Optimizing mechanical performance and water absorption behavior of silica-reinforced recycled thermoplastic composites for rooftile applications

Abstract

This investigation was conducted to synthesize and characterize sustainable composite materials for construction applications—specifically roof tiles—by utilizing recycled thermoplastics (HDPE, PP, PET) reinforced with silica sand microparticles. This methodology offers the dual advantage of mitigating plastic-waste accumulation and providing a cost-effective alternative to conventional, resource-intensive construction materials. Employing a Taguchi L9 orthogonal array, the study systematically optimized the effects of thermoplastic type, filler weight fraction (30–50 wt.%), and particle size (150–450 µm) on mechanical performance and moisture-ingress resistance. Experimental analysis revealed that the composite comprising HDPE, 30 wt.% silica, and 150 µm particles (A1B1C1) exhibited superior energy-absorption capacity (impact strength: 5.83 J) and minimal water absorption (0.15%). Conversely, the most favorable mechanical integrity was achieved with the PET, 50 wt.% silica, and 300 µm particle configuration (A3B3C2), which maximized hardness (60.8 HRB), compressive strength (57.9 MPa), and flexural strength (49.56 MPa). Multi-response optimization via Grey Relational Analysis (GRA) confirmed that the A3B3C2 combination provided the most robust overall solution, yielding a maximum Grey Relational Grade of 0.833. Consequently, the successful fabrication of these optimized recycled thermoplastic composites demonstrates their strong potential for real-world use in roof-tile production, thereby supporting a more sustainable and resource-efficient construction industry.

Introduction

Composite materials provide several advantages over traditional construction materials, including high strength-to-weight ratio, corrosion resistance, and design flexibility. Plastics, in particular, have become widely used due to their lightweight nature, durability, and ease of processing. However, global plastic production has increased sharply in recent decades, with packaging accounting for a major proportion of consumption. Driven by its wide ranging applications, plastic production has grown exponentially, increasing from about 5 million tons in the 1950s to over 400 million tons in 2022, with global demand expected to quadruple by 2050. ¹ Plastics constitute approximately 8–12% of municipal solid waste, amounting to nearly 190 million tons generated annually. ² Due to their non-biodegradable nature, limited recycling rates, and challenges in waste management, plastic waste has emerged as a major environmental concern, contributing to pollution and ecological degradation.^1,3,4 The rise in plastic consumption is driven by population growth, low production costs, and the wide range of applications, ⁵ yet only a small portion is recycled, making sustainable plastic waste management a global priority. ⁶ In recent years, the construction industry has shown increasing interest in utilizing recycled plastics as alternative raw materials. Recycled polymer-based composites are particularly promising due to their durability, ease of molding, and resistance to degradation. When polymers are combined with fillers, their structural, mechanical, and thermal properties can be significantly improved. ⁷ Fillers such as silica sand are commonly incorporated to enhance strength, hardness, dimensional stability, and cost efficiency.^8–12

Several studies have reported the potential of plastic–sand composites for construction applications. Khadka et al. ¹³ developed plastic–sand floor tiles with low water absorption (0.04%) and compressive strength of 25.4 MPa, suitable for non-traffic pavements. Bamigboye et al. ¹⁴ demonstrated that roof tiles manufactured from recycled PET and river sand exhibited better performance at 40–50% PET content than conventional cement tiles. Christopher et al. ¹⁵ investigated LDPE-bonded sand blocks and observed improved compressive strength (27 MPa) when smaller silica particles were used. Similarly, Omosebi et al. ¹⁶ produced floor tiles by substituting cement with shredded PET and sand, obtaining higher compressive and flexural strengths compared to conventional cement-based tiles. However, there remains a research gap regarding the comparative use of multiple recycled thermoplastics combined with varying silica sand particle sizes and filler proportions, particularly for roofing applications.

In this study, three common waste thermoplastics—high-density polyethylene (HDPE), polypropylene (PP), and polyethylene terephthalate (PET)—were selected to produce silica-reinforced composites for roof tile applications. The main objective is to optimize production parameters and evaluate their effects on compressive strength, flexural strength, hardness, impact energy absorption, and water absorption properties. Specifically, the effects of thermoplastic type, silica sand weight fraction, and silica particle size were systematically investigated. The novelty of this research lies in the comparative evaluation and rigorous multi-objective optimization of silica-reinforced composites produced from three different recycled thermoplastics—HDPE, PP, and PET—for roof-tile applications. Unlike previous studies that primarily focused on a single polymer system or relied on trial-and-error approaches, this work employs a robust optimization framework combining Taguchi experimental design with Grey Relational Analysis (GRA) to simultaneously evaluate the effects of thermoplastic type, silica sand particle size, and filler weight fraction. This integrated approach quantitatively identifies an optimal composite formulation that simultaneously achieves enhanced and more consistent mechanical and water-resistance performance, outperforming non-optimized mixes reported in earlier studies and overcoming the limitations of single-property optimization.^17,18 By elucidating the interactions between polymer type, filler size, and composition, the study provides new insights into performance-driven mix design. Furthermore, the use of locally available river silica sand enhances cost efficiency and sustainability, while the comparative assessment of multiple recycled thermoplastics offers practical guidance for selecting the most suitable waste polymer for roofing applications. Overall, this work demonstrates that systematic optimization is a decisive factor in achieving high-performance recycled plastic–sand composites, thereby contributing to sustainable construction and advancing circular economy objectives.

Materials and methods

Materials

The materials used in this study included waste thermoplastics and silica sand. The thermoplastic materials—high-density polyethylene (HDPE), polypropylene (PP), and polyethylene terephthalate (PET)—were sourced from post-consumer packaging waste and other discarded products collected from solid waste disposal sites in Mekelle City, Tigray, Northern Ethiopia. Silica sand was collected from three riverbed locations in the Tigray region, namely Edagahamus, Gerebgba, and Shket. The sand is predominantly composed of silicon dioxide (SiO₂), which provides high strength, durability, and suitability as a reinforcing filler for composite production.

