To prepare WCSA clinker, four kinds of industrial solid wastes were used with complementary matching of their ingredients. Carbide slag was sampled from the Tai’an Jiepu Technology Science and Technology Ltd. (Tai’an, China); aluminum dust was sampled from Xinfa Group (Liaocheng, China); red mud was sampled from Xinfa Group (Liaocheng, China); flue gas desulfurization (FGD) gypsum was sampled from Liaocheng Coal-fired Power Plants (Liaocheng, China). The chemical compositions and x-ray diffraction (XRD) patterns for the raw materials are provided in Supplementary Table S1 and Supplementary Figure S1, respectively. Standard sand was used as the aggregate for RRM. Four kinds of additives, including polycarboxylic acid water reducer, lithium carbonate, boric acid, and dispersible latex powder, were used to adjust the hydration characteristics of RRM, i.e., fluidity, setting time, and mechanical strengths.

The schematic overview of this study is shown in Figure 1. The above-mentioned industrial solid waste materials were proportionally mixed based on the following three modulus values in Eqs 1–3. The three modulus values were set as Cm = 0.98 , P = 1.85 , and N = 3.09 , and the corresponding proportion of solid wastes is provided in Supplementary Table S2. The mixture of solid waste materials was caked and calcined at 1,250°C (holding 30 min) to prepare the WCSA clinker. Then, the clinker was mixed with 5 wt% anhydrous gypsum, ground together, to prepare WCSA. Cm = CaO − 0.7 Ti O 2 1.87 Si O 2 + 0.73 ( A l 2 O 3 − 0.64 F e 2 O 3 ) + 1.40 F e 2 O 3 (1) P = A l 2 O 3 S O 3 (2) N = A l 2 O 3 Si O 2 (3)

The prepared WCSA was mixed with standard sand at a weight ratio of 1:1 and then the mixture was blended with water at a water-to-binder ratio (W/B) of 0.28. Different amounts of the four additives were added into the mortar to adjust its hydration characteristics one by one: (i) polycarboxylate superplasticizer was added to adjust the fluidity at a low W/B ratio (Li et al., 2021; Tian et al., 2019), (ii) lithium carbonate was added to improve the early-age compressive and flexural strengths (Zhang Y. et al., 2018), (iii) boric acid was added to adjust the setting time (Cau Dit Coumes et al., 2017a), and (iv) dispersible latex powder was added to improve the bonding strength (Shi et al., 2021). With the optimized amounts of additives, the WCSA-based RRM was successfully prepared.

The compressive and flexural strengths of the prepared RRM were tested according to International Standard ISO 679-2009 “Cement - Test methods - Determination of strength”^[24]. The prepared RRM was mixed with water at a water-to-binder ratio of 0.28, and then stirred using a mixer. After stirring, the mortar was cast into 40 mm × 40 mm × 160 mm molds and de-molded after 24 h. The de-molded samples were cured at 95% humidity and 20 ± 2°C. The compressive and flexural strengths of the cured samples were tested using an electronic universal testing machine (YAW-300C; Zhongluchang Co., Ltd., China) after 2-h, 1-day, 3-day, and 28-day curing ages. The compressive and flexural strengths (1 day, 3 days, and 28 days) of the prepared WCSA were also tested according to International Standard ISO 679-2009, and the mass ratio of WCSA:standard sand:water was 1: 3: 0.5.

The bonding strength of the prepared RRM was tested according to Chinese Industry standard JC/T 2381-2016 “Repairing mortar, 2016”^[25]. A base mortar with dimensions of 70 mm × 70 mm × 20 mm was prepared using OPC and cured at 95% humidity and 20 ± 2°C for 28 days. A forming frame was placed on the base mortar and then the RRM slurry was poured into the forming frame and cured at 95% humidity and 20 ± 2°C. The bonding strength was determined after 1-day curing age.

The fluidity test for the prepared RRM was conducted on the basis of Chinese national standard GB/T 2419-2005 “Test method for fluidity of cement mortar”^[26]. The RRM slurry was added to truncated cone dies, where the round die was slowly filled, allowing the slurry to flow undisturbed and freely before it eventually stopped. Then, measurements of the largest diameter of the slurry diffusion and the vertical height were conducted, and the average value was calculated. The slurry was stirred for 6 min and measured again to complete the test.

