Development of a Geopolymer for 3D Printing Using Submerged Arc Welding (SAW) Slag

Abstract

Reducing the carbon footprint of the construction sector is a growing priority. This study explores the potential of using submerged arc welding (SAW) slag as a precursor in the development of low-carbon geopolymeric materials for 3D printing. The influence of potassium hydroxide (KOH) molarity, partial replacement of ground granulated blast furnace slag (GGBFS) with SAW slag, and water-to-binder (w/b) ratio was evaluated in terms of fresh and hardened properties. Increasing KOH molarity delayed setting times, with the longest delays at 10 M and 12 M. The highest compressive strength (48.5 MPa at 28 days) was achieved at 8 M; higher molarities led to strength losses due to excessive precursor dissolution and increased porosity. GGBFS replacement increased setting times due to its higher Al₂O₃ and MgO content, which slowed geopolymerization. The optimized formulation, containing 20% SAW slag and activated with 8 M KOH at a w/b ratio of 0.29, exhibited good workability, extrudability, and shape retention. This mixture also performed best in 3D printing trials, strong layer adhesion and no segregation, although minor edge irregularities were observed. These results suggest that SAW slag is a promising sustainable material showing for 3D-printed geopolymers, with further optimization of printing parameters needed to enhance surface quality.

1. Introduction

Interest in 3D printing within the construction industry has been steadily increasing in recent years [1,2]. Compared to conventional construction methods, 3D printing offers various benefits such as reducing costs, saving time and materials, minimizing waste, providing design flexibility, and improving overall efficiency and productivity [3,4,5].

Despite the benefits of this technique, it currently faces the challenge of reducing the amount of Portland cement consumed, since it is the most commonly used material in both traditional construction and 3D printing. It is valued for its good compressive strength, workability, low cost, and durability [6,7,8]. However, it has some drawbacks, such as CO₂ and solid particle emissions, water pollution, noise generation, and waste. Cement production alone contributes about 33% of global CO₂ emissions [9]. Therefore, many studies have been conducted to reduce or eliminate cement in construction materials, adopting alternative binders like incorporating industrial by-products such as fly ash and slag or geopolymers, which are crucial for achieving sustainable and low-carbon building practices [10,11,12]. The utilization of by-products in geopolymer materials offers the potential to decrease CO₂ emissions by up to 80% in comparison to Portland cement [13].

Geopolymers, also known as alkali-activated materials, are inorganic polymers synthesized through the reaction of aluminosilicate materials and alkaline activators [14]. Therefore, industrial wastes rich in aluminosilicates, such as fly ash (FA) and -ground granulated blast-furnace slag (GGBFS) are used as the most common precursors.

There are also other by-products, rich in silica and alumina, which can be used as precursors for geopolymers, such as biomass ash or rice husk ash [11,15,16,17,18]. Other industrial waste could be submerged arc welding (SAW) slag. SAW is a process in which heat is provided by an electric arc generated between one or more electrodes and the workpiece.

This electric arc is immersed in a granular flux layer that covers it completely, protecting the metal deposited during welding. Once the process is completed, the modified flux material, known as submerged arc welding (SAW) slag, is collected, crushed, and sieved to obtain particles comparable in size to the original flux.

The generation of SAW slags has increased over time. An example of this is India, where about 2500 t of weld slag was generated during the year of 1982 which rose to 10,000 t during the year 2006 [19].

In particular, SAW slags have been investigated as a partial replacement for fine aggregates in mortar [20,21,22]. Experimental studies have shown that compressive strength is maximized when 10% of fine aggregates are replaced by SAW slag. Likewise, the replacement of up to 15% sand with GGBFS or SAW improves resistance at an early age [19,21]. In addition, the workability of the concrete is not affected with substitution levels of up to 15% and water absorption remains stable with 10% SAW slag in concrete [22].

The use of welding slag in concrete not only improves certain mechanical properties but also brings economic and environmental benefits. The incorporation of this industrial waste in construction makes it possible to reduce costs, minimize waste disposal, and promote the circular economy. In this context, SAW slag is presented as a viable resource for the manufacture of building materials, contributing to more sustainable and efficient construction.

