Effects of steel slag content on the performance of alkali-activated coal gangue-slag mortar

Abstract

To address the significant shrinkage issue of alkali-activated materials (AAMs) and to utilize industrial solid waste, this study used steel slag (SS), coal gangue (CG), and ground granulated blast furnace slag (GGBFS) as the raw materials, maintaining the proportion of GGBFS at 50% to guarantee fundamental strength, and substituting a portion of CG with SS to formulate alkali-activated coal gangue-slag mortar (ACGM). Based on existing research, this study examined the mechanical characteristics and desiccation contraction of ACGM with various SS proportions. In addition, microscopic analyses were employed to analyze the impact of SS on the reaction processes and structural evolution of ACGM. The results demonstrated that when 40% SS was added, more gel and ettringite were produced in the system. These products interwove to form a dense microstructure while inhibiting gel deformation, leading to a 41.7% increase in the 28-day compressive strength regarding ACGM. Furthermore, the 28-day drying shrinkage associated with ACGM was reduced by 28.9%. This improvement was attributed to the addition of SS, which slowed the decrease in internal relative humidity. Additionally, the formation of ettringite inhibited matrix deformation and reduced drying shrinkage. However, incorporating 50% SS was found to inhibit both mechanical strength development and drying shrinkage reduction. The utilization of SS in ACGM provided a pathway for large-scale application of SS in AAMs, offering novel insights for mitigating shrinkage in AAMs.

Introduction

Alkali-activated materials (AAMs) have attracted significant attention from researchers since their emergence in 1895 due to their green, low-carbon, and environmentally friendly characteristics^1,2. The precursors of AAMs typically include industrial solid waste, such as ground granulated blast furnace slag (GGBFS), coal gangue (CG), steel slag (SS), phosphogypsum, red mud, and fly ash³. After appropriate excitation treatment, these materials exhibit excellent working performance, such as high compressive strength, durability, resistance to chloride ion erosion, high temperature resistance, and low hydration heat^4,5. In addition, producing 1 ton of AAMs can reduce CO₂ emissions by nearly 1 ton compared to producing 1 ton of OPC^6,7. Currently, AAMs can be categorized into high-calcium systems, including GGBFS and SS, and low-calcium systems, which encompass fly ash and CG, based on the calcium content in the precursor⁸. The primary hydration product in high-calcium systems is C-A-S-H gel, while N-A-S-H gel dominates in low-calcium systems⁹. GGBFS and CG serve as common precursors in AAM research¹⁰, combining elevated-calcium and reduced-calcium systems. The hydration products mainly comprise C-(N)-A-S-H gel and ettringite, thereby utilizing the advantages of both high-calcium systems (i.e., higher compressive strength) and low-calcium systems (i.e., stronger acid resistance)^11,12.

Although AAMs possess numerous excellent working properties, their drying shrinkage is typically more than double that of OPC, hindering the widespread adoption of AAMs. The elevated contraction rate of AAMs can lead to the uneven deformation of the material, resulting in harmful cracks. These damaging cracks facilitate the entry of corrosive substances into the interior of materials, reducing their load-bearing capacity and structural durability^13,14. To reduce the high shrinkage of AAMs, researchers have proposed several mitigation strategies, mainly including the following: adjusting the alkali equivalent and alkali activator type; adding mineral additives, chemical additives, and fiber reinforcement agents; using porous aggregates; and improving maintenance conditions^14,15. Although these methods can reduce the shrinkage rate of AAMs, numerous cracks caused by shrinkage can still be observed in AAMs, and these methods also face issues such as high cost and narrow application range. Xie et al.¹⁶ noted that the primary constituents of alkali-activated mineral powder fly ash systems consisted primarily of N-A-S-H gel and minor crystalline phases. Zhang et al.¹⁷ highlighted that the lack of crystalline species including ettringite, portlandite, and hydrotalcite represented a key factor contributing to the elevated drying shrinkage detected in AAMs. This was because the crystal was more stable than the gel and could be used as a fine aggregate in the matrix to alleviate shrinkage. In addition, some crystals such as ettringite and hydrotalcite may further partially compensate for the shrinkage of the matrix due to their own micro expansion effect¹⁸, resulting in a more stable structure with reduced shrinkage.