Processing of silica sand

The collected silica sand was first placed in plastic containers and thoroughly washed with water to remove dirt and impurities. The washing process was repeated several times to ensure complete cleaning, after which the sand was oven-dried to eliminate residual moisture. Subsequently, the dried sand was sieved in accordance with ASTM C33/C33M-08 standards using ISO sieve meshes, and three particle size ranges—450 µm, 300 µm, and 150 µm—were selected for this study.

Silica characterization technique

Silica sand samples were collected from four locations in Tigray, Northern Ethiopia—Edagahamus I, Edagahamus II, Gerebgba, and Shket—for X-ray fluorescence (XRF) analysis. Determining the oxide composition of the sand is essential to evaluate its potential influence on the mechanical performance of the resulting composites. For the analysis, 12 g of finely ground sand from each site was prepared and compacted under a pressure of 200 kN. Elemental characterization was conducted using an XRF spectrometer (PANalytical PW 4400/24, 230 V, 5 KVA, 50/60 Hz) at Messebo Cement Factory, Mekelle, Tigray, Ethiopia.

Fabrication of composites

After cleaning and preparation, the shredded waste plastics were fed into a furnace and melted at controlled temperatures to produce a homogeneous molten material suitable for molding. The melting temperatures applied to melt the recycled HDPE, PP, and PET are 135°C, 175°C, and 265°C, respectively. Silica sand micro particles were gradually incorporated into the molten thermoplastics to ensure uniform dispersion while maintaining the integrity of the polymer matrix, with a mixing time of 10 minutes at a mixing speed of 150 rpm. ¹⁹ The resulting mixture of molten plastic and sand particles was subsequently shaped and compacted using a compression molding machine at a pressure of 20 MPa. ¹³

Testing methods

Impact strength

The impact behavior of the composites was evaluated using a Charpy pendulum impact tester in accordance with ISO 179-1:2010, ²⁰ using notched specimens with dimensions of 80 × 10 × 4 mm³. A minimum of five specimens were tested for each composition, and the average value was reported to ensure reliability.

Flexural strength

The flexural strength of the composites was measured in accordance with ASTM D790 ²¹ using a universal testing machine (Microcomputer-Controlled Electro-Hydraulic Servo Universal Testing Machine, Model SI-1000 KN) at Mekelle University, Tigray, Ethiopia. Tests were performed on five specimens for each composition, and all measurements were conducted at ambient temperature.

Compression strength

The compressive strength of the composites was evaluated in accordance with ASTM D695-96 ²² using a universal testing machine (Microcomputer-Controlled Electro-Hydraulic Servo Universal Testing Machine, Model SI-1000 KN) at Mekelle University, Tigray, Ethiopia. Rectangular specimens with dimensions of 25 × 25 × 10 mm were prepared following the standard requirements. Each specimen was tested at ambient temperature, and a minimum of five replicates were used for each composition to ensure repeatability and reliability of the results.

Hardness

Specimens for Rockwell B-scale hardness testing were prepared with standardized dimensions of 25 × 25 × 10 mm, following ASTM D785. Hardness measurements were conducted at an ambient temperature of 23 ± 2°C using a Rockwell hardness tester. For each composite formulation, a minimum of five measurements were taken at different locations on the specimen surface, and the average value was reported to ensure accuracy and reproducibility.

Water absorption

Water absorption of the developed composites was determined by immersing the specimens in distilled water, following the procedure outlined in ASTM D570-98. Each specimen was weighed before immersion to obtain its initial (conditioned) weight and reweighed after immersion to obtain the wet weight. The percentage of water absorption was then calculated using equation (1). All tests were conducted at ambient temperature, with a minimum of five replicates per composite formulation to ensure reproducibility. The average value was reported, allowing assessment of the effects of silica sand content, particle size, and thermoplastic type on the water uptake behavior of the composites.

$\mathbf{IW}=\frac{\mathbf{WW}-\mathbf{CW}}{\mathbf{CW}}\times \mathbf{100}\%$

(1)

where, IW = increase in weight (%), WW = weight of the specimen after immersion (g), and CW = conditioned (initial) weight of the specimen before immersion (g).

Design of experiment

Taguchi analysis

Taguchi analysis is widely used to identify the optimal combination of input parameters that yield the best possible results while minimizing the number of experiments. A key component of this method is the signal-to-noise (S/N) ratio, which is evaluated based on three quality characteristics: “larger the better,” “nominal the best,” and “smaller the better.” In the present study, an L9 orthogonal array was employed to investigate three process parameters—type of thermoplastic, weight fraction of silica sand, and particle size of silica sand—each varied at three levels, as summarized in Tables 1 and 2. The corresponding S/N ratios were calculated from the experimental response values using equations (2) and (3). ²³ Waste thermoplastics such as HDPE, PP, and PET are abundantly available in Ethiopia, and their utilization addresses critical environmental concerns related to plastic waste. In this study, the “smaller the better” criterion was applied to minimize water absorption (WA), whereas compressive strength (CS), impact strength (IS), hardness (H), and flexural strength (FS) were optimized using the “larger the better” criterion. Silica sand weight fractions of 30 wt.%, 40 wt.%, and 50 wt.% were selected based on prior studies,^4,14 as these levels provide improved mechanical performance suitable for developing durable rooftop tiles from recycled thermoplastic wastes. Furthermore, these ranges were chosen considering interfacial bonding and processing limitations: lower silica contents ensure sufficient polymer matrix continuity for toughness, while higher contents enhance stiffness and load transfer without causing excessive brittleness or processing difficulties. Particle size levels (150–450 µm) were selected to balance effective filler-matrix interaction, dispersion, and workability during composite fabrication.

Table 1.

Process parameters with their levels.

