Although a large body of literature has existed and shown the promise of geopolymer mortars based on industrial and agricultural by-products, the available literature largely concerns binary systems, individual performance parameters, and high alkali geopolymer formulations. Minimal focus has been given to ternary blended geopolymer mortars combining with calcium-rich GGBS, aluminosilicate-rich metakaolin, and waste-based substances, like paper sludge ash, especially at low-to-moderate molarity conditions to enhance the sustainable frame work.

Moreover, investigations on fresh, mechanical, or durability properties are commonly divided into most of the reported studies, which do not examine their interactions and trade-offs. The lack of multi-criteria and data-driven optimization models limits the determination of practically feasible mix designs with balanced performances.

Therefore, there is an obvious necessity to design a study that:

Scientifically examined ternary blended geopolymer mortar using GGBS, metakaolin, and paper sludge ash.

The production of a low ternary geopolymer mortar system based on GGBS-metakaolin-paper sludge ash (PSA), in which PSA is used as a value-added waste binder and not as a filler.

This study employed a unique mixing process initiated by combining the source materials with a NaOH solution to dissolve the aluminium and silicon in the raw material. Subsequently, Na₂SiO₃ solution was added to enhance the binding ability, resulting in superior strength compared with alternative production methods, as shown in Figure 6 (Hanjitsuwan et al., 2014). The process commenced for 3 min with the blending of GGBS, either individually or together with MK and PSA, and sand in a bowl. A pre-prepared NaOH solution was gradually added to the dry mixture, and the mixture was maintained under wet conditions for 3 minutes. The compressive strength was measured by placing the freshly prepared mixtures into cube moulds with dimensions of 50 mm × 50 mm × 50 mm. The cast samples were then positioned on a vibrating table and subjected to vibration for 10 s to eliminate any entrapped air. Subsequently, the specimens were cured at an ambient temperature of 20 °C ± 5 °C. Following the curing process, the samples were stored under standard environmental conditions until they attained testing ages of 7, 14, and 28 days.

Taguchi method is fractional factorial experiments technique owing to its strength, ease, and capacity to reduce the number of required experimental trials. This method is used to conduct experiments using orthogonal arrays (OAs) that are able to clearly investigate two or more factors at a time while greatly reducing the amount of work required in the experimental process. This process begins with the identification of control factors and allocation of relevant levels, followed by the selection of a suitable OA to model these combinations. This approach is effective for minimising the number of experimental runs while considering variables that are difficult to manipulate [30]. The Taguchi design enables the simultaneous and independent testing of multiple factors with the least number of resources, time and financial resources required to conduct experiments through the use of orthogonal arrays. To determine the difference between the desired and observed values, a loss function was developed from which the signal-to-noise (S/N) ratio η was obtained. Three types of S/N ratios may be used, depending on the characteristics under investigation: (i) lower-the-better (LB) when minimisation of the response is required, (ii) nominal-the-better (NB) when the target value is known, and (iii) higher-the-better (HB) when the maximum response is required. A lower-the-better approach was used in this study for water absorption. The compressive strength should be maximised in terms of workability.

A ternary blended geopolymer mortar was fabricated according to the Taguchi method. The four identified factors and their levels are presented in Table 3. Factors and levels for this study were chosen based on previous studies. Four factors, A-D = GGBS (A), metakaolin (B), Paper Sludge Ash (C), and molarity (D), are listed in Table 4. An L₁₆ (4⁴) orthogonal array was generated using the Taguchi design for 16 experiments.

This study employed a unique mixing process initiated by combining the source materials with a NaOH solution to dissolve the aluminium and silicon in the raw material. Subsequently, Na₂SiO₃ solution was added to enhance the binding ability, resulting in superior strength compared with alternative production methods (Hanjitsuwan et al., 2014). The process commenced for 3 min with the blending of GGBS, either individually or together with MK and PSA, and sand in a bowl. A pre-prepared NaOH solution was gradually added to the dry mixture and the mixture was maintained under wet conditions for 3 minutes. The compressive strength was measured by placing the freshly prepared mixtures into cube moulds with dimensions of 50 mm × 50 mm × 50 mm. The cast samples were then positioned on a vibrating table and subjected to vibration for 10 s to eliminate any entrapped air. Subsequently, the specimens were cured at an ambient temperature of 20 °C ± 5 °C. Following the curing process, the samples were stored under standard environmental conditions until they attained testing ages of 7, 14, and 28 days.

