Towards Ultra-Rapid and High-Toughness Cementing: A Synergistic Acceleration Leveraging Aluminum Sulfate and Sodium Alginate Copolymer Along with Glass Fibers

Abstract

This study synthesizes two highly water-soluble copolymers, p(SA-co-SMAS) and p(SA-co-SMAS-co-AMPS) using sodium alginate (SA), sodium 2-methylprop-2-ene-1-sulfonate (SMAS), and 2-acrylamido-2-methylpropane sulfonic acid (AMPS, with or without addition) as precursors. Under ball milling, these copolymers are blended with aluminum sulfate and glass fibers to produce two series of cement admixtures. Compared to systems without admixtures or with pure aluminum sulfate as sole admixture, the admixture obtained from p(SA-co-SMAS) and aluminum sulfate significantly shortens the initial setting time (4.47 vs. 33.59 and 29.51 min) and final setting time (8.46 vs. 45.26 and 35.12 min), while markedly improving compressive strength (9.2 vs. 3.5 and 4.3 MPa) and flexural strength (3.5 vs. 1.0 and 1.1 MPa). This enhancement is attributed to the formation of a unique boehmite (AlO(OH)) phase in synthesized admixture, which rapidly reacts with tricalcium silicate, gypsum, and water in cement to form ettringite (Ca₆Al₂(SO₄)₃(OH)₁₂·26H₂O). The ettringite interlocks with the two-dimensional C–S–H gel, creating a stable three-dimensional network. Further blending this admixture with 200-mesh glass fibers yields a new admixture containing Al₄SO₄(OH)₁₀·36H₂O. Compared to boehmite, this phase further reduces setting times and increases average compressive strength (10.2 vs. 9.2 MPa). The admixture derived from p(SA-co-SMAS-co-AMPS) and aluminum sulfate shows even better performance: setting times are further shortened and flexural strength is significantly enhanced, owing to the presence of the more effective Al₄SO₄(OH)₁₀·36H₂O phase. Incorporating 200-mesh glass fibers into this system results in the shortest setting times (initial: 2.24 min, final: 5.73 min) and an excellent 24 h compressive strength (9.4 MPa), likely due to a unique and unexpected pore-filling effect. In contrast to conventional uses of sodium alginate as a retarder, glass fibers as mere reinforcements, and aluminum sulfate as a strength-impairing accelerator, this work demonstrates a synergistic strategy, which enables an ultra-rapid and high-strength cement setting process, offering highly significant scientific and practical value.

1. Introduction

In the application of cement, effective control over its hydration (setting) time and mechanical strength is crucial to the success of construction projects [1,2]. For instance, in the construction of oil wells or buildings in cold regions, if the cement sets too quickly, it may lead to structural deformation, significantly impairing oil and gas transport efficiency or building usability [3]. In such cases, retarders, superplasticizers, or water-reducing agents are often incorporated during cement hydration to slow down the setting process [3].

Conversely, in the construction of bridges, tunnels, and culverts, where shotcrete is commonly used, rapid setting of cement along with adequate mechanical strength is required. Here, accelerators are added to the cement hydration process [4]. In general, the setting time of pure cement without any additives is far too slow to meet the requirements of shotcrete applications, yet considerably faster than what is needed in scenarios requiring delayed setting. Therefore, the research and application of cement hydration admixtures have become an urgent priority.

Currently, aluminum sulfate-based accelerators, as representative chloride-free and alkali-free agents, are widely favored in both consumer markets and engineering applications. The mechanism by which aluminum sulfate accelerates cement hydration is as follows: in the alkaline environment of cement hydration, aluminum sulfate releases Al³⁺ ions, which hydrolyze to form [Al(OH)₄]⁻ [5]. This anion subsequently reacts with Ca²⁺ to form tricalcium aluminate (C₃A, 3CaO·Al₂O₃). The C₃A then reacts with gypsum (CaSO₄·2H₂O) to form ettringite (calcium sulfoaluminate hydrate, 3CaO·Al₂O₃·3CaSO₄·32H₂O, AFt) [5]. Ettringite exhibits a micro-nano fibrous morphology, which combines with similarly micro-nano scaled two-dimensional C–S–H gel sheets to form a three-dimensional network structure [5,6]. This significantly enhances the cement setting rate and has the potential to improve the mechanical strength of the hardened cement [5,6].

Although aluminum sulfate accelerators can significantly shorten the setting time, they also present several notable drawbacks. Firstly, while accelerating the setting process, they may lead to concentrated early release of hydration heat, increasing thermal stress and the risk of microcracking, which can potentially impair long-term strength and durability [7]. Secondly, aluminum sulfate exhibits limited solubility in water—only 36.5 g per 100 g of water at 20 °C [8]. Concurrently, the hydration process continuously produces calcium hydroxide (portlandite, Ca(OH)₂) from tricalcium silicate (C₃S), resulting in a steady rise in the pH of the system [9].

Consequently, Al³⁺ ions from aluminum sulfate do not remain stable in the cement hydration environment and are highly susceptible to alkaline hydrolysis, forming inactive Al(OH)₃ and Al₂O₃. Therefore, in practical applications, engineers often incorporate water-soluble organic or inorganic ligands, such as diethanolamine [10], triethanolamine [10], or sodium fluoride [11], to complex with Al³⁺, thereby controlling its hydrolysis rate and maximizing the accelerating effect of aluminum sulfate.

When investigating the process through which the added organic ligands enhance the performance of aluminum sulfate accelerator from a deeper mechanistic perspective, diethanolamine and triethanolamine serve as illustrative examples. Experimental results indicate that diethanolamine can stabilize the sparingly soluble aluminum sulfate in an aqueous solution by coordinating with Al³⁺ [10,12,13,14]. Subsequently, this complex reacts with Ca²⁺ and SO₄²⁻ ions released from gypsum dissolution to form AFt (ettringite). Upon releasing diethanolamine, the ligand further interacts with Ca²⁺ in the cement hydration system to form a water-soluble complex. This process significantly increases the concentration of soluble Ca²⁺, thereby promoting the formation of Ca(OH)₂ and accelerating the hydration of C₃S [10]. As a result, the cement hydration time is notably shortened, and mechanical strength is enhanced [10].

In contrast, the behavior of triethanolamine differs. Its coordination ability with Al³⁺ is stronger than that with Ca²⁺, and it also exhibits greater Al³⁺ affinity compared to diethanolamine [10,14]. Consequently, when triethanolamine is used as an auxiliary additive in aluminum sulfate accelerators, it accelerates the rapid formation of AFt. However, excessive and rapid AFt generation can encapsulate the surface of C₃S grains, hindering their hydration. As a result, the accelerating effect of triethanolamine is often less pronounced than that of an equivalent dosage of diethanolamine. Nevertheless, due to the relatively delayed hydration reaction mediated by triethanolamine, it may help to bridge micro-cracks formed during cement hydration, leading to improved mechanical strength at 24 h [10].

It is evident that the research strategy of incorporating additional organic ligands to stabilize aluminum sulfate accelerators is effective. However, polydentate ligands derived from the same central atom (with three or more coordination sites) likely exhibit lower accelerating efficiency compared to bidentate ligands derived from the same atom or linear ligands with more extended and flexible coordination geometries.

Overall, however, substantial practice and effort are still required to regulate the form and release pattern of aluminum sulfate in the cement hydration process through the addition of auxiliary agents, thereby facilitating the smooth progression of hydration and overcoming the limitation of insufficient early strength associated with conventional aluminum sulfate accelerators.

The role of natural or synthetic polymers in the cement hydration process is receiving increasing attention from both academia and industry. Generally, commonly used polymers such as sucrose [15], sodium lignosulfonate [16], and polycarboxylate ether (PCE) [17] function more as retarders than accelerators in cement hydration. These additives often serve as superplasticizers or water reducers, significantly slowing hydration to improve mortar fluidity and workability. However, natural or synthetic polymers that exhibit accelerating effects are quite rare. One possible explanation is that polymers introduced during cement hydration may coat the early-formed C–S–H particles, reducing the hydration reaction rate [18].

From another perspective, the selection of highly water-soluble polymers containing multiple coordinating atoms as additives for cement hydration may, when combined with aluminum sulfate, not only effectively coordinate with Al³⁺ (thereby preventing the ineffective hydrolysis and precipitation of aluminum sulfate) but also engage in coordination interactions with Ca²⁺. Furthermore, such polymers can encapsulate early-formed C–S–H particles. Owing to their high water solubility, they can thoroughly disperse these particles within cement pores, potentially not only accelerating the hydration reaction but also bridging microcracks generated during cement hydration, thereby enhancing the mechanical strength of the cement.

In the search for and development of polymers with cement-accelerating properties, special attention should be given to polymers or polymeric monomers featuring multiple coordinating atoms, as they hold significant potential and promising application prospects. For instance, sodium alginate (SA) is extracted from alginic acid in algal cell walls through an alkalization process [19]. It exhibits unique properties including biodegradability, non-toxicity, pH sensitivity, and non-immunogenicity [19]. With large-scale availability and low cost, it is widely used in the food, pharmaceutical, and nutraceutical industries [19]. Chemically, sodium alginate is a hydrocolloid—a linear copolymer composed of β-D-mannuronate (M) and α-L-guluronate (G) monomers [20]. It undergoes rapid cross-linking in the presence of Ca²⁺ ions, forming stable gels and films, which accounts for its broad industrial applicability [21].

For a long time, SA has been categorized among saccharides in cement hydration applications, where it exhibits a distinct retarding effect [22]. Extensive research further confirms that SA does not accelerate setting but acts solely as a retarder, albeit with the ability to enhance the mechanical strength of hardened cement [22]. Given its high sensitivity to Ca²⁺ and the abundance of Ca²⁺ ions in the cement hydration system, introducing SA into this environment could be meaningful for regulating the setting process. However, using SA alone—though potentially beneficial for strength development—offers limited functionality. Breakthrough improvements are expected only through functional modification of SA.

2-Acrylamido-2-methylpropane sulfonic acid (AMPS) is a water-soluble functional monomer containing both sulfonic acid and amide groups. Its polymers, such as poly(AMPS) or copolymerized derivatives, serve multiple important functions in cement-based materials. Firstly, the sulfonic acid group (–SO₃H) of AMPS fully ionizes to –SO₃⁻ in the high-alkaline environment of cement paste, disrupting the flocculated structure of cement particles through strong electrostatic repulsion [23]. Meanwhile, its polymeric chains provide steric hindrance, significantly reducing the water-to-cement ratio and enhancing fluidity [24]. Consequently, AMPS plays a significant role in maintaining mortar rheology and inhibiting rapid cement setting [23].

The amide and sulfonic acid groups in AMPS act as multidentate ligands, capable of coordinating with metal ions such as Al³⁺ and Fe³⁺. In composite accelerator systems, such as when combined with aluminum sulfate, AMPS may suppress premature cation precipitation, thereby optimizing the efficiency of setting control. Additionally, the amide group contributes to stabilizing mortar rheology and enhances the monomer’s penetration into cement [25]. As a result, AMPS helps optimize cement particle packing and moisture distribution, reduces the proportion of large pores formed during hydration, and accelerates early strength development [26].

Sodium 2-methylprop-2-ene-1-sulfonate, also known as Sodium Methallyl Sulfonate (SMAS), is an olefinic monomer containing a sulfonate group (–SO₃⁻Na⁺). In cement chemistry, it serves primarily as a functional monomer for synthesizing high-performance polycarboxylate ether (PCE)-based superplasticizers or other copolymer additives [27]. The sulfonate group (–SO₃⁻) of SMAS remains highly ionized in the alkaline environment of cement paste, effectively dispersing cement particles and preventing flocculation through strong electrostatic repulsion. Its allylic structure enhances the chemical stability of the polymer chain during polymerization, particularly under high-temperature or highly alkaline conditions [27]. The sodium sulfonate group also contributes to salt tolerance [28]. Furthermore, the branched chain on the double bond of SMAS is less bulky than that of AMPS, offering greater flexibility in tailoring copolymer flexibility and spatial conformation. Therefore, copolymerizing SMAS with functionalized components such as AMPS is expected to yield copolymers that integrate the advantages of each constituent while mitigating their individual drawbacks, thereby producing copolymers with effective acceleration and strengthening properties.

On the other hand, glass fiber (GF) was initially used as a reinforcement material for polymers, where it demonstrates outstanding performance in enhancing strength, toughness, and corrosion resistance due to its “bridging effect” and “crack-arresting effect” [29]. It has been widely applied in transportation, aerospace, marine engineering, and wind power generation. In recent years, the use of GF in cement-based materials has gained increasing attention. Firstly, incorporating glass fiber improves the mechanical strength and durability of cement [30]. Secondly, the cylindrical morphology and low surface roughness of GF contribute to enhancing the workability of cement mixtures [29]. Thirdly, glass fibers effectively fill microcracks formed during cement hydration, significantly increasing the compressive, flexural, and split tensile strengths [30]. However, achieving uniform dispersion of glass fibers in cement hydration systems typically requires a higher water-to-cement ratio [29]. This issue can be mitigated by combining GF with water-soluble polymers.

Based on the foregoing analysis, and to meet the requirements for shotcrete in specialized applications such as tunnels and bridges—where both rapid setting and high mechanical strength are essential—this work develops a novel accelerator system through the following procedure: a series of water-soluble coordination copolymers are synthesized via free-radical copolymerization in an aqueous solution, using persulfate to initiate the reaction between SMAS, AMPS (or absent), and SA. This copolymer is then complexed with aluminum sulfate, and GFs of different mesh sizes are incorporated.

This study aims to synthesize a copolymer bearing multiple, diverse coordinating groups through the copolymerization of sodium alginate with SMAS and/or AMPS, followed by its complexation with Al³⁺. This coordination model differs from the single-mode coordination exhibited by diethanolamine or triethanolamine with Al³⁺, which is prone to excessive sensitivity, leading to massive AFt formation that coats C₃S surfaces and impedes its hydration [10,12,13,14]. In contrast, the linear distribution of multiple and varied coordinating groups in the synthesized copolymer is expected to coordinate not only with Al³⁺ but also with Ca²⁺, thereby promoting the formation of both AFt and Ca(OH)₂ and consequently accelerating the cement hydration process from an alternative pathway.

