Introduction

Freezing–thawing (F-T) and heaving are significant factors causing soil volume change in cold regions. These factors have destructive effects on many engineering applications such as pipelines, highways, embankments, or water conveyance channels (Lu et al., 2019). In cold regions, at least one freezing and thawing cycle occur in a year (Kalkan, 2009; Yarbaşı et al., 2007). The F-T cycles have effects on soil based on physical, chemical, and biological aspects (Liu et al., 2021). The primary reason for the negative effect of the F-T on the physical and mechanical properties of soil is gradually enlarging void spaces of soil due to frozen moisture in the voids. Besides, the F-T cycles lead to fragmentation of coarse particles and aggregation of fine soil particles (Zhang et al., 2016). In cold regions, engineering behaviors of soil are the most affected parameters by freezing and thawing cycles. There is a wide range of studies examining the effect of F-T on various soil engineering behavior such as density (Konrad, 1989), dynamic properties (Mahya et al., 2021), hydraulic permeability (Viklander, 1998; Yıldız and Soğancı, 2012), stress–strain relation (Eskişar et al., 2015; Lu et al., 2020; Sahlabadi et al. 2021), mechanical properties of soils (Hohmann-Porebska, 2002; Liu et al., 2019; ShahriarKian et al., 2021; Jamshidvand et al. 2021), and dynamic characteristics (Cui et al., 2014; Wang et al., 2015).

Generally, the substantial effects of F-T on soil properties are seen in shallow soil layers due to the low thermal conductivity of the soil. Thus, in most studies, it was recommended to overcome the negative effects of F-T cycles using shallow stabilization methods. Some of the widely used shallow soil stabilization methods based on stabilizers can be categorized as conventional additives (Hotineanu et al., 2015; Saadat and Bayat, 2019; Wu et al., 2020), geofibers and synthetic fluids (Zaimoglu, 2010; Tiwari et al., 2020; Liu et al., 2021), non-conventional chemical additives (Güllü, 2015; Uzer, 2016; Yılmaz and Fidan, 2018; Taş et al. 2018), alkali activated materials (Amulya et al., 2020; Karozou et al., 2020; Syed et al., 2020), and recycled materials (Aydin et al., 2020; Jafari and Esna-ashari, 2012; Yilmaz et al., 2015).

Zeolite minerals which have a wide range of uses are produced by several countries in the world (Flanagan, 2016). In some studies, it was suggested to use zeolite minerals as a soil stabilization agent (Mola-Abasi and Shooshpasha, 2016; Savas, 2016; Eyo et al., 2020; ShahriarKian et al., 2021; Chenarboni et al., 2021). Bayburt province is one of the places where zeolite-rich tuffs are found in Turkey. The zeolite-rich tuffs are commonly used as restoration or structural materials (Cicerali et al., 2020). Yılmaz et al. (2015) stated that large amounts of waste material are generated during cutting operations, which could be used for soil stabilization.

From the early 1980s, the CT scanning has been used in the civil engineering discipline. As the CT enables to examine 3D structure of soils and is a rapid and non-destructive method, CT scanning to examine soil structure is gradually increasing in geotechnical engineering (Peth et al., 2008; Liu et al., 2021; Xu et al., 2021). CT scanning is used to examine pore morphology (Han et al., 2020), pore size distribution (Akhondzadeh et al., 2020), and micro–macro cracks in soil media (Wang et al., 2020). The pore structure of the soil is a primarily examined factor in geotechnical studies concerning the freezing–thawing effect. Therefore, X-ray and CT scanners are functional to observe and quantify pore characteristics (Liu et al., 2021). The HU is a dimensionless unit universally used in CT scanning to express CT numbers in a standardized and convenient form. As the HU value is a distinctive parameter in CT scanning, materials can be specified based on the HU values.

