Monitoring Early-Stage Evolution of Free Water Content in Alkali-Activated Slag Systems by Using ¹H Low-Field NMR

Abstract

In the present study, the evolution of free water content in five different alkali-activated slag (AAS) systems was continuously monitored and compared using ¹H low-field NMR. The alkali activators used were waterglass solutions with three different moduli (1.2, 1.4, and 1.6), sodium hydroxide solution, and sodium sulfate solution. The findings reveal that the type of activator significantly affected the dynamic changes in the relative free water content. Notably, an increase in free water content was observed in the early stages of hydration of all AAS systems except for those activated by sodium sulfate solution. Additionally, this study investigated the relationship between changes in free water content and hydration heat in the AAS systems, dividing the initial 24 h of AAS hydration into three stages. The results demonstrate that free water can serve as an effective probe for monitoring the hydration process in fresh AAS pastes, offering valuable insights alongside traditional thermal signals.

1. Introduction

Alkali-activated slag (AAS) is recognized as a sustainable cementitious material and has been proposed as one of the viable substitutes for Portland cement in recent decades. Its use has the potential to significantly reduce energy consumption and carbon emissions associated with clinker production [1,2,3,4]. However, certain performance limitations of AAS, such as very short setting time [5,6] and high shrinkage [7,8], have constrained its widespread adoption. To enhance the utilization of AAS, many researchers have employed various techniques to investigate its hydration mechanisms and microstructure. These studies aim to understand the changes in macroscopic properties and identify potential modification methods [9]. However, due to the diverse sources of raw materials for AAS and variations in alkali activators and precursor properties [10,11], the hydration mechanism of AAS remains not fully understood.

In recent years, the ¹H low-field NMR technique has become extensively used in the research of cement-based materials. Water, a fundamental component in all cementitious systems, undergoes various transformations during hydration, shifting from free water to chemically bound water, physically absorbed water, and pore water. By correlating these transitions with hydration kinetics, the hydration process can be analyzed through the proton signals of water molecules in their different binding states [12]. Liu et al. [13] quantitatively assessed the hydration degree of cement pastes with water–cement ratios of 0.25, 0.35, and 0.45 under various curing conditions using ¹H low-field NMR to analyze transverse relaxation time (T₂) distributions. Yang et al. [14] employed ¹H low-field NMR to calculate the saturation degree and pore structure of the cement paste under different saturation methods. In our previous study, the liquid absorption-release behavior of SAPs in cement pastes and AAS pastes was compared using ¹H low-field NMR [15]. The influence of the pre-absorbed water amount and the water–cement ratios of paste on the water absorption-release behavior of SAPs in fresh cement paste was investigated by ¹H low-field NMR [16]. In addition to cement-based materials, the ¹H low-field NMR technique has gradually been applied to the study of alkali-activated materials (AAM) systems. Liang et al. [17] used ¹H low-field NMR to monitor the evolution of the state and relative contents of the water in metakaolin-based geopolymers prepared from waterglass activators of different moduli with curing time. Jiang et al. [18] used ¹H low-field NMR to monitor the AAS at various ages and estimated the average gel pores size by calculating the weighted average of transverse relaxation time. Zhang et al. [19] employed ¹H low-field NMR to study the hydration process of waterglass-activated fly ash-slag pastes, finding that the paste activated with low modulus showed a higher degree of hydration at an early age. Compared to cement-based materials, AAM systems are highly complex, with their hydration processes significantly influenced by the composition of the cementitious materials and the type of activators used [20,21]. Currently, there is a lack of study on the application of ¹H low-field NMR in investigating the early hydration process of AAM. Addressing this gap is essential for enhancing the understanding and comprehension of AAM hydration mechanisms.

