Compaction Characteristics of Kaolin Reinforced with Raw and Rubberized Oil Palm Shell

This paper presents an attempt to evaluate the suitability of oil palm shell (OPS) and rubberized OPS (ROPS), an alternative bio-material, as reinforcement in kaolin. OPS was surface coated with rubber, and its water absorption potential was studied in 5 media involving water and kaolin samples (with different water contents). The water absorption data measured in the laboratory was used as an indirect measure to verify the degradability of ROPS samples when used as reinforcements in kaolin. The surface treatment of OPS with rubber was found to perform well, with around a fivefold decrease in water absorption, thus making it an ideal treatment procedure to this end. Kaolin-ROPS mixtures with different OPS and ROPS proportions (0, 5%, 10%, 20%, and 30% by weight) were prepared in laboratory to evaluate their compaction behaviors. Both standard proctor compaction and mini-compaction procedures were adopted in this study to ensure applicability of the findings across a wide range of compaction methods adopted in the laboratory. Compaction curves obtained for both kaolin-OPS and kaolin-ROPS mixes showed a decreasing trend in the maximum dry density values with increasing proportions of OPS and ROPS. Optimum water content of kaolin-OPS mixtures did not show a significant variation, while kaolin-ROPS mixture showed a downward trend with increasing ROPS contents, thereby signifying improvement in the compaction characteristics after OPS reinforcement in kaolin.


Introduction
Recent decades have witnessed the transition of research and industry towards sustainable developments. Solid wastes or biomass have seen increased utilization/reuse for reinforcement and stabilization of problematic soils. Reduction in greenhouse gas emission and a greater use environmentally friendly materials are the key reasons for this development. Biomass predominantly used for soil reinforcement typically involves fibrous and shell-type materials, which prove to be a cost-effective and eco-friendly alternative to synthetic grids and strip-type geotextiles. Among the various locally available bio-wastes, jute, bamboo, coir, and oil palm-based fiber are the frequently studied materials [1], and have demonstrated notable efficiency in improving the strength, ductility, and stiffness in clayey soils. Oil palm-derived fibers have displayed inherent ability to interlock the soil particles together, resulting in a high-strength coherent matrix [2]. Oil palm shell (OPS) is one such bio-mass or by-product commonly obtained from palm oil mills. OPS, also known as palm kernel shell, is derived from the palm oil extraction process, mostly in the form of raw shell fractions. OPS is well known for its properties such as low specific gravity (typically 1.14 to 1.62) and low bulk density (typically 4.9 to 5.9 kN/m 3 ) depending on its age and species [3].
OPS, an agricultural solid waste, inevitably undergoes biodegradation. Its decomposition rate is subjective and varies depending on the presence of moisture and air [4]. As expected in a wide range of natural materials, water absorption most likely initiates the degradation process in OPS as well. The cause of OPS's high-water absorption is partly attributed to the presence of fibers and pores on the surface of the shell [5,6]. Moreover, water absorption of OPS is highly influenced by the species of palm tree [3]. Studies estimate 14% to 33% water absorption when OPS are water soaked for 24 h [6,7], while another study shows that the average period of decomposition for untreated oil palm residues, including empty fruit bunches, leaflets, rachis, eco-mat, and palm fronds, is around eight months [8]. However, with a low nutrient content, which is optimum for decomposition as compared to other oil palm residues [9], OPS is likely to decompose at a much faster rate. For this reason, processing OPS into a less biodegradable form is crucial before utilizing and exposing it to moist soil environments. Various treatment methods have been tried to improve and reduce the water absorption of OPS and natural fibers. Studies involving OPS in concrete applications show pre-treatment of OPS using preservatives such as sodium dichromate solution, ferrous sulphate solution, and polyvinyl alcohol solutions [4]; in which studies polyvinyl alcohol (PVA) showed promising results by reducing movement of water into OPS. A recent study demonstrates heat treatment as another promising approach to reduce water absorption [10]. While these studies use advanced methods, traditional rubber-based coating, which has demonstrated its effectiveness in corrosion and chemical resistivity for steel items [11], is likely an eco-friendly approach. Notwithstanding this, rubber-based liquid sealants are resistant to water and moist air; moreover the physical treatment of OPS with synthetic liquid rubber can help decelerate the biodegradation process, while generating the least disturbance to the chemical and structural properties of the OPS.
Soil compaction is one of the commonly used soil improvement methods to enhance structural and mechanical properties by reducing the air-phase volume in the soil mass without changing the water volume within the pore space. Soil compaction typically aims to increase the bearing capacity and shear strength of soil, while reducing permeability and settlement of the soil mass by reducing its porosity [12,13]. The compaction characteristic in terms of optimum moisture content (w opt ) and maximum dry unit weight (γ dmax ) is rather important to ensure preparation of the soil base-layer with sufficient bearing for almost all construction works. Nonetheless, it is also crucial to identify the effect of varying compaction efforts, as this can play an important role in applicability of the results. Most of the published research bases its discussions mainly on compaction efforts using standard proctor and modified proctor compaction tests. The mini-compaction procedure [14] is another method which is highly preferred for fine grained soils. It should be noted that very limited studies have reported the compaction characteristic of natural fibers and shell-like material reinforced soil using the mini-compaction method. In addition, the difference in results between the standard proctor compaction and mini-compaction methods, due to variation in energy transmission during the compaction process, raises concern [15], which has led to a research gap requiring comparison of results from both these test procedures to arrive at reliable conclusions.
The main aim of this study is to evaluate the suitability of the ROPS samples as another effective reinforcement material for kaolin in improving the compaction characteristics and behavior of the ROPS-kaolin samples, using laboratory compaction tests. Effectiveness and performance of coating layer(s) of rubber sealant on the OPS's surface is evaluated against resistance to water absorption. The compaction characteristics and behavior of thus prepared OPS-kaolin and ROPS-kaolin samples are evaluated in the laboratory. In this process, a comparative study between standard proctor compaction and mini-compaction [14] procedures are also conducted to identify any variation in results due to varying compaction efforts and procedure. The paper therefore presents an elaborate discussion of the results in comparison to relevant literature.

