Properties of cement–bentonite mortar developed by mineral additives for primary firm secant pile by Taguchi method

To make construction materials more sustainable and greener, companies must consider the environmental impact when sourcing materials include developing methods for sustainable recycling and encouraging design practices that prioritize the use of green materials using additives. Geotechnical applications including construction of cutoff walls require development of enhanced materials that possess certain attributes to ensure strain compatibility with surrounding soil. The study aims to explore the utilization of different materials such as bentonite, cement dust, fly ash, lime and polypropylene fibers to produce cement–bentonite mortar with low permeability and sufficient strength. For optimizing the process parameters in the experimental domain, an orthogonal array by Taguchi method was used, and thirty-two experimental runs were performed. The properties investigated included flow%, displaced volume rate, compressive, splitting tensile, flexural, shear strengths, elastic modulus and permeability coefficient. The test results demonstrated that cement dust, fly ash and lime could achieve optimal performance in terms of low permeability and sufficient strength. Additionally, polypropylene fibers up to 0.3% could be effectively used to achieve sufficiently low elastic modulus without affecting other conductive strengths significantly. The findings of the regression model demonstrated that the developed models could account for how the independent variable affected the necessary responses. This study could provide engineers with insights into selecting the appropriate materials to achieve the desired performance characteristics for some geotechnical applications considering sustainability.