On the other hand, cement hydration inherently generates numerous microcracks, significantly compromising mechanical strength. Building upon the ball-milled mixture of the copolymer and aluminum sulfate, this research further introduces glass fibers of varying sizes via ball milling to formulate a novel accelerator. Given the progressively alkaline environment of cement hydration, the incorporated glass fibers (SiO₂) are anticipated to mitigate crack propagation through covalent bonding and physical filling, thereby enhancing the mechanical strength of the cementitious matrix.

In summary, within the accelerator synthesized in this work, the coordination between the copolymer and metal ions (Al³⁺ and Ca²⁺) is designed to regulate the cement hydration rate, with aluminum sulfate serving as the source of Al³⁺ and glass fibers functioning to improve mechanical strength. The synergistic interaction among these three components is expected to yield a significant enhancement in cement hydration performance. In addition, detailed characterization of cement paste and mortar will be conducted to elucidate the mechanisms underlying the setting behavior and strength development. This project holds considerable scientific and practical value for advancing high-performance accelerators suited for specialized engineering applications.

2. Materials and Methods

2.1. Materials

Sodium 2-methylprop-2-ene-1-sulfonate (Sodium Methallyl Sulfonate, SMAS, 98%), 2-acrylamido-2-methylpropane sulfonic acid (AMPS, 98%), ammonium persulfate ((NH₄)₂S₂O₈, 98%), aluminum sulfate octadecahydrate (Al₂(SO₄)₃·18H₂O, AS, 99%), and glass fiber powders (GF, 200 and 2000 mesh, both have molecular formulas SiO₂, molecular weights 60.08, purities ≥ 98%) are all bought from Shanghai Aladdin Biochemical Technology Co., Ltd. (Shanghai, China). Sodium alginate (SA; specification: QB/F(W)0051-2018; CAS No.: 9005-38-3; molecular formula: (C₆H₇NaO₆)_n; relative molecular mass: (198.11)_n; viscosity range: 1.05–1.15 Pa s; pH: 7–8; water-insoluble matter: 0.1%) is sourced from Fuchen (Tianjin) Chemical Reagent Co., Ltd. (Tianjin, China).

Ordinary Portland Cement (OPC, grade P·O 42.5) is sourced from China National Building Materials Academy Co., Ltd. (Beijing, China). The Chinese ISO standard sand (GB/T 17671-2021 [31]) is provided by Xiamen ISO Standard Sand Co., Ltd. (Xiamen, China). Laboratory-prepared distilled water is used.

2.2. Instruments

Ball milling is performed using a YJKS high-energy grinder (Foshan Tenghao Instrument Technology Co., Ltd., Foshan, China) equipped with dual jars and operated at 220 V and 370 W, employing 18 mm zirconia (ZrO₂) milling beads. ¹H NMR spectra are recorded on a Bruker ADVANCE III 400 MHz spectrometer (Bruker Corporation, Billerica, MA, USA) using DMSO-d₆ as the solvent.

Gel Permeation Chromatography (GPC) analysis is carried out on a DAWN HELEOS II system (Wyatt Technology, Santa Barbara, CA, USA) using DMF as the mobile phase at a flow rate of 1.000 mL min⁻¹, equipped with a HELEOS light scattering detector and a 659.0 nm laser.

X-ray photoelectron spectroscopy (XPS) is conducted on a Kratos Axis Ultra DLD spectrometer (Kratos Co., Manchester, UK) with monochromatic Al Kα radiation (1486.6 eV). The C 1s peak at 284.8 eV is used as the reference for binding energy calibration. Spectral deconvolution is performed using a mixed Gaussian–Lorentzian function with 30% Lorentzian character.

Wide-angle X-ray diffraction (XRD) patterns are acquired on a Philips X’Pert Pro diffractometer (PANalytical, Almelo, The Netherlands) with Cu Kα radiation (λ = 1.5418 Å), scanning over a 2θ range from 5° (or 10°) to 80° at a rate of 0.05° s⁻¹.

Optical microscopy is performed using an SZ760 zoom stereo microscope (Chongqing Aote Optical Instrument Co., Ltd., Chongqing, China). Scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDS) are performed on a GeminiSEM 500 instrument (Carl Zeiss, Shanghai, China) at an accelerating voltage of 15 kV, offering resolutions of 0.6 nm (at 15 kV) and 0.9 nm (at 1 kV), with a magnification range from ×20 to ×2,000,000.

FT-IR spectra are collected on a Bruker VERTEX 70 spectrometer (Bruker Corporation, Billerica, MA, USA) in ATR mode over the range of 400–4000 cm⁻¹. UV-Vis spectra are obtained using a PerkinElmer Lambda 950 spectrophotometer (PerkinElmer, Waltham, MA, USA), scanning from 250 to 800 nm. Samples are dispersed in distilled water at a concentration of 2.5 mg mL⁻¹, using distilled water as the blank reference.

Thermogravimetric analysis (TGA/DTG) is conducted on a METTLER TOLEDO instrument (Mettler-Toledo International Inc., Greifensee, Zurich, Switzerland) under an air atmosphere, from 35 to 800 °C, with a temperature accuracy of ±0.3 °C and a heating rate of 10 °C min⁻¹. Differential scanning calorimetry (DSC) is performed using a METTLER TOLEDO DSC 3 instrument (Mettler-Toledo International Inc., Greifensee, Zurich, Switzerland), scanning from 30 to 300 °C at a heating rate of 10 °C min⁻¹ with a temperature precision of ±0.1 °C.

Inductively coupled plasma optical emission spectrometry (ICP-OES) is carried out on an Agilent 5110 spectrometer (Agilent Technologies, Santa Clara, CA, USA) under the following operating parameters: pump rate of 60 rpm, plasma gas flow of 12.0 L min⁻¹, nebulizer flow of 0.70 L min⁻¹, auxiliary gas flow of 1.0 L min⁻¹, RF power of 1250 W, and a stabilization time of 20 s. Particle size and zeta potential measurements are performed using a Zetasizer Nano ZSE (Malvern Panalytical, Malvern, Worcestershire, UK).

The setting times of cement pastes, mixed using an NJ-160A mixer (Wuxi Xiyi Building Material Instrument Factory, Wuxi, China), are determined with a Vicat apparatus (Wuxi Xiyi Building Material Instrument Factory, Wuxi, China). Mortars are prepared in a JJ-5 mixer (Wuxi Xiyi Building Material Instrument Factory, Wuxi, China). The compressive and flexural strengths of the mortar specimens are evaluated on a WAY-300B fully automatic testing machine (300 kN capacity, Wuxi Xiyi Building Material Instrument Factory, Wuxi, China) with an EHC-2300 control system (Wuxi Xiyi Building Material Instrument Factory, Wuxi, China), applying a loading rate of 48 N s⁻¹. All specimens are cured in an HBY-40B standard conservation box (Wuxi Xiyi Building Material Instrument Factory, Wuxi, China) at 20 °C and 90% relative humidity.

2.3. Synthesis of Admixtures

The flow chart of admixture preparation is presented in Figure 1.

2.3.1. Synthesis of T2

T2 is synthesized in a 1 L three-neck flask equipped with a reflux condenser, a constant-pressure dropping funnel, a magnetic stirrer, and an oil bath. SA (10.0 g), SMAS (10.0 g, 0.063 mol), n-butanol (100 mL), and distilled water (250 mL) are charged into the flask. The mixture is then heated to 80 °C with magnetic stirring, followed by the slow dropwise addition of an aqueous (NH₄)₂S₂O₈ solution (3.0 g, 0.013 mol of (NH₄)₂S₂O₈ dissolved in 50 mL of distilled water) over a period of 2 h. Special attention is paid to maintaining temperature control and reaction stability during the addition. After complete addition, the reaction mixture is stirred under reflux at 80 °C for an additional 2 h and then cooled to room temperature. The resulting mixture is concentrated to dryness using a rotary evaporator (RE-52AA, Shanghai Yarong Biochemical Instrument Factory, Shanghai, China) to obtain T2.

2.3.2. Synthesis of T2a

T2 (50 g), AS (50 g, 0.075 mol), and distilled water (50 g) are charged into a ball mill jar along with zirconia (ZrO₂) milling beads. The mixture is ball-milled at 25 °C for 1 h. The resulting product is then dried using a rotary evaporator to obtain T2a.

2.3.3. Synthesis of T2a-200

T2a (50 g), GF (50 g, 200 mesh), and distilled water (50 g) are charged into a ball mill jar along with zirconia (ZrO₂) milling beads. The mixture is ball-milled at 25 °C for 1 h. The resulting product is then dried using a rotary evaporator to obtain T2a-200.

2.3.4. Synthesis of T2a-2000

T2a-2000 is synthesized using a procedure identical to that of T2a-200, except that the 50 g of 200-mesh GF is substituted with an equal mass of 2000-mesh GF.

2.3.5. Synthesis of T3

T3 is synthesized using a procedure identical to that for T2, with the sole modification being the addition of 10 g (0.048 mol) of AMPS to the three-neck flask at the beginning of the reaction.

2.3.6. Synthesis of T3a

T3a is synthesized using a procedure identical to that for T2a, with the sole modification that T2 (50 g) is replaced with an equal mass of T3.

2.3.7. Synthesis of T3a-200

T3a-200 is synthesized using a procedure identical to that for T2a-200, with the sole modification being the replacement of T2a (50 g) with an equal mass of T3a.

2.3.8. Synthesis of T3a-2000

T3a-2000 is synthesized using a procedure identical to that for T2a-2000, with the sole modification being the replacement of T2a (50 g) with an equal mass of T3a.

2.4. Measurement of Setting Times of Cement Pastes

The initial and final setting times (IST and FST) of the cement pastes are determined using a Vicat apparatus on pastes prepared with an NJ-160A mixer, following the procedure specified in Chinese Standard GB/T 35159-2017 [32]. A cement paste is prepared by mixing 400 g of cement with 148 g of distilled water, corresponding to an admixture dosage of 6 wt.% relative to cement. The cement and water are first blended at low speed for 30 s. Then, 24 g of the admixture (6 wt.%) is added, followed by low-speed mixing for 5 s and high-speed stirring for 15 s. For the blank reaction without admixtures, 24 g of distilled water is added to the reaction system in place of any admixture. The resulting paste is immediately placed into a round mold, lightly vibrated for compaction, and smoothed with a scraper. The water-to-cement ratio for the cement paste series is as follows: 0.43 for P-non (entry 1,

Table 1. Setting times of cement pastes and mechanical strengths of mortars using different admixtures.

Entry	Paste (P-Admixture- GF Mesh)	Setting Time (ST, min, Paste) a		Mortar (M-Admixture-GF Mesh)	Compressive Strength (MPa, Mortar) b at: 24 h (28 d, R28 c)	Flexural Strength (MPa, Mortar) b at: 24 h (28 d)
		Initial (IST)	FST (FST)
1	P-non	33.59 ± 0.30	45.26 ± 0.10	M-non	3.5 ± 0.25 (20.8 ± 0.85)	1.0 ± 0.15 (2.9 ± 0.35)
2	P-AS	29.51 ± 0.67	35.12 ± 0.48	M-AS	4.3 ± 0.18	1.1 ± 0.15
3	P-T2a	4.47 ± 0.67	8.46 ± 0.21	M-T2a	9.2 ± 1.87	3.5 ± 0.25
4	P-T2a-200	3.56 ± 0.10	6.77 ± 0.02	M-T2a-200	10.2 ± 1.81	3.5 ± 0.11
5	P-T2a-2000	3.68 ± 0.08	6.15 ± 0.03	M-T2a-2000	8.0 ± 0.79	3.3 ± 0.15
6	P-T3a	2.95 ± 0.02	6.30 ± 0.04	M-T3a	8.9 ± 1.28	3.9 ± 0.1
7	P-T3a-200	2.24 ± 0.04	5.73 ± 0.03	M-T3a-200	9.4 ± 1.07	3.6 ± 0.05
8	P-T3a-2000	2.58 ± 0.02	5.55 ± 0.02	M-T3a-2000	7.2 ± 0.79 (32.3 ± 1.27, 155%)	3.3 ± 0.05 (11.0 ± 0.55)

^a Experimental details as in Section 2.4; for each paste, the initial and final setting times were measured three times, with the average value ± SD (standard deviation) reported. The p-values are obtained from Student’s t-test for intergroup comparisons. ^b Experimental details as in Section 2.5; for each mortar, compressive strength was measured six times, and flexural strength three times (three-point loading method), data format: average value ± SD (standard deviation); the p-values are obtained from Student’s t-test for intergroup comparisons. Examples of mechanical strength reports of mortars at 24 h as in Figures S16–S19, Supplementary Materials. ^c Retention ratio after 28 days, R₂₈ = f_{t, 28}/f_{r, 28} × 100%, f_{t, 28}: comprehensive strength of tested mortar at 28 d (MPa, admixture-introduced), f_{r, 28}: comprehensive strength of standard mortar (MPa, M-non, entry 1, Table 1, admixture-blank) at 28 d, according to Chinese standard GB/T 35159–2017 [32].

), and 0.37 for P-AS, P-T2a, P-T2a-200, P-T2a-2000, P-T3a, P-T3a-200, and P-T3a-2000 (entries 2–8, Table 1). During the preparation of the cement paste, as the admixture actively participates in the hydration process to an extent comparable to water, the water-to-cement ratio is adjusted to maintain a constant mass percentage of cement within the overall paste mixture.

The IST and FST are measured at 10 s intervals using a Vicat apparatus, which employs a needle of fixed cross-section under a constant load. The IST is defined as the time from the release of the needle in free fall until it penetrates to a depth of 4 ± 1 mm above the base. The FST is recorded as the period from the end of the IST until the needle no longer leaves a visible impression on the paste surface.

2.5. Measurement of Mechanical Strengths of Cement Mortars

Mortar specimens are prepared using a JJ-5 mortar mixer (Wuxi Xiyi Building Material Instrument Factory, Wuxi, China). Their compressive and flexural strengths are determined on a WAY-300B testing machine (Wuxi Xiyi Building Material Instrument Factory, Wuxi, China) in accordance with Chinese Standard GB/T 35159-2017 [32]. The mortar is prepared by mixing 900 g of cement with 468 g of distilled water (corresponding to a 6 wt.% admixture dosage) in a JJ-5 cement mortar mixer. The mixture is stirred at low speed for 30 s, followed by an additional 30 s of low-speed stirring while 1350 g of Chinese ISO standard sand is gradually added. It is then mixed at high speed for 30 s, rested for 90 s, and mixed again at high speed for 30 s. Subsequently, 54 g of admixture (6 wt.%) is incorporated, followed by low-speed mixing for 5 s and high-speed mixing for 15 s. For the blank reaction without admixtures, 54 g of distilled water is added to the reaction system in place of any admixture. The water-to-cement ratio for the cement paste series is as follows: 0.58 for M-non (entry 1, Table 1), and 0.52 for M-AS, M-T2a, M-T2a-200, M-T2a-2000, M-T3a, M-T3a-200, and M-T3a-2000 (entries 2–8, Table 1). During the preparation of the cement mortar, as the admixture actively participates in the hydration process to an extent comparable to water, the water-to-cement ratio is adjusted to maintain a constant mass percentage of cement within the overall mortar mixture. The fresh mortar is immediately cast into a 40 mm × 40 mm × 160 mm mold and cured in a standard conservation box (Wuxi Xiyi Building Material Instrument Factory, Wuxi, China) at 20 °C and 90% relative humidity for designated periods of 24 h and 28 days.

