Keywords

1 Introduction

Expansive soils are a significant hazard for lightweight building foundations and highway pavements, because they experience changes in volume due to seasonal moisture fluctuation, swelling during wet seasons and shrinking during dry periods. If a pavement rests on expansive soil, longitudinal surface cracks may occur because of the seasonal volume change of the subgrade expansive soil [1]. Infrastructure damage due to expansive soils is commonly reported in many countries, such as Australia, Canada, England, China, India, and the USAs. Consequently, the need for change in conventional foundation construction systems is becoming a very important requirement to construct more sustainable houses and buildings with low maintenance costs over their lifetime.

Soil stabilization is an effective way to enhance the durability, mechanical characteristics and to reduce or eliminate the amount of volume change in expansive soils. Using lime and cement for soil stabilization are less cost effective and not environmentally friendly strategies due to the use of energy, resources and carbon footprint produced during the manufacturing process. Moreover, investigations revealed that cyclic wetting and drying cause arresting volume change behavior to be lost after the first wet–dry cycle, and consequently swelling potential increases after each cycle due to the formation of expansive material such as ettringite in calcium-based stabilized soils [2].

Much attention has been given to reducing the utilization of natural resources in cement and other traditional construction materials. For example, the construction sector generated roughly 92.5 million tonnes of asphalt concrete and 4.6 million tons of cement in 2013 and 2015, respectively [3, 4], necessitating alternative construction materials to minimize environmental impact and to preserve natural resources. Hence, the main goal of research by both scientists and engineers has been reducing the demand for natural resources as well as to minimizing the disposal of wastes such as glass [5,6,7]. Significant research has been performed into the use of waste glass in construction. For example, cullet has been tested as aggregate in the construction sector ranging from concrete and cementitious materials to roadway and asphalt construction. In addition, waste glass can be utilized in the manufacture of ceramic-based products [8].

The mechanical behavior of clay soils can be significantly improved by adding glass powder to the raw soil [9]. Adding fly ash and cement together with recycled glass powder increases the shear strength and CBR of soil [10,11,12]. Individually, glass powder can increase the strength of cement-stabilized expansive soil and decrease the plasticity index of the soil mixture. Furthermore, the addition of glass powder increases dry density, CBR and UCS with a reduction in optimum moisture content [13]. However, the effect of crushed glass on soaked CBR can be less pronounced compared with unsoaked conditions. In terms of other additives, Phanikumar showed that the addition of fly ash to expansive soils can modify the pore orientation of the soil while significantly improving its compaction behavior [14].

Previous studies identified the optimum mix design using waste glass and secondary additives, proposing specifications for a novel capping layer that can be applied to minimize the impact of soft subgrades on foundations [15]. A series of mechanical tests (UCS, standard compaction test and direct shear test), hydraulic conductivity test (permeability) and microscopy test (X-ray diffraction, scanning electron and porosity test) were carried out to explore the optimum combination of glass wastes and secondary additives for field trials in expansive soils.

The aim of current study is to evaluate the performance of a capping layer by applying it in large-scale tests as verification of the proposed novel capping layered foundation for buildings and roads. A laboratory experiment was carried out to investigate the performance of a foundation slab placed on an expansive soil. The proposed capping layer was placed between the foundation and weak subgrade clays to evaluate the foundation’s performance under environmental and operational loads. Performance of the foundation and soil conditions were carefully monitored across a period of time under operational loads for verification of the proposed capping layered foundation system.

2 Methods

2.1 Materials

2.1.1 Soil Characteristics

Soil was obtained from a land excavation site in Melbourne, Australia. The soil was categorized as fine-grained (CL) [16] with a high degree of expansion. All the physical properties of the soil are shown in Figs. 1, 2 and Table 1.

Fig. 1
A graph of percentage passing percent versus particle size. The curve extends between (0.3, 11) and (30, 100). The pointed nodes are (1, 30), (5, 50), and (10, 90). The values are approximate.

Particle size distribution of soil

Full size image
Fig. 2
A graph of dry density versus moisture content percent. A downward parabola extends between (14.2, 1.79) and (21.8, 1.70), passing through (17, 1.83) and (18, 1.82). The values are approximate.

Soil standard compaction test result

Full size image
Table 1 Summary of soil characteristics
Full size table

2.1.2 Crushed Recycled Glass Powder

Crushed glass behaves like natural rock and is totally inert and non-biodegradable. Glass powder (GP) was used in this study. Particle size distribution and samples are shown in Figs. 3 and 4 respectively.

Fig. 3
A line graph of percentage passing percent versus particle size. Glass aggregate extends between (90, 0) and (12000, 100). The glass powder extends between (2000, 0) and (12000, 100). The values are approximate.

Particle size distribution of glass powder and glass aggregate [15]

Full size image
Fig. 4
A photograph of a closer view of glass power particle size.

Glass powder used in the study [15]

Full size image

2.1.3 Fly Ash

Class F fly ash from a local supplier was the primary additive used in the capping layer. For all types of fly ash, the particle size distribution is similar to that of silt (Fig. 5). Typical physical and mechanical properties for fly ash are shown in Table 2. Other characteristics of the fly ash including chemical composition can be found in Karami et al. [17].