Levels	Process parameters
	Type of thermoplastic (A)	Weight percentage of silica sand (B)	Silica particle size (C)
1	HDPE	30 wt. %	150 µm
2	PP	40 wt. %	300 µm
3	PET	50 wt. %	450 µm

Table 2.

L9 orthogonal array with input parameters.

Experimental runs	Process parameters
	A	B	C
1	1	1	1
2	1	2	2
3	1	3	3
4	2	1	2
5	2	2	3
6	2	3	1
7	3	1	3
8	3	2	1
9	3	3	2

Smaller is better:

$\frac{\mathit{S}}{\mathit{N}}=-\mathbf{10}\hspace{0.17em}\mathbf{log}\left(\frac{\mathbf{1}}{\mathit{n}}{\displaystyle \sum }_{\mathit{k}=\mathbf{1}}^{\mathit{n}}{{\mathit{y}}_{\mathit{i}}}^{\mathbf{2}}\right)$

(2)

Larger is better:

$\frac{\mathit{S}}{\mathit{N}}=-\mathbf{10}\hspace{0.17em}\mathbf{log}\left(\frac{\mathbf{1}}{\mathit{n}}{\displaystyle \sum }_{\mathit{k}=\mathbf{1}}^{\mathit{n}}\frac{\mathbf{1}}{{{\mathit{y}}_{\mathit{i}}}^{\mathbf{2}}}\right)$

(3)

Where y_i represents response variables and “n” donates the number of experiments.

Analysis of variance

Analysis of variance (ANOVA) is a statistical method used to evaluate the significance of factors affecting response variables, based on contributions, F-tests, and p-values.^24–27 In this study, ANOVA was applied to the experimental results to determine the statistical significance of each process parameter on the measured responses. The results are reported as percentage contributions, which indicate the relative influence of each factor on the performance characteristics of the composites. A significance level of α = 0.05 was adopted, corresponding to a 95% confidence level, ensuring that the probability of the observed differences arising from random variation is limited to only 5%, thereby confirming the reliability of the results.

Multi response optimization

Although the Taguchi method is efficient, it does not inherently offer a complete solution for determining optimal parameters when dealing with multiple objectives at the same time. Hence, the Grey Relational Analysis (GRA) technique is considered appropriate for optimizing multiple response variables by accounting for the relative significance of each quality characteristic.^26,28,29 This study applied GRA to optimize the development of composites, focusing on type of thermoplastic wastes, weight percentage of silica sand particles and size of particles of silica sand as the main process parameters.

GRA follows these procedures

Step 1: Grey relational generation

Which involves normalizing the experimental data according to the nature of the response. When the desired outcome is a maximum value, the “higher the better” normalization formula is applied, whereas for minimizing a response, the “smaller the better” normalization equation is used.

Smaller is better

${\mathbf{y}}_{\mathbf{i}}\left(\mathbf{k}\right)=\frac{\mathbf{max}\hspace{0.17em}{\mathbf{x}}_{\mathbf{i}}^{\mathbf{0}}\left(\mathbf{k}\right)-{\mathbf{x}}_{\mathbf{i}}^{\mathbf{0}}\left(\mathbf{k}\right)}{\mathbf{max}\hspace{0.17em}{\mathbf{x}}_{\mathbf{i}}^{\mathbf{0}}-\mathbf{min}\hspace{0.17em}{\mathbf{x}}_{\mathbf{i}}^{\mathbf{0}}\left(\mathbf{k}\right)}$

(4)

Larger is better

${\mathbf{y}}_{\mathbf{i}}\left(\mathbf{k}\right)=\frac{{\mathbf{x}}_{\mathbf{i}}^{\mathbf{0}}\left(\mathbf{k}\right)-{\mathbf{minx}}_{\mathbf{i}}^{\mathbf{0}}\left(\mathbf{k}\right)}{\mathbf{max}\hspace{0.17em}{\mathbf{x}}_{\mathbf{i}}^{\mathbf{0}}-\mathbf{min}\hspace{0.17em}{\mathbf{x}}_{\mathbf{i}}^{\mathbf{0}}\left(\mathbf{k}\right)}$

(5)

Where $y_i\left(k\right)$ is normalization value of gray relational generalization, $x_i^0\left(k\right)$ is the later gray relational formation value $\mathit{max}\hspace{0.17em}{x}_i^0\hspace{0.17em}and\hspace{0.17em}\mathit{min}\hspace{0.17em}{x}_i^0\left(k\right)$ are maximum and minimum value $x_i^0\left(k\right)$ and $x_i^0\left(k\right)$ is target value.

• Step 2: Grey relational coefficient

The grey relational coefficient represents the degree of correlation between the ideal (best) and the observed experimental results. It is determined using the formula provided in equation (6). equation (7) defines the deviation between ${\mathrm{y}}_0\left(\mathrm{k}\right)$ and ${\mathrm{y}}_{\mathrm{i}}\left(\mathrm{k}\right)$ , whereas equation (8) indicates the maximum deviation value, and equation (9) specifies the minimum deviation value.

${\mathit{\xi }}_{\mathit{i}}\left(\mathit{k}\right)=\frac{{\mathrm{\Delta }}_{\mathit{min}}+{\varnothing \mathrm{\Delta }}_{\mathit{max}}}{{\mathrm{\Delta }}_{\mathit{o}\mathit{i}}\left(\mathit{k}\right)-\varnothing {\mathrm{\Delta }}_{\mathit{max}}}$

(6)

${\mathrm{\Delta }}_{\mathit{o}\mathit{i}}={\mathit{y}}_{\mathbf{0}}\left(\mathit{k}\right)-{\mathit{y}}_{\mathit{i}}\left(\mathit{k}\right)$

(7)

${\mathrm{\Delta }}_{\mathit{max}}=\mathit{max}\hspace{0.17em}{\mathit{y}}_{\mathbf{0}}\left(\mathit{k}\right)-{\mathit{y}}_{\mathit{i}}\left(\mathit{k}\right)$