3. Results and Discussion

3.1. Characterizations of Admixtures

3.1.1. Structures of Copolymers T2 and T3

The synthetic routes for T2 and T3 are illustrated in Figure 1. XPS survey scans (Figure 2) and the corresponding surface elemental composition with binding energies (Table S1, Section S1, Supplementary Materials) confirm the presence of chemical elements from the starting materials in both T2 and T3 (Figure 2a,e). Compared to T2, the molar percentages of carbon and nitrogen in T3 increase, while those of sulfur and sodium decrease (entries 5 vs. 1, Table S1). This is attributed to the fact that, relative to the SMAS monomer, AMPS exhibits a higher molar percentage of carbon and contains nitrogen, whereas SMAS does not (T3 vs. T2, Figure 1f). These results indicate that the AMPS monomer has been successfully copolymerized into T3 (Figure 1f). Trace amounts of Si detected in T2 and T3 are likely attributable to by-products from the sodium alginate extraction process (Figure 2a,e; entries 1 and 5, Table S1).

In the high-resolution C 1s spectrum of T2, three peaks are observed at 284.6 eV, 286.0 eV, and 287.6 eV (Figure S1a), corresponding to saturated carbon (sp³ hybridized, C–H or C–C), carbon in C–O bonds, and carboxyl carbon, respectively [33]. A similar profile is observed in the C 1s spectrum of T3 (Figure S1e vs. Figure S1a). The N 1s spectrum of T2 displays two peaks at 399.4 eV and 401.7 eV (Figure S2a), assigned to NH₄⁺ from (NH₄)₂S₂O₈ (Figure 1f) and organic nitrogen impurities in SA, respectively [34]. In T3, the N 1s peaks are shifted relative to those in T2 (Figure S2e vs. Figure S2a), with the peak at 401.6 eV indicating the influence of the amide nitrogen from the AMPS monomer (Figure 1f).

The S 2p spectra of T2 and T3 exhibit doublet peaks at 168.5/167.6 eV and 168.8/167.8 eV, respectively (Figure S3a and Figure S3e), where the higher binding energy component corresponds to S 2p_1/2 and the lower to S 2p_3/2 [35], both originating from –SO₃Na and –SO₃H groups (Figure 1f). The Na 1s peaks for T2 and T3 appear at similar positions (Figure S4a,e), suggesting that Na⁺ exists primarily as free cations in both materials (Figure 1f) [36].

In the O 1s spectrum of T2, peaks are observed at 531.6 eV, 532.7 eV, and 536.0 eV (Figure S5a), assigned to oxygen in C–O, C=O, and hydroxyl groups, respectively [37]. The O 1s profile of T3 is similar to that of T2 (Figure S5e). Collectively, these results indicate that the key functional groups from the precursor monomers are retained in the copolymers T2 and T3 (Figure 1f).

Furthermore, wide-angle XRD analysis of T2 reveals the presence of two distinct crystalline phases with sharp diffraction peaks. The first phase is identified as mirabilite (PDF No. 11-0674, Na₂SO₄·10H₂O; Figure 3a), while the second corresponds to D-glucitol (PDF No. 19-1504, C₁₇H₂₇NO₉; Figure 3a). The SO₄²⁻ ions in the first phase originate from the use of (NH₄)₂S₂O₈ (Figure 1a–c). The persulfate anion (S₂O₈²⁻) undergoes hydrolysis to form SO₅²⁻ and SO₄²⁻, with SO₅²⁻ further hydrolyzing to yield HO₂⁻. HO₂⁻ then reacts with S₂O₈²⁻ to generate SO₄⁻· radicals, which initiate the subsequent aqueous free-radical copolymerization (Figure 1d–f) [38,39]. The second phase likely results from the strong oxidative effect of (NH₄)₂S₂O₈ on the pyranose rings in sodium alginate.

In addition, T2 exhibits a broad diffraction hump in the 2θ range of 10°–65° (Figure 3a), consistent with previously reported XRD patterns of sodium alginate [40]. In contrast, the XRD pattern of T3 is dominated by the characteristic features of sodium alginate [40], though a thenardite phase (PDF No. 75-0914, Na₂SO₄) is also detected (Figure 3e). These results confirm that the fundamental backbone structure of sodium alginate is retained in both T2 and T3 (Figure 1f).

Gel permeation chromatography (GPC) reveals that T2 exhibits a high number-average molecular weight (Mₙ = 1.591 × 10⁸ g mol⁻¹) with a narrow molecular weight distribution (M_w/Mₙ = 1.238) (Figure S6 and Table S2). In comparison, T3 shows a two-orders-of-magnitude lower M_n (6.529 × 10⁶ g mol⁻¹) and an even narrower distribution (M_w/Mₙ = 1.103) (Figure S7 and Table S2). This trend suggests that the shorter side chain on the carbon–carbon double bond in SMAS imposes less steric hindrance during polymerization than the bulkier substituent in AMPS (Figure 1), thereby favoring the formation of longer polymer chains. Conversely, the more sterically demanding side chain in AMPS promotes greater regularity in chain growth. As a result, the incorporation of AMPS into the copolymerization system with SMAS and SA yields a copolymer with more uniform chain length and a narrower molecular weight distribution (Figure 1f) [41,42].

Meanwhile, the ¹H NMR spectrum of T2 indicates that the molar ratio of the SMAS segment to the SA segment is approximately 1:1.5, based on the ratio of the integral area of the methyl hydrogen in the SMAS fragment (δ = 1.24 ppm, Figure 4a–c) to that of the hydrogen atoms in the SA fragment (δ = 4.82–4.73 ppm, H_G-1 and H_GM-5, Figure 4a–c) [20,40]. In addition, the ¹H NMR spectrum of T3 reveals the presence of groups such as the methyl group in the AMPS fragment (δ = 1.33 ppm, Figure 4d–f).

SEM analysis reveals that T2 is composed of micron-scale lamellar structures (Figure 5f–h), whereas T3 consists of large, smooth-surfaced membranous substances (Figure 5o,p). EDS mapping confirms the uniform distribution of C, N, O, S, and Na on the surfaces of both T2 and T3, albeit with differing relative mass percentages (Figure 6 and Figure 7). Notably, the sodium content on the T3 surface is significantly lower than that on T2 (Figure 7b vs. Figure 6b). These results demonstrate the successful formation of copolymers T2 and T3, and indicate again that the incorporation of AMPS reduces the sodium content on the copolymer surface (Figure 1e; entries 5 vs. 1, Table S1).

3.1.2. Structures of Admixtures T2a and T3a

Upon loading AS onto T2, the resulting product T2a exhibits a significant decrease in carbon and sodium content, a substantial increase in oxygen content, and the emergence of aluminum (entries 2 vs. 1, Table S1). In the C 1s spectrum of T2a, the carboxyl carbon peak appears at 288.5 eV, which is significantly higher than that in T2 (287.6 eV, Figure S1b vs. Figure S1a) [33]. This indicates that in T2a, Al³⁺ coordinates with carboxyl groups, causing the electron cloud of the C–O bond in the carboxyl group to shift toward the oxygen atom. This reduces the repulsion between the outer electrons of the carbon atom, making it more difficult for C 1s photoelectrons to escape (Figure 1f) [43].

In the Na 1s spectrum of T2, a single peak at a binding energy of 1071.3 eV (Figure S4a) corresponds to Na⁺ in the –SO₃⁻Na⁺ groups of the SMAS segment [36]. In contrast, the Na 1s spectrum of T2a exhibits two peaks at binding energies of 1072.1 eV and 1070.7 eV (Figure S4b). This suggests that after mixing T2 with aluminum sulfate, some –SO₃⁻Na⁺ groups in the SMAS segment lose Na⁺ and coordinate with Al³⁺, while the released Na⁺ combines with anions such as SO₄²⁻ in the solution to form salts (corresponding to the peak at 1072.1 eV in T2a, Figure S4b). Meanwhile, other –SO₃⁻Na⁺ groups in the SMAS segment retain Na⁺ but still coordinate with Al³⁺ (corresponding to the peak at 1070.7 eV in T2a, Figure S4b) [44].

Additionally, the S 2p spectrum of T2 shows two peaks at binding energies of 168.5 eV and 167.6 eV (Figure S3a), mainly reflecting the influence of sulfur derived from the –SO₃⁻Na⁺ groups in the SMAS segments (Figure 1f) and from the by-product inorganic salt Na₂SO₄·10H₂O (Figure 3a) [35]. In contrast, the S 2p spectrum of T2a exhibits two peaks at binding energies of 170.4 eV and 169.1 eV (Figure S3b), reflecting the influence of aluminum sulfate incorporation (Figure 1f).

Concurrently, the Al 2p spectrum of T2a shows two peaks at binding energies of 75.2 eV and 75.8 eV (Figure S8a). The former is characteristic of Al³⁺ in an environment of lower electronegativity, likely coordinated Al³⁺ or aluminum sulfate, while the latter corresponds to Al³⁺ in a higher electronegativity environment, possibly aluminum oxide or hydroxide [45]. XRD analysis further confirms that T2a contains not only Al₂(SO₄)₃·14H₂O (PDF No. 49-1097, Figure 3b) but also boehmite (PDF No. 83-2384, AlO(OH), Figure 3b), which aligns well with the XPS findings.

In FT-IR spectrum, T2 exhibits three peaks at wavenumbers of 3670 cm⁻¹, 2980 cm⁻¹, and 2923 cm⁻¹ (Figure 8(Aa)), which are attributed to the O–H stretching vibration of hydroxyl groups, the asymmetric stretching vibration of C–H bonds in methyl groups, and the asymmetric stretching vibration of C–H bonds in methylene groups, respectively [19]. The peak at 1748 cm⁻¹ suggests the possible formation of an ester bond between the carboxyl and hydroxyl groups in sodium alginate. The two peaks at 1646 cm⁻¹ and 1400 cm⁻¹ correspond to the asymmetric and symmetric stretching vibrations of the –COO⁻ group in sodium alginate [40]. The peak at 1218 cm⁻¹ represents the asymmetric stretching vibration of the S=O bond in the sulfonate group, while the peak at 1053 cm⁻¹ is assigned to the symmetric stretching vibration of the S=O bond [40]. The peak at 887 cm⁻¹ can be ascribed to the asymmetric stretching vibration of the G-unit in sodium alginate (Figure 1f) [40].

After the loading of AS onto T2 (Figure 1f), the resulting product T2a exhibits an infrared spectrum largely consistent with that of T2. However, the vibration peak originally at 1053 cm⁻¹ in T2 shifts to a higher wavenumber in T2a (Figure 8(Ab) vs. Figure 8(Aa)), which is a result of the incorporation of SO₄²⁻ in T2a [46].

Microscopy observations reveal a higher degree of aggregation in T2a compared to T2 (Figure S9d vs. Figure S9c). Further SEM analysis shows that T2a exhibits larger lamellar structures than those of T2 (Figure 5i,j vs. Figure 5f–h). Additionally, EDS mapping indicates a sparse distribution of sodium but a high density of aluminum on the surface of T2a (Figure S10g,h). Consequently, the sodium-containing phase in T2a (PDF No. 71-2270, NaHSO₄·H₂O, Figure 3b) is likely located in the inner layer, whereas the aluminum-containing phases (PDF No. 49-1097, Al₂(SO₄)₃·14H₂O; PDF No. 83-2384, AlO(OH); Figure 3b) are predominantly situated on the sample surface.

During the formation of T3a from T3, the observed changes in T3a—compared to T3—closely mirror those observed during the conversion of T2 to T2a. This similar trend is evident across multiple analytical dimensions: chemical elemental composition (entries 6 vs. 5, Table S1), chemical states of oxygen (Figure S5f vs. Figure S5e), sodium (Figure S4f vs. Figure S4e), aluminum (Figure S8d) and sulfur (Figure S3f vs. Figure S3e), phase composition (Figure 3f vs. Figure 3e), microscopy (Figure S9h,i vs. Figure S9g), SEM morphology (Figure 5q vs. Figure 5o,p), EDS analysis (Figure S11), and infrared spectroscopy (Figure 8(Af) vs. Figure 8(Ae)).

3.1.3. Structures of Admixtures T2a-200 and T2a-2000

When glass fibers are further loaded into the synthesis of admixtures, according to SEM observations, the 200-mesh glass fibers (GF-200) used in this work exhibit a length of approximately 80–100 μm and a cross-sectional diameter of about 8–10 μm, with uniform size distribution and smooth surfaces (Figure 5a–c). In contrast, the 2000-mesh glass fibers (GF-2000) have a length of about 10–30 μm and a similar cross-sectional diameter of 8–10 μm but display non-uniform size distribution and rough surfaces (Figure 5d,e).

The molar percentages of Si, Al, and C on the surface of T2a-200 increase significantly compared to those of T2a (entries 3 vs. 2, Table S1). Compared to the C 1s spectrum of T2a, the C 1s spectrum of T2a-200 shows the carboxyl carbon peak splitting into a doublet at binding energies of 288.4 eV and 289.1 eV (Figure S1c vs. Figure S1b) [33]. The former still corresponds to the carboxyl carbon of the alginate segment in T2a-200, while the latter may indicate that, under ball-milling conditions, some carboxyl groups of the alginate segment in T2a-200 undergo dehydration with surface hydroxyl groups on GF-200, forming ester linkages (Figure 1f) [47].

The S 2p spectrum of T2a-200 closely resembles that of T2a (Figure S3c vs. Figure S3b), indicating no significant change in the chemical environment of sulfate groups before and after loading onto GF-200 (Figure 1f). In contrast, the Na 1s spectrum of T2a-200 exhibits an additional peak at a binding energy of 1073.7 eV compared to that of T2a (Figure S4c vs. Figure S4b). This likely suggests that NaHSO₄·H₂O present in T2a (Figure 3b) undergoes an ion-exchange reaction with surface hydroxyl groups of GF-200 (SiO₂) during ball-milling, where Na⁺ replaces H⁺.