Material and Methodology
The kaolin sample used in this study was procured commercially. The kaolin sample was tested in the laboratory to evaluate its basic geotechnical properties. Table 1 tabulates the laboratory procedures used and the geotechnical properties of the kaolin sample. OPS samples used this study were procured from an oil palm mill/plantation located in Klang region of the Selangor state in Peninsular Malaysia. The OPS samples thus transported to the Soil Research Laboratory located at Monash University Malaysia were initially washed with clean water to remove any oil residues. The samples were then dried to remove moisture using oven-drying method in the laboratory-Under a constant temperature of 105 • C for 24 h. Table 2 tabulates the key physical properties of the OPS samples used in this study. Please note that the variations in the properties are expressed as a range denoting the top and bottom limits of the measured properties.  Commercially available liquid rubber (Flex seal, www.flexsealproducts.com) was used in this study for the surface coating of the OPS samples. The steps/procedure included dipping individual OPS samples into the rubber sealant followed by curing for a period of 24-30 h at room temperature. For double layer coating, the dipping process was repeated once after the previous layer of coating was set (curing for a period of around 24-30 h). In this study, single layer rubber-coated OPS and double layer rubber-coated OPS are identified as ROPS SL and ROPS DL , respectively. Figure 1 shows the images of the OPS and ROPS samples. The ROPS samples were then preserved in air-tight containers before using them in the laboratory experiments as detailed further in this section. The dimensions of the OPS used in this study typically ranged between 4.75 mm and 15 mm. The thickness of the rubber layer for single-layer and double-layer coating was observed to vary between 0.04-0.8 mm and 0.15-2 mm respectively. Thickness values were basically influenced by the surface irregularity of OPS.

Water Absorption Test
Water absorption tests were conducted as per the recommendations detailed in ASTM D570 [21]. Other than water, the water absorption performance of OPS, ROPSSL, and ROPSDL were conducted in the kaolin with 32%, 48%, 64%, and 128% moisture content, corresponding to 0.5wL, 0.75wL, 1wL, and 2wL of kaolin. A total of 15 scenarios with an average of 8 samples each were prepared for each scenario. Table 3 details the schedule used for the water absorption tests in this study. The test procedure starts with recording the dry mass of OPS, ROPSSL, and ROPSDL samples using a standard weighing balance with accuracy of 0.001 g. The specimens were then immersed in preselected medium and moisture conditions as per the scenarios. The test was conducted at controlled laboratory temperatures of 25 ± 3 °C for up to 20 days (480 h). During the testing period, water absorption of the specimens was continuously monitored at fixed intervals of six minutes. For

