The Performance and Reaction Mechanism of Untreated Steel Slag Used as a Microexpanding Agent in Fly Ash-Based Geopolymers

Abstract

Steel slag is an industrial by-product of the steelmaking process, which is under-utilized and of low value due to its characteristics. Alkali-activated technology offers the possibility of high utilization and increased value of steel slag. A geopolymer composition was composed of steel slag, fly ash, and calcium hydroxide. Four experimental groups utilizing steel slag to substitute fly ash are established based on varying replacement levels: 35%, 40%, 45%, and 50% by mass. The final samples were characterized by compressive strength tests, and Fourier-transform infrared spectroscopy measurements, thermogravimetric measurements, scanning electron microscopy with energy dispersive spectroscopy, X-ray diffraction, and mercury intrusion porosimetry were used to investigate the chemical composition and microstructure of the final products. Higher steel slag/fly ash ratios lead to a lower bulk density and lower compressive strength. The compressive strength ranges from 3.7 MPa to 5.6 MPa, and the bulk density ranges from 0.85 g/cm³ to 1.13 g/cm³. Microstructural and energy-dispersive X-ray spectroscopy analyses show that the final geopolymer products were a type of composite consisting of both calcium aluminate silicate hydrate and sodium aluminate silicate hydrate, with the unreacted crystalline phases acting as fillers.

1. Introduction

Traditional building materials such as ordinary Portland cement (OPC) are potentially harmful to energy consumption and the environment, and this has stimulated the sustainable development of the construction industry [1,2]. The development of sustainable and low-energy geopolymer technology for the reutilization of industrial wastes or by-products has attracted increasing attention in today’s resource-constrained and critical environment [3,4,5,6,7,8,9,10,11,12,13,14]. Geopolymers fabricated in an alkaline solution and aluminosilicate materials fabricated by dissolution, reorientation, polycondensation, and polymerization reactions possess three-dimensional silico-aluminate frameworks where all the shared oxygen atoms are linked with SiO₄ and AlO₄ tetrahedra [15,16]. The presence of alkali cations balances the charge deficit of aluminum ions in four-fold coordination. Generally, geopolymerization consists of three essential steps: (1) the alkali solution dissolves the amorphous phases derived from solid raw materials to produce reactive species such as reactive silica and alumina; (2) the dissolved reactive metal ions form monomers after the reorientation reaction, which includes transportation and the orientation process; and (3) the monomers form three-dimensional aluminosilicate gels during polycondensation and polymerization reactions [17]. The resulting geopolymer with excellent performance (e.g., high early compressive strength and resistance to high temperatures, acids, and freeze–thaw cycles) may be an alternative to OPC in infrastructure construction. In general, industrial waste or by-products consisting of amorphous silica or alumina can be utilized for geopolymer production to relieve the pressure of depleting natural resources [18].

Steel slag (SSL) is an industrial by-product of steel-making, created when molten steel is separated from impurities consisting of liquid oxides containing iron and carbon. SSL contains crystalline compounds, and the predominant minerals include magnetite, tricalcium silicate, and dicalcium silicate, with a handful of free magnesium oxide and calcium oxide owing to the sufficiently low cooling rate of SSL [19,20,21]. The proportions of crystalline compounds in SSL are determined by manufacturing conditions, such as the cooling rate, cooling methods, and steel-making furnaces. The presence of free magnesium oxide and calcium oxide that are not consumed entirely in SSL normally leads to the expansion of SSL under humid conditions. However, the high autogenous shrinkage amount of geopolymers hampers their utilization [22,23,24]. Shrinkage of geopolymers not only reduces their mechanical properties but also their service life. It has been reported that the initial expansion can reduce the shrinkage amount of geopolymers at an early age [25]. The drastic difference between magnesium oxide and calcium oxide is the required time for expansion, as calcium oxide can expand to a larger volume within a short time; however, magnesium oxide needs much more time for expansion than calcium oxide. The expansive properties of SSL limit its application in the fabrication of OPC aggregates, which may cause concrete to crack due to the high expansion degree caused by SSL [20,26]. SSL has always been utilized widely as an aggregate in road construction owing to its poor hydraulic properties, high crushing strength, and fantastic affinity to bitumen.

