Recycling Spent Fluorescent Lamp Glass Waste in Calcium Aluminate Cement: Effects on Hydration and Mechanical Performance

Featured Application

This study presents a novel application of spent fluorescent lamp glass (SFLG) waste as a partial replacement for calcium aluminate cement (CAC), offering a sustainable solution for hazardous waste reuse and reducing CAC’s environmental impact. Mortars with 25–50 wt.% SFLG showed enhanced long-term strength due to strätlingite formation, which limited harmful phase conversion. Residual mercury in SFLG was safely encapsulated, ensuring environmentally safe waste management.

Abstract

Calcium aluminate cement (CAC) offers rapid strength development, chemical durability in harsh environments, and high-temperature resistance, but its long-term performance may be compromised by the conversion of metastable hexagonal hydrates into stable cubic phases. Concurrently, recycling spent fluorescent lamp glass (SFLG) is limited because of its residual mercury content. This study investigates the use of manually (MAN) and mechanically (MEC) processed SFLG as partial CAC replacements (up to 50 wt.%). Both SFLG types had irregular morphologies with mean particle sizes of ~20 µm and mercury concentrations of 3140 ± 61 ppb (MAN) and 2133 ± 119 ppb (MEC). Moreover, the addition of SFLG reduced the initial and final setting times, whilst MEC waste notably extended the plastic state duration from 20 min (reference) to 69 min (50 wt.% MEC). Furthermore, strength development was accelerated, with SFLG/CAC mortars reaching peak strengths at 7–10 days versus 28 days as in the CAC reference. CAC and 15 wt.% SFLG mortars showed strength loss over time by reason of their phase conversion, whereas mortars with 25–50 wt.% SFLG experienced significant long-term strength gains, reaching ~60 MPa (25 wt.%) and ~45 MPa (35 wt.%), respectively, after 365 days, with strength activity indexes (SAI) near 90% and 70%, respectively. These improvements are attributed to the formation of strätlingite (C₂ASH₈), which stabilized hexagonal CAH₁₀ and mitigated conversion to cubic katoite (C₃AH₆). Mercury leaching remained below 0.01 mg/kg dry matter for all mixes and curing ages, classifying the mortars as non-hazardous and inert under Spanish Royal Decree 646/2020. The results suggest that SFLG can be safely reused as a sustainable admixture in CAC systems, enhancing long-term mechanical performance while minimizing environmental impact.

1. Introduction

Calcium aluminate cement (CAC) is a hydraulic binder known for its rapid strength development, high durability in sulphate-rich environments, excellent resistance to high temperatures (exceeding 1600 °C), and strong resistance to chemical attack and abrasion [1,2,3]. However, CAC hydration is highly sensitive to any given time–temperature conditions. At ambient temperature, the primary hydration products formed are hexagonal calcium aluminate decahydrate (CAH₁₀) and octahydrate (C₂AH₈), along with aluminum hydroxide (AH₃, Al(OH)₃), as shown in Equations (1) and (2), where C = CaO, S = SiO₂, A = Al₂O₃, and H = H₂O [1,3,4]. C₂AH₈ may form either directly from the hydration of CA (CaAl₂O₄) or via transformation from CAH₁₀ (Equation (3)).

CA + 10H → CAH₁₀

(1)

2CA + 11H → C₂AH₈ + AH₃

(2)

2CAH₁₀ + 9H → C₂AH₈ + AH₃

(3)

Although CAC can be used in prefabricated components made of mass concrete or non-structural reinforced concrete, its application in structural elements was prohibited due to the potential transformation of the initial hydration products under specific temperature and humidity conditions. This conversion process results in a more porous structure, resulting in mechanical strength reduction [2,5]. As extensively demonstrated in the literature [1,3,6,7,8,9,10,11], the hexagonal hydrates CAH₁₀ and C₂AH₈ are metastable at room temperature and gradually convert into the thermodynamically stable phases AH₃ and C₃AH₆ (cubic, also known as hydrogarnet) over time.

Conversion occurs according to Equations (4) and (5) [1,2,7], releasing AH₃ gel and water during the reaction. This process occurs in the hardened concrete and strongly depends on temperature, humidity, and time. The conversion of metastable hydrates into the stable cubic phase increases the material’s porosity and permeability, resulting in reduced mechanical strength and diminished protection of embedded steel reinforcement, which becomes significantly more susceptible to environmental degradation [1,7,8,10].

3CAH₁₀ → C₃AH₆ + 2AH₃ + 18H

(4)

3C₂AH₈ → 2C₃AH₆ + AH₃ + 9H

(5)

To prevent or mitigate conversion and its detrimental effects on the binder, previous studies partially replaced CAC with silica-rich additions to stabilize CAC hydrates [6,7,9,11,12,13,14]. For example, Pacewska et al. [12] investigated the partial replacement of CAC with fluid catalytic cracking (FCC) catalyst residue, while Hidalgo et al. [7] explored the use of silica fume and fly ash as supplementary materials. In both studies, silicate ions reacted with hexagonal calcium aluminate hydrates (CAH₁₀ and C₂AH₈), promoting an alternative reaction pathway that led to the formation of strätlingite (C₂ASH₈, also known as hydrated gehlenite), a stable phase at room temperature. According to Pacewska et al. [12], the extent of strätlingite formation strongly depends on the silica availability and its release rate from the additive. Moreover, Hidalgo et al. [7] further noted that when using mineral admixtures such as FA, high CAC replacement levels (>40%) are required to observe any impact on C₂AH₈ formation and its subsequent conversion to C₃AH₆. Additionally, other studies, such as that of Son et al. [2], have incorporated sulphates in order to stabilize the hexagonal phases through the formation of ettringite and calcium monosulfoaluminate [2].

Discharge and fluorescent lamps emit light through the use of chemical elements such as mercury and phosphor powder. On the strength of the high toxicity of mercury, European Directive 2011/65/EU [15] imposes restrictions on its use in electrical and electronic equipment (EEE), establishing maximum allowable mercury concentrations based on lamp type and power. Additionally, European Directive 2012/19/EU [16], transposed into Spanish law via Royal Decree 110/2015 [17], governs the management of waste electrical and electronic equipment (WEEE), including mercury-containing discharge and fluorescent lamps. This legislation requires the separate collection and appropriate treatment of end-of-life EEE, requiring that at least 80% of such lamps be recycled.

