Using Vacuum Mixing for 3D Printing of Mortars Made with Recycled Sand

Abstract

This study investigates the use of recycled concrete aggregates as a replacement for natural sand in printable mortars, comparing the properties of both fresh and hardened states. Two types of mortars were considered, natural mortar and recycled mortar, with further variations based on mixing methods under ordinary atmospheric pressure and vacuum pressure. The experimental approach included air content, mini-slump, printability, and various hardened state tests such as compressive strength and porosity measurements using both water absorption and mercury intrusion porosimetry (MIP). The results showed that mortars made with recycled sand exhibited higher fluidity, as evidenced by an increase in slump of approximately 50 to 70 mm across 30 min, compared to those made with natural sand. This difference was attributed to the pre-saturation of recycled sand, which, as a hypothesis, may increase with the amount of free water available while mixing under vacuum. Additionally, mortars containing recycled sand exhibited higher water-accessible porosity (approximately +7% compared to natural mortars) and lower compressive strength, with a reduction of about 5 to 10% for printed and cast samples, with the decrease being more pronounced in printed specimens. However, vacuum mixing was found to significantly reduce entrapped air content, by about 53% in natural mortars and 62% in recycled ones, and to enhance the workability of both types. The pore size distribution indicated that recycled mortars had a more complex pore network, with pores in the ranges of [0.01–0.1] mm and [0.1–1] mm, contributing to increased porosity and reduced mechanical strength. Overall, this study demonstrates the potential of using recycled sand in mortar formulations, with proper control of pre-saturation and mixing conditions to optimize performance in both fresh and hardened states.

1. Introduction

Three-dimensional printing has become the latest technology to have developed rapidly in many sectors, including the construction sector. Three-dimensional concrete printing could be seen as the future of construction, representing a significant innovation that has already revolutionized the industry. However, it is still considered to be a relatively new technology in the construction industry. Many companies are developing 3D printers aiming to build at low costs and in a short time. In addition, 3D printing allows for the building of complex architectural forms, whether on a small or large scale [1,2,3]. Despite its potential, several challenges continue to hinder the adoption and further development of this technology [4,5,6,7]. One of the main obstacles concerns the supply and formulation of printable cementitious materials [8,9,10]. Most printable materials consist of mortars made from cement, sand, water, and admixtures. Due to the limitations regarding nozzle size, aggregates are often excluded, making sand the most essential natural component in these mortars [11]. The construction industry needs exceptionally large quantities of materials issued from natural resources. Sand is one of these materials and is the third most frequently used natural resource by humans, after air and water. As demand for construction materials continues to rise, alternatives to natural sand sources must be used. Recycled sand (RS), derived from construction and demolition waste, can partially or totally replace the natural sand in the formulations [12,13,14]. By incorporating recycled sand into 3D printable mortars, the environmental impact of sand extraction can be significantly reduced while promoting a circular economy. The use of recycled sand not only helps conserve natural resources but also minimizes waste, contributing to a more sustainable construction industry. However, the use of RS in construction materials presents challenges related to material properties, including variability in water absorption, porosity, and particle shape, which can affect the performance of the final product [15,16].

The preparation process of mortars, including the mixing procedure, plays a crucial role in the final properties of the material [17,18,19]. Methods such as vacuum mixing have been explored to improve consistency and reduce the amount of entrapped air in cementitious materials [20,21]. These advances are especially important when dealing with recycled materials, where inconsistencies in aggregate properties may otherwise lead to undesirable effects on workability and strength [22,23].