Despite chemical and mineralogical structure of the waste zeolites being similar to widely used waste materials in stabilization applications such as fly ash and blast-furnace slug, few study was examined durability performance of the tuffs considering the inner structure. Zeolite-rich tuffs can be mined in various parts of the world, such as Turkey, Greece, Italy, Yemen, Iran, the USA, and Japan (Kitsopoulos et al.,1996; Pham et al., 2019; Rajabi and Ardakani, 2020). The studies showed that there are little differences in oxide components and mineralogical structures between these tuffs (Eyo et al., 2020; ShahriarKian et al., 2021). The effect of these differences on the mechanical and durability performances of the stabilized samples is a gap in the literature. Thus, three different of subtypes of the tuffite formation was taken into account to examine the effect of differences of tuff structure on the stabilization performance. Ultrasonic pulse velocity (UPV) is one of the widely used nondestructive tests methods. However, UPV values are prone to inconsistencies originating from the experimental process. In a wide range of studies about soil stabilization assessments, CT scanning is used to examine pore structure. On the other hand, there are very few studies on assessing the usability of the HU values as a non-destructive strength assessment method. This study aimed to investigate the effect of waste zeolites on swelling pressure, mechanical properties, and durability of stabilized clayey soil subjected to freezing–thawing cycles. Within this scope, SEM, CT scanning, and image processing methods were used to examine the micro and macro effects of the waste zeolites. The specific objectives of this paper were to (1) investigate statistical significance between unconfined compressive test and HU values of stabilized soils and (2) evaluate the effects of freeze–thaw cycles on HU values and unconfined compressive strengths of the samples.

Material

Natural soil

Clayey soil was taken at a depth of 2 m from the trial pit excavated at the Demirözü district of Bayburt city (Turkey). Some of the geotechnical properties of the soil are presented in Table 1.

Table 1 Physical properties of the soil
Full size table

Grain size distribution, compaction parameters, specific gravity, and Atterberg limits of the soil were determined considering ASTM D 422, ASTM D698, ASTM D 854, and ASTM D 4318, respectively. The grain size distribution of the natural soil is presented in Fig. 1a. The organic matter content was determined as 5% (ASTM D 2974). It was seen that the soil is suitable for lime stabilization applications based on the organic matter content. X-ray diffraction (XRD) analysis was performed for natural soil using Rigaku/StamartLab X-Ray diffractometer (ASTM D 5381). The estimated mineralogical composition of the soil is shown in Fig. 1b. Based on the estimation, it is seen that the main mineralogical components of the soil are montmorillonite (M), quartz (Q), and calcite (C).

Fig. 1
figure 1

a Grain size distribution of the natural soil. b X-ray diffraction pattern of the natural soil

Full size image

Lime and waste zeolites

In the context of this study, natural soil was stabilized with lime and three types of natural tuff mixtures. Commercially available hydrated lime used in the stabilization was obtained from Karsan Lime Company. X-ray fluorescence (XRF) analyses were performed to determine the oxide composition of the lime and waste zeolites considering ASTM D 934. There are three types of tuffs locally known as Bayburt Stone (BS). The zeolite-rich tuffs are commonly used as restoration or structural materials (Cicerali et al., 2020). During the mining and cutting process, a large amount of waste material occurs. As the average rejection rate is over 60%, disposing of the waste material is required to prohibit environmental pollution. Some of the physical properties and chemical components of lime and the tuffs are given in Table 2. The siliceous nature of the tuffs can be seen from the table, and it is also seen that SiO2, Al2O3, and Fe2O3 oxides contain more than 80% of the total compound. The tuffs do not contain harmful chemical matters such as nitrate, sulfate, radioactive elements, or heavy metals. Therefore, it can be qualified as a harmless soil stabilization agent.

Table 2 Oxide composition and physical properties of lime and the tuffs
Full size table

Color is the most notable distinctive feature of the tuff types. In Table 2, green (GTW), white (WTW), and yellow waste zeolites (YTW) are denoted as Tuff-1, Tuff-2, and Tuff-3, respectively. Within the scope of the study, waste zeolites shown in Fig. 2a were grinded, and the effect of powdered waste zeolites (Fig. 2a) on mechanical and durability properties of stabilized soil was investigated.

Fig. 2
figure 2

Tuff wastes used in the study. a Raw tuff wastes. b Powdered tuff wastes

Full size image

As one of the major components is zeolitic materials, the tuffs contain a high silica structure. XRD patterns of the powdered waste zeolites are shown in Fig. 3. The chief mineralogical components of the tuffs are quartz (Q), feldspar (F), and periclase (P) for GWT and WWT (Fig. 3a and b), quartz (Q), and feldspar (F) for YWT (Fig. 3c).