Fig. 5
A graph of percentage passing versus particle size. The curve extends between (1.00, 0) and (120.00, 100) by passing through (10.00, 20). The values are approximate.

Fly ash particle size distribution [17]

Full size image
Table 2 Physical properties of fly ash [14]
Full size table

2.1.4 Lime

Hydrated lime, which is the most concentrated form of lime, was used as a secondary additive. Lime decreases the liquid limit of soil mixtures, resulting a reduction in plasticity index. Typical physical and engineering properties of hydrated lime are similar to what was used in this study are shown in Table 3.

Table 3 Physical and chemical properties of hydrated lime [14]
Full size table

2.2 Experimental Set-Up

2.2.1 Stage 1: Control Test

Preparation of the soil required an extremely thorough and careful process to obtain as homogeneous material as possible. In preparation for the test, the natural soil was first sequentially sieved with a 25 mm sieve. A large-scale soil box (L: 1900 mm, W: 750 mm, H: 1000 mm) was built to enable the control of hydraulic and mechanical processes of the soil by changing the water table in the soil. A drain pipe and a valve system were installed at the bottom of the box (i.e., at a depth of 1000 mm) to allow drainage and to control the water table inside the box. Sieved soil was mixed with water in a concrete mixer until a homogenous mixture with optimum moisture content was obtained. The soil was then placed in the test box in 4 layers and compacted in 100 mm rises (Table 4). Figure 6 shows the details of the test box setup. A geotextile sheet was placed over the bottom surface of the box, a 100 mm thick drainage layer made of gravel was placed on top and finally a filter paper before the soil. The top soil surface at the top was left exposed to atmospheric conditions to facilitate evaporation and infiltration. The experiment began late December 2022 during the summer, and continued for 6 months.

Table 4 Soil layer characteristics and thickness for stages 1 and 2
Full size table
Fig. 6
Three illustrations. a. The top view of the sensor locations measured 750 by 1900 millimeters. B and c. The layers are measured from 0 to 670 and 0 to 820, the labels are marked, and the digital replacement gauge and U B I B O T are labeled.

Proposed sensor locations for prototype test: a top view, b stage 1, c stage 2

Full size image
Table 5 Instruments
Full size table
Table 6 Load calculations
Full size table

As shown in Fig. 6 and Table 5, six moisture probes and three digital movement sensors were installed in different locations. Two temperature/humidity sensors were also installed in the test box to record both the soil and room temperatures. Data were recorded in the controller at a frequency of 1 h and transferred to a computer daily. In addition, weekly manual readings of the movement were taken. A sprinkler system was constructed above the sandbox to simulate rain events similar to field conditions. Water spray was based on the water table depth at specific times and amounts. Water percentage was obtained from the swelling test results for raw soil according to Australian Standard (AS 1289.7.1.1-2003). The amount of water was calculated based on the raw soil amount in the box (Table 7).

Table 7 Water added to the sample
Full size table

For Stage 1, a concrete slab scaled down to L: 750 mm, W: 550 mm, and H: 100 mm was placed on top of the compacted soil to apply the dead loads during the test period. The size of the concrete block was scaled down from an actual designed slab for a residential building following AS 2780. In this test, the dead load was added to the slab for increasing the thickness of the slab from 100 to 170 mm for a residential-purpose building. The live load was applied to the soil by specific weights at the end of wetting and drying periods. The standard load amounts and the calculated applied load based on the slab area are shown in the Table 6.

2.2.2 Stage 2: Test with Capping Layer

In Stage 2, all steps in Stage 1 were repeated, but a capping layer was constructed on top of the raw soil (Fig. 6c). The design mix of the capping layer was based on previous studies [15]. Materials were mixed with soil in percentages of 25% WG powder, 7.5% fly ash and 3% lime by weight compared with raw soil. To prepare the stabilized soil, all ingredients were mixed properly in the dry state in a mechanical mixer. Water was added to the mixture up to the optimum moisture content according to the Standard Compaction test and well mixed. Similar to Stage 1, the soil (subgrade soil) was uniformly compacted in the box. The size of the installed capping layer was same as the raw soil in length and width to cover the whole soil surface area. Thus, the size of the capping layer was L: 1900 mm, W: 750 mm, and H: 100 mm.

3 Results and Discussion

3.1 Slab Settlement

At the beginning of both stages of the prototype test, the slab started settling down for the first few days after it was placed on the soil surface, due to the weight of both the slab and the capping layer on the raw soil, which caused a secondary compaction process during those first few days. The results show that settlement in Stage 2 was greater than in Stage 1 at early stages. The maximum settlement in Stage 1 was 0.06 mm and for Stage 2, it was 0.10 mm after week 1 for the control and stabilized capping layer tests respectively (Fig. 7). The larger settlement in Stage 2 resulted from elastic compression of the capping layer, which consisted of coarser materials than raw soil. However, this increment in initial settlement was negligible and did not imply a reduction in bearing capacity.