To explore a new method of applying SSL to facilitate management and reduce the severe environmental risk caused by the occupation of an increasing amount of land from the accumulation of SSL, trials have investigated the incorporation of SSL in geopolymer binders [19,20,21,26,27,28,29,30]. To date, various solid materials, such as high-calcium fly ash (FA), low-calcium FA, and metakaolin, have been used in combination with SSL to generate geopolymers to induce synergistic effects, leading to the influence of complementary advantages because geopolymerization and the chemical composition and microstructure of geopolymers are determined by natural characteristics, such as the amorphous phase and reactivity of the starting raw solid wastes [15,31]. However, the use of SSL in large amounts will cause the development of strength to be unsatisfactory, which limits its further utilization. Previous investigations have demonstrated that calcium hydroxide positively impacts the development of strength and the formation of a denser microstructure in geopolymer binders due to accelerated geopolymerization and the generation of a complex calcium-containing geopolymer [32,33,34,35].

In this paper, the current work aims to utilize untreated steel slag as a microexpanding agent in the production of FA–steel slag-based geopolymers. The goal is to solve the shrinkage problem of geopolymers and improve the utilization of SSL. In this investigation, a constant amount of calcium hydroxide is added to geopolymer paste to enhance the early strength of geopolymers. The understanding of the replacement of FA by SSL is improved to assess the suitability of SSL. The compressive strength and microstructure characteristics are determined by scanning electron microscopy (SEM) with energy-dispersive spectroscopy (EDS), and mercury intrusion porosimetry (MIP), X-ray diffraction, thermogravimetric analysis, and FTIR analysis are also performed to obtain a more thorough understanding.

2. Experimental Program

2.1. Materials

2.1.1. Geopolymer Binder

The primary materials used to fabricate the geopolymer binder, as aluminosilicate sources, were Class F FA (obtained from Gangrun Group, China) and SSL (supplied by Huaiyin Iron and Steel Company, China). The chemical compositions and loss on ignition (LOI) values of the precursor materials, determined by X-ray fluorescence, are exhibited in

Table 1. Chemical composition of FA and SSL.

Oxides (%)	SiO2	CaO	MgO	Fe2O3	Al2O3	SO3	Na2O	K2O	TiO2	LOI
FA	46.36	9.17	5.41	3.27	27.13	0.39	0.47	0.73	-	7.07
SSL	20.94	39.88	6.63	20.74	6.13	0.04	-	-	-	5.64

. The particle size distribution of the cementitious materials determined by dynamic light scattering (DLS) is shown in Figure 1. The particle size of FA is in the range of 0.892 µm to 456.5 µm, with a median grain size (d₅₀) of 36.42 µm, and that of SSL is between 1.261 µm and 95.97 µm, with a d₅₀ of 15.62 µm. White crystalline calcium hydroxide powder (analytical grade, ≥95 wt% purity), including trace impurities such as sediment of ammonium hydroxide, was provided by Damao Chemical Reagent Co., Ltd., China.

The prepared alkaline activator with a modulus ratio of 1.4 was a mixture of sodium silicate solution with an original silicate modulus of 3.3 (7.9 wt% Na₂O, 26.1 wt% SiO₂, and 66% wt% H₂O) and anhydrous sodium hydroxide pellets (AR, ≥95 wt% purity). The produced activator solution was cooled to room temperature before sample preparation.

2.1.2. Mixture Preparation

The characteristics of raw materials, such as the reactivity of source materials and the quantity and shape of unreactive phases derived from raw materials, significantly affect the mechanical properties and microstructure of geopolymer samples. Accordingly, the replacement of FA with SSL was the only variable considered in investigating the influence of SSL content on the geopolymer samples. The detailed program designs are shown in

Table 2. Mix proportions of raw materials.

Specimen	FA (g)	SSL (g)	Alkaline Activator (g)	Water-Binder
FASSL1	260	140	240	50.78%
FASSL2	240	160	240	50.78%
FASSL3	220	180	240	50.78%
FASSL4	200	200	240	50.78%

.

