Add sodium silicate and deionized water into a beaker, stir well with a glass rod, then add accurately predesigned amount of sodium hydroxide, stir until the particles are completely dissolved, cover with plastic wrap to prevent moisture volatilization. Make the activator stand by for about 4 h, until the solution cools to room temperature and becomes clear before use. For every 100 g of sodium silicate solution prepared into an alkali activator of modulus n, the mass x of sodium hydroxide required is determined by Equation 1. n = 26 / 60 8.2 / 62 + x / 80 = 0.433 0.132 + x / 80 (1)

The alkali activator and precursor materials were mixed according to the designed proportions. The mixture was first stirred at a low speed for 1–3 min to ensure uniform wetting, followed by high-speed mixing for an additional 2–5 min to achieve full homogenization. Immediately after mixing, the fresh paste was injected into the test molds. Compressive strength specimens were prepared using silicone molds with dimensions of 20 mm × 20 mm × 20 mm, with three replicates for each group. The paste was manually compacted and then placed on a vibrating table for 30–60 s to remove entrapped air, followed by surface leveling. After casting, the specimens were covered with plastic film and cured at 25 °C for 24 h. They were then demolded and transferred to standard curing conditions (25 °C ± 2 °C and RH ≥ 95%.) until the designated testing ages. Compressive strength tests were performed at the specified curing times.

To evaluate the influence of metakaolin based polymer composition on compressive strength and fluidity, orthogonal test was designed. It studied the combination of several factors at different levels to determine the optimum ratio conditions. In the orthogonal test, the modulus of sodium silicate (factor A), silica fume content (factor B) and liquid-solid ratio (factor C) were selected. Combined with the previous research results from the research group (Wang et al., 2025; Ma et al., 2020), the levels of each factor are shown in Table 2.

The L⁹(33) orthogonal experimental design with 3 factors and 3 levels was employed. The experimental schemes are shown in Table 3.

To investigate the influence of different factors on compressive strength of metakaolin based polymer, this study calculated the heat collection efficiency and maximum cooling value ki for each factor at level i, as well as the range R between all ki of the same factor (with i values for factors A, B, and C being 1 to 3). The specific calculation methods are shown in Equations 2, 3. k i = ∑ v i n (2) R = max k 1 , k 2 , k 3 − min k 1 , k 2 , k 3 (3) where kᵢ represents the compressive strength when a factor is at level i; R value reflects the degree of influence of the factor on compressive strength. The larger the R value, the more significant the influence on compressive strength of geopolymer.

Compressive strength tests of the metakaolin-based geopolymer at the ages of 3, 7, and 28 days were conducted using an electrohydraulic servo press with a capacity of 300 kN and a loading rate of 2 mm/min. The testing procedure followed the requirements of ASTM C109.

To evaluate the working performance of metakaolin geopolymer paste, the fluidity of metakaolin geopolymer was measured according to the cement paste fluidity measurement method specified in the standard GB/T8077-2012 “Test Method for Homogeneity of Concrete Superplasticizer.”

To study the influence of superplasticizer on bleeding volume of geopolymer slurry, reference experiments were carried out with high and low additive contents. In high dosage group, melamine, sodium lignosulfonate and naphthalene superplasticizer were 2.5%, polycarboxylic acid 1%, silane coupling agent KH550 0.2 mL, while in low dosage group, melamine, sodium lignosulfonate and naphthalene superplasticizer were 1%, polycarboxylic acid 0.5%, silane coupling agent 0.1 mL. Weigh 20 g metakaolin and the corresponding amount of superplasticizer into a measuring cylinder, stir evenly with a glass rod, add sodium silicate solution with modulus of 0.9 and mix well to make a constant volume of 100 mL. The bleeding volume is the supernatant volume of the slurry, record every 20 min.

