Introduction

The cemented sand-gravel mixtures are widely used for fast and economical construction of a specific kind of embankment used in hardfill dams and road embankments. The first idea of the hardfill dam was proposed by Londe and Lino [1]. Since the last two decades, many hardfill dams and road embankments have been constructed around world concerning many extensive advantages in comparison with conventional embankments [2,3,4,5,6,7]. While the technology of hardfill dams was established in recent decades, many researchers focused on developing and optimizing new aspects of hardfill dam design [8,9,10,11,12,13,14,15].

In the majority of hardfill embankments, the material of the embankment body is prepared by blending clean coarse-grained soils with cementing agents in the presence of low water content [2]. In some cases, however, the base soil available in the site is comprised of coarse-grained soil with a considerable amount of fine grains (finer than a #200 sieve). The previous studies on cemented soil mixtures focused on the admixture of cement and clean sand or cement and clay and there are few studies on the mixtures in which the base soil in the mixture is comprised of clayey sands or silty sand. Mahasneh and Shawabkeh [16] explored the strength of various admixtures of sand-clay-cement. According to their results, a mixture containing 25% of sand, 50% of cement and 25% of clay owned the maximum strength with comparatively less permeability. Besides, the maximum compressive strength was observed in a sample involving 25% of sand and 75% of cement without clay. Venkatarama Reddy et al. [17] explored the effect of clay particles on the durability, strength and water absorption of soil–cement mixtures and showed that a mixture having clay particles of 14–16% had the greatest strength. Iravanian and Bilsel [18] compared the strength of sand-bentonite admixtures with and without adding the cementing agent and concluded that adding cement to the sand-bentonite mixtures led to an increase in the strength and deformation modulus. Greatest value of strength was observed in a sample involving 75% of sand and 25% of cement without clay. Yongfeng et al. [19] measured the permeability and compressive strength of marine clay-cement mixtures enriched by a polymer. The mixtures contained cement with a weight ratio of 3, 5 and 7%. The test results implied that increasing cement in mixtures led to a decrease in the permeability of mixtures. Besides, the polymer additive increased the failure strain and deformability of mixtures. Yang et al. [20] conducted some triaxial tests on the CSG mixtures comprised of various cement contents and different gradations and contents for aggregates. They concluded that the cohesion of the mixtures increases with cement and aggregate contents. However, the friction angle has been slightly influenced by cement content. Moon et al. [21] measured the compressive strength, ultrasonic pulse velocity (UPV), and shear wave velocity (Vs) of the sand-cement mixtures containing low contents of kaolin particles up to 5%. According to the results of this research, the strength and stiffness of cemented sand mixture increased with increasing in fine content due to improving the density of the mixture with kaolinite particles. Barati et al. [22] studied an admixture of the iron ore tailing enriched with bentonite and cementing agents for stabilization purposes. They found that adding stabilized agents of bentonite or cement to the tailing material increased the optimum water content and decreases the maximum dry density in the compaction test. Besides, the cement had a distinctive role in enhancing the compressive strength of mixtures and, in turn, bentonite caused less increase in the strength of mixtures.

The effect of the cement on the improving the strength and microstructure of stabilized soils and relevant mechanism of cement performance can be explained by exploring the microstructure of hardened soil–cement matrix. After reacting the cement powder with water in the fresh soil–cement mixture, the hydration process is started and some new compounds and hydration products such as initial ettringites, Portlandite (calcium hydroxide or Ca(OH)2), calcium silicate hydration (CSH), and calcium aluminate hydrate (CAH). These the main products of cement hydration are detected in the form needle-shaped bonds of ettringite, honey comb form of CSH, and platy sheets of Portlandite in scanning electron microscopy (SEM) images of soil–cement [23,24,25]. The calcium silicate hydration (CSH) compound is one of the main products of the cement hydration process that is responsible for bonding between grains and strength evolution in the concrete or cement-stabilized soil [26]. Stark [27] summarized all previous researches about the cement hydration and microstructure development. Recently some researchers investigated the microstructure of cement-stabilized soils containing other additives such as fly ash [28], nano CSH gels [29], nano-pozzolan [30], SiO2 nanoparticles [31], and Blast Furnace Slag [25] by implementing SEM–EDX and XRD analyses.

Despite the various studies about the mechanical specifications of cemented sand or cemented clay mixtures, there is a lack of comprehensive investigations about the behavior of CSG mixtures in which the base soil is comprised of clayey or silty sands, and the previous research [21] was conducted on the CSG mixtures containing a very limited fine content (up to 5%) and one type of fine particle. This paper, therefore, aims to investigate the compaction, deformation modulus and strength of CSG mixtures enriched with various contents of bentonite and kaolinite powders in the presence of different cement contents. Besides, the microstructure of mixtures was analyzed to find out how fine-grained additives influence the CSG mixture.

Testing procedure

Material specifications

In this study, a naturally clean sandy soil, designated with the ‘S’ symbol, was used as the native soil in mixtures. The native soil is gravelly sand comprising 65% sand and 35% gravel with the original fine content (passing through a #200 sieve) less than 1%. The S soil is classified as the SW in the unified soil classification system. The particle size distribution and other specifications of the native soil of S are presented in Fig. 1 and Table 1, respectively.