Fig. 7
Three graphs of slab movement and moisture content percentage versus time. It depicts slab movement average stabilized, control, surface moisture stabilized, and control. All denote an increasing trend.

Slab movement and moisture profiles for stage 1 and stage 2 during the wetting period from week 1–9. Horizontal profiles: a at soil surface level, b 200 mm below the soil surface level, and c at the base of the soil layer for Stage 1 and 100 mm above the base for stage 2

Full size image

It is important to note that the subsequent slab movement due to swelling was reduced when the capping layer was introduced. The difference in maximum slab movement between the two stages was 0.2 mm. More importantly, the rate of increase in slab movement was less with the capping layer in place. In the control test, slab movement reached 0.4 mm in less than 5 weeks, whereas it took nearly 9 weeks to move by the same amount under stabilized conditions. Therefore, it can be concluded that the capping layer was effective in mitigating the magnitude and the rate of the vertical slab movement.

3.2 Expansive Soil Movement (Swelling)

The soil started swelling and increasing in volume as expected for any expansive soil. During the wetting period, the soil in the control test swelled by 0.6 mm on average (maximum 0.72 mm for individual gauges). In Stage 2, with the capping layer in place, the slab moved by only 0.397 mm (maximum 0.61 mm for individual gauges) (Fig. 7a–c). As shown in Fig. 7, the swelling in Stage 2 of the test with the stabilized capping layer was lower than with the control soil during the wetting period.

Because the capping layer was more permeable than the control soil, it assisted water to drain away faster and remove it from the slab and the expansive soil underneath. Moreover, the load applied to the soil from the weight of the capping layer helped to arrest or reduce the swelling of the expansive soil.

3.3 Moisture Distribution

Water was sprayed on top of the soil with the same pattern (time, amount, and size of spray nozzle) for both stages of the test. As shown in Fig. 7a, the capping layer tended to be mostly dryer than the raw soil in the control test, because of the higher permeability of the capping layer (Fig. 8). In the control test, water pooled on the surface or remained at a shallow depth (based on visual observations) after a wetting event whereas more water infiltrated the soil through the capping layer in Stage 2.

Fig. 8
A scatterplot of permeability versus glass percent. The glass powder G P is high at (25, 0.000001) and low at (0, 1 E 09). Soil plus 7.5 percent F A plus 3 percent plus lime plus G P high at (0, 0.00001) and low at (25, 0.000001). The values are approximate.

Permeability results for soil–glass powder and soil–fly ash–lime–glass powder mix [17]

Full size image

This was evident from the moisture variation at 200 mm from the surface (Fig. 7b). Moisture content at the raw soil-capping layer interface was higher than the moisture content recorded at 200 mm within the raw soil in the control test (Fig. 9). This was due to the rapid flow of water through the capping layer. However, the capping layer was effective in delaying water reaching the raw soil layer. Without the capping layer, the moisture content at the surface of raw soil was >30% in less than 3 weeks. When the capping layer was in place, it took almost 5 weeks for the water content at the top of the raw soil layer (interface between the raw soil and capping layer) to reach >30%. Consequently, the raw soil layer remained drier for a longer period, which will reduce the vertical movement of the slab.

Fig. 9
A graph of increasing depth versus moisture content. The title reads concrete slab 700 times 550 times 160 millimeters. It depicts control test W 2, W 6, W 8, and stabilized capping layers W 2, W 6, and W 8.

Variation of water content in control soil and stabilized capping layer in the wet season of the test: vertical profile

Full size image

The moisture variation shown in Fig. 7c corresponded to the condition at the base of the control test and at 400 mm depth below the surface for Stage 2. Interestingly, the difference in moisture content at these locations was narrow, indicating the raw soil layer remained drier in Stage 2 throughout the wetting period. Thus, it was evident that the capping layer kept the water away from the expansive soil underneath.

4 Conclusions

This research investigated a novel sustainable foundation system that could mitigate the adverse influence of expansive soil conditions for construction. The proposed approach is based on a capping layer constructed from recycled glass waste and other sustainable additives and placed directly above the existing expansive soil. A prototype foundation was constructed in the laboratory using an optimum stabilization mix design derived from a detailed investigation of mechanical and hydraulic characteristics. The slab movement and soil conditions were monitored over 6 months under simulated dry/wet moisture fluctuations and operational loads. Results showed that the foundation’s performance on the novel capping layer was significantly productive compared with control conditions. The higher efficiency of the new capping layer system was evidenced by 35% reduction in slab displacement during wet season simulation, which is a significant improvement in foundation serviceability on expansive soils. The outcome from this research will have a significant impact on minimizing infrastructure maintenance costs, which are a heavy burden for asset managers in any country. The proposed capping layer, which incorporates waste materials (32.5% of total), is a waste management strategy that minimizes the serviceability concerns of civil infrastructure while introducing a value-added benefit for waste materials. It is to be noted that the results provided in this document are based on the referred tests/materials conditions and should not be used in field applications without appropriate verification.