3.1 Effect of CCR and NS ratio on UCS
Figure 3 shows the variation in the strength of solidified mud at various CCR/NS ratio for different curing time. When the CCR/NS ratio ranged from 10:0 to 2:8, the strength increased significantly with decreasing CCR/NS ratio until the maximum value at optimal CCR/NS ratio (6:4) was reached and then decreased steeply. In this mixture, the optimal CCR/NS ratio did not change with the increase of curing time and remained at 6:4, for which the highest 7-day, 14-day, and 28-day strength were 0.9MPa, 1.6MPa, 2.2MPa, respectively. However, when CCR/NS ratio was reduced to 2:8, the strength was too low to measure at 7 days of curing and did not increase from 14 days to 28 days, which was similar to that at 3:7 and 4:6. Other than that, the strength increased with curing time for all CCR/NS ratio due to that the high content of NS only bounded the soil particles together in the early term and did not attain an increase in strength with time.
Densification and pozzolanic reaction are the essential reasons for early and long term strength development, respectively. After the addition of pulverized granular CCR and FA, the water content per unit mass and fluidity of the untreated mud decrease. The spherical and porous CCR and FA particles having similar particle size to clay fill the pores and around the clay particles, resulting in the maximum short-term strength related to the maximum dry unit weight. Meanwhile, the clay and FA particles contain alumino-silicates layers with negative surfaces which attached a swarm of cations (such as K+, Na+, and Mg+). The Ca2+ ion hydrated by CCR replaces the monovalent cation in the double layer cation with an ion exchange behavior, performing a higher polymerization power due to the weaker inter-ionic repulsion compared with monovalent cation. Then clay particles agglomerate with each other and form a stronger clump together with FA.
In addition to filling and agglomeration, the time-dependent pozzolanic reaction is the trump contributor to strength improvement. CCR leaches aluminum and silicon in the amorphous phase of FA, accelerating the pozzolanic reaction to produce cementitious gels of CSH and CAH. The Ca2+ ion plays the role of charge-balancing matter within the polymeric structure and supplies extra nucleation sites for rapid hardening and precipitation of coagulable particles (Phetchuay et al. 2016). However, the Ca2+ ion plays a role at relatively low alkaline system, and its dissolution will be hindered at too high OH− concentration. Such that, when the CCR/NS ratio was more than 8:2, the polycondensation reaction was hindered and early aluminosilicate gel precipitation was produced by excessive OH− ion concentration (Phummiphan et al. 2017). As a consequence, this early precipitation impeded further reaction by encasing the FA particles, resulting in negligible strength gain at a high CCR/NS ratio. This strength optimization of CCR/NS activated FA-based geopolymer is owing to the reason that the content of CCR, which accelerates the hydration process, determines the activation of FA and final quantity of cementitious products, while NS is responsible for the degree of consolidation of soil particles. Such that, in the low alkaline system, the quantity of OH− was insufficient to dissociate the alumina and silica to produce alumino-silicate gels. It makes sense that when the ratio continued to decrease below 5:5, pozzolanic reaction almost rarely happened and the growth in strength originated only from NS.
3.2 Variations of water content and pH value
The water content of the untreated mud (without binder) and solidified mud with different CCR/NS ratio at different time are shown in Fig. 4. When 14% of CCR and 14% of FA (without NS addition) were added, the water content decreased sharply from 141–83%, which was due to the water absorption of dry powder CCR and FA in the mixture. And then, with the increase of NS content, the water content increased again due to the decrease of CCR content and the introduction of NS aqueous solution. When CCR/NS ratio was 5:5, the water content was 95%.
After curing for 7 days, the water content of the untreated mud decreased slightly, from 141–137%, which was consistent with the fact that the untreated mud stored at the construction site is not easy to evaporate and precipitate. In contrast, the water content decreased from 83–74% after 7 day of curing with the addition of CCR and FA. At the same time, with the increase of NS proportion, the water content increased inversely, but all of them had different degrees of reduction compared with the initial water content. After 28 days of curing, the water content of untreated mud was maintained at 130%, compared with the effective reduction of water content of the mud with the addition of curing agent. When the CCR/NS ratio was varied from 9:1 to 5:5, the water content at 28 days were 74%, 75%, 76%, 80%, and 85%, respectively. It indicates that the CCR/NS ratio determines the initial water content, and the pozzolanic reaction that leads to the strength increase also makes the decrease of water content with curing time.
