Mix Proportion Design and Performance Regulation of 3D-Printing Phosphogypsum-Based Geopolymer Paste

Abstract

Building 3D printing technology exhibits remarkable construction advantages, with solid waste-based 3D printing slurry emerging as a research hotspot in the field. Phosphogypsum is compatible with diverse solid wastes for the fabrication of geopolymer, whereas its feasibility as a 3D printing material merits further investigation. In this study, calcium carbide slag (CS), ground granulated blast-furnace slag (GGBS), recycled concrete powder (RCP), and phosphogypsum (PG) underwent co-activation. The mix proportion received optimization via response surface methodology (RSM), and printability assessment proceeded based on the optimized proportion. Key conclusions include the following: PG exerts a role in optimizing the internal structure within the geopolymer matrix. The 28-day compressive strength of the composite geopolymer exceeds 25 MPa. Application as a 3D printing material facilitates enhancement of slurry stability in the later stage. Excessive PG addition elevates the shear stress and viscosity of the 3D printing paste, shortens the paste open time, and impedes paste extrusion and molding. Based on a comprehensive analysis of printability and the performance of printed specimens, the optimal mix proportion of the phosphogypsum-based geopolymer 3D printing paste was determined as follows: CS: 22.5%; GGBS: 45%; RCP: 22.5%; PG: 10%; W/b: 0.4.

1. Introduction

Building 3D printing features a short construction period, low cost, and capacity to satisfy personalized construction requirements [1,2,3], with material performance serving as the core key to guaranteeing its smooth implementation. For cement-based cementitious materials, fluidity, extrudability, constructability, and high mechanical properties prove indispensable to ensure smooth transportation, continuous extrusion, and structural stability throughout the printing process [4,5]. Driven by dual demands for efficient solid waste resource utilization and low-carbon development in the construction industry, the application of solid waste as a substitute for cementitious materials grows increasingly urgent. Numerous existing studies employ solid waste-based materials, including fly ash [6,7], waste concrete powder [8], and calcined clay [9], to replace cement in 3D printing material preparation. These materials demonstrate favorable printability and mechanical properties, while also offering distinct advantages of low carbon footprint, environmental friendliness, and cost reduction.

Phosphogypsum, a solid waste by-product generated during phosphoric acid production, requires urgent resourceful utilization [10]. Existing studies indicate that the combination of phosphogypsum with other solid waste-based materials promotes hydration product formation via sulfate ion (SO₄²⁻) provision, thereby accelerating early strength development of the cementitious system. Bheem Pratap et al. [11] fabricated cementitious materials with a compressive strength up to 47.97 MPa using two solid wastes—phosphogypsum and fly ash. In the field of building materials, phosphogypsum also finds application in the preparation of artificial gypsum particles, cement mortar, saline–alkali soil conditioners, cement, and gypsum boards [12,13,14,15]. Calcium carbide slag [16,17,18,19], ground granulated blast-furnace slag [20], and recycled powder [21,22], as waste by-products from acetylene production, blast-furnace ironmaking, and building demolition projects respectively, achieve enhanced reactivity through synergistic activation. Therefore, the development of phosphogypsum-based geopolymer as cementitious materials via the combination of phosphogypsum with multiple solid wastes holds great significance for their resource utilization.

Extensive existing studies focus on the application of solid waste-based geopolymer in 3D printing pastes. The technical challenges of phosphogypsum 3D printing focus on three dimensions: impurity interference, slurry performance regulation, and process adaptation. At the impurity level, phosphate and fluoride in phosphogypsum not only cause abnormal setting [23] but also affect the stability of hydration products, leading to long-term strength degradation [24]. The corresponding solutions include neutralization with quicklime, calcination pretreatment (175 °C), and compounding with supplementary cementitious materials (ground granulated blast-furnace slag, fly ash), which improve material stability through the dual effects of chemical neutralization and hydration promotion. In terms of slurry performance, the balance between fluidity and setting time is the core contradiction; excessive addition will lead to decreased fluidity [24,25], while unstable dehydration reaction will affect slurry uniformity [26]. Researchers have improved the printing adaptability of slurry by controlling the phosphogypsum content (2.5–7.5 wt%) and optimizing the water–binder ratio (0.67) and water temperature (15–20 °C). In terms of process adaptation, geopolymer rheological properties must satisfy the extrusion and molding requirements of 3D printing, with setting and hardening time matched to printing speed [27]. Panda B et al. [28] utilized geopolymer Bingham parameters to characterize extrudability, identifying a yield stress range of 0.6–1.0 kPa as optimal for slurry extrusion. Alghamdi H et al. [29] adopted the Benbow–Bridgewater model, determining a shear yield stress of 700 Pa as the upper limit of extrudability.