The Advanced Materials Characterisation Laboratory at VIT University, Vellore, used X-ray fluorescence analysis to determine the chemical compositions of paper sludge ash, metakaolin, and GGBS. Physical properties, including specific gravity, fineness, and particle size, were analysed in accordance with ASTM C188 (ASTM C188-17, 2023). The mineralogical phases of the materials were analysed using X-ray diffraction (XRD), and the morphologies of GGBS, metakaolin, and paper sludge ash were examined using scanning electron microscopy (SEM). The physical characteristics of the sand, including the specific gravity, water absorption, fineness modulus, and particle size, were measured in accordance with ASTM C136 (ASTM C136-06, 2015) and evaluated according to ASTM-C33 specifications. The flowability of the freshly prepared geopolymer mortars was assessed using the flow table method specified in ASTM C1437-07 (ASTM C230/C230M-14, 2020). The compressive strength of the geopolymer mortar was evaluated after 7, 14, and 28 days of curing in accordance with ASTM standards C109/C109M-08 (ASTM C109/C109M-20, 2020). Three samples were tested, and the outcomes were averaged. To examine the microstructure, samples of the geopolymer mortar that had been subjected to compressive strength testing at 28 days were selected for SEM analysis.

The experiments were performed using an L₁₆ (4⁴) orthogonal array, and four parameters were evaluated in 16 experiments. Second, the L₁₆ orthogonal array was used to optimise and analyse all the mortar attributes together. The notations and quantities for the factor levels are listed in Table 3.

The optimisation of Class C fly ash silica fume geopolymers using the optimal conditions was determined using the Taguchi method, which involved a weight-to-solid ratio of 0.35, an alkaline-to-binder ratio of 0.40, and 10 M NaOH, which caused a significant reduction in the number of experimental runs and allowed the use of ANOVA to analyse the factors (Karslioğlu-Kaya and Onur, 2025). Hybrid Taguchi-GRA-PCA methods have also been optimised on advanced designs like one-part geopolymer concrete (OPGC) (Jain et al., 2025). Ultra-high-performance geopolymer concretes (UHPGC) have achieved a strength of up to 126 MPa with GGBS-based binders optimised using the Taguchi method (Alameri et al., 2025). Silica fume, eggshell ash, and ultra-pulverised palm oil fuel ash (UPOFA) were also used in other optimisation experiments, and the control of the curing temperature and NaOH concentration was investigated using Taguchi L₉ orthogonal arrays (Mashri et al., 2023). According to Abdul Rahim et al., a Taguchi technique incorporating a utility method enhanced fly ash-high-performance concrete that was exposed to temperatures reaching 800 °C, and cement content was found to be the biggest factor in the maintenance of strength (Rahim et al., 2013). Similarly, Vignesh et al. used the Taguchi L9 technique on a quaternary binder system consisting of metakaolin, GGBS, SCBA, and OPC and determined the best water-to-binder ratio of 0.32 to enhance strength, durability, and workability (Vignesh and Abdul Rahim, 2024). Xavier et al. used a Taguchi L₁₆ orthogonal array to optimise a ternary combination of GGBS, metakaolin, and calcined iron-rich material to obtain a compressive strength of 65 MPa and verified that the liquid-to-binder ratio was one of the dominant factors (Belarmin Xavier and Abdul Rahim, 2023). The Taguchi–TOPSIS optimisation and decision-making framework consists of six main components, which are mathematically defined and sequentially presented through (Equations 1–5).

Step 1: Normalised Decision Matrix X ¯ i j * = X i j ∑ i = 1 n X i j 2 (1)

Step 2: Calculate the Weighted Normalised Matrix V i j = X ¯ i j × W j (2)

Here, X ¯ i j the S/N ratio of the ith alternative for the jth response, and W j is the corresponding normalised weight of the S/N response matrix.

Step 3: Calculate the Ideal best and worst value.

Step 4: Separation from Positive Ideal Solution S i + = ∑ j = 1 m V i j − V j + 2 (3)

Step 5: Separation from Negative Ideal Solution S i − = ∑ j = 1 m V i j − V j − 2 (4) where V _j ⁺ and V _j ⁻ represent the positive and negative ideal normalised matrix values, respectively.