Figure 5 shows the variation of pH value of the untreated mud (without binder) and solidified mud with different CCR/NS ratio. The pH value of the solidified samples increased sharply after the addition of Ca(OH)2-rich CCR, and the pH of the sample was 12.29 at the initial moment when 14% CCR was added (without NS). The pH value decreased gradually with increasing the proportion of NS and reached 11.95 at the optimal CCR/NS ratio of 6:4. The results indicate that the optimal pH of CCR-NS as activator for the pozzolanic reaction is about 12, which can improve the activity of FA and clay particles with maximum efficiency. After curing for 7 days, the pH value of all the solidified samples decreased due to the consumption of OH− by the pozzolanic reaction, while there was no significant change in pH of the untreated mud. And then the pH value of the solidified samples decreased further after 28 days of curing time. When the CCR/NS ratio was 9:1,8:2, 7:3, 6:4, and 5:5, the pH was 11.89, 11.78, 11.41, 11.2, and 11.4, respectively. Notably, the effect of CCR/NS ratio and curing time on pH shows the same rule as on water content that the CCR/NS ratio determines the initial pH, and then the curing time shows the great impact.
3.3 Deformation characteristics
The typical stress-strain curve and appearance characteristics of solidified mud are shown in Fig. 6. As an anisotropic, non-homogeneous, and elasto-plastic material, solidified mud contains numerous internal pores showing a nonlinear relationship between stress and strain. Four phases of deformation that go through in stress-strain curve, so have been referred to as compaction, elastic, failure-developing, and failure phases.
Because of the presence of pores which delay the development of strength, the soil skeleton can not function until the pores are completely closed. So compaction phase is when it is mostly that the pores between the particles are compressed as the strain increase but the stress gains are relatively slow. A reduction in the vertical length of the sample is observed under the infliction of pressure, such that the loose structure becomes dense with preparing for the second development phase of strength. The second phase is that the elastic phase where there is a linear relationship between stress and strain with the increase of the strain after macro-pores compressed. The denser structure is responsible for the steady enhancement of strength. In the third phase, the stress and strain no longer maintain a linear relationship, i.e., the stress increases slowly with strain loading until it reaches a peak value. It is observed that multiple vertical cracks appear on the surface of the samples under compression, along with obvious plastic deformation. The fourth phase is the failure phase, where more micro-cracks are generated and the large vertical penetrating cracks rapidly propagate and expand within the samples, resulting in an obvious lateral slippage and plastic failure. As a consequence, the stress declines steeply with the increase of strain, but there is still residual strength presented in the plastic failure phase.
The variations in stress-strain curves of the samples cured for 7, 14, and 28 days containing different CCR/NS ratio are shown in Fig. 7. When the CCR/NS ratio was 9:1, the stress-strain curve at 7-day performed as a work hardening type (Li et al. 2021), whose failure strain reached 7% without significant peak stress. In addition, the failure mode showed plastic splitting damage, where the specimen was squeezed around under vertical pressure producing tympanites. The resistance against vertical compressive force was small and constant in the failure developing phase, such that, the specimen length was shortened apparently.
With the increase of peak stress, failure strain gradually decreased to about 3%. In addition, the residual stress after reaching the peak stress had a significant tendency to decrease while its descent speed was positively correlated with the magnitude of peak stress. The brittleness of the samples was revealed by the fact that the samples suddenly failed under vertical pressure, with vertical or arc-shaped cracks on the side and showing an hourglass shape of damage after sliding off the surrounding debris.
The brittleness or plasticity reflects two characteristics of damage destruction to materials, which is measured by εf, the strain corresponding to the ultimate compressive strength. The former is characterized by a small deformation and sudden destabilization after reaching the peak stress with large hazards; while the instability of the latter is gradual and its hazards can be predictable and evitable. It is seen that solidified mud is a suitable plastic material as subgrade or foundation fill, which breaks down after relatively large deformation and where the residual strain dissipates slowly.