As noted in previous studies [18,19], some mercury in discharge and fluorescent lamps is responsible for generating visible light, and the rest interacts with the lamp’s glass and phosphor powder. During processing at lamp recycling facilities, the glass is separated for recovery, and the mercury-laden phosphor dust is collected for potential reuse in the manufacturing of new lamps. Rey-Raap and Gallardo [18,19] found that spent fluorescent lamps (SFLs) contain an average mercury concentration of 2.37 ± 0.50 mg/g. Of this mercury, approximately 85.76% is retained in the phosphor powder, 13.66% is diffused into the glass, and 0.58% vaporized. Consequently, the glass waste often contains mercury from both diffusion and residual dust and is classified as a hazardous waste under Spanish law MAM/304/2002 [20].

Ambilamp, a Spanish non-profit association (www.ambilamp.es) [21], reported collecting 2161 tons of lamps in 2023. About 86% of these (approximately 1858.5 tons) were non-light-emitting diode lamps that contained mercury. In 2019, around 5475 tons of lamps were placed on the Spanish market [22], and an estimated 4708.5 tons (86%) contained mercury. Since glass constitutes roughly 85% of a lamp’s weight [18,23], this corresponds to around 4000 tons of mercury-contaminated glass that will eventually be generated when these lamps reach their end of life.

The presence of mercury in spent fluorescent lamp glass (SFLG) significantly restricts its reuse, particularly in high-temperature applications where mercury could volatilize. By reason of the high cost and technical challenges associated with reducing the concentration of mercury in SFLG, its reuse is generally limited, often resulting in its disposal in hazardous waste landfills. Despite advances in techniques to reduce mercury levels in SFLG waste, completely removing the phosphor powder, which contains approximately 85.76% of the mercury, remains a major challenge.

Numerous recent studies have investigated the incorporation of glass powder, primarily derived from domestic glass containers, into a variety of cement-based construction materials. These studies highlight its effectiveness in applications such as concrete aggregates, partial replacement of Portland cement (PC), recycled aggregate in asphalt, and the production of clay bricks, lightweight aggregates, and geopolymers [23,24,25,26,27,28,29].

To date, however, no research has explored the use of SFLG waste as a partial replacement for CAC, with the dual aim of enhancing structural stability and mitigating the conversion effects characteristic of CAC. This study addresses this research gap by proposing a novel approach that both promotes the functional reuse of SFLG waste and contributes to the development of more durable and sustainable CAC-based materials.

In this work, two types of untreated SFLG waste were used to partially replace CAC. The substitution pursued three objectives: (i) reducing hydrate conversion in CAC, (ii) promoting the reuse and valorization of hazardous waste, and (iii) lowering the environmental impact of CAC production. The SFLG waste was used in its raw form, without any pre-treatment to reduce its mercury content. To ensure environmental safety, leaching tests were performed to assess whether the mercury in the SFLG waste was effectively encapsulated within the cementitious matrix, thereby enabling the final material to be classified as inert and non-hazardous upon disposal.

2. Experimental Process

2.1. Materials, SLFG Waste Preparation, and Characterization

In this study, 0 to 50 wt.% CAC was replaced with two types of SFLG waste, processed manually (MAN) and mechanically (MEC). The CAC was supplied by Cementos Molins Industrial S.A. and complied with UNE-EN 14,647 specifications [30]. The two types of glass waste were provided by VAERSA, an environmental management company based in the Valencian region of Spain, and were derived from end-of-life fluorescent lamp processing using two distinct separation techniques. Figure 1 shows spent fluorescent lamps prior to manual (MAN) and mechanical (MEC) processing. On the one hand, the MAN waste was obtained through manual dismantling, which involved the hand-operated separation of lamp components. On the other hand, the MEC waste was produced by thermally cutting the ends of the lamps using a flame, followed by the recovery of the phosphor powder, originally applied to facilitate light emission, using air-blowing techniques. Subsequently, the glass tubes were broken under negative pressure, and the remaining components (glass, plastic, and metal) were separated using electromagnetic and density-based methods. These differing processing methods, along with variations in the original lamp characteristics, are expected to result in SFLG wastes with distinct properties and potential contaminant profiles, which may influence their behavior in cementitious matrices.

As previously reported [31], the reactivity of the SFLG waste was enhanced by reducing its particle size through crushing in a jaw crusher BB200 (manufactured by Retsch, Haan, Germany) and subsequent milling. In this regard, milling conditions had been previously optimized, and SFLG particles were milled for 8 h in an Orto-Alresa ball mill (Álvarez Redondo S.A., Madrid, Spain) using alumina media. Their particle size distribution was analyzed using a Mastersizer 2000 (Malvern Instruments, Malvern, UK). The morphology of the milled particles was examined by way of scanning electron microscopy (SEM) using a JEOL JSM-7001F (JEOL Ltd., Tokyo, Japan) equipped with EDX-WDX. Their chemical composition was assessed using X-ray fluorescence (XRF) with a Bruker S4 Pioneer spectrometer (Bruker AXS GmbH, Karlsruhe, Germany), and the mineralogical phases were identified through X-ray diffraction (XRD) with a Bruker AXS D4 Endeavor (Bruker AXS GmbH, Karlsruhe, Germany), from 5° to 70° 2θ, using Cu Kα radiation at 20 mA and 40 kV. The mercury concentration in the SFLG waste particles was measured using a LECO AMA-254 total mercury analyzer (LECO Corporation, St. Joseph, MI, USA), which allows for the direct analysis of mercury in both liquid and solid samples without the need for sample pretreatment. This analyzer provides quantitative measurements across a wide concentration range, from parts per billion (ppb) to parts per million (ppm).

2.2. Preparation of Pastes and Mortars

Pastes and mortars were prepared using both types of SFLG waste powder, according to the conditions outlined in

Table 1. Sample designations, SFLG waste percentages, and curing conditions.

SFLG Waste Type	Designation	Binder/Sand/Water Ratio	CAC Replacement wt.%	Curing T °C	Curing Age Days
-	CAC	1:3:0.5	0	20	Up to 365
Manual	MAN15		15
	MAN25		25
	MAN35		35
	MAN50		50
Mechanical	MEC15		15
	MEC25		25
	MEC35		35
	MEC50		50

. The samples were designed based on the type and amount of SFLG waste used, such that MAN35 refers to a sample in which 35 wt.% of CAC was replaced with manually processed SLFG. A reference sample, with no CAC substitution, was also prepared as a reference (denoted as CAC). Moreover, the microstructure was examined in pastes, and the compressive strength evolution was assessed in mortars, which were prepared according to UNE EN 196-1:2005 specifications [32]. Additionally, the mortar samples were made using siliceous sand with a maximum particle size of 2 mm and a fineness modulus of 2.74. All samples were cured in a temperature-and humidity-controlled chamber at 20 °C and 95% relative humidity until the testing age (up to 365 days).

