The Impact of Incorporating Shell Aggregates and Coconut Fibres on the Physical, Thermal and Mechanical Properties of Cement Mortars

The accumulation of shellfish waste is an urgent environmental issue in most seafood-consuming countries. Challenges include poor biodegradability, the release of toxic gases during the decomposition of organic matter on the surface, and the significant financial burden on public waste management systems. Consequently, the recycling of these or other types of waste is the focus of interest for many researchers around the world. On the other hand, cement mortar is a mix of cement, sand and water, and is an essential composite in construction for masonry work. It is used to join construction elements together and to fill gaps between basic elements. Sand is a granular constituent that plays a structural role in mortar. There are two types: natural sand (from dunes, rivers, ...) and artificial sand, which comes from the disaggregation of rocks. The form and particle size of the sand are very significant for its use in construction. Shapes that are too round or too fine are not recommended, which means that desert sand cannot be used in construction. Due to the high demand for structural aggregates in construction, some countries use sand directly from beaches, the seabed or quarries adjacent to beaches under construction. This often leads to uncontrolled sand extraction, with no regard for the environmental consequences, causing beaches to disappear, deforming landscapes and salinizing groundwater.

Recycling shell waste to produce a partial sand replacement can therefore be a relevant and suitable solution to the two environmental problems mentioned above. Our work focuses on the reuse of this residue in building materials. Seashells are mainly made up of calcium carbonate CaCO₃, in very high percentages (over 90% by mass). When calcined at high temperatures, its chemical composition resembles that of limestone CaO [1, 2]. The presence of organic matter in the structure of the shell acts as a matrix that binds the aragonite tablets, forming structured layers with these two constituents. The shell is generally made up of three parts: the inner layer (the nacre), the intermediate layer (the prismatic) and the outer layer (the periostracum) [3].

Furthermore, natural fibres (coconut fibre, date palm fibre, alfa, bamboo, hemp, etc.) are becoming increasingly usable in construction worldwide [4-6]. This is thanks to their intrinsic properties, mechanical properties, low weight and cost. This use is suitable for several products, including lightweight load-bearing walls, insulation materials, floor and wall coverings, and roofing. In general, the fibres have a porous structure that improves the thermal properties of the materials. Coconut fibres have been studied by several authors used in cement mortar [7-11]. It has been found that when fibres are added, the materials become porous which affects the mechanical properties especially the compressive strength. This behavior is mainly due to the hydrophilic nature of the fibres.

Research to date has focused on the recycling of shell waste to formulate concrete [12-17] and mortars, either in the form of fine and coarse aggregates to replace sand and gravel respectively, or in the form of calcined powder [18-22], which can be used as an additive or replacement of cement. For mortars, Safi et al. [23] used size between (0-5 mm) of crushed shells as a replacement sand in the formulation of a self-compacting mortar (SCM) with percentages of 0%, 10%, 20%, 50% and 100% by weight. Ez-Zaki et al. [24] worked with mussel shells and glass powder as a substitution of sand in cement mortar with percentages of 20%, 40% and 60% by mass. They used a granulometry of (0-1 mm) for the shell sand. Martínez-García et al. [25] used a particle size distribution similar to that of natural sand (0-4 mm), by mixing the mass fractions of the two types of shell sand, fine (0-1 mm) and coarse (0-4 mm). They replaced the natural sand with the following percentages 25%, 50% and 75% by volume of shell sand. Merlo et al. [26] investigated the effects of substitute natural sand with Acanthocardia tuberculata shell sand on the physical and mechanical properties of mortars, using the following substitution rates 0%, 5%, 10% and 15% by mass, with a grading (0-3 mm) of shell sand. In general, the physical, mechanical and thermal properties of mortars and concretes are greatly affected by the grain size and shape of the shell aggregates [27].

Shell waste and coconut fibre waste are abundant in our region of Agadir in Morocco. They are discarded either by human activities or by natural phenomena, causing negative impacts on the environment, such as bad smells, attracting insects, water pollution, etc. The recycling of natural waste in the construction materials industry has become a major objective for researchers around the world. This approach aims to promote environmentally friendly construction, also known as eco-construction. It helps to reduce the impact of buildings on the climate and the environment. At the same time, it minimises the construction sector's dependence on increasingly scarce natural resources and recycles natural waste discarded into the environment. To this end, this study has been carried out to examine its feasibility as a building material.