The stress-strain curves can be used to obtain not only characteristic parameters such as ultimate unconfined compressive strength (UCS) and failure strain (εf) but also deformation modulus (E50), which are considered as the references for the toughness of the material. The deformation modulus E50 is a stiffness parameter for inelastic material, defined as the secant modulus of the point corresponding to 50% of ultimate stress in the stress-strain curve.
Figure 8 shows the variation between UCS and εf of the solidified samples with their fitted curve. The fitted equation that is established by regression analysis with a power function model has a high correlation coefficient (named R2) of 0.887. It can be seen that the failure strain decreases exponentially with the increase of UCS, reaching a minimum strain of 2.8% at UCS of 2.2MPa. Since the strength of the solidified samples at early-term of curing with non-optimal ratio admixture is mostly below 1 MPa, such that, the corresponding εf is concentrated between 3.5% and 7%. When the UCS is greater than 1.5 MPa, the εf drops to about 3% meeting the recommended deformation of the solidified soils for filling projects application. So this relationship can be used as a quick approximation for estimating strain value corresponding to a given strength where the high water problematic soils needed to be treated.
As shown in Fig. 9, there is an approximately linear relationship between the E50 and UCS. After establishing a linear model with UCS as an independent variable, the rate between E50 and UCS is gained for measuring the toughness of this material. The multiplicative relationship between E50 and UCS may vary for different types of soils, such as 18 to 53 for cement-stabilized zinc-contaminated soil (Du et al. 2013), 51 to 80 for sandy soil (Wang et al. 2018), or 53 to 95 for basalt fiber and metakaolin-based cement clay (Huang et al. 2020). In this study, the linear ratio of E50 to UCS is about 39 with a correlation degree of 0.938, and the difference of that may be related to the high water content and clay composition. These parameter relations of the solidified mud for foundation application should be considered based on the strength requirement, which provides parameter guidance to derive deformation and settlement values as well as to estimate the toughness of the solidified mud. As a ductile material, solidified mud has the ability to resist deformation and withstand loads for using in engineering, while achieving resource utilization of wastes.
3.4 XRD and SEM analysis
As shown in Fig. 10, the XRD pattern indicates the mineral component of FA-based geopolymer solidified mud with the CCR/NS ratio of 6:4 at 60 days of curing. The main clay mineral present in solidified mud was quartz (whose peaks were observed at 2θ = 22.2°, 27.0°, 39.8°, 50.4°), with a small amount of calcite (2θ = 29.7°) also being detected. In addition, the broad and amorphous peaks in the 2θ region of 9.5° to 30.5° were observed, which revealed the presence of alumino-silicate gel including gismondine (zeolites) and chabazite-Na (Sukmak et al. 2013). The appearance of several cementitious products was observed for the sample at several 2θ angles of 30° to 33.7°, and 48.5°, which were corresponded to the crystalline reflection of CSH (Sukmak et al. 2015. In addition, the reflection peaks at 29.5° and 34.0° implied the presence of CAH, which was similar to that observed in the stabilized WK (Latifi et al. 2016).
The geopolymer mechanism can be interpreted as the fact that Al- and Si-rich clay and FA particles are dissolved in an alkaline medium provided by CCR, and then form a 3D network structure of silica-aluminate chains after the polymerization reactions (Phummiphan et al. 2016). The following formulas (2) and (3) show the processes of pozzolanic reactions which produce cementitious gels of CSH and CAH. The role of NS is to produce more soluble CSH in the presence of Ca2+, the process is shown in formula (4) (Brykov et al. 2002). And then reacting with soluble Al2O3 in the alkaline solution to form silica-aluminate gel (NASH), as shown in formula (5). The cementitious gels CSH and NASH are prevalent in high Ca2+ and high Na+ systems, respectively. However, in this system, the coexistence of CSH and NASH was found, which was in agreement with the previous researches reported by Sukmak et al. (2013) and Phummiphan et al. (2017).