2.3. Paste and Mortar Characterization

Table 2. Tests conducted to characterize the SFLG/CAC blended binders and curing ages at which they were performed.

SFLG Waste	Setting Time	TG	XRD	SEM	Compressive Strength	Lixiviation
MAN	Performed	3, 7, 28, 90, 365	3, 7, 28, 90, 365	28	1, 3, 7, 10, 28, 90, 365	3, 7, 28, 90
MEC	Performed	3, 7, 28, 90, 365	3, 7, 28, 90, 365	28	1, 3, 7, 10, 28, 90, 365	3, 7, 28, 90

summarizes the tests performed to characterize the pastes and mortars developed with each type of SFLG waste and varying percentages of CAC substitution. The setting time was determined in pastes, in accordance with UNE EN 196-3:2005 specifications [33]. The microstructural evolution with the curing time, SFLG waste type, and content was evaluated in pastes via microstructural analyses, including thermogravimetry (TG), XRD, and SEM. TG analyses were conducted using a TGA SDTA851e/LF/1600 thermobalance (Mettler-Toledo GmbH, Greifensee, Switzerland), with sealed aluminum crucibles with manually perforated lids, under a nitrogen atmosphere, from 35 to 600 °C, at a heating rate of 10 °C/min. SEM and XRD analyses were performed following the equipment and conditions outlined in Section 2.1.

Lixiviation tests were conducted on mortar samples following the UNE EN 12457-4:2003 standard [34]. This procedure is intended for testing the leaching behavior of granular waste materials with particle sizes under 10 mm, using a liquid-to-solid ratio of 10 L/kg and an agitation period of 24 h. The resulting leachate was filtered through a 45 µm filter, and mercury concentrations were subsequently determined using a Total Mercury Analyzer. The instrument analyzes liquid or solid samples of 50 µg, with a detection limit of 0.01 ppb. Six replicates were performed for the analysis of mercury in the SFLG waste powder, whereas three replicates were conducted for the lixiviation tests.

Compressive strength was determined for each curing age, type of SFLG waste, and CAC percentage of substitution as the average of six tested specimens. The tests were conducted following the UNE EN 196-1:2005 specifications [32] in an MEH-3000 PT/W testing machine (manufactured by S.A.E. Ibertest, Madrid, Spain). The strength activity index (SAI) was calculated to evaluate the relative compressive strength of the SFLG/CAC mortars with respect to the reference 100 wt.% CAC mor.

3. Results and Discussion

3.1. Raw Material Characterization

Figure 2 presents SEM micrographs of the MAN and MEC SFLG waste powders. Both powders exhibit irregular morphologies characterized by smooth surfaces and sharp edges. As previously reported in [31], the mean particle diameter of the milled powders was approximately 20 µm, with 90 vol.% of the particles smaller than 60 µm and 10 vol.% smaller than 2 µm.

Table 3. Chemical composition of the SFLG waste, manually and mechanically processed, wt.%.

	MAN	MEC
SiO2	75.7	74.2
Na2O	7.6	12.2
CaO	2.9	4.4
Al2O3	5.2	3.1
MgO	1.0	1.9
K2O	2.6	1.8
BaO	2.4	1.3
Other	1.92	0.86
LOI *	0.65	0.36

* Loss on ignition at 1000 °C.

presents the chemical composition of the MAN and MEC SFLG waste materials. Both types of glass waste were predominantly composed of SiO₂ (~75%) and exhibited some relatively high Na₂O contents (7.6% for MAN and 12.2% for MEC). This composition aligns with typical soda–lime glass formulations, such as those used in containers and light bulbs, which, according to Mohajerani et al. [25], generally contain 66–75% SiO₂ and 12–17% Na₂O. Moreover, the XRD spectra of the MAN and MEC SFLG waste, previously reported in [31], confirmed that both MAN and MEC SFLG samples exhibited a primarily amorphous structure, suggesting potential reactivity. As such, minor crystalline phases identified in the diffractograms, such as vaterite (CaCO₃), hydrotalcite (Mg₆Al₂CO₃(OH)₁₆·4(H₂O), and natron (Na₂CO₃·10H₂O), are likely the result of carbonation processes occurring during handling and storage.

The mercury concentration in the SFLG MAN powder was 3140 ± 61 ppb, whereas the MEC powder exhibited a lower concentration of 2133 ± 119 ppb. The lower mercury content in the MEC sample was attributed to the glass cleaning machinery employed, which effectively removed a larger quantity of phosphor powder.

3.2. Setting Time

Variations in the initial (IST) and final setting times (FST) with the different types and amounts of SFLG waste are reported in Figure 3. Both the IST and FST decreased as the quantity of SFLG waste increased, whatever the type of SFLG used, MAN or MEC. For a given amount SFLG waste, the reduction was generally more pronounced with MEC waste compared to MAN waste. Additionally, the pastes made with the MAN waste exhibited a plastic state duration (FST-IST) similar to that of the CAC sample. In contrast, the plastic state duration for pastes made with MEC waste was extended, reaching up to 69 min, compared to the 20 min for the CAC control paste. The results were consistent with prior research by Nowacka and Pacewska [14], who had investigated the effects of two different aluminosilicates on the early hydration of CAC. Their study found that both types of aluminosilicates accelerated the setting and hardening of CAC at room temperature. Sengül and Erdogan [4] also observed that replacing 25% to 75% of CAC with ground perlite generally shortened the initial and final setting times. However, when SFLG waste was used to partially replace Portland cement (CEM I 42.5R), the evolution of the setting time was different, since, regardless of the SFLG waste type used (MAN or MEC), the IST generally remained the same or was extended with 25 wt.% to 50 wt.% MEC, while the FST was usually prolonged. Hence, this resulted in an overall extension of the plastic state, reaching up to 70 min compared to the 45 min of the reference 100 wt.% PC sample.

