Using Waste PET Shreds for Soil Stabilization: Efficiency and Durability Assessment

Encountering weak fine-grained soil presents a significant challenge during highway construction. The conventional approach to address this challenge involves incorporating calcium-based stabilizers, particularly cement, for stabilization. However, despite its widespread use, the application of cement for soil stabilization has adverse environmental consequences. Accordingly, finding alternative methods to minimize cement usage has become a prominent area of research, from researchers worldwide. This study evaluates the effectiveness of utilizing discarded polyethylene terephthalate (PET) bottle shreds as soil reinforcement alongside lower cement contents. To this end, soil samples were mixed with varied contents of two grades of shredded PET. The ground PET pellets, which displayed diverse shapes, demonstrated properties resembling fibres. Compacted samples were subjected to CBR to determine optimum PET content. Strength, small-strain stiffness and durability of soils in original state, mixed with cement and PET were measured by a programme of CBR, wetting–drying cycles, and ultrasonic pulse experiments. Findings suggest a 28–91% increase in strength of soil upon its mixing with cement—PET shreds. Substituting cement with cement—PET shred led to a drop in accumulated loss of mass (ALM). In twelve cycles of wetting and drying, the small-strain stiffness initially decreased, but then stabilized at approximately the same value in the subsequent cycles. The porosity to binder ratio was adopted as an index for CBR, ALM and the maximum shear modulus. The latter was finally proposed as a measure for durability.


Introduction
Clay and silt are very common types of soil found worldwide, and it is very frequent that soil structures such as slopes, and highway embankments are built from this type of soil [1,2]. For subgrades, building on natural weak finegrained soil could cause serious damage, risk uncontrolled erosion, and decrease the service life of the earth structure [3]. It is known that the engineering properties of soil can be improved using the process of soil stabilization [4]. Practice of mixing cement with the soil is one of the most common methods in soil treatment [5][6][7][8][9][10][11][12][13][14][15]. Cement stabilization is proved effective in enhancing the geotechnical properties of soil [16,17]. However, the process of cement production is harmful to the environment due to the use of substantial amounts of energy and natural resources, as well as the release of massive volumes of carbon dioxide (CO 2 ) [18][19][20]. The carbon dioxide (CO 2 ) emissions from cement manufacturing from 1928 to 2018 were approximately 38.3 GTs, globally, according to Andrew [21]. As such, many researchers [e.g., [22][23][24][25] have examined alternative, more sustainable materials and methods to decrease the overall emission throughout the construction sector [e.g., 22 and 23, also see 24 and 25]. A particular challenge with soil-cement stabilization method is the durability of the treated soil against degradation. Durability can be defined as a material's ability to withstand the weathering conditions, wetting-drying and freeze-thaw cycles, while retaining its integrity and stability over extended period of time [26]. The diminishing benefits of cement stabilization after weather exposure cycles [14], provide additional support for exploring alternative stabilizing materials as substitutes for cement.
Furthermore, the growing concern over solid waste generation necessitates effective management strategies to address this environmental issue [27,28]. Polyethylene terephthalate (PET) is one of the most widely used materials to manufacture products like bottles for beverage and other containers. After single use, PET bottles are discarded and become PET waste [29,30]. Numerous researchers have investigated the utilization of PET bottles as construction materials, revealing a wide range of advantages in soil improvement applications. These benefits have been observed in both cohesive fine-grained soils and cohesionless coarse-grained soils [31][32][33]. Ferreira et al. found that in the case of sandy soil, the introduction of PET fibres contributes to increased soil strength, reduced deformation in vertical and lateral directions, and improved stiffness [34]. Incorporating waste plastic bottle shreds, particularly PET, along with cement offers a potential waste disposal method while reducing the cement requirement for soil stabilization, leading to a decrease in CO 2 emissions associated with cement manufacturing. Hence, several studies have investigated the engineering properties of cement stabilized soil reinforced with PET shreds [35][36][37][38][39][40][41]. Bozyigit et al. [42] investigated the strength of cement stabilized PET strips reinforced clay, and stated that adding PET strips improve the unconfined compressive strength of clay samples.
