Fire Resistance of Lightweight Aggregate Concrete Containing Spent Garnet as Partial Fine Aggregate Replacement

2.1. Materials

The concrete specimens comprised ordinary Portland cement (OPC), river sand, POC, spent garnet, water, and superplasticizer (SP). The river sand used in this study was air-dried and had a maximum particle size of 2.36 mm after passing through a sieve. POC was a lightweight coarse aggregate from a palm oil mill plant's dumpsite. Initially, the large POC chunks were crushed to the required size using a jaw crusher, resulting in POC coarse aggregate ranging from 5 mm to 10 mm. The spent garnet waste was sourced from an East Coast Peninsular Malaysia facility associated with brownfield services for the oil, gas, and petrochemical industries. Since the collected spent garnet was damp, it was oven-dried at 105 °C ± 5 °C for 24 hours to eliminate moisture content. After removing impurities, the garnet was screened through a 600 µm sieve to replace fine aggregate (river sand) partially. The visual representations and particle dispersion of POC and the spent garnet utilized in this research can be found in Figs. (1 and 2), respectively. The physical characteristics of POC, river sand, and spent garnet were assessed and are provided in detail in Table 1. Clean tap water was used for both concrete specimen preparation and curing. A 1% addition of SP by cement was incorporated as a water reducer to enhance the workability of the concrete. The SP used in this research is classified as Type F according to ASTM C494/C494M-19 [30].

2.2. Mix Proportion and Testing Method

Ordinary Portland Cement, fine aggregate (natural river sand), and POC were used for the mix design in the trial mix method. The fine aggregate was partially replaced with spent garnet. The replacement percentages were mixed in several batches, with the control sample being normal concrete without spent garnet. For a series of POC LWAC mixes, the range for percentage replacement of spent garnet was 10%, 20%, 30%, 40%, and 50% (based on the fine aggregate weight), as tabulated in Table 2. The concrete mixes were designed to achieve a target compressive strength of 45 MPa. The interactions between cement, aggregates, and additives like spent garnet and POC can influence concrete strength. The optimal dosage of these additives may vary depending on the specific combination of materials used in the mix. Once the concrete had been thoroughly mixed, it was poured into the steel mold, which measured 100 mm x 100 mm x 100 mm. All the hardened concrete specimens were demoulded after 24 hours and placed in the curing water tank for 28 days. After 28 days of curing, the specimens underwent compressive strength and water absorption tests. Then, after exposure to elevated temperatures of 300 °C, 500 °C, and 700 °C for 1 hour in a furnace, compression tests were conducted on the POC LWAC specimens. The testing procedure was based on the approach used by Abdul Rahim Zai and Salhotra [31]. The furnace switch was shut down at the end, and the vent hole opened. This process helps them cool naturally for 24 hours. Afterwards, an evaluation was conducted on visual inspection, mass reduction, and decrease in residual compressive strength. Fig. (3) shows the specimen's preparation and testing process. The mass loss and reduction in compressive strength values of specimens exposed to elevated temperatures were evaluated using Eqs. (1 and 2), respectively.

(1)

where:

M_i = mass of the specimen before exposed to elevated temperature, g

M_t = mass of the specimen after exposed to elevated temperature, g

(2)

where:

f_{c(28 days)} = 28 days compressive strength of the concrete mixture, MPa

f_{c (t days)} = compressive strength at (t) days after being exposed to elevated temperature, MPa

2.3. Respond Surface Methodology

The effect of different elevated temperatures and levels of spent garnet replacement on POC LWAC properties, exploiting mass loss and strength reduction, was also analyzed statistically using Design Expert 13 software. The analysis aimed to establish the possible relationship between factors (spent garnet content and elevated temperatures) and responses (mass loss and strength reduction). The response surface methodology (RSM) model produced a surface response graph and generated the possible equation. The regression equation derived from the analysis establishes the correlation between the factors and the response. Once the model was validated, the best model to fit the correlation was verified and discussed. The analysis provided insights into how variations in spent garnet content and exposure to elevated temperatures affect the mass loss and strength reduction of POC LWAC specimens.