$$\text{S}\text{i}{\text{O}}_{2} + \text{n}\text{C}\text{a}(\text{O}\text{H}{)}_{2} + \text{x}{\text{H}}_{2}\text{O}\to \text{n}\text{C}\text{a}\text{O}·\text{S}\text{i}{\text{O}}_{2} ·\text{x}{\text{H}}_{2}\text{O}$$
2
$$\text{A}{\text{l}}_{2}{\text{O}}_{3} + \text{m}\text{C}\text{a}(\text{O}\text{H}{)}_{2} + \text{y}{\text{H}}_{2}\text{O}\to \text{m}\text{C}\text{a}\text{O}·\text{A}{\text{l}}_{2}{\text{O}}_{3}·\text{y}{\text{H}}_{2}\text{O}$$
3
$$\text{C}{\text{a}}^{2+} + \text{N}{\text{a}}_{2}\text{S}\text{i}{\text{O}}_{3}·\text{w}{\text{H}}_{2}\text{O} \to \text{C}\text{a}\text{O}·3\text{S}\text{i}{\text{O}}_{2}·\text{w}{\text{H}}_{2}\text{O} + 2\text{N}{\text{a}}^{+}$$
4
\(\text{S}\text{i}{\text{O}}_{2} + \text{A}{\text{l}}_{2}{\text{O}}_{3} + 4\text{O}{\text{H}}^{-} + 3{\text{H}}_{2}\text{O} \to 2\left[\text{A}\text{l}\right(\text{O}\text{H}{)}_{4}{]}^{-} + \left[\text{S}\text{i}{\text{O}}_{2}\right(\text{O}\text{H}{)}_{2}{]}^{2-}\) (Na)-A-S-H (5)
For observing the microstructure and pore characteristics of solidified mud, SEM was executed at the specimens with the CCR/NS ratio of 6:4 at 28 days and 60 days of curing. Figure 11(a) and (b) are the magnified 500 and 5000 times SEM images of the sample cured for 28 days, respectively, with MIP image of magnified 5000 SEM as shown in Fig. 11(c). Accordingly, Fig. 11(d), (e), and (f) are magnified 500 SEM, magnified 5000 SEM, and magnified 5000 MIP images of the sample cured for 60 days, respectively. As shown in Fig. 11(a) and (d), the FA particles (marked by the green circle) were distributed in the mixture, with a larger size than clay and bentonite particles (red circle). A large number of cementitious products CSH and CAH etched around FA and soil particles or in the pores were observed, which contributed to the development of strength. In addition, the huge cracks were also being seen in the whole microstructure. Figure 11(b) and (e) indicate that the more aggregated frameworks are formed by the rearrangement of clay and bentonite particles with flocculent hydration products and colloidal matter (correspond to NASH), in which flocculent products show dense matrix while NASH acts as an adhesive. Other than that, the porosity of the mud is equally influential in the development of strength, the pores depicted by the red-filled portion are shown in Fig. 11(c) and (f). Although the appearance of cementitious products produced by the pozzolanic and polymerization reactions stuffs the pores in the structure, those pores are still present and cause deficiency of the strength.
3.5 Natural dehydration
The variations of water content and fracture characteristics were used to study the development of shrinkage fractures in clayey mud with alkaline additive during the natural dehydration test. Figure 12(a), (b), and (c) show the appearance development of dehydration fracture for the solidified mud with the CCR/NS ratio of 6:4 in the containers with the depth of 7cm, 5cm, and 3cm, named 64 − 7, 64 − 5, and 64 − 3, respectively. For the samples of 64 − 7 and 64 − 5, the appearance of cracks can be described as numerous minor cracks distributed among the main cracks. In the later period, the main cracks are scattered on the surface and the minor cracks are squeezed inconspicuously. For the sample of 64 − 3, there are only a few slender cracks on the surface.
On the contrary, the appearance development of dehydration cracks for untreated mud in the containers with the depth of 7cm, 5cm, and 3cm are shown in Fig. 12(d), (e), and (f), named UN-7, UN-5, and UN-3, respectively. It indicates that the development of the cracks on untreated mud has experienced three levels during the cracking stage: main crack, secondary crack, and third-level crack. As the depth of the container decreases, the main and secondary cracks become wider, and the third-level or higher-level cracks become more obvious. The appearance development of the cracks can be portrayed as the process that the main cracks firstly form in the middle of the containers and gradually grow longer and wider, and then the secondary cracks or third-level cracks develop from the middle of the main cracks with an angle of 90°. It is noteworthy that for the untreated mud, all cracks are deep into the bottom of the containers and split the soil into ruptured blocks of different sizes. With the decrease of water content, these ruptured blocks separate from the inner wall of the containers and then shrink significantly.
