrained triaxial shear tests and triaxial permeability tests were performed. These tests were performed on compacted samples which were stored for differing time periods. Compaction was carried out using the miniature Harvard test apparatus (shear strength testing) and the standard Proctor method (permeability testing).

The following conclusions can be drawn:

  • The strain modulus and material rigidity increase with increasing aging time.

  • The strain at failure is not as affected or influenced by the aging time.

  • Undrained cohesion increased as time of storage increases.

Mechanical behavior change is consistent with Zeevaert (1949), Shmertmann (1991), and Mesri (1975). This increase in shear strength was attributed to aging by thixotropic hardening during the sample storage. The effect of aging should be accounted for when determining the properties of compacted and/or natural materials stored at the laboratory for extended periods of time before testing. Shear strength is commonly overestimated in aged soils. The increase in the deformation modulus and undrained cohesion can reach up to 100 % for a storage period of about 328 days. Even though in-situ soil modification takes place under drained conditions, (which differ from the laboratory conditions), this study provides a helpful reference point in describing and quantifying the aging time effect on undrained silty soils.

The soil permeability decreases as storage time increases. This is not consistent with Mitchell et al. (1965), who reported an increase in permeability due to the change in soil structure with time. Mitchell et al. (1965) also argued that the increase in permeability is due to same change in structure responsible for the increase in strength. Dunn and Mitchell (1984) suggest that the reduction in permeability during the testing period (while the sample is saturating and permeating) is caused by a variety of phenomena including micro-organisms growth, secondary sample consolidation and progressive filling by very fine particle migration. These phenomena might have negated the expected increase in permeability with aging time.