Steel slag is an alkaline solid waste generated as a by-product of the steelmaking process, and it is commonly categorized into two categories: basic oxygen furnace slag and electric arc furnace slag. For every 1 ton of crude steel manufactured, approximately 0.1–0.15 tons of SS are produced along with the release of 1.85 tons of CO₂. The accumulation of SS also leads to numerous environmental problems, such as dust accumulation and groundwater and air pollution, as well as occupying significant portions of land, posing economic and environmental challenges for developing countries^19,20. However, due to its high hardness, density, and chemical similarity to cement (main minerals consisting of C₂S, C₃S, Fe₂O₃, and SiO₂), SS has the potential to prepare cement-based materials, especially basic oxygen furnace slag^21–23. The inclusion of SS may also enhance the flowability of cement, and small inclusions of SS in concrete structures do not have adverse effects on structural strength. Therefore, SS is often used as a partial substitute for sand and cement in civil engineering^24–26. At present, SS has been used to fabricate asphalt pavement for highways, as railway ballast, and as concrete aggregate²⁷. Han et al.^28,29 found that the hydration products of concrete or AAMs mixed with SS, including C-S-H and ettringite, contributed to enhancing the material’s strength and minimizing shrinkage. However, f-MgO and f-CaO in SS frequently lead to volume expansion, resulting in the cracking of concrete and reducing strength and durability³⁰. Nevertheless, the volume expansion of f-MgO and f-CaO may be beneficial. Brand et al.³¹ found that the micro expansion effect of f-MgO and f-CaO during hydration could alleviate the high shrinkage problem of AAMs. Therefore, SS could serve as a potential material for reducing the shrinkage rate of AAMs. However, a research gap remains regarding the effects of the hydration product of SS in AAMs, with few published studies on the impacts of altering SS content on the mechanical strength and drying shrinkage of AAMs.

Therefore, to improve the high shrinkage rate of AAMs and expand the recycling of SS, as well as to determine the role and reaction principle of SS in AAMs, this study used CG and GGBFS as precursors and water glass and sodium hydroxide as the alkali activators to prepare alkali-activated coal gangue-slag mortar (ACGM). We subsequently assessed the feasibility of improving the drying shrinkage and other working properties of ACGM with different SS contents. This article explored the mineral composition, microstructure, and morphology of ACGM treated with SS using X-ray powder diffractometry (XRD), Fourier transform infrared spectroscopy (FTIR), thermogravimetric/derivative thermogravimetric analysis (TG/DTG), scanning electron microscopy (SEM), the hydration heat and nitrogen adsorption techniques, and mercury intrusion porosimetry. These methods were employed to determine the mechanisms by which SS improved the working performance and drying shrinkage of ACGM. Compared to previous studies, this study focused on directly utilizing existing SS without additional treatment such as carbonization. Finally, this study offers suggestions for the application of SS in AAMs, providing new directions for the recycling and utilization of SS and the research on AAMs.

Materials and methods

Raw materials

The CG employed in this study was sourced from a regional mine in Guizhou Province, whereas the SS (basic oxygen furnace slag) and GGBFS originated from Henan Borun Casting Materials Co., Ltd. (Gongyi, China). The density and specific surface area of CG were 2.54 g/cm³ and 350 m²/kg, respectively, those of SS were 3.25 g/cm³ and 390 m²/kg, respectively, and those of GGBFS were 3.00 g/cm³ and 412 m²/kg, respectively. The chemical composition and particle size distribution of the base materials are presented in Table 1 and Fig. 1, respectively. The XRD patterns of CG and GGBFS are shown in Fig. 2, where the primary constituents of CG consisted of quartz, anorthite, zeolite, and a small amount of calcite³². The primary components of GGBFS were the glass phase and calcite³³. The XRD pattern of SS is shown in Fig. 3, with its main phases consisting of SiO₂, C₂S, and C₃S³⁴. Combining Fig. 1 and Fig. 4a,b, CG exhibited a D50 value of 19.6 µm, with particle morphology characterized by irregular flakes or blocks³⁵. GGBFS had a D50 value of 7.2 µm, while SS had a D50 value of 49.3 µm, indicating loosely aggregated solid blocks with irregular shapes and small particles attached to the surface³⁶. The alkaline activator used in the experiment was solid water glass (34.3% Na₂O and 63.4% SiO₂) and solid sodium hydroxide with a purity exceeding 98%. The particle size range of the sand employed in this study was 0.075–1.180 mm.