Water Absorption Test
Water absorption tests were conducted as per the recommendations detailed in ASTM D570 [21]. Other than water, the water absorption performance of OPS, ROPS SL , and ROPS DL were conducted in the kaolin with 32%, 48%, 64%, and 128% moisture content, corresponding to 0.5w L , 0.75w L , 1w L , and 2w L of kaolin. A total of 15 scenarios with an average of 8 samples each were prepared for each scenario. Table 3 details the schedule used for the water absorption tests in this study. Table 3. Experimental program used in the water absorption tests.

Scenarios
Descriptions (Sample + Medium) The test procedure starts with recording the dry mass of OPS, ROPS SL , and ROPS DL samples using a standard weighing balance with accuracy of 0.001 g. The specimens were then immersed in preselected medium and moisture conditions as per the scenarios. The test was conducted at controlled laboratory temperatures of 25 ± 3 • C for up to 20 days (480 h). During the testing period, water absorption of the specimens was continuously monitored at fixed intervals of six minutes. For this purpose, a few OPS/ROPS S /ROPS DL samples were extracted out from the medium at every interval and wiped with clean cloth to remove excess surface moisture, and mass of the sample was then measured immediately to avoid unnecessary moisture loss from the samples. Each reading was taken thrice and their mean value was recorded to ensure representative measurement. This process was repeated until the saturation was achieved-That is, the consecutive measurements showed no additional water absorption. The percentage water absorption, M t , was then determined using the following Equation: where w n refers to the mass of sample after immersion, and w d refers to the dry mass of samples before immersion.

Compaction Tests
Compaction behavior of kaolin supplemented with different OPS or ROPS DL compositions, ranging from 0-30%, were analyzed using both the standard proctor compaction [20] and mini-compaction [14] test procedures. In both compaction test procedures, kaolin mixtures (kaolin + OPS or ROPS DL ) were mixed thoroughly with a pre-determined amount of water. The mixture was then cured in plastic seal bags for 24 h prior to the compaction tests to ensure consistent distribution of water within the mixtures. For standard proctor compaction, the mixtures were compacted in three layers with 25 blows/layer using the conventional 2 kg rammer as stated in ASTM D698-12 [20]; whereas in mini-compaction, the mixtures were compacted in three layers with 36 blows/layer using 1 kg mini-compaction rammer. Since the specimens were prepared as a mixture of kaolin and OPS or ROPS DL , the specific gravity of these soil samples, G ss was estimated theoretically using the below relationship [22].
where M d1 is the mass of dry soil (kaolin in this study); M d2 is the mass of OPS or ROPS DL ; G s1 is the specific gravity of the soil; and G s2 is the specific gravity of OPS or ROPS DL . For both the compaction test procedures, a total of 8 samples across the dry and wet side of the optimum were measured to calculate the dry unit weight-water content relationships. These compaction curves were then used to estimate the optimum moisture content (w opt ) and maximum dry unit weights (γ dmax ). Figure 2 shows the variation of water absorption over a duration of 500 h. In general, the linear trend of water absorption observed in the OPS soaked in water with time can be explained in accordance with Fickian diffusion theory, which is defined as the diffusion of water from an area of higher concentration to one of lower concentration [23,24]. Discussions in this theory also indicate the possibility of an initial linear relationship between the water content (water absorption of material) and time, but the linearity may fade away, with the relationship tending to flatten as the saturation point is approached.

