Introduction

The rapid development of new energy batteries is a key initiative supporting global low-carbon strategies. Among these, lithium-ion batteries have emerged as the dominant energy storage technology due to their high energy density and long cycle life, driving a surge in lithium mining1,2,3. China endowed with abundant and geographically concentrated lithium reserves, primarily extracting lithium resources from Sichuan, Jiangxi, and Tibet provinces4. However, lithium extraction and processing methods vary across regions due to differences in raw materials and refining techniques5. For instance, Sichuan province predominantly employs the acid roasting method to produce lithium carbonate from spodumene ore. While lithium mining generates significant economic benefits, it also poses urgent environmental challenges. Notably, producing one ton of lithium carbonate generates 8–10 tons of lithium slag (LS), the disposal of which remains a critical issue6,7. Currently, LS is primarily managed through open-air stockpiling, which not only wastes land resources but also risks ecological contamination8,9,10. Thus, developing efficient LS recycling strategies has become an indispensable ecological imperative for the sustainable advancement of new energy technologies.

The resource utilization of various industrial wastes in building materials is considered one of the most effective disposal methods11,12. Some researches showed that LS exhibits certain pozzolanic effects. Its main oxide components (SiO2, CaO, Al2O3, etc.) can further react with Ca(OH)2 from ordinary Portland cement hydration to form products such as calcium silicate hydrate, leading to studies exploring LS applications in building materials such as supplementary cementitious materials, cement production, and ceramic manufacturing13,14,15. As such, past research findings indicate that while LS used alone as an admixture enhances 28-day compressive strength and carbonation resistance of concrete, it may reduce workability and chloride ion penetration resistance16. Li et al. demonstrated that raw materials containing less than 5% LS effectively promote early strength of white Portland cement5. Moreover, high-temperature sintering of LS and fly ash mixtures can produce lightweight aggregates with low weight, high strength and low leaching toxicity17. Zhou et al. successfully prepared green, low-carbon lightweight aggregates from LS through alkali activation and confirmed its high reactivity and application potential through comparative performance analysis with other solid waste-derived aggregates18. As such, utilizing LS resource in building materials becomes an increasingly prominent research focus due to its inherent advantages and environmental benefits.

It is undeniable that despite numerous researchers having explored various feasible approaches for the resource utilization of LS in building materials, its actual utilization rate remains relatively low19,20. For instance, China’s overall LS utilization rate still falls below 30%4. The limited application of LS in construction materials can be primarily attributed to several factors. Firstly, Chinese Standard (GB175-2023) “Common Portland Cement” has imposed restrictions on the use of LS in cement production, which was previously considered the most valuable and highest utilization approach for LS resource recovery21,22. Secondly, the relatively high sulfur content in LS may compromise the durability of cement-based materials, consequently limiting its application as an admixture. Finally, certain inherent characteristics of LS, such as its high-water absorption and fine particle size, adversely affect the performance of cement-based materials, thereby restricting its allowable dosage23,24. Therefore, to further expand the application avenues of LS in building materials and enhance its utilization rate, it is imperative to explore more solutions to address both the inherent limitations of LS and its negative impacts on the performance of cement-based materials.

The primary hydration product of calcium sulfoaluminate cement (CSC) is ettringite, and its hydration process consumes sulfate ions. Therefore, incorporating LS as an admixture in calcium sulfoaluminate cement-based materials may potentially promote ettringite formation or mitigate the detrimental effects of sulfur elements in LS on material durability25. Meanwhile, cement particles of different sizes significantly influence the performance and hydration process of cement-based materials26. However, research remains scarce regarding how LS affects the hydration of cement with varying particle sizes.

Therefore, based on previous research, this study aims to explore in detail the influence law of LS on the hydration performance of different types of cement slurry through the form of external admixture. This study systematically investigates the effects of LS content (0%, 10%, 20%, and 30%) on the properties of cement pastes with different cement types (OPC and CSC) and particle sizes (OPC and UPC), including fluidity, setting time, water absorption, and compressive strength. Furthermore, X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FTIR), and scanning electron microscopy (SEM) are employed to analyze the microstructural and chemical characteristics of various cement pastes, aiming to elucidate the underlying mechanisms by which LS influences the performance of different cement paste systems.