Mix 0.1 g additive with 8 mL sodium silicate solution with different modulus (0.8, 1.4, 2.0), observe the stability of additive in appearance, such as agglomeration, separation, or color change. According to the principle of maximum wavelength selection, the optimum determination wavelengths of melamine, sodium lignosulfonate, naphthalene series, polycarboxylic acid and KH550 were 236 nm, 428 nm, 427 nm, 272 nm and 301 nm respectively; Secondly, prepare several additive solutions with known concentrations and test their absorbance at corresponding detection wavelengths; mix 0.1 g of additive with 8 mL of sodium silicate solution with different moduli (0.8 and 1.4), test the absorbance of supernatant with ultraviolet visible spectrophotometer and substitute it into the standard curve fitting equation to obtain the actual concentration C'. Calculate the insolubility η as per Equation 4. η = C 0 − C ′ C 0 × 100 % (4)

Fourier transform infrared spectroscopy (FT-IR) was employed to analyze the structural change of superplasticizers and characterize the chemical stability of different superplasticizers. The measurement range was 4,000 to 400 cm⁻¹ with the resolution of 4 cm⁻¹ and 32 scans (Hu et al., 2024).

To study the effect of superplasticizer on the surface tension of slurry, a sodium silicate solution with modulus of 0.9 was selected and mixed according to the proportion of 0.1 g additive per 8 mL sodium silicate solution. After mixing, the solution was allowed to stand for 24 h, and then the supernatant was processed by centrifuge. The surface tension was measured by ADVANCEKRUSS (Germany) instrument. Water solution was selected as the reference.

The zeta potential of metakaolin suspensions was measured to evaluate the influence of different superplasticizers under varying sodium silicate moduli. In the preparation process, the alkali activator, metakaolin, and the designated additive were successively added into a beaker and stirred magnetically for 2 min to ensure uniform dispersion. Subsequently, a 2 mL aliquot of the suspension was immediately extracted and transferred into the sample cell for zeta potential measurement. All tests were conducted at room temperature to minimize environmental interference, and each measurement was repeated three times to ensure accuracy and reproducibility.

Because the mechanical performance of geopolymers is highly sensitive to mix-design variables, this subsection compares the influence of silicate modulus, silica fume content, and liquid-to-solid ratio on compressive strength and strength development. The objective is to identify the formulation that maximizes early and long-term performance. It can be seen from Table 4 that the combination of A2B1C1 (i.e., silicate modulus 0.9, silica fume content 15%, liquid-solid ratio 0.75) showed the highest compressive strength at all ages, which was significantly better than other ratio combinations. Combined with the range and variance analysis results in Table 5, it can be seen that the liquid-solid ratio (C) and the silicate modulus (A) have significant effects on the strength development (p < 0.05), where the liquid-solid ratio is the dominant factor, the optimal level is C1 (0.75), and the optimal level of sodium silicate modulus is A2 (0.9). In contrast, the effect of silica fume content (B) is not significant, but the strength performance is better at B1 (15%). Therefore, A2B1C1 is not only verified statistically, but also shows mechanical advantage in measured properties. In addition, although the fluidity (132 mm) of the combination is lower than that of the higher liquid-solid ratio group, it is still in an acceptable range and can meet the construction requirements.

To further analyze the increasing trend of compressive strength of geopolymer, the results of nine groups of specimens at different ages are shown in Figure 2. The compressive strength of metakaolin geopolymer at 3d age can reach 40%∼70% of 28d strength, and the compressive strength at 7d can generally reach more than 80% of 28d compressive strength, which has significant early strength. When the strength level is lower (<20 MPa), the compressive strength of 3 days is similar with that of 28 days. When the strength level is higher (>20 MPa), the compressive strength of 7 days is obviously higher than that of 3 days, but the increase of 28 days is limited, which indicates that the strength of metakaolin-based polymer mainly develops rapidly in the early stage and tends to be stable.