The initial and final setting times of the prepared RRM were determined according to International Standard ISO 9597-2008 “Cement-Test methods-Determination of setting time and soundness, NEQ”^[27]. The mortar was mixed and blended to a workable consistency. A Vicat apparatus was then used to perform needle penetrations at fresh locations for each measurement. The initial and final setting times were taken as the elapsed time required to achieve a penetration of 4 ± 1 and 0.5 mm, respectively.

The chemical compositions of raw materials and WCSA were characterized by x-ray fluorescence spectrometry (XRF; Spectro XEPOS 05C, SPECTRO, Germany). The mineral phases of the raw materials, WCSA clinkers, and RRM hydration products were characterized by x-ray diffraction (XRD; Rigaku Dmax-2500 PC, Rigaku, Japan) using Cu-Ka radiation with 50 kV voltage, 100 mA current, and a scanning speed of 2.4°/min over a range of 5°–65°. The morphologies and structures of RRM samples were characterized with scanning electron microscopy coupled with an energy-dispersive spectrometer (SEM/EDS; Zeiss Supra 55, Carl Zeiss MERLIN Compact, Germany). The hydrated WCSA samples were characterized using a thermogravimetric analysis combined with the differential scanning calorimetry (TGA/DSC; TGA/DSC 1, Mettler Toledo, Switzerland) under nitrogen atmosphere with a heating rate of 10°C/min from 30°C to 400°C.

Life cycle assessment (LCA) was conducted using the SimPro ReCiPe2016 H 1.01 model according to ISO 14040 (ISO, 2006). The goal of this LCA study was to appraise the environmental impact of SCSA, CSA, and OPC clinker production chains, and the boundary considered in this study was from the cradle to the gate of the materials’ production facility. The processes included mainly raw material transportation, raw material preparation, and production of the three cementitious materials. The function unit for the three techniques was established as 1 ton production of clinker. LCA inventory data were collected for each step in each material’s production process. The inventory included mainly inputs of energy and source inputs, and outputs of waste and pollutants. Diesel was used for raw material transportation, and the vehicle used was assumed to be a diesel truck with 28 tons, and the transportation distance of each raw material was assumed to be 50 km from the mining facility or the solid waste facility to the gates of the production facility. Electricity was used mainly for raw material preparation and equipment, and coal was mainly used for material production. Data for raw material consumption, energy consumption, waste, and pollution emission were obtained mainly from industry statistics and literature (Huntzinger and Eatmon, 2009; Li et al., 2014; Li et al., 2019a; Chen et al., 2015; Jeswiet and Szekeres, 2016; Ren et al., 2017b; Ren et al., 2020). Detailed system boundary, function unit, and life cycle inventory are provided in Supplementary Table S3.

The main mineral components of WCSA were ye’elimite (Ca₄Al₆O₁₂SO₄, C₄A₃Ŝ), dicalcium silicate (C₂S, mainly α′-C₂S and β-C₂S), and anhydrite (CaSO₄, CŜ) (Supplementary Figure S2), which were similar to those of ordinary CSA cement (Ren et al., 2020). During the hydration process of WCSA, C₄A₃Ŝ can quickly react with H₂O and CŜ to form ettringite (Eq. 4) (Telesca et al., 2014), which can contribute to much of the early-age mechanical strength. C₂S can also react with H₂O to form C-S-H gel (Eq. 5) (Li et al., 2019b; Yao et al., 2019), which can contribute to the later strength development. The mechanical properties of the WCSA mortar are shown in Table 1. The prepared WCSA had compressive strengths of 41.9, 59.9, and 73.4 MPa and flexural strengths of 7.2, 8.0, and 9.1 MPa after a curing age of 1 day, 3 days, and 28 days, respectively. The ratio of 3-day compressive strength to 28-day compressive strength was above 0.8, owing to the rapid hydration of the main mineral C₄A₃Ŝ and the relatively lower hydration reactivity of C₂S(Wang et al., 2016; Li et al., 2019b; Yao et al., 2020). The initial and final setting times of the WCSA were only 20 and 28 min, respectively. These results indicated high early strength, high later strength, and fast hardening properties of the WCSA, which were suitable for RRM preparation. C 4 A 3 S ^ + 2 C S ^ + 38 H →   C 3 A ⋅ 3 C S ^ ⋅ 32 H + 2 A H 3 (4) 2 C 2 S + n H 2 O → x C 2 S ⋅ y H + ( 2 − x ) CH (5)