(8)

${\mathrm{\Delta }}_{\mathit{min}}={\mathit{m}\mathit{i}\mathit{n}\mathit{y}}_{\mathbf{0}}\left(\mathit{k}\right)-{\mathit{y}}_{\mathit{i}}\left(\mathit{k}\right)$

(9)

Where, ${\xi }_i\left(k\right)$ represents the grey relational coefficient, Δ_oi denotes the deviation between $y_0\left(k\right)$ and $y_i\left(k\right)$ , where $y_0\left(k\right)$ is the reference sequence and $y_i\left(k\right)$ , is the comparative sequence. ${\mathrm{\Delta }}_{\mathit{max}}$ and ${\mathrm{\Delta }}_{\mathit{min}}$ correspond to the maximum and minimum deviation values, respectively. The symbol $\varnothing$ refers to the identification coefficient, which lies within the range 0 < $\varnothing$ < 1 and is commonly assigned a value of 0.5.

• Step 3: Grey relational grade

The overall GRG is calculated using equation (10)

${\mathbf{\gamma }}_{\mathbf{i}}=\frac{\mathbf{1}}{\mathbf{n}}{\sum }_{\mathbf{i}}^{\mathbf{n}}{\mathit{\xi }}_{\mathit{i}}\left(\mathbf{k}\right)$

(10)

Where n is the number of quality characteristics, ${\gamma }_i$ is the overall GRG.

• Step 4: Grey relational ordering

In this stage, different parameter combinations are ranked according to their overall performance considering multiple objectives.

Results and discussion

Elemental composition

The elemental composition of the river sand samples, as determined by XRF analysis, is summarized in Table 3 and reveals notable geochemical variability among the sites. Sand collected from Edaga Hamus I exhibits the highest SiO₂ content (87.80%), indicating a highly mature, quartz-rich material with minimal contributions from alumina- and iron-bearing minerals. Silica, predominantly occurring as silicon dioxide (SiO₂), is the most widely utilized non-metallic mineral due to its global abundance, low silt content, and advantageous physical properties, including high thermal resistance, remarkable hardness, and overall cost effectiveness. In contrast, the Edaga Hamus II and Gerebgba samples display lower SiO₂ contents (65.93% and 69.63%, respectively) along with elevated levels of Al₂O₃, Fe₂O₃, CaO, and MgO, suggesting greater contributions from feldspars, ferromagnesian minerals, and possibly carbonate-rich lithologies. The Shket sand, with a SiO₂ content of 82.36% and low concentrations of other oxides, more closely reflects the chemical maturity of Edaga Hamus I. Overall, the observed compositional variations reflect differences in mineralogical maturity, provenance, and the degree of weathering influencing each sampling location. Given its exceptionally high silica concentration, the Edaga Hamus I sand was therefore selected for further investigation in this study.

Table 3.

Elemental composition of four type of sand.

Sample type	Sample code	SiO2 (%)	Al2O3 (%)	Fe2O3 (%)	CaO (%)	MgO (%)
Sand	Edaga hamus (I)	87.80	0.91	3.38	2.58	0.00
Sand	Edaga hamus (II)	65.93	6.75	5.57	4.55	0.23
Sand	Gerebgba	69.63	3.17	4.85	7.43	0.59
Sand	Shket	82.36	1.88	3.81	3.85	0.04

Mechanical properties

The mechanical properties of the thermoplastic–silica sand composites, presented in Table 4, showed clear variations across the nine experimental runs, reflecting the combined effects of polymer type, silica loading, and silica particle size. Overall, increasing silica content from 30 wt.% to 50 wt.% led to notable improvements in hardness and compressive strength. This enhancement is attributed to the rigid silica particles reinforcing the polymer matrix, restricting polymer chain mobility, and increasing resistance to localized deformation. Samples containing 50 wt.% silica (e.g., Runs 3, 6, and 9) consistently exhibited the highest hardness and compressive strength, confirming the reinforcing role of mineral fillers. Flexural strength followed a similar trend, increasing with higher silica loading, particularly in composites based on stiffer polymers. PET-based composites (Runs 7–9) demonstrated the strongest flexural performance due to PET’s high crystallinity and modulus, which facilitate effective load transfer and interaction with silica particles. PP-based composites showed moderate improvements, whereas HDPE-based samples exhibited the lowest flexural strength, consistent with HDPE’s lower stiffness and weaker filler–matrix adhesion. Impact strength, however, displayed an opposite trend. As silica content increased, impact resistance decreased because rigid filler particles promote brittleness and reduce the matrix’s capacity to absorb fracture energy. The greatest reductions occurred in samples containing 50 wt.% silica. Nevertheless, HDPE-based composites consistently exhibited higher impact strength compared to PP- and PET-based counterparts, owing to HDPE’s superior ductility, toughness, and energy-absorbing capability. ³⁰ Particle size also influenced performance: smaller particles (150 µm) improved dispersion, slightly mitigating the reduction in impact resistance, whereas larger particles (450 µm) introduced stress concentration points and increased brittleness. Similar trends have been reported for recycled polypropylene composites reinforced with natural fillers, where optimized formulations achieved enhanced mechanical, thermal, and interfacial properties. ¹⁸

Table 4.

Experimental results of mechanical properties of thermoplastic rooftile composite.