The effect of combining the two types of waste in the same mix comes from a preliminary study in the literature. Several researchers have studied the effect of replacing natural sand with shell sand in mortars and concrete [14, 15, 17, 24]. They have observed that by increasing the replacement rate, porosity increases and mechanical strength is adversely affected, but an improvement in thermal properties is obtained. Moreover, in previous studies, the use of coconut fibres in mortar resulted in a slight increase in mechanical strength, particularly flexural strength [8, 9, 28]. Thanks to the mechanical properties of coconut fibres, which led to the idea of incorporating both types of waste in the same mix. Both materials have intrinsic properties that favour them. Assembling them in the same composite is an innovative idea that can add new results to the literature.

Following this documentary analysis, we observed that most authors use all of the crushed shell sand without extracting the fine fraction. This practice can result in a grain size distribution that is very different from that of natural sand. Natural sand generally contains only a small proportion of fines. In addition, the use of shell aggregates with natural fibres in the same formulation has not been studied in the literature. This choice makes it possible to combine the intrinsic properties of the two materials. For this, we chose an approach to study the effects of replacing natural sand with shell sand and coconut fibres on the properties of the mortar.

Our approach consists of using a grain size of (0.08-2 mm). This particle size was chosen to closely match the dimensional characteristics of the fine natural sand traditionally used in mortars. This choice ensures high compactness and homogeneity of the mixture. It also allows the specific effect of very fine shell particles on the mechanical properties of the mortar to be evaluated. In addition, the fine particles of crushed shell sand must be removed. This reduces the amount of fine organic matter it may contain. A high organic matter content acts as an air entrainment agent, creating more voids within the paste. On the other hand, it also acts as a setting retarder [29]. In parallel, the coconut fibres were used with a length of 4 cm. This length was selected for the coconut fibres because it is sufficient to create crack bridging effects and improve the mortar's resistance to bending and cracking without excessively compromising its homogeneity or workability. Most authors are more interested in the mechanical properties of cement mortars than in their thermal properties, so our study consists in determining the thermal behaviour as well. The thermal aspect is a very important axis that must be treated in order to estimate heat transfers within the building and to try to reduce energy requirements.

The purpose of this study is to evaluate the effect of partially replacing natural sand with shell sand (grain size 0–0.08 mm) and 4 cm of coconut fibres on the mechanical and physical properties of mortars. The goal is to determine whether this combination can offer performance equal to or superior to traditional formulations.

To this end, we have prepared cement mortars by substituting natural sand with shell sand and coconut fibres in percentages of (0%, 15%, 25% and 45%) and (0%, 8% and 16%), respectively, for the two types of incorporation. The study examined the physical properties of mortars in their hardened state, as well as their mechanical, thermal and capillary characteristics. In addition, microstructural analyses were carried out in order to establish a link between structural observations and measured properties.

3.1 Physical properties

Figure 12 shows the development of dry and saturated density of cement mortar and its errors determined by the difference between the maximum and average values for each composition. The density in the dry state, is determined by the quotient of the mass of the specimens which has been taken after drying them in an oven at 60℃ ± 5℃ to a constant mass, by the volume it occupies when immersed in water (measured by hydrostatic weighing). The saturated density is calculated after the dry mortars have been immersed in water, it being reached when two successive weighings, carried out 15 minutes apart during immersion, do not differ by more than 0.2% by mass EN 1015-10 [38].

Figure 12. Evolution of saturated and dry densities of cement mortars

Figure 12 shows that the dry density decreased as the rate of substitution of natural sand by shell sand increased, whereas this decrease became more significant as the rate of substitution by coconut fibres increased. On the other hand, there was an opposite trend in saturated density. The rate of reduction in the dry state is 7.78% and 9.31% for CM450 and CM4516, respectively, compared with CM00.