Fig. 1
figure 1

The grains size distribution of native soils (S soil)

Full size image
Table 1 The specifications of native soil (S) and fine additives (kaolinite and bentonite)
Full size table

Two fine-grained additives of kaolinite (denoted by K) and bentonite (denoted by B) were separately blended in the mixture. The bentonite additive is mostly comprised of montmorillonite mineral and is exhibited as a white powder with a high plastic index of about 114%. The specifications of the bentonite additive are presented in Table 1. The plastic limit, liquid limit and plastic index, maximum dry density and optimum water content of the kaolinite additive are presented in Table 1. The maximum dry density and optimum water content of native soil and bentonite and kaolinite additives were measured in accordance with the ASTM D698 standard [32]. The X-ray crystallography (XRD) graphs of bentonite and kaolinite particles are presented in Fig. 2a,b, respectively. The scanning electron microscopy (SEM)image of soil grains is shown in Fig. 3.

Fig. 2
figure 2

XRD graph of fine grained additives used in mixtures: a bentonite powder, b kaolinite powder

Full size image
Fig. 3
figure 3

SEM image of grains of native soil used in the CSG mixture

Full size image

The Ordinary Portland type II cement was used as the cementation agent in cemented mixtures. It is supplied from the north west of Iran and has the Blaine value of 3049 (cm2/gr) and initial and final setting times of 116 and 175 min, respectively. The mineral composition of cement powder is presented in Table 2.

Table 2 The chemical composition (in percent) of cement agent used in mixtures
Full size table

Testing program

Different base soils are prepared by blending the native sandy soil of S with a different amount of kaolinite or bentonite additives. Then, the cementing agent with a content of 5, 10, and 15% is added to each base soil to produce the soil–cement mixtures. The cement content in mixtures is expressed as the ratio of cement weight to the dry weight of base soil (including a summation of native soil and additives). Furthermore, the fine content in the mixture refers to the dry weight ratio of the fine additive to the native soil and is selected as 0, 10, 20 and 30%.

Considering the mentioned ranges of cement and fine additive contents, 21 mixture designations are prepared and tested. All of the test designations are presented in Table 3. Each mixture designation is identified with the symbol of C with a number representing the cement content in percentage, followed by the symbol of B or K with a number that characterize the fine additive type and additive content in the mixture, respectively. For instance, the mixture designation of S-C5-B30 refers to a mixture containing native soil of S, and the bentonite additive is added by a dry weight equal to 30% that of base soil. Then, the cement with a weight equal to 5% of all aggregates is added to the mixture. The mixture with the designation of S-C5 refers to a mixture with a cement content of 5% without any fine additive.

Table 3 The designation of mixture and testing program
Full size table

The maximum dry density and optimum water content of all mixture designations are obtained for the fresh state of the mixture in accordance with the ASTM D558 [33]. For the uniaxial compression test, the test specimens are prepared in the cylindrical mold with a height of 11.64 cm and diameter of 10.16 cm and compacted with optimum water content to achieve the maximum dry density. After curing ages of 7 and 28 days, the uniaxial compression strength and deformation modulus properties are measured for all mixture designations in accordance with the ASTM D1633 standard [34]. SEM images and XRD patterns of some mixtures containing low and high cement contents are captured to interpret the test results.

Sample preparation

Before compaction tests, 5 kg of native soil of S was weighted and aggregates of soil were separated into two coarse and fine fractions based on the passing and retaining on sieve No. 4. The coarse fraction of native soil, which is passed through \(\frac{3}{4}\) inch and retained on No. 4 sieves was cured to achieve the saturated surface dried (SSD) state. The fine fraction of native soil, which is finer than the No. 4 sieve was dried. Then, the predetermined amount of the dried fine additive (kaolinite or bentonite) was supplied to attain the specific weight ratio of fine additive in 5 kg of native soil. The fine additive was added to the fine fraction of native soil, and then the predetermined weight of cement was added to a combination of the fine fraction of native soil and fine additive. The dry mixture was carefully blended to make a high degree of uniformity. Next, water was added to the dry admixture and then a coarse fraction of the native soil in the SSD state was added to the wet admixture to prepare the final state of the mixture. The fresh sand-fine additive-cement mixture with predetermined water content was poured into the compaction mold with a volume of 944 cm3 in three layers. In each layer, 25 blows of a hammer drop with a weight of 2.5 kg and a falling height of 30 cm were applied to the fresh mixture. Then, the dry density of the fresh mixture prepared with at least five water contents was measured and maximum dry density and optimum water content were determined from a density-moisture curve.

The aforementioned procedure was applied for blending the materials to prepare the test specimens for the uniaxial compressive strength (UCS) test, in which two samples were prepared for each mixture designation. The fresh mixture prepared with optimum water content was poured into the mold of the UCS test in three layers and compacted to reach the maximum dry density. After initially hardening the fresh sample inside the test mold, the test sample was detached from the mold. Then, the sample was left inside the humid room and cured under a constant relative humidity of 100% and a temperature of 24 °C. After completing the intended curing age and before the UCS test, the sample was brought out and the top and down surfaces of the test specimen were treated with a thin and smooth layer of capping material. After curing ages of 7 and 28 days, the samples were loaded in the UCS test apparatus. The loading jack is controlled by a stress rate of 120 kPa/s, which falls within the acceptable range of the ASTM standard. As the loading system of UCS apparatus used in this research is stress-control. therefore, the post peak part of the stress–strain curve for CSG mixtures cannot be traced.

At the same time as preparing the specimens for UCS, some cubic specimens with the dimension of 5 cm were prepared with the same compaction condition and cured in the identical condition as the USC test specimen. After the age of 28 days, the microstructure composition of cubic samples is determined by conducting tests of scanning electron microscopy (SEM), an energy dispersive X-ray (EDX) analysis, and X-ray crystallography (XRD). The SEM images of the CSG mixture are prepared in the laboratory of Razi Applied Science Foundation in Iran by using a TESCAN MIRA3 high-performance variable pressure analytical electron microscope with a high resolution of 5.0 nm using a voltage of 1 kV, along with EDX (energy dispersive X-ray) analysis. High-resolution images are produced during SEM analysis at magnification from 5X to 500KX.