Experimental run	Mechanical properties
	IS (J)	H (RHB)	CS (MPa)	FS (MPa)
1	5.83	41.27	30.50	24.70
2	4.10	46.67	33.10	28.67
3	3.65	51.60	37.20	35.75
4	4.60	44.3	36.10	28.60
5	3.12	49.9	41.80	35.35
6	2.80	56.8	46.80	45.35
7	3.70	48.65	44.41	36.53
8	2.90	54.60	53.60	43.55
9	2.10	60.80	57.90	49.56

Overall, the results indicate a trade-off between rigidity and toughness. PET-based composites with moderate to high silica content and intermediate particle sizes provide excellent hardness, compressive strength, and flexural performance, while HDPE-based composites excel in impact resistance. Since roof tiles must withstand compressive, bending, and impact forces, further optimization is required to balance these mechanical properties effectively. Strategies may include adjusting silica content, selecting an optimal particle size, or exploring hybrid polymer blends to develop composites that combine high structural strength with sufficient toughness. These findings highlight the potential of silica-reinforced thermoplastic composites for roof-tile production, while emphasizing the importance of continued refinement to achieve an optimal balance of mechanical performance for real-world applications.

Taguchi analysis

Impact strength

The main effects plot for the signal-to-noise (SN) ratios, shown in the Figure 1, illustrates the influence of the three process parameters—polymer type (A), silica content (B), and silica particle size (C)—on the impact strength of the thermoplastic–silica sand composites. According to the SN ratio analysis, higher values indicate better performance (i.e., higher impact strength). From the graph, it is evident that Type of thermoplastic (A) has the significant effect on impact strength. The SN ratio is highest at level 1 (HDPE), decreases at level 2 (PP), and reaches its lowest value at level 3 (PET). This indicates that HDPE-based composites provide superior impact resistance, while PET-based composites are the most brittle. This can be attributed to differences in polymer chain flexibility and toughness, where HDPE’s ductile nature allows better energy absorption and crack blunting, whereas PET’s stiffer chains facilitate crack propagation and brittleness. Silica content (B) also strongly influences impact strength. The plot shows the highest SN ratio at the lowest filler content (30 wt.%), decreasing progressively with 40 wt.% and 50 wt.% silica. This trend indicates that increasing silica loading reduces the composite’s capacity to absorb impact energy, which can be attributed to the brittle nature of rigid mineral fillers. Higher filler content can create stress concentration points within the matrix, reducing load transfer efficiency and promoting premature fracture. Previous studies have reported that increasing the weight fraction of silica particles incorporated into thermoplastics can lead to a noticeable reduction in absorbed impact energy. ¹⁰ Silica particle size (C) appears to have a comparatively minor effect on impact strength. The SN ratios show only slight variation across the three particle sizes, indicating that particle size has a limited influence on impact behavior relative to polymer type and silica content. Smaller particles may slightly improve interfacial adhesion and local load transfer, but the overall effect is minimal compared to polymer ductility and filler fraction.

Figure 1.

The main effects plot S/N ratios on impact strength.

In summary, the main effects plot highlights that HDPE with low silica content exhibits the highest impact strength, while increasing silica content and using stiffer polymers like PET decrease impact resistance. Particle size has only a marginal effect. These insights are important for optimizing composite formulations for applications requiring high toughness, such as roof tiles.

Hardness

As illustrated in Figure 2, the hardness of recycled thermoplastic composites reinforced with silica sand microparticles is strongly influenced by the type of thermoplastic (A), silica content (B), and silica particle size (C). The plot for hardness indicates that the type of thermoplastic (A) has the significant effect. Among the three polymers, HDPE-based composites (level 1) exhibit the lowest hardness, followed by PP-based composites (level 2) with moderate hardness, while PET-based composites (level 3) achieve the highest hardness. This trend reflects the differences in crystallinity: PET has the highest crystallinity, followed by PP, and HDPE has the lowest. Higher crystallinity results in a stiffer molecular structure and greater resistance to indentation, explaining the superior hardness of PET-based composites. ³¹ Additionally, the higher stiffness of PET enhances load transfer between the polymer matrix and the rigid silica particles, contributing to overall composite reinforcement. Silica content (B) also strongly influences hardness. Increasing the silica content from 30 wt.% (level (1) to 50 wt.% (level (3) leads to a significant increase in hardness across all polymer types, consistent with the intrinsic rigidity of silica particles, which reinforce the polymer matrix and restrict chain mobility. ³² Higher filler content also improves the load-bearing capacity of the composite and limits polymer chain deformation under applied stress, further increasing surface hardness. Silica particle size (C) has a moderate effect on hardness. Composites containing intermediate particle size (300 µm, level (2) exhibit the highest hardness, whereas smaller (150 µm, level (1) or larger (450 µm, level (3) particles produce slightly lower values. This suggests an optimal particle size that maximizes particle-matrix contact, improves stress transfer efficiency, and prevents stress concentration points that could weaken the surface. PP composites show intermediate performance. Both polymer type and silica content dominate the hardness response, whereas particle size has a secondary effect. These findings provide guidance for optimizing hardness in recycled thermoplastic composites for roof-tile applications. Selecting PET as the matrix, incorporating higher silica content, and using an appropriate particle size can maximize surface hardness, ensuring improved stiffness, resistance to indentation, and wear performance essential for durable, long-lasting roofing materials. These observations are consistent with previously reported findings, where recycled HDPE reinforced with silica exhibited a maximum hardness of approximately 44 HV, confirming the reinforcing effect of silica particles on thermoplastic matrices. ¹³ Furthermore, literature reports indicate that PET composites containing 50 wt.% silica with intermediate particle sizes achieve the highest hardness, whereas HDPE composites with lower silica content and finer particles show the lowest hardness, reinforcing the conclusion that increasing silica content enhances composite hardness. ³³

Figure 2.

The main effects plot S/N ratios on hardness.