Step 6: Closeness Coefficient P i = S i − S i   + + S i   −   (5) where S ⁺ is the positive separation matrix and S ⁻ is the negative separation matrix.

The experimental results of the Taguchi method (Table 5) were optimised using the Technique of Order of Preference by Similarity to Ideal Solution (TOPSIS) to determine the most effective combination of parameters that affect the performance of the ternary blended geopolymer mortar (GGBS + MK + PSA). The multi-output optimization process was based on five response parameters: initial setting time, final setting time, flow, compressive strength, and water absorption of the samples.

Weights were assigned to each response based on its relative importance (Table 6). The compressive strength had the highest weight (0.30) because it is an important parameter for determining structural performance, followed by flow (0.15) and water absorption (0.15), which are the workability and durability parameters, respectively. The setting times were given moderate weights (0.23 and 0.15) based on their effect on handling and strength development at an early age.

A normalised and weighted decision matrix (Table 7) was created to normalise all the responses and use their weights. Thereafter, the Euclidean distances between the positive ideal solution (S⁺) and negative ideal solution (S⁻) were calculated to assess the proximity of each mix to the performance conditions, which are most likely to result in optimal performance. The closeness coefficient (P_i) was calculated using the above equation.

According to the outcomes (Table 7), Mix TBM-16 (P_i = 0.604) had the highest closeness coefficient, and all results indicated the most balanced performance in terms of strength, setting behavior, workability, and durability. TBM-6 (P_i = 0.564) and TBM-13 (P_i = 0.549) also performed successfully, indicating that they could be used for the given application demands. Mixed with lower Pi (e.g., TBM-7 and TBM-2) indicated poor overall optimisation because of the comparatively low strength and high absorption values. Consequently, the best mix was determined to be TBM-15 because of its excellent compressive strength and sufficient workability with limited water absorption, resulting in a balanced combination of mechanical and durability properties. This proved the accuracy of the Taguchi-TOPSIS method for multi-criteria optimization in geopolymer systems.

The predicted values were found to agree well with the experimental results, with all parameters having less than 5 per cent deviation, as predicted by the Taguchi TOPSIS hybrid optimisation model, as shown in Table 8. This close relationship reaffirms the accuracy and predictability of the designed model for maximising both the fresh and hardened characteristics of the ternary-blended geopolymer mortar. The small difference can be explained by experimental errors and material dispersion, which indicates that the proposed model can be successfully used to predict the performance trend with high accuracy.

The initial and final setting times of the geopolymer pastes were determined using a Vicat apparatus, in compliance with ASTM C191 (ASTM C191-08, 2021). The commencement of the setting period was noted upon the needle’s penetration to a depth of 5–7 mm, whereas the conclusion of this period was marked when the needle reached a depth of 30–35 mm, as shown in Figure 7. Metakaolin and paper sludge ash were found to be very slow in ternary blended GGBS pastes. Although geopolymer pastes are usually set quickly owing to an increase in polymerisation, the presence of these additive materials prolongs the time. The setting time for Mix-I was 34 min at 1 M and 19 min at 4 M. However, the setting time for Mix-4 was 75 min at 1 M. These results show that metakaolin and paper sludge ash are highly significant in delaying GGBS-based alkali activation from the paste-setting time.

The workability of the fresh geopolymer mixes was evaluated using the flow table test in accordance with ASTM C230 (ASTM C230/C230M-14, 2020) (Figure 8).