Introduction
In order to prevent the ground and groundwater from entering between the reinforced concrete "male" piles in a secant pile wall assembly, the secondary "male" pile's bore must be cut into the concrete of the "primary" female pile.Concrete with a sturdy pile, in other words filling the space between the "hard" secondary piles, is "firm" primary pile concrete, which is weaker than secondary pile concrete.To allow the secondary pile to be carved into the surrounding piles, the concrete mixture must be made to be more elastic [1,2].
The desired results are achieved in the construction of piles by determining the lowest characteristic strength of the initial concrete, as it plays an important role in determining the accuracy and efficiency of the cutting process and the verticality of the piles.The strength of primary pile concrete can vary widely due to the intended long-term function of the pile, potentially ranging from 0.5 MPa for nonstructural piles to around 40 MPa for structural piles.It is one of the various elements that can affect the accuracy and efficiency of the cut and pile verticality, according to research evidence [1].
Slurry walls, made of traditional or plastic concrete, are currently one of the most popular geotechnical techniques and are frequently employed for both structural and hydrogeological purposes (cutoff walls).The latter are commonly used in many different civil engineering applications, such as tunnels, dams, reservoirs and dewatering facilities, and are designed to prevent and/or improve the flow of groundwater [3].
Plastic concrete is a sort of substance that behaves in a manner that is roughly halfway between that of typical concrete and that of a soil component.Furthermore, in geotechnical applications, plastic concrete regarding cutoff walls is buried inside the earth.There are two key qualities that are necessary.First, they should be sturdy enough to support the weight of the dam body, and second, they should have a low elastic modulus that blends in with the surrounding natural soil [4][5][6][7].It is strongly recommended that the cutoff wall's elastic modulus be approximately four times that of the surrounding soil [8].If the elastic modulus of a cutoff wall is much higher than that of the surrounding soil, both the soil and the wall may lack load-bearing capacity for unified action.This can result in the cutoff wall cracking and losing its performance [8].Normal concrete, which has a high elastic modulus, is not typically used for constructing cutoff walls due to this reason.Since the elastic modulus and the strength of concrete are intimately correlated with one another, an increase in the strength of concrete also results in an increase in the elastic modulus [5,8,9].To ensure the adequate load-bearing capacity of retaining structures like cutoff walls and sheet pile walls, designers consider factors such as soil properties and end bearing capacity.
Barrier walls have a specification that limits the maximum permeability to 1 × 10 -8 m/s [2,10], and the unconfined compression tests were recommended [11].
Bentonite is a clay mineral which exhibits a little bit of swelling behavior and obeys pozzolanic properties [12,13].It seems that there have been extensive studies on the factors that may affect the strength and deformability parameters of a material include mix design, restricting pressure, age and curing conditions of concrete/mortar specimens.The studies analyzed the failure modes, stress-strain curves, peak stress and failure criteria of the concrete specimens using various testing methods such as triaxial tests.Pisheh et al. [14] studied the mechanical and engineering characteristics of plastic concrete including bentonite content using triaxial compression tests and found that the concrete specimens containing bentonite showed higher ductility and energy absorption capacity compared to control specimens.
Researchers have been actively working on creating inexpensive, effective additives and recycling materials to enhance the mechanical properties and lessen the hydraulic conductivity of cutoff walls for the past ten years.These materials include blast furnace slag, cement bypass dust, fly ash, lime, waste tire powder, zeolite and activated carbon.By incorporating these materials, researchers hope to create an economical cutoff wall with properties that are well suited for various geotechnical applications [2,[15][16][17][18][19][20].
Laboratory tests were conducted to enhance the mechanical properties of various soil types in geotechnical applications [21,22].In addition, a nonlinear regression analysis was performed to establish the exact relationship between the friction angle of the soil and the characteristics of the soil particles [23].
The modified plastic concrete suggested by He et al. [15] that is used for cutoff walls was made up of 7.7% silica fume, 77.5% clay, 6.8% bentonite and 8% cement (dry component).The resulting values for this mixture included a coefficient of permeability of 2.36 × 10 -9 m/s, a compressive strength of 0.896 MPa and an elastic modulus of 1244.5 MPa [15].
Cement kiln dust (CKD) is a fine powder material that resembles cement and it can be parted and restored to the kiln through a dust collection procedure.Studies have reported variations of CKD quantity generated, with some studies showing up to 25% of CKD being produced.Recycling more CKD to the kiln would, therefore, result in less net CKD generation, which, in turn, would lead to a reduction in the dust quantity.There are many potential beneficial applications of CKD that could reduce the amount of waste generated from cement production or alternatively the total quantity of dust produced can be recycled or placed in a landfill.The average amount of CKD produced per ton of clinker was between 0.06 and 0.07 ton [16].
There is limited research available on the qualities of cement mortars made with CKD and supplementary cementitious materials for use in infrastructure [17].However, in order to build a cutoff wall that is economical and has beneficial geotechnical properties, it is possible to use cement bypass and include CKD in soils with a low plasticity index to enhance the unconfined compressive strength.These findings are discussed in reference [2].The effect of adding CKD to soils with a low plasticity index on the unconfined compressive strength of the soil can substantially improve its performance [16,18].The use of CKD has been found to be an effective stabilizer for improving the strength and permeability of dune sands when added to cement mixtures.Studies have shown that CKD can reduce the permeability of dune sands more significantly than cement alone.In addition, CKD can offer a notable economic advantage as an alternative to cement in stabilizing collapsing soil due to its ability to reduce the collapse potential of treated soil.These findings suggest that the use of CKD can be a viable alternative for soil stabilization, with potential benefits such as reduced costs and improved performance [16].
Produced low strength with lime-cement for nonstructural concrete has many geotechnical applications [19].The results confirmed that lime-cement concrete incorporating rubber powder up to 10% could be significantly used without any changes to its geotechnical properties and strength, even when exposed to ground moisture.However, further research may be needed to evaluate its behavior in a variety of conditions [19].
The study conducted by Zhao et al. [24] demonstrated that bentonite and fly ash are useful for improving the longterm strength of designed cementitious composites, which can help to become more sustainable.
On the other hand, the incorporation of polyvinyl alcohol fiber volume fraction in plastic concrete can improve its ductility and toughness properties, thus saving the rigid sections from destruction completely.Additionally, the study found that the use of fiber (~ 0.2%) in plastic concrete maintained its strength and did not rapidly droop when exposed to varying loads [25].
Tan et al. [26] conducted a study to investigate the impact of silica fume, fly ash and bentonite on cement-based grouts bleeding using the Taguchi method.The conclusion from the results showed that the Taguchi method was a promising tool to achieve the ideal conditions for cement-based grouts.
Limited studies on the simultaneous use of bentonite, cement dust, fly ash, lime and polypropylene fibers as well as the use of design of experiment [27][28][29] in the use of these materials with cement mortar.
In the current study, the influencing factors are highlighted and conducted.By investigating ideas that have not been extensively researched previously, researchers can generate new insights and solutions to address issues or challenges in various fields.This can help to improve the understanding of different phenomena and provide opportunities for further research and development.Additionally, identifying gaps in literature can lead to innovative research ideas and contribute to the advancement of knowledge in different domains.The experiments of this work are designed by Taguchi method and depict the performed test results to evaluate the influence of various parameters on the behavior of cementitious mortar in fresh and hardened states of the corresponding specimens.