Several previous studies have investigated the enhancement of fine-grained soil through the incorporation of cement, slag, lime, and gypsum [43][44][45][46][47][48][49]. The investigations have yielded findings indicating that the addition of cement induces brittle behaviour in the stabilized soil, with limited or negligible plastic deformation [50]. Additionally, there have been experimental trials involving fibre-reinforced cemented soil. Indeed, a significant challenge encountered in PET fibre/cement soil stabilization is the dependency of the final mixture's performance on factors such as the consistency of preparation and mixing procedures, soil type, compaction energy, and the shape and dimensions of the added fibres [51][52][53][54].
Khattak and Al-rashid [55] considered the durability of PET fibres reinforced cemented fine-grained soil and found that PET fibres improved the durability of the mixtures. Zhang, et al. [56] carried out a study on the durability of cemented silty clay soil, the study revealed that the loss of mass decreases as the cement content increases. Furthermore, Consoli, et al. [57], established an effective approach for stabilized cement treated soils based on the porosity/cement (binder) ratio as a strength parameter. The porosity/binder index could be explained as the ratio of porosity of the dense mixture to the percentage of the binder by volume [57]. Further research investigated the feasibility of using the porosity/binder ratio as durability and stiffness factors in cement stabilized soil samples, by considering mass loss after multiple cycles of wetting and drying [11,58,59]. In addition, Hanafi et al. [44] studied the porosity/binder ratio for cemented soil mixes from the mass loss perspective.
The primary goal of this study is to evaluate the practicality of PET shreds as a reinforcing medium and to determine the ideal percentage range of PET shreds in terms of strength, CBR, durability and small-strain stiffness.

Materials
The fine-grained soil used in this study was collected from Duhok governorate (36°59′24.5"N, 42°39′20.7"E) in northern Iraq and is categorized as low plasticity clayey silt, ML, according to Unified Soil Classification System (USCS) [60]. The characterization test results are presented in Table 1 and the grain-size distribution curve is presented in Fig. 1.
Type I ordinary Portland cement (OPC), in accordance with ASTM C150/C150M-18 [61] was used in this study.
The PET shreds utilized in the current research were made from waste PET plastic water bottles. The bottles were first cut into small pieces using scissors, then ground into shreds using a disc mill grinder, and finally were screened to obtain two grades. The first grade sized 2-4.75 mm and the second grade sized 42-841 μm, these grades will be mentioned as size no. 10 and size no. 40, respectively. The process of grinding and screening the PET resulted in having fibre-like pellets in random shapes which is given the general name of "shreds" throughout this work. Figure 2 illustrates the two PET grades.

Methods
To determine the optimum PET shreds content range based on improvement in California bearing ratio (CBR) values, CBR tests were conducted. For CBR test, cylindrical samples were utilized as per ASTM D1883-16 [62]. The quantities of needed dry soil and PET shreds were prepared based on each mixture's optimum moisture content (OMC) and maximum dry density (MDD) that was found through standard Proctor compaction test ( Table 2). The CBR values were determined by applying pressure to the soil using a cylindrical piston. The piston was inserted into the soil at a constant  To determine the effect of adding PET shreds to cemented soil's behaviour, durability test in terms of loss of mass, and stiffness test in terms of ultrasonic pulse velocity (UPV) were performed as per ASTM D559/D559M-15 [63] and ASTM D2845-08 [64] respectively, on samples with dimensions of 11.62 cm in height and 15.24 cm in diameter. After determining the dry mass of the soil sample (M S ) based on the predefined dry density (ρ d ), the proportion of cement and PET contents was measured as a percentage of the dry mass of soil. The selected dry density for all samples was 1.4 g/cm 3 . The samples were prepared by combining the required weight of soil, PET shreds and cement in a sealed zip-lock bag. To prevent agglomeration, great care was taken to thoroughly mix the soil and shreds by rotating the ziplock bag in all directions for no less than five minutes. The materials in dry state were blended till a visual uniformity was reached. The predetermined quantity of distilled water was later added and the mixing continued till a homogeneous mixture was obtained. The prepared samples were compressed in five uniform layers in the CBR moulds. The extruded samples were subsequently wrapped in cling film and placed inside a climatic chamber set at a temperature of 24° ± 2 °C and were left for curing for 7 and 28 days, following the guidelines outlined in ASTM C511-19 [65].