The mix-15 and mix-11 exhibited the highest flow value of 60%. The findings indicate that an optimum relationship between binder substitution and activator concentration may lead to an improvement in the mobility of the fresh mix; however, the absence of such an optimal relationship is likely to decrease workability. Conversely, the lowest flow rate was recorded as 20% for mix 4, which represents the stiffening action of increased activator concentration combined with poorer proportions of the binder (Nath and Sarker, 2014). In particular, when 15%–20% of GGBS-MK was substituted, the amount of binder had a greater impact. Higher flow values were always observed for compositions with moderate PSA content (2.5%–7.5%). This enhancement in workability is due to the fine particle size and reactivity of MK, which contributed to the dispersion of the particles and decreased the interparticle friction, thereby enhancing the fluidity of the paste (Olivia and Nikraz, 2012). Workability decreased as the PSA exceeded 7.5 and was lower than 7.5. This tendency is likely explained by the porous and uneven structure of PSA, which enhances water uptake and reduces the amount of free liquid available for lubricating the paste matrix, ultimately decreasing the flow (Olivia and Nikraz, 2012). Another factor influencing the workability was the alkalinity of the solution used. Activation of mixes with 2 M tended to yield higher flow values than activation mixes with 4 M. The aluminosilicate species dissolved more quickly at higher molarities and increased the rate of geopolymerization, resulting in a stiffer and less workable mix. On the other hand, reduced molarity offers sufficient dissolution kinetics and liquid phase to make the particles movable and lubricate efficiently, which is converted into enhanced flow behaviour (Xu and Van Deventer, 2000). An interesting behavior was observed in the interaction between the binder composition and activator molarity. Blended mix 2 performed well at 2 M and considerably decreased to 4 M. Similarly, mix 4 was very good even at 4 M, indicating that a moderate PSA content can be used to overcome the phenomenon of molarity stiffening. This shows that the synergism between the binder constituents and the level of activator is crucial, and not the effects of the individual parameters (Zhang et al., 2009). In general, the findings indicated the minimal workability was obtained when GGBS was partially replaced with 15 per cent MK, PSA was added at various levels of 2.5%–5%, and the molarity of NaOH and Na₂SiO₃ solution was maintained between 2 and 4 M at different PSA levels. These results were consistent with those of a previous report, which highlighted the need to balance the concentration of the activators and the synergy of the binders to increase the flow behaviour of geopolymer systems (Fernández-Jiménez and Palomo, 2005; Nath and Sarker, 2014; Zhang et al., 2009). The observed trends confirm that the workability of geopolymer mortars is not determined by a single parameter but is a result of a complex interaction between the binder type, binder proportion, and molarity of the activators.

The tested samples compressive strength of the ternary blended geopolymer mortars, as shown in Figure 9A, the results consistently increased with curing age, indicating an enhancement in the formation and densification of the gel during geopolymerization. In group A, which had the highest (92.5%) GGBS content, and its concentration increased from 1 M to 4 M, the strength significantly increased from 13.51 N/mm² (TBM-1) to 46.25 N/mm² (TBM-4) at 7 days and from 29.10 N/mm² to 53.67 N/mm² at 28 days, respectively. This indicates that higher alkalinity favors the dissolution of Ca–Si and Al species, resulting in the promotion of more C–A–S–H gel formation. For Group B (85% GGBS), an equal proportion combination of metakaolin and PSA yielded medium to high strengths; TBM-6 attained 60.07 N/mm² at 28 d, representing an optimum activator concentration of 4 M with 10% PSA for better reactivity and gel connectivity. However, for Group C (77.5% GGBS), the strength ranged between 32.17 N/mm² and 65.56 N/mm², suggesting that although lower calcium levels are associated with early age gain, higher molarity can be improved by accelerating geopolymerization. Group D had the lowest GGBS content (70%), but a higher metakaolin content (15%–20%), indicating a significant synergistic effect. TBM-15 achieved the highest compressive strength of 83.20 N/mm² in 28 days. Such a high performance is due to the increased aluminosilicate reactivity and densified pore structure stemming from the hybrid gel network in the C–A–S–H phases. In general, the compressive strength increased with the Molarity and MK content up to an optimum level, whereas PSA behaves as a reactive filler, contributing positively to microstructural densification. Therefore, TBM-15 was considered to be the optimum mix with a good combination of strength, reactivity, and durability, which proved that the formulation based on a ternary geopolymer blend was effective in this study, and mix design parameters played a key role in strength development.

The plot of the signal-to-noise (S/N) ratios, as shown in Figure 9B, shows the interaction between each control factor (GGBS, metakaolin, paper sludge ash (PSA), and molarity) and the overall performance response of the ternary blended geopolymer mortar. Because the objective function with the larger the better was chosen, an increase in the S/N ratio led to better compressive strength and overall performance.