Concurrently, in the Al 2p spectrum, the peak position of T2a-200 shifts relative to that of T2a (Figure S8b vs. Figure S8a). In the Si 2p spectrum, T2a-200 exhibits three peaks at 101.2 eV, 102.2 eV, and 103.1 eV (Figure S12a), corresponding to interfacial silicon species, silicate, and SiO₂, respectively [48]. In the O 1s spectrum, the doublet peaks of T2a-200 shift toward lower binding energies compared to those of T2a (Figure S5c vs. Figure S5b), further reflecting the influence of the incorporated glass fiber components.

Furthermore, XRD analysis reveals that the aluminum-containing phases in T2a, when mixed with GF-200 and subjected to ball milling, transform into Al₂Si₂O₅(OH)₄ (PDF No. 14-0164, Figure 3c), and Al₄SO₄(OH)₁₀·36H₂O (PDF No. 08-0076, Figure 3c), along with residual SiO₂ (PDF No. 42-1401, Figure 3c).

From another perspective, SEM analysis reveals significant morphological differences between T2a-200 and T2a. First, the fibers of GF-200 are coated with granular substances (Figure 5k). Second, the surface of T2a-200 exhibits abundant lamellar wrinkles, contrasting sharply with the large sheet-like structures of T2a (Figure 5l vs. Figure 5i,j). EDS spectra demonstrate high concentrations of aluminum, oxygen, and sulfur, along with low concentrations of silicon, carbon, and sodium on the surface of T2a-200 (Figure S13), suggesting that aluminum-containing species encapsulate the copolymer and silica components.

The formation of T2a-2000 from T2a follows a process similar to that of T2a-200. First, both T2a-2000 and T2a-200 contain the same crystalline phases: Kaolinite-1A (PDF No. 14-0164, Al₂Si₂O₅(OH)₄) and SiO₂ (PDF No. 42-1401) (Figure 3d vs. Figure 3c). Second, they exhibit comparable trends in morphological evolution (Figure 5m,n vs. Figure 5k,l vs. Figure 5i,j). Third, T2a-2000 also shows evidence of aluminum-containing species encapsulating the copolymer and silica components (Figure S14).

Simultaneously, compared to T2a, the surface silicon and aluminum content of T2a-2000 increases, while the sulfur, nitrogen, and oxygen content decreases (entries 4 vs. 2, Table S1), reflecting the influence of GF-2000 (SiO₂) loading (Figure 1f). In the C 1s spectra, the carboxyl carbon peak in T2a-2000 shifts to a higher binding energy relative to that in T2a (288.8 eV vs. 288.5 eV, Figure S1d vs. Figure S1b). This also suggests that carboxyl groups in the alginate segment may undergo esterification grafting with surface hydroxyl groups on GF-2000 under ball-milling conditions [47].

The S 2p spectra of T2a-2000 and T2a are very similar (Figure S3d vs. Figure S3b), indicating nearly identical chemical environments for the sulfur-containing species in both materials. Upon comparing the O 1s spectra, the peaks at 533.0 eV and 532.1 eV in T2a shift to 533.4 eV and 532.0 eV, respectively, in T2a-2000 (Figure S5d vs. Figure S5b). This suggests that the O 1s binding energy of oxygen in Si–O bonds (from GF-2000) is lower than that in C–O bonds. Moreover, the increased binding energy for oxygen in C=O bonds may result from the dehydration reaction between carboxyl groups in alginate and surface hydroxyl groups on GF-2000 during ball-milling, forming ester linkages [47].

Additionally, the Al 2p spectrum of T2a-2000 closely resembles those of T2a and T2a-200 (Figure S8c vs. Figure S8d,a), confirming that the aluminum-containing phases—or at least the chemical environment of aluminum—are essentially consistent among T2a-2000, T2a-200, and T2a (Figure 3d vs. Figure 3c,b).

3.1.4. Structures of Admixtures T3a-200 and T3a-2000

The formation processes of T3a-200 and T3a-2000 from T3a closely resemble those of T2a-200 and T2a-2000 derived from T2a. Notably, both T3a-200 and T3a-2000 contain hydroxylated aluminum sulfate (PDF No. 25-1491, Al₄SO₄(OH)₁₀·5H₂O, Figure 3g,h). Simultaneously, after ball-milling T3a with GF-200, the resulting T3a-200 exhibits a morphology highly similar to that of T2a-200, with numerous petal-like microstructures observed in both (Figure 5r vs. Figure 5l). In contrast, T3a-2000 shows a distinctly different morphology (Figure 5t vs. Figure 5r). This difference may be attributed to the distinct particle sizes of GF-200 and GF-2000. In addition, the surface of T3a-200 also exhibits encapsulation of the copolymer and silica components by aluminum-containing species (Figure S15).

From the perspective of surface element distribution, the molar percentages of silicon on the surfaces of T3a-200 and T3a-2000 are higher than that on T3a, whereas the molar percentages of aluminum, sulfur, nitrogen, and sodium on the surfaces of T3a-200 and T3a-2000 are lower than that of silicon on T3a (entries 7 and 8 vs. 6, Table S1). This indicates that the primary component of glass fibers GF-200 and GF-2000 is SiO₂, and during the synthesis of T3a-200 and T3a-2000, T3a does not encapsulate GF-200 or GF-2000 (Figure 1f).

In the C 1s spectra of T3a-200 and T3a-2000, the binding energies of carboxyl carbon (288.4 eV, Figure S1g; 288.3 eV, Figure S1h) are higher than that of T3a (288.1 eV, Figure S1f). This suggests that after ball-milling with GF-200 and GF-2000, the carboxyl groups of the alginate segments in T3a likely undergo dehydration condensation with surface hydroxyl groups on the glass fibers to form ester bonds.

Compared with the N 1s spectrum of T3a, the N 1s spectra of T3a-200 and T3a-2000 shift only slightly toward lower binding energy (Figure S2g,h vs. Figure S2f), indicating that covalent bonding likely occurs between T3a and the glass fibers during synthesis (Figure 1f).

Similarly, the S 2p spectra of T3a-200 and T3a-2000 shift slightly toward lower binding energy relative to that of T3a (Figure S3g,h vs. Figure S3f), further supporting the likelihood of covalent linkage between T3a and the glass fibers (Figure 1f). In the O 1s spectrum of T3a, the peak at 533.5 eV represents oxygen in the C=O bonds of carboxyl groups from the alginate segment (Figure S5f) [37]. This peak shifts to lower binding energy in both T3a-200 and T3a-2000 (Figure S5g,h vs. Figure S5f), implying that hydrogen bonding may form between the C=O groups of alginate and surface hydroxyl groups on the glass fibers due to covalent linkage.

In the Al 2p spectrum of T3a, peaks at 75.1 eV and 76.2 eV (Figure S8d) are assigned to Al³⁺ in Al₂(SO₄)₃·14H₂O (PDF No. 22-0023) and Al₃(HSO₄)(SO₄)₄·9H₂O (PDF No. 86-1777), and to Al³⁺ in Al₄SO₄(OH)₁₀·36H₂O (PDF No. 08-0076), respectively (Figure 3f) [45]. In contrast, the Al 2p spectrum of T3a-200 shows two distinct photoelectron peaks: one at 74.9 eV (Figure S8e) corresponding to Al³⁺ in Al₄SO₄(OH)₁₀·5H₂O (PDF No. 25-1491) (Figure 3g), and the other at 75.2 eV (Figure S8e) corresponding to Al³⁺ in Al₂Si₂O₅(OH)₄ (PDF No. 14-0164) (Figure 3g) [45]. For T3a-2000, the Al 2p spectrum exhibits peaks at 75.2 eV (Figure S8f) for Al³⁺ in Al₄SO₄(OH)₁₀·5H₂O (PDF No. 25-1491) (Figure 3h), at 74.5 eV (Figure S8f) for Al³⁺ in Al₂O₃·2SiO₂·2H₂O (PDF No. 03-0052), and at 75.9 eV for other hydroxyl-rich aluminum-containing complex oxides [45].

In the Si 2p spectrum of T3a-200, peaks at 102.2 eV and 101.4 eV (Figure S12c) represent Si in SiO₂ (PDF No. 42-1401, Figure 3g) and Si in Al₂Si₂O₅(OH)₄ (PDF No. 14-0164, Figure 3g), respectively [48]. In T3a-2000, peaks at 103.1 eV (Figure S12d) and 101.7 eV (Figure S12d) correspond to Si in SiO₂ (PDF No. 42-1401, Figure 3h) and Si in Al₂O₃·2SiO₂·2H₂O (PDF No. 03-0052, Figure 3h), respectively [48]. These XPS data are in excellent agreement with the XRD results, further confirming the compositional profiles of T3a-200 and T3a-2000.

FT-IR spectra of T3a-200 and T3a are very similar, except that the peak at 3670 cm⁻¹ is more pronounced in T3a-200 than in T3a (Figure 8(Ag) vs. Figure 8(Af)), which arises from the surface hydroxyl groups of added GF-200 (Figure 1f) [49]. However, the FT-IR spectrum of T3a-2000 differs somewhat from that of T3a-200 (Figure 8(Bh) vs. Figure 8(Ag)). First, T3a-2000 exhibits peaks at 2974 cm⁻¹ and 2906 cm⁻¹, assigned to the asymmetric stretching vibration of C–H bonds in methyl groups and methylene groups, respectively [19]. These wavenumbers are lower than the corresponding peaks in T3a-200 (2980 cm⁻¹ and 2923 cm⁻¹). This suggests that due to the different sizes of GF-2000 and GF-200, the surface dislocations and defects of the two glass fibers differ, leading to variations in the folding of T3a polymer segments and their degree of connection and bending on the fiber surfaces after blending (Figure 5s,t vs. Figure 5r).

In summary, blending T3a with GF-200 and GF-2000 yields products with essentially identical elemental valence states and chemical compositions, but with minor differences in morphology and molecular-scale polymer conformation.

3.2. Effects of Admixtures

The effects of the admixture-free control and various admixtures are summarized in Table 1. In the absence of any admixture, the cement exhibits prolonged initial and final setting times, along with slow development of mechanical strength (entry 1, Table 1). When AS is used as the admixture, the initial setting time of the paste is significantly reduced compared to the admixture-free control (29.51 ± 0.67 vs. 33.59 ± 0.30 min, p = 0.00323 < 0.05; entries 2 vs. 1, Table 1). The 24 h compressive strength of the mortar also shows a marked increase (4.3 ± 0.18 vs. 3.5 ± 0.25 MPa, p = 1.3769 × 10⁻⁴ < 0.05; entries 2 vs. 1, Table 1). However, the use of AS alone fails to meet common industrial standards in terms of both setting time and mechanical strength (Chinese standard GB/T 35159–2017: IST ≤ 5 min, FST ≤ 12 min, 24 h compressive strength ≥ 7.0 MPa) [32].

When T2a is used as an admixture at a 6% dosage, it demonstrates significant advantages over AS, both in terms of the initial and final setting times of the paste and the 24 h flexural and compressive strengths of the mortar (entries 3 vs. 2, Table 1). These improvements can likely be attributed to the boehmite (PDF No. 83-2384, AlO(OH), Figure 3b) present in T2a, which may play a major role in accelerating setting and enhancing the mechanical strength of cement.

The incorporation of 200-mesh glass fibers (GF-200) into T2a yields the admixture T2a-200, which further significantly reduces both the initial and final setting times compared to T2a alone (3.56 ± 0.10 vs. 4.47 ± 0.67 min, p = 0.00524 < 0.05; 6.77 ± 0.02 vs. 8.46 ± 0.21 min, p = 0.00525 < 0.05; entries 4 vs. 3, Table 1). Meanwhile, the 24 h compressive strength of mortar with T2a-200 is higher than that with T2a, though the difference is not statistically significant (10.2 ± 1.81 vs. 9.2 ± 1.87 MPa, p = 0.418 > 0.05; entries 4 vs. 3, Table 1). The flexural strengths, however, are nearly identical for the two admixtures (entries 4 vs. 3, Table 1). These results suggest that the hydroxylated aluminum sulfate (PDF No. 08-0076, Al₄SO₄(OH)₁₀·36H₂O) and aluminosilicate (PDF No. 14-0164, Al₂Si₂O₅(OH)₄) phases present in T2a-200 (Figure 3c) may contribute more effectively to setting acceleration than the boehmite (PDF No. 83-2384, AlO(OH)) contained in T2a (Figure 3b).

When 2000-mesh glass fibers (GF-2000) replace 200-mesh fibers during ball milling with AS and T2, the resulting T2a-2000 shows no significant improvement over T2a-200 in terms of initial setting time (3.68 ± 0.08 vs. 3.56 ± 0.10 min, p = 0.19792 > 0.05; entries 5 vs. 4, Table 1), compressive strength (8.0 ± 0.79 vs. 10.2 ± 1.81 MPa, p = 0.02419 < 0.05; entries 5 vs. 4, Table 1), or flexural strength (3.3 ± 0.15 vs. 3.5 ± 0.11 MPa, p = 0.1447 > 0.05; entries 5 vs. 4, Table 1). However, it demonstrates a marked reduction in final setting time (6.15 ± 0.03 vs. 6.77 ± 0.02 min, p = 2.5079 × 10⁻⁵ < 0.05; entries 5 vs. 4, Table 1). XRD analysis reveals that T2a-2000 and T2a-200 share nearly identical phase composition. Nevertheless, the hydroxylated aluminum sulfate in T2a-2000 contains less crystalline water (PDF No. 08-0055, Al₂SO₄(OH)₄·7H₂O, Figure 3d), which may contribute to its enhanced setting acceleration activity (Figure 3d vs. Figure 3c).