This paper goes a step further by presenting the development of a 3D printing geo-polymer in which SAW slag is used not merely as an aggregate replacement, as in previous studies, but as a precursor in the binder phase. This work introduces a novel application of welding slag that has not yet been explored in the context of additive manufacturing. The developed material fulfills the fresh state requirements of 3D printing, such as consistency, extrudability, continuity, and buildability, while also exhibiting adequate setting times and mechanical performance.

2. Materials and Methods

Submerged arc welding (SAW) slag was procured from Castolin Eutectic SL, Madrid, Spain, and ground granulated blast-furnace slag (GGBFS) was procured from Cementos Levante from Cartagena, Spain. Both materials were used as precursors. The precursor was activated with alkali hydroxide of KOH with 97% purity (ITW Reagents, Barcelona, Spain). The properties of the alkali activator provided by the manufacturer are given in

Table 1. Properties of potassium hydroxide.

Properties	Potassium Hydroxide
Chemical formula	KOH
Molecular weight (g/mol)	56.11
Appearance	Solid
Relative density (g/cm3)	2.04

. Sand used in the development of the geopolymer 3D printed is a CEN standard sand according to EN 196-1 [23].

The SAW slags and GGBFS fineness were measured using Blaine measurements according to EN 196-6 [24]. GGBFS fineness was obtained after 3 h in a ball mill (Matest, model D0544/A) at a constant speed of 60 rpm. The values of density and Blaine from SAW slag and GGBFS are shown in

Table 2. Properties of SAW slag and GGBFS.

Properties	SAW Slag	GGBFS
Density (g/cm3)	3.1	2.9
Blaine (cm2/g)	4968	4068

.

Chemical analyses were performed using an S8 Tiger Bruker spectrometer (XRF), from Malvern Panalytical (Madrid, Spain). The chemical composition of all the used materials has been presented in

Table 3. Chemical composition of the main oxides in SAW slag and GGBFS (wt.%).

Oxides (%)	Na2O	MgO	Al2O3	SiO2	K2O	CaO	MnO	TiO2	Fe2O3	P2O5	SO3	LOI
SAW slag	2.43	15.71	19.55	20.99	1.09	18.22	8.48	2.76	2.77	0.03	0.05	7.92
GGBFS	0.17	4.26	13.58	30.83	0.35	46.58	0.15	0.65	0.87	0.03	1.63	0.90

. The composition of the SAW slag and GGBFS given in Table 3 shows that SiO₂, Al₂O₃, and CaO were the major compounds. Significant amounts of MgO were also identified.

The alkaline solution was prepared the day before by thoroughly mixing distilled water and KOH pellets and then allowed to cool in laboratory conditions (temperature: 20 ± 2 °C and humidity relative: 35 ± 5%). Four different molarities, 6 M, 8 M, 10 M, and 12 M were used to prepare alkaline solutions.

SAW slag, GGBFS, and sand were mixed. Once the mixture was homogenized, an alkaline solution was poured into the mix twice. The mixing continued for about 3 min until a homogeneous mixture was achieved. The prismatic specimens of dimensions 160 × 40 × 40 mm³ were cast for the mechanical properties. The fresh mortar mixes were cast into the standardized molds and covered with a plastic film to avoid evaporation and carbonation of the alkaline solution. The specimens were demolded 24 h after casting and stored at 21 ± 2 °C and 95 ± 2% relative humidity in a controlled chamber. The mixed proportions of the alkali-activated slag mortar (AASM) specimens are presented in

Table 4. Mix proportions (binder in wt.%, other components in parts by weight).

Samples	GGBFS	SAW Slag	Sand	KOH (M)	w/b
S1-1	100	0	124	6	0.37
S1-2	100	0	124	8	0.37
S1-3	100	0	124	10	0.37
S1-4	100	0	124	12	0.37
S2-0	100	0	124	8	0.32
S2-1	90	10	124	8	0.32
S2-2	80	20	124	8	0.32
S2-3	70	30	124	8	0.32
S2-4	60	40	124	8	0.32
S2-5	50	50	124	8	0.32
S3-1	90	10	124	8	0.29
S3-2	80	20	124	8	0.29
S3-3	70	30	124	8	0.29

. GGBFS and SAW slag are expressed as relative weight percentages of the binder (total = 100 wt.%). Sand and KOH are given in parts by weight with respect to the binder; w/b denotes the water-to-binder ratio. The geopolymer specimens were designated depending on the molarity and the SAW slag ratio. For each test, three specimens were prepared and evaluated.