Compaction specifications of mixtures

After conducting the compaction test of the fresh mixture, the compaction curve (i.e. dry density versus water content) is measured for all measures. Then, the maximum dry density (MDD) and optimum water content (OWC) are determined from the compaction curve. In order to avoid increasing the paper length, only the typical compaction curve of mixtures without fine content and other mixtures with 5% cement and various bentonite and kaolinite contents are shown in Fig. 4. Detailed data about the compaction specifications of CSG mixtures was presented by Aghajani et al. [35]. Moreover, the MDD and OWC of all CSG mixtures are presented in Table 3.

Fig. 4
figure 4

The dry density-moisture curve of the CSG mixtures: a mixtures with various cement contents and no fine additive, b mixtures containing of 5% cement content in which the base soil containing various bentonite (B) contents: b mixtures containing of 5% cement content in which the base soil containing various kaolinite (K) contents

Full size image

The mixtures involving identical cement content were categorized in one group and the graphs of MDD and OWC for the CSG mixtures enriched with bentonite additive and with the same cement are presented in Fig. 5a, b, respectively. Regarding the high degree of water absorption in bentonite particles, the majority of moisture in the fresh mixture is absorbed with bentonite powder and, thus the OWC increases with an increase in bentonite content. Besides, the bentonite particles are swelled by absorbing moisture. The swelled bentonite particles with less weight occupied more volume spaces in the fresh mixture, and MDD of the mixture is reduced with increasing the bentonite content.

Fig. 5
figure 5

The graph of maximum dry density and optimum water content for the CSG mixtures containing bentonite additive

Full size image

The graphs of the MDD and OWC of mixtures with the same cement content and different kaolinite contents are sketched in Fig. 6. The compaction specifications of mixtures marginally depend on the kaolinite content. In a group of mixtures with the same cement, the greatest OWC occurs in the mixture with 20% kaolinite. However, the difference of OWC between the mixtures with the same cement and various kaolinite contents is limited up to 25%. The kaolinite particles have a lower tendency to absorb water in fresh mixtures and, thus do not considerably affect the OWC of mixtures.

Fig. 6
figure 6

The graph of maximum dry density and optimum water content for the CSG mixtures containing kaolinite additive

Full size image

The MDD of mixtures with kaolinite additive is related to the interaction of kaolinite particles with the void space between the grains of native sandy soil. In mixtures with a cement content of 5%, the greatest MDD occurs in a mixture with a kaolinite content of 10%. The kaolinite considerably fills the void spaces between coarse grains in this mixture and causes the emergence of the best packing of grains inside the mixtures. By increasing the kaolinite percentage in the mixture, the kaolinite particles with less weight, occupy more volume in the mixture, reducing the MDD of the mixture. In the group with a higher cement content of 15%, the mixture without kaolinite has the greatest MDD, which decreased by adding kaolinite to the mixtures.

Results of uniaxial compressive strength and deformation tests

The CSG mixtures without fine additive

The stress–strain curve for the CSG mixtures with cement contents of 5, 10, and 15% and cured at the two ages of 7 and 28 days is presented in Fig. 7. The cement content considerably influenced not only the general shape of the stress–strain curve but also the strength, strength evolution after curing, and failure strain, and deformation modulus of mixtures. The mixture with low cement of 5% shows a ductile behavior accomplishing with lower compressive strength and higher failure strain. By increasing the cement content, the mixture behavior turns to a brittle type and failure occurs at a higher stress level. For comparison, the uniaxial compressive strength (q), failure strain and tangential deformation modulus are measured at the stress level of 0.5q, and the graph of these parameters versus cement content are presented in Fig. 8. By increasing the cement content from 5 to 15%, the uniaxial compressive strength increased from 2.197 to 9.274 MPa with a difference rate of 320% at the age of 28-day. Furthermore, the same increase in cement content causes a considerable rise in the mixture stiffness, and the deformation modulus of the mixture S-C15 increases about 421% in comparison with the mixture of S-C5.

Fig. 7
figure 7

The stress–strain curve for CSG mixtures (without fine additives) and cured at the ages of 7 and 28 days

Full size image
Fig. 8
figure 8

The curves of ultimate stress and ultimate strain versus cement content for CSG mixtures (without fine additives) cured at the ages of 7 and 28 days

Full size image

The development of strength and deformation modulus in mixtures after completing the curing process is entirely related to the cement content in the mixture. By comparing the stress–strain graph of the mixture of S-C5 after two curing ages of 28 and 7 days, it can be seen that the deformation modulus is raised about 68%, with a decrease in the failure strain. However, the uniaxial compressive stress of this mixture at the 28-day curing age is raised only 50% compared to the 7-day cured sample. In contrast, when the cement content increases in the mixture, the raise of uniaxial compressive strength due to the curing completion seems most distinctive in comparison with the raise of deformation modulus. For instance, the difference of compressive strength in mixtures of S-C10 and S-C15 between two curing ages of 28 and 7 days is about 90.5 and 110.8%, respectively.

The failure modes of samples at the curing age of 28 days are shown in Fig. 9. Accordingly, various rupture lines with near-vertical direction and some fragments occur in the mixture of S-C5 at the failure moment. The bonds between the grains are weak and cannot suffer axial loading and many of the bonds between the grains are broken. In the mixture of S-C15, however, while most of the specimen is kept in a stable condition and the bonding between grains are not broken, only two major slip surfaces with inclined alignment emerge in the specimen, which fails under localized shear stress along the slip surface.