Compressive strength

Based on the analysis of the S/N ratios main effects plot shown in Figure 3, the thermoplastic type (A) is the most significant factor influencing the compressive strength of the thermoplastic–silica sand composites, followed by the silica content (B), while the silica particle size (C) has only a minor effect. The plot reveals a clear trend showing that PET-based composites (Level 3) achieve the highest compressive strength, followed by PP (Level 2), and then HDPE (Level 1). This trend can be attributed to PET’s higher crystallinity and stiffness, which enhance load-bearing capacity and improve stress distribution under compression. The S/N ratio also increases with higher silica content, indicating that 50 wt.% (Level 3) provides the most effective reinforcement. The rigid silica particles restrict polymer chain mobility, improve stress transfer between filler and matrix, and increase the overall stiffness of the composite, resulting in higher compressive strength. Previous studies on recycled thermoplastic roof tiles have similarly reported that composites with higher silica content exhibit superior compressive strength, consistent with the present results. ¹³ Conversely, silica particle size has only a minor effect, showing slight variations across the tested levels. In summary, the optimal formulation for maximum compressive strength—which is critical for roof-tile applications—utilizes PET combined with the highest silica content. Compressive strength decreases in HDPE composites with lower silica content due to lower matrix stiffness and less effective stress transfer, emphasizing the importance of polymer type and filler content in load-bearing performance.

Figure 3.

The main effects plot S/N ratios on compressive strength.

Flexural strength

The main effects plot for the flexural strength (FS) S/N ratios, shown in Figure 4, identifies the key factors governing the composite’s resistance to bending, with polymer type (A) and silica content (B) emerging as the dominant parameters. In the S/N ratio analysis, higher values indicate better performance, corresponding to higher flexural strength. The plot shows a steep increase in S/N ratio for polymer type (A) from HDPE to PET (A3), indicating that the high inherent stiffness and crystallinity of PET provide a superior matrix capable of withstanding greater strain under bending stress. PET’s rigid molecular structure enhances stress distribution, delays crack initiation, and improves load transfer from the matrix to the embedded silica particles. Silica content (B) further complements this effect, with the highest S/N ratio observed at 50 wt.% (B3). At this loading, the rigid silica particles maximize filler-matrix contact, restrict polymer chain mobility, and enhance overall composite stiffness, resulting in higher flexural strength prior to failure. Silica particle size (C) has a comparatively minor effect, showing only slight variations across the tested levels, indicating that moderate particle sizes are sufficient to achieve effective stress transfer without introducing significant stress concentrations. The experimental results validate these trends, showing that the optimal combination for maximum flexural strength is A3B3C2: PET (A3), 50 wt.% silica (B3), and 300 µm particle size (C2). This optimized formulation provides the mechanical robustness required for roof tiles, ensuring they can resist bending stresses from wind loads and handling during installation, thereby enhancing durability and long-term performance.

Figure 4.

The main effects plot S/N ratios on flexural strength.

Water absorption

The analysis of water absorption utilized the “smaller is better” criterion for the Signal-to-Noise (S/N) ratio, where higher S/N values correspond to lower percentages of water absorbed, representing optimal composite durability. The experimental results, detailed in Table 5 and illustrated in Figure 5, demonstrate significant variation in water uptake across the nine combinations, ranging from a minimum of 0.15% (A1B1C1) to a maximum of 0.54% (A3B3C2). The optimal combination producing the lowest water absorption (0.15%) was A1B1C1: HDPE thermoplastic (A1), the lowest silica content (30 wt.%, B1), and 150 µm silica particle size (C1). This outcome is strongly linked to material properties. HDPE is a non-polar, semi-crystalline polymer with excellent inherent moisture resistance, which minimizes water diffusion into the matrix. Additionally, lower silica content reduces the volume of hydrophilic particles within the composite, limiting sites for water absorption. As the filler content increases, water uptake rises (e.g., 0.54% at A3B3C2) due to the greater affinity of silica particles for moisture, which promotes water migration along particle-matrix interfaces. Taguchi analysis successfully identifies that minimizing hydrophilic filler content while maximizing water-resistant polymer content is the most effective strategy for low moisture uptake. These results are consistent with previous studies on PET/sand composites, which reported comparable water absorption values. ³⁴ Low water absorption is critical for roof-tile applications, as it enhances durability, reduces the risk of swelling or degradation under environmental exposure, and ensures long-term performance in wet conditions.

Table 5.

Experimental results of water absorption.

Experimental run	Percentage of water absorbed (%)
1	0.15
2	0.23
3	0.40
4	0.21
5	0.33
6	0.44
7	0.30
8	0.37
9	0.54

Figure 5.

The main effects plot S/N ratios on water absorption.

ANOVA

Impact strength

The Analysis of Variance (ANOVA) results for the impact strength (energy absorption) of the recycled waste plastics reinforced with micro silica sand particles, summarized in Table 6, definitively establish the statistical influence of the process parameters. The analysis reveals that the weight fraction of silica sand (B) is the most dominant factor, contributing 56.17% to the total variation in impact strength, closely followed by the type of thermoplastics (A), which contributes 41.25%. Crucially, the p-values calculated for both the type of recycled plastic (A) and the weight fraction of silica sand (B) are below the 0.05 significance level (corresponding to a 95% confidence level). This statistical evidence demonstrates that these two factors significantly influence the impact strength of the developed composite. However, the p-value for silica particle size (C) is 0.226, which is greater than the 0.05 significance level. This confirms that Silica Particle Size (C) is statistically insignificant in controlling the impact strength, aligning with the visual observation from the S/N ratio plot that showed only slight variation across its levels. In conclusion, the composite’s energy absorption capability is overwhelmingly controlled by the material’s base ductility (A) and the volume of the brittle reinforcing filler (B).

Table 6.

ANOVA results for impact strength.

Source	DF	Adj SS	Adj MS	Contribution (%)	F-value	p-value
A	2	4.05449	2.02724	41.25	70.86	0.014
B	2	5.52016	2.76008	56.17	96.47	0.010
C	2	0.19616	0.09808	2.00	3.43	0.226
Error	2	0.05722	0.02861	0.58
Total	8			100

Hardness

Table 7 summarizes the ANOVA results for the hardness of recycled thermoplastics reinforced with micro silica sand particles. The analysis shows that the type of thermoplastic and the silica sand weight percentage are the dominant factors, contributing 32.78% and 66.69%, respectively. Both parameters have p-values below the 0.05 significance level, confirming their statistically significant influence on the hardness behavior of the recycled thermoplastic–silica sand composites. However, the p-value for silica particle Size (C) is 0.329, which is greater than the 0.05 significance threshold. This indicates that factor C is statistically insignificant in affecting the hardness response, consistent with the S/N ratio plot that revealed only slight variation across its levels. In conclusion, the composite’s hardness performance is predominantly controlled by the inherent stiffness of the base thermoplastic (A) and the quantity of the brittle silica filler (B), while particle size (C) plays a negligible role.