This study investigated the influence of formula modifications compared to the reference one in three Series (S1, S2, and S3) regarding their material properties (mechanical properties and setting time). In Serie 1, the influence of KOH molarity was studied, molarity was increased from 6 M to 12 M. In Serie 2, the influence of SAW slag was studied. In addition, in Serie 3, the optimization of the water-to-binder (w/b) ratio was studied regarding the 3D printing properties.

The tests carried out on the dosages in Table 4, both in the fresh and hardened state, are shown below.

Flow test

The flow of the mixes was determined according to EN 1015-3 [25], which specifies a method for measuring the consistency of fresh mortars by means of a flow table. In this test, the spread diameters of the fresh mortar mixes were measured along four perpendicular lines on the table surface.

Setting time

The setting time of paste samples was evaluated following EN 1015-6 [26], which establishes procedures for determining both the initial and final setting times of mortars by means of a Vicat needle apparatus.

Mechanical properties

The compressive and flexural strengths of the specimens were determined at 5, 7, and 28 days of hydration using a Matest Jolting Apparatus (Model E130), in accordance with EN 196-1 [23]. The flexural tests were performed at a loading rate of 5 kg/s, while the compressive tests were carried out at a loading rate of 240 kg/s. This standard describes the reference method for testing cement mortars, using prismatic specimens (160 × 40 × 40 mm³), and prescribes the procedures for flexural and subsequent compressive loading of the halves. For each age, three specimens were tested and the average strength values were recorded.

Extrudability and buildability

The ability of a geopolymer to be extruded and to maintain its shape (buildability) is a critical requirement for 3D printing applications. To evaluate the capacity of the mixes to retain their geometry during successive layer deposition, a shape retention test was performed, adapted from the procedure proposed by B. Nematollahi et al. and F. Fernández et al. for geopolymers and ECC mortars [27,28]. In this method, the fresh mortar was first placed into a mini-slump cone, which was removed after one minute to define the initial geometry. A static load of 350 g was then applied for 45 s, allowing the material to deform. Additional weights were successively added until a cumulative load of 2000 g was reached. After each loading step, the spread of the mortar was recorded along two perpendicular axes. A smaller spread diameter indicates a higher capacity of the mixture to maintain its shape.

In this study, the 3D printing parameters for TRL 1 to 3 were explored. A manual extrusion technique was used as an approximation technique for the 3D printing process. A caulk gun was used during the process (TRL 1-2), with a rectangular geometry nozzle (4 mm × 2 mm), as in D. G. Soltan and F. Fernández et al. [28,29].

3D printing test

To carry out the next step to the scale up, the 3D printing test was carried out in a 3D printed “SMART 2500” model from Be More 3D (Valencia, Spain) (Figure 1). The printer consists of a gantry structure with dimensions of 2.5 by 3.0 by 2.5 m and a hopper with a worm screw that pushes the material into the extruder. The 3D printer was equipped with an extruder with a 40 mm circular nozzle. The printing speed was 85 mm/sec, and the time between layers was 37 s. The height of the printed figure is 20 cm with a layer width of 6 cm and a layer height of 1 mm.

3. Results

3.1. Influence of KOH Molarity

The setting time and mechanical properties of the Serie 1 mixes (S1-1, S1-2, S1-3, and S1-4) were studied.

The initial and final setting time results of the mixtures are shown in Figure 2. The setting time of the geopolymer is primarily affected by the type and concentration of activators and the water-to-binder ratio [30]. The initial setting time increased from 19 min at 6 M (S1-1) and 8 M (S1-2) to 22 min at 10 M (S1-3) and 25 min at 12 M (S1-4). Similarly, the final setting time showed an increasing trend, with values of 34 min for S1-1, 39 min for S1-2, and 52 min for both S1-3 and S1-4. These results indicate that higher KOH molarity delays the setting process.

The compressive strength of a geopolymer activated with potassium hydroxide (KOH) at different molar concentrations was evaluated at 5, 7, and 28 days, Figure 3. At 5 days, the strength ranged from 27.9 MPa (S1-4) to 33.55 MPa (S1-2), while at 7 days, it varied between 30.9 MPa (S1-4) and 37.45 MPa (S1-2). At 28 days, the highest strength was achieved at S1-2 (48.5 MPa), followed by S1-1 (44.4 MPa), S1-3 (44.3 MPa), and S1-4 (40.2 MPa). These results indicate that 8 M KOH (S1-3) leads to the highest compressive strength development over time, whereas increasing the molarity beyond this point negatively affects mechanical performance.