In the present study, phosphogypsum serves as the primary raw material, compounded with solid wastes including calcium carbide slag, ground granulated blast-furnace slag, and recycled powder. Mix proportions undergo design based on response surface methodology, with optimal selection conducted via comprehensive consideration of indicators such as compressive strength, spread diameter, spread diameter after vibration, and slump after vibration. Rheological properties and slurry water content states receive investigation centered on the optimized mix proportion. The feasibility of 3D printing undergoes comprehensive evaluation through indicators including extrudability, open time, and constructability. The innovations of this paper are as follows: 1. The cementitious material is prepared entirely from solid wastes without cement incorporation, which significantly reduces carbon emissions. 2. The synergistic complementary relationship and interaction mechanism among different solid wastes are explored, which not only realizes the resource utilization of solid wastes but also provides a certain theoretical basis for the feasibility of using solid waste-based materials as 3D printing materials.

2. Materials

The phosphogypsum (PG) used in the experiment was supplied by Hubei Sandi Environmental Protection New Materials Co., Ltd., Yichang, China and the Calcium carbide slag (CS) was sourced from Henan Haohua Yuhang Chemical Co., Ltd., Jiaozuo, China. Both materials were ground in a ball mill for 30 min prior to application. The Regenerated micro-powder (RCP) was derived from waste concrete collected from a construction site in Zhengzhou; it was produced via crushing in a jaw crusher, separation of coarse aggregates, and subsequent grinding in a ball mill for 1 h. The ground granulated blast-furnace slag (GGBS) was purchased from Wuhan Wuxin New Building Materials Co., Ltd. The chemical composition of the raw materials is shown in

Table 1. Chemical composition of PG, GGBS, CS and RCP (%).

	SiO2	Al2O3	Fe2O3	CaO	MgO	TiO2	Na2O	K2O	P2O5	SO3	Loss
PG	5.54	0.88	0.82	37.24	0.13	0.09	0.07	0.14	1.69	52.33	0.07
GGBS	29.83	12.38	0.47	46.39	6.25	0.93	0.37	0.60	-	1.86	0.92
CS	3.23	1.08	0.24	67.9	0.12	0.32	0.13	1.1	-	-	25.88
RCP	30.6	9.2	4.2	49.5	1.7	0.8	-	-	-	1.4	2.6

, and the XRD patterns of the raw materials are presented in Figure 1.

3. Optimal Ratio Design of Phosphogypsum-Based Geopolymer Response Surface Methodology

In this study, the formulated system (CS + GGBS + RCP + PG) belongs to the category of calcium-rich alkali-activated materials. It is referred to as “geopolymer” in this study in view of its dominant geopolymerization reaction during curing and its consistent performance characteristics with geopolymer.

3.1. Experimental Methods

The mix ratio of this experiment is based on the Box–Behnken principle of the response surface methodology [30]. CS/(CS + RCP), GGBS/(CS + RCP + GGBS), and PG/(CS + RCP + GGBS + PG) are selected as influencing factors, which are respectively denoted as A, B, and C. The experimental mix ratio was designed using the software Design-Expert 13, and the software automatically generated 17 sets of response surface methodology mix ratios. Calculate the percentage dosage of each cementitious material, select a water-binder ratio of 0.4, aim to compare the effects of phosphogypsum in different geopolymer mix proportion systems and determine the 17 combinations of phosphogypsum-based geopolymer as shown in

Table 2. Phosphogypsum-based geopolymer mix ratio.