Water Absorption of OPS and ROPS Samples
As is evident from Figure 2a, the water absorption rate for OPS is higher in the first 24 h of immersion in water. After the first 24 h, the rate of water absorption starts to flatten and remain constant over a prolonged time. Higher water absorption rates can be explained by capillary action due to water diffusion and swelling of OPS surface/fibers. The capillarity action is hypothesized as the flow of water molecules along fiber-matrix interfaces and water diffusion through the OPS's bulk matrix (surface fiber). Immersion of OPS in water understandably creates a concentration gradient between the dry OPS and water medium. Thus, diffusion owing to concentration gradient is the most likely reason for the increasing rates of water absorption in OPS samples. Literature presents a similar explanation, relating the increasing water absorption to water uptake by capillary action in micro-pores and surface fibers [25,26]. Thus, a potentially swollen OPS surface is prone to micro-cracks, which can further escalate the capillarity and transport of water; as also explained by Dhakalet [27]. The trends and observations from this experimental study agree with trends presented by Sreekala et al. [26]-A study on the water absorption behavior of oil palm fibers. With very limited data and research discussions in this specific field of study, comparison with the behavior of oil palm fibers was considered as the most appropriate option. Figure 2a also reveals that the maximum values of water absorption for OPS range from 17.4% to 31.2% (with soaking medium of 0.5w L kaolin to water-soaked sample respectively), as compared to values of water absorption (up to 39%) recorded during the initial stages of this research, presented in Loi et al. [28].
Minerals 2020, 10, x FOR PEER REVIEW 6 of 13 explanation, relating the increasing water absorption to water uptake by capillary action in micropores and surface fibers [25,26]. Thus, a potentially swollen OPS surface is prone to micro-cracks, which can further escalate the capillarity and transport of water; as also explained by Dhakalet [27]. The trends and observations from this experimental study agree with trends presented by Sreekala et al. [26]-A study on the water absorption behavior of oil palm fibers. With very limited data and research discussions in this specific field of study, comparison with the behavior of oil palm fibers was considered as the most appropriate option. Figure 2a also reveals that the maximum values of water absorption for OPS range from 17.4% to 31.2% (with soaking medium of 0.5wL kaolin to watersoaked sample respectively), as compared to values of water absorption (up to 39%) recorded during the initial stages of this research, presented in Loi et al. [28].  Observations from Figure 2a also highlight the fact that that the OPS immersed in kaolin slurry prepared at 0.5w L showed lowest water absorption values as compared to other mediums, with OPS immersed in water recording the highest water intake values. The average water absorption values observed are 14.3%, 24%, 25.5%, and 26.1% for kaolin prepared at liquid limits of 0.75w L , 1w L , and 2w L , respectively. This shows that increasing the water content of kaolin will result in increased water intake of OPS. This behavior clearly relates to water content in the kaolin samples. Water within the soil voids accelerates and catalyzes the diffusion of water between the soil medium and dry OPS. Similar observations are presented by Eskander and Saleh [29]. The lower water absorption of OPS in 0.5w L kaolin samples is partly attributed to low permeability of the compacted kaolin samples, because permeability of clays typically is inversely proportional to the degree of saturation. Thus, low permeability potentially retards the diffusion or movement of water between soil and OPS samples, which in higher w L (≥1) would be the other way.
The results of the water absorption test for OPS also clearly suggest the possibility of higher biodegradation, due to the presence of water. This potentially is unavoidable without surface treatment, due to the higher water absorption nature of OPS. Hence, further studies were conducted using rubber coating of OPS. Figure 2b,c shows the water absorption behavior of ROPS SL and ROPS DL samples. Observation from ROPS DL samples shows that the water absorption rate reduced drastically in comparison to OPS-kaolin samples, showing a decrease of up to five times using kaolin prepared at 0.5w L as a soaking medium. Moreover, providing a second layer of rubber coating reduced the water intake rate by at least 5% in all mediums. Overall, both ROPS SL and ROPS DL have lower water absorption as compared to OPS. For instance, Figure 2d gives a clear insight into the effect of coating, by comparing the OPS with ROPS SL and ROPS DL immersed in kaolin prepared at 0.75w L . The water absorption of ROPS DL reduced to as low as 2.6% when compared to water absorption of 22.5% observed in OPS. This is nearly an eight-fold decrease. Thus, the rubber sealant selected to coat the OPS aided in reducing water movement or diffusion, which theoretically signifies the capability of ROPS to perform as a relatively long term reinforcement for kaolin samples.