Research significance
As the sustainability of the construction materials is being nowadays essential when developing and selecting the construction materials, the authors carried out this study to develop enhanced cement-bentonite mixtures that can effectively satisfies the varied geotechnical applications.Several materials may be considered to enhance the performance of traditional cement-bentonite mixtures.In order to enhance the performance, cement bypass dust, hydrated lime, fly ash, and polypropylene fiber (PP) were utilized.Each material was chosen based on its believed contribution: cement bypass dust and fly ash, as recycled fine materials, were expected to minimize permeability and benefit from the chemical reaction due to the contained oxides; hydrated lime, as a fine material, aimed to minimize permeability and enhance ductility; and polypropylene fiber, as a reinforcement, was expected to enhance ductility and reduce permeability.To optimize the target mixtures considering the pervious proposed variables with different levels, very large numbers of experiments need to be carried out which is both time-and money-consuming.The Taguchi method was used to minimize number of trials to save money and time provided that it can yield the same conclusion.The main objective of this work involves evaluating the effect of utilizing a combination of bentonite, cement dust, fly ash, lime and polypropylene fibers on the performance of cement mortar.The study focuses on developing regression models that can effectively represent the fresh and hardened properties of cementitious bentonite mortar.The test results of this study could potentially help engineers and researchers in the field of geotechnical applications to optimize cement-based materials that contain bentonite and recycled materials.

Experimental work
An experimental plan to evaluate the characteristic of modified cement-bentonite mortar by different additives is designed.

Materials
All used materials in this investigation were purchased from the local Egyptian marketplace and specified as the following section.
Cement: Portland cement type (CEM I 42.5N) was used in this work.The Portland cement used agrees to the Egyptian Standard Specifications (ES) (4756-1/2013) and EN 197-1/2011 requirements and obtained from Suez Cement Company.
Bentonite: A greenish-yellow sodium bentonite powder was used comply with API-13A (American Petroleum Institute), and it seems to be the most practical and beneficial to use as reported in the previous work [30][31][32].
Cement dust: Cement bypass dust obtained from Suez Cement Company was used.
Fly ash: Class F fly ash (FA) was used as low-calcium FA specified by ASTM C618-08.
Hydrated lime: Hydrated lime in accordance with the requirements of ASTM C 207 type N and EN 459-1 was used.Table 1 gives the chemical compositions, and physical and mechanical properties of cementitious materials used.
Fine aggregates: Well-graded sand of fineness modulus 2.3 was used and followed to the ES (1109-2002) requirements (zone II-medium sand).

Page 4 of 19
Fibers: Polypropylene fiber (PP) (FIBERMESH) with 18 mm and 0.04 mm in length and equivalent diameter, respectively, was used.The average tensile strength, modulus of elasticity of the PP fiber and the ultimate strain were 600 MPa, 6 GPa and 8%, respectively, as given by the supplier.