The influence of vacuum mixing on the fresh and hardened performance of 3D printable mortars made with recycled sand (RS) remains unexplored. In particular, the combined influence of these factors on printability characteristics (including buildability, extrudability, and interlayer adhesion) as well as on the mechanical and durability performance of printed and cast elements has not been explored. Further research is therefore required to elucidate these interactions and to optimize the formulation and processing of printable mortars for additive manufacturing applications. The influence of different mixing procedures and mixer types on fresh properties of the prepared cementitious materials should be better understood. One of the parameters is the uncontrollable air inlet during concrete mixing, leading to undesirable fluctuations in the properties of fresh and hardened concrete. Thus, this paper aims to study the effect of the mixing procedure on the rheological and mechanical properties of printable mortar prepared with natural sand (NS) and recycled concrete sand (RS). In order to perform this study, some steps were required. The NS and RC were first characterized, and the NS was recomposed to be close to the RS to eliminate the effect of particle size distribution. The study started with a reference printable mortar composed of NS and Portland cement. In a second step, the granular fraction (NS) was replaced with RS. The main challenge of replacing NS by RS while keeping the other components is estimating the amount of water absorbed by RS. The water absorption of RS (very high compared to NS) was considered to control the amount of effective water.

2. Materials and Methods

2.1. Materials

CEM I 52.5 N Portland cement complying with NF EN 197-1 [24] was used, with a real dried density of 3.16 g/cm³. The chemical composition of the used cement is shown in

Table 1. Chemical composition of cement CEM I 52.5 N.

Chemical Component	Results %
CaO	64.3
SiO2	18.3
Al2O3	5.2
Fe2O3	4.0
MgO	1.4
Na2O	0.32
K2O	0.43
SO3	3.5
Cl−	0.06
Loss-on-ignition	2.3
Insoluble residue	0.4

.

In all mixtures, the effective water to cement ratio (W_eff/C) by mass was 0.27. A Polycarboxylic Ether-based (PCE) superplasticizer, Master Glenium 51complies with EN 934-2 [25], was used as a water Reducing Agent (WRA) as to increase the flowability of the mortar. The mass of PCE was the mass in liquid form delivered by the supplier. The solid content of the WRA was 20% over the total mass in liquid form. Two types of sand were used: natural river sand (NS) and fine recycled concrete aggregate from the “Gonesse recycling platform” (RS). The particle size distribution analysis showed that both sand types had different particle distributions. Therefore, the natural sand was sieved and recomposed to obtain a particle size distribution similar to that of the recycled sand.

Figure 1 and

Table 2. Characteristics of natural and recycled sand (WA₂₄: water absorption coefficient, ρ_rd: real dried density and ρ_SSD: saturated surface dried density).

Characteristics	NS	RS
Water Absorption (%)	0.7 ± 0.04	8.3 ± 0.46
ρrd (g/cm3)	2.49 ± 0.02	2.11 ± 0.04
ρSSD (g/cm3)	2.50 ± 0.01	2.30 ± 0.05

show particle size distributions and characteristics of natural recomposed and recycled sand, respectively. It is important to highlight that the two types of sand did not have similar grain shapes, which was an uncontrollable parameter but could have a non-negligible influence on the behavior in the fresh state.

Table 2 summarizes the most significant characteristics of the material used. The water absorption coefficient (WA₂₄), the real dried density (ρ_rd) of the sands and the saturated surface dried density (ρ_SSD) were determined according to European standard EN 1097-6. The results shown in Table 2 represent the average of three samples tested for each sand.

The determination of water absorption (noted WA₂₄) is of great importance for the characterization of RS. Up to now, there has been no particular standard for the measurement of the water absorption coefficient of RS. Therefore, standards used for natural aggregates such as EN 1097-6 [26] and ASTM C127 [27] (for coarse aggregates) and ASTM C128 [28] (for fine aggregates) were used for the characterization of RS. The principle of these methods is similar, and the water absorption coefficients for NS and RS were determined based on the water content at saturated surface dry (SSD) state after total immersion in water for 24 h. The SSD state of sand dried progressively under warm air was identified using a cone test and according to the EN 1097-6 [26]. The sand was then dried in an oven at a temperature of 75 °C instead of 105 °C, as mentioned in the standard EN 1097-7 [29], until reaching the constant mass of oven-dried aggregates (noted MOD). In the case of recycled concrete aggregates, the drying temperature was reduced to avoid the outflow of chemically bound water with adherent cement paste [30].