As one of the representative systems of AAM, the early hydration process of AAS has garnered significant attention. Gebregziabiher et al. [22] used an in situ isothermal calorimetric method to characterize the early reaction kinetics and microstructure development of AAS pastes. Cao et al. [23] performed in situ monitoring of the early reaction process of AAS pastes using an ultrasonic monitoring system. In our previous study, the volume change in AAS pastes was measured by helium pycnometry, and the relationship between the volume change in AAS pastes and early hydration was analyzed in conjunction with the hydration exotherm [24]. In this study, the evolution of free water content during the early hydration process of the different AAS systems was continuously and comparatively monitored using ¹H low-field NMR with water as a probe. Additionally, the relationship between the changes in free water content and hydration heat in various AAS systems was explored.

2. Materials and Methods

2.1. Materials

In this study, ground granulated blast furnace slag (GGBFS) with a d₅₀ of 10.9 μm was used as a precursor. Its chemical composition (determined by X-ray fluorescence) is given in

Table 1. Chemical composition of GGBFS (wt.%).

Oxides	CaO	SiO2	Al2O3	MgO	SO3	Na2O	Others
Content	34.06	33.80	17.67	9.56	2.38	0.71	1.82

. There are 5 different AAS systems, and their alkali activators are waterglass (Na₂SiO₃) solution with different moduli (SiO₂/Na₂O = 1.2, 1.4, and 1.6), sodium hydroxide (NaOH) solution, and sodium sulfate (Na₂SO₄) solution.

2.2. Mixture Proportions and Mixing Procedures

The alkali equivalent (W_Na2O/W_slag) for all AAS pastes was 5%, and the water-GGBFS ratio was 0.40. GGBFS was mixed with the prepared alkali activator solutions for 1 min. The codes WG1.2, WG1.4, WG1.6, SH, and SS indicate alkali activator was waterglass with a modulus of 1.2, 1.4, and 1.6, NaOH solution, and Na₂SO₄ solution, respectively.

2.3. H Low-Field NMR

An amount of 20 g of the fresh AAS pastes was poured into a quartz glass vial and sealed with polytetrafluoroethylene film for ¹H low-field NMR testing. The vial was placed into the sample chamber of the ¹H low-field NMR instrument with the chamber cap secured for continuous sampling, with a sampling interval of 5 min and a total duration of 24 h, by which time the free water content change in the early stage of the AAS systems had largely stabilized. Figure 1 shows the ¹H low-field NMR instrument. The magnetic field and radiofrequency (RF) coil of the 1H low-field NMR instrument (PQ001, Niumag, China) were 0.5 T and 25 mm, respectively. The CPMG (Carr–Purcell–Meiboom–Gill) sequence was used to measure the transverse relaxation time (T₂) of the paste. Then, the T₂ relaxation curves were fitted to a multi-exponential curve using the inverse Laplace transform algorithm. The echo time (τ1 = 0.302 ms) and the number of scans (NS = 4) were kept constant throughout the testing.

2.4. Isothermal Calorimetry

The hydration temperature rise of AAS pastes was measured using isothermal calorimetry in accordance with GB12959-2008 [25]. The exothermic of these paste samples was continuously recorded over a 24 h hydration period.

3. Results and Discussions

3.1. Transverse Relaxation Time (T₂) Distribution of AAS Pastes

Figure 2 shows the T₂ distribution of AAS pastes at 5 min, 15 min, 50 min, and 24 h after mixing. T₂ is related to the size of the pores where free water is located; the smaller the T₂, the smaller the corresponding pore size. Therefore, it is possible to distinguish water in different pore sizes according to the T₂ distribution. The T₂ values range from short to long, corresponding to interlayer and gel water (0.01~1 ms), capillary water (1~100 ms), and surface water (100~10,000 ms). For all samples, as hydration time increases, the T₂ peak gradually shifts towards shorter relaxation times, indicating the gradual refinement of the pore space in the pastes. The T₂ peak area reflects the relative content of free water—the larger the area, the higher the free water content. Thus, changes in the free water content in AAS pastes can be analyzed based on the changes in the total T₂ peak area.