The water physically bound to the cement mortar evaporates completely after drying. As the rate of substitution increases, the cement mortar becomes more fluid and, as a result, the pores appear more intensely after drying (the density decreases). The quantity of water absorbed is considered to be linked to the specific surface area, also to the fineness of the particles, so the effect of shell aggregates on the fluidity of the paste varies according to these first three parameters. For example, Kuo et al. [39] noted that shells do not improve the workability of samples, since they used shell aggregates with a lower fineness modulus than natural aggregates, so shell particles absorb water more than sand. However, Martínez-García et al. [14] have attributed the reduced workability of concrete to the form of the laminated shell granulates, also to the presence of organic substances that increase the viscosity of the paste. The effect of reduced fluidity is also observed by several other authors [40-42]. The increase in fluidity observed during mixing of mortar in our experiment may also be due to the elimination of the fraction below 0.08 mm from the shell sand, which may contain more free fine organic substances and consequently the paste becomes less viscous compared with other studies using sand with this granular fraction (< 0.08 mm). The lower density observed in Figure 12 may also be due to the flaky shapes of the shell aggregates (Figure 5(b)) and their larger sizes (Table 1), which create more pores with the cement mortar. On the other hand, the effect of the reduction in dry density when coconut fibres are substituted is due to the low density of the fibres compared with natural sand (Table 1).

Water-accessible porosity P (%) is calculated using Eq. (4) according to Nguyen et al. [13], with m_dry the dry mass of the specimens after drying at 60℃ in the oven, m_wat mass of the specimens under water, V_t  volume of the specimen by hydrostatic weighing, and ρ_water density of the water.

$P(\%)=\left[1-\left(\frac{m_{d r y}-m_{w a t}}{\rho_{w a t e r} \times V_t}\right)\right]$              (4)

Figures 13 and 14 illustrate the variation in water-accessible porosity and water absorption by immersion of the specimens. Porosity increases with the percentage of substitution (of shell sand and coconut fibre), recording a value of 22.80% for CM00 and an increase of 56.58% and 66.93% for CM450 and CM4516, respectively.

Figure 13. Water-accessible porosity of different mortars

Figure 14. Absorption of water by immersion of different mortars

The absorption of water by immersion is determined by the ratio of the difference between the saturated and dry masses to the dry mass of the specimens as a percentage. This variable increases with the rate of substitution by shell sand, ranging from 9.72% for CM00 to 22.53% and 26.09% for CM450 and CM4516, respectively.

Figures 15 and 16 show the correlation between the absorption of water by immersion and the water-accessible porosity on the one side, and the correlation between the water-accessible porosity and the dry density of the specimens on the other side. They are given by the following equations, respectively:

$W(\%)=1.02 P(\%)-13.38$               (5)

$P(\%)=-87.79 \rho_{d r y}+194.47$                 (6)

We therefore deduce that the increase in the rate of substitution of natural sand by shell sand or coconut fibres at the same time affects the dry densities of cement mortars and consequently the other physical properties, since they are all linearly proportional to each other. The trend of increasing porosity and water absorption has been found by several authors [23, 25, 27, 39, 42].

Figure 15. Evolution of absorption of water by immersion as influenced by water-accessible porosity of different mortars

Figure 16. Evolution of water-accessible porosity as influenced by dry density of different mortars

3.2 Microstructure

Figure 17 shows the secondary electron images observed by scanning electron microscopy SEM of the fracture faces of CM00, CM250 and CM450 cement mortars for 0%, 25% and 45% substitution, respectively. From these pictures, we can clearly see the increase in pore rates within the volumes of the composites, which justifies the evolution of porosity presented in paragraph 3.1. The porosity is the natural consequence of the quantity of water added in addition to that required for hydration and any voids present in the aggregates. The increase in the rate of substitution of natural sand by shell sand resulted in an excess of mixing water in the mixture, due to low absorption of water capacity of shell sand compared with natural sand (Table 1). This quantity of water, which increases with the rate of replacement, evaporates when drying at 60℃, leaving pores inside the composites.