Fig. 9
figure 9

The failure mode of CSG mixture samples (without fine grained additive) at the curing age of 28 days: a S-C5 sample, b S-C10 sample, c S-C15 sample

Full size image

The bonding between grains is weak in the mixture with low cement due to the low cement content. It causes the hydration process of cement gel to occur with low intensity, and less strength is added to the mixture after completing the curing. This weak bonding between grains is easily broken under loading. In contrast, strong cementation and bonds occur between grains in the mixture with a higher cement content of 15% and these bonds are progressively completed with time.

The sand-bentonite-cement mixtures

The stress–strain graphs for three groups of mixtures with cement contents of 5, 10, and 15% and containing various amounts of bentonite and cured at the ages of 7 and 28 days are sketched in Fig. 10. The parameters of uniaxial compressive strength, failure strain, tangential deformation modulus for the group of mixtures with identical cement content are shown in Fig. 11.

Fig. 10
figure 10

The stress–strain curve for CSG mixtures containing bentonite additive and cured at both ages of 7 and 28 days: a the family of mixtures with 5% cement, b the family of mixtures with 10% cement, a the family of mixtures with 15% cement

Full size image
Fig. 11
figure 11

The graph of uniaxial compressive strength graph a, failure strain graph b, deformation modulus graph c versus bentonite content for CSG mixtures containing bentonite additive and cured at both ages of 7 and 28 days

Full size image

In a family of mixtures with a cement content of 5%, adding the bentonite additive to the mixture improves the mechanical response of the sample and the deformation modulus and uniaxial strength of all mixtures having bentonite additive are greater than the S-C5 mixture without bentonite. The maximum influence of the bentonite additive is observed in the mixture with 10% bentonite. This mixture exhibits a brittle behavior where the compressive strength and deformation modulus increase about 17% and 94%, respectively, and the failure strain decreases about 41% in comparison with the S-C5 mixture. The bentonite additive restricts the deformation of the mixture and all mixtures involving the bentonite additive are failed under lower failure strain than the S-C5 mixture.

The bentonite additive influences the failure modes of mixtures with a low cement content of 5% (Fig. 12). Even though the S-C5 specimen is fragmented at the failure moment and most of the bonding between the grains is broken (Fig. 9a), all specimens containing the bentonite additive are failed with emerging a slip surface while the other parts of the specimen are in an intact and unbroken state.

Fig. 12
figure 12

The failure mode of CSG mixtures with 5% cement, cured at age 28 day and containing: a 10% bentonite (S-C5-B10 sample), b 20% bentonite (S-C5-B20 sample), c 3 0% bentonite (S-C5-B30 sample)

Full size image

Increasing the cement content in the mixtures changes the role of bentonite in the mechanical response of mixtures, and adding the bentonite to the mixtures containing higher cement reduces the strength and deformation modulus. In the family of mixtures with 15% cement, for instance, the compressive strength of mixtures containing 10, 20 and 30% of bentonite additives decreases by about 29, 53.6, and 59% in comparison with the S-C15 mixture without bentonite. Moreover, the reduction of deformation modulus of S-C15-B20 and S-C15-B30 mixtures is about 5 and 47% respectively, and the deformation modulus of the mixture with 10% bentonite (S-C15-B10) increases about 31%. However, similar to the family of mixtures with low cement, all mixtures containing bentonite additive and high cement are failed under lower failure strain. With increasing the bentonite content, the trend of a gradual decrease in the mixture compressive strength is observed in the group of mixtures with 10% cement. In this group of mixtures, the mixture of S-C10-B10 with 10% bentonite has a higher stiffness, and the deformation moduli of S-C10-B20 and S-C10-B30 mixtures are lower than the S-C10 mixture with no bentonite.

The bentonite additive influences the failure mode of mixtures with 15% cement where the failure images are shown in Fig. 13. In contrast to the mixture group with 5% cement, adding bentonite to the mixture group with 15% cement imposes weakness in the sample and reduces the integrity of the specimen at the failure moment. As shown in Fig. 13c, various cracks with limited length along several directions emerge in the S-C15-B30 mixture at the failure moment. These cracks are gradually developed in the specimen when the stress level exceeds 0.5q and causes a reduction in the stiffness of the sample. As shown in Fig. 9c, however the S-C15 mixture with no bentonite is ruptured by imposing a limited number of the slip surface with no fine cracks or instability in all parts of the specimen.

Fig. 13
figure 13

The failure mode of CSG mixtures with 15% cement, cured at age 28 day and containing: a 10% bentonite (S-C15-B10 sample), b 20% bentonite (S-C15-B20 sample), c 3 0% bentonite (S-C15-B30 sample)

Full size image

One of the key parameters for exploring the role of cement in the mixture is the investigation of strength and stiffness evolution between two curing ages. In mixtures with identical cement of 5% and different bentonite contents, the increasing rate of the compressive strength between two curing ages of 7 and 28 days varied between 50 and 72%. The utmost influential of the curing process is found in deformation specifications in such a way that all mixtures are failed at a relatively lower strain by completing the hydration process of cement agent after 28 days. Besides, the deformation modulus of all mixtures with 5% cement increases in the range between 35 and 56% after the curing age of 28 days.

In mixtures with 15% cement, the bentonite additive changes the evolution rate of the strength and deformation parameters due to curing completion. By completing the hydration process after 28-day, the strength and stiffness of the S-C15 mixture with no bentonite are considerably enhanced, the compressive strength increases from 4.4 to 9.27 MPa, and the deformation modulus is enhanced from 4 to 7.29 MPa. The increase rate of strength in S-C10 mixture between two curing ages of 7 and 28 days is about 90.5%. This means that the strength of these mixtures is almost doubled after the curing age of 28 days. However, when the bentonite is added to the CSG mixtures with higher cement content, the evolution rate of strength between two curing ages reduced. For instance, the percentage of strength increase after completion of the curing in mixtures with 15% cement and the bentonite additive is limited to 50%. Besides, the increasing rate of strength after curing age in the mixtures with 10% cement and various bentonite content is limited to 45%. Nevertheless, all mixtures having bentonite are failed with lower strain after the 28-day age than the 7-day cured specimen.