Compared to T2a, T3a significantly reduces the initial and final setting times of cement paste (IST: 2.95 ± 0.02 vs. 4.47 ± 0.67 min, p = 1.79372 × 10⁻⁴ < 0.05; FST: 6.30 ± 0.04 vs. 8.46 ± 0.21 min, p = 7.60288 × 10⁻⁵ < 0.05; entries 6 vs. 3, Table 1) and increases the average flexural strength of mortar (3.9 ± 0.1 vs. 3.5 ± 0.25 MPa, p = 0.11 > 0.05, entries 6 vs. 3, Table 1), while showing no significant effect on compressive strength (8.9 ± 1.28 vs. 9.2 ± 1.87, p = 0.02419 > 0.05, entries 6 vs. 3, Table 1). XRD analysis indicates that the hydroxylated aluminum sulfate phase in T3a (PDF No. 08-0076, Al₄SO₄(OH)₁₀·36H₂O, Figure 3f) likely contributes to more effective setting acceleration than the boehmite phase present in T2a (PDF No. 83-2384, AlO(OH), Figure 3b).

Compared to T3a, the use of T3a-200 significantly shortens the initial and final setting times of cement paste (IST: 2.24 ± 0.04 vs. 2.95 ± 0.02 min, p = 6.03544 × 10⁻⁵ < 0.05; FST: 5.73 ± 0.03 vs. 6.30 ± 0.04 min, p = 8.89332 × 10⁻⁵ < 0.05; entries 7 vs. 6, Table 1) and enhances the 24 h average compressive strength of mortar (9.4 ± 1.07 vs. 8.9 ± 1.28 MPa, entries 7 vs. 6, Table 1). However, it leads to a reduction in flexural strength (3.6 ± 0.05 vs. 3.9 ± 0.1 MPa, p = 0.03565 < 0.05; entries 7 vs. 6, Table 1). XRD analysis indicates that the hydroxylated aluminum sulfate (PDF No. 25-1491, Al₄SO₄(OH)₁₀·5H₂O, Figure 3g) in T3a-200 contains less crystalline water than that (PDF No. 08-0076, Al₄SO₄(OH)₁₀·36H₂O, Figure 3f) in T3a, which may contribute to its higher setting acceleration activity.

Furthermore, in terms of both setting time and mechanical properties, T3a-2000 exhibits inferior performance compared to T3a-200 (entries 8 vs. 7, Table 1). This difference may be attributed to the lower setting and strengthening activity of the Al₂O₃·2SiO₂·2H₂O phase (PDF No. 03-0052, Figure 3h) in T3a-2000 relative to the aluminosilicate phase (Al₂Si₂O₅(OH)₄, PDF No. 14-0164, Figure 3g) present in T3a-200.

Based on the experimental results, it is evident that in terms of accelerating cement hydration and enhancing mechanical strength, the synergistic effect between copolymer T3 and aluminum sulfate (T3a) is superior to that between copolymer T2 and aluminum sulfate (T2a). This likely indicates that copolymer T3, synthesized with AMPS as one of its monomers (Figure 1f), possesses excellent permeability and binding affinity within the cement hydration system. Furthermore, the synergistic effect between 200-mesh glass fibers and T2a (similarly for T3a) outperforms that between 2000-mesh glass fibers and T2a (or T3a). This suggests that 200-mesh glass fibers, with their larger particle size compared to 2000-mesh fibers, may more effectively reinforce the cement matrix by bridging micro-cracks and mitigating crystalline defects formed during hydration (Figure 1f).

Finally, with T3a-2000 as an admixture, the 28-day compressive strength ratio of the resulting cement mortar reaches 155% (entries 8 vs. 1, Table 1), indicating that the T3a-2000 additive has no adverse effect on the long-term strength of cement paste. Moreover, when T2a-200 and T2a-2000 are used as admixtures (entries 4 and 7, Table 1), both the initial and final setting times of the cement paste and the 24 h compressive strength of the mortar significantly exceed the requirements of the Chinese National Standard GB/T 35159-2017 (initial setting time ≤ 5 min, final setting time ≤ 12 min, and 24 h mortar compressive strength ≥ 7 MPa) [32]. This suggests that this series of admixtures holds considerable potential for acceleration application.

3.3. Characterizations of Cement, Pastes and Mortars

3.3.1. Structures of Cement and Glass Fiber-Free Pastes and Mortars

Structures of Cement, M-non, M-AS, P-T2a and M-T2a

The surface elemental composition of the cement raw material used in this work is shown in Figure 2i and entry 9 of Table S1, while its chemical composition is provided in Table S3. The cement itself contains potassium, as evidenced by the K 2p spectrum, which exhibits a doublet at 295.6 eV and 292.8 eV, corresponding to the 2p_1/2 and 2p_3/2 photoelectrons of K⁺ (Figure S20a) [50]. Additionally, the silicon in the cement is represented by a single peak in the Si 2p spectrum at a binding energy of 101.80 eV (Figure S12e), likely attributable to silicon in tricalcium silicate (C₃S) [48]. The sulfur in the cement displays a doublet in the S 2p spectrum at binding energies of 169.8 eV and 168.7 eV (Figure S3i), corresponding to the 2p_1/2 and 2p_3/2 photoelectrons of sulfur in SO₄²⁻ [35]. In the N 1s spectrum, the cement exhibits two peaks at binding energies of 407.7 eV and 405.0 eV (Figure S2i), representing nitrogen in NO₃⁻ and NO₂, respectively [34]. The O 1s spectrum of the cement shows two peaks at binding energies of 531.5 eV and 533.7 eV (Figure S5i), with the former attributed to oxygen in tricalcium silicate and the latter to oxygen in organic matter within the cement [37].

XRD pattern of the cement reveals that its primary component is tricalcium silicate (PDF No. 73-0599, Ca₃(SiO₄)O, Figure 9a). SEM observations further indicate that the cement consists of particles ranging in size from 2 to 20 μm, corresponding to tricalcium silicate (Figure 10a).

When no admixtures are introduced, that is, when cement is hydrated with pure water only, the resulting 24 h mortar (M-non) examined by XRD shows only the presence of dicalcium silicate monohydrate (PDF No. 29-0373, Ca₂SiO₄·H₂O, Figure 9b). Meanwhile, SEM observations reveal that the morphology of M-non closely resembles that of the raw cement, both exhibiting massive particles; however, the particle size of M-non is somewhat larger than that of the raw cement (Figure 10b vs. Figure 10a). These findings indicate that although cement can undergo hydration with pure water alone, the process is highly incomplete, with dicalcium silicate monohydrate being the main by-product.

When AS is used as the sole admixture, the phase composition of the resulting 24 h mortar M-AS undergoes significant changes compared to the admixture-free case. In addition to the by-products of incomplete tricalcium silicate hydration—dicalcium silicate (PDF No. 20-0237, Ca₂SiO₄, Figure 9c) and SiO₂ (PDF No. 85-0865, Figure 9c)—portlandite (Ca(OH)₂, PDF No. 72-0156, Figure 9c) is also detected. Portlandite is a key hydration product of tricalcium silicate and contributes significantly to enhancing the mechanical strength of the hardened cement [51]. Simultaneously, SEM observations reveal that the particles in M-AS no longer exhibit the sharp edges seen in M-non (Figure 10c vs. Figure 10b).

When T2a is used as an admixture, the resulting P-T2a contains portlandite, unreacted tricalcium silicate, and ettringite (PDF No. 41-1451, Ca₆Al₂(SO₄)₃(OH)₁₂·26H₂O, Figure 9d). The same ettringite phase is also identified in the 24 h mortar M-T2a (Figure 9e). In the Al 2p spectrum, T2a exhibits two photoelectron peaks at binding energies of 75.2 eV and 75.8 eV (Figure S8a), corresponding to aluminum in Al₂(SO₄)₃·14H₂O and AlO(OH), respectively (Figure 3b) [45]. The cement shows a peak at 73.9 eV (Figure S8g), which may be attributed to aluminum in Al₂O₃. In contrast, P-T2a displays three peaks at binding energies of 71.3 eV, 73.7 eV, and 75.7 eV (Figure S8h), assigned to free Al³⁺, Al³⁺ in ettringite (Figure 9d), and aluminum in AlO(OH), respectively [45]. In the O 1s spectrum, T2a shows two peaks at 532.1 eV and 533.0 eV (Figure S5b), representing oxygen in SO₄²⁻ and AlO(OH), respectively [37]. Meanwhile, P-T2a exhibits peaks at 531.2 eV and 533.1 eV (Figure S5j), corresponding to oxygen in ettringite and portlandite, respectively (Figure 9d) [37].

In the C 1s spectra of P-T2a and M-T2a, the binding energies of carboxyl carbon are 288.8 eV and 289.0 eV, respectively, both higher than that of T2a (288.5 eV) (Figure S1j,k vs. Figure S1b). This indicates that during cement hydration, the Al³⁺ originally coordinated with carboxyl groups of organic polymers in T2a may transform into inorganic aluminum-containing compounds [45].

In the N 1s spectrum of T2a, peaks at 403.3 eV, 401.7 eV, and 400.1 eV correspond to nitrogen in –NO, NH₄⁺, and C–N groups, respectively (Figure S2b) [34]. However, in the N 1s spectrum of P-T2a, nitrogen in NH₄⁺ and C–N groups appears at 400.5 eV and 398.6 eV (Figure S2j), while in M-T2a, they are observed at 401.7 eV and 399.3 eV (Figure S2k). Overall, after cement hydration, the N 1s binding energies of NH₄⁺ and C–N groups shift to lower values or remain unchanged. This likely suggests that nitrogen-containing inorganic ions or organic species transition from a relatively free state to a more covalently bound state with cement components (including hydrogen-bonding interactions), resulting in electron delocalization around the nitrogen atom and a consequent decrease in N 1s binding energy.

A similar trend is observed in the S 2p spectra of these samples (Figure S3j,k vs. Figure S3b), further confirming that SO₄²⁻ or –SO₃⁻ groups present in T2a become tightly and covalently bound to cement components after hydration. In contrast, the behavior of Na⁺ is opposite. Comparison of the Na 1s spectra of P-T2a, M-T2a, and T2a shows that the Na 1s binding energy shifts to a higher value after cement hydration (Figure S4i,j vs. Figure S4b), indicating that Na⁺ becomes hydrated or incorporated into the cement hydration products.

The Si 2p peak representing silicon in tricalcium silicate in the raw cement splits into two peaks after hydration to form paste, both shifting to higher binding energies (Figure S12f vs. Figure S12e). In mortar formed after hydration, the Si 2p peak also splits into two components relative to that in cement clinker, with one shifting to lower and the other to higher binding energy (Figure S12g vs. Figure S12e). This suggests that the standard sand (SiO₂) added during mortar preparation retains its original chemical state [48], primarily serving as a reinforcing filler without participating in the hydration reaction (Section 2.5; PDF No. 46-1045, SiO₂, Figure 9e).

Comparison of the FT-IR spectra of T2a, P-T2a, and M-T2a reveals that mortar M-T2a contains more SiO₂ and organic components than paste P-T2a and raw cement. This is evidenced by more pronounced peaks in the M-T2a spectrum at 3670 cm⁻¹ (O–H stretching of surface hydroxyls on SiO₂), 2974 cm⁻¹ and 2906 cm⁻¹ (asymmetric C–H stretching of methyl and methylene groups), and 1069 cm⁻¹ (asymmetric S–O stretching of –SO₃⁻) compared to T2a and P-T2a (Figure 8(Bk) vs. Figure 8(Bi,Bj)) [19,40,46].

Furthermore, UV-Vis spectra of both P-T2a and M-T2a exhibit a distinct absorption peak at 287 nm, attributed to π–π* electronic transitions in organic compounds [52], while peaks in the 300–400 nm range correspond to ligand-to-metal charge transfer (LMCT) in metal complexes present in the samples (Figure 8(Da,Db)) [53]. In summary, T2a actively participates in the cement hydration process, thereby significantly influencing both the hydration kinetics and the mechanical strength of the resulting cement mortar.

Meanwhile, SEM observations reveal that both paste P-T2a and mortar M-T2a contain fibrous substances with lengths of 1–2 μm and cross-sectional diameters of several tens of nanometers (blue circles, Figure 10e,f). These fibers are identified as ettringite (PDF No. 41-1451, Ca₆Al₂(SO₄)₃(OH)₁₂·26H₂O, Figure 9d,e), and their morphology is consistent with previously reported characteristics of ettringite [54]. EDS analysis of the surfaces of P-T2a and M-T2a shows a high distribution density of Al, Si, and O, while Na, C, and N exhibit relatively low distribution densities (Figures S21 and S22), indicating a generally uniform distribution of ettringite.

More importantly, ettringite (PDF No. 41-1451, Ca₆Al₂(SO₄)₃(OH)₁₂·26H₂O), a key binding component, is present in all T2- and T3-derived pastes and 24 h mortars (Figure 9d–o and Figure 10d–x). The cross-sectional diameters of ettringite in various pastes and mortars show minimal variation (Table S4 and Figure 9p) [55]. Simultaneously, SEM observations reveal that C–S–H (stacked or folded nanocrystalline sheets [6]) coexists with nanoscale rod-like ettringite, mutually supporting each other to form a three-dimensional network structure (Figure 9d–o; red and blue circles, Figure 10d–x). The formation of this three-dimensional network likely explains the significantly reduced setting times of T2- and T3-derived pastes and the substantially enhanced compressive and flexural strengths of the corresponding 24 h mortars (entries 3–8 vs. 1, 2, Table 1). The reaction for ettringite formation induced by T2a could be shown in Equation (1) (Figure 3b and Figure 9d):

$$\begin{array}{c}2(\mathrm{AIO}(\mathrm{OH})+(3\mathrm{CaO}·\mathrm{Si}{\mathrm{O}}_2)+3(\mathrm{C}\mathrm{a}\mathrm{S}{\mathrm{O}}_4)+\stackrel{\mathrm{T}2\mathrm{a}(\mathrm{C}\mathrm{a}\mathrm{t}\mathrm{a}\mathrm{l}\mathrm{y}\mathrm{s}\mathrm{t})}{\to }3\mathrm{C}\mathrm{a}\mathrm{O}·\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3·3\mathrm{C}\mathrm{a}\mathrm{S}{\mathrm{O}}_4·32{\mathrm{H}}_2\mathrm{O}+\mathrm{S}\mathrm{i}{\mathrm{O}}_2\\ \mathrm{B}\mathrm{o}\mathrm{e}\mathrm{h}\mathrm{m}\mathrm{i}\mathrm{t}\mathrm{e}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}{\mathrm{C}}_3\mathrm{S}\hspace{1em}\hspace{1em}\hspace{1em}\mathrm{G}\mathrm{y}\mathrm{p}\mathrm{s}\mathrm{u}\mathrm{m}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\mathrm{E}\mathrm{t}\mathrm{t}\mathrm{r}\mathrm{i}\mathrm{n}\mathrm{g}\mathrm{i}\mathrm{t}\mathrm{e}(\mathrm{A}\mathrm{F}\mathrm{T})\hfill \\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\mathrm{a}\mathrm{l}\mathrm{s}\mathrm{o} \mathrm{as} \mathrm{C}{\mathrm{a}}_6\mathrm{A}{\mathrm{l}}_2{(\mathrm{S}{\mathrm{O}}_4)}_3{(\mathrm{O}\mathrm{H})}_{12}·26{\mathrm{H}}_2\mathrm{O}\hfill \end{array}$$

(1)

Since both M-non and M-AS do not contain the ettringite (Ca₆Al₂(SO₄)₃(OH)₁₂·26H₂O) phase (Figure 9b,c), while this phase is observed in both P-T2a and M-T2a (Figure 9d,e), it indicates that the incorporation of T2a may play a crucial role in ettringite formation. The inorganic phase AlO(OH) (PDF No. 83-2384, Figure 3b) present in T2a precedes Al₂(SO₄)₃·14H₂O (PDF No. 49-1097, Figure 3b) and reacts with tricalcium silicate and gypsum during hydration to form ettringite (Equation (1)). In this process, the water-soluble organic polymer components in T2a likely serve to bind various reactive particles and facilitate the reaction (Figure 1f).