According to the plot, the S/N ratio increases considerably with increasing GGBS content, which means that the presence of calcium-rich GGBS positively impacts the creation of C-A-S-H gel, resulting in the strengthening of the material and densification of the matrix. The metakaolin factor exhibited a relatively low level of variation, indicating that in the examined range (5%–20%), MK modulates a stable aluminosilicate network, which does not introduce a significant deviation in response. The slight modification of the PSA curve is a reason to believe that PSA is acting as a reactive filler, adding Si and Ca to enhance the continuity of the geopolymer gel instead of initiating significant reactions.

The molarity factor has the highest positive slope, which proves that it is the most dominant parameter. This drastic increase in the S/N ratio between 1 M and 4 M indicates that an increase in the alkali concentration increases the rate at which the precursor dissolves and forms geopolymer gels, leading to a high mechanical performance.

In general, the role of molarity and GGBS content was proven to be the most significant according to the main effects plot, followed by PSA and metakaolin. The parameter combination that showed optimum geopolymer mortar performance was identified as GGBS (Level 4) - metakaolin (Level 2) - PSA (Level 4) - Molarity (Level 4), which is a mixture of high calcium level, moderate level of aluminosilicate, and high level of alkaline activation that can attain the best results in terms of compressive strength and durability.

The current experiment has a greater range of compressive strength (15–60 N/mm²) at 7 days than the reference works, which shows the combined effect of the alkalinity of the molarity (1–4 M) and the different contents of the ternary binders, as shown in Figure 10. Although other mixes demonstrated reduced early age strength compared to Sharmin et al. (2017), who used 14 M NaOH and reported strengths generally in the range of 25–45 N/mm², several optimized mixes in the current study already obtained similar or even higher strengths. Conversely, the 8-M-activated studies of Deng et al. (2022) and Babu and Thangaraj (2023) are more balanced at the early age strength development, indicating an accelerated initial geopolymerization because of the increased alkalinity. However, some current-study combinations can compete with these advantages at much lower molarities, with the positive contribution of precursor synergy. The strength development was also more pronounced at 14 d, with several mixes developing a strength above 50–60 N/mm² that is higher than the levels of strength reported by Sharmin et al. (2017), which were mostly below 45 N/mm², even with a high molarity. The increasing strength trend remained positive in the current study, indicating sustained geopolymerization and the progressive formation of C-S-H and C-A-S-H gels. The relatively flatter trend observed by Sharmin et al. (2017) indicates the limited long-term performance of high-alkali activation, unless the binder is well designed. The findings are also similar to those of Deng et al. (2022), and Farag Gaddafi et al. (2024), revealing that intermediate-age strength can be optimally improved by ternary compositions and not by alkalinity. The advantage of the current study is tangible at 28 d. Some of these mixes have compressive strengths well above 70–80 N/mm², which is much higher than the compressive strengths (30–45 N/mm²) reported by Sharmin et al. (2017), whose compressive strengths do not exceed around 28 days, and also above the range of compressive strengths reported by Deng et al. (2022), Babu and Thangaraj (2023), and Farag Gaddafi et al. (2024), which range between 45 and -60 N/mm². This improvement confirms that an optimized ternary binder composition and controlled low-molarity activation enhance a denser microstructure and more successful gel formation than high-molarity systems. Overall, the findings indicate that geopolymer mortars with high compressive strength can be obtained without using highly concentrated alkaline solutions, providing a more sustainable and practical alternative to traditional high-alkali geopolymer formulations. In general, this study demonstrates that high-performance geopolymer mortars with low alkaline activator molarities (1–4 M) can be produced by optimizing the ternary binder design. The proposed mixes had similar early age strength and much better long-term compressive strength when compared to the higher molarities (8–14 M) used in the published literature, with maximum strengths of over 80 N/mm² in 28 days. This performance improvement can be attributed to the synergetic effect of GGBS, metakaolin, and paper, which enhances efficient geopolymerization, dense gel development, and fine microstructure. These results confirm that well-monitored proportions of precursors and molarity can compete well with traditional high-alkali systems and provide a more sustainable, safer, and environmentally friendly method for producing geopolymer mortar.

A quantitative comparison with Afrin et al. (14 M NaOH) indicates that the current study results in a compressive strength increase of between −59% and +93% after 7 d, between −31% and +110% after 14 d, and between −27% and +138% after 28 d. TBM-15 (∼93% at 7 d and 138% at 28 d) and TBM-9 (∼110% at 14 d) showed the greatest improvements. Notably, the mean strength gain was positively correlated with curing age, with a value of over 50% at 28 d, even though a much lower molarity of activators was used (1–4 M). These findings are quantitatively helpful to confirm that an optimised ternary binder composition is superior to high-molarity activation in delivering high-quality long-term compressive strength.