Table 7.

ANOVA results for hardness.

Source	DF	Adj SS	Adj MS	Contribution (%)	F-value	p-value
A	2	100.264	50.132	32.78	187.69	0.005
B	2	203.998	101.999	66.69	381.88	0.003
C	2	1.087	0.544	0.36	2.04	0.329
Error	2	0.534	0.267	0.17
Total	8

Compressive strength

As shown in Table 8, the ANOVA results for the compressive strength of recycled thermoplastics reinforced with micro silica sand particles reveal that the thermoplastic type and the silica sand weight percentage are the dominant factors, contributing 74.41% and 23.38% of the total variation, respectively. Both parameters have p-values below the 0.05 significance threshold, confirming their statistically significant influence on the composite’s compressive strength. In contrast, the p-value for silica particle size (C) is 0.383, which exceeds the 0.05 significance level. This indicates that particle size is statistically insignificant in determining compressive strength, a finding consistent with the minimal variation observed across its levels. Overall, the compressive strength of the composite is governed primarily by the thermoplastic type (A) and the silica sand weight fraction (B), while the effect of particle size (C) remains negligible.

Table 8.

ANOVA results for compressive strength.

Source	DF	Adj SS	Adj MS	Contribution (%)	F-value	p-value
A	2	509.15	254.57	74.41	87.63	0.011
B	2	159.96	79.981	23.38	27.53	0.035
C	2	9.35	4.675	1.37	1.61	0.383
Error	2	5.81	2.905	0.85
Total	8			100

Flexural strength

The ANOVA results presented in Table 9 indicate that the flexural strength of recycled thermoplastics reinforced with micro silica sand particles is predominantly influenced by the type of thermoplastic and the silica sand weight percentage, which contribute 48.57% and 49.60% of the total variation, respectively. Both factors exhibit p-values below the 0.05 significance threshold, confirming their statistically significant effect on the composite’s flexural behavior. In contrast, the p-value for silica particle size (C) is 0.117, exceeding the 0.05 significance level and indicating that particle size has no statistically significant influence on flexural strength. This result aligns with the S/N ratio plot, which showed only minimal variation across particle size levels. Overall, the statistical outcomes are fully consistent with the S/N trends: the composite’s flexural performance is primarily governed by the thermoplastic type (A) and the silica sand weight fraction (B), while particle size (C) plays a negligible role.

Table 9.

ANOVA results for flexural strength.

Source	DF	Adj SS	Adj MS	Contribution (%)	F-value	p-value
A	2	273.646	136.823	48.57	227.40	0.004
B	2	279.438	139.719	49.60	232.22	0.004
C	2	9.124	4.562	1.62	7.58	0.117
Error	2	1.203	0.602	0.21
Total	8			100

Water absorption

The ANOVA results presented in Table 10 indicate that the water absorption of recycled thermoplastics reinforced with silica sand microparticles is primarily influenced by the thermoplastic type and the silica sand weight percentage, which contribute 25.64% and 73.26% of the total variation, respectively. Both parameters exhibit p-values below the 0.05 significance threshold, confirming their statistically significant effect on the composite’s water absorption behavior. In contrast, the p-value for silica particle size (C) is 0.350, which exceeds the 0.05 significance level. This indicates that particle size has no statistically significant influence on water absorption, consistent with the minimal variation observed in the S/N ratio trend. Overall, the composite’s water absorption response is governed mainly by the thermoplastic type (A) and the silica sand weight fraction (B), while particle size (C) plays a negligible role.

Table 10.

ANOVA results for water absorption.

Source	DF	Adj SS	Adj MS	Contribution (%)	F-value	p-value
A	2	0.030867	0.015433	25.64	66.14	0.015
B	2	0.088200	0.044100	73.26	189.00	0.005
C	2	0.000867	0.000433	0.72	1.86	0.350
Error	2	0.000467	0.000233	0.39
Total	8			100

Multi response optimization

Grey Relational Analysis (GRA) was employed to optimize multiple responses by enhancing key mechanical properties—impact strength, compressive strength, hardness, and flexural strength—while minimizing water absorption, as excessive moisture uptake can compromise durability and performance. These properties are particularly critical for roof tile applications, where the material must withstand mechanical loads, impact from environmental factors, and prolonged exposure to moisture. The effects of process parameters on all target responses were systematically evaluated to identify the most effective combination. Grey Relational Grade (GRG) values and their rankings were used to determine the optimal parameter settings, with higher GRG values indicating conditions closer to ideal performance. As shown in Tables 11 and 12, the maximum GRG value obtained was 0.833, representing the condition that simultaneously achieves minimal water absorption and maximizes impact strength, hardness, compressive strength, and flexural strength—properties essential for durable and reliable roof tiles. The optimal process parameters were identified as A3B3C2, corresponding to PET, 50 wt.% silica sand, and a 300 µm particle size. PET provides high stiffness, toughness, and thermal stability, ensuring the composite can absorb impacts and withstand environmental stress without brittle failure. The 50 wt.% silica sand content enhances hardness, compressive strength, and flexural strength, improving load-bearing capacity while maintaining good processability for molding. A particle size of 300 µm ensures proper interfacial bonding between the silica and PET matrix, optimizing mechanical reinforcement and minimizing water absorption. Together, this combination produces a high-strength, low-water-absorption composite with the durability and mechanical performance required for roof tile applications. This optimized PET/silica sand roof tile outperforms or is comparable with most previously reported cement-based and composite roof tiles in terms of mechanical strength and water resistance, indicating its potential for sustainable, durable, and cost-effective use in real-world roofing applications.^35–38

Table 11.