The durability tests were held to evaluate the loss of mass produced by repeated twelve wetting/drying cycles. Each cycle started with samples submerging fully in distilled water for five hours at room temperature. Subsequently, the samples were taken out of the water and subjected to oven drying for forty-two hours at 71 °C. Afterward, the samples were brushed using a wire brush with an applied force of approximately 13.3 N. The accumulated loss of mass (ALM) percentage was calculated for each sample following the completion of wetting and drying cycles. The masses of the samples were recorded after each wetting-drying-brushing cycle to determine the mass loss for each individual cycle. The ALM after the twelfth cycle was then obtained by summing the masses recorded at each cycle.
Ultrasonic pulse velocity (UPV) tests were employed to evaluate the maximum shear modulus (G max ) for the same samples used in the durability test. The UPV measurements were conducted before initiating the cycles and after each individual cycle of the repeated 12 wetting/drying cycles. Each sample was measured for UPV after drying and before submerging it in water for the next cycle. Two transducers were attached to both the upper and lower ends of the samples using a coupler gel, with the precise distance between the transducers being measured. The device was then operated to determine the time taken for the waves to travel through the sample between the transducers, enabling the measurement of velocity. This measurement of velocity allowed for the calculation of maximum shear modulus (G max ).
A modified version of Eq. (1) of porosity was utilized in this study to calculate the porosity, this equation was initially proposed by Consoli et al. [66].
where, is the porosity, d is the dry density, M S , M C , and M PET are the mass content of the soil, cement, and PET shreds respectively, and S , C , and PET are the density of soil, cement, and PET shreds respectively.
To predict the behaviour of soil samples stabilized with cement, Consoli et al. suggested a ratio between the porosity and cement index (η/C iv ) [67]. Ekinci et al. offered a general index (X iv ), that is applicable for different types of binder [49]. In the current research, the general index was used to predict the ALM and G max of the mixes, in which Xiv is determined using the Eq. (2).
where, X iv is the adjusted general index for this study, V is the total volume of the soil sample, V C is the volume of cement and V PET is the volume of PET shreds, M c is the mass of cement content, M PET is the mass of PET shreds content, c is the density of cement utilized, PET is the density of the PET shreds utilized.
An external exponent, α, must be adjusted to the parameter X iv to determine the relationships between strength η/ Xiv. The external exponent of adjusted porosity/binder η/ (X iv )0.32 index for this research was selected as the best fitting power. The selection of exponent of 0.32 was based on a number of former studies in the literature. A number of studies which were carried out for different kinds of soil demonstrated that the exponent may range between 0.28 and 0.35 [44,49,68]. In addition, in other studies it was stated the exterior exponent of 0.32 to be the best fitting exponent [57][58][59].
The data of all samples prepared are shown in Table 3.

Control of Shreds on Strength
The results of the CBR tests are shown in Fig. 3. The findings demonstrate that the addition of PET shreds has enhanced the CBR values irrespective of the grade and (1) percentage added. Considering the grade of the PET shreds, the results show that the grade 2-4.75 mm (size no. 10) presents better performance than the grade sized 42-841 μm (size no. 40) regarding the CBR strength values. In addition, the PET shreds content of 0.7% exhibited the optimum improvement for coarser PET grades, while the content of 0.6% provided the optimum improvement for the finer PET grades. Concluding from the results the percentage range of 0.6-0.8 is the optimum PET shreds range to be utilized in reinforcing cemented soil in this study. Figure 4 shows ALM for all the mixes after 12 wetting-drying cycles. The mixes have different cement content and PET shreds percentages, and are cured for varying durations. The results after 12 cycles show an increase in the durability of all of the PET sherds enhanced samples regardless of the curing duration and percentage addition. The percentage ALM being reduced in PET shreds enhanced samples can be explained by the reinforcement property that the PET shreds add to the soil, which hold the soil particles together and therefore make it more durable. Nevertheless, the other key factor in durability is the cement content, with higher cement content lower ALM was observed. It is important to mention that the porosity is a principal factor to consider regarding the ALM percentage, typically samples with lower porosity values show lower ALM percentage. This phenomenon can be related to the higher coordination number and particle contact which allow the cement to develop more bonds [15].