Fig. 3
figure 3

X-ray diffraction patterns of a green waste zeolites (GWS), b white waste zeolite (WWT), c yellow waste zeolite (YWT)

Full size image

Sample preparation

Soil-tuff and soil-tuff-lime mixtures were prepared considering optimum water content. Pulverized tuff was used in the stabilization applications. The pulverization of the tuff wastes was done with stone crushing and grinding machines. The grinding process was continued until 90% of the grains were under 100-micron material. Stabilization studies were started after obtaining 90% of the material below 100 microns with the granulometry analysis. The pulverized tuff and optimum amount of lime were mixed in the Hobart mixer for 5 min. The pH value of the samples was taken into account to determine the optimum amount of lime. Suspensions of the mixtures were prepared by adding lime to the natural soil at an increasing rate of 1% between 0 and 10%, in the test carried out at a pH measurement temperature of 21.7 °C. The optimum lime ratio criterion was taken as 12.4, the pH value of the lime and water solution. The minimum lime ratio value corresponding to this value in natural sample and lime mixture solutions was determined as the optimum lime ratio of the soil (6%). The prepared mixture was mixed with soil for 5 min. The dry matters were mixed with a predetermined amount of water and mixed for 5 min to sustain homogeneity.

The mixtures were compacted into a standard compaction mold with standard compaction energy. Samples used for various tests are taken from the mold using special thin wall samplers. Cylindrical samples of 76 mm length and 38 mm diameter were used for unconfined compressive strength, durability tests, and CT imaging. The 20 mm length and 75 mm diameter specimens were used for swelling pressure and one-dimensional swelling percentage tests. The samples were wrapped with plastic film and cured for 28 days (unconfined compressive strength, durability tests, and CT imaging) or 14 days (swelling pressure and one-dimensional swelling percentage tests) under constant humidity (97% ± 2 °C) and temperature (20 °C ± 3 °C).

Method

At the first stage of the study, optimum lime content for stabilizing the soil was determined as 6%, considering ASTM D 6276. The waste zeolites were grinded, and percentages of tuffs by dry weight of soil were determined. In the study, soil and lime were represented with S and L. In the definition of the tuffs, color (G: green, W: white, and Y: yellow) and additive percentage (5%, 10%, 15%, and 20%) were taken into account. Notations used to procure detailed information about the samples are given in Table 3.

Table 3 Notations used in the experimental study
Full size table

Mechanical properties and durabilities

The optimum moisture content (wopt) and maximum dry density (ρmax) of the mixtures were determined with standard compaction test (ASTM D 698). Liquid limits (wL), plastic limits (wP), and plasticity index (PI) of the mixture were determined considering ASTM D 4318. Optimum moisture content was considered in sample preparation.

In the context of the study, the effect of waste zeolite amount on unconfined strength and durability parameters was investigated. Within this scope, natural soil, soil-waste zeolite, soil–lime, and soil–lime-waste zeolite mixtures are taken into account. After a 28-day curing period, unconfined compressive strengths of the samples were determined (ASTM D 2166). The soil samples were subjected to 12 F-T cycles (ASTM D 560) to investigate durability properties of the stabilized soils. Immediately, after F-T cycles, CT imaging was performed and unconfined strengths of samples were determined. Closed system freezing–thawing device was used to subject F-T cycles. The specimens were subjected to freezing at − 23 °C for 24 h and then thawing phase + 21 °C for 24 h. One F-T cycle comprises freezing and thawing phases.

The effect of waste zeolite on the volumetric change of stabilized soil was investigated with swelling tests. Swelling pressure and free swell index tests were performed on soil (S), soil-waste zeolite (SW, SG, and WY), soil–lime (SL), and soil–lime-waste zeolite (SLW, SLG, and SLY) mixtures considering ASTM D 4546.