Normalized values and deviation of responses.

Experiment run	Normalization values of responses					Deviation
	IS	H	CS	FS	WA	IS	H	CS	FS	WA
1	1	0	0	0	1	0	1	1	1	0
2	0.5361	0.276	0.094	0.159	0.794	0.463	0.723	0.905	0.840	0.2051
3	0.415	0.528	0.244	0.444	0.358	0.584	0.471	0.755	0.555	0.641
4	0.670	0.155	0.204	0.156	0.846	0.329	0.844	0.795	0.843	0.153
5	0.273	0.441	0.412	0.428	0.538	0.726	0.558	0.587	0.571	0.461
6	0.187	0.795	0.594	0.830	0.256	0.812	0.204	0.405	0.169	0.743
7	0.428	0.377	0.507	0.475	0.615	0.571	0.622	0.492	0.524	0.384
8	0.214	0.682	0.843	0.758	0.435	0.785	0.317	0.156	0.241	0.564
9	0	1	1	1	0	1	0	0	0	1

Table 12.

Grey relational coefficients and grey relational grades of responses.

Experiment run	Grey relational coefficient					Gray relational grade
	IS	H	CS	FS	WA	GRG
1	1	0.333	0.333	0.333	1	0.500
2	0.518	0.408	0.355	0.373	0.709	0.414
3	0.461	0.514	0.398	0.473	0.4382	0.461
4	0.602	0.371	0.385	0.372	0.764	0.433
5	0.407	0.472	0.459	0.466	0.520	0.451
6	0.381	0.709	0.552	0.746	0.402	0.597
7	0.466	0.445	0.503	0.488	0.565	0.476
8	0.388	0.611	0.761	0.674	0.469	0.608
9	0.333	1	1	1	0.333	0.833

SEM analysis

The scanning electron microscopy (SEM) image shown in Figure 6 provides visual confirmation of the structural integrity of the optimal composite formulation identified through Grey Relational Analysis (A3B3C2: PET, 50 wt.% silica, 300 µm). The micrograph shows a relatively uniform dispersion of silica sand particles within the PET matrix, which is critical for ensuring that applied loads during mechanical testing, such as compressive and flexural stresses, are effectively and evenly transferred from the ductile polymer matrix to the rigid silica particles throughout the composite. This efficient load transfer underpins the high strength, stiffness, and hardness observed in the mechanical tests. Additionally, the SEM image reveals irregular surfaces on some silica particles, which promote mechanical interlocking with the PET matrix. This micro-scale interlocking enhances interfacial bonding and load transfer efficiency, providing a structural basis for the composite’s superior mechanical performance and durability, making it well-suited for demanding applications such as roof tiles.

Figure 6.

SEM image of PET silica reinforced composite.

Conclusion

Recycling plastic waste into sustainable construction materials, such as floor and roof tiles, reduces the demand for virgin raw materials, lowers construction costs, and promotes environmental sustainability. This study developed high-performance composites by reinforcing recycled thermoplastics with locally sourced silica sand, achieving significant improvements in mechanical properties along with acceptable moisture resistance. Notable increases in hardness, compressive strength, and flexural strength validate the reinforcement strategy, while further optimization is required to determine the best combination of thermoplastic type, composition, and sand particle size for roof-tile applications.

The key findings are summarized as follows.

The highest impact strength (5.83 J) and the lowest water absorption (0.15%) were recorded for the composite fabricated using recycled HDPE, 30 wt.% silica sand, and 150 µm particle size. This indicates the sensitivity of energy absorption and moisture resistance to the filler content. In contrast, the impact strength decreased by 64 % for recycled PET with 50 wt.% silica sand compared to the highest impact strength observed in the 30 wt.% silica sand reinforced HDPE composite.

The maximum values for hardness (60.80 HRB), showing a 47.3% increase compared to the lowest hardness, compressive strength (57.9 MPa) with an 89.8% increase, and flexural strength (49.56 MPa) with a 100.6% increase, were obtained for PET reinforced with 50 wt.% silica sand and 300 µm particle size (Experimental Run 9). These results demonstrate the effectiveness of this configuration in enhancing load-bearing performance.

Notably, water absorption varied across experimental runs, with the lowest value of 0.15% recorded for HDPE with 30 wt.% silica sand and 150 µm particle size, representing a 72.2% reduction compared to the highest value of 0.54% observed for PET with 50 wt.% silica sand and 300 µm particle size.

Grey Relational Analysis identified PET with 50 wt.% silica sand and 300 µm particle size (A3B3C2) as the optimal condition, yielding a maximal Grey Relational Grade of 0.833. This combination offered the best overall balance across all measured properties, including compressive strength (57.9 MPa), hardness (60.80 HRB), flexural strength (49.56 MPa), impact strength (2.1 J), and water absorption (0.54%). These values indicate that the optimized composite can withstand typical mechanical loads and resist moisture uptake, making it suitable for practical roof-tile applications where both structural integrity and durability against environmental exposure are critical.

In conclusion, careful selection of polymer type, filler content, and particle size is essential for achieving a balance between mechanical performance and moisture resistance in recycled thermoplastic composites. The results highlight the potential of these materials as sustainable alternatives for roof-tile production. Limitations of the current study include the lack of comprehensive durability evaluations, such as long-term aging, UV resistance, thermal aging, and weathering, as well as limited assessment of emissions during plastic melting and challenges associated with large-scale manufacturing. Building on this foundation, future research should address these limitations by evaluating environmental emissions, conducting detailed durability testing, optimizing manufacturing processes, and ensuring compliance with construction standards. These efforts are essential to support environmental sustainability, long-term performance, and successful industrial implementation of the composites.