Table 3. Mix design for normal and recycled mortar.

kg/m3	NM	RM
CEMI 52.5	938.3	938.3
Weff	247.6	247.6
PCE SP	13	13
RS	-	963.1
NS	1126	-
Wabsorbed	7.88	79.94

shows the mix design of the mortar with natural sand (noted NM) and mortar with recycled sand (noted RM). Since water absorption for the recycled sand was higher than that of natural sand, recycled sand was used in the saturated surface dry condition. The RS was immersed into water (W_eff + W_absorbed) for 24h in containers kept at 20 °C. For both types of studied mortars, the same volumes of sand, effective water and cement were used.

2.2. Mixing Procedure

Figure 2 shows the vacuum mixer used for the tests. Four mixtures, denoted NMO, NMV, RMO and RMV, were prepared. The first letter refers to the type of sand: natural (N) or saturated recycled (R). The letter M refers to mortar, while the last letter indicates whether the mixing was performed under ordinary atmospheric pressure (O) or under vacuum pressure (V).

The mixing procedure depended on the type of sand used.

For NMO and NMV, dry sand and cement were first added to the mixer pan and mixed for 20 s at a minimum speed of 260 rpm. Then, the total amount of water (added water + absorbed water) mixed initially with SP was poured into the dry mix for 10 s. In the vacuum-mixed batches, the materials were added in the same sequence as in the standard mixing procedure. The vacuum pump was activated 80 s after mixing began, and the pressure stabilized at 145 mbar within approximately 60 s. The total mixing duration for all mortars was 300 s, meaning that the vacuum-mixed samples were mixed under vacuum for about 160 s. Throughout the process, the mixer speed was maintained at a constant 260 rpm.

For RMO and RMV, different mixing protocols were followed because the recycled sand was used in saturated condition for better control of the real effective water while mixing compared to NMO and NMV. First, the recycled sand was dried in a conventional oven (105 °C for 24 h). Then, the sand was immersed in the total amount of water (W_added + W_absorbed) in a sealed container. The container was then rolled 20 times back and forth horizontally to homogenize the mixture. The sealed container was then kept in an air-conditioned room at 20 °C for 24 h. For mixing, the first small amount of water was extracted from the container and mixed with the total amount of SP to ensure better homogeneity. Then, the rest of the water with recycled sand and dry cement were poured in the mixer pan and mixed for 20 s. The amount of water mixed with SP was then added during 10 s of mixing. The remaining mixing procedure was similar to that described for NMO and NMV.

For all prepared mortars, the total mixing time was 300 s. During mixing, the mixer speed was kept constant at 260 rpm. After mixing, a volume of 4 L of fresh mortar was poured into a container to perform characterization and printing tests.

2.3. Fresh and Hardened Tests

The fresh properties of the test mixtures were assessed by measuring air content and slump and by performing printing tests. Hardened mortars were tested for their mechanical properties, and porosity by means of water absorption and mercury intrusion.

2.3.1. Fresh Tests

• Air content

The air content test was performed according to standard EN 12350-7 [31]. All air content tests were performed 5 min after mixing.

• Mini-Slump

The mini-slump test was performed according to standard EN 1015-3 [32] to evaluate the workability of the fresh mortar. A mini cone with a bottom diameter of 100 mm, an upper diameter of 70 mm and height of 60 mm was used for testing. After mixing, the mortar was immediately poured into the cone, and the latter was lifted slowly. The spread diameter was then measured in four directions after 15 blows with the jolting table and the arithmetic mean of the measured diameters was calculated.

In order to evaluate the workability loss of the prepared mortars, the slump flow measurement was repeated at various times with a gap of 10 min. For the same sample, the arithmetic mean of the measured diameters was calculated at 5, 15, 25 and 35 min. For the measurements taken at different times, the batch was placed in a bucket covered with a damp cloth and manually mixed before each test. The same sample was used for all these measurements. The effect of the type of sand and vacuum mixing on the consistency of the studied mortars was assessed by comparing the relative slump according to Equation (1) [33].

$$s_r={\left({\displaystyle \frac{d}{d_0}}\right)}^2-1$$

(1)

where S_r is the relative slump, d is the average of 4 measured diameters of the spread and d₀ is the cone bottom diameter.

For each formulation, the test was repeated three times (3 batches).