3.3. Compressive Strength of SFLG/CAC Mortars

Figure 4 and Figure 5 present the compressive strength results (MPa) and the strength activity indexes (SAI, %) for mortars incorporating 0 to 50 wt.% of MAN and MEC SFLG waste. Figure 4 shows the mean compressive strength values derived from six mortar specimens for each data point (defined by curing age, SFLG waste type, and CAC replacement percentage), together with the corresponding standard deviations. Both MAN and MEC SFLG types exhibited similar performances at equivalent CAC substitution levels and curing ages, indicating consistent mechanical behavior regardless of the waste type. Also, a rapid increase in strength was observed, with less variation in strength with the curing age when increasing the amount of SFLG. Although all mixes exhibited a peak strength followed by a decline, the maximum strength was achieved earlier with higher amounts of SFLG waste. Thus, the strength of the reference CAC sample progressively increased up to 28 curing days, while the blended mortars reached their maximum strength after 7–10 curing days. This observation aligns with previous studies conducted by Pacewska et al. [6] and Konáková et al. [3], indicating that the addition of SFLG waste accelerates the hydration process. Furthermore, after 365 curing days only the reference 100 wt.% CAC mortar and those blended with 15 wt.% SFLG maintained the strength development trends observed up to 90 days. In contrast, mortars containing 25 wt.% to 50 wt.% SFLG showed significant strength gains, achieving nearly 60 MPa with 25 wt.% SFLG and 45 MPa with 35 wt.% SFLG. As detailed in Section 3.5 and Section 3.6, the evolution of compressive strength is consistent with the microstructural changes observed over the curing time, which are influenced by the amount of SFLG waste incorporated.

The SAI, defined as the relative strength between the SFLG/CAC and the reference CAC mortars, is reported in Figure 5. The blended mortars consistently showed lower compressive strength than the 100 wt.% CAC mortar, with the strength loss generally comparable to or slightly exceeding the corresponding replacement level. This indicates that the SFLG waste did not significantly contribute to strength development, and the reduction was primarily attributed to the dilution of CAC in the binder matrix. Up to 90 days, the SAI typically peaked at 3 curing days and gradually declined thereafter. This decline became more pronounced with increasing CAC substitution, so after 90 curing days, strength reductions approached 20% in mortars with 15 wt.% SFLG (5% above the substitution level) and nearly 70% in mortars with 50 wt.% SFLG (≈20% above the substitution level). However, the trend changed with longer curing periods. Among the SFLG/CAC mortars cured for 365 days, only those containing 15 wt.% SFLG followed the pattern observed up to 90 curing days, showing a slight reduction in SAI. In contrast, mortars with 25–50 wt.% SFLG demonstrated significant strength gains over time, reaching their highest SAI values after 365 days. Notably, mortars containing 25 wt.% and 35 wt.% SFLG recorded SAI values exceeding 90% and 70%, respectively. This behavior is attributed to the strength loss experienced by the CAC reference sample and the strength gained in the blended SFLG/CAC mortars. As highlighted by Pacewska et al. [12], the strength values used for structural design should be based on long-term results, obtained after the completion of the conversion process.

The obtained results are consistent with those previously published by Pacewska et al. [12] in FCC/CAC blended systems, who also observed that at early curing ages, the strength of CAC samples containing 25 wt.% FCC (20 MPa after 1 curing day) was significantly lower compared to the reference CAC sample (close to 47 MPa after 1 curing day). The authors [12] also observed that both the reference CAC sample and that containing 5% FCC experienced a decrease in strength after the conversion period, whereas the strength of the sample with 25% FCC showed a slight improvement.

As extensively documented by the scientific community [1,6,7,8,9,10,12,35], the strength reduction observed in the CAC reference sample and that containing 15 wt.% SFLG waste after extended curing periods can be attributed to the conversion of hexagonal into cubic hydrates, typically observed in CAC cements. Therefore, the microstructural analyses performed, reported in Section 3.5 to Section 3.7, will be essential to confirm this transformation and assess whether the addition of SFLG waste influences the conversion process.

3.4. Lixiviation Results

The leaching tests performed aim to determine if the mercury present in the SFLG waste is effectively encapsulated within the binding matrix. This is an essential requirement for classifying the SFLG/CAC blended cements as inert materials at the end of their useful life. The results of the leaching tests, performed according to the UNE-EN 12457-4:2003 standard [34], are presented in

Table 4. Lixiviation of the SFLG/CAC blended mortars, developed with the MEC and MAN SFLG waste, cured at 20 °C.

		MAN				MEC
		15	25	35	50	15	25	35	50
Hg concentration in blended mortars, mg/kg		0.1047	0.1745	0.2443	0.3490	0.0711	0.1185	0.1660	0.2371
3 days	Lixiviation, mg/kg dry matter	0.0011	0.0009	0.0013	0.0008	0.0012	0.0009	0.0009	0.0009
	VC, %	12.86	4.71	4.99	6.56	7.25	4.61	10.15	2.34
7 days	Lixiviation, mg/kg dry matter	0.0005	0.0007	0.0007	0.0008	0.0006	0.0006	0.0006	0.0006
	VC, %	13.40	6.15	3.82	14.32	10.69	12.85	17.24	5.94
28 days	Lixiviation, mg/kg dry matter	0.0006	0.0005	0.0003	0.0005	0.0017	0.0019	0.0026	0.0043
	VC, %	1.23	21.06	6.24	18.20	18.77	9.52	7.67	10.50
90 days	Lixiviation, mg/kg dry matter	0.0004	0.0007	0.0007	0.0008	0.0006	0.0007	0.0011	0.0011
	VC, %	7.07	23.93	22.43	9.18	14.89	6.64	10.57	2.26

and Figure 6. These tests were conducted in compliance with Spanish Royal Decree 646/2020 [36], which defines leaching limits for classifying waste as inert, non-hazardous, or hazardous. For mercury, the threshold values are 0.01, 0.2, and 2 mg/kg of dry matter, respectively.