Controls on ALM
Statistical analysis was used to investigate the effects of each factor on ALM, namely; Shreds size, cement content, and curing days. This was done by using Analysis of Variance (ANOVA). Figure 5a illustrates the effect of the PET shreds size on loss of mass, ALM, from the results it can be comprehended that addition of size no. 10 PET resulted in better performance of the samples than addition of size no. 40, as the loss of mass is reduced in all mixes. Figure 5b illustrates the effect of the content of added PET shreds on ALM percentage. Three percentages of 0.6, 0.7, and 0.8 of PET shreds were added to the samples. For size no. 10 best performance was observed at percentage of 0.7 followed by 0.6 and 0.8%. On the other hand, for size no. 40 PET shreds, the best performance was observed at the percentage of 0.6 followed by 0.7 and eventually 0.8%. Figures 4 and 5 show that the increase in cement content reduces the ALM at the end of each cycle of wetting and drying. Increase in content of the cement in the samples cause more solid cementation bonds between the mixture units, therefore it reduces the ALM of the sample which is aligned with the findings of Consoli et al. [15,70]. Additionally, Fig. 5c demonstrates that 28 days of curing results in higher durability value than the 7 days. The ASTM standard requires the ALM percentage after twelve cycles not exceeding 7% for fine grained soils, and in this study all the mixes satisfied this requirement. Consequently, the cement content of 7% can be considered as an optimum percentage to be added to the mixtures. Figure 6 illustrates the relation between the ALM and the adjusted porosity/binder index (η/Xiv0.32) of the samples in this study. Considering the results, it is evident that an increase in porosity reduces durability, Fig. 6 indicates that porosity reduction due to the integration of PET shreds, results in a decrease in ALM at all curing durations. Moreover, it can be noticed that the 7 and 28 days of curing had a higher effect at high porosity. However, the gain in durability due to PET shreds addition seems to be more significant for higher values of (η/Xiv0.32). For the purpose of regularising the results of different mixes and curing conditions, the method proposed by Consoli et al. is applied. This equation divides the established power equations to a specific value of CBR, stiffness, or loss of mass, matching to a value of [porosity/binder] = ∇ within the studies ranges, to predict the mechanical performances of cemented sand [71]. Extending the model to determine the outputs of cement and PET stabilized clayey silt, a normalization model is introduced in Fig. 6. This method divides the generated equations of each mixture by a particular ALM value. In this research, samples were selected with ALM at (η/Xiv0.32) = ∇ = 32 for all mixes. A sum of 42 ALM results were utilized to adjust Eq. (3), permitting the formation of a correlation between the ALM and the adjusted porosity/binder index. A reasonable determination's coefficient (R 2 ) of 0.87 was acquired. It can be presumed that the obtained correlation can be used to predict ALM in wetting and drying cycles using only the Porosity/Binder Index.

Small Strain Stiffness
In Fig. 7 the variation of small strain stiffness along cycles is presented. The results show that soil without PET enhancement in the case of curing period of 7 days, has a decrease in the stiffness during the initial cycles occurs however it will regain some of it toward the 12th cycle. While, in the case of addition of size no. 10 PET shreds, a reduction in the initial stiffness can be observed compared to the plain cement-soil, which is eventually recovered in the upcoming cycles. However, in the case of PET shreds size no. 40, the increase in stiffness continues up to the 12th cycle. In the 28 days cured samples, a constant decrease can be noticed in the stiffness, yet, this decrease ratio is less in the case of samples enhanced with PET shreds in both sizes. The initial stiffness loss in 7 days-cured samples, is potentially a result of retracting and resultant cracking of the samples throughout temperature rise in the drying section of the cycles and confirms the observations in comparable previous research [69][70][71]. In the case of samples enhanced with PET shreds of size-no. 40, there was no decrease in stiffness. The reason behind that is that the stabilizing material in these samples helped the soil grains to have their optimum bonding and interlocking orientations, therefore, the initiation of cracks was prevented, and the soil sample kept its integrity throughout the cycles.