Figure 11 shows the water absorption behaviour of the ternary blended geopolymer mortar mixes at 7, 14, and 28 days. A consistent decrease in water absorption was reported as the curing age increased, and this was shown to be a result of gradual geopolymerisation and refining of the pores in the matrix. In Group A, where the highest GGBS proportion was (92.5%), the absorption of TBM-1 was approximately 9.5% at 7 days, and that of TBM-4 was below 2% at 28 days, which declined dramatically with an increase in molarity, that is, between 1M and 4M. This is a clear indication that an increase in alkalinity increases the rate of dissolution of the reacting species and is associated with an increase in C–A–S–H gels, resulting in a denser microstructure. Reduced absorption values were also observed in Group B (85% GGBS), and TBM-6 demonstrated slight water uptake at 28 days, indicating that the combination of GGBS, MK, and PSA in a balanced ratio enhanced the compactness of the matrix. Group C (77.5% GGBS) exhibited slightly higher water absorption than the other groups, presumably because of its lower calcium level; however, the steady decline over time also indicates successful pore refinement. Group D had lower GGBS (70%) but higher metakaolin content, and mixes TBM-15 and TBM-16 recorded the lowest water absorption after 28 d, which implies the presence of a thick hybrid gel. In general, the water absorption decreased with increasing molarity and GGBS content, and the curing time increased, proving that the ternary geopolymer system formed a highly compact and strong matrix. This is supported by the lowest absorption within TBM-15, which proves its better microstructural integrity and the combination of strength and durability characteristics.

ANOVA is a statistical tool used to determine the effect of each of these factors, that is, GGBS, metakaolin, paper sludge ash (PSA), and molarity, on the various performance responses of the ternary blended geopolymer mortar. ANOVA measures the extent of each factor to the total variation in the experimental outcomes and then concludes with F-values and P-values, indicating statistical significance. The higher the F-value, the lower the P-value, which is below 0.05, indicating that the specific factor has a statistically significant impact on the response at a 95% confidence level. The percentage contribution also determines the relative contribution of each parameter in controlling a given mortar property, as shown in Figure 12.

The ANOVA results tabled in Tables 9–11 indicate that molarity was the most dominant factor, with the highest percentage contributions and lowest p-values (less than 0.05), which verified that it was statistically significant. In terms of flow, the molarity had the greatest contribution (71.1), indicating that the concentration of the alkali solution determines the paste viscosity and workability. Similarly, molarity had the highest impact on compressive strength (58.9%), followed by GGBS (19.3%), which once again proved that alkaline activation and calcium availability are crucial in the formation of a dense geopolymer matrix. The causes of changes in the initial and final setting times were 44.7% and 51%, respectively, by molarity. This demonstrates the significance of accelerating geopolymerisation and gel formation. There was a moderate response from Metakaolin and PSA to the setting behaviour. Similarly, the maximum contribution was on molarity (63.6%), followed by GGBS (14.8%), which shows that increased alkalinity levels and calcium levels cause a decrease in porosity and an increase in matrix densification. Altogether, ANOVA results prove that molarity is the most significant statistic that determines the impact on fresh and hardened properties, and GGBS is the second most important factor in the strength and durability increase. Metakaolin and PSA also add value to the microstructure, making the geopolymer matrix more stable over the long term. Equations 6–16 are used to find the mean, standard deviation, degrees of freedom, standard error of the mean, confidence interval (95%), sum of squares, mean square, F-statistic, and percentage contribution.