Table 1.

Chemical compositions of the raw material.

Chemical analysis of CG, GGBFS, and SS (wt%)
Oxides	SS	GGBFS	CG
SiO2	19.1	29.5	47.4
Al2O3	6.1	15.9	26.5
MgO	6.9	9.3	1.9
Fe2O3	22.1	0.3	9.9
CaO	40.6	39.7	2.3
TiO2	1.3	1.4	6.9

Fig. 1.

Particle size distributions of GGBFS, CG, and SS.

Fig. 2.

Crystal composition of GGBFS and CG.

Fig. 3.

Crystal composition of SS.

Fig. 4.

(a) SEM image of CG. (b) SEM image of SS.

The activation of coal gangue entailed both mechanical and thermal activation. Initially, raw coal gangue was pulverized into small fragments with a mallet to obtain blocky coal gangue. Then, the crushed coal gangue was ground in a ball mill until it could pass through a 200-mesh sieve. Finally, the ground coal gangue powder was calcined in a muffle furnace at 700 °C for 2 h to obtain the activated coal gangue powder. Both SS and GGBFS were sieved through a 200-mesh sieve.

Experimental methods

The experimental design is summarized in Table 2, with a water-to-binder ratio of 0.4. The experiment used sodium silicate (Na₂O: 34.3 wt%, SiO₂: 63.4 wt%) as the activator, with a modulus of 2.0. The modulus of sodium silicate was adjusted to 1.5 through the incorporation of sodium hydroxide with a purity level of 98%, sustaining a mass ratio of sodium silicate to sodium hydroxide of 8.27. ACGM was prepared at room temperature (20 °C). First, the precursor materials CG, SS, GGBFS, and sand were mixed and stirred for 2 min; then, the activator solution was added and stirred for 3 min. The solution was then immediately combined with adhesive sand and poured into 40 mm × 40 mm × 40 mm triple molds. The molds were shaken on a vibration table for 60 s, and the surface of the mold was covered with a layer of polyethylene film after vibration completed; then, the mold was placed in a curing box (20 °C, 95% humidity) for 24 h. The experimental flowchart is shown in Fig. 5.

Table 2.

Mix proportions of ACGM (wt%).

Mix no	CG (%)	SS (%)	GGBFS (%)	Precursor (%)	Activator/precursor (%)	Sand/precursor (%)	Water/precursor (%)	Modulus
C5S0	50	0	50	100	15	100	40	1.5
C4S1	40	10	50	100	15	100	40	1.5
C3S2	30	20	50	100	15	100	40	1.5
C2S3	20	30	50	100	15	100	40	1.5
C1S4	10	40	50	100	15	100	40	1.5
C0S5	0	50	50	100	15	100	40	1.5

Fig. 5.

Experimental flowchart.

Experimental tests

Compressive strength test

This study carried out compressive strength tests according to the GB/T 17671-2021 standard³⁷. Following the mix ratio specified in Table 2 and the experimental procedure detailed in "Experimental methods" secction, the ACGM specimens were tested at their designated curing ages (7, 14, and 28 days), where the loading rate of the experiment was 2.5 kN/s.

Drying shrinkage test

The drying shrinkage test in this study was conducted with reference to ASTM C596³⁸. The mortar samples used for measuring drying shrinkage were cast into a mold with dimensions of 25 mm × 25 mm × 285 mm. The samples were initially cured under ambient conditions over a 24-h period, subsequently removed from the molds, and submerged in water to proceed with further curing over an additional 24-h duration. Prior to monitoring the dimensional variations, the baseline measurements of the samples were documented. Next, the samples were placed in a curing box for curing, where the temperature of the curing box was 25 °C and the relative humidity was 50%. The length of the specimens was measured daily for the first 7 days of the experiment and once a week after 7 days. The drying shrinkage of the specimens was calculated as follows:

1

where represents the drying shrinkage (με), is the initial specimen length (mm), and is the length at the time of measurement (mm).