3.2. Total T₂ Peak Area and Relative Free Water Content

To clearly illustrate the changes in relative free water content in different AAS pastes, Figure 3 presents the total T₂ peak area per unit weight of AAS pastes over time. The free water content in all AAS pastes decreases rapidly during the first ten minutes, likely due to the wetting and dissolution of GGBFS particles and the initial formation of C-S-H [22,23,26]. In the subsequent 2~4 h, the free water content in the three WG systems begins to increase rapidly, even surpassing the initial levels. An increase in free water content was also observed in the SH system, although the increase was significantly less than that in the WG systems. However, no increase in free water content was noted in the SS system.

For WG, the predominant silicate monomers in solutions are Si(OH)₃O⁻ and Si(OH)₂O₂²⁻ [27,28,29,30], and a small amount of Si(OH)₄ [28,29]. When the waterglass solution interacts with GGBFS particles, its high pH environment breaks the Ca-O, Mg-O, Al-O, and Si-O bonds on the surface of the GGBFS particles. This leads to the release of Ca, Mg, Al, and Si into the pore solution in the forms of Ca²⁺, Ca(OH)⁺, Mg²⁺, Al(OH)₄⁻, Si(OH)₃O⁻, and Si(OH)₂O₂²⁻ (Equation (1)) [27,30,31]. These introduced ions disrupt the electrostatic repulsion between the original Si(OH)₃O⁻ and Si(OH)₂O₂²⁻ in the pore solution, thereby facilitating the aggregation of the silicate structure and leading to dehydration condensation and the release of water (Equation (2)) [32]. Simultaneously, Al in the form of Al(OH)₄⁻ formed a complex with Si simplified to silicate monomers Si(OH)_4−yO_y^y− by a condensation reaction (Equation (3)) [31]. The above two processes are the primary sources of free water generation. The free water content increases with the elevation of modulus, which can be attributed to the fact that when the alkali equivalent for all AAS pastes is constant, a higher modulus introduces more Si(OH)₃O⁻ and Si(OH)₂O₂²⁻ into the system, thereby facilitating the aforementioned reactions.

$$\mathrm{GGBFS}\to {\mathrm{Ca}\left(\mathrm{OH}\right)}_{\mathrm{x}}^{\left(2-\mathrm{x}\right)+}+{\mathrm{Mg}}^{2+}+{\mathrm{Al}\left(\mathrm{OH}\right)}_4^{-}{+\mathrm{Si}\left(\mathrm{OH}\right)}_{4-\mathrm{y}}{\mathrm{O}}_{\mathrm{y}}^{\mathrm{y}-}\mathrm{x}=0,1 \mathrm{y}=1,2$$

(1)

$$2\mathrm{n}\equiv \mathrm{Si}-\mathrm{OH}\to \mathrm{n}\equiv \mathrm{Si}-\mathrm{O}-\mathrm{Si}\equiv +\mathrm{n}{\mathrm{H}}_2\mathrm{O}$$

(2)

$${\mathrm{Al}\left(\mathrm{OH}\right)}_4^{-}+{\mathrm{Si}\left(\mathrm{OH}\right)}_{4-\mathrm{y}}{\mathrm{O}}_{\mathrm{y}}^{\mathrm{y}-}\to {\left(\mathrm{OH}\right)}_3\mathrm{Al}-{\mathrm{OSi}\left(\mathrm{OH}\right)}_{3-\mathrm{y}}{\mathrm{O}}_{\mathrm{y}}^{\left(\mathrm{y}+1\right)-}+{\mathrm{H}}_2\mathrm{O}\mathrm{y}=1,2$$

(3)

For SH, sodium hydroxide creates a high pH environment that breaks the Ca-O, Mg-O, Al-O, and Si-O bonds on the surface of the GGBFS particles. However, in these systems, the presence of Al(OH)₄⁻, Si(OH)₃O^−, and Si(OH)₂O₂²⁻ arises solely from the dissolution of GGBFS, without additional introduction of these species. Consequently, the consideration reaction is limited, producing much less free water compared to the WG system.