In terms of hydrodynamic size and zeta potential, the transition from T2 to T2a results in a gradual increase in hydrodynamic size (1639 nm vs. 1062 nm, Figure 11(Ab) vs. Figure 11(Aa)) and a significant improvement in colloidal stability (1.80 mV vs. −29.0 mV, Figure 11(Bb) vs. Figure 11(Ba)), indicating the coordination of aluminum ions with T2 not only increases the volume of the resulting polymeric complex but also enhances its stability in aqueous environments (Figure 1f).

Furthermore, the hydrated particle size follows the order of T2a (1639 nm, Figure 11(Ab)) < P-T2a (3521 nm, Figure 11(Ca)) < M-T2a (4418 nm, Figure 11(Cb)), while the zeta potential sequence is T2a (1.80 mV, Figure 11(Bb)) > M-T2a (−0.0217 mV, Figure 11(Db)) > P-T2a (−4.69 mV, Figure 11(Da)). Consequently, the particle volumes of both the paste and mortar in aqueous solution increase compared to T2a, demonstrating that T2a can integrate into the inorganic gel network formed during cement hydration. The larger particle volume of the mortar compared to the paste reflects the influence of incorporating standard sand in the mortar formation process. The stability of both paste and mortar in aqueous solution decreases relative to T2a, indicating that T2a rapidly forms stable micelles upon dissolution in water, whereas the hydration process of the cement components in P-T2a and M-T2a can persist for an extended duration. Additionally, the particles of mortar M-T2a in aqueous solution exhibit greater stability than those of paste P-T2a, suggesting that the incorporation of standard sand during the formation of M-T2a contributes to stabilizing the cement components.

On the other hand, within the heating range of 35–250 °C, the weight loss of P-T2a is 15.40 wt.%, while that of M-T2a is 15.05 wt.% (Figure 12(Ab) vs. Figure 12(Aa); entries 2 vs. 1, Table S5). This indicates that the mass percentage of water content (including adsorbed water, coordinated water, and crystalline water) in M-T2a is lower than that in P-T2a, which may be attributed to the incorporation of standard sand (SiO₂) during the formation of mortar M-T2a (Section 2.5). Within the 250–800 °C range, the weight loss of M-T2a is higher than that of P-T2a (8.99 wt.% vs. 6.39 wt.%, Figure 12(Ab) vs. Figure 12(Aa); entries 2 vs. 1, Table S5), probably suggesting that M-T2a contains a higher proportion of organic components compared to P-T2a. However, it should be noted that the weight loss peaks observed around 450 °C for both P-T2a and M-T2a include contributions from the dehydroxylation of Ca(OH)₂ [56]. This is likely due to the strong adsorption capacity of standard sand (SiO₂) for water-soluble polymers during the formation of M-T2a (Figure 1f).

Simultaneously, P-T2a exhibits faster weight loss at 60 °C, 89 °C, and 423 °C, whereas M-T2a shows accelerated weight loss at 62 °C, 92 °C, and 432 °C (Figure 12(Bb) vs. 12(Ba)). The weight loss at the first two temperatures corresponds to the thermal release of water, while the latter represents the thermal decomposition of organic components (including the weight loss derived from dehydroxylation of Ca(OH)₂ [56]). The slightly delayed rapid weight loss observed in M-T2a suggests a stronger interaction between T2a and standard sand (Section 2.5).

Additionally, based on DSC analysis, within 30–300 °C, P-T2a displays two endothermic peaks at 82 °C and 108 °C, while M-T2a exhibits peaks at 90 °C and 108 °C (Figure 12(Cb) vs. 12(Ca); entries 2 vs. 1, Table S5). These peaks correspond to the endothermic release of adsorbed water and coordinated (or crystalline) water, respectively.

Structures of P-T3a and M-T3a

During the synthesis of T3a from T3 and the subsequent formation of P-T3a and M-T3a, a notable distinction emerges when compared to the T2-derived series. Specifically, the AlO(OH) present in T2a is replaced in T3a by highly hydroxylated aluminum sulfate (PDF No. 08-0076, Al₄SO₄(OH)₁₀·36H₂O, Figure 3f vs. Figure 3b), a transformation probably attributed to the introduction of AMPS in the synthesis of T3 (Figure 1f).

The paste P-T3a, mediated by T3a, exhibits a faster setting rate than P-T2a, while the mortar M-T3a demonstrates higher flexural strength than M-T2a (entries 6 vs. 3, Table 1). This indicates that Al₄SO₄(OH)₁₀·36H₂O in T3a more readily promotes the formation of ettringite compared to the AlO(OH) phase in T2a. The reaction equation is proposed as follows:

$$\begin{array}{l}\mathrm{A}{\mathrm{l}}_4\mathrm{S}{\mathrm{O}}_4{(\mathrm{O}\mathrm{H})}_{10}·36{\mathrm{H}}_2\mathrm{O}+(3\mathrm{C}\mathrm{a}\mathrm{O}·\mathrm{S}\mathrm{i}{\mathrm{O}}_2)+3(\mathrm{C}\mathrm{a}\mathrm{S}{\mathrm{O}}_4)\stackrel{\mathrm{T}3\mathrm{a}(\mathrm{Catalyst})}{\to }3\mathrm{C}\mathrm{a}\mathrm{O}·\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3·3\mathrm{C}\mathrm{a}\mathrm{S}{\mathrm{O}}_4·32{\mathrm{H}}_2\mathrm{O}\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}{\mathrm{C}}_3\mathrm{S} \mathrm{G}\mathrm{y}\mathrm{p}\mathrm{s}\mathrm{u}\mathrm{m}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\mathrm{E}\mathrm{t}\mathrm{t}\mathrm{r}\mathrm{i}\mathrm{n}\mathrm{g}\mathrm{i}\mathrm{t}\mathrm{e}(\mathrm{A}\mathrm{F}\mathrm{t})\\ \hfill \mathrm{also} \mathrm{as} \mathrm{C}{\mathrm{a}}_6\mathrm{A}{\mathrm{l}}_2{(\mathrm{S}{\mathrm{O}}_4)}_3{(\mathrm{O}\mathrm{H})}_{12}·26{\mathrm{H}}_2\mathrm{O}\end{array}$$

(2)

From the perspective of surface element distribution and their chemical valence, in the C 1s spectrum, the binding energy of carboxyl carbon in both P-T3a and M-T3a increases to 288.9 eV compared with that in T3a (Figure S1p,q vs. Figure S1f). This indicates that the Al³⁺ originally coordinated with carboxyl groups in T3a is released during cement hydration and participates in the formation of aluminum-bearing cement phases, such as ettringite (PDF No. 41-1451, Ca₆Al₂(SO₄)₃(OH)₁₂·26H₂O; Figure 9j,k). The carboxyl groups in T3a transition from an Al³⁺-coordinated state to either a free/uncoordinated state or a Ca²⁺-coordinated state, resulting in an increased C 1s binding energy for the carboxyl carbon [43].

In the N 1s spectrum of T3a, peaks at 399.8 eV and 401.5 eV correspond to nitrogen in C–N bonds and in NH₄⁺, respectively [34]. During cement hydration (i.e., in P-T3a and M-T3a), both peaks shift toward lower binding energies (Figure S2p,q vs. Figure S2f). This also suggests that the amide nitrogen in T3a changes from an uncoordinated or Al³⁺-coordinated state to a Ca²⁺-coordinated state after hydration. A similar shift trend is observed for the S 2p photoelectrons in T3a, P-T3a, and M-T3a (Figure S3p,q vs. Figure S3f).

The Na 1s photoelectron in T3a shifts toward higher binding energy after cement hydration (i.e., in P-T3a and M-T3a; Figure S4o,p vs. Figure S4f), indicating that Na⁺ in T3a becomes either hydrated by water molecules or incorporated into the cement lattice (e.g., by binding with highly electronegative oxygen anions) [36,44]. In the O 1s spectrum of T3a, peaks at 532.0 eV (oxygen in organic C–O bonds) and 533.5 eV (oxygen in organic O–H bonds) both shift to lower binding energies after cement hydration (i.e., in P-T3a and M-T3a; Figure S5p,q vs. Figure S5f), which is attributed to the introduction of SiO₂ and CaO components during hydration [37].

The Al 2p spectrum of T3a exhibits two peaks at 75.1 eV and 76.2 eV, corresponding to Al³⁺ from inorganic salts (Figure 3f) and Al³⁺ coordinated with the copolymer (Figure 1f) [45]. After T3a-mediated cement hydration, both peaks in P-T3a and M-T3a shift significantly toward lower binding energies (Figure S8n,o vs. Figure S8d), confirming the formation of a new inorganic aluminum-containing phase (ettringite, Ca₆Al₂(SO₄)₃(OH)₁₂·26H₂O; Figure 9j,k).

The Si 2p spectrum of P-T3a shows peaks at 103.5 eV and 101.2 eV (Figure S12l), assigned to silicon in tricalcium silicate (PDF No. 86-0402, Ca₃SiO₅; Figure 9j) and in C–S–H gel, respectively [48]. In M-T3a, the Si 2p peaks appear at 102.4 eV and 101.1 eV (Figure S12m), corresponding to silicon in SiO₂ (PDF No. 89-1961, Figure 9k) and in C–S–H gel [48].

From the perspective of sample morphology and surface element distribution, analysis reveals that both P-T3a and M-T3a contain abundant ettringite nanorods and C–S–H gel nanosheets (Figure 10q–t). The density of these ettringite nanorods and C–S–H nanosheets is significantly higher than that observed in P-T2a and M-T2a (Figure 10d–g). This difference explains why P-T3a and M-T3a exhibit faster setting performance and superior flexural strength compared to P-T2a and M-T2a, respectively (entries 6 vs. 3, Table 1).

EDS analysis of P-T3a indicates that the polygonal aggregates correspond to C–S–H gel nanosheets, where the distribution of sodium, aluminum, sulfur, carbon, and nitrogen elements is notably sparse (Figure S23). In M-T3a, EDS mapping identifies large particles (>10 µm) as SiO₂, micrometer- to nanometer-scale sheet-like structures as C–S–H gel nanosheets, and the presence of aluminum-containing phases, such as ettringite (Figure S24a). Notably, the aluminum-containing phases are almost entirely concentrated on one side of the C–S–H gel nanosheets (Figure S24h).

A comparison of the FT-IR spectra of P-T3a and M-T3a reveals that they are largely identical. The only notable distinction is the vibration peak at 659 cm⁻¹ in P-T3a, which corresponds to the stretching vibration of the Ca–O bond [57]. In M-T3a, this peak shifts to a higher wavenumber (Figure 8(Cq) vs. Figure 8(Cp)). This shift indicates that the chemical environment of the calcium oxide component differs between P-T3a and M-T3a (Figure 9j vs. Figure 9k).

UV-Vis spectra of P-T3a and M-T3a nearly overlap in the range of 250–500 nm, suggesting that both contain essentially the same organic matter and exhibit almost identical coordination of this organic matter with metal ions. However, in the 500–800 nm region, the UV absorption of P-T3a is higher than that of M-T3a, indicating differences in the lattice structures of their inorganic components (Figure 8(Dg) vs. Figure 8(Dh)).

The hydrated particle sizes of T3a, P-T3a, and M-T3a exhibit the following order: M-T3a (2735 nm) > T3a (2205 nm) > P-T3a (2176 nm) (Figure 11(Ch,Af,Cg)), while their zeta potentials follow the sequence: T3a (2.05 mV) > P-T3a (−0.802 mV) > M-T3a (−4.52 mV) (Figure 11(Af,Cg,Ch)). Herein, during T3a-mediated paste hydration, T3a decomposes and incorporates into the paste matrix while hydration of cement particles continues, thereby leading to instability. Furthermore, although M-T3a exhibits the largest hydrated particle size among the three, it shows the lowest stability in aqueous solution. This implies that the standard sand (SiO₂) doped during the synthesis of M-T3a promotes maximal particle size in solution, while the porous structure of the introduced sand likely exhibits strong water-absorption capacity, resulting in the lowest stability of M-T3a in aqueous medium.

P-T3a exhibits weight losses of 16.39% in the 35–250 °C range and 7.61% in the 250–800 °C range, both of which are higher than those of P-T2a (15.40% in 35–250 °C; 6.39% in 250–800 °C) (Figure 12(Da) vs. Figure 12(Aa), entries 7 vs. 1, Table S5). This further demonstrates that during paste synthesis, T3a incorporates more water and organics readily into the cement hydration products than T2a, leading to significantly shorter initial and final setting times for P-T3a relative to P-T2a (entries 6 vs. 3, Table 1). This effect is attributed to the AMPS copolymer unit in T3a, reaffirming the excellent permeability of AMPS (Figure 1f).

Similarly, M-T3a shows a lower weight loss than M-T2a in the 35–250 °C range but a higher weight loss in the 250–800 °C range (Figure 12(Db) vs. Figure 12(Ab), entries 8 vs. 2, Table S5). These data likewise indicate that the mortar synthesized with T3a contains a higher organic content than that synthesized with T2a, thereby enhancing its flexural strength (entries 6 vs. 3, Table 1) and further confirming the superior permeability of AMPS.