Figure 13(a) and (b) show the variations of the cracks area ratio of the solidified mud and untreated mud over time. It indicates that the samples have been subject to three stages of cracking: early, accelerated, and stable stage. In early cracking stage, soil surface water is continuously lost and the dehydration cracks are not obvious. In the accelerated cracking stage, the surface cracks increase the contact area between the soil and the air, leading to accelerated water loss, which in turn results in more cracks. In the final stage, the soil almost no longer loses water, and the crack development tends to be stable. As shown in Fig. 13(a), the cracks area ratio of the solidified samples basically ceased to change after 7 days. As the container depth decreases, the final cracks area ratio gradually decreases, and that ratio of 64 − 7, 64 − 5, and 64 − 3 at 9-day is 13.7%, 7.5%, and 4.6%, respectively. For the untreated samples of UN-5 and UN-3, the cracks area ratio increases sharply within 2 to 9 days and then stabilizes at 31% and 36.3%, respectively. Due to the deeper container, the UN-7 sample maintains the supersaturated state for the longest time, starting to crack after 7 days, with the cracks area ratio of 25.1% after 24 days.
Figure 14(a) shows the variation of water content of the solidified mud and untreated mud with different container depths. In addition, Fig. 14(b) and (c) show the water loss ratio of the solidified mud and untreated mud per day at different periods, respectively. The effect of the depth of the container on the water content shows the same rule that during the dehydration process the shallow the soil, the less the water content. The water content of solidified mud with the depth of 7cm, 5cm, and 3cm decreases from 93–51.2%, 41.2%, and 18.5%, respectively. While the water content of UN-7, UN-5, and UN-3 drops from 140–43.7%, 23.6%, and 8.2%, respectively. As shown in Fig. 14(b), the water loss ratio of the solidified mud with a container depth of 3cm is larger than that of 5cm and 7cm. The daily water loss ratio reaches a maximum of 9.3% at 3 to 4 days and is below 1% after 13 days, indicating that the soil water content maintains at a relatively constant value after 13 days of dehydration. Figure 14(c) shows that the water loss ratio of the untreated samples with depths of 5cm and 3cm is higher than that of 7cm during the first 11 days, reaching the maximum value of 18% between 6 and 7 days. In contrast to the samples of UN-5 and UN-3, the UN-7 still has a water loss ratio of 5.4% per day for 16–19 days, with the water loss ratio of 4–8% in the early stage.
In the dry environment, different from the humidity in the standard curing environment, the samples has obvious dehydration and cracking. The evaporation of free water from the upper part of the specimens causes a lower surface water content than the overall water content, which results in the initial crack. Subsequently, the cracks extend in the cracking stage due to deficient matrix suction and negative pore pressure between the soil particles (Yuan et al. 2021). However, after the free water content continuously decreases, the water turns into bound water which is more difficult to volatilize, such that the water loss ratio no longer increases in the shrinkage stage and the cracks morphology is relatively fixed. The effect of alkaline additive on the cracks is mainly expressed in the following aspects. The addition of Ca2+-rich CCR changes the original electrical double-layer structure of the clay layer, and then the clay particles adsorb cations and form the thin layer of bound water film on their surface due to the electrostatic adsorption. In additon, the additive enhances the condensation force between particles to resist the cracking caused by shrinkage and water loss.
In in-situ field applications, the subgrade enhancement starts with the addition of FA and CCR to problematic soils with high water content, followed by the liquid alkaline additive NS. After several days of maintenance, the soil is ready for compaction when it has reached the optimum water content. Although the curing time of the geopolymer on the soils may be longer than that of cement, the FA-based geopolymer treated soils are environmentally friendly materials in terms of energy consumption and waste utilization, with the prospect of replacing conventional cement-cured soils in the subgrade applications.