3.1. Compressive Strength

Fig. (4) displays compressive strength variations in LWAC made from POC with varying proportions of spent garnet replacing sand. Following a 28-day curing period, the lightweight aggregate concrete (LWAC) incorporating 20% spent garnet exhibited the highest compressive strength, reaching 53.7 MPa. This surpassed the compressive strength of the control specimen, which was at 51.0 MPa. Strength gradually declined as the spent garnet percentage increased: 52.4 MPa at 30%, 48.6 MPa at 40%, and 44.6 MPa at 50% garnet replacement. The 20% garnet mixture achieved the highest strength by filling pores and creating a denser concrete structure. The rough, angular texture of spent garnets also contributed to increased strength. Research by Muttashar et al. [32] emphasized the benefit of this texture in enhancing particle grip and reducing strength loss, resulting in higher strength in self-compacting geopolymer concrete. However, caution is needed to avoid excessive use of spent garnet, as 30%, 40%, and 50% levels led to reduced compressive strength in POC LWAC. The reduction in the compressive strength was mainly due to lesser integrity between the cement paste and the aggregate, as mentioned by previous researchers [33]. Table 3 comprehensively compares compressive strength results from multiple studies, each investigating the impact of different substitution percentages of spent garnet or expanded perlite on lightweight aggregate concrete. The table encompasses findings from the present research, highlighting the variations in compressive strength across different substitution levels and the additional insights provided by each study.

3.2. Water Absorption

The impact of incorporating spent garnet as a partial replacement for fine aggregate on the water absorption of POC LWAC is illustrated in Fig. (5.) After 28 days of curing, the water absorption rates are as follows: 2.8%, 2.7%, 2.0%, 2.5%, 2.6%, and 3.5% for 0%, 10%, 20%, 30%, 40%, and 50% substitution of spent garnet in POC LWAC, respectively. Notably, the water absorption values for all POC concrete mixes with varying proportions of spent garnet are below 4%. As per Neville [34], since the water absorption remains below 10% for all specimens, they appear to be of high-quality concrete. Incorporating 20% spent garnet leads to a lower water absorption rate than other substitution levels, possibly because the fine particles of spent garnet act as fillers, filling the internal voids/ pores and resulting in a more compact structure. However, when garnet is substituted at 50%, it exhibits the highest water absorption, consistent with Abdul Shukor Lim et al. [27], who investigated the effect of excess spent garnet in mortar. Zhang et al. [35] attribute the reduction in water absorption primarily to the gradual densification of the inner structure. However, in cases where spent garnet is utilized, the packing level of the aggregate may be inadequate, leading to voids within the POC LWAC. These unfilled voids facilitate water infiltration, exacerbating the absorption. Table 4 presents a comparative analysis of water absorption percentages from various studies investigating the substitution of garnet or iron ore in concrete mixtures. The studies include the present research, Abdul Shukor Lim et al. [27], and Zhang et al. [35]. The table highlights the impact of different substitution levels on water absorption rates and offers additional insights into the structural densification effects observed in the concrete mixtures.

3.3. Visual Assessment

Figs. (6 to 8) illustrate the visible cracking and altered surface color of POC LWAC specimens with varying proportions of spent garnet following a 1-hour exposure at 300 °C, 500 °C, and 700 °C. The heated POC LWAC containing 0%, 20%, and 50% of spent garnet was chosen for visual assessment. At 300 °C, all the surfaces of the heated concrete specimens remained unchanged. When the temperature increases to 500 °C and 700 °C, fractures in the concrete begin to appear. This is due to the effect of heat stress caused by water evaporation [36]. No explosive spalling was witnessed in specimens subjected to heating up to 700 °C; however, crack formations were observed in all samples. Notably, POC LWAC, with a 50% spent garnet replacement, exhibited a more pronounced occurrence of surface cracking at higher temperatures. Upon visual evaluation, it was observed that upon heating to 500 °C, the specimen's color transitioned from light grey to reddish and brownish. Furthermore, at 700 °C, the surface of POC LWAC took on a pinkish hue. In the case of 50% spent garnet replacement, the POC LWAC appeared white with evident surface cracks. Ahsan et al. [37] reported similar findings when using excessive seashell powder in concrete.