As shown in Figure 2, the A2B1C1 mixture exhibited the highest compressive strength and growth rate among all specimens. Its strength reached nearly 40 MPa at 3 days, increased to about 55 MPa at 7 days, and further rose to nearly 60 MPa at 28 days, showing a typical pattern of rapid early development followed by stabilization at later ages. This superior performance is attributed to the optimized silicate modulus (0.9), silica fume content (15%), and liquid-to-solid ratio (0.75), which promote aluminosilicate dissolution and dense gel formation. Similar results were reported by Hattaf et al. (2021) and Gao et al. (2014), who found that a moderate silicate modulus facilitates early geopolymerization and improves structural compactness, leading to enhanced strength and long-term stability.

It can be seen from Table 5 that, in terms of fluidity, the effect of sodium silicate modulus (A) has statistical significance at the significance level of 0.05 (p = 0.0363), while the effect of silica fume content (B) and liquid-solid ratio (C) does not reach the significance level within the investigation scope of this study (p = 0.7019 and 0.2051). The results of range analysis were consistent with the results of variance analysis. The order of factors affecting fluidity was A (R = 15.00) > C (R = 6.33)>B (R = 2.00). According to the trend of average values, the average values of A from A1 to A3 are 117.33, 131.67 and 132.33 respectively, showing a monotonic increasing trend; the average values of C from C1 to C3 are 124.33, 126.33 and 130.67 respectively, also showing an increasing trend; and the average values of B from B1 to B3 are 128.33, 126.67 and 126.33 respectively, showing a slight decreasing trend overall. According to the analysis, the optimum combination is A3B1C3 (modulus 1.0, silica fume 15%, liquid-solid ratio 0.95) for highest working performance.

It should be noted that when the strength and fluidity are optimized simultaneously, the optimal levels of A and C are not consistent. The silicate modulus influences the solution viscosity and polymerization behavior by regulating the SiO₂/Na₂O ratio. When the modulus is too low and the alkalinity is excessively high, the solution exhibits high viscosity, which hinders the polycondensation process. If the modulus is too high, the reaction rate will be insufficient. Moderate modulus is more conducive to balancing mechanical properties and workability.

Moreover, reducing the liquid-to-solid ratio increases reactant concentration and matrix compactness, improving strength but reducing fluidity due to limited free water. In contrast, increasing this ratio enhances workability but weakens strength because of dilution and higher porosity (Albidah et al., 2021; Luan et al., 2021). And optimization of activator modulus together with water content is critical for achieving balanced mechanical and rheological performance (Park et al., 2022). Based on these findings and the results summarized in Table 4, a silicate modulus of 0.9, silica fume content of 15%, and liquid to solid ratio of 0.75 were selected as the reference mix for subsequent additive performance evaluation.

The analysis in this subsection focuses on establishing how different superplasticizers modify the initial workability of metakaolin-based geopolymer pastes. According to the strength test results, the dosage of sodium lignosulfonate, naphthalene-based additives and melamine are 0.5%, 1%, 1.5%, respectively. The dosages of polycarboxylic acid additives are 0.1%, 0.4%, 0.7%, respectively, and the silane coupling agent KH550 are 0.2 mL, 0.5 mL, 0.8 mL, 1.1 mL, respectively. The results of fluidity test are shown in Figure 3. The initial fluidity of the metakaolin-based polymer without superplasticizer is 122 mm.