Experimental program

Raw materials

The LS used in the experiment is obtained from Meishan City, Sichuan Province, which is processed through drying and 1-hour ball milling. Conch brand OPC, 800-mesh UPC, and 500-mesh CSC were all purchased from Jiuqi Building Materials Company. A polycarboxylate superplasticizer is used as the water reducing admixtures. The chemical compositions of the four raw materials are presented in Table 1. The particle size distribution curves and crystalline phase analysis results of the four materials are shown in Figs. 1 and 2, respectively. The particle size distribution reveals that the materials can be ranked from finest to coarsest as follows: LS (Dv(50) = 5.55 μm), UPC (Dv(50) = 7.84 μm), PC (Dv(50) = 21.36 μm), and CSC (Dv(50) = 25.45 μm). XRD analysis indicates that the main crystalline phases in PC and UPC are tricalcium silicate (C3S), dicalcium silicate (C2S), and tetracalcium aluminoferrite (C4AF). The primary components of CSC are identified as calcium sulfoaluminate (C4A3\(\:\stackrel{-}{\text{S}}\)) and calcium carbonate (CaCO3). LS is found to consist mainly of gypsum dihydrate (CaSO4·2H2O), lithium aluminum silicate (LAS), and leached spodumene (LSP).

Table 1 Chemical composition of Raw materials/wt%.
Full size table
Fig. 1
Fig. 1
Full size image

Particle size distribution curves of the four raw materials.

Fig. 2
Fig. 2
Full size image

Crystalline phase analysis results of the four raw materials.

Sample preparation

The water-to-cement ratio is fixed at 0.5 for all cement pastes. LS is incorporated at 10%, 20%, and 30% by mass of cement, with detailed mix proportions listed in Table 2. Given the finer particle sizes of UPC and CSC, a polycarboxylate-based superplasticizer is added to maintain workability. Fresh pastes are prepared using an SYJ-10 high-speed cement mixer for thorough homogenization. After measuring fluidity, the mixtures were cast into 40 mm cubic molds and cured at ambient temperature for 24 h. Demolded specimens are subsequently stored in a standard curing chamber (20 ± 2 °C, ≥ 95% RH) until testing ages. Figure 3 illustrates the sample preparation and testing protocol.

Table 2 Mix proportions of experimental groups/kg m− 3.
Full size table
Fig. 3
Fig. 3
Full size image

Flow chart of sample preparation and performance testing.

Test methods

The fluidity of fresh pastes is tested according to Technical Specification for Cement Grouting Construction of Hydraulic Structures DL/T 5148 − 202127. Fresh paste is poured into a truncated cone mold (upper diameter: 36 mm, lower diameter: 60 mm, height: 60 mm) until level with the top, then the mold is vertically lifted. After stabilization, the diameters in two perpendicular directions are measured, with fluidity taken as the average value of the diameter value. Following DL/T 5148 − 2021, a portion of fresh paste is poured into a 100 mL graduated cylinder. After stabilization, the volume of separated water is recorded to calculate bleeding rate. Another portion is used for setting time testing per Test Methods for Water Requirement of Normal Consistency, Setting Time and Soundness of Cement GB/T 1346–201128. The remaining paste is cast into 40 mm cubic molds. After demolding, 1-day compressive strength is immediately tested. Other specimens are cured in a standard chamber until 7- and 28-day tests. All reported data represent averages from triplicate tests. Crushed fragments from compressive tests are dried and ground for microstructural, crystalline phase, and chemical bond analyses to elucidate mechanistic effects of LS on hydration and strength development.

Test results and discussions

Fluidity

Figure 4 presents the effects of LS content on the fluidity of different cement pastes. Overall, the incorporation of LS reduces the fluidity of cement pastes, with a more pronounced decrease observed at higher LS dosages. This is because the addition of LS effectively decreases the water-to-binder ratio in the paste system29. In addition, LS particles exhibit finer size distribution (Fig. 1), porous structure, and rougher surface morphology compared to cement particles (Fig. 3), resulting in higher water absorption capacity16,30. From the perspective of cement types, LS demonstrates the most significant impact on the fluidity of CSC, followed by OPC, with UPC showing the least sensitivity. Specifically, when the LS content increases from 0% to 30%, the fluidity of OPC, UPC, and CSC pastes decreases by 55.76%, 46.15%, and 63.08%, respectively. The relatively smaller fluidity reduction in UPC can be explained by similar particle size distributions of UPC and LS, making UPC less susceptible to the physical effects of fine LS particles. The most substantial fluidity decline in CSC systems stems not only from the physical characteristics of LS but also from chemical interactions. The gypsum present in LS accelerates the formation of ettringite, a process that consumes additional water and consequently further impairs workability25.