CT and SEM imaging

Besides mechanical properties, CT scanning and SEM imaging were performed to examine the effect of F-T on volume chance and pore characteristics of samples. CT scanning was performed with Toshiba Alexion industrial tomograph. Energy levels of 100 kV and 200 mA were applied for scanning. Each sample was divided approximately 152 pieces of slices (0.5 mm slice thickness) with 512 × 512 pixels. The volume and HU value of each sample were determined with 3D Doctor software using image processing technique. Boundary conditions and HU values of the sections were determined as shown in Fig. 4. The HU values of the samples were determined as average HU value of the sections. Characteristic HU value of the 28-day cured samples was determined before and after the F-T cycles.

Fig. 4
figure 4

Determination of boundary condition of the scanned samples

Full size image

As enlarging pore size is the primary reason for strength loss originating from F-T cycles, the effect of waste zeolite on the volume change of stabilized soil was investigated. In the context of the study, the volume change of the samples was examined with CT imaging. Volumes of the samples were calculated considering pore volumes with digital image processing software (3D doctor), as it is shown in Fig. 5. The effects of curing time and F-T cycles on volume change of natural soil (S), soil–lime (SL), and soil–lime-waste zeolite (SLW) mixtures were investigated. Volumes of the 1-day cured and 28-day cured samples were determined to examine the effect of curing on volume. The volume of the 28-day cured samples subjected to the F-T cycles was determined to examine the durability performance of the waste zeolites.

Fig. 5
figure 5

Determination of pores of the sample

Full size image

Scanning electron microscope (SEM) images of the soil–lime (SL) and soil–lime-waste zeolite (SLW) mixtures were analyzed to examine the effect of curing (28 days) and F-T cycles on the mechanism of small scale behavior of waste zeolite and lime mixtures in the stabilized soil. The analyses were conducted using Zeiss EQV LS-10 electron microscope.

Statistical analyses

The HU values (F-Ts) of the stabilized soils were determined from the CT scanning, and statistical significance between HU values and unconfined compressive strength was examined with SPSS 18.0 statistical software. Analysis of variance was performed in the evaluation of the values. Pearson’s correlation analysis and Spearman’s Rho were performed to measure the strength of association between two variables.

Results and discussion

The effect of waste zeolite and lime on compaction parameters and consistency limits

Compaction parameters and consistency limits of the soil, soil–lime, and soil–lime-waste zeolite mixtures are shown in Table 4. The liquid limit and plasticity index of the soil-waste zeolite mixtures decreased with increasing tuff content. Sharp decrease of liquid limit and plasticity index values were seen in lime-stabilized soil, and lower wL and PI values were obtained with soil–lime-waste zeolite mixtures. Waste zeolite waste zeolite-lime mixtures lead to a small amount of decrease in maximum dry unit weight and optimum moisture content of stabilized soils.

Table 4 Percentage of tuff and lime used in mixture preparation and some physical properties of the mixture
Full size table

The effect of waste zeolite content on swelling characteristics of stabilized soil

Swelling pressure tests were performed on natural soil (S), soil-waste zeolite mixtures (SW), and soil–lime-waste zeolite mixtures cured for 14 days at 20 °C. Figure 6a and b show the effect of waste zeolite and lime on free swell index and swelling pressure, respectively. From the figures, it was seen that free swell index and swelling pressure decreases with increasing tuff content, and lime plays a decisive role on swelling characteristics. Lime and lime-waste zeolite mixtures lead to eliminate the volumetric behavior of soil.

Fig. 6
figure 6

The effect of waste zeolite, lime, and waste zeolite-lime mixtures on a free swell index of samples and b swelling pressure of samples

Full size image

The effect of F-T cycles on unconfined compressive strength

In this part of the study, the effect of the amount of tuff wastes on unconfined compressive strength was examined. Within this scope, a high plasticity clayey soil was stabilized waste zeolite and lime-waste zeolite mixtures with various tuff contents. Unconfined compressive strengths of soil (S) and soil-waste zeolite mixture (SG, SW, SY) are presented in Fig. 7a. A slight increase in unconfined compressive strength was observed with increasing waste zeolite amount. A similar trend in USC values was observed in ShahriarKian et al. (2021). In this study, it was intended to stabilize a low plasticity soil with zeolite, cement, and cement-zeolite mixtures. UCS values of the soil-zeolite mixture were higher with respect to the soil. In both of the studies, a slight increase in UCS can be explained by the role of materials. Natural tuff and zeolite acted as filler material.