Fluidity test

The fluidity of ACGM was evaluated according to GB/T 2419-2005³⁹. The evaluation was executed with fresh mortar sustaining a constant solid-to-liquid ratio of 0.4 and sand-to-binder ratio of 3. Using a standard flow table device, the newly mixed mortar was placed in two layers into the test mold, and each layer was appropriately compacted. After removing the mold sleeve, the table was shaken 25 times at a frequency of 1 jolt per second. Finally, a ruler was used to assess the fluidity.

Microstructural tests

The paste specimens that had solidified for 28 days were crushed into small fragments measuring approximately 1 cm, which were then immersed in anhydrous ethanol for 48 h to stop hydration, and subsequently the samples were dehydrated in a vacuum drying oven at 40 °C over a period of 48 h. The desiccated samples were preserved in sealed bags for future microscopic analysis.

The following microscopic analysis was conducted. (1) The samples were examined by using a Rigaku Ultima IV X-ray diffractometer, employing a Cu target configured for a scanning range of 5–90° (2θ) at a scanning velocity of 2°/min. (2) The chemical functional groups of ACGM were investigated by using a Bruker Tensor 27 Fourier transform infrared spectrometer. (3) The samples were also subjected to thermogravimetric analysis using a TA TGA55 thermogravimetric analyzer, with a temperature range of 30–800 °C. (4) Nitrogen adsorption analysis of the pore structure of ACGM was performed using a Micromeritics ASAP 2460 automated specific surface area and porosity analyzer (BET). (5) Mercury intrusion porosimetry was performed on ACGM using a Micromeritics AutoPore IV 9510 fully automatic mercury intrusion analyzer. (6) After gold spraying, the microstructure of ACGM was analyzed by using a Zeiss Gemini 300 scanning electron microscope integrated with an energy dispersive X-ray spectrometer (EDS).

Results and discussion

Fluidity analysis

The fluidity test results of ACGM at varying levels of SS are presented in Fig. 6. As shown in Fig. 6, as the concentration of SS increased, the fluidity of the ACGM mortar demonstrated a trend of first increasing and then decreasing compared to the control group C5S0. After adding 10–40% SS, the fluidity of the mortar increased by 13.3–83.3%. This enhancement is attributed to several factors. First, SS has a higher density and lower water absorption than CG⁴⁰. In an equal-mass substitution, this reduces the absolute volume of solid particles, thereby increasing the effective water-to-binder ratio⁴¹. Second, SS particles can fill the gaps within the hydrated matrix, increasing the packing density of the mixture. This releases water originally trapped in voids to lubricate particles, enhancing fluidity⁴².However, when the proportion of SS reached 50%, the fluidity of the mortar dropped by 8.6% relative to the 40% SS proportion. This decrease occurs because SS particles exhibit higher fineness and a larger specific surface area than CG, which reduces the water film thickness around each particle. Additionally, the irregular, blocky morphology of SS lacks a ball-bearing effect. At this high content, these adverse effects on fluidity outweigh the beneficial ones⁴³. Therefore, the content of SS should not exceed 40% in practical applications. Furthermore, since GGBFS and CG are more reactive than SS in early hydration, a lower SS content can lead to a faster initial reaction rate, promoting flocculation and reducing fluidity⁴⁴.

Fig. 6.

Fluidity test results.

Compressive strength analysis

Figure 7 illustrates the effect of varying SS content on the compressive strength of the ACGM mortar. As shown in Fig. 7, the blank control (C5S0) exhibited compressive strength values of 45.1, 49.5, and 53.7 MPa after 7, 14, and 28 days, respectively. With identical SS content, the compressive strength of ACGM gradually increased as the curing duration was prolonged, achieving 7.3–14.3% (10% SS), 7.8–22.2% (20% SS), 11.5–25.7% (30% SS), 15.5–41.7% (40% SS), and − 0.7% to 31.5% (50% SS), respectively. When the SS content reached 40%, the 28-day compressive strength of ACGM reached a maximum value of 76.1 MPa. This strength enhancement can be attributed to two factors. First, SS particles are harder and denser, and unreacted SS can fill pores and reinforce the structure. Second, SS incorporation increases the calcium content, which generates more Ca(OH)₂, elevates alkalinity, and promotes strength development. Nevertheless, when the SS proportion increased to 50%, the compressive strength declined (7-day strength reduced by 0.7% relative to the control group, while the 28-day strength dropped by 7.2% compared to 40% SS). This decline in the compressive strength of C0S5 resulted from the deficiency in alumina and silica ions within the pore solution. When the coal gangue content was equal to zero, the total alumina and silica concentration in solution underwent a marked decline, consequently inhibiting the formation of C-A-S-H or N-A-S-H gels⁴⁵.