It is interesting to note that when a mixture is enriched with 10% bentonite regardless of the cement content, the mixture inherently has higher deformation modulus at both curing ages and the lowest increase in deformation modulus with curing completion occurs in a mixture with 10% bentonite among the mixtures with the same cement content. This observation implies that the mixture with 10% bentonite has strong packing in base soil which leads to a low tendency to deformation in the mixture.

Based on the mentioned results, it can be argued that the role of the bentonite in the strength and deformability of mixtures depends on the cement content. As mentioned previously, a weak bonding is developed between grains in the mixture with low cement content due to an insufficient cementing agent in the mixture. When bentonite is added to this mixture, the bentonite particles help to fill the void spaces between coarse grains of native soil and impose a strong texture in the mixture, besides the bentonite particles have low interference with the hydration process of cement. Nevertheless, the bentonite agents increase the bonding between grains by imposing an extra cohesion force. Thus, the mixtures with low cement containing the bentonite additive tend to have less deformation and more integrity at the failure moment.

In contrast to the aforementioned process, the strong bonding between grains inherently exists in the mixture with a high cement content of 15% and the cementing agent governs the behavior of the mixture. When the bentonite additive is added to this mixture, the bentonite absorbs the moisture in a fresh state of the mixture and has a disruptive effect on the hydration process of cement. This weakens the cementation bonding between grains, which is reflected in the failure mode. Furthermore, the progress of strength and stiffness development due to curing completion is disrupted by the bentonite. Thus, the mixture with high cement containing bentonite gains less strength value rather than the pure sand-cement mixture.

The sand- kaolinite-cement mixtures

The stress–strain graph of the CSG mixtures enriched by the kaolinite additive is illustrated in Fig. 14. The graphs of uniaxial compressive strength, deformation modulus and failure strain parameters for the group of mixtures with identical cement content and different kaolinite percentage are presented in Fig. 15.

Fig. 14
figure 14

The stress–strain curve for CSG mixtures containing kaolinite additive and cured at both ages of 7 and 28 days: a the family of mixtures with 5% cement, b the family of mixtures with 10% cement, a the family of mixtures with 15% cement

Full size image
Fig. 15
figure 15

The uniaxial compressive strength graph a, failure strain graph b, deformation modulus graph c versus kaolinite content graphs for CSG mixtures containing kaolinite additive and cured at both ages of 7 and 28 days

Full size image

The performance of the kaolinite additive in the mixtures with low cement is similar to the role of the bentonite additive and adding any amount of the kaolinite to the mixture with 5% cement improves both the compressive strength and deformation modulus and reduces the failure strain. However, the influence intensity of the kaolinite additive is considerably distinguished relative to the bentonite additive. For instance, even though the uniaxial compressive strength of the S-C5 mixture is measured at about 2.197 MPa at the curing age of 28 days, the uniaxial strength of mixtures of S-C5-K10, S-C5-K20 and S-C5-K30 mixtures increases to 3.8, 3.19 and 2.57 MPa, respectively. Furthermore, the deformation modulus of these mixtures is measured at 7.15, 4.82 and 4.35, respectively. The deformation modulus of all mixtures with the kaolinite additive is considerably greater than that of S-C5 mixture with a value of 1.55 MPa. Moreover, similar to the family of mixtures with 5% cement containing bentonite, adding kaolinite to the mixture with 5% cement reduces the failure strain and the maximum decrease of failure strain occurs about 49% in the S-C5-K10 mixture.

The kaolinite additive influences the failure mode of mixtures with 5% cement which is shown in Fig. 16. The S-C5-K10 mixture is failed by imposing an inclined slip surface with a sharp slope where most parts of the specimen remain stable with no major crack. The localized slip surface is observed in other mixtures enriched with further kaolinite with the exception that a limited number of narrow cracks are observed in these mixtures. On the other hand, the S-C5 mixture with no kaolinite is failed by the fragmentation of specimen and imposing many wide cracks (Fig. 9a).

Fig. 16
figure 16

The failure mode of CSG mixtures with 5% cement, cured at age 28 day and containing: a 10% kaolinite (S-C5-K10 sample), b 20% kaolinite (S-C5-K20 sample), c 30% kaolinite (S-C5-K30 sample)

Full size image

The stress–strain curves of mixtures with 5% cement and different kaolinite content at both curing ages of 7- and 28 days are presented in Fig. 14a. The increasing rate of uniaxial strength and deformation modulus between the two curing ages in all mixtures varies between 50 and 80%. However, while the reduction rate of failure strain is measured about 25% in the S-C5 mixture after completing the curing age, the mixtures with 10% kaolinite additive show similar failure strain in both curing ages of 7 and 28 days. Also, the decreasing rate of failure strain in the mixtures with 20% and 30% kaolinite is less than 11% after curing completion.