• Printing

The printability of fresh mortar was evaluated using a 2D printer developed at laboratory scale for cementitious materials in the Magnel-Vandepitte Laboratory for Structural Engineering and Building Materials at Ghent University [Figure 3]. A mortar gun consisting of a hopper, auger, vibrator rod and nozzle, developed initially for masonry joints, was adjusted to print fresh mortar. The hopper has a capacity of 2 L. The vibrator-fed auger drive assures constant mortar bead. A variable speed drill is mounted and connected to the gun. The gun drill-mate is fixed in vertical position; its vertical position can be manually adjusted.

First, the barrel was filled with fresh mortar, and then the material was delivered in a controlled and consistent manner by adjusting the drill speed. The material flowed from a specially designed nozzle with 4 interchangeable steel tips and was deposited on a linear motion guide. The effective stroke of the rail guide was 450 mm, the max speed 140 mm/s and the maximum holding weight was 40 kg. All the printing tests were performed 10 min after mixing and with a linear speed of 15 mm/s. The thickness of the printed layer was 10 mm while the nozzle size was 25 mm. The maximum number of layers that could be printed using this setup was 15.

The results showed that NMO, NMV and RMO were printable. Fifteen layers were printed without noticeable deformation or cracking, as shown in Figure 4. However, RMV showed unprintable behavior due to its high fluidity.

2.3.2. Hardened Tests

The hardened tests included water absorption, mercury intrusion porosity (MIP) and mechanical strength tests. Hardened tests were conducted on mortars cast into molds and printed as well.

For molded mortars, two molds of 3 samples of 4 × 4 × 16 cm³ each were filled in two layers. Then, the fresh mortar was compacted using a standard jolting apparatus. The samples were unmolded after 24 h and stored in a moist atmosphere (Relative Humidity RH 90% and temperature 20 °C) until hardened tests. The specimens were taken from the moist room at 28 days, broken in flexure into two halves and then one half tested for strength in compression and the other half kept for porosity tests (water absorption and mercury intrusion). Every mold (3 samples) had 3 flexural, 3 compression and 3 porosity tests at 28 days.

For printed samples, different protocols were followed. Two samples were first printed, one of 3 layers (S3) and a second of 15 layers (S15). Due to the high fluidity of the RMV mortar, it was not possible to print more than 3 layers. The three printed layers allow for a comparison with the cast samples (of the same height), as shown in the results presented in Section 3.

After 24 h, the printed samples were stored in a moist atmosphere (RH 90%) until strength testing. At 28 days, compression and porosity tests were performed on cubic samples prepared by sawing S3 and S15. For S3, the cubic samples (3 × 3 × 3 cm³) were obtained directly by sawing the printed samples (Figure 5). For S15 (Figure 6), the printed samples were first divided into 5 columns. Then, every column was split into 6 pieces, numbered from 1 (bottom) to 6 (top). Finally, every piece was sawn into a cubic shape, (3 × 3 × 3 cm³) for cubes 1 and 2 and (2 × 2 × 2 cm³) for the cubes 3 to 6, as shown in Figure 6.

• Water-accessible porosity test

Water-accessible porosity ($P_{acc,water})$ was measured by imbibition under vacuum and hydrostatic weighing according to French standard NF P 18-459 [34]. The hardened mortars were first cut into equal cubes. The samples were kept under vacuum for 4 h and then stored in water for 24 h to determine the water-accessible porosity. The porosity measurements were performed on cubes extracted from mortars cast in standard molds and on printed layers as well (Figure 6).

• MIP test

The porosity and pore size distribution were determined using a mercury porosimeter (Micrometrics Autopore IV) with low- and high-pressure positions, allowing access to a pore radius between 6 nm and 600 µm. The samples were kept under vacuum for 4 h before the MIP test. The MIP tests were performed on small cubes extracted from mortars cast in standard molds and on printed layers as well (Figure 6).

• Mechanical Strength

Mechanical tests were performed according to EN 196-1 [35]. The mechanical behavior of each of the formulations studied was evaluated by simple compression tests. The tests were carried out using an Instron^® type mechanical press with a capacity of 150 KN.