Specific Gravity of Kaolin Mixed with OPS and ROPS Sample
The variation in γ d values is predominantly related to the changing specific gravity (G S ). Thus, additional studies were conducted to determine the reduction in G S values of the kaolin-OPS and kaolin-ROPS DL samples. Table 4 therefore tabulates the G S values of every variation of the test samples used in this study. The values, as anticipated, follow a monotonic decrease depending on the composition of OPS or ROPS DL in the kaolin mixtures. The higher the OPS or ROPS DL , the lower the G s . Further, as evident from the table, there is a nearly 20% decrease in the G S values at higher OPS/ROPS DL contents. This clearly relates to the lower G S values of the OPS and ROPS DL samples. Nonetheless, lower G S values of kaolin-OPS and kaolin-ROPS DL encourage utilization of OPS and ROPS DL as lightweight construction material to substitute for conventional soil particles in geotechnical engineering works.  Figure 3 shows the compaction curves for the OPS-kaolin samples, where the OPS content varied from 0-30% by weight using both standard proctor compaction and mini-compaction test procedures. General observations show typical compaction curves, with dry unit weight (γ d ) values increasing with increasing water contents (w) till the optimum, followed by decreasing γ d with further increase in w. Observations show that the inclusion of OPS in kaolin decreases the maximum dry unit weight (γ dmax ) of the soil-OPS mix. For comparison, the γ dmax values for kaolin and kaolin reinforced with 30% OPS were 13.95 kN/m 3 and 12.65 kN/m 3 , corresponding to w opt values of 24.8% and 24.9% respectively. The trend of unit weight values is inversely proportional to the increasing OPS contents observed in this study for both standard proctor compaction and mini-compaction tests; which compare well with the literature [2,28,30,31]. The moderate reduction in the γ dmax values can be related to the substitution of soil mass with OPS, which has a lower specific gravity (see Section 3.2.1) in comparison to kaolin. Though γ d values show a clear decreasing trend with increasing OPS contents, the variation pattern of w opt is insignificant through different combinations of kaolin-OPS samples, especially for the standard proctor compaction test results. The w opt of kaolin-OPS for all the chosen percentage of OPS nonetheless ranges between 23.9% and 26.8% for standard proctor compaction tests and 24.4% to 25.5% for mini-compaction tests.

Compaction Characteristics of OPS-Reinforced Kaolin
Minerals 2020, 10, x FOR PEER REVIEW 8 of 13 Figure 3 shows the compaction curves for the OPS-kaolin samples, where the OPS content varied from 0-30% by weight using both standard proctor compaction and mini-compaction test procedures. General observations show typical compaction curves, with dry unit weight (γd) values increasing with increasing water contents (w) till the optimum, followed by decreasing γd with further increase in w. Observations show that the inclusion of OPS in kaolin decreases the maximum dry unit weight (γdmax) of the soil-OPS mix. For comparison, the γdmax values for kaolin and kaolin reinforced with 30% OPS were 13.95 kN/m 3 and 12.65 kN/m 3 , corresponding to wopt values of 24.8% and 24.9% respectively. The trend of unit weight values is inversely proportional to the increasing OPS contents observed in this study for both standard proctor compaction and mini-compaction tests; which compare well with the literature [2,28,30,31]. The moderate reduction in the γdmax values can be related to the substitution of soil mass with OPS, which has a lower specific gravity (see Section 3.2.1) in comparison to kaolin. Though γd values show a clear decreasing trend with increasing OPS contents, the variation pattern of wopt is insignificant through different combinations of kaolin-OPS samples, especially for the standard proctor compaction test results. The wopt of kaolin-OPS for all the chosen percentage of OPS nonetheless ranges between 23.9% and 26.8% for standard proctor compaction tests and 24.4% to 25.5% for mini-compaction tests.

Compaction Characteristics of ROPSDL-Reinforced Kaolin
Similar to OPS, both standard proctor compaction and mini-compaction tests were conducted on five different compositions of ROPSDL in the kaolin-ROPSDL samples. Figure 4 shows the compaction curves of ROPSDL-reinforced kaolin. In general, the experimental trends are comparable to that of kaolin-OPS mixes. The γdmax of the kaolin-ROPSDL samples for the compaction curves obtained from both standard proctor compaction and mini-compaction tests show a decreasing trend with increasing ROPSDL content. For compaction curves obtained from standard proctor compaction tests, the γdmax decreases from 13.95 kN/m 3 to 12.97 kN/m 3 and 12.54 kN/m 3 at 20% and 30% ROPSDL contents, respectively, while wopt shows a decreasing trend from 24.8% to 23.9% and 22.9%, respectively. More importantly, the compaction curve at optimum exhibits a downward-leftward shift over the γd-w dimension, indicating a significant reduction in both γdmax and wopt, as highlighted in Figure 4a,b. Decrease in wopt values with increasing ROPSDL content is mainly attributed to the hydrophobic nature and low water absorption capabilities of the ROPSDL, particularly because, for ROPSDL, the surface fibers have been covered by the rubber coating, thus resulting in lesser water intake. The increasing proportion of ROPSDL therefore replaces the kaolin (which would have