Fig. 4
Fig. 4
Full size image

Effect of LS content on fluidity of different cement pastes.

Setting time

The initial and final setting times of three types of cement pastes containing different LS contents are tested, and the results are shown in Fig. 5. Without LS, the initial setting times of OPC, UPC and CSC are 302 min, 384 min and 40 min respectively, while the final setting times are 673 min, 457 min and 130 min respectively. Compared with OPC, UPC exhibits longer initial setting time but shorter final setting time. This is because UPC has finer particle size and larger specific surface area. In the early stage, hydration products are more likely to adhere to the surface of cement particles, thus affecting the initial setting time31. However, in the later stage, the higher specific surface area promotes the reaction rate of cement particles32. Compared with OPC and UPC, CSC shows significantly shorter setting times, which is attributed to the much faster hydration rate of calcium sulfoaluminate minerals in CSC than that of dicalcium silicate and tricalcium silicate minerals in OPC and UPC. The addition of LS shortened the setting times of cement pastes, and higher LS content resulted in shorter setting times. This is because LS absorbs part of the water, effectively reducing the water-cement ratio of the paste. In addition, the gypsum contained in LS can be used as raw material for accelerating the formation of ettringite, thereby shortening the setting time25. When the LS content increased from 0 to 30%, the initial setting times of OPC, UPC and CSC decreased to 237 min, 329 min and 31 min respectively, while the final setting times reduced to 574 min, 392 min and 84 min respectively.

Fig. 5
Fig. 5
Full size image

Effect of LS content on setting times of different cement pastes: (a) initial setting time; (b) final setting time.

Bleeding rate

The bleeding rate of cement paste is closely related to the stability of the paste. During testing, no bleeding phenomenon is observed in UPC and CSC pastes, while OPC paste exhibits bleeding. The bleeding rate of OPC paste with varying LS contents is measured, as showing in Fig. 6. Without LS, the bleeding rate reaches 3.9%. As LS content increases, the effective water-to-cement ratio decreases, leading to a reduction in bleeding rate. At 30% LS content, the bleeding rate dropped to 0.2%, effectively enhancing the stability of the paste.

Fig. 6
Fig. 6
Full size image

Variation of bleeding rate in OPC paste with LS content.

Compressive strength

Figure 7 shows the effect of LS content on the 1d, 7d and 28d compressive strength of different cement pastes. The 1d compressive strength, representing the early-age strength of pastes, can characterize the hydration rate of cement and is closely related to the setting time. When the LS content is 0%, the 1d strengths of OPC, UPC and CSC pastes are 7.24 MPa, 5.10 MPa and 21.54 MPa, respectively. These results indicate that OPC and UPC have relatively slow hydration rates, while CSC has a fast hydration rate, which is consistent with the setting time test results. With increasing LS content, the early-age strength of all pastes shows an increasing trend. When the LS content increases from 0% to 30%, the 1d strengths of OPC, UPC and CSC pastes increase by 4.83%, 32.75% and 38.39%, respectively.

After 7 days of curing, the compressive strengths of OPC, UPC, and CSC pastes increase to 29.29 MPa, 43.61 MPa, and 28.56 MPa, respectively. Notably, UPC paste exhibits the most significant strength development, attributable to the more complete hydration of its finer cement particles. In contrast, CSC shows the lowest strength gain, as most of its reactive components has already hydrated during early stages. With increasing LS content, all pastes demonstrate continued strength enhancement. When LS content rises from 0 to 30%, the 7-day strengths of OPC, UPC, and CSC pastes increase by 61.35%, 6.67%, and 52.28%, respectively. Unlike its effect on 1-day strength, LS shows minimal influence on the 7-day strength of UPC paste. This trend persists for 28-day strength development. At 28 days of curing, the strengths of OPC, UPC, and CSC pastes reach 42.88 MPa, 59.39 MPa, and 31.66 MPa, respectively. Corresponding strength increases of 79.41%, 7.22%, and 54.49% are observed when LS content increases from 0% to 30%. The improved later-age strength primarily stems from LS filler effect and the participation of its reactive components in hydration reactions, leading to increased formation of hydration products33,34,35. This explanation can be further confirmed by the results of microstructure and mineral composition tests. Combined with microstructural analysis, it reveals that the UPC paste without LS had the greatest structural density due to more complete hydration. Therefore, the improvement space for adding LS was limited. Conversely, both OPC and CSC pastes without LS exhibit noticeable microcracks, which are progressively mitigated with higher LS incorporation.