Polycarboxylate superplasticizer was added into the prepared RRM to adjust its fluidity at a low W/B ratio. The influence of different amounts of polycarboxylate superplasticizer on the fluidity and 2-h strength development of the WCSA-based RRM was studied. As shown in Figure 2A, the fluidity of the RRM increased with the increase of the amount of polycarboxylate superplasticizer added. The SEM images of RRM (Supplementary Figure S3) indicated that the minerals were more dispersed at 0.2% polycarboxylate superplasticizer than those at 0% polycarboxylate superplasticizer, and more needle-like ettringite was observed. Polycarboxylate superplasticizer preferentially could absorb on aluminate phases that have a positively charge surface and modify the fluidity of the fresh mixture by producing electrostatic repulsion between the anhydrous and hydrated cement particles (Tian et al., 2019; Li et al., 2021), which contributed to the increase of fluidity of RRM with polycarboxylate superplasticizer added. A recent study reported that the use of polycarboxylate superplasticizer did not significantly influence the hydration process of ye’elimite after the main hydration period and could reduce expansion after 4 days of curing (Li et al., 2021). With these effects, this work used polycarboxylate superplasticizer to promote the constructability and durability of the RRM.

The existing literature has shown that the adsorption of polycarboxylate superplasticizer on ettringite could inhibit the growth of ettringite crystals and delay the setting time of CSA cement in the early-age hydration period (Sun et al., 2011; Ma et al., 2014; Su et al., 2019; Li et al., 2021), which may affect the early-age strengths of RRM. In this work, the 2-h compressive and flexural strengths of RRM fluctuated with the addition of polycarboxylate superplasticizer, and the maximum values occurred when the addition amount was 0.2% (Figure 2B). However, the overall compressive and flexural strengths of RRM at 2-h hydration age were relatively low and were decreased significantly when the amount of polycarboxylate superplasticizer content was higher than 0.2%. The XRD patterns of hydrated WCSA samples with different amounts of polycarboxylate superplasticizer at 2-h hydration age (Figure 3) showed that the main mineral phases of all samples were ettringite, anhydrite, and ye’elimite. When the polycarboxylate superplasticizer was 0.2%, the ettringite peak was the highest and the anhydrite peak was the lowest, indicating the highest degree of hydration among the samples. The TG/DTG curves (Supplementary Figure S4A) indicated that, with the increasing amounts of added polycarboxylate superplasticizer, the amounts of ettringite first increased and then decreased, which was consistent with the change trend of the 2-h compressive and flexural strengths of RRM.

The above results indicated that a small content of polycarboxylate superplasticizer could increase the fluidity of WCSA effectively. However, excessive polycarboxylate superplasticizer would inhibit the hydration of WCSA and the formation of ettringite, thereby reducing the early-age strength of RRM.

When the amount of added polycarboxylate superplasticizer was set at 0.2%, to deal with the decline of the early-age strength of RRM caused by polycarboxylate superplasticizer, lithium carbonate was added. The influence of different amounts of lithium carbonate on the initial setting time and 2-h strength development of the WCSA-based RRM is shown in Figure 4. The 2-h compressive and flexural strengths obviously increased with the addition of lithium carbonate and then dropped when the addition amount was over 0.2% (Figure 4A). Correspondingly, the initial setting time of RRM initially decreased and then increased (Figure 4B). The highest 2-h compressive and flexural strengths reached 36.5 and 8.5 MPa, respectively, which were much higher than those without lithium carbonate addition, indicating that lithium carbonate strongly accelerated the early-age hydration of WCSA in RRM.