When the cement content increases in the mixture, the kaolinite additive keeps a positive role in enhancing the strength and stiffness of the mixture. By adding 10% kaolinite to the S-C15 mixture, the uniaxial strength is increased from 9.27 to 11.27 MPa with an increase rate of 21%. However, the deformation modulus of the S-C15-K10 mixture becomes close to the S-C15 mixture. Then, increasing the kaolinite content to 20% reduces the increasing rate of mixture strength, and the uniaxial strength of the S-C15-K20 mixture raised only 6.9% greater than the S-C15 mixture. Finally, adding further kaolinite to the mixture has a negative influence on both strength and stiffness specifications in such a way that the uniaxial strength of the S-C15-K30 mixture is reduced by about 44% in comparison with the S-C15 mixture. It should be noted that all mixtures with 15% cement containing kaolinite are failed under lower failure strain relative to the mixture with no kaolinite additive. A similar trend is observed in the mixture group with 10% cement and the utmost influence of kaolinite additive on the strength and stiffens is found in the mixture with 10% kaolinite. Then, adding further kaolinite to the mixture drops the uniaxial strength and deformation modulus to levels lower than the mixture with no kaolinite additive.

The failure modes of the mixture group with 15% cement and various kaolinite contents are shown in Fig. 17. The mixture with a higher kaolinite content of 30% is failed by imposing numerous cracks with limited length and wide openings and the specimen is divided into smaller pieces. In contrast, the mixture with 10% kaolinite is failed in a different mode and an aligned slip surface with a sharp angle emerged in the sample while the bonding between grains in the rest of the sample are kept in a stable and unbroken condition.

Fig. 17
figure 17

The failure mode of CSG mixtures with 15% cement, cured at age 28 day and containing: a 10% kaolinite (S-C15-K10 sample), b 20% kaolinite (S-C15-K20 sample), c 30% kaolinite (S-C15-K30 sample)

Full size image

The stress–strain graph of mixtures with 15% cement and two curing ages of 7 and 28 days are compared in Fig. 14c. The rate of increase in uniaxial strength between 7-day and 28-day ages varied between 91 and 110% for S-C15-K30 and S-C15 mixtures, respectively. Besides, the increasing rate of deformation modulus with the curing process is close in the mixtures with 0, 20, and 30% kaolinite. However, the S-C15-K10 mixture has the lowest increasing rate of deformation modulus after the hydration completion and the mixture initially has a higher deformation modulus at the earlier curing age of 7 days. This result was previously observed in the mixtures with 10% bentonite where the 10% fine additive imposes the best packing in base soil of mixtures and makes a mixture with a strong matrix.

The similar rate of stiffness and strength evolution with the curing process in mixtures with and without the kaolinite additive implies that the kaolinite has no negative and disrupted effect on the cement hydration process. Nevertheless, kaolinite fills the void spaces between grains in base soil, but the intensity of the kaolinite additive effect on the uniaxial strength and deformation modulus of the mixture depends on the cement content of the mixture. In the mixtures with low cement content, the initial bonding between grains is comparatively weak, and adding any amount of kaolinite improves the texture and enhances the strength of mixtures. On the other hand, the strength of mixtures with higher cement content principally depends on the cementing agent, and by adding an extra kaolinite content greater than 20% increases the specific surface of grains in the mixture. Thus, the present cementing agent is not sufficient to cover all aggregates of base soil in the mixture and the emerged bonding between grains is weak in comparison with the mixture having no kaolinite additive. The best performance for enhancing the strength and stiffness of cemented mixtures is found in the mixture with 10% kaolinite additive in all cement contents.

Microstructure analyses of the mixtures by SEM images

The sand-cement mixture with 5% cement

By focusing on the result of strength tests, it can be seen that the strength and stiffness of cemented mixtures are related to the interaction between the three parameters of cement content, additive type, and additive content in the mixture. In mixtures with a low cement content, both additive types have a positive impact on enhancing the strength and stiffness of mixtures. In mixtures with a higher cement content, the trend of strength and deformation modulus depends on the additive type and content. For interpreting the consequences, the microstructures of sand-cement mixtures with 5% and 15% cement enriched with both additives are thoroughly investigated by analyzing the SEM images and XRD pattern of mixtures. The SEM images of mixtures with 5 and 15% cement are shown in Figs. 18–21. Besides, the percentages of basic elements that existed in the bulk of mixtures are detected by EDX analysis of SEM images and presented in Table 4.

Fig. 18
figure 18

The SEM images of mixtures with cement of 5% and different fine additives under 150 × magnification: a S-C5 mixture, b) S-C5-B10 mixture, c S-C5-K10 mixture, d S-C5-B30 mixture, e S-C5-K30 mixture

Full size image
Table 4 The quantitative results of EDX-SEM analysis including the weight percentage of basic elements in the bulk of mixture
Full size table

In the S-C5 mixture (Fig. 18a), the cement paste covers the coarse grains of native soil and fills the void spaces between grains. In a closer view of SEM image with the scale of 5 μm and 5000 × magnification (Fig. 19a), the needle-shaped cement paste emerges between grains of sand soil which causes grain bonding. These needle-shaped bonds are referred to ettringite and calcium hydroxide crystals which are the main products of cement hydration in concrete and soil–cement mixtures [23,24,25].

Fig. 19
figure 19

The SEM images of mixtures with cement of 15% and different fine additives under 5000 × magnification: a S-C15 mixture, b S-C15-B10 mixture, c) S-C15-K10 mixture, d S-C15-B30 mixture, e) S-C15-K30 mixture

Full size image

The sand-bentonite-cement mixtures with 5% cement

In the S-C5-B10 mixture (Figs. 18b, 19b), the bentonite minerals are exhibited in the form of swelled plates that are extended into the void space between sand aggregates. However, the intensity of needle-shaped crystals decreases in the cement paste and bentonite reduces the impact of the cement hydration in the mixture. The influence of bentonite plates on the cement hydration changes the participant of basic elements in the mixture. The participant ratio of calcium to silicon elements in the cemented admixture (denoted by the Ca/Si ratio) accounts for an indication for the hydration of cement and the ratio of Ca/Si increases by completing and intensifying the cement hydration in the mixture [, , 36,37,38]. From the EDX analysis, the ratio of Ca/Si in the S-C5-B10 mixture is reduced to 0.639 and, in turn, an increase in the participant of Na and Mg elements occurs. Therefore, the hydration process of cement is weakened by presenting the bentonite particles. The phenomenon of a low Ca/Si ratio due to montmorillonite interaction with cement hydration in cemented materials was previously reported by Fernández et al. [36].