Fig. 7
figure 7

Unconfined compressive strength of the samples a stabilized with waste zeolite before the F-T cycles, b stabilized with waste zeolite after the F-T cycles, c stabilized with waste zeolite-lime mixtures before the F-T cycles, and d stabilized with waste zeolite-lime mixtures after the F-T cycles

Full size image

Besides, durability performances of the waste zeolites were investigated by subjecting the samples to strength tests after F-T cycles. In Fig. 7b, unconfined compressive strengths of the mixtures after subjecting F-T cycles were shown. The strength of the waste zeolite-soil mixture showed better performance than natural soil. However, a notable increment of unconfined strength and durability was not observed.

Unconfined compressive strength of natural soil (S), soil–lime (SL), soil–lime-waste green tuff (SLG), soil–lime-waste white tuff (SLW), and soil–lime-waste yellow tuff (SYG) mixtures are presented in Fig. 7c. From the figure, it was seen that waste zeolite increased the unconfined compressive strength of the samples up to 50%. The reason for the strength increment can be explained by containing a high amount of SiO2, Al2O3, and Fe2O3 oxides of waste zeolites. The strength increased with increasing tuff content up to a constant rate; beyond this rate, strength of the samples decreased with increased tuff content. Optimum contents of the SLG, SLW, and SLY are 5%, 10%, and 5% respectively. The effect of zeolite-rich material on the UCS value of the samples becomes more explicit when they are mixed with CaO supply such as cement or lime. Figure 7c is compatible with the findings of ShahriarKian et al. (2021) and Chenarboni et al., (2021). These studies show that the cement-zeolite mixture increases the UCS values, and zeolite has positive effects on the UCS values of the stabilized samples.

It can be seen from Fig. 7d that tuff wastes reduce the negative effects of F-T cycles on strength. This finding is compatible with ShahriarKian et al. (2021) and Jamshidvand et al. (2021). These studies show that zeolite or zeolite-rich tuff alleviates the negative effects of F-T cycles on strength. It is apparent from this figure that the strength reduction of lime and SLW mixture is the lowest. Because yellow waste zeolite contains much more Na2O than other types of tuffs, NaOH is formed as a result of the reaction of Na2O with water. The NaOH leads to an increase pH value of the soil media and increases the stability of CSH products.

Macro-microstructure of the samples

This part of the study was intended to investigate the relation between the microstructure of the soil samples and unconfined compressive strengths. The HU value was considered in the evaluation of the macrostructure of the samples. The HU values of the 28-day cured samples before and after the F-T cycles were determined. The determined HU values of soil (S), soil–lime (SL), soil–lime-waste green tuff (SLG), soil–lime-waste white tuff (SLW), and soil–lime-waste yellow tuff (SYG) mixtures are presented in Fig. 7a. After subjecting F-T cycles, the HU values of the 28-day cured samples were determined, as it is shown in Fig. 7b. From Figs. 6 and 7, it was seen that tendency of the HU value and unconfined compressive strengths was similar.

Simple correlation analyses were performed to examine statistical significance between unconfined compressive strength and HU values before and after the F-T cycles. As a result of the Spearman’s rho test, it was determined that there is a significant, positive correlation between HU value and unconfined compressive strength (r = 0.99, p < 0.01). Regression analyses were performed to examine the statistical predictability of unconfined compressive strength using HU values. As a result of ANOVA, it was determined that unconfined compressive strength could be estimated using HU values (r = 0.96, r2 = 0.93, p < 0.01).

Figure 9 shows the effect of curing time and amount of waste zeolite on the volume change of stabilized soils. From the figure, it was seen that volumes of the SK and SLW samples increased with curing time. Volumetric increment of the solid phase of the samples can be explained with synthesized high strength materials as a result of pozzolanic reactions by considering the increase of strength and HU value shown in Figs. 7 and 8, respectively. F-T cycles decreased the volume of all samples. By considering HU loss in Fig. 7b, the volumetric reduction can be explained with an increment of pore size. From Fig. 9, it was seen that waste zeolites are effective on volumetric stability of soils.