Mean: x ̄ = 1 ¯ n   ∑ i = 1 n x i (6)

Standard Deviation: s = √ ∑ i = 1 n x i − x ̄ 2 / n − 1 (7)

Degree of Freedom: df = n − 1 (8)

Standard Error of the Mean: S E = s / n (9)

Confidence Interval (95%): CI = x ̄   ±   t α / 2 , df     *  SE (10)

For a 95% confidence level: t 0.025 , d f (11)

Sum of Squares (ANOVA): S S   b e t w e e n = ∑ j = 1 k n j x ¯ j − x 2 (12) S S   w i t h i n = ∑ j = 1 k ∑ i = 1 n j x i j − x ¯ j 2 (13)

Mean Square: MS = SS / df (14)

F-statistic: F = MSbetween MSwithin (15)

Percentage Contribution Contributio n   % = S S factor S S total × 100 (16)

The SEM micrograph of the optimum mix (OM) (Figure 14) reveals that the binder is compact, with fewer voids than the control mix, indicating that the pores are refined and more compressible. When metakaolin (MK) interacts with alumina, silica, and Ca²⁺ in GGBS, it forms C-S-H and C-A-S-H gels owing to its pozzolanic reactivity. EDAX analysis confirmed that the Si, Ca, and Al phases were the most common polymerization products, consistent with the XRD data. The silicate phase was formed owing to the presence of a large amount of silicon, and the polymerisation was facilitated by Ca, which led to the densification of the matrix. Moreover, GGBS is a fine particle that enhances void filling, decreases porosity, and strengthens the structure. The combination of GGBS, MK, and PSA leads to the formation of a smaller microstructure with superior mechanical properties.

The XRD profiles show gradual development of hydration products in GGBS-, G-MK, and PSA-based geopolymer mortars, as shown in Figure 15. The control mixes consisted primarily of crystalline quartz (Q) and calcite (C), along with a minor amount of calcium silicate hydrate (C–S–H). The peaks observed with the inclusion of both metakaolin and PSA were more pronounced, particularly the C-S-H and calcium aluminosilicate hydrate (C-A-S-H) phases, particularly at extended curing periods (14 and 28 days). The data indicated a progressive enhancement in geopolymerisation and gel formation over time, with the 28-day mix (G92.5MK5PSA10) demonstrating the highest intensities, signifying notable advancement in the structural development of the binding gels.

These findings are supported by the FTIR spectra, which display wide and distinct O-H and H-O-H bands associated with bound water in the hydration products along with strong Si-O-T and Al-O-Si bands, indicating the formation of a geopolymeric gel. The modified blends exhibited stronger band intensities and shifts than the control mix, particularly in the Si–O–Si and Al–O–Si regions, indicating that the silicate and aluminate species were polymerised into C-A-S-H networks. Therefore, the XRD and FTIR results are consistent, showing that the addition of metakaolin and PSA increases the dissolution–polymerisation rate, resulting in an increase in gel formation and enhancement of structural densification with increasing gel age.

The FTIR spectra of the control and blended mixes showed the chemical changes in the geopolymeric gels over time period (Figure 16). The control mix exhibited a broad band in the region of 950–1100 cm^-1, which is a stretch related to Si-O-T (T = Si or Al) asymmetric stretching vibrations, which are known to characterise CSH-type gels developed as a result of GGBS hydration (Fernández-Jiménez and Palomo, 2005). A weak band at 450–600 cm^-1 is a sign of Al-O-Si bending vibrations, which show a lack of aluminosilicate network formation in the absence of metakaolin or PSA. The band observed at 1630 cm^-1 corresponds to the bending of H-O-H in physically bound water, while the broad band ranging from 3200 to 3600 cm^-1 is associated with the O-H stretching vibrations of hydroxyl groups and the adsorbed water molecules present within the gel structure (Bakharev, 2005).

The addition of metakaolin and paper sludge ash resulted in an increased intensity of the Si-O-T stretch (950–1100 cm^-1) and a slight shift in position to lower wavenumbers, indicating improved polymerisation and replacement of Si in the aluminosilicate skeleton with Al (Davidovits, 2008). The A-O-Si bending vibrations in the 450–600 cm^-1 region were also stronger than those of the control, which proved that more Al was involved, as well as PSA. The intensity of the Si–O –T band at 14 days was further enhanced, with a corresponding weakening in the intensity of the O-H- and H-O-H-related peaks, which showed that free water was gradually consumed as the gel was formed. After 28 d, Si-O-T band peaks were observed, indicating a highly interconnected aluminosilicate structure and stable C-A-S-H gel formation. Meanwhile, the decreased intensity of the hydroxyl-related peaks (1630 and 3200–3600 cm^-1) also served as an indicator of a reduced number of unreacted hydroxyl groups and increased bound water in the hardened matrix (Provis, 2018).