Specimen preparation and curing
The proportions of the mortar samples were mixed using a 60-L pan mixer.All dry powders (cement, bentonite, cement dust, fly ash and lime) were first mixed for 1 min.The appropriate needed amount of water is gradually applied to the dry powder materials and continue mixing for 3 min.Then, sand was combined and continuously mixed for an additional 3 min.Then PP fibers were gradually added to the mixture and continued mixing until the mixture becomes glossy and well blended (about 3 min).Finally, after mixing, the cement mortar specimens were prepared for flowability and penetration tests.After that, all specimen molds were cast for hardened tests (see Table 2).
All the specimens prepared and cast for the conducted hardened properties were stayed in their molds and covered with plastic sheets for two days after casting in laboratory at temperature (25 ± 2) °C and 55% RH; then, they removed from the molds and remarked with IDs remaining in ambient temperature at (25 ± 2) °C and 55% RH up to testing at 7 and 28 days.where C is the cement (kg/ m 3 ), B bentonite (kg/ m 3 ), D cement dust (kg/ m 3 ), F fly ash (kg/ m 3 ), L lime (kg/ m 3 ) and P PP fibers (%) Level Variables Tests properties

Fresh properties
The soil-bentonite mortar wall backfill material shall require a slump between 100 and 150 mm for optimal workability and efficient consistency in agreement with [2,33].In this study, water is supplemented to reach a practical high workability for in situ placement which is enough to be pumpable the flow % within the range (170-300%).
The fresh properties were evaluated by using the flowability and penetration methods.The consistency or flowability of the mortar mixtures was observed by the flow table test method in agreement with ASTM C1437.
The cone penetration test was applied by the hollow cone with own weight 370 g, 220 mm diameter and 180 mm height, and it was allowed to fall freely and statically penetrate the investigated mortar mixes according to [34,35].The depth of penetration was observed as depicted in Fig. 1.Based on the determined depths of penetration and the value of the related own weight, the displaced volume rate cm 3 /g was computed using Eq. ( 1) [34].
where v is the displaced volume, cm 3 , and h is the depth of penetration, cm.

Hardened properties
Mechanical properties Flexural strength and compressive strength of cement-based mixtures were determined by using prismatic specimens with dimensions of 40 × 40 × 160 mm 3 that were organized according to EN 196-1.The flexural strength was determined by using a three-point load bending test setup applied to each prism specimen, and the compressive strength was determined by applying an axial compres- Fig. 1 Tests setup sion load to the two halves of the prism specimen after the flexure test until failure (Fig. 1).Moreover, the splitting tensile test was conducted on cylindrical specimens 50 × 100 mm 2 based on Brazilian Standard NBR 7222 and ASTM C496 as illustrated in Fig. 1.Direct shear test was performed on standard prismatic samples 40 × 40 × 160 mm 3 as shown in Fig. 1.The compression, flexure, split and shear tests were performed using digital hydraulic compression-flexure machine with 300 kN capacities for all mortar specimens at the age of 7 and 28 days.
Alternatively, the modulus of elasticity test was performed using cylindrical specimens with 50 × 100 mm 2 according to ASTM C469 using a 300 kN capacities Universal Testing Machine with data acquisition system.The test results were based on at least the average of three test specimens for each test age.

Permeability test
The free-falling head permeability test is a widely used testing method that aims to establish the permeability of fine-grained soils or/and materials with medium to minimal permeability, such as clays and silts.The test requires connecting a short sample to a graduated standpipe, which is then used to supply the head of water and measure the amount of water flowing through the sample.The mortar samples were first saturated with water.The sample must be fully saturated before initiating the test [2,15].Permeability test is carried out on saturated mortar specimens with dimensions 60 mm in diameter and 30 mm in height (Fig. 1).The coefficient of permeability is computed for mortar samples by standard falling head cell conforming the Egyptian Code (ECP) 202-2001, Eq. ( 2) [15]: where: k: permeability coefficient, cm/S, a: cross-sectional area of the water pipe, cm 2 , l: length of the mortar specimen, cm, A: cross-sectional area of the mortar sample, cm 2 , T: t1-t2, S, h 1 : the distance to the bottom of the flask before the test, cm, and h 2 : the distance to the bottom of the flask before the test, cm.