During the hydration process of RRM, Li⁺ introduced by lithium carbonate would induce a fast precipitation of amorphous Li-containing Al(OH)₃ and the consumption of aluminate ions from the solution (Cau Dit Coumes et al., 2017b). Consequently, it would induce a fast dissolution of ye’elimite. Meanwhile, Li-containing Al(OH)₃ would provide seeds for the heterogeneous nucleation of amorphous Al(OH)₃, leading to massive precipitation of hydrates (mainly ettringite) (Cau Dit Coumes et al., 2017b). Therefore, the early-age WCSA hydration was significantly accelerated and the mechanical strengths were quickly increased. However, the acceleration effect of lithium carbonate was not linear with the addition amount. The XRD patterns and TG/DTG curves of hydrated WCSA samples with different amounts of lithium carbonate at 2-h hydration age (Supplementary Figures S4B, S5A) showed that, the ettringite peak intensity was the highest at 0.2% addition amount, rather than increased with the addition amount of lithium carbonate. As shown in Figure 4A, the compressive and flexural strengths decreased when the addition amount exceeded 0.2%. Cau Dit Coumes et al. (2017a) found that, the duration of period of low thermal activity produced by WCSA hydration tended to decrease at higher lithium additions, indicating the lack of sustainability of the acceleration effect induced by excessive lithium carbonate. Parr et al. (2004) also mentioned that the higher the lithium concentration, the faster the hydration of calcium aluminate cement until a limit above which a plateau effect occurred. These might explain the limited increase in compressive and flexural strengths induced by excessive lithium carbonate.

The above results indicated that small addition of lithium carbonate could effectively increase the early-age mechanical strengths of RRM and reduce the setting time. However, as shown in Figure 4B, the initial setting time was decreased to less than 15 min, which was unfavorable to the actual engineering application of RRM. Borate compound is well known for acting as set retarder during the hardening of cement (Hu et al., 2017; Chen et al., 2019). In this work, boric acid was added to extend the initial setting time of RRM (added 0.2% polycarboxylate superplasticizer and 0.2% lithium carbonate). As shown in Figures 4C,D, the initial setting time of RRM was prolonged with the addition of boric acid, but the compressive and flexural strengths decreased quickly as the addition amount increased.

In Figure 4C, small addition of boric acid (≤0.06%) slightly increased the mechanical strengths of RRM. It might be explained that a small amount of boric acid neutralized OH⁻ and promoted the release of Li⁺, which accelerated the hydration of RRM. However, a higher amount of boric acid significantly reduced the mechanical strengths of RRM. It has been reported that a poorly crystallized borate compound [ulexite, NaCaB₅O₆(OH)₆A 5H₂O] was formed during the CSA cement hydration, which covered the surface of CSA clinker particles and prevented the further dissolution of ye’elimite, producing a strong retarding effect (Champenois et al., 2015; Chen et al., 2019). The SEM images of RRM (Figure 5) showed that, some substances were attached to the surface of the hydration products when 0.16% boric acid was added, indicating the formation of poorly crystallized borate compound. Meanwhile, excessive H⁺ in boric acid neutralized OH⁻ in Al(OH)₃ and inhibited the formation of ettringite. These effects led to the obvious decrease in mechanical strengths of RRM.

When the addition amount of boric acid was 0.16%, the initial setting time was 28 min, which was close to the requirement of ≤30 min in JC/T 2381-2016 “Repairing mortar, 2016”^[25]. Meanwhile, the compressive and flexural strengths still reached 36.1 and 8.5 MPa at 0.16% addition amount, respectively. Therefore, 0.16% addition amount of boric acid might be more suitable for the RRM and was chosen in this study.

When the addition amount of polycarboxylate superplasticizer, lithium carbonate, and boric acid added were set at 0.2%, 0.2%, and 0.16%, respectively, latex powder was added to increase the bonding strength of RRM. Redispersible latex powder could increase the flexural strength and bonding strength of the mortar, but it would reduce the compressive strength (Wang Peiming, 2018; Liu et al., 2019; Fan Chengwen, 2020; Lin et al., 2020). In this work, the influencing effects of latex powder on the 1-day bonding strength and 2-h mechanical strengths of RRM were studied (Figure 6). In the presence of an increased amount of latex powder, the bonding strength and flexural strength of the RRM generally increased; however, the 2-h compressive strength showed an overall downward trend. When the addition amount of latex powder exceeded 3%, the compressive strengths significantly decreased and the flexural strengths also decreased slightly.