3.3.2. Structures of Glass Fiber-Facilitated Pastes and Mortars

Structures of P-T2a-200 and M-T2a-200

T2a-200 mediates a paste with a faster setting rate compared to that mediated by T2a alone (entries 4 vs. 3, Table 1). This acceleration is likely due to the formation of the Al₄SO₄(OH)₁₀·36H₂O phase in T2a-200, which exhibits higher hydration-promoting activity than the AlO(OH) phase present in T2a (Figure 3c vs. Figure 3b). Meanwhile, the average 24 h compressive strength of M-T2a-200 is also higher than that of M-T2a (entries 4 vs. 3, Table 1).

From the perspective of phase composition, P-T2a-200 and P-T2a are essentially identical, with the only difference being the polymorph of residual tricalcium silicate present in each (Figure 9f vs. Figure 9d). Further investigation, however, reveals that the ettringite phase in P-T2a-200 exhibits a significantly smaller particle size compared to that in P-T2a (entries 3 vs. 1, Table S4; Figure 9f vs. Figure 9d; Figure 10i vs. Figure 10e). Consequently, the role of GF-200 in T2a-200 likely involves promoting the accelerated formation of ettringite during paste hydration, thereby shortening the initial and final setting times (entries 4 vs. 3, Table 1), which also results in the smaller ettringite particle size.

The phase composition of M-T2a-200 and M-T2a is fundamentally consistent, with minor distinctions observed in the crystallinity of silica and tricalcium silicate (Figure 9g vs. Figure 9e). More in-depth analysis indicates that the ettringite phase in M-T2a-200 is larger in size than that in M-T2a (entries 4 vs. 2, Table S4; Figure 9g vs. Figure 9e; Figure 10k vs. Figure 10f). This increase in ettringite size can be attributed to the combined effect of GF-200 from T2a-200 and the addition of standard sand (SiO₂) during the preparation of mortar M-T2a-200. This synergistic action ultimately enhances the average 24 h compressive strength of the mortar (entries 4 vs. 3, Table 1).

The molar percentages of carbon, sulfur, silicon, and aluminum on the surface of P-T2a-200 are lower than those on P-T2a (entries 12 vs. 10, Table S1). This indicates that the admixture T2a-200 penetrates more readily into the internal pores of the cement paste hydration products compared to T2a, a difference attributable to their distinct phase compositions (Figure 3c vs. Figure 3b). Additionally, P-T2a-200 contains more large-scale SiO₂ particles than P-T2a (Figure 10h,i vs. Figure 10d,e; Figure S25 vs. Figure S21). These observations explain the faster setting time of P-T2a-200 relative to P-T2a (entries 4 vs. 3, Table 1).

The molar percentages of carbon, sulfur, and calcium on the surface of M-T2a-200 are lower than those on M-T2a. In contrast, the molar percentages of oxygen, nitrogen, silicon, and aluminum are higher (entries 13 vs. 11, Table S1). This suggests that T2a-200 also penetrates more effectively into the internal pores of the mortar’s hydration products. Furthermore, the surface morphology of M-T2a-200 differs significantly from that of M-T2a: M-T2a-200 displays numerous granular particles (Figure 10j–l and Figure S26), whereas M-T2a is characterized by abundant layered and film-like structures (Figure 10f,g and Figure S22). These findings indicate a higher content of cement gel components in M-T2a-200 compared to M-T2a, which accounts for its superior 24 h compressive strength (entries 4 vs. 3, Table 1).

Analysis of the C 1s spectra reveals that the binding energy of carboxyl carbon in P-T2a-200 is higher than that in P-T2a (Figure S1l vs. Figure S1j). A similar increase is observed for the carboxyl carbon binding energy in M-T2a-200 compared to M-T2a (Figure S1m vs. Figure S1k). This indicates that the carboxyl groups in T2a-200 are more prone to dissociate from Al³⁺ coordination during the formation of paste and mortar.

In the N 1s spectra, P-T2a-200 exhibits three peaks at binding energies of 398.0 eV, 399.7 eV, and 401.2 eV, corresponding to nitrogen in C–N bonds, NH₄⁺, and –NO bonds, respectively (Figure S2l) [34]. In contrast, P-T2a shows two peaks at 398.6 eV and 400.5 eV, also representing nitrogen in C–N bonds and NH₄⁺ (Figure S2j) [34]. Compared to P-T2a, the N 1s photoelectrons for both C–N and NH₄⁺ in P-T2a-200 appear at lower binding energies. This suggests that during the T2a-200-mediated paste synthesis, the presence of GF-200 facilitates stronger coordination between the amide groups or NH₄⁺ in T2a-200 with Ca²⁺ or O²⁻ from the cement hydration products. A similar trend is observed when comparing the N 1s spectra of M-T2a-200 and M-T2a (Figure S2m vs. Figure S2k).

The variation in sulfur binding energy follows a trend analogous to that of nitrogen (Figure S3l vs. Figure S3j,m vs. Figure S3k), indicating that the sulfonate or sulfate groups in T2a-200 undergo comparable changes in their coordination environment. A similar binding energy shift is also observed for sodium ions (Figure S4k vs. Figure S4i,l vs. Figure S4j) and oxygen (Figure S5l vs. Figure S5j,m vs. Figure S5k), further supporting the tendency of oxygen in T2a-200 organic components to coordinate with Ca²⁺ during cement hydration, thereby reducing the O 1s photoelectron binding energy. The binding energy shift of aluminum aligns with the trend observed for nitrogen (Figure S8j vs. Figure S8h,k vs. Figure S8i). This similarly indicates that Al³⁺ coordinated with copolymers in T2a-200 is more likely to form coordination bonds with O²⁻ from cement hydration components under the influence of GF-200.

In Si 2p spectrum, P-T2a displays two peaks at 101.9 eV and 102.2 eV, representing silicon in hydrated (or unhydrated) tricalcium silicate and silicon in SiO₂, respectively (Figure S12f) [48]. P-T2a-200 shows three peaks in its Si 2p spectrum at 100.3 eV, 101.3 eV, and 102.4 eV. The first peak corresponds to silicon in C–Si bonds, while the latter two align with the two peaks in P-T2a, albeit with decreased and increased binding energies, respectively [48]. This shift is attributed to the incorporation of GF-200 (noting that the Si 2p photoelectron binding energy for pure SiO₂ typically ranges between 103.0 eV and 104.0 eV [58]). Meanwhile, compared to M-T2a, both peaks in the Si 2p spectrum of M-T2a-200 shift toward higher binding energies (Figure S12i vs. Figure S12g), further reflecting the influence of GF-200 incorporation.

T2a-200 displays a larger hydrodynamic particle size and improved colloidal stability relative to T2a (Figure 11(Ac) vs. Figure 11(Ab); Figure 11(Bc) vs. Figure 11(Bb)). P-T2a-200 exhibits a smaller hydrodynamic particle size yet higher stability than P-T2a (Figure 11(Cc) vs. Figure 11(Ca); Figure 11(Dc) vs. Figure 11(Da)). The same trends are observed when comparing mortar M-T2a-200 with M-T2a (Figure 11(Cd) vs. Figure 11(Cb); Figure 11(Dd) vs. Figure 11(Db)). The experimental results indicate that GF-200 not only stabilizes the admixture T2a-200 but also contributes to the stability of both cement paste and mortar during hydration.

Within 35–250 °C, P-T2a-200 exhibits lower weight loss than P-T2a, whereas in 250–800 °C, P-T2a-200 shows higher weight loss compared to P-T2a (Figure 12(Ac) vs. Figure 12(Aa); entries 3 vs. 1, Table S5). A similar trend is observed in the thermogravimetric behavior between M-T2a-200 and M-T2a (Figure 12(Ad) vs. Figure 12(Ab); entries 4 vs. 2, Table S5). These results indicate that the incorporation of GF-200 into the admixture reduces the water content in the resulting cement while enhancing the uptake of polymeric components from the admixture.

Structures of P-T2a-2000 and M-T2a-2000

Substituting GF-200 with GF-2000 in the admixture significantly shortens the final setting time of the paste, while prolonging the average initial setting time and reducing both the average compressive strength and average flexural strength (entries 5 vs. 4, Table 1). This can be attributed to the fact that both T2a-2000 and T2a-200 contain the same SiO₂ and Al₂Si₂O₅(OH)₄ components, but T2a-2000 additionally contains Al₂SO₄(OH)₄·7H₂O (PDF No. 08-0055, Figure 3d), which differs from Al₄SO₄(OH)₁₀·36H₂O (PDF No. 08-0076, Figure 3c) present in T2a-200, the latter containing more hydroxyl groups. Comparing the XRD patterns of P-T2a-2000 and P-T2a-200 reveals similar phase compositions, with the primary distinction being the crystalline form of tricalcium silicate (Figure 9h vs. Figure 9f). In contrast, a comparison of the XRD patterns of M-T2a-2000 and M-T2a-200 indicates the absence of unhydrated tricalcium silicate in M-T2a-2000 (Figure 9i vs. Figure 9j).

It is observed that compared to P-T2a-200, P-T2a-2000 exhibits an increase in the molar percentages of carbon and calcium, while the contents of oxygen, nitrogen, silicon, and aluminum decrease (entries 14 vs. 12, Table S1). Similarly, M-T2a-2000 shows higher molar percentages of carbon, oxygen, calcium, and silicon but lower contents of nitrogen, sulfur, and aluminum relative to M-T2a-200 (entries 15 vs. 13, Table S1). These results indicate that GF-2000 has a higher mesh number and larger specific surface area than GF-200. Consequently, when T2a-2000 is used as an additive instead of T2a-200, the surface silicon content in the resulting cement paste and mortar decreases. Moreover, because SiO₂ itself is not hydrophilic, the increase in its specific surface area makes it more difficult for the organic copolymers in the additive to penetrate into the cement hydration products. Instead, these copolymers tend to remain on the surface of the paste and mortar. At the same time, however, Al³⁺ coordinated with the organic copolymers enters into the cement hydration products and becomes part of them.

In C 1s spectra, the carboxyl carbon peak for P-T2a-2000 appears at 288.8 eV, which is lower than that for P-T2a-200 at 289.0 eV (Figure S1n vs. Figure S1l) [33]. Similarly, the carboxyl carbon peak for M-T2a-2000 is observed at 288.9 eV, also lower than that for M-T2a-200 at 289.1 eV (Figure S1n vs. Figure S1l). These results indicate that in both paste and mortar mediated by GF-2000, the carboxyl groups of the organic copolymers undergo a greater degree of Al³⁺ decoordination compared to those mediated by GF-200. In the S 2p spectra, the doublet peaks of P-T2a-2000 versus P-T2a-200 (Figure S3n vs. Figure S3l) and of M-T2a-2000 versus M-T2a-200 (Figure S3o vs. Figure S3m) exhibit the same trend as observed in the C 1s spectra. This further confirms that the sulfonate groups of the organic copolymers in GF-2000-mediated systems experience a higher extent of Al³⁺ decoordination than those in GF-200-mediated systems [45].

In Al 2p spectra, both Al 2p photoelectron peaks of P-T2a-2000 shift significantly toward lower binding energy compared to those of P-T2a-200 (Figure S8l vs. Figure S8j), and a similar trend is observed in the comparison between M-T2a-2000 and M-T2a-200 (Figure S8m vs. S8k). These results indicate, from a cationic perspective, that the degree of Al³⁺ dissociation from organic copolymers in both paste and mortar mediated by GF-2000 is higher than that mediated by GF-200. The dissociated Al³⁺ subsequently incorporates into the inorganic cement hydration phases [45].

The two Si 2p photoelectron peaks of P-T2a-2000 shift toward lower binding energy compared to the corresponding peaks of P-T2a-200 (Figure S12j vs. Figure S12h). A similar shift is observed between M-T2a-2000 and M-T2a-200 (Figure S12k vs. Figure S12i). This can be explained by the fact that GF-2000 has a higher mesh number and thus a larger specific surface area than GF-200, along with a greater concentration of crystal dislocations and defects. Consequently, during the hydration of cement paste and mortar under alkaline conditions, GF-2000 (SiO₂) participates more readily in the cement hydration reactions compared to GF-200, resulting in a reduction in the binding energy of Si 2p photoelectrons.

The FT-IR peaks at 1400 cm⁻¹ (corresponding to the symmetric stretching vibrations of the –COO⁻ group in sodium alginate) and 1241 cm⁻¹ (corresponding to the asymmetric stretching vibration of the S=O bond in the sulfonate group) are more pronounced in P-T2a-2000 than in P-T2a-200 (Figure 8(Bn) vs. Figure 8(Bl)) [40]. This observation further confirms that the extent of Al³⁺ dissociation from organic copolymers in paste mediated by GF-2000 is higher than that mediated by GF-200.

Upon comparing the morphology of P-T2a-2000 with that of P-T2a-200, it is observed that the block-like objects in P-T2a-2000 are relatively smaller in volume, while P-T2a-200 contains larger block-like objects (Figure 10m,n vs. Figure 10h,I, Figure S27 vs. Figure S25). A similar morphological trend is noted between M-T2a-2000 and M-T2a-200 (Figure 10o,p vs. Figure 10j,k, Figure S28 vs. Figure S26). This difference can be attributed to the volume distinctions between GF-2000 and GF-200 (Figure 5a–e), as well as between T2a-2000 and T2a-200 (Figure 5k–n). In conjunction with the performance test results of paste and mortar (entries 5 vs. 4, Table 1), it is evident that while smaller-volume glass fibers participate more actively during cement hydration—thereby reducing the final setting time of paste—the enhancement of compressive strength and flexural strength in mortar is more effectively achieved by using larger-volume glass fibers.

The following observations are made regarding aqueous particle size: P-T2a-2000 (4734 nm) > P-T2a-200 (2622 nm) (Figure 11(Ce) vs. Figure 11(Cc)), and M-T2a-2000 (2546 nm) > M-T2a-200 (2287 nm) (Figure 11(Cf) vs. Figure 11(Cd)). In terms of zeta potential: P-T2a-2000 (−5.31 mV) < P-T2a-200 (−2.21 mV) (Figure 11(De) vs. Figure 11(Dc)), and M-T2a-2000 (−7.15 mV) < M-T2a-200 (0.523 mV) (Figure 11(Df) vs. Figure 11(Dd)). It is evident that the paste particles prepared with GF-2000 exhibit larger hydrated particle sizes but lower stability in aqueous solution compared to those prepared with GF-200. A similar trend is observed for the mortar. These findings explain why the overall setting-acceleration and strength-enhancing performance of T2a-200 is superior to that of T2a-2000 (entries 4 vs. 5, Table 1).