3. Results and Discussion

3.1. Fresh State

For the same theoretical W_eff/C ratio, the workability of mortars containing natural sand is much less than that of mortars containing saturated recycled sand, as shown in Figure 7. The theoretical compositions of mortars of natural sands and recycled sand are very close. The two sands have similar particle size distributions, the same envelope volumes, and the same quantities of cement and theoretical effective water. Thus, the differences in workability between the two materials can come from the geometry of the particles and the free water in the paste, and can possibly also be due to the mixing procedure.

As displayed in Figure 7, the relative slump increases when recycled sand is used compared to natural mortar. The recycled sand was saturated in water with the amount of water required to reach the same effective water level as the normal sand. However, it was not certain that all the pores were filled with water in the saturation phase. This difference may also be partially attributed to the mixing procedure itself and to the slight excess of saturation water in the sands. An amount of water can be released from these pores during the mixing process. This phenomenon reduces the actual saturation level of the aggregates at the time of mixing, potentially increasing the amount of free water in the mixture.

Figure 8 shows the air content measurements. Three measurements were taken in each case, and the results are reproducible with low standard deviations. The results show that under ordinary mixing, natural mortar exhibits higher entrapped air content (3.3%) than recycled mortar (2.0%). This is attributable to the mixing procedure. For natural mortar, water is added during mixing, which can increase the amount of entrapped air. Additionally, the difference in behavior between natural and recycled mortars regarding water addition can be attributed to the inherent properties of the aggregates. Natural sand typically has lower porosity and water absorption compared to recycled sand, allowing for more predictable water and air content. In contrast, the higher porosity of recycled aggregates necessitates careful management of water content to control entrapped air and ensure adequate hydration. Recycled mortars, particularly those incorporating recycled aggregates, exhibit higher water absorption due to the porous nature of the aggregates. Without proper pre-wetting, the quantity of water absorbed by aggregates during mixing could be lower than the theoretical amount of absorbed water, leading to a higher effective water-to-cement ratio and potentially reducing the strength. Pre-wetting the aggregates can mitigate this issue by reducing their absorption capacity during mixing [18,36].

The results show that mixing under a vacuum reduces the entrapped air content for natural (53%) and recycled (62%) mortars. However, entrapped air reflects the voids present in the mortar due to mixing and consolidating. In this study, using vacuum mixing allows the entrapped air content to be reduced, as shown in Figure 7.

3.2. Hardened State

• Porosity

The characterization of porosity of recycled concrete is essential for understanding its durability and performance. Two commonly used methods are water-accessible porosity measurement and mercury intrusion porosimetry (MIP) [37]. Water-accessible porosity reflects the volume of interconnected pores that can be filled with water under low pressure, providing information about the material’s permeability and capacity for fluid transport. In contrast, mercury intrusion porosimetry allows for a more detailed analysis of pore size distribution and includes both connected and isolated pores, often under high pressure. However, the water-accessible porosity better represents the pores relevant to real-world environmental exposure, such as those affecting freeze–thaw resistance and chloride ingress. Therefore, combining both methods provides a more comprehensive understanding of the pore structure and potential durability of recycled concretes.

The results of water-accessible porosity tests for the studied mortars are shown in Figure 9. The results were obtained by printing only three layers. Indeed, the mix RMV was fluid and not buildable for more than three layers. For a given mortar there is no noticeable difference between the measured porosity for cast or printed mode. This shows that the porosity of the interlayer zone does not have a significant effect on the porosity of the printed specimen. It is shown that recycled mortar has the highest averaged porosity, at ~23%, while the natural mortar has an average value of ~15%. It is also found that vacuum has no significant effect on porosity. Given the mix design, as for a similar paste, the variations in porosity can be explained from the porosity of the sands used. However, it is also possible that some of the additional porosity observed for recycled mortars is directly related to the uncertainty of the calculated volume of saturation water in the sands.