Compaction Characteristics of ROPS DL -Reinforced Kaolin
Similar to OPS, both standard proctor compaction and mini-compaction tests were conducted on five different compositions of ROPS DL in the kaolin-ROPS DL samples. Figure 4 shows the compaction curves of ROPS DL -reinforced kaolin. In general, the experimental trends are comparable to that of kaolin-OPS mixes. The γ dmax of the kaolin-ROPS DL samples for the compaction curves obtained from both standard proctor compaction and mini-compaction tests show a decreasing trend with increasing ROPS DL content. For compaction curves obtained from standard proctor compaction tests, the γ dmax decreases from 13.95 kN/m 3 to 12.97 kN/m 3 and 12.54 kN/m 3 at 20% and 30% ROPS DL contents, respectively, while w opt shows a decreasing trend from 24.8% to 23.9% and 22.9%, respectively. More importantly, the compaction curve at optimum exhibits a downward-leftward shift over the γ d-w dimension, indicating a significant reduction in both γ dmax and w opt , as highlighted in Figure 4a,b. Decrease in w opt values with increasing ROPS DL content is mainly attributed to the hydrophobic nature and low water absorption capabilities of the ROPS DL , particularly because, for ROPS DL , the surface fibers have been covered by the rubber coating, thus resulting in lesser water intake. The increasing proportion of ROPS DL therefore replaces the kaolin (which would have absorbed and held water in pores) with an equivalent volume, which leads to an additional decrease in values of w opt . The reduction in γ dmax values can be attributed to the lower specific gravity of kaolin-ROPS DL samples (see Section 3.2.1). Hence, the replacement of kaolin with increasing ROPS DL proportion is expected to reduce the specific gravity, which further leads to the reduction in γ dmax values. It should be noted that there are studies which also relate the decrease in γ dmax values to the loss of compaction efficiency caused by the elastic response of rubber during compaction [32]. But in this study, the inclusion of ROPS DL caused only a 0.8% decrease in γ dmax values as compared to the γ dmax values of OPS, which is negligible. Overall, the obtained compaction studies prove reliable, since the observed compaction characteristics from this study are consistent and in good comparison with similar literature [32][33][34][35][36].
Minerals 2020, 10, x FOR PEER REVIEW 9 of 13 absorbed and held water in pores) with an equivalent volume, which leads to an additional decrease in values of wopt. The reduction in γdmax values can be attributed to the lower specific gravity of kaolin-ROPSDL samples (see Section 3.2.1). Hence, the replacement of kaolin with increasing ROPSDL proportion is expected to reduce the specific gravity, which further leads to the reduction in γdmax values. It should be noted that there are studies which also relate the decrease in γdmax values to the loss of compaction efficiency caused by the elastic response of rubber during compaction [32]. But in this study, the inclusion of ROPSDL caused only a 0.8% decrease in γdmax values as compared to the γdmax values of OPS, which is negligible. Overall, the obtained compaction studies prove reliable, since the observed compaction characteristics from this study are consistent and in good comparison with similar literature [32][33][34][35][36].