Leaching results for the SFLG/CAC blended mortars cured for up to 90 days, developed with the MAN and MEC SFLG waste, indicate minimal mercury leaching, with all values falling below the regulatory threshold of 0.01 mg/kg of dry matter. As a result, this blended cement can be classified as inert waste at the end of its useful life. The percentage of leached mercury in the MEC mortars ranges from 0.25% to 1.81%, demonstrating very low levels of leachability. The variation coefficient (VC, %) ranges from 2.26% to 18.77% (with a standard deviation (SD) between 2 × 10⁻⁵ and 4.5 × 10⁻⁵ mg/kg dry matter). Additionally, no clear correlation was observed between the increasing SFLG content in the blended mortars and mercury leaching, nor between curing age and mercury leachability. Despite the higher mercury concentration in the MAN SFLG/CAC blends, no significant differences in leaching behavior were noted between the two types of mortars. The MAN SFLG/CAC blended mortars exhibited similar leaching behavior, with levels ranging from 0.12% to 1.05%. The VC for these mortars varied from 1.23% to 23.93% (with an SD in the range of 1 × 10⁻⁵ to 1.7 × 10⁻⁵ mg/kg dry matter). The high variance in some samples could result from several factors: the small volume of sample analyzed (50 µg), insufficient liquid homogeneity before analysis, the low mercury concentration (near the 0.01 ppb detection limit), and the limited number of replicates (three). Nevertheless, the highest concentrations remain well below the 0.01 mg/kg dry matter leaching limit set by Spanish Royal Decree 646/2020 for inert waste classification [36].

The results suggest that these mortars are unlikely to present environmental or health risks during their service life. Once they reach the end of their service life, they can be classified as inert (non-hazardous) waste, which offers advantages for both recycling and landfilling. Furthermore, the leaching levels remained very low, even at early curing ages (3-day-cured mortars), emphasizing the safety of the material for handling by construction workers at the end of its lifecycle.

3.5. X-Ray Diffraction (XRD) Studies

The diffractograms of SFLG/CAC pastes containing up to 50 wt.% SFLG waste, cured at 20 °C for up to 365 days, are presented in Figure 7 (MAN up to 90 days), Figure 8 (MEC up to 90 days), and Figure 9 (365 curing days, MAC and MEC). The diffraction patterns of the milled SFLG waste, also included as a reference, denoted a predominantly amorphous structure, with minor crystalline phases identified as hydrotalcite (T, Mg₆Al₂CO₃(OH)₁₆·4(H₂O), PDFcard 14-1191) and natron (R, Na₂CO₃·10H₂O, PDFcard 15-800), likely resulting from slight carbonation during handling and storage of the sample. In consonance, the SFLG/CAC blended pastes exhibited increasing baseline deviations with higher SLFG content. The XRD patterns were similar regardless of the waste type (MAN or MEC), with the same crystalline phases present; however, pastes containing MAN waste typically showed slightly more intense peaks compared to those with MEC.

The intensity of the diffraction signals associated with unreacted CA (CaAl₂O₄, PDFcard 34-440) decreased with curing time and, at any given curing age, was significantly reduced with increasing SFLG waste content. This trend aligns with the compressive strength results presented in Figure 4, which showed that the maximum compressive strength was achieved earlier with increasing amounts of SFLG waste. These findings are consistent with previous studies conducted by Pacewska et al. [6,12] and Wu et al. [13], which also concluded that incorporating silica-rich waste into CAC systems accelerates the kinetics of the reaction.

CAC pastes cured for 3 days were mainly composed of hexagonal decahydrate (D, CAH₁₀, CaAl₂O₄·10H₂O, PDFcard 12-0408) and octahydrate (O, C₂AH₈, 2CaO·Al₂O₃·8H₂O, PDFcard 11-0205). However, the incorporation of SFLG waste modified the hydration products, promoting the formation of strätlingite (S, C₂ASH₈, Ca₂Al₂SiO₇·8H₂O, PDFcard 29-0285) in samples containing ≥25 wt.% SFLG. The intensity of strätlingite peaks increased with higher SFLG contents and longer curing times so that in samples with 15 wt.% SFLG (both MAN and MEC), these signals were only distinctly observed after 90 days of curing. Strätlingite was not detected in the CAC reference paste, regardless of the curing age.

In the reference CAC paste, hexagonal CAH₁₀ and C₂AH₈ hydrates were clearly present up to 28 curing days, but their intensity decreased over time. After 90 curing days, only CAH₁₀ remained and insignificant amounts were detected after 365 curing days. In contrast, in the SFLG/CAC blended pastes, the intensity of the peaks attributed to the decahydrate CAH₁₀ generally increased up to 28 days and thereafter was either maintained or slightly increased in pastes with 35–50 wt.% SFLG waste while being slightly reduced with lower contents. Then, after 365 curing days, CAH₁₀ was still present in the SFLG/CAC pastes containing ≥25 wt.% MAN and ≥35 wt.% MEC SFLG waste, with higher amounts of CAH₁₀ with increasing SFLG waste contents. Conversely, the intensity of peaks associated with octahydrate C₂AH₈ progressively decreased with higher waste contents and longer curing time, becoming undetectable in all samples after 365 curing days.

The evolution of the cubic phase katoite (K, C₃AH₆, 3CaO·Al₂O₃·6H₂O, PDFcard 02-1124), also known as hydrogarnet, exhibited marked differences between the 100 wt.% CAC and the SFLG/CAC blended pastes. Although katoite was detected in minor quantities across all systems at early curing ages (3 and 7 curing days), after 90 and 365 curing days its diffraction peaks were most intense in the reference CAC paste and progressively diminished with increasing SFLG content. In the 100 wt.% CAC paste cured for 365 days, katoite was the predominant crystalline phase formed, signals associated with the hexagonal C₂AH₈ hydrate, initially prominent, were no longer detectable, and the intensity of the CAH₁₀ peaks, which had been very significant at 28 curing days, significantly reduced. As predicted by Equations (4) and (5), strong gibbsite (G, Al(OH)₃, PDFCard 12-0460) signals confirmed its formation during the conversion process.

These phase transformations align with the well-documented microstructural evolution of CAC systems [1,7,8,9,10], where the initially formed metastable hexagonal hydrates convert over time into the stable cubic C₃AH₆ phase. This conversion is accompanied by increased porosity and a consequent reduction in mechanical strength, as described by Equations (4) and (5). The microstructural evolution is consistent with the compressive strength results presented in Figure 4, where mortars with 100 wt.% CAC reached peak strength at 28 days, followed by a decline at later curing ages.

The microstructure of pastes that incorporated significant amounts of SFLG waste developed differently, and in doing so, the formation of the hexagonal decahydrate CAH₁₀ and strätlingite progressively increased with higher SFLG waste content, a trend that persisted over extended curing periods. Strätlingite, CAH₁₀, and the cubic phase katoite were the main phases observed in the 50 wt.% SFLG pastes cured for 90 or 365 days.