Fig. 7.

Compressive strength test results.

.

Hydration heat analysis

According to the experimental results presented in "Fluidity analysis" and "Compressive strength analysis" sections, the effect of SS contents of 0%, 10%, 40%, and 50% on the hydration heat of ACGM was investigated. As illustrated in Fig. 8a,b, the hydration heat development pattern of the ACGM sample was similar to that of OPC. Figure 8a shows a primary peak (at 0.05 h) during the initial induction period, which is associated with the exothermic dissolution of the precursors⁴⁶. As the proportion of SS ranged from 10 to 50%, the primary peak value progressively declined with increasing SS content (25.9–55.9%). This decrease is attributed to the lower early-age reactivity of SS, as its main components (e.g., C₂S, C₃S) generally contain less active SiO₂ and Al₂O₃ compared to GGBFS and CG. Consequently, SS exhibited low activity and released less heat in the early stages⁴⁷. During the initial phase of the induction period, the exothermic rate of hydration slowed down. This was caused by the gel layer generated through hydration during the initial phase of induction, which coated the outer layer of the precursor. This layer slowed down the decomposition of the precursor, resulting in a reduced hydration rate¹⁶. Then, a second exothermic peak emerged between 9 and 24 h, resulting in an acceleration period. However, the peak value was not pronounced, and the arrival time was delayed. This delay can be explained by the fact that the polymerization rate of the initial reaction products surpasses the dissolution rate of SS, reducing the amount of available unreacted material and thus slowing the overall reaction. The curves of C5S0 and C4S1 were similar, with the peak value occurring at 21 h, indicating that 10% SS did not exert a marked influence on the hydration progression of the ACGM paste within the acceleration phase. However, with 50% SS content, the secondary reaction peak was reduced by 8.6%, suggesting that the SiO₂ and Al₂O₃ concentrations in the ACGM system exhibited marked variations reduced by replacing CG with SS, and the reaction activity of the paste could be reduced, resulting in the hydration reaction significantly slowing down.

Fig. 8.

(a) Heat evolution curves and (b) cumulative heat release curves.

The cumulative heat release curve is shown in Fig. 8b. Using C5S0 as the reference sample, the cumulative heat release values of 10%, 40%, and 50% SS over 48 h were reduced by 0.7%, 4.5%, and 11.3%, respectively. During the initial 0–0.5 h period, the cumulative heat release values of 10%, 40%, and 50% SS were lowered by 19.1%, 45.4%, and 53.8%, respectively, relative to the reference sample.

Drying shrinkage analysis

Figure 9 illustrates the drying shrinkage of the ACGM mortar containing varying levels of SS. The results indicated that the drying shrinkage of ACGM developed quickly during the first 7 days, followed by a gradual progression. Relative to the blank control sample C5S0 (3522 με), the 28-day drying shrinkage of the remaining samples increasingly declined with an increase in SS content, with values of 5.7% (10% SS), 9.1% (20% SS), 17.6% (30% SS), 28.9% (40% SS), and 27.6% (50% SS). When the proportion of SS reached 40%, the drying shrinkage of ACGM was reduced by 28.9%, reaching a minimum value of 2501 με, and the improvement in drying shrinkage was the most significant. This was attributed to the following reasons. The addition of SS slowed the initial hydration reaction rate of the mortar, reduced the consumption of free water, maintained the internal moisture of the mortar, and thereby reduced the drying shrinkage⁴⁸. In addition, the hydration of SS facilitated the formation of ettringite, and the microscopic expansion of ettringite crystals occupied the pores, thus minimizing drying shrinkage⁴⁹. Furthermore, f-MgO and f-CaO in SS could generate Ca(OH)₂ and Mg(OH)₂ after reacting with water, causing the volume of SS to expand slightly, thus reducing the drying shrinkage to a certain extent⁵⁰. When the content of SS reached 50%, there was an increase in the drying shrinkage of ACGM, which was possibly attributed to the reduced stiffness of the matrix, especially in the early stages, due to decreased early hydration with an excess of SS.