The solid raw materials, including SSL, FA, and calcium hydroxide, in proportion, were mixed for approximately 3 min at a low speed to ensure homogeneous mixing using an NJ-160A mixer. Then, the prepared alkaline solution and deionized water were added gradually for an additional 10 min at a high speed. The fresh slurry was added in three layers to polystyrene molds and compacted using a vibration table. The specimens were cured in an oven at 80 °C for 24 h to achieve a denser microstructure and better properties at an early age. The heated cured samples were demolded and then stored in a curing box (whose temperature was kept at 20 ± 2 °C and relative humidity was kept at 98%) before testing.

2.2. Testing Procedures

2.2.1. Mechanical Properties

The compressive strength of cubic samples with dimensions of 100 mm × 100 mm × 100 mm were cured for 7 days in accordance with GB/T 11969-2008 [36] by using an HG-YH200F-type servo-controlled pressure testing machine. Tests were performed three times to minimize error. Additionally, the surfaces of the selected specimens were polished and smoothed.

2.2.2. Microstructural Characterization

All characterization tests were carried out on specimens cured for 7 days. The test samples were crushed into powder for X-ray diffraction (XRD) analysis, thermogravimetry, and Fourier-transform infrared spectroscopy (FTIR), and small pieces were prepared for mercury instruction porosimetry (MIP) and scanning electron microscopy (SEM). Alcohol was used to prevent the hydration of specimen fractures, followed by drying in a vacuum oven.

XRD experiments were performed with a DX-2700 powder X-ray diffractometer with Cu Kα radiation in the range of 2θ 10–60° with a scan speed of 5° per minute at a step size of 0.02° to determine the crystalline phase composition of solid raw materials and specimens.

FTIR measurements were conducted over the wavelength range of 400–4000 cm⁻¹ with a resolution of 4 cm⁻¹ using a NEXUS 670 spectrometer to investigate the chemical bonds in the reaction samples.

Thermogravimetric measurements were conducted using STA 449C equipment (Germany) to analyze the reaction products. A handful of around 12 mg of powder was stored in a small crucible, followed by heating to 1000 °C in a nitrogen atmosphere at a heating rate of 10 °C per minute. Moreover, the calibration was performed before the TGA instrument.

The pore structures and pore size distribution derived from the fractures of specimens were characterized by mercury intrusion porosimetry instead of nitrogen adsorption, owing to their large pore structure and high porosity. The measurements were conducted on an AutoPore 9500 instrument with a 130° contact angle.

The microstructures and chemical compositions of the final samples were determined by scanning electron microscopy with energy-dispersive X-ray spectroscopy (EDX) using a Nano SEM450 instrument (Hitachi, Japan). The accelerating voltage was kept at 10 kV for SEM and 15 kV for EDX. Both measurements were performed under high-vacuum mode with a constant 5 mm working distance. All experimental programs have been presented in Figure 2.

3. Results and Discussion

3.1. Compressive Strength and Bulk Density

The compressive strength and bulk density of specimens with the replacement of varying SSL contents with FA are illustrated in Figure 3. The variation trend of compressive strength is proportional to that of bulk density, and the values of compressive strength and bulk density decrease with increasing SSL content. After seven days of curing, these curves and histograms exhibit a common feature: the samples with a higher content of SSL show poorer mechanical properties and lower bulk densities. The observed phenomenon is related to the expansion of SSL particles in which the free lime derived from SSL can react with water to fabricate Ca(OH)₂, resulting in a volume increase [20,21]. Therefore, the greater the SSL content is, the higher the expansion volume, leading to a lower bulk density. It is well known that bulk density has a significant favorable effect on the development of compressive strength, and a greater bulk density results in a denser structure that yields better mechanical properties [37]. This is one of the reasons for the changes in bulk density and compressive strength.

Furthermore, the low early reactivity and irregular angular shape of SSL particles, as microaggregates, allow them to react with the filling effect to reduce the shrinkage of geopolymer binders and improve the flexural strength of specimens with their incorporation, as reported in previous studies [38,39].