The theoretical water-accessible porosity for the aggregates NS and RS is calculated according to Equation (2) [38]: 1.7% for the NS and 17.5% for the RS. So, it can be deduced that the theoretical difference in porosity of NS and RS leads to the difference in porosity of the mortars.

$$P={\displaystyle \frac{WA.{\rho }_{rd}}{{\rho }_w}}$$

(2)

Knowing the porosities of aggregates and the total porosities of the mortars, the porosities of the pastes (P_paste) can be estimated using Equation (3) [11]:

$$P_{mortar}={XP}_{paste}+\left(1-\mathrm{X}\right)P_{sand}$$

(3)

Here, X represents the volume fraction of paste, which can be estimated from the mix compositions, neglecting the volume of entrapped air. $P_{mortar}$ is the total mortar porosity (as determined in Figure 9), and $P_{sand}$ is the sand porosity (recycled or natural) calculated using Equation (2)).

For NMO and NMV, X = 0.56, and for RMO and RMV, X = 0.55. According to Equation (3), the calculated P_paste values for NMO and NMV are 26% and 25%, respectively, using P_mortar values of 15.3% and 14.8% (from Figure 9). For RMO and RMV, P_paste values are 26.4% and 24.6% respectively, with P_mortar values of 22.4% and 21.4% (Figure 9). These results are summarized in

Table 4. Results for the calculation of P_paste for cast mortars.

Mortars Cast	X	Pmortar (%)	Ppaste (%)
NMO	0.56	15.3	26
NMV	0.56	14.8	25
RMO	0.55	22.4	26.4
RMV	0.55	21.4	24.6

.

These calculations confirm that both NMO and RMO mortars have nearly identical actual effective water-to-cement ratios, and that the higher porosity observed in the RM is indeed attributable to the greater porosity of the recycled sand. RM showed around 7% more water-accessible porosity, which aligns with the higher water absorption of the recycled sands, estimated at approximately 70 L/m³.

Figure 10 shows the water-accessible porosity results of the printed fifteen-layer mortar samples, comparing the mortars from the first three printed layers to those from the last three layers. The results show no significant difference between the samples from the first layers and those from the last layers. When comparing the results from Figure 9 to those of Figure 10, no noticeable difference is observed in the porosity values measured for each formulation.

Figure 11, Figure 12, Figure 13 and Figure 14 show the results of the MIP test.

The results show that the water-accessible porosity is higher than that measured with the MIP test. Several studies highlight that the porosity measured by mercury intrusion is systematically lower than that obtained by the water-accessible porosity method [38]. This can be explained by the fact that mercury intrusion only concerns pores of a certain size, while the water test takes into account all pores. Indeed, this discrepancy can be attributed to the ink-bottle effect: large pores connected by narrow throats are only filled at pressures corresponding to the throat diameter during MIP, leading to an underestimation of the total pore volume. In contrast, water can access these pores more freely, so the water-accessible porosity captures the full pore network. The required correction factor is significantly greater for RM samples than in NM, likely indicating a difference in their pore network complexity. This suggests that the RM samples have a more pronounced ink-bottle effect (i.e., a larger volume of large pores connected by small necks) and/or a greater volume of micro-pores/damage not measured by the standard MIP procedure, compared to the NM samples.

According to the literature, RS shows a higher porosity than NS, mainly as a consequence of the amount of old cement paste in RS [13]. Figure 13 and Figure 14 present the pore size distribution for mortars cast in standard molds and for mortars’ S3 printed layers.

Figure 13 presents the pore size distributions for mortars cast in standard molds. Note that the pore size distributions obtained from MIP do not match the reality due to an overestimation of the smaller pores and an underestimation of the larger voids due to the limitations of MIP, due to factors such as the ink-bottle effect [37].

Figure 13 shows that the porous network of mortars cast in molds was mainly composed of one class of pores [0.01–0.1]. Vacuum mixing does not have a significant influence on the capillary porosity.

Figure 14 presents the pore size distributions for S3 printed mortars. The results show that the porous network was mainly composed of one class of pores, [0.01–0.1], for natural mortars and two classes of pores, [0.01–0.1] and [0.1–1], for recycled mortar.