Experimental plan
Thirty-two different mortar mixtures were cast and tested to cover the experimental plan as presented in Table 2.In this investigation, six main parameters were managed as control variables, which are defined as: cement content (C), bentonite content (B), cement dust content (D), fly ash content (F), lime content (L) and polypropylene fibers % (P).The Taguchi method was used to design an experimental investigation related to cementitious bentonite mortar.The investigation used an orthogonal array (OA) L32 OA to estimate the effects of different variables on the characteristics of cementitious bentonite mortar, as specified in Table 2.The mixes proportions are specified in Table 3.
A statistical software program (Minitab V16) was used to create and analyze the experimental plan and the response data, respectively, as stated in references [27][28][29].Overall, the study aimed to explore the effects of different variables on the performance of the cementitious bentonite mortar.Using a full factorial design, this would lead to a total of 2048 (2 1 × 4 5 ) possible combinations being considered.However, allowing Taguchi, the samples could be managed into only 32 groups, and treating them as separate groups would still produce outcomes with the same degree of confidence.This suggests that it may be possible to reduce the number of combinations that need to be considered without sacrificing the accuracy or validity of the results, by applying Taguchi's approach which emphasizes maximizing the signal-to-noise ratio.
In the Taguchi technique, the mean square deviation (MSD) is commonly used to optimize the quality of a product or process by minimizing the deviation between experimental and required values (Ross, [36]).It is a measure of the variability of the data points around the mean value or target value and can be used to assess the quality of a process or product.
The aim is to minimize the variation from the desired value to enhance the quality of the design or procedure being investigated.The Taguchi method utilizes orthogonal arrays to determine the appropriate combination of input variable levels that will reduce the variation and improve the overall quality of the output.In this study, the key target is to maximize the flow, displaced volume rate, compressive strength, splitting tensile strength, flexural strength [24], and, consequently, "larger-is-better," and to minimize the shear strength, elastic modulus and coefficient of permeability [2,15], therefore, "smaller-is-better" quality performance were selected.
The MSD of the larger-the-better can be given by Eq. ( 3): The MSD of the smaller is better can be stated as Eq. ( 4): where y denotes the investigational detected value of the experiments and n is the number of times repetition of each experiment.
The signal-to-noise ratio (S/N) is commonly used in the Taguchi technique to evaluate the quality of a design or process.The value of the loss function is first calculated, and this value is afterward converted into a signal-to-noise ratio using a specific formula.The signal in the formula denotes the desirable value, typically the mean value, while the noise represents the undesirable value, typically the standard deviation.The S/N ratio is expressed in Eq. ( 5) and it is used to measure the amount of deviation performed in the performance.Higher values of the S/N ratio indicate better performance, while lower values indicate poor performance:

Flow and cone penetration test results
The test results of the flow% and the displaced volume rate (calculated by Eq. 1) are presented in Table 4 and shown  where C is the cement (kg/m 3 ), B bentonite (kg/m 3 ), D cement dust (kg/m 3 ), F fly ash (kg/m 3 ), L lime (kg/ m 3 ) and P polypropylene fibers (kg/m It is likely to isolate the effect of each (single) control variable or process parameter of the response at various levels because the experimental design plan is orthogonal.Regarding the determination of optimum cement-bentonite mortar variables for each response (single-response optimization) to estimate the reputation of the influence and contribution of each individual (single) control factor, the response table for mains from Taguchi method was used as presented in Table 5 and plotted in Figs. 4 and 5.The main effects of each control factor show the impact of each factor on the response.Despite the type of performance characteristics, the better performance is related to the largest corresponds.Therefore, the level with the highest mean is the best level of the process parameters (see bold italic values in Table 5).The deviation involving the maximum value and the minimum value (delta) of the response of main according to the results, the highest ranking controllable factor was the one with the maximum value, Rank 1.This means that the optimum level for the control factors is the one with the largest mean value.From Table 5, the optimum combination of flow% was obtained for (C2B1D4F4L4P1) (300 kg/m 3 cement, 50 kg/m 3 bentonite, 60 kg/m 3 cement dust, 60 kg/m 3 fly ash, 60 kg/m 3 lime and 0.1% PP fibers) which was recommended.The ratio of the influence of the controlling parameters affecting the flow% is C (13.4%), B (20.85%), D (17.77%), F (16.97%), L (22.07%) and P (8.94%), which means that the content of lime is the most significant effect on flow% (Rank 1) [37].
On the other hand, the optimal combination of displaced volume rate was also obtained for Experiment No. 20 (C2B1D4F1L4P1) (300 kg/m 3 cement, 50 kg/m 3 bentonite, 60 kg/m 3 cement dust, 0 kg/m 3 fly ash, 60 kg/m 3 lime and 0.1% PP fibers) which was recommended.The proportion of the contribution of the controlling factors affecting the displaced volume rate is C (10.89%), B (13.76%), D (25.6%), F (12.75%), L (10.61%) and P (26.40%), which means that the percentage of PP fibers is the most significant effect on displaced volume rate (Rank 1) [35].
From Figs. 4 and 5, the relationship between the bentonite and PP fiber contents and the flow % is inversely proportional.It is noted that the increase in the content of bentonite up to 125 kg/m 3 and 0.3% polypropylene fibers does not produce an appreciable change in this because of the sufficiency of the used mixing water for maintaining the workability of the mix in the range of 170-300%.

Compressive, splitting tensile and flexural strengths
The compressive, splitting tensile and flexural strengths are investigated at 7th day and 28th day, and the main target for these properties (the best performance) is the larger is better.The test results are given in Table 4 and depicted in Figs. 6, 7 and 8.The test results at 7 days show that the observed values of compressive strength of all the investigated mixtures satisfy the requirements of ECP 202 which stated that the required compressive strength at 14 days is in the range of 1.5-3 MPa and previous works [1,2,10].
It is clearly noted that for cement-bentonite mortar mixes, the compressive strength decreases with the higher content of bentonite at different ages, and these results agree with the previous results [2,14,24].The maximum recorded values are 5.7 MPa and 7.7 MPa, respectively, at 7 and 28 days, for Experiment No. 20 (C2B1D4F1L4P1).
Moreover, the effect of the considered control factors in the quality performance of the compressive strength can be arranged in an ascending order related to the percentages of contribution as C (4.53%) Rank 6, L (9.20%) Rank 5, P (9.48%) Rank 4, F (15.19%) Rank 3, D (28.15%) Rank 2 and B (33.45%) Rank 1.
Based on the search results provided [38], it is possible that the use of certain clay minerals, such as kaolin and bentonite, in cement-based composite materials can have an impact on the microstructure and performance of the resulting material.However, it is important to note that the information in the search results is not conclusive, and further research may be required to determine the exact effect of these factors on cement bonding and structure.
The same trend was observed for splitting tensile and flexural strengths which recorded 1 MPa and 3.9 MPa, respectively, at 28th day, for Experiment No. 20 (C2B1D4F1L4P1).
From the analysis given in Table 6 for compressive, splitting tensile and flexural strengths at 28th day, it can be specified that the optimum reflected combination of the control factor levels can be shortly given as C2B1D3F1L3P1, C2B1D4F4L4P4 and C2B1D4F1L4P4, respectively, with factor C (cement content), B (bentonite content), D (cement dust content), F (fly ash content), L (lime content) and P (polypropylene fiber %).Despite the type of performance characteristics, the better performance is related to the largest corresponds.Therefore, the level with the highest mean is the best level of the process parameters (see bold italic values in Table 6).
The maximum percentage contributions of the control factors influencing the compressive, splitting tensile and flexural strengths are 33.45%,51.66% and 40.76%, respectively, which mean the most significant effect on the performance is given for factor B (bentonite content) (Rank 1).
Figures 9, 10 and 11 show the main effects plot for compressive, splitting tensile and flexural strengths, respectively, at 28 days.From these figures, the relationship between the bentonite and the compressive, splitting tensile and flexural strengths is inversely proportional.It is noted that increasing the content of bentonite from 50 to 125 kg/m 3 produces an appreciable change up to 39.3%, 53.1% and 50.24% reductions for compressive, splitting tensile and flexural strengths, respectively.Fiber could improve the splitting and flexural strengths at high percentage (0.2%) because of interfacial connections between the cementitious composite matrix and fibers [39].