Group	PG (%)	GGBS (%)	CS (%)	RCP (%)	W/b
P1	10	22.5	0	67.5	0.4
P2	10	22.5	67.5	0	0.4
P3	10	67.5	0	22.5	0.4
P4	10	67.5	22.5	0	0.4
P5	0	50	0	50	0.4
P6	0	50	50	0	0.4
P7	20	40	0	40	0.4
P8	20	40	40	0	0.4
P9	0	25	37.5	37.5	0.4
P10	0	75	12.5	12.5	0.4
P11	20	20	30	30	0.4
P12	20	60	10	10	0.4
P13	10	45	22.5	22.5	0.4
P14	10	45	22.5	22.5	0.4
P15	10	45	22.5	22.5	0.4
P16	10	45	22.5	22.5	0.4
P17	10	45	22.5	22.5	0.4

.

The compressive and flexural strengths of the cement mortars were tested in accordance with GB/T 17671-2021 [31]; 17 groups of proportions were prepared into 40 mm × 40 mm × 40 mm size slurry test blocks at room temperature, and then sealed and cured with cling film for 24 h. After demolding, the bagged sealing method is continued for curing until the specified age is reached. Then, a pressure testing machine with a maximum pressure output of 300 kN is used, and the loading rate is set at 2400 ± 200 N/s to measure its compressive strength.

To investigate the influence of factor interaction on the fluidity of alkali-excited phosphogypsum ground geopolymer slurry, in this experiment, the pre-vibration expansion diameter, post-vibration expansion diameter, and vibration subsidence amount were taken as response values and marked as R1, R2, and R3 respectively. The uniformity of the concrete admixtures was evaluated in accordance with GB/T 8077-2023 [32]. The well-mixed slurry is quickly poured into a truncated cone die with an upper diameter of 36 mm, a lower diameter of 60 mm, a height of 60 mm and a smooth and seamless inner wall. It is then placed on a jumping table. While lifting the truncated cone die vertically, the stopwatch is started to allow the slurry to flow freely on the glass plate for 30 s. Finally, measure the maximum diameters of the flowing part in two vertical directions with a ruler, with the recording unit being millimeters (mm), and calculate the average value as the fluidity value of the slurry, that is, the expanded diameter before vibration. After the slurry no longer flows freely, it is vibrated 25 times by jumping the table. Then, the maximum diameters in the two vertical directions are measured again with a ruler, and the average value is calculated as the expanded diameter of the slurry after vibration. The vibration subsidence amount is the height difference in the slurry before and after vibration.

3.2. Compressive Strength

As shown in Figure 2, the activation of GGBS reactivity is highly dependent on an alkaline environment, and CS serves as the core alkaline source in this system. Groups P1, P3, P5, and P7 were not incorporated with CS, resulting in a lack of sufficient OH⁻ in the system to break down the vitreous structure of GGBS. This hinders the dissolution of reactive SiO₂ and Al₂O₃, thus preventing the formation of C-S-H gel or geopolymer gel through polycondensation reactions. Therefore, most of the mixtures failed to form cementitious products and develop strength at both the 7-day and 28-day curing ages. Group P3 had a GGBS content as high as 67.5%, representing the highest proportion of reactive components among all mixtures. Meanwhile, the low dosage of RCP (22.5%) effectively mitigated the dilution effect of inert components. Under such conditions, GGBS was able to undergo a slow potential hydration reaction by virtue of the trace alkali metal oxides in its vitreous structure. A small amount of C-S-H gel was gradually generated after 28 days of curing, thereby endowing the mixture with a certain level of strength. When a certain amount of carbide slag is incorporated, the strength is significantly improved, while the strength improvement effect weakens when the content of carbide slag exceeds 30%. The strength of alkali-activated cementitious materials mainly comes from the hydration reaction of the cementitious materials to generate substances with relatively high strength after hardening, such as C-S-H gels. CS, acting as an alkaline activator, and RCP achieve synergistic activation of the reactivity of GGBS (P9–P11, P13), facilitating the proceeding of its hydration reaction and thus enabling strength development [33]. Moreover, GGBS is rich in reactive SiO₂, Al₂O₃ and CaO components. In the alkaline environment provided by the alkali activator, its vitreous structure dissociates rapidly, releasing a large number of silicate–aluminate tetrahedral ions. These ions form a three-dimensional network of amorphous aluminosilicate gel through polycondensation reactions, which constitutes the core skeleton for the strength of the system [34]. When the content of GGBS exceeds 70% (P10), the proportion of alkali activator will be reduced accordingly, which fails to fully activate the reactivity of GGBS and thus leads to the decrease in early strength. RCP fills the voids between other cementitious materials, playing a certain physical filling role, making the material structure more compact. Moreover, RCP particles feature abundant angularities and irregular edges [35], forming a relatively stable structure with C-S-H gels, etc., which to a certain extent also helps to improve the early strength of the material (P13–P17). However, if the dosage of RCP is too high (P9, P11) and it replaces the alkali-activated material, CS, it will result in the inability to provide a sufficient alkaline environment, thus preventing the hydration reaction from occurring [36]. When 10% PG was incorporated, the strength of the paste was improved (P4, P13–P17). This is because an appropriate amount of PG in this system can generate ettringite to form crystal nuclei, which has a strength-enhancing effect, while the strength decreased significantly when the compound content of PG with other solid wastes reached 20% (P12). Excessive PG will dilute the proportion of reactive aluminosilicate components in the system, reduce the concentration of effective reactants for the alkali-activation reaction, decrease the yield of hydration products, and thus fail to form a dense matrix structure.