For SS, the activator does not contain Si(OH)₃O⁻ and Si(OH)₂O₂²⁻, and the weak alkalinity of this system cannot break the Ca-O, Mg-O, Al-O, and Si-O bonds on the surface of the GGBFS particles [27,33]. Therefore, it does not cause a significant increase in free water content.

3.3. Evolution of Relative Free Water Content and Exothermic Hydration

Figure 4 presents the comparative curves of free water relative content and hydration temperature rise for WG1.2, WG1.4, and WG1.6. To validate the effectiveness of the ¹H low-field NMR technique in characterizing the hydration process of AAS pastes, both the free water content change curves and hydration temperature rise curves were fitted for different stages (fitting parameters are indicated in

Table 2. Polynomial fitting parameters of free water content curves.

Line	Intercept	B1	B2	B3	B4	Adjust R-Square
WG1.2-T1′	355.85957	−92.24463	109.31328	−53.5594	9.81192	0.90438
WG1.2-T2′	660.8592	−573.7087	350.52311	−88.11199	7.95984	0.99458
WG1.2-T3′	232.71932	84.8228	−18.95517	1.72678	−0.04988	0.98493
WG1.4-T1′	357.707	−21.07717	−1.95053	10.66271	−2.36412	0.96426
WG1.4-T2′	7199.57682	−10,297.61	5766.71376	−1423.75936	130.95654	0.999157
WG1.4-T3′	234.53892	106.46055	−29.40542	3.61868	−0.16683	0.99389
WG1.6-T1′	-	-	-	-	-	-
WG1.6-T2′	340.47114	34.03166	−39.23835	20.97967	−3.29253	0.99491
WG1.6-T3′	241.07024	111.56595	−31.58026	4.18513	−0.21238	0.98795

Note: Equation: y = Intercept + B1 × x + B2 × x2 + B3 × x3 + B4 × x4.

and

Table 3. Polynomial fitting parameters of hydration temperature rise curves.

Line	Intercept	B1	B2	B3	B4	B5	Adjust R-Square
WG1.2-T1	−0.68378	16.27285	−19.78803	10.0948	−1.86767	0	0.99951
WG1.2-T2	10.77092	−11.5658	7.52827	−2.3281	0.28242	0	0.99453
WG1.2-T3	−169.2431	107.5423	−23.50679	2.23757	−0.07947	0	0.99839
WG1.4-T1	0.38426	12.72088	−12.41088	5.1256	−0.7771	0	0.91368
WG1.4-T2	−0.33207	6.90352	−2.51546	−0.0326	0.10002	0	0.97141
WG1.4-T3	−140.0046	94.63683	−21.59335	2.12849	−0.07791	0	0.99572
WG1.6-T1	0.11221	14.984	−14.73204	6.62838	−1.45539	0.12918	0.97498
WG1.6-T2	0.11221	14.984	−14.73204	6.62838	−1.45539	0.12918	0.97498
WG1.6-T3	−80.70401	46.00336	−7.13968	0.22497	0.01618	0	0.98267

Note: Equation: y = Intercept + B1 × x + B2 × x2 + B3 × x3 + B4 × x4 + B5 × x5.

). In the hydration temperature rise curves, the points T₁, T₂, and T₃ represent the zero points of the first derivative, indicating the peak moments of hydration temperature rise. In the ¹H low-field NMR curves, T_1′ marks the zero point of the first derivative, corresponding to the moment of minimum free water content. The T_2′ and T_3′ are the zeros of the second derivative, signifying the turning points in the rate of change in free water content.

Based on these two curves, the hydration process of WG is divided into three stages: the dissociation stage (Stage I), the transition stage (Stage II), and the diffusion-controlled stage (Stage III). The transition stage is further subdivided into the deceleration stage (Stage II-a) and the acceleration stage (Stage II-b).