(a)

(b)

(c)

Figure 17. SEM Micrographs (Secondary Electrons) of the fracture faces of: (a) CM00, (b) CM250, (c) CM450

(a)

(b)

(c)

(d)

(e)

Figure 18. SEM Micrographs (Secondary Electrons) of the fracture faces of mortars incorporated with shell sand

(a)

(b)

Figure 19. SEM Micrographs (Secondary Electrons) of the fracture faces of mortars embedded with coconut fibres

The pores may also be due to the shape of the shell sand aggregates, as shown by the scanning electron microscopy images. We can clearly see in Figure 18(a) the weak adhesion between the smooth inner surface of the shell and the cement mortar due to the void created between the two. On the other side, in Figure 18(b) and (c), the pores may be due to the creation of cavities around the grains of flattened and elongated shell sand or due to the lower specific surface area compared with that of natural sand (Table 1), which results in a poor arrangement of the aggregates. However, there are cases where there is good adhesion between the shell grains and the cement mortar in cases where these grains are almost rounded, especially with the periostracum part Figure 18(d) and (e).

The presence and increasing percentages of coconut fibres contribute to the increase in pores in cement mortars. When the mortar is prepared, the fibres are first immersed in water, which causes them to swell. When the mortar dries, the fibres detach from the cement matrix, leaving cavities in the surrounding area as noted in Figure 19. On the other hand, there are cases where pores are created due to the entanglement of the fibres with natural sand or shell aggregates.

3.3 Thermal properties

The thermal tests were carried out on the dried prisms of dimensions (10 × 10 × 2.5 cm³) under Laboratory conditions. Multiple analyses were carried out for each composite in order to obtain an average value for the thermal properties of each sample.

Table 3 shows the thermal conductivity and diffusivity of mortars. These thermal properties decreased when natural sand was replaced by shell sand and/or coconut fibre. For the first three CM0 mortars this reduction is not very significant, but when replacing with shell sand, it becomes more noticeable.

Table 3. Conductivity and thermal diffusivity of mortars

Thermal conductivity serves as an indicator of a material's insulating capacity. By replacing 45% of the sand with shell waste, it is estimated that the material's thermal insulation improves by 21.70% and 23.57% for CM450 and CM4516 respectively compared to standard cement mortar.

Figure 20 demonstrates the variation in thermal conductivity as a function of dry density for different mortars. Thermal conductivity decreases as the density decreases, a proportionality between the two values is observed, given by the following equation:

$\mathrm{y}=1.70 \mathrm{x}-1.97$               (7)

Figure 21 illustrates how thermal conductivity varies with the porosity of composites. An increase in porosity corresponds to a decrease in thermal conductivity. The proportionality between these two parameters given by the following equation:

$y=-0.02 x+1.77$               (8)

We can conclude that the thermal behaviour observed during substitution can be directly explained by the increase in the porosity rate within the mortar volume, this increase in porosity is demonstrated, on the one side, by the decrease in the density of the composites (Figure 12) and, on the other, by the observation of fracture faces using scanning electron microscopy (Figure 17). The effect of improving thermal properties when shell aggregates are incorporated has been observed in various formulations [24, 31, 43] either in plasters, compressed earth blocks or cement mortars respectively.

Figure 20. Evolution of thermal conductivity as influenced by dry density of different mortars

Figure 21. Evolution of thermal conductivity as influenced by water-accessible porosity of different mortars

3.4 Mechanical properties

Mechanical properties are estimated by compressive and flexural tests after 28 days of specimen curing. Three specimens of each mortar mixture were used to determine the mean value of the multiple analyses. Figures 22 and 23 indicate compressive and flexural strength of different types of mortar and their errors, calculated as the difference between the maximum and average values.

Figure 22. Compressive strength evolution

Figure 23. Flexural strength evolution

Compressive strength decreased slightly with increasing levels of shell sand, but the decrease was greater with the addition of coconut fibres. The rate of reduction ranged from 20.44% for CM450 to 31.49% for CM4516 compared with CM00. Flexural strength is not greatly affected, but it can be seen that the replacement of 8% coconut fibre recorded values very close to the reference mortar, even in the presence of shell sand, especially for CM0, CM15 and CM25. This optimal behaviour can be explained by good fibre dispersion at this incorporation rate, allowing effective bridging of microcracks and better redistribution of stresses within the matrix. On the other hand, higher fibre contents lead to increased porosity and agglomeration phenomena, which reduce the effectiveness of the fibre-matrix interface and limit stress transfer. The reduction in strength observed in this study is less pronounced than that reported in the literature, probably due to the natural roughness of coconut fibres and shell aggregates, which promotes mechanical interlocking and partially compensates for the lower rigidity of shell sand.