Shear strength and modulus of elasticity
The shear strength and elastic modulus were investigated at 7, 28 days and 28 days, respectively, and the main target for these properties (the best performance) is the smaller   4 and depicted in Figs. 12 and 13, respectively.It is clearly seen that the shear strength and modulus of elasticity decrease with the increase in the content of bentonite (as given in Table 6).The minimum documented values for shear strength are 0.01 MPa and 0.33 MPa, respectively, at 7 and 28 days, for Experiments No. 15 and 31 (C1B4D3F4L2P1 and C2B4D3F1L2P4).
From the analysis specified in Table 6, shear strength and modulus of elasticity at 28 days, it can be directed that the best blend of the conducted controlling factor levels can be presently given as (C2B4D1F1L1P1and C1B4D1F2L2P4), which taken into consideration the smaller is better to produce the best performance [2,15].
The maximum percentages of the contribution of the controlling factors influencing the shear strength and elastic modulus are 22.07%and 41.19%, respectively, which the strongest effect on the performance is specified for factor B (bentonite content) (Rank 1).
Figures 14 and 15 show the main effects plot, the shear strength and elastic modulus.From these figures, the relationship between the bentonite and shear strength and elastic modulus is inversely proportional.It is noted that increasing the content of bentonite from 50 to 125 kg/ m 3 produces an appreciable change up to 44.15% and 48.83% reductions for shear strength and elastic modulus, respectively.

Coefficient of permeability
The coefficient of permeability was investigated at 7 and 28 days, and the main target for this property (the best performance) is the smaller is better [2,15].The test results are given in Table 4 and plotted in Fig. 16.The minimum documented value for coefficient of permeability is 6.7 × 10 −7 and 1.02 × 10 -6 cm/s, respectively, at 7 and 28 days, for Experiments No. 4 and 28 (C1B1D4F4L4P4 and C2B3D4F2L2P4).It is obviously noticed that the permeability coefficient reduces with the increase of percentage using the polypropylene fibers (as given in Table 6).Ryan and Day [10] specified that the lowest value of coefficient of permeability for cutoff walls required for soil-cement-bentonite is 5 × 10 -6 cm/s.The test results show that the coefficient of permeability recorded values of all the investigated cement-bentonite mortar mixtures conformed to the requirement of ECP 202.
From the analysis detailed in Table 6, coefficient of permeability at 28 days, it can be concluded that the optimum combination of the conducted controlling variables and their levels can be presently given as (C2B3D2F3L4P4), which taken into consideration the smaller is better to produce the best performance.
The maximum proportion of the contribution for the controlling factors affecting the coefficient of permeability is 26.22%, which the strongest effect on the performance is specified for factor P (PP%) (Rank 1).
Furthermore, the effect of the considered control factors in the quality performance of the coefficient of permeability can be arranged in an ascending order related to the percentages of contribution as C (2.89%) Rank 6, B (11.57%) Rank 5, L (17.72%) Rank 4, D (19.2%) Rank 3, F (22.42%) Rank 2 and P (26.22%)Rank 1.
Figure 17 presents the main effects plot for the coefficient of permeability.It is observed from this figure that there is an inverse relationship between the volume fraction % of polypropylene fibers and the coefficient of permeability.It is noted that with the increase of the % polypropylene fibers from 0.1% to 0.3% produced an appreciable change up to 38.37% decrease.This results in a lower coefficient of permeability by using the fibers because the fibers cut and obstruct the capillary flow in the mortar mixes this result agrees with the results obtained by Li et al. [40].
However, due to the high volume and low strength, hexagonal crystals of CH are frequently surrounded by CSH gel and are difficult to directly detect because the original shape is easily distorted during the sampling procedure.Due to the use of mineral additives, a significant amount of gel was formed by the cement, which was continuously hydrated.On the other hand, it joined the particle fragments together to increase their solidity.However, the gel filled in the gaps to increase the compaction.So, it makes sense that adding additives to cement-bentonite mortar would increase its impermeability and compactness as seen in SEM image (Fig. 18).