Fig. 9.

Drying shrinkage test results.

Microstructure of ACGM

DTG/TG analysis

The DTG/ TG analysis results of the pastes subjected to 28-day of curing are shown in Fig. 10a and b. At approximately 100 °C, all samples exhibited a heat peak attributed to the evaporation of free water and the dehydration of ettringite. The second peak appeared at 500 °C, signifying the dehydroxylation process of Ca(OH)₂ within the system. The third thermal peak emerged at 600 °C, denoting the elimination of chemically bound water by the C-A-S-H gel within the system⁵¹. In the temperature range of 600–800 °C, slight mass loss was observed in all groups of specimens, due to the removal of CO₂ by calcite, indicating that the specimens underwent slight carbonization⁵². Notably, the mass loss of C1S4 was the highest between 0 and 200 °C, indicating that the most ettringite was generated in ACGM with 40% SS content. Ettringite formation improves strength, inhibits matrix deformation, and thereby reduces drying shrinkage.

Fig. 10.

(a, b) DTG/TG of ACGM.

XRD analysis

Figure 11 shows the XRD pattern of the ACGM pastes cured for 28 days. In all samples supplemented with SS, the diffraction peaks corresponding to certain crystal phases associated with SS, including CaO, MgO, C₂S, and C₃S, disappeared. The disappearance of the diffraction peaks of crystalline substances in SS indicated that these mineral components were involved in hydration reactions. Notably, as the content of SS increased, the diffraction peak of ettringite was significantly enhanced. This indicated that adding SS to ACGM helped to generate more ettringite. As the content of SS increased, the diffraction peak of SiO₂ exhibited a notable decline, suggesting that the inclusion of SS enhanced the involvement of SiO₂ in the hydration process in the system, leading to a higher yield of hydration products. Moreover, distinct peaks of Ca(OH)₂ and CaCO₃ were detectable, due to the metal cations flowing through the pores to the surface of the specimen in an alkaline environment and undergoing carbonization in air, causing Ca(OH)₂ to transform into CaCO₃. Therefore, both Ca(OH)₂ and CaCO₃ coexisted in the system⁵³. C1S4 and C0S5 also exhibited more extensive amorphous phase regions than C5S0 and C4S1, indicating the generation of more amorphous phase gel products. The increased formation of ettringite and gel products enhanced the strength of C1S4, ultimately contributing to a decrease in drying shrinkage. Moreover, the peak of calcite characteristic in C0S5 was the most prominent, indicating the deepest carbonation experienced by C0S5, which could be the reason for its strength reduction.

Fig. 11.

XRD patterns of ACGM.