Overall, the conversion process led to a microstructure in the reference CAC paste cured for 365 days predominantly composed of the cubic phase katoite. In contrast, in blended pastes, the intensity of peaks attributed to strätlingite increased significantly with the SFLG waste content, while the peaks due to katoite progressively decreased. Consequently, while katoite remained the dominant crystalline phase in the 100 wt.% CAC and 15 wt.% SFLG samples cured for 365 days, the pastes incorporating 35 wt.% and 50 wt.% SFLG waste were primarily composed of CAH₁₀ and C₂ASH₈. This microstructural evolution aligns with the long-term compressive strength values, which, in pastes containing ≥25 wt.% SFLG waste, were significantly higher than expected (Figure 4). Conversely, the reference CAC sample and those blended with 15 wt.% SFLG followed the typical strength evolution associated with CAC systems, achieving peak strength at early ages (≤28 days) and subsequently undergoing strength reduction thanks to the conversion of metastable hydrates into the cubic C₃AH₆ phase.

When comparing the diffractograms obtained after 365 curing days based on the type of waste used, either MAN or MEC, the peaks corresponding to CAH₁₀ and strätlingite were more pronounced in the samples containing 50 wt.% MAN compared to those made with an equal amount of MEC waste (see Figure 9). In the MAN/CAC blended pastes, strätlingite became significantly present when 25% SFLG waste was added. By contrast, for the pastes blended with MEC waste, both strätlingite and CAH₁₀ started to show notable presence with 35 wt.% SFLG. Additionally, the hexagonal CAH₁₀ phase was relatively reduced, while the amount of katoite was very high in the MEC25 paste, which indicates conversion from the hexagonal hydrate to the cubic form. These microstructural differences align with the strength values recorded after 365 curing days (Figure 4), which were slightly higher for the mortars prepared with MAN waste than those made with MEC.

No distinct signals attributable to crystalline CaCO₃ were clearly distinguished in any of the SFLG/CAC blended pastes, regardless of their curing age, waste type, or percentage of substitution. These findings contrast with those reported by García Alcocel et al. [5], who observed substantial CaCO₃ formation, minimal CAH₁₀ content, and the absence of C₃AH₆ in CAC mortars exposed to accelerated carbonation for 90 days. Notably, their experimental process was intentionally designed to promote porosity and facilitate CO₂ ingress by way of using a high water-to-cement ratio (w/c = 0.7) and, as such, accelerating both carbonation and alkaline hydrolysis.

As outlined in prior studies [1,5,37,38], alkaline hydrolysis is a progressive degradation mechanism that affects CAC systems exposed to both alkalis and CO₂. This process involves a cyclic carbonation mechanism, initiated when water-soluble alkali compounds (e.g., sodium from SFLG waste) dissolve and form sodium hydroxide (NaOH). As shown in Equation (6), NaOH can react with CO₂, either from the atmosphere or calcareous aggregates, to form sodium carbonate (Na₂CO₃). The resulting carbonates can subsequently react with calcium aluminate hydrates (CAH₁₀, C₂AH₈, and C₃AH₆), producing calcium carbonates and soluble alkali aluminates (Equation (7)). These aluminates can hydrolyze in water (Equation (8)), releasing aluminum hydroxide (Al(OH)₃) and regenerating NaOH, thereby perpetuating the reaction cycle. This mechanism can lead to the progressive decomposition of the calcium aluminate hydrates, the primary binding phases in CAC, compromising the durability of the cementitious matrix.

2Na(OH) + CO₂ → Na₂CO₃ + H₂O

(6)

Na₂CO₃ + CAH₁₀ → CaCO₃ + Na₂O·Al₂O₃ + 10 H₂O

(7)

Na₂O·Al₂O₃ + 4 H₂O → 2Al(OH)₃ + 2Na(OH)

(8)

In the present study, two crystalline carboaluminate phases were identified after 365 curing days: phase E (Ca₄Al₂(CO₃)(OH)₁₂·5H₂O, PDFcard 87-0493) and phase A (3CaO·Al₂O₃·3CaCO₃·32H₂O, PDFcard 41-0215). These phases were clearly identified in the 100 wt.% CAC reference paste and in samples containing up to 35 wt.% MEC waste but only in minor amounts in pastes incorporating MAN waste. This difference is consistent with the lower Na₂O content of MAN (7.6%) compared to MEC (12.2%), as detailed in Table 3. As reported by García Alcocel [5], carboaluminate formation can inhibit or limit the conversion of metastable hexagonal hydrates into the stable cubic calcium aluminate form, but it does not prevent carbonation. Minor amounts of sodium carbonate (N, Na₂CO₃, PDFcard 25-815) were detected only in MEC35 and MEC50 pastes, which also aligns with the higher alkali content of the MEC waste.

Giner-Juan [1] reported that the final products of the alkaline hydrolysis cycle are CaCO₃ and Al(OH)₃. However, the XRD spectra of SFLG/CAC pastes cured for 365 days did not reveal significant amounts of either compound. Only minor amounts of gibbsite were detected, and its content progressively decreased with increasing SFLG substitution. Similarly, the content of katoite (C₃AH₆), a phase typically formed during the conversion process, which is typically promoted by alkalis [5], also decreased with higher SFLG contents.

These results, along with the strength gains observed from 90 to 365 days in the SFLG/CAC pastes containing ≥25 wt.% SFLG waste and the persistent presence of CAH₁₀ and strätlingite (C₂ASH₈) after long curing periods, strongly suggest that alkaline hydrolysis did not occur under the studied conditions. The results indicate that the alkali concentration in the pore solution provided by the SFLG was insufficient to initiate the cyclic carbonation mechanism, while the formation of strätlingite contributed to microstructural stabilization. These findings are consistent with Schwarz and Neithalath [39], who had reported minimal sodium ion release into the solution when glass powder was used as a pozzolanic admixture.

Although no significant Na₂CO₃ formation or degradation from alkali-induced reactions was observed, it remains essential to assess whether alkaline hydrolysis could occur under more aggressive environmental conditions, particularly in systems with higher Na₂O contents such as MEC35 and MEC50.