Based on the strength analysis, the proportion of each component in the cementitious system is determined as follows: CS: 15%~25%; GGBS: 30%~45%; RCP: 15%~25%; PG: 10%~15%.

3.3. Analysis Based on the Response Surface Method of Liquidity

Based on the test data using the Design-Expert 13 software, the response surface diagram of the influence of factor interaction on the three response values was drawn and analyzed, as shown in Figure 3, Figure 4 and Figure 5.

As shown in the graph, in terms of a single factor, the influence of CS on the expansion diameter before and after vibration and the amount of vibration settlement far exceeds the other two factors, and its influence is very significant. This is because when CS content is high, as an alkaline activator, it causes the pH value of the slurry to rise, promoting the progress of hydration reactions, while consuming a large amount of water, thereby reducing the fluidity of the slurry. When the dosage of GGBS is 20% to 40%, it shows a promoting effect on the fluidity of the slurry. When the dosage exceeds 40%, its fluidity shows a downward trend. This is because GGBS requires the Ca(OH)₂ generated by the hydration of cement to activate its activity. The early hydration rate is much lower than that of cement. When the dosage is low, it reduces the heat release and setting speed of cement hydration, releases more free water and improves fluidity. When the dosage is high, due to the lag of hydration, the system lacks sufficient hydration products to lubricate the particles, and more water is needed to meet the wetting requirements of the particles, which is prone to cause deterioration of fluidity.

For PG, when the dosage ranges from 0 to 10%, its fluidity improves to a certain extent. However, when the dosage rises to 20%, its fluidity performance declines. This is because on the one hand, PG particles fill the voids between the particles, and on the other hand, CaSO₄·2H₂O in PG reacts with tricalcium aluminate (C₃A) to form ettringite (AFt). AFt covers the surface of the cementitious material particles to form a coating layer, which hinders the further hydration of the cementitious material, thereby increasing the free water content of the system and making the slurry flow more easily. When the dosage of PG is too high, due to its low activity, it will increase the friction between material particles, raising the yield stress of the slurry and being unfavorable for the flow of the slurry.

4. Analysis of Printability of Phosphogypsum-Based Geopolymer

Based on the comprehensive analysis results of strength and response surface methodology, as well as the observation of the flow state of fresh paste, six mix proportions shown in

Table 3. Experimental mix ratio.

Group	PG (%)	GGBS (%)	CS (%)	RCP (%)	W/b
D1	10	67.5	22.5	0	0.4
D2	0	50	50	0	0.4
D3	20	40	40	0	0.4
D4	0	75	12.5	12.5	0.4
D5	20	20	30	30	0.4
D6	10	45	22.5	22.5	0.4

were optimized for subsequent experiments.

The standard consistency water requirement and setting time of phosphogypsum geopolymers were determined in accordance with "Test Methods for Water Requirement of Normal Consistency, Setting Time and Soundness of Cement" [37]. According to the slurry preparation process of phosphogypsum-based geopolymer, the yield stress and plastic viscosity of phosphogypsum-based geopolymer were determined by using the RST-SST rheometer of Brookfield Company, MA, USA. The test program first maintains a shear rate of 100 S⁻¹ for 1 min and then reduces it to 0 S⁻¹ to pre-shear the slurry. Subsequently, the shear rate gradually increases from 0 S⁻¹ to 10 S⁻¹ within 1 min, and then rapidly increases from 10 S⁻¹ to 100 S⁻¹ within 1 min. The decrease and increase rates change in the same way. The hydration process of the fresh slurry was tested by low-field nuclear magnetic resonance equipment (LF-NMR), and the distribution of internal moisture in the early stage of the fresh slurry was observed and analyzed.