In the wetting cycles, water fills the voids of soil and ran between the miniscule fissures, widening them and causing initiation of new cracks in the sample. This process can cause reduction in stiffness when undergoing repetitive cycles of wetting and drying. In the state of 7 days curing duration, for samples reinforced with PET, the stiffness improved after the loss in initial cycles, which could be related to the fact that the cementation bonds continue developing up to 28 days. However, after 28 days, most of the cementation bonds being completed, loss of stiffness takes place by subjecting the samples to further cycles. After passing 7 days the strength gain gets slower and the cementation bonds almost stop developing after 28 days, therefore fissuring and disintegration would occur in latter cycles. Previous studies also confirmed that silt samples treated with cement gained more than half of its overall strength and stiffness before 28 days [72]. Moreover, it can be seen that PET shreds treated samples showed better stiffness throughout the cycles compared to the control mix of soil and cement. This can be attributed to the reinforcement action of PET shreds that limits the cracks development during the cycles. Figure 8 illustrates the correlation among the maximum shear modulus (G max ) at 12 wetting/drying cycles and the adapted porosity/binder index and (η/X iv 0.32) of the treated soil mixes in this study. As it can be noticed, the G max gain is lower compared to the increase of η/(X iv )0.32 value. It can be observed that a having high porosity and low volumetric binder contents result in lower G max values. The results of the analysis of ALM was performed also for G max , and it revealed that (η/X iv 0.32) correlates with G max at 12 cycles. Similar to ALM, G max can be normalized by dividing the generated equations for each mixture by a certain G max value. The relation between normalized G max at 12 cycles [G max /(G max for η/(Xiv)0.32 = 32)] and the adapted porosity/ binder index is depicted in Fig. 8 (4) could introduce a great contribution in soil stabilization with cement and PET area, by determining the G max value of cured fine grained soil treated with cement and PET shreds, using the results of a single experiment.

Conclusions
This study evaluated the effect of reinforcing low content cement stabilized soil with PET shreds in different ranges by carrying experimental and analytical work and the following conclusions were obtained: (1) The PET shreds reinforcing of soil is a helpful solution to improve the small strain stiffness and durability of soil. The inclusion of different percentages of PET shreds improved CBR values from 28.5 to 90.7% compared to plain compacted soil. In general, PET shreds of 2-4.75 mm demonstrated better results compared to 42-841 μm. The optimum percentage range of PET shreds to be added was found to be 0.6, 0.7, and 0.8% of the selected sizes. (2) The soil stabilized with cement and PET shreds presented better durability to wetting and drying exposure compared to soil stabilized with only cement, it also ensured the satisfaction of mass loss according to standards. (3) Reinforcing soil with cement and PET shreds reduced brittleness of soil due to the ductility properties of PET shreds. In addition, the inclusion of PET shreds reduced the stiffness loss through wetting /drying cycles.
(4) The use of the adapted porosity/binder index (η/Xiv) to extract the performance of the mixes investigated in this study can be regarded as successful. When G max and ALM values were associated with the modified porosity/binder index, high coefficients of determination were obtained. G max and ALM of mixtures can be predicted using the proposed porosity/binder index calculations. (5) Decrease in the porosity/binder index presented a direct correlation with Accumulated Loss of Mass, ALM, and an inverse correlation with stiffness, therefore, lower porosity combined with a higher binder content resulted in more durable and sturdier samples. (6) This research suggests an eco-friendly and cost-effective solution by reducing the usage of cement, minimizing the environmental impacts of soil disposal, and lowering CO 2 emissions resulting from the transportation of unwanted soil. Additionally, this technique is cost-effective as the waste material is utilized in construction work instead of incurring transportation costs. Moreover, cement manufacturing's environmental impacts can be reduced by reinforcing cemented soil with PET shreds, which also makes it more costeffective by lowering cement usage. (7) Correlations between the ALM, G max , porosity, and amount of cement and PET shreds used was formed to predict the behavior of soil after facing cycles of wetting and drying.
In this research, a formula to predict the ALM through stiffness was proven valid. Future study may be performed for validating this formula and prove its applicability more so that it can be used in practical areas.