These results can be mainly attributed to the porosity of the material, the smooth surface of the shell particles which limits their adhesion to the cement paste, and the irregular stacking of the aggregates resulting from their flattened and elongated shape (Figure 18).

These results are in concordance with those of other authors who have studied the effect of replacing natural sand with shell sand in cement mortars [23, 24, 33, 44, 45]. The rate of reduction in compressive strength for Martínez-García et al. [25] reached values between (60-70%) for 50% of the replacement. However, Woon et al. [44] achieved a rate of 54.88% for 50% replacing natural sand with sand from mussel shells. Ez-zaki et al. [24] estimated a reduction rate of 6.5% and 27% for 40% and 60% of the replacement respectively. These results seemed to be related to the granulometry used for the shell sand, note that Martínez-García et al. [25] used shell aggregates between 0 and 4 mm, while these studies used aggregates with a maximum size of less than 1 mm [24, 44]. Shells are known for their flattened shape, with a less efficient granular arrangement. Crushing minimises the flattening of particles while generating a wide range of granulometric fractions, then crushing plays a very significant role in the mechanical properties of composites, the larger the grains the more the shape of the shells is rather flattened and hollow and generates more pores and therefore the mechanical properties are affected by the study of Nguyen et al. [27]. Also, the elimination of the fraction < 0.08 mm may be one of the reasons why the mechanical properties were not greatly affected in our study, it is possible that this fraction contains fine organic substances.

The coconut fibres did not greatly improve the mechanical properties, but they did contribute in other ways to increasing the porosity of the mortar volume, which had a negative effect on compressive strength. This behaviour has been observed by several researchers, who have used coconut fibres as an additive or as a replacement in cement mortars [10, 11].

3.5 Water absorption through capillary action

Figure 24 shows the capillary water absorption coefficient of the different types of mortar. It can be seen that this coefficient decreases over time as the rate of substitution by shell sand increases, reaching a rate of 58.33% for CM450 compared to CM00. On the other hand, it can be seen that it increased when coconut fibre was substituted, reaching a rate of 30.56% for CM016 compared to CM00 and 40% for CM4516 compared to CM450. These composites can be classified as W₀ in accordance with EN 998-1 [46], except for composite CM4516 which is classified W₁; its coefficient is less than 0.4 Kg/m².min^0.5.

Figure 24. Capillary absorption coefficient of different mortar mixes

Martínez-García et al. [25] found the same trend when using shell particles as a partial substitution of sand in the formulation of cement mortars, noting that substitution of up to 75% by volume reduces this parameter to a value classified as W₁ according to EN 998-1. This type of waste can therefore be used in the formulation of insulating coatings to prevent water rising by capillary action.

Shell aggregates improve the durability of mortars, because of the flattened, smooth and lengthened shapes of the mussel shell grains, which act like a wall against the rising of water by capillary action. However, the presence of coconut fibres contributes to the fluctuation of this coefficient due to its hydrophilic nature.

The research activities were primarily carried out at the Laboratory of Thermodynamics and Energetics (LTE). Thermal property measurements were conducted at the Laboratory of Materials and Renewable Energies (LMER). Structural analyses, including X-ray diffraction and SEM, were performed at the Scientific Research Centre, Faculty of Sciences, Ibn Zohr University, Agadir, Morocco. Preparation of the mussel shells, including crushing and grinding, took place at the Laboratory of Mechanics, Processes, Energy, and Environment (LMPEE) at the National School of Applied Sciences, Ibn Zohr University, Agadir. Mechanical testing was completed at the Public Laboratory for Tests and Studies (LPEE), Agadir. The authors gratefully acknowledge the technical support provided by the staff of all mentioned laboratories, which was essential to the success of this work.