As shown in Figs. 18d, 19d, the majority of the S-C5-B30 mixture is occupied by the bentonite swelled plates and the void space in the mixture is efficiently filled by these platy particles. However, the considerable decreased ratio of Ca/Si and reduced intensity of the needle-shaped cement paste in SEM images suggests that the higher percentage of bentonite powders disrupts the cement hydration in the mixture. This issue is reflected in the strength and failure mode of the mixture as discussed previously.

The sand- kaolinite -cement mixtures with 5% cement

In the S-C5-K10 mixture (Fig. 18c), the kaolinite powders are observed between coarse grains of sand soil and fill the void space of the specimen. Furthermore, the needle-shaped crystals of the cement paste are found around both kaolinite powders and sand grains and result in the adhesion between sand and kaolinite. Furthermore, some needle crystals of cement paste emerged throughout kaolinite powders which implies that kaolinite powders participate in the hydration process of cement (Fig. 19c). From the EDX analysis, the ratio of Ca/Si in the mixtures of S-C5 and S-C5-K10 is measured at 1.139 and 1.221, respectively, implying that the kaolinite slightly increases the calcium product in the mixture and, thus, has a marginally positive influence on the cement hydration.

The negative influence of an extra amount of the fine additive on the cementation process is observed in the S-C5-K30 mixture with a considerably decreased ratio of Ca/Si. However, kaolinite additive is found in the form of a powder (Figs. 18e, 19e) that covers all coarse grains of the native sandy soil in the mixture. Regarding the increase in the specific surface of base soil due to the presence of extra kaolinite powders in this mixture, the cement is not sufficient to cover the base soil grains, and thereby reducing the cement paste concentration.

The sand-cement mixture with 15% cement

As the cement content increases in the S-C15 mixture, a higher intensity of the needle shape of cement paste is found in the SEM image (Figs. 20a, 21a). Moreover, this mixture has the highest participant of calcium, and consequently, the highest Ca/Si ratio is observed among the mixtures with the same cement content.

Fig. 20
figure 20

The SEM images of mixtures with cement of 5% and different fine additives under 5000 × magnification: a S-C5 mixture, b S-C5-B10 mixture, c) S-C5-K10 mixture, d S-C5-B30 mixture, e S-C5-K30 mixture

Full size image
Fig. 21
figure 21

The SEM images of mixtures with cement of 15% and different fine additives under 150 × magnification: a S-C15 mixture, b S-C15-B10 mixture, c S-C15-K10 mixture, d S-C15-B30 mixture, e) S-C15-K30 mixture

Full size image

The sand-kaolinite-cement mixtures with 15% cement

The fine powders of the kaolinite fill the void space in the S-C15-K10 mixture (Figs. 20c, 21c) and improve the structure and texture of the base soil in the mixture. Besides, the contribution of kaolinite minerals in the hydration process of cement causes the emergence of some new needle-shape crystals of cement paste from the kaolinite particle. Hence, some bonds are developed between coarse grains of sand and fine powders of kaolinite, leading to an emerging a strong structure in mixture. Even though this issue does not influence the Ca/Si ratio in the S-C15-K10 mixture, some new bonds are detected in the XRD graph of the mixture which will be discussed later. Similar to the S-C5-K30 mixture, the current cement content is not sufficient for the further kaolinite content in the S-C15-K30 mixture and, thus, most of the kaolinite powders remain in the original state and are not covered with the cement paste in the mixture (Figs. 20e, 21e). This causes a reduction in the Ca/Si ratio for this mixture in comparison with the S-C15 mixture.

The sand- bentonite-cement mixtures with 15% cement

The expanded platy shape of bentonite particles is largely founded in the S-C15-B10 mixture (Figs. 20b, 21b). These bentonite particles disrupt the hydration of cement, leading to a considerable decrease in the Ca/Si ratio and the intensity of the needle-shaped cement past decreases in comparison with the S-C15 mixture. In mixtures with 30% bentonite, the mixture has a less open void spaces, the majority of which is blocked by the swelled plates of bentonite. However, less needle-shaped cement paste is observed in this mixture, which leads to the reduction of cementation effect and, thus, the strength of the mixture decreases in comparison with the S-C15 mixture.

XRD Analyses of the mixtures

The XRD patterns of mixtures with 5% cement containing both additives types are illustrated in Figs. 22, 23, those of the mixtures with 15% cement are presented in Figs. 24, 25. The various phases in the XRD pattern of mixtures are identified by implementing the X’Pert Highscore Plus software [39], and the detected phases are marked on the XRD graph. The peaks that appear on the XRD graph are related to the bonds in the mixture. The calcium silicate hydration (CSH) and quartz are particularly characterized in the mixture with 5% cement. The number of CSH peaks increases in the S-C5-K10 mixture with 10% kaolinite. Furthermore, some new peaks recognized as quartz emerge in this mixture. These new peaks reveal that the kaolinite particle corporates in the hydration process and the kaolinite minerals are transformed into the cement hydration products such as quartz and CSH. This is confirmed by the SEM and EDX analyses of the S-C5-K10 mixture where the percentage of calcium in the mixture is higher than the S-C5 mixture with no kaolinite. The participant of the kaolinite minerals in the hydration process of cement and turning the kaolinite to CSH and other cement gel products were previously reported by Tabet et al. [40]. Moreover, Chrysochoou [41] observed the dissolution of the kaolinite mineral in stabilized soil with cement. The released kaolinite particles corporate in the cement hydration which leads to an increase in CSH compounds and improves the strength of stabilized soil.