Fig. 7
Fig. 7
Full size image

Effect of LS content on compressive strength of different cement pastes: (a) OPC; (b) UPC; (c) CSC.

Microstructural analysis

Figure 8 presents the 1-day microstructures of three cement pastes with 0% and 30% LS content. Figure 9 shows EDS analysis of some typical components in the slurry. For OPC paste without LS, the primary hydration products are calcium silicate hydrate (C-S-H) and calcium hydroxide (CH), consistent with observations in LS-free UPC paste. Both OPC and UPC pastes without LS exhibit relatively porous microstructures with evident voids, explaining their lower early-age strengths. With 30% LS addition, while the hydration products in PC paste remained largely unchanged, significant reduction in macro-pores is observed. This improvement, attributed to LS filler effect, corresponds to the moderate strength increase. Notably, UPC paste containing 30% LS shows distinct ettringite (AFt) crystal formation, transforming its originally porous structure into a denser matrix and accounting for its substantial strength enhancement. In contrast, CSC paste without LS already displays a dense microstructure composed mainly of ettringite crystals and aluminum hydroxide gel (AH3), explaining its superior 1-day strength exceeding 20 MPa. The incorporation of 30% LS further densifies the CSC matrix with increased AH3 content, facilitated by gypsum in LS which accelerated hydration reactions36. Compared with traditional admixtures such as fly ash, silica fume and mineral powder, LS containing gypsum is expected to further enhance the early strength of cement paste as an admixture13.

Fig. 8
Fig. 8
Full size image

1-day microstructure of different cement pastes.

Fig. 9
Fig. 9
Full size image

EDS analysis of typical substances in the slurry.

Figure 10 presents the 28-day microstructures of three cement pastes with 0% and 30% LS contents. Compared to their 1-day microstructures, all three cement pastes develop significantly denser microstructures. The UPC paste contains only minor pores, while both OPC and CSC pastes still exhibit visible cracks, explaining why UPC achieves the highest 28-day strength. With 30% LS addition, all three cement pastes show further microstructural densification, with substantial reductions in both pores and cracks. This microstructural improvement represents one of the key factors contributing to the observed strength enhancement.

Fig. 10
Fig. 10
Full size image

28-day microstructure of different cement pastes.

Crystalization analysis

Figure 11 presents the XRD results showing the effect of LS content on the crystalline phases of different cement pastes at 1 day. From Fig. 11(a) and 11(b), it can be observed that OPC and UPC pastes with the same LS content show similar crystalline compositions, primarily consisting of hydration product CH, unhydrated C3S and C2S, as well as components from LS. With increasing LS content, the diffraction peak intensities corresponding to LS components increased, while those of hydration products show minimal changes, indicating that LS primarily functions as a filler at early ages. These XRD results are consistent with the microstructural observations. In contrast to OPC and UPC pastes, CSC paste exhibits significantly different crystalline phases, dominated by hydration product ettringite, unhydrated calcium sulfoaluminate, and more abundant calcium carbonate. The more prominent calcium carbonate peaks in CSC paste can be attributed to two factors. First, higher initial calcium carbonate content in CSC compared to OPC and UPC. Moreover, the tendency of AFt in CSC paste to react with atmospheric CO2 to form additional calcium carbonate37,38,39. Similar to OPC and UPC pastes, LS addition has limited influence on the crystalline composition of CSC paste. Notably, the C-S-H gel observed in OPC and UPC pastes and AH3 gel in CSC paste through microstructural analysis does not show distinct characteristic peaks in Fig. 11, confirming their existence as amorphous phases.

Fig. 11
Fig. 11
Full size image

XRD results of different cement pastes at 1 day: (a) OPC; (b) UPC; (c) CSC.