FTIR analysis

The FTIR spectra of pastes with varying ratios after 28-day curing are shown in Fig. 12. The spectral band at approximately 460 cm⁻¹ was ascribed to the bending vibrations of the Al-O and Si–O bonds in AlO₄ and SiO₄, while the spectral band at around 620 cm⁻¹ was ascribed to the vibrational mode of the Ca-OH bonds in ettringite. The spectral band at approximately 878 cm⁻¹ was associated with the extensional vibration of Al–OH bonds in ettringite. The strongest vibration spectrum band in each group of specimens was approximately 1000 cm⁻¹, which exceeded that of the C-A-S-H gel (940 cm⁻¹) and was lower than that of the N-A-S-H gel (1050 cm⁻¹). These characteristic peaks were ascribed to the vibrational oscillations of the Si–O–Si and Si–O–Al linkages within the C-(N)-A-S-H gel, indicating that the reaction products of ACGM primarily consisted of aggregates of C-A-S-H and N-A-S-H gels. After incorporating SS, the distinguishing peak exhibited minor lateral displacement, attributed to modifications in Si/Al ratios within the system, leading to variations in Si–O–Al bond density, though the cumulative influence remained limited. The vibrational band proximate to 1116 cm⁻¹ was linked to the asymmetrical extension pattern of the S–O bonding in SiO₄²⁻ within ettringite, while the vibrational band located at 1480 cm⁻¹ indicated the C-O stretching motion of CO₃²⁻ in CaCO₃. The vibrational band adjacent to 1650 cm⁻¹ originated from the angular deformation of the H–O-H linkage in the interstratified moisture of the sample⁵⁴. A gentle peak emerged near 3450 cm⁻¹, corresponding to the interlayer water and water within the ettringite of the specimen, which aligned with the stretching vibration of the O–H bond. The change in the ettringite characteristic peak was particularly noteworthy, indicating that ettringite was indeed generated in the system upon the addition of SS. Combined with the DTG/TG and XRD analyses presented in previous sections, it can be concluded that although the incorporation of SS slowed the reaction rate in the early stages (within the first 24 h), it enhanced the hydration process in the medium and later stages. Specifically, the hydration of SS contributed additional C-S-H gel to the system and reacted with the aluminum from GGBFS to form ettringite. Ettringite formation generally has a positive effect on compressive strength and reduces drying shrinkage. In this study, the characteristic peak of ettringite in C1S4 was the highest, which is consistent with its highest 28-day compressive strength.

Fig. 12.

FTIR spectra of ACGM.

Pore structure analysis

Figure 13 shows the effect of SS on the pore structure. As illustrated in Fig. 13a, as the concentration of SS increased from 10 to 50%, the most developed pore size range in the material increased from 7.9 nm to 9.3, 12.3, and 10.8 nm, and the peak pore size reached its maximum value when the content of SS reached 40%. When the SS content reached 50%, the peak pore size decreased, suggesting that the integration of 50% SS inhibited the expansion of the critical pore size. As shown in Fig. 13b, relative to the control group (C5S0), the inclusion of 10%, 40%, and 50% SS led to a reduction in the cumulative pore volume of ACGM across the 0–200 nm range by 2.4%, 71.9%, and 64.6%, respectively. This suggested that the incorporation of SS could lower the total porosity of ACGM, but the effect of reducing porosity weakened when the content of SS exceeded 40%. As shown in Fig. 13c, with an increase in SS content, the pore volume in the mesopore range (1.25–25 nm) decreased by 11.7% (10% SS), 20.8% (40% SS), and 17.1% (50% SS) and correspondingly increased in the macroporous range (25–200 nm) compared to C5S0. This indicated that the addition of SS could reduce the proportion of mesopore volume in ACGM. This reduction is attributed to the high reactivity of SS during the mid to later hydration stages. The hydration of C₂S, C₃S, and other minerals promotes the formation of C-(N)-A-S-H gel and ettringite. These products interweave, and their crystallization helps inhibit gel deformation and prevents mesopore formation, thereby refining the pore structure⁵⁵. When the content of SS was 50%, the cumulative pore volume increased by 7.8% compared to the 40% SS content, the proportion of pores within the mesopore range increased by 4.7%, and the proportion of pores in the macroporous range consequently dropped. This could be due to the complete replacement of CG by SS and the insufficient content of Al and Si in the system, which weakened the hydration reaction and resulted in a decrease in gel and ettringite formation. In addition, the insufficient reaction of f-MgO and f-CaO in SS could have negatively impacted the pore structure⁵⁶. Overall, adding SS to ACGM generated more gel and ettringite, improved the pore distribution, and reduced the total porosity and drying shrinkage of ACGM.

Fig. 13.

(a) Pore size distribution differential curves of ACGM, (b) cumulative pore volume of ACGM, (c) porosity of ACGM in the mesopores (1.25–25 nm) and macropores (25–200 nm), (d) correlation analysis between drying shrinkage and mesopore volume.