These findings align with those of García Alcocel et al. [5], who demonstrated that alkaline hydrolysis becomes significant only under highly aggressive chemical conditions, such as exposure to 1 M NaOH solutions, compared to milder environments (0.01 M NaOH). In their study, exposure to concentrated solutions led to increased CaCO₃ formation, accelerated conversion of hexagonal hydrates into C₃AH₆, and a reduction in mechanical strength. In contrast, in the present study, the pastes sealed during curing, without intentional exposure to carbonation or the inclusion of calcareous aggregates, thus minimizing CO₂ ingress and the potential for initiating the cyclic degradation mechanism. It is also important to acknowledge the inherent limitations of XRD analyses in detecting low-concentration or poorly crystalline phases. Then, some carbonation or hydrolysis products may fall below the detection threshold or remain undetected due to low crystallinity. These compounds may also be challenging to identity through thermogravimetric analyses because of overlapping thermal events.

Consistent with leaching test results, no signals associated with soluble heavy metal salts were identified in any of the XRD diffractograms, which indicates that mercury was effectively stabilized within the cementitious matrix. Bignozzi et al. [40] proposed that when glass waste is used as a supplementary cementitious material, leaching occurs primarily through ion exchange mechanisms, wherein calcium ions from the cement matrix are replaced by heavy metals existing as soluble salts.

In summary, the results of this study confirm the chemical reactivity of SFLG waste and align with previous findings [7,11,12,13], which reported the formation of the alternative hydrate C₂ASH₈, which plays a key role in limiting the conversion of hexagonal hydrates CAH₁₀ and C₂AH₈ into the cubic phase C₃AH₆. The absence of CaCO₃, the minor and decreasing presence of gibbsite, the reduction in katoite formation, and the observed long-term strength gains in the SFLG/CAC blended pastes all indicate that alkaline hydrolysis did not occur under the experimental conditions employed. These results suggest that the alkalis present in the SFLG waste were neither sufficiently reactive nor mobile to initiate the cyclic carbonation process described in Equations (6)–(8). However, since the microstructural analyses were conducted on sealed paste samples without aggregates, future research ought to investigate whether alkaline hydrolysis may occur in systems with calcareous aggregates and subjected to longer-term exposure or more aggressive carbonation environments, particularly in mixes with the highest sodium contents (MEC35 and MEC50).

3.6. Thermogravimetric Analyses

Table 5. Thermogravimetric results: total weight loss (TWL) and weight losses recorded in the temperature ranges of 50–220 °C and 220–400 °C.

SFLG wt.%		3 d			7 d			28 d			90 d			365 d
		TWL	50–220 °C	220–400 °C	TWL	50–220 °C	220–400 °C	TWL	50–220 °C	220–400 °C	TWL	50–220 °C	220–400 °C	TWL	50–220 °C	220–400 °C
		%			%			%			%			%
CAC	0	14.4	6.6	8.2	23.8	12.4	4.9	25.5	12.7	6.8	17.4	2.4	13.5	18.7	3.2	14.0
MAN	15	15.2	7.5	7.1	19.6	8.4	5.2	18.3	7.4	7.4	17.5	5.7	10.3	19.2	3.2	10.9
	25	14.9	7.1	6.5	17.9	7.0	4.8	15.6	6.5	6.5	17.0	4.9	8.3	20.0	7.1	7.4
	35	14.3	8.0	4.6	18.2	8.7	3.7	14.9	8.8	4.1	16.4	7.1	5. 8	21.9	9.7	3.7
	50	13.1	6.7	3.3	16.4	7.5	2.8	13.9	7.5	3.6	15.7	7.5	2.6	16.1	7.7	3.1
MEC	15	12.3	2.9	9.9	17.3	7.9	6.0	18.1	6.6	7.9	17.3	2.9	11.2	16.9	2.9	11.6
	25	12.6	2.7	9.7	17.1	8.3	5.3	15.0	5.5	7.5	15.7	3.3	8.9	16.4	4.4	9.8
	35	11.5	4.2	5.9	16.4	6.8	4.6	13.3	5.6	6.2	15.1	5.2	5.9	19.1	7.2	5.0
	50	9.5	4.9	4.4	15.8	9.7	3.6	12.5	7.0	5.0	14.3	5.3	4.6	15.7	6.7	4.0

presents the total weight loss (TWL) along with the weight losses observed within the temperature ranges of 50–220 °C and 220–400 °C for pastes cured at 20 °C for up to 365 days. The corresponding differential thermogravimetric curves (DTG) for the developed SFLG/CAC pastes are shown in Figure 10, Figure 11 and Figure 12. The TWL values are also indicated on the respective curves. As expected according to the compressive strength results (Figure 4 and Figure 5), the TWL generally decreased with increasing SFLG waste content, which it is attributed to a dilution effect. In the 100% CAC pastes, the highest TWL was observed after 28 curing days, whereas the SFLG/CAC blended pastes typically reached their maximum TWL after just 7 curing days. In this context, the results are aptly consistent with the findings reported by Pacewska et al. [12], indicating that the addition of SFLG waste accelerates CAC hydration. The TWL values also correlate with compressive strength and microstructural development, as slightly lower TWL values were recorded in the MAN pastes compared to the MEC samples for a given amount of SFLG waste and curing age.

Although the thermal decomposition temperatures could vary depending on the experimental conditions employed during the thermogravimetric analysis [12], studies by Hidalgo et al. [7] and Pacewska et al. [12] indicate that bands below 120 °C are associated with the dehydration of Al₂O₃·xH₂O (CAH hydrates and alumina gel), while those occurring between 120 and 200 °C are attributed to the dehydration of the hexagonal CAH₁₀ and C₂AH₈ hydrates. The dehydration of CSH and CASH also occurs within this temperature range [41,42], while signals observed between 180 and 220 °C are attributed to strätlingite, carboaluminate phases, and further decomposition of the octahydrate C₂AH₈. Weight losses observed in the 220–280 °C range are associated with the decomposition of AH₃ (gibbsite), while those between 280 and 350 °C correspond to the dehydration of C₃AH₆ or C₃AS₃-xH₂O.

Up to 28 curing days, the signals associated with CAH and CASH hydrates, which peaked at approximately 150–160 °C, were significantly stronger in the reference CAC paste than in the SFLG/CAC blended pastes. This is consistent with the generally higher weight loss values recorded in the 50–220 °C range for the CAC reference sample when compared to the SFLG/CAC blended pastes, as shown in Table 5. Although the XRD results indicate that the addition of SFLG modifies the typical hydration pathway of CAC and promotes the formation of the stable phase strätlingite, this phase was not clearly identified in the DTG curves due to overlapping of the dehydration bands of CAH₁₀, C₂AH₈, C₃AH₆, and C₂ASH₈.