3.4. Mass Loss

The mass loss of POC LWAC is shown in Fig. (9). The mass loss of the concrete increased as the temperature increased from 300 °C, 500 °C and 700 °C. The mass loss of exposed specimens was found to be 6.76% to 7.59% (300 °C), 7.31% to 9.30% (500 °C), and 8.42% to 10.62% (700 °C). Moreover, the weight loss of the exposed specimens increased with increasing concrete porosity. When the temperature does not exceed 400 °C, the initial mass loss could be attributed to the evaporation of capillary water. Shukri, Mohamed, and Ishak [38] state that when specimens are subjected to temperatures between 600 °C and 800 °C, the moisture content will be lost from the tiny pores of cement pastes and aggregates, which would result in mass loss of the samples. Table 5 provides a comparative analysis of mass loss percentages observed in concrete specimens exposed to varying temperatures, as reported by the present research and Shukri, Mohamed, and Ishak [38]. The amount of utilized spent garnet generally impacts the mass loss. Concrete containing 20% of spent garnet has the lowest mass loss rate compared to other specimens. It is attributed to reducing concrete voids due to the incorporation of spent garnet as it fills all pores, making the concrete denser. Furthermore, excessive use of spent garnet will lead to a higher mass loss. It could be explained by the fact that the concrete specimen's void becomes larger as the amount of spent garnet increases at 50% replacement. As the void of the concrete matrix increases, it indicates that concrete is becoming less dense. Wang, Huang, and Xie [39] showed a similar pattern in terms of the mass reduction of concrete with the temperature rise.

The validation of the proposed work is meticulously depicted in Fig. (10), where the correlation between spent garnet content and varying elevated temperatures concerning mass loss in POC LWAC is analyzed using RSM. Fig. (10a) illustrates a 3D surface response plot, offering insights into this relationship. The plot distinctly reveals that an increased proportion of spent garnet correlates with a corresponding rise in mass loss in POC LWAC, particularly evident with the temperature escalation up to 700 °C. This trend highlights a proportional relationship between spent garnet content, elevated temperatures, and mass loss. Furthermore, the surface response plot emphasizes that incorporating 20% spent garnet in POC LWAC improves mass loss performance compared to plain POC and other variants containing 10%, 30%, 40%, and 50% spent garnet. Fig. (10b) supplements this analysis with a perturbation plot, facilitating a comparison of the individual effects of these factors on the response. Curve A (spent garnet content) exhibits a more pronounced curvature, indicating higher sensitivity than Curve B (elevated temperature). This underscores the varied content of spent garnet in POC LWAC significantly influences and dominates, resulting in mass loss. The ANOVA analysis further solidifies the strength of the quadratic regression, unveiling a robust relationship between mass loss and these factors, as evidenced by an R-squared value (R²) of 0.8743. Eq. (3) encapsulates the empirical relationship between mass loss and the multiple factors, denoted by the spent garnet content (A) and elevated temperatures (B).

(3)

3.5. Strength Loss

As illustrated in Fig. (11), the compressive strength of the samples noticeably decreases when exposed to higher temperatures, reaching up to 700 °C. This pattern is consistent and significant, indicating that higher temperatures lead to a uniform reduction in compressive strength across all test series. The principal factor contributing to this decline in compressive strength of POC LWAC under elevated temperatures is concrete dehydration, stemming from the elimination of free water. According to prior studies by Heap et al. [40], the hydrothermal influence makes strength loss inevitable, resulting in free and adsorbed water loss. Furthermore, exposure to elevated temperatures induces various physical and chemical alterations in concrete, affecting its matrix. These alterations encompass cement drying, resulting in voids, cracks, and paste shrinkage. Using a suitable percentage enhances the performance of POC LWAC upon exposure to elevated temperatures. Using 20% of the spent garnet resulted in the lowest percentage of strength reduction of concrete. Less void in POC LWAC can be attributed to its denser matrix, making it less vulnerable to damage when subjected to elevated temperatures than the 50% POC LWAC with a porous matrix.

The surface response plots presented in Fig. (12) were instrumental in examining the relationship between spent garnet content, elevated temperature, and strength loss in POC LWAC. In Fig. (12a), the surface response plot illustrates how variations in spent garnet content and temperature impact strength loss in POC LWAC. It becomes apparent that as both spent garnet content and temperature increase, a corresponding rise in strength loss, highlighting a clear connection between strength loss and elevated temperature. Fig. (12b) further elucidates this relationship through a perturbation plot, visually comparing the effects. Notably, Line A, representing spent garnet content, displays a steeper slope than Line B, suggesting that changes in spent garnet content as a fine aggregate replacement exert a more significant influence on enhancing strength loss in POC LWAC than variations in temperature levels. Using RSM, a linear regression model was established to capture this relationship accurately. The subsequent analysis employing ANOVA demonstrated a robust fit with an R-squared value (R²) of 0.8068, indicating a significant correlation. Eq. (4) expressed the relationship between strength loss of POC LWAC, spent garnet content (A), and elevated temperatures (B).

(4)