Fig. 22
figure 22

The XRD graph of mixtures with 5% cement enriched with kaolinite additive (Q: quartz, CSH: calcium silicate hydration, P: portandite, E: ettrinigite, K: kaolinite)

Full size image
Fig. 23
figure 23

The XRD graph of mixtures with 5% cement enriched with bentonite additive (Q: quartz, CSH: calcium silicate hydration, P: portandite, E: ettrinigite, K: kaolinite)

Full size image
Fig. 24
figure 24

The XRD graph of mixtures with 15% cement and enriched with kaolinite additive (Q: quartz, CSH: calcium silicate hydration, M: montmorillonite)

Full size image
Fig. 25
figure 25

The XRD graph of mixtures with 15% cement enriched with bentonite additive (Q: quartz, CSH: calcium silicate hydration, M: montmorillonite)

Full size image

In the XRD graph of the S-C5-K15 mixture, three particular peaks of kaolinite mineral recognized at the reflected angles of 2Q = 12.6, 25.04, and 36.2° are exactly coincident with those in the XRD graph of kaolinite powder shown in Fig. 2b. Thus, the participant of kaolinite particles declines in the hydration of cement because of the extra amount of kaolinite additive and insufficient cement in the S-C5-K15 mixture and the particles remain in their original state. This issue is observed not only in the SEM images but also in the XRD graph of this mixture.

By increasing the cement content in the S-C15 mixture, more peaks of quartz are detected in the XRD graph, revealing the strong bonds created by the cement hydration. However, the number of peaks increased by adding 10% kaolinite to the mixture and new CSH peaks are detected in the mixture in comparison with the S-C15 mixture. Based on the EDX analysis, the calcium content and the Ca/Si ratio of the S-C15 are greater than the S-C15-K10, which means that the cement hydration mainly occurs in the S-C15 mixture. However, the additional number of peaks in the S-C15-K10 reveals that kaolinite mineral reacts with cement and turns into new cemented materials. Thus, the formation of new bonds in the mixture with 10% kaolinite together with dense packing of grains in matrix observed in SEM images leads to developing a mixture with high strength and deformation modulus.

In contrast to mixtures containing kaolinite, the number of peaks in XRD graph decreased in mixtures with 15% cement containing bentonite and this issue is more distinctive in the S-C15-B30 mixture. This means that bentonite decreases the effect of cement hydration and the consequent cementation bonding becomes weak in this mixture. Therefore, the mixture with a higher bentonite content has the lowest strength. Besides, some peaks that are recognized as montmorillonite mineral are detected in the S-C15-B30 mixture at the reflecting angles of 2Ѳ = 19.9 and 22.12°.

Conclusions

The uniaxial compression strength, deformation modulus, failure strain and microstructure of CSG mixtures containing fine-grained additives were investigated in the current research, along with analyzing the mineralogy and SEM images of the mixture were analyzed. The following consequents were obtained from the test results:

  • The maximum dry density of CSG mixtures, regardless of fine-grained additives, does not considerably depend on the cement content. However, the cement content increases the optimum water content in the mixture.

  • The strength and deformation modulus of cemented mixtures are related to the interaction between the three parameters of cement content, additive type, and additive content in the mixture.

  • In mixtures with a low cement content of 5%, both types of fine-grained additives have a positive impact on enhancing the strength and stiffness of mixtures. The mixture enriched with 10% of the fine additive has the greatest compressive strength, the highest deformation modulus, and the lowest failure strain among the mixtures enriched with a similar type of fine additives.

  • In mixtures containing a higher cement content, the trend of strength and deformation modulus parameters of samples depends on the additive type.

  • In the family of mixtures with 15% cement enriched by bentonite, adding bentonite to the CSG mixture reduces the strength. However, all samples with bentonite additive are failed under a lower stain in comparison with the mixtures without an fine additive. Besides, the mixture enriched with 10% bentonite has the highest deformation modulus among the other mixtures due to acting the best packing of grains in base soil.

  • Adding 10% of kaolinite to the CSG mixture with 15% cement considerably increases the strength and deformation modulus, and then the strength slightly decreases by a further increase of kaolinite in the mixture.

  • The analysis the mineralogy and SEM images confirms the aforementioned trend of compressive strength and deformation modulus of mixtures. When the CSG mixture contains 10% of the additive, the structure of base soil in the matrix attains the best packing and a strong matrix is devolved inside the mixture. Furthermore, the kaolinite particles in the mixture containing 10% of the kaolinite additive contribute to the hydration process of cement and further bonds are established inside the matrix of the mixture. Therefore, the mixture with 10% of kaolinite results in the best performance in terms of strength and stiffness.

  • By adding the fine additive greater than 10% to the CSG mixture, the cement gel is not sufficient to cover all grains of base soil, and the hydration product intensity, resulting in decreased strength.

  • The bentonite particles have a disrupting effect on the hydration process of cement, especially in mixtures containing higher a cement content. This reduces the calcium percentage and the ratio of calcium to silicon elements in the mixture, deteriorating the bonding in the mixture.

  • It can be concluded that if the higher strength and low deformability are intended, the fine additive content in the mixture should be set to 10%. The maximum efficiency for strength and integrity of the mixture is attended by adding 10% kaolinite additive.