As shown in Figure 3, the improvement effect of various superplasticizers on the fluidity of metakaolin based polymer is as follows: sodium lignosulfonate > silane coupling agent > melamine > polycarboxylic acid > naphthalene series superplasticizer. Sodium lignosulfonate and silane coupling agents initially increased and then decreased with the increase of dosage. Specifically, when 0.5% sodium lignosulfonate was added, the fluidity of slurry increased to 127 mm; when the addition amount increased to 1.0%, the initial fluidity further increased to 130 mm, which was about 7% higher than that of the reference. When the dosage continued to increase to 1.5%, the fluidity decreased slightly. The optimal dosage of silane coupling agent was about 0.8 mL, and the fluidity improvement effect was similar to that of sodium lignosulfonate. In contrast, melamine, naphthalene series and polycarboxylic acid superplasticizer had negative effects overall. At a dosage of 0.5%, melamine achieved an initial flowability of 125 mm, which was slightly higher than that of the control group. However, its effect was noticeably weaker than that of sodium lignosulfonate. When the dosage increased to 1.5%, the flowability decreased to 118 mm. The naphthalene-based superplasticizer yielded an initial flowability of only 115 mm at 0.5% dosage, representing a 5.7% decrease compared with the reference, and further decreased to 107 mm at 1.5% dosage, a reduction of 12.3%. The polycarboxylate superplasticizer exhibited a similar trend: at a dosage of 0.7%, the flowability dropped to 112 mm, 8.2% lower than that of the reference. This may be attributable to the poor compatibility of melamine with highly alkaline systems, while naphthalene-based and polycarboxylate superplasticizers tend to adsorb or complex with reactive silicate ions in the metakaolin–sodium silicate system, weakening their dispersing effect (Derkani et al., 2022). This interaction promotes early structural flocculation of the paste, thereby suppressing flowability.

To further evaluate the time-dependent workability performance, this subsection examines the evolution of flowability loss under various superplasticizers. The bleeding volume of geopolymer slurry with five different superplasticizers at high and low superplasticizers is shown in Figure 4. Without superplasticizer, the bleeding volume of metakaolin geopolymer in 120 min is 2.5 mL ∼3 mL.

As shown in Figure 4, at high dosage, the five kinds of superplasticizer did not effectively inhibit slurry bleeding. Polycarboxylic acid additive had the largest bleeding volume, with bleeding volume of 8 mL at 120 min, followed by sodium lignosulfonate. This phenomenon can be attributed to the competitive adsorption of superplasticizer molecules on particle surfaces and in the solution at high dosages, resulting in the saturation of effective dispersion sites (Zhang et al., 2021; Partschefeld et al., 2023). Consequently, the electrostatic repulsion and steric hindrance effects are weakened, which in turn accelerated particle reaggregation and water bleeding (Plank and Winter 2008).

In contrast, at lower content, except polycarboxylic acid, other superplasticizers showed certain improvement effect. At 120 min, the bleeding volumes of melamine, sodium lignosulfonate, naphthalene series superplasticizer, polycarboxylic acid and KH550 silane coupling agent were 2.3 mL, 1.9 mL, 2.4 mL, 5.0 mL and 2.0 mL respectively. Sodium lignosulfonate > KH550 > melamine > naphthalene series superplasticizer > polycarboxylic acid, which was basically consistent with the results of fluidity test, further indicating the advantages of sodium lignosulfonate and KH550 in improving the working performance of geopolymer.

The test results show that the high effective superplasticizer commonly used in cement system can improve the fluidity of metakaolin based polymer. In contrast, sodium lignosulfonate and KH550 showed better dispersion and water reduction effects, but the dosage should be strictly controlled to avoid the negative effects caused by excessive addition, among which the optimum dosage of sodium lignosulfonate was 1.0%, and the optimum dosage of silane coupling agent KH550 was 0.8 mL/150g.

This subsection investigates the bleeding behavior and physical stability of the geopolymer slurry to further clarify the influence of superplasticizers on mixture uniformity. Figure 5 shows the appearance characteristics of superplasticizers in sodium silicate solutions. The silicate modulus from left to right in the figure is 0.8, 1.4 and 2.0 respectively.

The results showed that the lower the modulus of sodium silicate, the worse the solubility and the more unstable the appearance of the system. When the modulus is 2.0, all the systems were clear and homogeneous except for polycarboxylic acid superplasticizer. When the modulus is 0.8, precipitation or stratification occurred in all five superplasticizers. Under the conditions of this study, the appearance stability from high to low is sodium lignosulfonate, naphthalene series water reducer > KH550 silane coupling agent, melamine > polycarboxylic acid superplasticizer (Luukkonen et al., 2018; Ren et al., 2023). This result highlights the adaptation advantage of lignin superplasticizer in strong alkali silicate environment.