A statistical analysis confirmed that the 8 M mix formed a significantly higher-performing group compared to the 10 M and 12 M mixes, while the 6 M mix showed intermediate values without significant differences from either group. These results indicate that 8 M KOH provides the most favorable balance between dissolution of precursors and geopolymer gel formation, whereas higher molarities compromise performance due to excessive dissolution and porosity.

3.2. Influence of SAW Slag

The setting time and mechanical properties of the Serie 2 mixes (S2-0, S2-1, S2-2, S2-3, S2-4, and S2-5) were studied.

The initial and final setting time results of the mixtures are shown in Figure 4. The table presents the initial and final setting times for different replacement levels of slag with welding slag in a geopolymer matrix. As the percentage of welding slag increases, both setting times tend to increase. The reference sample (2-0) exhibits an initial setting time of 45 min and a final setting time of 58 min. With partial replacement (S2-1 to S2-3), the setting times gradually increase, reaching 63 and 94 min at S2-3. At higher replacement levels (S2-4 and S2-5), the setting times increase significantly, reaching 88 and 141 min for the highest replacement (S2-5).

Figure 5 shows the compressive strength evolution of Serie 2 mixes, measured at curing ages of 5, 7, and 28 days of hydration. An increase in the compressive strength with the hydration time was observed.

At early ages (five and seven days of hydration), the compressive strength remains relatively similar across the different dosifications until the substitution of SAW slag reaches or exceeds 40% of the mix. At this point, a notable decrease in strength is observed, with dosifications S2-4 and S2-5 reaching 30 MPa and 27.5 MPa, respectively. After 28 days of hydration, the compressive strength values converge across all dosifications, ranging between 45 and 50 MPa. Notably, mix S2-3 achieves the highest strength at this age, reaching 52 MPa, but statistical analysis confirmed that the 30% replacement mix (S2-3) achieved significantly higher strength (52.4 MPa) compared to S2-4 and S2-5. At 90 days, long-term curing further enhanced performance across all mixtures, with S2-2 and S2-3 forming the best-performing group (>62 MPa), while high-replacement mixes (S2-4 and S2-5) remained in the lower group (~59–60 MPa). These results indicate that partial replacement up to 30% does not compromise strength development, whereas higher SAW content reduces early age performance, although this difference tends to diminish with long-term curing.

3.3. Optimization of w/b Ratio for 3D Printing

To ensure good rheological behavior for the 3D printing of the S2 Serie, the w/b was adjusted.

The initial w/b ratio of 0.32 was progressively reduced to 0.29 to meet the specific workability requirements of 3D printable geopolymers, which demand a balance between sufficient flowability for extrusion and adequate buildability to maintain shape after deposition.

• Flow test

The flow test results are presented in

Table 5. Flow test results of S3-1, S3-2, and S3-3.

Consistency	S3-1	S3-2	S3-3
(mm)	185	183	183

. The flow values of S3-1, S3-2, and S3-3 keep constant with the increase in SAW slag percentage, maintaining values of 180 mm.

• 3D printing parameters

• Shape retention test

The shape retention test revealed good performance of the material when depositing weight on it in a short period. Figure 6 shows the evolution of the height.

The cone of material used in the consistency test before the performance of the test is shown, with an initial height of 4.8 cm.

2.

Layer-on-layer deposition

Figure 7 shows the buildability performance of S3-2 material through the use of a caulk gun. The total height of the seven-layer printed sample measured 7 cm. As for the printing width, all layers maintained the 4 cm height.

• 3D printing test

Figure 8 shows the 3D printer used in the scaling test. The purpose of this scaling test was to assess the feasibility of printing more complex geometries at a larger scale and to evaluate the material’s behavior during deposition. For the sample, 54 kg of the S3-2 dosage was mixed.