Comparison of Results Obtained from Standard Proctor Compaction and Mini-Compaction Tests
There are frequent debates about the varying compaction behaviors/characteristics evaluated using standard proctor compaction and mini-compaction test procedures. The variations are typically related to the way soil particles rearrange themselves owing to varying rammer size and compaction efforts [15]. This study also presents a detailed insight and compares the results obtained from both standard proctor compaction and mini-compaction tests. Figure 5 shows the compaction curves of kaolin and OPS/ROPSDL-reinforced kaolin samples obtained using both compaction procedures. As can be clearly observed in Figure 5a, there is no significant variation between the compaction curves from standard proctor compaction and mini-compaction test procedures. Similarly, even in the case of ROPSDL-reinforced kaolin as shown in Figure 5b, the mini-compaction results observed in this study are comparable and consistent with the one obtained from standard proctor compaction test procedure. Hence, based on the results obtained in this study, it can be concluded that the mini-compaction procedure using 1 kg rammer involving 36 blows/layer of compaction frequency, proposed by Sridharan and Sivapullaiah [14], can be considered suitable for both kaolin and kaolin reinforced with OPS and ROPSDL. This can be used as an alternative testing method to estimate the compaction characteristics over a shorter time duration compared to the standard proctor compaction test. However, due consideration is crucial owing to the smaller diameter mold used in the mini-compaction procedure with OPS and ROPS samples (maximum dimension extending up to 15 mm). For this reason, the mini-compaction procedure is advisable as a quick test to estimate the compaction behavior of kaolin-OPS and kaolin-ROPS samples.

Comparison of Results Obtained from Standard Proctor Compaction and Mini-Compaction Tests
There are frequent debates about the varying compaction behaviors/characteristics evaluated using standard proctor compaction and mini-compaction test procedures. The variations are typically related to the way soil particles rearrange themselves owing to varying rammer size and compaction efforts [15]. This study also presents a detailed insight and compares the results obtained from both standard proctor compaction and mini-compaction tests. Figure 5 shows the compaction curves of kaolin and OPS/ROPS DL -reinforced kaolin samples obtained using both compaction procedures. As can be clearly observed in Figure 5a, there is no significant variation between the compaction curves from standard proctor compaction and mini-compaction test procedures. Similarly, even in the case of ROPS DL -reinforced kaolin as shown in Figure 5b, the mini-compaction results observed in this study are comparable and consistent with the one obtained from standard proctor compaction test procedure. Hence, based on the results obtained in this study, it can be concluded that the mini-compaction procedure using 1 kg rammer involving 36 blows/layer of compaction frequency, proposed by Sridharan and Sivapullaiah [14], can be considered suitable for both kaolin and kaolin reinforced with OPS and ROPS DL . This can be used as an alternative testing method to estimate the compaction characteristics over a shorter time duration compared to the standard proctor compaction test. However, due consideration is crucial owing to the smaller diameter mold used in the mini-compaction procedure with OPS and ROPS samples (maximum dimension extending up to 15 mm). For this reason, the mini-compaction procedure is advisable as a quick test to estimate the compaction behavior of kaolin-OPS and kaolin-ROPS samples.  Figure 6 shows the variations of compaction characteristics evaluated with the OPS content in kaolin before and after double layer rubber coating. In general, the higher the OPS or ROPSDL content in the kaolin mixtures, the lower the γdmax values. Figure 6a further shows that the γdmax values follow a gradually decreasing trend for both cases, with kaolin-ROPSDL samples having a higher tendency to decrease, as signified by the relatively higher value of Δγdmax/ΔROPSDL equal to −0.045, as against Δγdmax/ΔOPS of −0.029. Rubber coating likely cause a mild decrease in the values of Δγdmax compared to that of OPS, which is still within an acceptable limit of less than 1%. Overall, reduction in the γdmax values basically relates to lower GS values of the OPS and ROPS, with loss of compaction efficiency due to rubber coating playing only a minor role, to our understanding. Nonetheless, further studies are suggested in this research focus.    Figure 6 shows the variations of compaction characteristics evaluated with the OPS content in kaolin before and after double layer rubber coating. In general, the higher the OPS or ROPS DL content in the kaolin mixtures, the lower the γ dmax values. Figure 6a further shows that the γ dmax values follow a gradually decreasing trend for both cases, with kaolin-ROPS DL samples having a higher tendency to decrease, as signified by the relatively higher value of ∆γ dmax /∆ROPS DL equal to −0.045, as against ∆γ dmax /∆OPS of −0.029. Rubber coating likely cause a mild decrease in the values of ∆γ dmax compared to that of OPS, which is still within an acceptable limit of less than 1%. Overall, reduction in the γ dmax values basically relates to lower G S values of the OPS and ROPS, with loss of compaction efficiency due to rubber coating playing only a minor role, to our understanding. Nonetheless, further studies are suggested in this research focus.  Figure 6 shows the variations of compaction characteristics evaluated with the OPS content in kaolin before and after double layer rubber coating. In general, the higher the OPS or ROPSDL content in the kaolin mixtures, the lower the γdmax values. Figure 6a further shows that the γdmax values follow a gradually decreasing trend for both cases, with kaolin-ROPSDL samples having a higher tendency to decrease, as signified by the relatively higher value of Δγdmax/ΔROPSDL equal to −0.045, as against Δγdmax/ΔOPS of −0.029. Rubber coating likely cause a mild decrease in the values of Δγdmax compared to that of OPS, which is still within an acceptable limit of less than 1%. Overall, reduction in the γdmax values basically relates to lower GS values of the OPS and ROPS, with loss of compaction efficiency due to rubber coating playing only a minor role, to our understanding. Nonetheless, further studies are suggested in this research focus.    Figure 6b shows the variation of w opt for kaolin samples reinforced with OPS and ROPS DL samples. As evident from the figure, kaolin-OPS samples show a contradicting trend compared to kaolin-ROPS samples. This is clearly evident from ∆w opt /∆OPS value of 0.035, signifying a positive trend, whereas ∆w opt /∆ROPS DL results for kaolin-ROPS samples show a decreasing trend with a gradient of −0.12. The surface treatment of OPS with rubber significantly reduces its water absorption potential, which is reflected well with w opt values showing a decreasing trend with increasing ROPS DL content in the kaolin, whereas the OPS-kaolin shows higher water absorption with increasing OPS content, which is related to the hydrophilic nature of exposed surface fibers on OPS [24]. To this end, the proposed rubber coating of OPS shows considerable improvement in terms of w opt , and a moderate yet acceptable trend in the case of γ dmax with increasing ROPS DL content in kaolin.