3.2. XRD Analysis

Figure 4 and Figure 5 compare the XRD patterns of FA, SSL, and four specimens to investigate their relationship. A broad hump in the range from 15° to 30° 2θ in the XRD pattern of FA, with some peaks, shows that the amorphous phases, including amorphous silica and alumina, are present in large quantities. The crystalline phases of mullite and quartz derived from FA are also detected [16,31,40]. It is well known that FA, as a calcined material in which high-temperature calcination is able to convert the crystalline phase to amorphous phases, is dominated by the amorphous phase. The predominant crystalline phase of SSL is mainly the RO phase, which is a solid solution mixed with bivalent metal oxides such as FeO, MgO, and MnO, calcite, tricalcium silicate (3CaO·SiO₂), and dicalcium silicate (2CaO·SiO₂). The diffraction peaks of the SSL crystalline phases are strong, indicating that various crystalline phases in SSL were mixed and overlapped together. The absence of pronounced broad humps in the XRD patterns of SSL demonstrates that SSL did not exist in large quantities.

The XRD results of four hardened samples at seven days with the replacement of different amounts of SSL with FA are presented in Figure 5. All four patterns are similar and show a broad and asymmetric hump from 2θ values of 20° to 35°, along with some detected crystalline phases, such as quartz, mullite, and gismondine. The sharp peaks of the specimens, corresponding to mullite and quartz inherited from FA, indicate that the crystalline phases do not participate in geopolymerization and are only utilized as unreactive fillers in the geopolymer binder [15,31]. Furthermore, the quartz and mullite peaks were less intense than those of the raw material FA, which can be ascribed to the replacement of FA with SSL leading to dilution [38,39]. Comparing these with the XRD pattern of the raw material SSL powder shows that diffraction peaks of the active ingredients of SSL, such as C₃S and C₂S, are absent in the XRD patterns of the samples, demonstrating the involvement of active compounds in the reaction process. The presence of the low crystalline mineral calcite may be induced by the low concentrations being detected. Another reason can be attributed to other side reactions during geopolymerization [16,31,41,42]. In the XRD pattern range of 20–35°, a clear bulging peak can be seen, which is the characteristic peak of the alkali-activated product.

3.3. FTIR Analysis

The different amounts of SSL have a relatively distinct effect on the formed gel structure, as exhibited by the FTIR spectra of specimens on day 7 in the range of 400 to 4000 cm⁻¹ in Figure 6. The bands observed around 3450 cm⁻¹ and 1640 cm⁻¹ are linked to the presence of interlayer-adsorbed water. This occurrence could result from residual water within the samples or a higher water content in the raw materials. These bands correspond to the molecular water’s bending vibrations and the symmetric stretching of O-H bonds, respectively [3,43,44]. The peak shoulders detected near 1500 cm⁻¹ are widely believed to be ascribed to the ν₃ carbonate [CO₃]²⁻ vibration mode. Additionally, the bands at approximately 800 cm⁻¹ are typical of the symmetrical stretching vibration of Si-O (Q1) bonds, which are accepted in the literature as being derived from traces of crystalline-phase quartz inherited from incompletely involved FA particles [3,45]. The broad humps displayed near 580 cm⁻¹ and 460 cm⁻¹ are associated with the presence of mullite and quartz derived from the raw FA material, respectively.

The main characteristic bands of FTIR spectra at 900–1200 cm⁻¹ are attributed to T-O asymmetric stretching and can be used to investigate the difference in the three-dimensional gel structures of alkali-activated materials [46]. More specifically, the main band located near 980 cm⁻¹ is assigned to the Si-O-Si bonds of SiO_n units (n = 2), which is the representative structure of the C-A-S-H-type gel [46]. The investigated reaction products should agree with the SEM–EDS analysis. A distinct difference in this region is that the visible signals shifted to lower numbers. The position changes could be ascribed to the different amounts of FA replaced with SSL, which results in different amounts of Al being substituted by Si in the Si-O-Si bonds. Another reason is that the smaller molecular vibration constants may lead to more T-O ionic bonds, as reported by E. Kränzlein [47].