Stage I: In the hydration temperature rise curve, an initial peak appears during this stage. In the free water content curve, a rapid decrease in free water content is observed. This decrease is attributed to the early wetting, dissolution of GGBFS particles, and the initial formation of C-S-H [22,23,26]. When GGFSS dissolves in an alkali environment, the bond energies of Si-O and Al-O are higher than those of Ca-O and Mg-O, causing Ca and Mg to dissolve first. As a result, a low Si-rich layer consisting of C-S-H, C-A-H, and C-A-S-H with a low Ca/Si ratio quickly forms on the surface of GGBFS, as shown in Figure 5a [27].

Stage II: As shown in Figure 5b, as the hydration of GGBFS processes, the silicate structure composing the Si-rich layer not only forms a complex with Al(OH)₄⁻ dissolved in solution (Equation (3)), producing a small amount of water, but also re-adsorbs Ca²⁺, Ca(OH)⁺, and Mg²⁺ ions that are dissolved in solution through strong electrostatic attraction. This results in the formation of a passivation layer with a positive surface charge and a negative internal charge. This passivation layer inhibits further hydration, marking the beginning of Stage II-a [31]. As hydration is suppressed, the temperature curve shows a decreasing trend, while the free water content curve exhibits a sharp increase. As previously mentioned, this can be attributed to dissolved ions from GGBFS disrupting the electrostatic repulsion between the original Si(OH)₃O⁻ and Si(OH)₂O₂²⁻ in the pore solution. This disruption facilitates the aggregation of the silicate structure, leading to dehydration, condensation, and the release of water (Equation (2)) [32]. Water that was originally bound gradually transforms into free water. Coupled with the impeded hydration, the rate of free water production exceeds its consumption. As shown in Figure 5c, with the continuous dissolution of GGBFS, Si(OH)₃O⁻ and Si(OH)₂O₂²⁻ in pore solution rapidly complex with dissolved ions (e.g., Al(OH)₄⁻ as seen in Equation (3)), thereby weakening the surface passivation and promoting further dissolution, marking the beginning of Stage II-b [34,35]. A second peak in the hydration temperature rise curve occurs at this stage, indicating the formation of hydration products in large quantities [36]. Notably, while the content of free water continues to increase at this stage, the rate of increase gradually decreases. This is because, in Stage II-b, ions in solution still undergo dehydration condensation reactions, but water consumption accelerates significantly due to the weakening of the passivation layer and the further increase in hydration. This results in a gradual decrease in the net rate of free water production.

Stage III: The hydration temperature rise curve begins to decline, and by 24 h, the temperature of the AAS pastes approaches room temperature. In the free water content curve, no significant changes are observed. At this stage, the reaction rate slows considerably and reaches a stable phase, with hydration predominantly controlled by the diffusion rate, as shown in Figure 5d.

4. Conclusions

This study aims to continuously monitor the early hydration process of AAS pastes using the ¹H low-field NMR technique. The hydration of AAS pastes at their early age was divided into three distinct stages based on changes in free water content and the hydration temperature rise. These findings offer valuable insights into the early reaction mechanisms of AAS. The key conclusions are summarized as follows:

The type of activator significantly affects the change in free water content. AAS pastes activated by waterglass show a substantial increase in free water content at the initial stages, with the increase being more pronounced at higher waterglass moduli. In contrast, AAS pastes activated with sodium hydroxide exhibit only a modest rise in free water content, while activation with sodium sulfate does not result in any increase.

The ions dissolved from the GGBFS in the alkaline environment disrupt the electrostatic repulsion between the ions in the waterglass, promoting further aggregation of the silicate structure. Additionally, these ions may react directly with the waterglass. These processes are likely contributors to the observed increase in free water content.

Based on the observed changes in free water content and the corresponding hydration temperature rise, the early hydration process of WG can be categorized into three stages: dissociation (Stage I), transition (Stage II), and diffusion control (Stage III). The high correlation between the changes in free water content and hydration temperature rise suggests that water may serve as a useful probe, in addition to heat signals, for monitoring the hydration process of fresh AAS pastes.