Regression model results
It was found that the optimal combination of factors/levels for one property or response may differ from that of another property or response.This is because different properties or response may have different sensitivity or response to changes in particular factors or levels and may require a different combination of inputs or conditions for optimal results.In such cases, it is important to understand and optimize the factors or levels that are most relevant to the specific property or outcome of interest [27,41].This suggests that an engineering decision based on experience may be necessary for accepted decision-making in the cementitious bentonite procedure as facing similar cases.Based on the experimental test results given in Table 4, the models for fresh and hardened properties were formulated as presented in Tables 5 and 6, respectively, through Eqs.(6 and 7) and Eqs.(8)(9)(10)(11)(12)(13), respectively.The results suggest that a linear regression model (general regression analysis) can be used and found to be adequate for all responses.

Conclusions
Based on the work performed in this investigation and the analysis conducted, the following conclusions could be obtained: 1.The introduction of the Taguchi method in this study for designing experiments is simple and effective.As well, the number of experiments conducted was appropriate.2. Based on the Taguchi analysis (single-response optimization), the relationship between the bentonite and the hardened properties (compressive, splitting, flexural, shear strengths, modulus of elasticity and coefficient of permeability) is inversely proportional regardless of the target property required.Moreover, it is considered the maximum percentages contribution of the controlling factors influencing the properties.3. The study confirmed that cement dust, fly ash and lime can effectively be used to attain optimum performance in terms of minimal permeability and sufficient strengths proportional regardless of the cement content.Furthermore, polypropylene fibers can be effectively used to achieve optimal performance in terms of low permeability and sufficiently low modulus of elasticity without considerable effect on other conductive strengths.
259 Page 18 of 19 4. The proposed regression models have reasonable accuracy in predicting both fresh and hardened properties. 5.The regression models showed that adding cement dust, fly ash and lime can be helpful for expansion in a certain direction and serve many environmentally friendly geotechnical applications due to the remarkable improvement in properties, in addition to being cost effective.6.Furthermore, further investigations include validation of the durability and long-term properties and the microstructures of such composites are still needed, trying more recycled materials, combinations of recycled materials as well as hydraulic conductivity under triaxial tests.

Fig. 2 Fig. 3
Fig. 2 Test results of flow test for different mixtures

Fig. 6
Fig. 6 Test results of compressive strength for different mixtures at different ages

Fig. 10 Fig. 11
Fig. 10 Mean effects plot for splitting tensile strength at 28 days

Fig. 12
Fig. 12 Test results of shear strength for different mixtures at different ages

Fig. 16 Fig. 17
Fig.16 Test results of coefficient of permeability for different mixtures at different ages

Fig. 18
Fig. 18 SEM image of cementbentonite mortar samples modified with different additives at 28 days

Table 1
Properties of cementitious materials used (chemical compositions, physical and mechanical)

Table 2
Experimental plan

Table 4
Test results of experimental work at different ages

Table 5
Response table for means and analysis and regression model (single-response optimization), fresh properties

Table 6
Response table for means and regression model (single-response optimization), hardened properties at 28 days