Fig. 8
figure 8

HU values of samples a stabilized with waste zeolite before the F-T cycles, b stabilized with waste zeolite after the F-T cycles, c stabilized with waste zeolite-lime mixtures before the F-T cycles, and d stabilized with waste zeolite-lime mixtures after the F-T cycles

Full size image
Fig. 9
figure 9

The effect of curing time and C-T cycles on volume change of the samples

Full size image

In this part of the study, the effects of the curing period and F-T cycles on the microstructure of the samples were investigated. Within this scope, the microstructures of the SL and SLW samples were considered. SEM images of the 1-day cured samples and 28-day cured samples (before and after C-T cycles) were examined. Figures 10 and 11 show SEM images of the lime-treated soil. From Figs. 10a and b and 11a and b, flocculation and agglomeration phases of stabilized soil are seen. As seen from Fig. 10b, flocculated and agglomerated particles lead to an increase in the volume of the solid structure. Thus HU value and unconfined strength of the samples increased. From Figs. 10c and 11c, it is seen that F-T cycles caused an increase of pore dimensions. Thus, HU values and the strength of the samples subjected to F-T cycles decreased (Figs. 7 and 8).

Fig. 10
figure 10

SEM images of SL samples a 1-day cured samples, b 28-day cured samples before F-T cycles, and c 28-day cured samples after F-T cycles (× 2500)

Full size image
Fig. 11
figure 11

SEM images of SLW samples a 1-day cured samples, b 28-day cured samples before F-T cycles, and c 28-day cured samples after F-T cycles (× 5000)

Full size image

Conclusions

This paper investigated the durability performances of stabilized high plasticity clayey soil with lime and waste zeolites additives. The contribution of additives was examined in macro and micro aspects. In this context, unconfined compressive strength, CT, and SEM images were taken into account. The following conclusions can be drawn from the results of the test:

  • HU values and unconfined compressive strength of the lime-stabilized samples increased remarkably with the curing period. Waste zeolite-lime mixtures showed better performance than lime-stabilized soils based on unconfined compressive performance. From the XRF, XRD, CT imaging, and unconfined compressive strength results, it was seen that waste zeolite contributed to pozzolanic reactions that occurred in the lime-stabilized samples due to comprising a high amount of active pozzolanic oxide components. CSH gels formed in the pores of the soil particles due to pozzolanic reactions between lime and waste zeolites. Volumetric increment of the samples can be explained with a new material synthesized between the soil particles. HU values increased depending on synthesized solid material, and increased UCS value confirmed that high strength material formed as a result of pozzolanic reactions.

  • Waste zeolite and lime mixture increased the UCS values up to 53% in comparison to lime-stabilized samples. Little differences in oxide composition of waste zeolites may affect mechanical, and durability performances of samples stabilized with lime and waste zeolites. Lime and white waste zeolite mixture showed a remarkable improvement in the UCS value. A similar trend was observed between the UCS and HU values.

  • The SEM analyses determined that the inner structure of the soils stabilized lime and waste zeolite mixture were denser than lime-stabilized soils. F-T cycles led to enlarge pore size. Adding waste zeolites led to increasing durability performance of soils stabilized with lime. Lime-green waste zeolite mixtures showed the best performance against freezing–thawing cycles. Although the UCS value of the lime-stabilized soil decreased 35% due to F-T cycles, the strength loss of the stabilized soils with lime and green waste zeolite mixture was 12.5%. Additionally, the UCS value of the samples subjected to F-T cycles was 33% higher than lime-stabilized soils did not subject to F-T cycles.

  • The strength of soils can be evaluated accurately considering HU values. As a result of the Spearman’s rho test, it was determined that there is a significant, positive correlation between HU value and unconfined compressive strength (r = 0.99, p < 0.01). Considering the ANOVA results, it was determined that unconfined compressive strength can be estimated using HU values (r = 0.96, r2 = 0.93, p < 0.01).

  • CT imaging has a considerable advantage in geotechnical applications due to giving an opportunity to examine 3D structures and evaluate the strength performances of soils without any disturbance.

The stabilizing behavior of tuff wastes with and without lime in various types of natural materials can be researched in future studies. Instead of lime, another major binder can be utilized. Experiments carried out in the laboratory can be backed up by field tests.