3.4. MIP Analysis

The nitrogen adsorption method and mercury intrusion porosimetry (MIP) are usually used to investigate the pore structure of specimens from preliminary research; however, the pore size distribution determines which of the two methods should be utilized: a fine pore structure covering the size range of 0.002 to 0.1 µm is studied by the nitrogen adsorption method, and a larger pore structure between 3 nm and 950 µm is studied by MIP [46]. The high-volume expansion SSL is studied with the MIP method to assess its pore size distribution, structure, and porosity. In MIP measurements, mercury is used to fill the pores of the investigated material, followed by applying pressure to determine the pore structure with the Washbum equation. Because pores have different morphologies, the assumed pores are all assumed to be cylindrical in shape, which might lead to an incorrect calculation of the pore size. Hence, MIP is a measurement tool for reference.

The relationships of pore size distribution with cumulative intrusion and log differential intrusion determined by MIP are illustrated in Figure 7 and Figure 8, respectively. On the one hand, the microexpanding specimens with larger amounts of SSL show higher cumulative intrusion, which shows that the pore volume increases with increasing SSL content and indicates that the inclusion of SSL leads to more expansion. This result is consistent with the porosities of the samples shown in

Table 3. The porosity of specimens.

Sample	Porosity (%)
FASSL1	45.16
FASSL2	49.63
FASSL3	52.85
FASSL4	57.49

. On the other hand, multistage pore geopolymer samples are shown in Figure 8. Two threshold pore diameters (also called the critical pore diameters, which define the significant change in the mercury intrusion rate per pore size distribution change) of each sample are shown in the log differential intrusion curves [38,39]. The first critical pore diameter of the small gel pores is approximately 0.01 µm, and the pore distribution shifted to a lower pore size, indicating a denser microstructure formation. The refinement of the pore size distribution may be caused by the uninvolved water in the reaction leaving space. The second threshold pore diameter was approximately 10 µm. The pore volume in this range increased sharply due to the increasing content of free magnesium oxide and calcium oxide achieved with increasing SSL content.

3.5. SEM-EDS Analysis

Figure 9 indicates the microstructure of a fractured surface inherited from specimens along with selected representative EDS curves. Porous and inhomogeneous micromorphological features with a certain amount of microvoids and cracks can be seen on the four surfaces of the samples. The microvoids may be ascribed to the following three reasons: (1) the initial mixing of the slurry led to residual air bubbles entering the sample precursor; (2) evaporated water left voids that were taken up by water; and (3) the free lime derived from SSL reacted with water, resulting in volume expansion [16,31,41,42]. On the other hand, the microcracks may be due to two possibilities, autogenous shrinkage in drying and load-induced cracks, according to the compressive strength test. It is well known that both are detrimental to the development of compressive strength. Thus, researchers have investigated removing air and suppressing bubbles to improve mechanical properties [31].

The unreacted FA microsphere particles and residual shells of FA particles embedded in the matrix are observed. An interesting spot with continuous clusters and a dense matrix without obvious particles is selected to investigate the chemical composition by EDS analysis. This area is characterized to find that the primary elements are Na, Al, Si, and O, which are the primary constituents of geopolymer binders. Hence, the bulky amorphous network base consists of geopolymer binders [16,31]. Furthermore, the added calcium hydroxide also participates in geopolymerization in the fabrication of the C-A-S-H-type gel and the coexisting N-A-S-H-type gel [48,49]. In addition, minor elements such as Ti and Fe, as impurities, are expected in the geopolymer. This is a defect that has a detrimental influence on the development of compressive strength.