Figure 12 shows the effect of LS content on the crystalline phases of different cement pastes at 28 days. Compared with the 1-day results, OPC, UPC and CSC pastes exhibit similar types of crystalline phases, and LS content shows comparable influence patterns on all three pastes. With extended curing age, the diffraction peaks of C2S and C3S in PPC and UPC pastes disappeared, indicating complete hydration of cement. The ettringite diffraction peaks in CSC paste significantly intensifies, demonstrating continuous hydration development. Furthermore, increasing LS content leads to reduced intensity of CH diffraction peaks in OPC and UPC pastes, suggesting that reactive components in LS consumed part of the CH through hydration reactions40,41,42. This observation confirms the pozzolanic activity of LS in these systems at later ages.

Fig. 12
Fig. 12
Full size image

XRD results of different cement pastes at 28 days: (a) OPC; (b) UPC; (c) CSC.

Chemical bond types

Figure 13 investigates the influence of LS content on chemical bond types in different cement pastes. Figure 13(a) displays the FTIR spectra of OPC pastes with four LS incorporation levels. The results demonstrate that LS content has minimal effect on the chemical bond types, with all pastes showing characteristic peaks primarily at 3450 cm− 1, 1650 cm− 1, 1424 cm− 1, 1115 cm− 1, 960 cm− 1, 880 cm− 1 and 450 cm− 1. The peaks at 3450 cm− 1 and 1650 cm− 1 correspond to O-H bond vibrations, indicating the presence of free or bound water in the pastes. The characteristic peaks at 1424 cm− 1 and 880 cm− 1 confirm the existence of calcium carbonate. The peak at 1115 cm− 1 represents [SO4]2− vibration, demonstrating the presence of gypsum. The peaks at 960 cm− 1 and 450 cm− 1 are attributed to Si-O bond vibrations originating from C2S (Ca2SiO4) in cement and LSP (LiAlSi2O6) in LS43. Consequently, the peak intensity at 960 cm− 1 in OPC paste increases with higher LS content.

Figure 13(b) presents the chemical bond characteristics of three cement pastes with 0% and 30% LS content. The results indicate nearly identical chemical bond types between UPC and OPC pastes. Notably, the characteristic peak at 1115 cm− 1 completely disappears in UPC paste without LS, confirming the absence of gypsum. In contrast, the presence of gypsum in both PC and LS results in distinct diffraction peaks at 1115 cm− 1 for OPC, OPC-3, and UPC-3 pastes. Significant differences emerge when comparing CSC paste with OPC and UPC pastes. First, the peak intensities at 3450 cm− 1 and 1650 cm− 1 show remarkable enhancement, attributable to the substantial bound water content in AFt crystals within CSC paste35,44. Meanwhile, similarly intensified peaks at 1115 cm− 1 result from the abundant gypsum and AFt content in CSC systems. Moreover, a new diffraction peak appears at 620 cm− 1, also corresponding to [SO4]2− vibrations45. Conversely, the peak intensity at 960 cm− 1 decreases significantly due to reduced C2S content in CSC paste.

Fig. 13
Fig. 13
Full size image

FTIR analysis of chemical bond types in different cement pastes at 28 days.

Conclusions

This study systematically investigated the effects of LS on the physico-mechanical properties of OPC, UPC, and CSC pastes, while exploring the underlying mechanisms through microstructural characterization. The main conclusions are as follows:

  1. (1)

    LS incorporation reduces paste fluidity and shortens setting times. At 30% LS content, the fluidity of all three cement pastes decreased by 46.15–63.08%, with initial and final setting times reduced by 9–65 min and 46–99 min, respectively.

  2. (2)

    LS addition enhances the 1d, 7d, and 28d compressive strengths of all cement pastes. With 30% LS content, the 1d strength increased by 4.83–38.39%, while the 28d strength improved by 7.22–79.41%.

  3. (3)

    SEM, XRD, and FTIR analyses reveal that the effects of LS stem from a filling effect and the activation of hydration reactions, where reactive components and gypsum in LS promote the formation of hydration products such as calcium silicate hydrate (C-S-H) and ettringite.

  4. (4)

    The addition of LS has performance improvement effects on slurries of different cement types, but the influence patterns are significantly different. This study provides a theoretical basis for the application of LS indifferent types of cement slurry.