The absorbance curves of the superplasticizer are shown in Figure 6, and the absorbance-concentration calibration fitting curves of the superplasticizer are shown in Figure 7.

Except KH55, the other superplasticizer have good linearity at their respective characteristic wavelengths (R² ≥ 0.995). The continuous hydrolysis-polycondensation of KH550 in water leads to the spectrum drifting with time, and it is challenging to obtain the steady-state concentration-absorbance relationship.

Table 6 shows the incompatibility rate of each additive in different modulus sodium silicate solutions. As shown in Table 6, with the decrease of silicate modulus (Ms), the insolubility of each additive increased, indicating that the solubility of the system decreased and the appearance stability deteriorated. Melamine had an insolubility of 94.06% and 97.73% when the silicate modulus was 1.4 and 0.8, respectively, indicating that its solubility was extremely low, making it difficult to reduce water content. Sodium lignosulfonate still maintains a certain solubility at low modulus (e.g., the insolubility was significantly lower than other superplasticizers when Ms = 0.8), which was consistent with its better workability than the same kind. Naphthalene superplasticizers were highly sensitive to modulus, rising sharply from 15.59% of Ms = 1.4 to 85.30% of Ms = 0.8, indicating that they are more suitable for medium and high modulus conditions. The solubility of polycarboxylic acid superplasticizer was poor, and the insolubility reached 78.47% and 92.16% respectively when Ms = 1.4 and 0.8. In general, the decrease of modulus was accompanied by the increase of ionic strength and pH, which easily initiated stratification and competitive adsorption, thus inhibiting the effective dissolution and dispersion of superplasticizer, while sodium lignosulfonate showed relatively better adaptability in this strong alkali silicate environment.

FTIR spectroscopy is employed to identify molecular-level interactions between the admixtures and the geopolymer network, providing insight into their chemical action mechanisms. Figure 9 illustrated the FTIR comparison of various water reducing agents in different modulus sodium silicate solutions.

L, N, P can be seen in water with their respective hydrophobic and hydrophilic fragments: aromatic (C=C/C-C) skeleton at around 1,600/1,510 cm⁻¹, C-N/C-O vibration at around 1,186 cm⁻¹, -CH/-CH stretching at around 2,881 cm⁻¹, and sulfonate characteristics (S=O at around 1,358 cm⁻¹, S-O at around 1,032 cm⁻¹), amino (N-H at around 3,469 cm⁻¹ and 3,416 cm⁻¹), and polyether segments (C-O-C at around 1,097 cm⁻¹). Upon exposure to sodium silicate, these peaks exhibited attenuation and slight shifts, signifying hydrolysis and ionic shielding effects under strongly alkaline and high-ionic-strength conditions. On one hand, labile bonds (e.g., ether or amide) underwent partial cleavage, resulting in reduced molecular weight and disrupted continuity of hydrophobic segments (Partschefeld et al., 2023). On the other hand, competitive adsorption of silicate species at polar sites impaired surface activation and dispersion efficacy, thereby corroborating the observed decline in workability (Ren et al., 2023).

The change of KH system was more typical: the signs of methoxy group to silanol (Si-OH) conversion (3,200–3,500 cm⁻¹ wide peak enhancement, 1,000–1,100 cm⁻¹ wide region Si-O-Si signal enhancement) indicate that silane hydrolysis-condensation occurs in sodium silicate, and Si-O-Si, Si-O-Al bonds were formed with the matrix (Issa and Luyt, 2019). Meanwhile, the change of C-H intensity indicates that the organic end groups were rearranged or partially broken. The interfacial reaction was beneficial to early network construction, but also shortened the machinable time, which was consistent with the result of low fluidity retention of KH group. Overall, KH-550 accelerates early network formation but leads to rapid loss of flowability, which accounts for its poor suitability for enhancing slurry workability.

This subsection assesses how each superplasticizer alters the surface tension of the system to clarify whether surface activity contributes to its water-reducing behavior. Select the sodium silicate solution with modulus of 0.9 to test the effect of additive type on the surface tension. The results are shown in Figure 10.