The SEM images of the FASSL4 sample are compared with those of FASSL1. The gel microstructure of FASSL1 is denser than that of FASSL4, which is ascribed to the reactivity difference of the raw materials. FA, as a calcined material, can be dissolved and participate easily in the geopolymer process; however, the irregularly shaped low-reactivity SSL particles contain amorphous phases that participate in quick geopolymerization. The partially unreacted irregular particles are observed at high magnification with the EDS curve, and the selected area contains mainly Na, Ca, Si, and Al, which are attributed to the unreacted SSL particles. This indicates that the final geopolymer products are composed of an amorphous geopolymer matrix and neoformed phases as fillers, such as FA and SSL particles. Meanwhile, the geopolymer products overlap with the various types of particles, indicating good bonding with each other. As reported, SSL particles can have a filling effect that enhances the connection between SSL particles and geopolymer binder [38,39]. However, this effect is inferior to the high reactivity of FA in fabricating more amorphous geopolymer gels. In fact, the complex chemical compositions of SSL may react in some side reactions, making it difficult to evaluate the mechanical properties [16].

3.6. Thermogravimetric Analysis

The mass loss of specimens cured for 7 days is detected using TG and differential thermogravimetry (DTG), as shown in Figure 10. The decrease in mass from room temperature to 1000 °C can be categorized into three distinct steps. The loss of mass occurring within the temperature range of room temperature to 105 °C is attributed to the release of physically bound water, such as adsorbed free water or interlayer water within the geopolymer. The sharp mass loss event between 105 °C and 300 °C within the major DTG peaks can be assigned to the decomposition of the aluminosilicate gel reaction products [3,50]. As mentioned above in the SEM analysis, the decomposition of the main reaction products in this range is ascribed to the C-(N)-A-S-H-type gel of geopolymer binders. In addition, cementitious systems in the same range are associated with the C-S-H gel. Then, the negligible decline in the curves along with the minor DTG peaks at approximately 600 °C is attributed to the polymerized/condensed water evaporated from the Si-OH and Al-OH groups [3]. Finally, the slight weight loss between 600 °C and 900 °C is attributed to the decarbonation of carbonated phases [15,40,46,51]. Then, the curves gradually stabilize, which indicates that the mass barely changed.

3.7. Expansion Mechanism

Fluxes, including lime or dolomitic lime, form in furnaces with scraps in the steelmaking process, which is why free lime (f-CaO) is detected in the XRD pattern of SSL. Free lime reacts with water to form Ca(OH)₂, resulting in a volume increase, which is always considered to be the major reason for the volume expansion of specimens. MgO in the form of wüstite in the glassy state derived from SSL is often mixed with a solid solution consisting of FeO and RO phases. The free form of MgO is usually formed under low-alkali conditions owing to its volumetric instability. Due to the similar radii of Mg⁺⁺, MgO, FeO, and Fe⁺⁺, they can form solid solutions in a high-basicity environment. This is another reason for expansion [52].

From a different perspective, both chemical and physical volume changes from the hydration of free lime are considered. As mentioned above for the chemical reaction, the reaction equation of f-CaO and water is as follows:

$$\mathrm{C}\mathrm{a}\mathrm{O}+{\mathrm{H}}_2\mathrm{O}\to \mathrm{C}\mathrm{a}{\left(\mathrm{O}\mathrm{H}\right)}_2\pm 15.5\mathrm{C}\mathrm{a}\mathrm{l}/\mathrm{M}\mathrm{o}\mathrm{l}$$

At the appropriate temperature, the reaction can proceed normally; however, at temperatures above 547 °C, the reaction proceeds in the opposite direction.

The volume expansion owing to physical changes can be attributed to the increase in void volume in conjunction with the increased solid phase content due to the hydration of free lime. As shown in Figure 11, the particles of free lime are assumed to be spherical before and after hydration. The relative contents (the solid-to-void ratio) are not associated with the particle size and remain constant. Then, the void volume changes with the solid phase. When the water molecules absorb free lime during the hydration process, the volume of the solid phase increases. This will lead to an absolute volume increase, although the solid-to-void ratio is constant. Then, the void volume will increase with the solid phase. The combination of increased solid phase and void contents will result in the volume expansion of the samples [52]. This can explain why increasing SSL content results in an increasing porosity owing to the increased void volume in the range of the increasing volume of the solid phase, which is consistent with Figure 8.