4. Discussion

4.1. Influence of KOH Molarity

The initial and final setting time results of the mixtures are shown in Figure 2. This behavior suggests that increasing KOH concentration affects the reaction kinetics of the geopolymerization process. A higher molarity may slow down the dissolution of aluminosilicate precursors, influencing gel formation and polymerization, thereby extending the setting time. This effect is particularly noticeable at S1-3 (10 M) and S1-4 (12 M), where the final setting time reaches 52 min, indicating a prolonged structural development phase.

The compressive strength of a geopolymer activated with potassium hydroxide (KOH) at different molar concentrations was evaluated at 5, 7, and 28 days, Figure 3. High molarity dosages (10 M and 12 M), S1-3 and S1-4, can generate an excess of dissolution of the precursors, generating an alteration in the reaction kinetics and increasing the porosity of the mixtures, reducing the compressive strength. Whereas, at low molarity (6 M and 8 M), S1-1 and S1-2, the compressive strength is higher. Therefore, a higher concentration of KOH can compromise structural stability, affecting both setting time, mechanical properties, and parameters of importance for 3D printing materials.

The S1-2 (8M) dosage had an initial setting time of 19 min and a final setting time of 39 min, these values ensure sufficient workability for printing. However, any inconvenience during the printing process, either during mixing or extrusion, could cause problems such as clogging of the extruder or hose. In addition, S1-2 has a compressive strength of 48.5 MPa at 28 days of hydration, the highest value among all dosages. Therefore, S1-2 has an optimal balance between setting time and mechanical properties, being the dosage chosen for the next phase.

4.2. Influence of SAW Slag

The setting time and mechanical properties of the Serie 2 mixes (S2-0, S2-1, S2-2, S2-3, S2-4, and S2-5) were studied.

An increase in setting time is observed as the replacement of GGBFS by SAW welds increases. This can be explained by the chemical composition of both materials, see Table 3, as the precursor composition has a major impact during the geopolymerization process. SAW slag possesses a higher value of Al₂O₃ content with respect to GGBFS (19.55% vs. 13.58%), the same happens with MgO content (15.71% vs. 4.26%). Whereas, it has a much lower value in CaO and SiO₂, 18.22% vs. 46.58% and 20.99% vs. 30.83%, respectively.

This difference in chemical composition affects the geopolymerization process. For example, low CaO content reduces calcium-aluminate-silicate-hydrate (C-A-S-H) gels, which are directly involved in setting times. In addition, high Al₂O₃ content can slow down the polymerization process by increasing the availability of alumina as they compete with silica during the geopolymerization process.

Therefore, the increase in SAW slag causes a slower geopolymerization process, increasing the setting time.

The evolution of compressive strength shows the influence of varying SAW 280 slag replacement GGBFS from early hydration days (5 days) to 28 days (Figure 5).

At early hydration ages (five and seven days), the compressive strength of the S2-0, S2-1, S2-2, and S2-3 dosages indicate that a replacement of up to 30% SAW does not noticeably affect compression. However, when replacement reaches 40% and 50% (S2-4 and S2-5), there is a noticeable decrease in mechanical properties, with a drop to 30 MPa and 27.5 MPa, respectively.

However, all dosages, at 28 days of hydration, show a similar value, reaching values between 45 and 50 MPa. This may be due to the pozzolanic activity of the SAW slag compensating for the loss of compressive strength at early ages. Dosage D2-3 shows the highest compressive value with a value of 52 MPa.

These compressive results show the potential of SAW slag as a partial replacement in geopolymers. It is also worth noting that the compressive strengths obtained in this study are significantly higher than those reported for other sustainable binders, such as geopolymers derived from ceramic slurry waste consolidated by moderate thermal treatment, which reached only about 17 MPa [31].

4.3. Optimization of w/b Ratio for 3D Printing

• Flow test

The flow test results are presented in Table 5. The material has a high fluidity, reaching values above 185 mm in the three dosages studied, S3-1, S3-2, and S3-3.

Visual observation during the production handling of mixtures confirmed the good appearance of the material during the measurements. It can be concluded that the workability of S3-1, S3-2, and S3-3 is satisfactory for the final 3D application.

• 3D printing parameters

• Shape retention test

The placement of the 350 g weight in S3-2 is shown in Figure 6a, which causes a minimal decrease to 4.7 g. This height remains constant when an additional 350 g is added (see Figure 6b). This height decreases to 3.9 cm when a total of 2800 g is added (Figure 6c). The S3-2 sample shows a slight deformation, managing to positively retain the shape. Despite undergoing deformation, the stress suffered by the consistency is stronger than that which it would undergo in the 3D printing process.