This experiment used the NELD-3D736 printer. The printing parameters are as follows: the nozzle diameter of the 3D printing extruder is 10 mm; software parameter settings: the layer height of the printed model is set at 6 mm, the height of the first layer is 60% of the layer height, the default printing speed is 28 mm/s, the movement rate is 100%, the extrusion ratio is 65%, and the filament diameter is fixed at 15 mm. The filament width after deposition is intentionally wider than nozzle diameter due to flattening. The method for testing the indicators is as follows.

Extrudability: By printing a hexagonal single-line model with an inscribed circle radius of 65 mm, a line width of 15 mm, and a height of 6 mm, observe whether the lines break. If no breaks occur, the extrusion continuity is met. Otherwise, this observation demonstrates inferior extrudability of the slurry. Under the premise of meeting the extrusion continuity, the Coefficient of extrusion is calculated through Formula (1).

$${\mathrm{C}}_{\mathrm{j}}={\displaystyle \frac{\mathrm{S}}{\mathrm{x}}}\times 100\mathrm{\%}=\sqrt{{\displaystyle \frac{{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{({\mathrm{x}}_{\mathrm{i}}-{\overline{\mathrm{x}}}_{\mathrm{m}})}^2}{\mathrm{n}}}}\times {\displaystyle \frac{\mathrm{n}}{{\mathrm{x}}_1+{\mathrm{x}}_2+\cdots +{\mathrm{x}}_{\mathrm{n}}}}\times 100\%$$

(1)

(${\mathrm{C}}_{\mathrm{j}}$ is the Coefficient of extrusion; $\mathrm{s}$ is the width standard deviation; $\mathrm{x}$ is the average width; n = 6).

Open time: Starting from the moment water begins to come into contact with the rubber material, through continuous printing of a circular ring model with an inner radius of 65 mm, an outer radius of 80 mm, and a height of 6 mm, the printing interval is 5 min until the printer can no longer extrude the slurry normally.

Constructability: Print six layers of 6 mm high circular cylindrical models. After printing is completed, record the actual height of the model at 0 min, 30 min and 60 min respectively. Calculate its rate of height change through Formula (2).

$$∆\mathrm{h}\prime =(1-{\displaystyle \frac{\mathrm{H}\prime }{\mathrm{H}}})\times 100\mathrm{\%}$$

(2)

($∆\mathrm{h}\prime$ represents the rate of height change of 3D printed products; ${\mathrm{H}}^{\prime }$ represents the actual height of the printed model; $\mathrm{H}$ represents the height of the printed model).

4.1. Performance Analysis of Fresh Slurry

The test results of water consumption for standard consistency and setting time are shown in Figure 6 and Figure 7.

As shown in the graph, when the dosage of CS is high (D2, D3), the overall water consumption for standard consistency is relatively high. CS, as an alkaline activator, provides sufficient Ca(OH)₂ for the hydration reaction when its content is high, promoting the progress of the hydration reaction. At the same time, it consumes a large amount of water, thereby increasing the water consumption [38]. Compared with the dosage of PG, PG reacts with GGBS, CS and other active components to generate ettringite and other hydration products, and the formation process of these products consumes a large amount of water [39], thereby increasing the water consumption of the system (D3). In terms of setting time, compared with (D1, D2, D3) and (D4, D5, D6), the initial setting time of the former is generally shorter. This is because the content of CS is relatively high, which increases the pH value of the geopolymer slurry. At the initial stage of hydration, it promotes the hydration reaction of PG and GGBS, generating hydration products, and thus accelerating the rapid setting of the slurry [40]. When the dosage of RCP increases (D4 > D6 > D5), the hydration activity of the RCP itself is relatively low, and the hydration reaction is slow, thereby delaying the setting time of the entire material system. When comparing the dosage of GGBS, if the proportion of GGBS is relatively high (D1, D4), the initial setting time will be prolonged. This phenomenon arises from the incomplete activation and utilization of the latent reactivity of ground granulated blast-furnace slag (GGBS) at the early stage, accompanied by the formation of limited hydration products, which consequently prolongs the initial setting time. By comparing the dosage of PG, excessive dosage of PG significantly shortens the setting time (D3, D5). The CaSO₄·2H₂O in PG and the main component Ca(OH)₂ of CS stimulate the potential hydraulic properties of the GGBS, promoting the rapid setting of the slurry.