The extent of drying shrinkage in AAMs was markedly affected by the amount and dimensions of capillary pores. As shown in Fig. 13a, the pores of ACGM were concentrated in the mesopore range (1.25–25 nm). Therefore, a correlation analysis was conducted between the drying shrinkage and pore volume in the mesopore range, as illustrated in Fig. 13d. A clear linear relationship was observed, indicating that drying shrinkage increases with mesopore volume, which confirms that mesopores are the primary factor influencing drying shrinkage in ACGM. In the mesoporous range, drying shrinkage increased as the pore volume increased, indicating that the drying shrinkage of ACGM was mainly influenced by the mesopores.

SEM analysis

According to the research results presented in "Fluidity analysis"–"Drying shrinkage analysis" sections, the microstructures of the ACGM samples with SS contents of 0%, 10%, 40%, and 50% after 28 days were analyzed by SEM–EDS. As shown in Fig. 14a, numerous unhydrated CG particles and a small quantity of irregularly shaped gel accumulated on the surfaces of the unhydrated particles. Based on EDS analysis, the dominant elements of the hydration product were Ca, Na, Al, and Si; thus, we inferred that the product was C-(N)-A-S-H gel. Moreover, more pronounced voids were observed, explaining the low strength of C5S0. As shown in Fig. 14b, unhydrated raw material particles and flocculent gels were present on the surface. No obvious unhydrated precursor particles were observed in Fig. 14c, indicating that the hydration degree of 40% SS was more complete than that of 0% and 10% SS. Furthermore, a significant number of ettringite (AFt) crystals formed and filled in the pores. In addition, the interweaving of ettringite crystals and gels inhibits gel deformation and reduces shrinkage⁵⁷, and the formation of CaCO₃ crystals showed that the specimens were carbonized during curing. The hydration products shown in Fig. 14d were essentially the same as those shown in Fig. 14c, which explained why C1S4 and C0S5 had similar results in terms of fluidity, compressive strength, and drying shrinkage, from a microscopic perspective. According to the EDS analysis (Fig. 14c and d), the content of Ca in the hydration products of the specimens with 40% and 50% SS content was higher than that of specimens with 0% and 10% SS content, and the increased calcium-to-silicon ratio enhances the water retention of the matrix, which is conducive to reducing shrinkage⁵⁸.

Fig. 14.

(a) SEM–EDS image of C5S0. (b) SEM–EDS image of C4S1. (c) SEM–EDS image of C1S4. (d) SEM–EDS image of C0S5.

Conclusions

The high shrinkage of AAMs has long been a significant challenge in the field. In this study, the effects of varying SS contents were examined on the fluidity, compressive strength, and drying shrinkage of ACGM, and the micro mechanism of the role of SS in ACGM was evaluated. The following key conclusions were obtained.

1. The partial replacement of CG with SS in ACGM was found to enhance fluidity and compressive strength while reducing drying shrinkage. The optimal improvement was achieved at an SS content of 40%.

2. The Ca(OH)₂ generated from SS hydration increases the initial alkalinity of the system. This, in turn, further activates the silicon and aluminum components in CG, promotes gel formation, and improves the early strength of the material.

3. The synergy between SS and GGBFS helps maintain a high alkalinity level, enabling continuous activation of the silicon and aluminum components in the raw materials. During early hydration, GGBFS hydration produces a significant amount of gels. At later stages, SS hydration contributes to ettringite formation, thereby enhancing strength, restraining matrix deformation, and reducing drying shrinkage.

4. This research primarily discussed the influence of SS on the drying shrinkage of ACGM over a period of 28 days. Subsequent studies will be conducted on chemical shrinkage, self-shrinkage, and long-term performance, and other solid waste materials may be considered for use. In addition, environmental benefits such as lifecycle assessment, energy consumption, and the environmental friendliness of AAMs prepared from solid waste will be investigated.

Author contributions

T.L.Huang and Q.Q.Xie wrote the main manuscript. J.H.Deng and B.Shen reviewed the manuscript.

Funding

Guizhou Provincial Basic Research Program (Natural Science) (No. QianKeHeJiChu-ZK[2024]YiBan094) and the National Natural Science Foundation of China (51068003).

Data availability

Any data, models, or code that underpin the findings of this study can be obtained from the corresponding author (jhdeng@gzu.edu.cn) upon appropriate inquiry.

Declarations

Competing interests

The authors declare no competing interests.