In line with the XRD and compressive strength results, the intensity of the bands attributed to CAH and CASH hydrates (ranging from 120 °C to 200 °C) remained stable or increased after 90 and 365 curing days in the blended pastes, particularly those containing MAN waste. In contrast, the 100% CAC reference sample showed a decrease in the intensity of these bands. Simultaneously, the signals associated with AH₃ (maximum at 265–295 °C) and C₃AH₆ (maximum at 280–325 °C) were significantly intensified in both the 100 wt.% CAC and 15 wt.% SFLG samples. This microstructural evolution originated that the sample with 35 wt.% SFLG waste exhibited the highest TWL among those cured for 365 days, surpassing even the reference sample. At this curing age, the weight loss recorded in the 50–220 °C range, primarily attributed to CAH₁₀ and strätlingite in the blended SFLG/CAC pastes, was also the maximum with 35 wt.% SFLG waste. Conversely, the intensity of the bands between 225 °C and 350 °C, attributed to AH₃ and C₃AH₆ or C₃AS₃-xH₂x compounds, was the highest in the 100 wt.% CAC sample and progressively decreased with increasing SFLG content (Table 5).

These TG results, in line with the XRD findings and the compressive strength reduction reported in Figure 4, confirm the conversion of the initially formed hexagonal C₂AH₈ and CAH₁₀ hydrates into cubic C₃AH₆ in both the CAC and 15 wt.% SFLG pastes. In this sense, the formation of the stable phase strätlingite suitably confirms that the incorporation of SFLG waste alters the formed hydrates and inhibits the conversion process. These observations agree with Pacewska et al. [12], who reported that FCC addition in FCC/CAC systems accelerates CAC hydration and restricts the conversion of hexagonal hydrates into cubic C₃AH₆ by promoting the formation of strätlingite.

For a given percentage of SFLG and curing age, no significant differences were observed in the DTG curves corresponding to both types of waste, MAN and MEC, indicating the formation of similar reaction products. Nonetheless, variations in the intensities of the dehydration bands suggest differences in the relative proportions of the phases formed. In line with the XRD results and the observed strength development, the intensity of the signals associated with AH₃ and katoite (220 °C to 400 °C) was lower in the 25 MAN samples cured for 365 days compared to the 25 MEC samples cured for the same period (7.4% vs. 9.8%, respectively, Table 5). Similarly, the weight loss attributed to strätlingite and hexagonal hydrates (50 °C to 220 °C) was higher in the 25 MAN pastes than in the 25 MEC pastes (7.1% vs. 4.4%, Table 5).

3.7. Scanning Electron Microscopy

Figure 13 illustrates the microstructures of both the 100 wt.% CAC reference paste and those incorporating 25 wt.% and 50 wt.% SFLG waste, after 28 days of curing at room temperature. As expected, the paste containing 50 wt.% SFLG waste displayed lower compactness relative to the reference CAC sample, which is attributed to the dilution of CAC within the binding matrix. SEM analyses confirmed the presence of plate-shaped C₂AH₈ hexagonal hydrates, elongated hexagonal CAH₁₀ prisms [43], cubic katoite C₃AH₆, and unreacted SFLG waste particles along with amorphous hydration compounds, all previously observed through XRD and TG analyses. Strätlingite was distinguished by its chemical composition and layered morphology. As reported by Santacruz et al. [44], although the microstructure of strätlingite in cement pastes strongly depends on hydration conditions, it consists of alternating tetrahedral and octahedral layers.

4. Conclusions

This research investigated the effects of replacing up to 50 wt.% of CAC with two different types of SFLG waste, MAN and MEC, leading to the following key conclusions:

• Both SFLG wastes, rich in amorphous SiO₂, demonstrated reactivity, modifying the CAC hydration process. At substitution levels ≥25 wt.%, the formation of stable strätlingite (C₂ASH₈) was promoted, reducing the conversion of metastable hexagonal hydrates (CAH₁₀ and C₂AH₈) into the cubic phase katoite (C₃AH₆). This stabilization is crucial for enhancing long-term durability in CAC systems.
• SFLG incorporation reduced both initial and final setting times, with MEC waste extending the plastic state duration. Moreover, SFLG waste accelerated hydration, promoting earlier strength development across all substitution levels.
• The partial CAC replacement with SFLG initially reduced compressive strength due to binder dilution. However, mortars containing ≥25 wt.% SFLG exhibited significant strength gains after 365 days, reaching ~60 MPa (25 wt.%) and ~45 MPa (35 wt.%). These long-term improvements are associated with the sustained presence of CAH₁₀ and strätlingite and reduced katoite formation, which enhanced microstructural stability. In contrast, the 100 wt.% CAC and 15 wt.% SFLG mortars experienced typical long-term strength loss due to conversion of hexagonal to cubic hydrates.
• Despite the sodium content in both SFLG types, no evidence of alkaline hydrolysis was observed after 365 days. Key reaction products (CaCO₃, Na₂CO₃, and Al(OH)₃) were not significantly detected, and only minor amounts of gibbsite and carboaluminates were found in the highest Na₂O pastes (35 wt.% and 50 wt.% MEC). The persistence of CAH₁₀ and strätlingite in blends with ≥25 wt.% SFLG, along with continued strength development, further support the absence of this degradation mechanism under the tested conditions. However, further research under more aggressive environments or with calcareous aggregates is essential, particularly for high-Na₂O systems such as MEC35 and MEC50.
• All SFLG/CAC mortars met the regulatory limits for inert waste classification, with mercury leaching well below the 0.01 mg/kg threshold. Mercury was effectively stabilized in the cementitious matrix, regardless of SFLG type or content, demonstrating that untreated SFLG waste can be safely reused in CAC systems without environmental risk.

This study has demonstrated that replacing at least 25 wt.% of CAC with SFLG waste (MAN and MEC) enhances the long-term durability of mortars by promoting the formation of stable strätlingite, which limits the harmful conversion of metastable hexagonal hydrates into cubic katoite. Despite initial strength reductions due to binder dilution, mortars with ≥25 wt.% SFLG exhibited significant strength gains after 365 days. All SFLG/CAC mixes met inert waste regulations, with mercury effectively stabilized within the binding matrix, confirming the safe reuse of untreated hazardous glass waste in sustainable construction materials and contributing to lower CAC-production emissions. Although no alkaline hydrolysis was observed under the tested conditions, further research is needed to assess the long-term performance of SFLG/CAC cements under more aggressive environments, particularly in high-sodium systems and with calcareous aggregates.