2.

Layer-on-layer deposition

It should be noted that the bottom layer of the printed sample was able to hold the weight of the upper layers, allowing layer upon layer to build without any deflection. Hence, as shown in Figure 7, S3-2 material does not noticeably deform during the shape retention test, furthermore, it does not decrease the layer thickness by caulking test.

Therefore, the S3-2 material showed good workability and rapid rebuilding that allowed both extrusion and buildability with minimal deformation of layers under the weight of subsequent layers.

• 3D printing test

A hexagon is chosen as the 3D printing shape. During the printing process, the S3-2 dosage formulation shows good extrudability, as no blockage occurred at any time during its placement in the hopper, despite the setting time of the sample. In addition, the interlayer adhesion was maintained during printing, and there were no signs of segregation. However, some irregularities are shown in the deposited layers, especially in the corners, where the nozzle direction changes. The printed figure remains stable without generating excessive deformation and without collapsing.

5. Conclusions

This study explored the potential of using submerged arc wielding (SAW) slag as a precursor in the development of a geopolymer for 3D printing. The research focused on evaluating the influence of potassium hydroxide (KOH) molarity, the addition of submerged arc welding (SAW) slag, and the water/solids ratio on the fresh state and hardened state of the geopolymer material for 3D printing. Unlike the majority of studies that rely on fly ash or ground granulated blast furnace slag (GGBFS), this work demonstrates for the first time the feasibility of valorizing SAW slag as a precursor for 3D-printed geopolymers. The novelty lies not only in the use of an underexplored industrial by-product but also in its validation at a pilot printing scale (54 kg batches in a gantry printer), bridging the gap between laboratory testing and practical application in digital construction. The following conclusions can be drawn from this study:

• Effect of KOH molarity: Increasing KOH molarity delayed setting times, with the longest values observed at 10 M and 12 M. The best mechanical performance was achieved with 8 M KOH, reaching 48.5 MPa at 28 days. Higher concentrations negatively affected strength due to excessive dissolution and porosity. These results provide new insight into the optimization of alkaline activator concentration for SAW-based geopolymers.
• Influence of SAW slag: Partial replacement of GGBFS with SAW slag slowed down setting time due to compositional differences (higher Al₂O₃ and MgO, lower CaO and SiO₂). However, compressive strengths above 45 MPa at 28 days were still obtained, and 30% replacement achieved the highest value (52.4 MPa). This confirms the potential of SAW slag to act as a sustainable supplementary precursor.
• Optimization of w/s ration and 3D printing parameters: The consistency, shape retention, and layer-by-layer extrusion tests show that the S3-1, S3-2, and S3-3 mixes meet the requirements for 3D printing. During the consistency test, the materials exhibited good consistency values and also showed a good appearance during the process. The shape retention test revealed that mix S3-2 showed almost no height deformation when subjected to a total weight of 2800 g, and the layer-by-layer deposition test demonstrated that the layers were extruded without any issues, showing good continuity and no deformation under the weight of the upper layers.
• 3D printing test: The optimized formulation (20% SAW slag, 8 M KOH, w/b = 0.29) showed adequate workability, extrudability, and buildability, meeting the requirements for 3D printing. In pilot-scale printing, this mix exhibited continuous extrusion, strong interlayer adhesion, and shape retention, with only minor edge irregularities. These results demonstrate, for the first time, the suitability of SAW slag–based geopolymers for additive manufacturing applications.
• Sustainability potential: The use of SAW slag as a precursor represents a significant step forward sustainability. This by-product can be valorized, not only in 3D printed geopolymers, but also as a binder substitute in mortars or in geopolymers for other applications. Its incorporation reduces waste disposal, decreases the demand for traditional binders, and supports circular economic strategies.

This study was limited to the evaluation of fresh-state behavior, setting time, and mechanical properties of SAW slag–based geopolymers for 3D printing. Microstructural investigations, such as SEM and XRF, were not included within the scope of this work, but they are essential to confirm the mechanisms of pozzolanic gel formation and will be addressed in future studies. Likewise, comparative evaluations of mechanical performance, including anisotropy in 3D-printed elements, are planned as part of the research team’s next investigations.