4.2. Analysis of the Rheological Properties of Fresh Paste

The rheological behaviors of fresh phosphogypsum-based geopolymer slurry match those of fresh cement slurry, with both systems allowing derivation of relevant rheological parameters via simulation through corresponding rheological models. To more intuitively present the experimental patterns in Table 3, this study averaged the data during the acceleration stage and selected 10 representative data points to analyze the variation patterns of shear stress and apparent viscosity.

As shown in Figure 8, by comparing the shear stress and apparent viscosity changes in the alkali-excited phosphogypsum-based geopolymer slurry under different mix ratios, the shear stress of the slurry tends to stabilize and is closely distributed near the fitting curve. For Figure 8a, when the content of CS is high and the shear rate is the same, the shear stress is relatively high, indicating that when the content of CS is large, the viscosity of the slurry is also high (D3). This is because the hydration activity of the RCP is low, while the activity of CS is relatively high. CS provides a large amount of Ca(OH)₂ for the hydration reaction of materials such as GGBS, accelerating the progress of the hydration reaction and forming more cementitious substances such as calcium silicate hydrate, which increases the viscosity of the slurry and is manifested as a large shear stress. However, as the dosage of GGBS decreases (D1, D2, D3), the shear stress increases at the same shear rate. This phenomenon stems from the fact that GGBS, as a latent hydraulic material, achieves full activation of its reactivity exclusively in an alkaline environment, with its hydration reaction proceeding at a relatively slow rate.

Compared with D1, D3, D5, D6, D2, and D4, the proportion of PG is different. The overall shear stress and apparent viscosity of the slurry with PG added are higher than those of the slurry without PG. On the one hand, PG undergoes a relatively sluggish hydration reaction while exhibiting distinct hygroscopicity. Increasing the content of PG will inhibit the hydration process of cementitious materials, and at the same time reduce the early fluidity of the slurry, resulting in an increase in the shear stress and viscosity of the slurry (D1, D6). In addition, PG provides raw materials for the hydration reaction to generate substances such as ettringite, and alters the internal microstructure, making the structure more stable, enhancing the resistance to shear deformation, and increasing shear stress and apparent viscosity (D3, D5).

4.3. Analysis of Moisture Migration in Paste Based on Low-Field Nuclear Magnetic Resonance

The hydration process of fresh slurry was tested by low-field nuclear magnetic resonance. The test results of low-field nuclear magnetic resonance of slurry with different mix ratios are shown in Figure 9.

The three peaks in D1 correspond respectively to these three different types of H atoms and also to three different types of pores, namely gel pores, microcapillary pores and macrocapillary pores. It is generally believed that the water in the gel pores and small capillary pores is adsorbed water, while the water in the large capillary pores is free water. The relaxation times of these pores are respectively 0.01 ms ≤ T2 ≤ 1 ms, 1 ms ≤ T2 ≤ 10 ms, and 10 ms ≤ T2 ≤ 1000 ms. The first and second peaks gradually move to the left over time, and the peak signals also gradually weaken. It indicates that the gel pores are gradually filled by hydration products as hydration proceeds, and the pores become smaller and smaller. The main peak gradually moves to the left as the hydration reaction proceeds, and the free water is gradually consumed. The main peak of D2 gradually shifts to the left over time, and the peak signal gradually strengthens, but the trend of movement is smaller than that of D1. This is because the free water in the slurry is not only absorbed by the hydration reaction, but also due to the extremely strong hygroscopicity of PG itself. A single characteristic peak appears for D3, a phenomenon attributed to the relatively low reactivity of phosphogypsum (PG). An increase in PG content decelerates the hydration process. Over time, the peak signal exhibits a pronounced attenuation trend, a result that further verifies the remarkable water-retention capacity of PG. Excessive PG will consume a large amount of free water in the slurry and accelerate its setting and hardening. The free water in D4 gradually decreases over time. This is because the addition of RCP promotes the activation effect of CS on the GGBS and accelerates the hydration rate. When the dosage of D5 PG increases, the free water content decreases and the peak signal weakens. Due to the excessive dosage of PG and CS, it had a counterproductive effect on the hydration process of the system, and there was no obvious change in the secondary peak. The main peak of D6 shows a relatively gentle trend over time compared to D5. The peak signal gradually weakens first, and the free water content decreases. The secondary peak shows no obvious trend of change, and the hydration rate is not much different from that of D5.