The latex powder was able to increase the mortar bonding strength and flexural strength because it could form an emulsion with water and became evenly dispersed in the RRM slurry. The polymer particles distributed in the emulsion then gradually deposited onto the cement particle surfaces during the hydration process. The polymer particles were connected by mutual fusion to form membranes, and they eventually formed a continuous network structure into which the RRM hydrates became intertwined (Wang Peiming, 2018). This process increased the contact area between the internal RRM particles and played a bridging role to decrease the tensile stress, and it was able to effectively absorb and transmit energy (Fan Chengwen, 2020; Wang Peiming, 2018). As a result, the bonding strength and flexural strength of the RRM was increased. However, the XRD patterns of hydrated WCSA samples with different amounts of latex powder added (Supplementary Figure S5B) showed that the anhydrite peak intensity generally increased when the addition amount increased from 1% to 4%, indicating the inhibitory effect of latex powder on the hydration of WCSA. In addition, Zhao et al. (2019) found that there were enlargement effects of redispersible polymer powders on porosity, average pore diameter, and mean pore diameter of pores in cement mortars. These effects probably led to the reduction of the compressive strength.

When the addition amount of latex powder was 2%, the 1-day bonding strength and 2-h compressive and flexural strength were 2.01, 32.5, and 9.2 MPa, respectively, which were much higher than the requirements for rigid repair mortar in JC/T 2381-2016. After curing age of 1 day, 3 days, and 28 days, the compressive strengths of RRM reached 57.2, 58.6, and 71.8 MPa, and the flexural strengths reached 12.6, 15.3, and 17.7 MPa, respectively. These results indicated that a high-performance RRM was successfully prepared via using solid waste-based CSA cement as binding material, standard sand as aggregate, and four additives as performance modifiers. Collectively, the optimal addition amounts of polycarboxylate superplasticizer, lithium carbonate, boric acid, and latex powder were set at 0.2 wt%, 0.2 wt%, 0.16 wt%, and 2 wt%.

Both CSA cement and OPC are usually used as the cementing materials for RRM. In order to evaluate the potential environmental benefits of using WCSA to prepare RRM, the LCA analysis of the production of WCSA, CSA cement, and OPC was conducted. The LCA results of 18 environmental impact categories are shown in Figure 7A. The LCA results indicated that, the WCSA production had a much lower overall environmental impact than traditional CSA cement and OPC production. It had obvious environmental benefits in the following major categories, i.e., marine ecotoxicity, terrestrial ecotoxicity, freshwater ecotoxicity, human carcinogenic and non-carcinogenic toxicity, and global warming. Previous studies also concluded that the process for preparing WCSA was more environment-friendly than that for producing traditional CSA cement (Ren et al., 2017b; Yao et al., 2019).

(ME1—Marine ecotoxicity; TE—Terrestrial ecotoxicity; TA—Terrestrial acidification; FE1—Freshwater ecotoxicity; HCT—Human carcinogenic toxicity; HNCT—Human non-carcinogenic toxicity; GW—Global warming; FRS—Fossil resource scarcity; OFTE—Ozone formation, Terrestrial ecosystems; FPMF—Fine particulate matter formation; OFHH—Ozone formation, Human health; FE2—Freshwater eutrophication; SOD—Stratospheric ozone depletion; WC—Water consumption; LU—Land use; IR—Ionizing radiation; MRS—Mineral resource scarcity; ME2—Marine eutrophication.)

The production of WCSA could consume large amounts of industrial solid wastes, reduce the use of high-grade raw materials, and increase the added value of industrial solid wastes. Production costs of WCSA, CSA, and OPC clinkers were also evaluated (Figure 7B), and the total production costs for WCSA, CSA, and OPC clinkers were estimated at 39.33$, 74.78$, and 48.42$, respectively (detailed data are provided in Supplementary Table S4). The use of industrial solid wastes significantly contributed to the cost reduction for WCSA production. In addition, a comparison of the mechanical properties among WCSA, CSA, and OPC (Supplementary Table S5) showed that WCSA had much higher compressive strengths than commercial CSA and OPC with lower cost, especially the early-age strengths of WCSA were two to three times that of OPC. The high early-age strength property of WCSA (i.e., 3-day hydration to achieve 80% of 28-day compressive strength) could effectively reduce the curing time of RRM, thereby improving repair efficiency. Therefore, WCSA obtained better performance at a lower cost and had greater advantages in preparing RRM. These results indicated that the production of WCSA would achieve large-scale utilization of industrial solid wastes and high performance, reducing environmental burden and production cost, which might have the potential to promote the industrial chain development of solid waste-derived high-performance RRM.