Conclusions
Oil palm shell (OPS) is one of the biomasses produced in palm plantations whose improper disposal can cause a nuisance to the surrounding environment. And whose incineration may contribute to greenhouse gases. Thus, this experimental study presents a novel attempt to evaluate the suitability of OPS and rubberized OPS samples for soil stabilization in the construction industry. For this purpose, a series of laboratory experiments were conducted on OPS and rubber-coated OPS samples to assess their water absorption and compaction characteristics when mixed with kaolin. Following are the key conclusions derived based on the detailed observations and discussions presented in this paper:

1.
High water absorption of OPS is generally detrimental to the OPS, leading to material degradation. Thus, the less the water's interaction with the OPS biomass, the less the degradation. To this end, surface coating using rubber sealant is deemed to be a favorable and eco-friendly alternative. The water absorption of the ROPS in water reduced to as low as 1.5% of the OPS, which recorded the water absorption of 31.2%. The water movement or diffusion was observed to be less when OPS was coated with rubber sealant, thereby signifying the capability of ROPS DL to perform as a relatively long term soil reinforcement for kaolin samples. However, more research evaluating the performance ROPS DL and exploring other options for surface treatment is strongly advised.

2.
For kaolin-ROPS samples, the compaction curves obtained using both standard proctor compaction and mini-compaction tests show γ dmax to decrease with increasing OPS contents. This is related to the substitution of kaolin particles with the equivalent volume of OPS. The lower specific gravity of OPS leads to decrease in γ dmax . However, variation of w opt is insignificant in this case.

3.
For kaolin-ROPS DL samples, the compaction curve at optimum exhibits a downward-leftward shift over the γ d-w dimension, indicating a significant reduction in both γ dmax and w opt . The lower specific gravity along with minor loss of compaction efficiency due to the elastic response of rubber on OPS during compaction could have led to this reduction in γ dmax , while the lower w opt clearly relates to reduced water absorption capability of ROPS DL samples.

4.
This study also evaluated the possibility of using the mini-compaction test (a relatively quick test) to estimate the compaction characteristics of shell-like material in soil. The mini-compaction test results were compared with the standard proctor compaction test. The compaction curves for OPS and ROPS DL -reinforced kaolin samples obtained using mini-compaction apparatus showed a comparable and consistent results with the results abstained using standard proctor compaction tests. Results from this study therefore suggests that the mini-compaction test procedure proposed by Sridharan and Sivapullaiah [14] can be considered to estimate the compaction behavior of kaolin-OPS and kaolin-ROPS samples.

Conflicts of Interest:
The authors declare no conflict of interest.