4.4. Analysis of Printability

The results of the printable performance experiment are shown in Figure 10 (D3 is not printable).

During the experiment, multiple breaks occurred in the lines printed with the D3 mix ratio. This was due to the high dosage of PG, which excessively generated crystals such as ettringite, significantly increasing the viscosity of the slurry. The yield stress was 128.4 Pa, and the extrusion resistance exceeded the driving force of the printer, clearly not meeting the extrudability requirements. Therefore, it was not printable. As shown in Figure 11,The remaining five groups all extruded continuously, meeting the requirements of extrusion continuity. The uniformity coefficients of extrusion in the remaining five groups were all less than 15% through measurement and calculation, indicating that they exhibit favorable extrudability. A few cracks appeared in the D2 printing. The excessive amount of CS caused the cementitious material to undergo a rapid hydration reaction, resulting in the production of a large amount of calcium silicate hydrate (C-S-H gel), etc. The viscosity of the slurry was high and the distribution was uneven. Due to the excessive GGBS content and insufficient early hydration of D4, the slurry undergoes severe deformation due to its own weight, with a height change rate as high as 45.28%. Therefore, its constructability is extremely poor. The height change rates of the other four groups were all less than 15%, indicating that they exhibit favorable constructability. Analysis reveals, on the one hand, that the high content of CS stimulates the GGBS to form C-S-H gel (D1, D2). On the other hand, the main component of PG, calcium sulfate dihydrate, reacts to form needle-rod-shaped ettringite crystals, which form a more stable and compact structure with C-S-H and other gels, enhancing the anti-deformation ability (D5, D6). Moreover, the open time of all four groups was more than 30 min, meeting the printing requirements. Analysis shows that when the dosage of CS is relatively high, the open time is significantly shorter (D5). This is because a high dosage of CS accelerates the progress of the hydration reaction, enabling the slurry to harden rapidly. However, RCP inherently exhibits relatively low hydration reactivity and shows limited tendency toward hydration reactions. When the dosage of RCP increases, it will slow down the hydration reaction rate of the entire material system (D6). The higher the dosage of PG, the shorter the open time of the slurry. This is because the CaSO₄·2H₂O and Ca(OH)₂ in PG stimulate the potential hydraulics of the slag, and the PG rapidly undergoes water formation (D5).

Therefore, an appropriate amount of PG is conducive to enhancing the buildability of the paste. A proper content of CS and GGBS can improve the extrudability of the paste. The incorporation of RCP optimizes the printable time.

5. Conclusions

In this research, we compounded GGBS, CS, RCP, and PG and evaluated the 3D printing performance of the selected mix ratio samples. The main conclusions are as follows.

(1)

Phosphogypsum (PG), compounded with calcium carbide slag, recycled powder, and ground granulated blast-furnace slag, fabricates geopolymer with 28-day compressive strength exceeding 25 MPa, thus demonstrating feasibility for application as 3D printing materials.

(2)

Phosphogypsum (PG) engages in synergistic reactions with calcium carbide slag and ground granulated blast-furnace slag to form ettringite, thereby optimizing the internal microstructure of the geopolymer. Excessive PG addition, however, depletes substantial free water within the slurry, inducing elevated shear stress and viscosity, as well as diminished slurry fluidity.

(3)

A moderate dosage of PG enhances the later-stage structural stability of the slurry and promotes the printability of 3D printing systems. Excessive PG addition, in contrast, shortens the slurry open time, thus impeding slurry extrusion and molding. The optimal PG dosage for 3D printing material preparation falls approximately at 10%.

(4)

Based on a comprehensive analysis of printability and the performance of printed specimens, the optimal mix proportion of the phosphogypsum-based geopolymer 3D printing paste was determined as follows: CS: 22.5%; GGBS: 45%; RCP: 22.5%; PG: 10%; W/b: 0.4.