It can be seen from Figure 10 that the addition of superplasticizer reduced the surface tension of aqueous solution, among which melamine has the most significant effect, and the surface tension of aqueous solution decreases from about 72 mM/m to 65 mM/m (9.7% decrease). Sodium lignosulfonate, naphthalene series superplasticizer, polycarboxylic acids and KH550 also decreased by varying degrees. On the contrary, in sodium silicate medium, except KH550, the surface tension of other systems was higher than that of sodium silicate body (about 71 mN·m⁻¹). Moreover, the surface tension of the same superplasticizer in sodium silicate solution was significantly higher than its value in aqueous solution. These observations suggest that the improved fluidity of metakaolin-based geopolymers is not governed by surface-tension reduction, but is instead primarily driven by particle dispersion, electrostatic repulsion, and interfacial chemical interactions (Fan et al., 2023; da Silva et al., 2026). This behavior contrasts with that reported for cementitious systems (Lu et al., 2021), where surface activity plays a more prominent role, further confirming that surface tension is not the controlling mechanism for workability in alkaline geopolymer environments.

Zeta potential (ZP) represents the electrokinetic potential at the shear plane of suspended particles, indicating their surface charge and electrostatic interactions. A higher absolute ZP implies stronger repulsive forces between particles, leading to improved dispersion stability and enhanced slurry workabilit. The ZP of metakaolin-based polymer slurries with different superplasticizers are shown in Figure 11.

It can be seen from Figure 11 that, under a silicate modulus of 1.2, all superplasticizers significantly increase the absolute ZP of metakaolin particles, thereby enhancing electrostatic repulsion. The differences among the admixtures are evident. Sodium lignosulfonate and the KH-550 silane coupling agent show the largest increases in absolute ZP, rising from approximately 21 mV to about 40 mV and 45 mV, respectively. This notable increase aligns well with the observed improvement in slurry workability. Polycarboxylic acid also exerts a moderate effect, increasing the absolute ZP to around 30 mV. In contrast, when the silicate modulus is 0.8, the enhancement of electrostatic repulsion is generally weaker. Under this condition, sodium lignosulfonate and KH550 perform slightly better than melamine, whereas polycarboxylic-acid- and naphthalene-based superplasticizers have the weakest effect, with absolute ZP values even lower than that of the blank sample. This trend is consistent with the results of the macroscopic fluidity tests.

Consistent with the FTIR and surface-tension analyses, the zeta-potential results further confirm that changes in alkalinity significantly influence interparticle interactions in metakaolin-based geopolymers. When the silicate modulus decreases from 1.2 to 0.8, the absolute ZP of the blank sample increases from approximately 21 mV–30 mV, indicating that particle surface charging and electrostatic repulsion are enhanced with increasing alkalinity. In contrast, sodium lignosulfonate and KH550 exhibit a notable decrease in absolute ZP (about 20%), suggesting that their dispersion efficiency is suppressed under more alkaline conditions. The variation in polycarboxylic acid is relatively small, implying that it is less sensitive to modulus changes, whereas the slight increase observed for naphthalene- and melamine-based superplasticizers is mainly attributed to the intensified ionization of the sodium silicate matrix at lower modulus.

As the silicate modulus decreases, the electrostatic activity of the alkaline solution increases, but the resulting dispersion behavior varies markedly among different superplasticizers. KH-550 and sodium lignosulfonate show pronounced attenuation in dispersion performance due to electrostatic shielding under high alkalinity and elevated ionic strength, indicating that their effectiveness may be compromised in on-site geopolymer mixtures with strong alkaline activators. In contrast, polycarboxylic acid maintains relatively stable dispersion behavior, while naphthalene- and melamine-based superplasticizers exhibit only limited responsiveness to changes in modulus. These differentiated responses highlight the importance of selecting alkali-tolerant superplasticizers to ensure reliable workability in practical metakaolin-based geopolymer construction, particularly when activator modulus fluctuates during field mixing.