Structures of P-T3a-200 and M-T3a-200

T3a-200 mediates a paste with a faster setting rate compared to that mediated by T3a. The 24 h mortar also exhibits an increase in average compressive strength but a decrease in flexural strength (entries 7 vs. 6, Table 1). From XRD analysis, T3a-200 and T3a exhibit distinct differences. T3a-200 contains Al₄SO₄(OH)₁₀·5H₂O (PDF No. 25-1491, Figure 3g) and Al₂Si₂O₅(OH)₄ (PDF No. 14-0164, Figure 3g), which are likely key components contributing to its superior setting-acceleration and strength-enhancing properties (Figure 3g vs. Figure 3f). Moreover, the phase compositions of P-T3a-200 and P-T3a are essentially identical (Figure 9l vs. Figure 9j). However, the transition states involved in their formation likely differ, as indicated by the smaller size of the ettringite phase in P-T3a-200 compared to that in P-T3a (entries 9 vs. 7, Table S4). In contrast, notable differences exist in the phase composition between M-T3a-200 and M-T3a. Specifically, M-T3a-200 contains residual unhydrated tricalcium silicate (Figure 9m vs. Figure 9k).

From a morphological standpoint, P-T3a exhibits abundant rod-like ettringite and sheet-like C–S–H phases (Figure 10q,r and Figure S23). In P-T3a-200, however, the presence of GF-200 likely contributes to the deformation of these rod-like ettringite and sheet-like C–S–H structures, along with the appearance of numerous granular particles, which may originate from GF-200 (Figure 10u and Figure S29). M-T3a contains a substantial amount of sheet-like material with dimensions on the order of tens of micrometers (Figure 10s,t and Figure S24). In comparison, M-T3a-200 displays nanoscale ettringite and C–S–H phases (Figure 10v and Figure S30).

Both Si 2p peaks in the spectrum of P-T3a shift to lower binding energies in P-T3a-200 (Figure S12n vs. Figure S12l). The two Si 2p peaks in the spectrum of M-T3a shift toward higher binding energies in M-T3a-200 (Figure S12o vs. Figure S12m). These results indicate that the incorporation of GF-200 effectively facilitates its own transformation into components of cement hydration products. The higher Si 2p binding energy observed in M-T3a-200 primarily reflects the synergistic influence of GF-200 and standard sand.

It can be observed that the hydrated particle size of P-T3a-200 is larger than that of P-T3a (Figure 11(Ci) vs. Figure 11(Cg)), while the stability of P-T3a-200 in water is higher than that of P-T3a (Figure 11(Di) vs. Figure 11(Dg)). Similarly, the comparison between M-T3a-200 and M-T3a shows the same trend in hydrated particle size and stability in water (Figure 11(Cj) vs. Figure 11(Ch), Figure 11(Dj) vs. Figure 11(Dh)).

Evidently, the admixture composed of T3a and GF-200 significantly promotes particle growth in cement paste and mortar and enhances their stability in aqueous solution compared to the T3a admixture alone. This finding corroborates the experimental results that the overall performance of P-T3a-200 is superior to that of P-T3a, and M-T3a-200 is superior to M-T3a (entries 7 vs. 6, Table 1).

Within 35–250 °C, P-T3a-200 exhibits higher thermal weight loss than P-T3a, whereas in 250–800 °C, P-T3a-200 shows lower thermal weight loss than P-T3a (Figure 12(Dc) versus 12(Da); entries 9 versus 7, Table S5). A similar trend is observed between M-T3a-200 and M-T3a (Figure 12(Dd) versus 12(Db); entries 10 versus 8, Table S5). In contrast, the opposite pattern is seen for P-T2a-200 versus P-T2a (Figure 12(Ac) vs. Figure 12(Aa); entries 3 versus 1, Table S5) and for M-T2a-200 versus M-T2a (Figure 12(Ad) vs. Figure 12(Ab); entries 4 vs. 2, Table S5). These results indicate that when AMPS is used as a comonomer, the resulting copolymer T3a, likely due to its long side chains containing amide, methyl, and sulfonic acid groups, tends to encapsulate the particles of cement hydration products rather than penetrate into their internal pores (Figure 1f).

Structures of P-T3a-2000 and M-T3a-2000

Replacing 200-mesh glass fibers with 2000-mesh fibers leads to a prolonged initial setting time and a shortened final setting time in paste P-T3a-2000, along with reduced compressive and flexural strengths in mortar M-T3a-2000 (entries 8 vs. 7, Table 1). Given that T3a-200 and T3a-2000 exhibit highly similar phase compositions, T3a-200 contains Al₂Si₂O₅(OH)₄ (PDF No. 14-0164; Figure 3g), while T3a-2000 contains Al₂O₃·2SiO₂·2H₂O (PDF No. 03-0052; Figure 3h). In contrast, XRD analysis reveals that P-T3a-200 and P-T3a-2000 share identical phases (Figure 9n vs. Figure 9l). Similarly, M-T3a-200 and M-T3a-2000 also contain the same phases, except for differences in the polymorph of tricalcium silicate present (Figure 9o vs. Figure 9m). On the other hand, P-T3a-2000 contains a greater amount of larger porous block-like material (Figure 10w and Figure S31), whereas P-T3a-200 possesses more granular or flaky particles with sizes below 1 µm (Figure 10w and Figure S31).

In S 2p spectra, the binding energies of the doublet peaks in P-T3a-2000 are higher than those in P-T3a-200 (Figure S3t vs. Figure S3r). Similarly, the binding energies of the doublet peaks in M-T3a-2000 are higher than those in M-T3a-200 (Figure S3u vs. Figure S3s). Compared with GF-200, GF-2000 more effectively promotes the removal of coordinated Al³⁺ from the sulfonate groups in T3a, leading to an increase in the binding energy of S 2p photoelectrons.

In the Na 1s spectra, the binding energies of the doublet peaks in P-T3a-2000 are higher than those in P-T3a-200 (Figure S4s vs. Figure S4q). This suggests that GF-2000 facilitates the incorporation of Na⁺ from T3a into the hydration products of cement paste more effectively than GF-200.

In the Al 2p spectra, the doublet peaks of P-T3a-2000 at 73.3 eV and 74.0 eV show a decrease compared with those of P-T3a-200 at 73.5 eV and 74.7 eV (Figure S8r vs. Figure S8p). Moreover, in the comparison between M-T3a-2000 and M-T3a-200, the peak at 75.8 eV present in M-T3a-200 disappears in the Al 2p spectrum of M-T3a-2000 (Figure S8s vs. Figure S8q). These observations demonstrate that GF-2000 promotes the incorporation of highly free Al³⁺ from T3a into the cement hydration components more effectively than GF-200.

In the Si 2p spectra, consistent trends are observed between P-T3a-2000 and P-T3a-200 (Figure S12p vs. Figure S12n) and between M-T3a-2000 and M-T3a-200 (Figure S12q vs. Figure S12o). This further confirms that GF-2000 enhances the hydration of silicon-containing components more effectively than GF-200.

P-T3a-2000 exhibits a larger hydrated particle size than P-T3a-200 (Figure 11(Ck) vs. Figure 11(Ci)). However, the hydrated particles of P-T3a-200 show higher stability than those of P-T3a-2000 (Figure 11(Di) vs. Figure 11(Dk)). In contrast, M-T3a-2000 has a smaller hydrated particle size than M-T3a-200 (Figure 11(Cl) vs. Figure 11(Cj)), while the hydrated particles of M-T3a-200 demonstrate greater stability than those of M-T3a-2000 (Figure 11(Dj) vs. Figure 11(Dl)). Compared to GF-2000, GF-200 more effectively enhances the stability of particles in cement paste and mortar. Consequently, the overall performance of cement paste and mortar mediated by GF-200 surpasses that mediated by GF-2000 (entries 7 vs. 8, Table 1).

Finally, M-T3a-2000 exhibits higher weight loss than M-T3a-200 in 35–250 °C, but lower weight loss in 250–800 °C (Figure 12(Df) vs. Figure 12(Dd); entries 12 vs. 10, Table S5). Therefore, M-T3a-2000 contains more water than M-T3a-200, while M-T3a-200 retains a higher amount of organic matter compared to M-T3a-2000. In other words, GF-200 likely promotes the adsorption of organic matter by cement mortar more effectively than GF-2000, thereby contributing to the enhanced overall mechanical performance of the cement mortar (entries 7 vs. 8, Table 1).

4. Conclusions

Based on the findings of this study, the following conclusions can be drawn:

• Using sodium alginate (SA), sodium 2-methylprop-2-ene-1-sulfonate (SMAS), and 2-acrylamido-2-methylpropane sulfonic acid (AMPS, with or without addition) as copolymer precursors, two highly water-soluble copolymers—p(SA-co-SMAS) and p(SA-co-SMAS-co-AMPS)—were synthesized. The p(SA-co-SMAS) copolymer exhibits a high molecular weight of Mₙ = 1.591 × 10⁸ g mol⁻¹ and a narrow molecular weight distribution (M_w/Mₙ = 1.238), while p(SA-co-SMAS-co-AMPS) has a lower molecular weight (Mₙ = 6.529 × 10⁶ g mol⁻¹) and an even narrower distribution (M_w/Mₙ = 1.103), due to the steric hindrance introduced by the longer branched chains of AMPS during polymerization. ¹H NMR analysis reveals that the molar ratio of SA to SMAS units in p(SA-co-SMAS) is approximately 1.5:1.
• Under ball milling, blending p(SA-co-SMAS) with aluminum sulfate yields an admixture containing a boehmite (AlO(OH)) phase, in which aluminum-containing phases such as boehmite encapsulate the p(SA-co-SMAS) copolymer. Compared with acceleration reactions without admixtures or with aluminum sulfate as admixture, the setting time of the resulting paste is significantly shortened. Moreover, the mortar mediated by this synthesized admixture exhibits substantially enhanced compressive and flexural strengths. This improvement is probably due to the rapid formation of ettringite (Ca₆Al₂(SO₄)₃(OH)₁₂·26H₂O) through the reaction of boehmite with tricalcium silicate, gypsum, and water in the cement. The resulting ettringite interlocks with the two-dimensional C–S–H gel, forming a stable three-dimensional network structure.
• Further blending the synthesized admixture with 200-mesh glass fibers under ball milling produces a new cement admixture containing Al₄SO₄(OH)₁₀·36H₂O. Compared to the boehmite (AlO(OH)) phase in the previous admixture, this phase further significantly shortens the initial and final setting times of the cement paste and increases the average compressive strength of the mortar. Similar to AlO(OH), Al₄SO₄(OH)₁₀·36H₂O promotes the formation of ettringite, but it exhibits superior accelerating effectiveness.
• The substitution of 200-mesh glass fibers with 2000-mesh glass fibers in the synthesis of the admixture with p(SA-co-SMAS) does not result in improved setting acceleration or reinforcement. Firstly, the newly formed active intermediate, Al₂SO₄(OH)₄·7H₂O, contains fewer hydroxyl groups than the original Al₄SO₄(OH)₁₀·36H₂O, leading to a weaker acceleration of cement hydration. Additionally, although the 2000-mesh glass fibers promote ion migration and enhance the extent of cement hydration more effectively than the 200-mesh fibers, the significantly larger volume of the 200-mesh fibers contributes to fewer voids and a more robust structure in the resulting cement mortar. Lastly, the cement paste and mortar prepared with the admixture synthesized from 200-mesh glass fibers and p(SA-co-SMAS) exhibit superior stability in aqueous solution.
• The admixture obtained by blending p(SA-co-SMAS-co-AMPS) with aluminum sulfate demonstrates better setting acceleration performance than that derived from p(SA-co-SMAS) and aluminum sulfate: both initial and final setting times are further reduced, and the flexural strength of the mortar is significantly improved. This enhancement is attributed to the presence of the Al₄SO₄(OH)₁₀·36H₂O phase in the AMPS-mediated accelerator, which exhibits superior accelerating effectiveness compared to AlO(OH). Therefore, in terms of accelerating cement hydration and enhancing strength, a positive synergistic effect exists between the synthesized copolymer and aluminum sulfate. Introducing AMPS as a synthetic monomer into the copolymer contributes to the formation of a rapid-setting agent composition with high early-strength and high reinforcing activity.
• Incorporating 200-mesh glass fibers into the admixture composed of p(SA-co-SMAS-co-AMPS) and aluminum sulfate further shortens the initial and final setting times of the cement paste and enhances the 24 h average compressive strength of the mortar, compared to an admixture consisting solely of p(SA-co-SMAS-co-AMPS) and aluminum sulfate. This improvement is likely attributed to the effective pore-filling by the glass fibers and organic components during cement hydration. In comparison with an accelerator formulated from p(SA-co-SMAS), aluminum sulfate, and 200-mesh glass fibers, the current admixture similarly and significantly reduces the setting times of the cement paste, although it does not lead to a marked increase in the mortar’s mechanical strengths. The underlying reason is that p(SA-co-SMAS-co-AMPS), which contains the highly water-soluble AMPS monomer rich in coordinating atoms, promotes the formation of larger and more stable hydrated particles in cement paste and mortar than does p(SA-co-SMAS). Thus, a clear positive synergistic effect is observed among the copolymer, aluminum sulfate, and 200-mesh glass fibers in accelerating the setting of cement hydration.
• Substituting 200-mesh glass fibers with 2000-mesh glass fibers in the admixture containing p(SA-co-SMAS-co-AMPS) has no significant effect on shortening the initial setting time of cement paste or on improving the mechanical strength of mortar, and only slightly reduces the final setting time. This result further confirms that the size of the glass fibers in the admixture markedly influences both the hydration rate and the strength development of cement.

In contrast to conventional approaches where sodium alginate typically acts as a retarder, glass fibers serve solely as reinforcing agents without accelerating effects, and aluminum sulfate functions as an accelerator at the expense of mechanical strength, this study presents a synergistic strategy. By modifying sodium alginate with superplasticizer monomers SMAS and AMPS, and combining them with aluminum sulfate and glass fibers, a cooperative interaction among the components facilitates an ultra-rapid and high-strength cement acceleration process. The preliminary findings of this research warrant attention from experts and engineers in the field of construction materials science.

Further investigations, including the long-term mechanical strength effects of the proposed admixture system and more detailed mechanistic studies on cement acceleration, are currently ongoing in our laboratory. Related outcomes are expected to contribute continuously to the advancement and sustainable development of cement admixture technologies.