Figure 15 shows the compressive strength of the cast and printed (S3) samples. Cast mortars exhibit higher strength than printed mortars. In addition, for cast mortars, the compressive strength is higher (~100 MPa) for natural mortar compared to the recycled one (~71 MPa). Regarding the printed samples (S3), the natural mortar exhibits higher strength (~70 MPa) compared to the recycled mortar (~52 MPa). The drop in strength due to printing is very significant for the natural sand mortar. In contrast, this decrease is less pronounced for recycled sand.

Figure 15 also shows that vacuum mixing improves the compressive strength of mortars with natural sand, with an increase in strength of about 5 to 10% for printed or cast samples, respectively. On the other hand, for recycled mortars, the effect of vacuum mixing is much more limited. This is probably due to the fact that the vacuum was applied, in our experiments, on the mortar with saturated recycled sand. The low pressure probably led to part of the absorbed water present in the RS being extracted, increasing the efficient water in the mortar by a little.

It is important to note that, for printed samples, the strength test was conducted in a direction perpendicular to the layers. Several studies have shown anisotropy in printed elements, which reduces the compressive strength compared to molded samples. The anisotropy is caused by zones of tension and expansion that are perpendicular to the layers [39]. While the air content and porosity show minimal variation between printed and cast samples, there is a significant difference in strength. The processing index, calculated as $f_{c,\mathrm{casted}}/f_{c,\mathrm{printed}}$, is approximately 1.5 for NMV and NMO, and around 1.3 for RMO and RMV. This could be explained by the fact that the additional porosity is located between the layers. The pore size distribution indicated that recycled mortars in printed samples had a more complex pore network with pores in the ranges of [0.01–0.1] mm and [0.1–1] mm, contributing to increased porosity and reduced mechanical strength.

Figure 16 shows the results of the compressive strength of the samples taken from the first and last of the 15 printed layers. The results show homogeneous strength with no significant variation between the first and the last layers.

4. Conclusions

This study has successfully evaluated the impact of using recycled sand in cementitious mortars for additive manufacturing, specifically focusing on the comparison between ordinary and vacuum mixing techniques. The results of both fresh and hardened state tests indicated several critical insights.

• Recycled mortars exhibited higher fluidity compared to natural mortars, as demonstrated by the increased relative slump values, which can be reasonably attributed to the release of water from the aggregates during mixing. This difference was primarily due to the need to pre-saturate the recycled aggregates with high water absorption. Vacuum mixing was shown to reduce the entrapped air content, enhancing the flowability and workability of both natural and recycled sand mortars.
• Mortars containing recycled sand exhibited a higher water-accessible porosity (~23%) compared to those with natural sand (~15%), aligning with the inherent characteristics of recycled aggregates. This increase in porosity, particularly in recycled mortars, may influence durability aspects, such as freeze–thaw resistance and permeability. Notably, no significant differences in porosity were found between cast and printed samples, implying that the printing process does not substantially alter the porosity of the material.
• The compressive strength of mortars containing natural sand was higher than for mortars with recycled sand, with a notable decrease in strength for printed samples due to the layer-by-layer construction. This can be reasonably attributed to the release of water from the aggregates during mixing. It is likely that the water initially retained within the recycled aggregates is released into the mix and remains outside the grains until hardening occurs. This phenomenon may explain the observed decrease in compressive strength. The strength of printed mortars was significantly lower, especially for natural sand, indicating that the printing process can negatively affect the structural integrity of the material. However, recycled mortars exhibited a more consistent strength reduction, suggesting that the incorporation of recycled aggregates could offer more stable performance in additive manufacturing applications.
• Vacuum mixing demonstrated a positive effect on reducing entrapped air content in both natural and recycled sand mortars. Although it did not significantly alter the overall mechanical properties, vacuum mixing improved the fluidity of the materials.

In conclusion, recycled sand-based mortars, although showing some performance limitations compared to natural sand mortars, demonstrate promising results in terms of workability and porosity. The use of vacuum mixing further enhances material performance. However, it should be noted that the cement content in this study was relatively high to ensure printability and mechanical stability. For practical and sustainable applications, future work should focus on optimizing mix designs, reducing cement content, and exploring alternative binders while maintaining suitable performance. Overall, this research provides valuable insights into the use of recycled aggregates in 3D printing and highlights the need for rational and sustainable material formulations in additive manufacturing.