The high-doping SSL has more free calcium oxide, which reacts with water to produce a large volume of calcium hydroxide. At the same time, magnesium oxide formed at low alkalinity will be converted to a solid solution at high alkalinity, which is the reason for the increase in volume. And the increase in volume makes the sample swell, porosity increase, and dry density decrease. The low active SSL and the decrease in dry density do not counteract the refinement of the pore size by SSL, so the compressive strength decreases with an increase in SSL content.

The theory of crystallization pressure as the primary expansion mechanism is also considered. Lavalle showed that a certain weight can be lifted owing to crystals growing from solutions. An energy mismatch in the range of the surface of the growing crystal and the walls owing to the existence of electrostatic and van der Waals forces during crystallization exists in porous media. A thin film between the crystal and the pore wall transfers ions in crystal growth because there is no direct contact between them. To simplify various coarse pores, three main pore models are described in Figure 12, Figure 13 and Figure 14 [53].

Pore Model 1: There is a spherical crystal in the cylindrical pore along with a supersaturated solution, and then, the crystal grows cylindrically with hemispherical sides. Crystal growth will reach equilibrium, and the curvature of the hemispherical side is derived from the crystal. Afterward, the solubility of the tips of the crystal is higher than that of the cylindrical portion, leading to an increase in the chemical potential from the cylindrical part until equilibrium is reached, indicating that crystal growth ends.

Pore Model 2: Crystal growth takes place within smaller cylindrical pore openings located in a spherical pore. The solution concentration at the low-curved portion of the crystal equals that at the small pore entrances of the developing crystal. As the spherical crystal grows, it elevates the pressure until the chemical potentials align with those of the highly curved pores within the crystal.

Pore Model 3: It has the same mechanism as Pore Model 2, and the only difference is the large portion of the pore; the pore in Model 2 is spherical, and the pore in Model 3 is cylindrical.

In brief, the size of the pore entry determines the generated maximum pressure. Cylindrical pores can generate lower pressure, and spherical and cylindrical pores with small pore entries can achieve higher pressure.

Various porous media can fill the porous pore in crystal growths, resulting in the formation of smaller pores, which is in accordance with Figure 9.

4. Conclusions

This paper presents an experimental study aimed at transforming fly ash and steel slag into an environmentally friendly construction material for green building applications through geopolymerization. This study focused on a synthesis parameter, the FA/SSL ratio, to assess its impact on the resulting geopolymer binder products. The final products, cured for 7 days, underwent analysis using X-ray diffraction, FTIR measurement, thermogravimetric measurement, scanning electron microscopy with energy-dispersive X-ray spectroscopy, mercury intrusion porosimetry, and mechanical testing. Drawing from the experimental findings, the following conclusions can be derived:

(1)

The untreated SSL was successfully utilized in fly ash–steel slag geopolymer, and the final geopolymer products were a type of composite containing C-A-S-H, N-A-S-H, and unreacted crystalline phases as fillers.

(2)

The low dosage of SSL reduces the compressive strength in the geopolymer reaction, which is due to its lower reactivity than FA. On the contrary, the high dosage of SSL makes the samples micro-expanded, with a decrease in the dry density and refinement of the pore size.

(3)

The expansion of samples was caused by chemical and physical changes in free lime and crystal growth in porous media.

The utilization of untreated steel slag in fly ash-based geopolymers significantly influences environmental protection and economic development in the construction and energy industries. Specifically, when employing substantial amounts of SSL, it can be transformed into an alkali-activated foam material, thereby serving the objectives of environmental conservation and energy efficiency.

At present, foam materials are often used in building insulation, fire prevention, sound insulation, and other fields. The required performance mainly includes high porosity, low apparent density, high strength, low thermal conductivity, etc. Organic thermal insulation materials have low density and good thermal insulation performance but poor fire resistance; they are not fireproof and are easy to decompose when heated. With the vast majority of inorganic thermal insulation materials, although fire resistance is better, there are problems such as poor thermal insulation, high production energy consumption, and so on. Compared with traditional cement-based foamed materials, foamed geopolymer materials tend to have lower thermal conductivity and higher strength. Subsequent research on FA-SSL-based geopolymers is poised to move toward alkali-activated foam materials.