Evaluating the Printability and Rheological and Mechanical Properties of 3D-Printed Earthen Mixes for Carbon-Neutral Buildings

Abstract

This research delves into the vital aspect of adapting local soil properties for 3D printing and explores the mix design of collected earthen materials to tackle challenges in printability, shrinkage, and mechanical properties. Initially, soil samples from six local sites underwent characterization based on Atterberg’s limits, focusing on identifying the most suitable high-clay-content soil. The soil with a higher clay content was used for further study, and its clay type was determined using an X-ray diffraction (XRD) analysis, revealing the presence of 49% kaolinite, 15% nontronite, and 36% illite clay minerals. Four earthen mixes were designed by including stabilizers (i.e., hydrated lime), natural pozzolana, and degradable natural fibers (wheat straw fiber). Subsequently, the study examined their rheological properties, shrinkage behavior, compressive and flexural strength, and printability (including extrudability and buildability). The pure soil mixture excelled in printing quality and mechanical strength, but suffered from cracking and drying shrinkage due to its high nontronite clay content. The existence of 15% nontronite clay mineral in the soil resulted in significant shrinkage and extensive cracking of specimens. However, fiber incorporation effectively mitigated large cracks and reduced shrinkage to as low as 2.6%. Despite initial expectations, introducing lime and pozzolana as soil stabilizers did not improve strength, prevent shrinkage, or improve the printability of soil mixes.

1. Introduction

In recent years, the construction industry has witnessed technological advancements during the fourth industrial revolution, leading to efforts to boost productivity, streamline supply chains, lower labor costs, and automate operations [1,2,3,4,5]. Additionally, labor shortages and construction-related risks have further driven the search for advanced solutions [3]. Lately, additive manufacturing, also known as 3D printing, has shown promise in addressing these concerns [6,7,8,9]. This technology was introduced as early as 1997 [10], with most research focused on 3D printing using concrete for construction purposes [11,12,13,14]. On the other hand, the construction industry faces numerous sustainability issues, including resource depletion, energy consumption, waste generation and carbon emissions [15,16,17]. Addressing these sustainability issues requires combining innovative construction techniques, green building materials and practices, energy-efficient designs, and responsible resource management throughout the construction lifecycle. In this context, carbon-neutral building practices seek to reduce construction project carbon footprints with a promising approach, such as the utilization of local resources like soil [18,19]. In recent years, there has been growing interest in additive manufacturing (or 3D printing) of soil and earthen materials, combining traditional materials with new technology to create eco-friendly humanitarian shelters. This study also centers on the 3D printing of local soil, and assesses both the fresh and hardened properties of the locally collected earthen materials.

While most research in construction technology focuses on 3D concrete printing, it is noteworthy that cement production contributes significantly to global carbon dioxide emissions (5–7%) [20]. The global cement industry’s production has surged more than 2.5-fold in less than three decades, as reported by the European Cement Association in 2020 [21] and Worrell et al. in 2008 [20]. Although advancements in clinker technology have reduced ${\mathrm{C}\mathrm{O}}_2$ emissions’ intensity during cement production, overall global ${\mathrm{C}\mathrm{O}}_2$ emissions from construction are projected to rise [12]. Consequently, there is growing interest in using sustainable resources, emphasizing the importance of sustainable building materials. In the pursuit of sustainability, research into 3D printing with alternative construction materials like earthen materials, hemp hurds, and bio-based materials has garnered significant interest [22,23,24,25,26,27,28,29,30].

Earthen materials, dating back over 9000 years [31,32,33,34], serve as a green alternative to conventional construction materials. Traditional construction methods with earthen materials include adobe, rammed earth, cob, light clay, and pit houses [35]. Soil, as a building material, offers several advantages, primarily its ready availability on construction sites. This reduces transportation costs and ensures accessibility with minimal construction expenses and energy usage. Although, the soil needs some stabilization such as fiber reinforcement to meet construction requirements [36]. In continue, adobe construction is particularly cost-effective due to its straightforward preparation and building techniques, making it an ideal choice for affordable housing. Adobe typically consists of soil, plant-based fibers for reinforcement, and optional additives or stabilizers as required. Soil can also be economically stabilized using hydrated lime, asphalt emulsion, or Portland cement. Additionally, soil possesses excellent thermal properties, remaining comfortably cool in summer and retaining heat well during the winter [35].

While these materials have shown promise, they do exhibit certain limitations when exposed to gravity or lateral seismic forces. Notably, soil lacks inherent ductility properties, making it prone to cracking under tensile and shear forces. Furthermore, it is susceptible to damage from earthquakes and other dynamic external loads. These vulnerabilities have historically resulted in significant buildings and human losses in seismic-prone regions. Recent earthquakes in Iran, Nepal, Pakistan, Portugal, and Turkey, for instance, inflicted severe damage on stone–mud masonry and adobe masonry structures. In comparison to buildings constructed with reinforced concrete, steel, or wood, those relying on soil-based construction proved to be particularly vulnerable to seismic loads [37,38,39,40].

Furthermore, achieving consistency in earth-based construction poses challenges due to the inherent heterogeneity of soil composition characteristics. Consequently, building codes for earthen construction have limited parameters for control. Notably, the International Building Code (IBC) [41] and the American Concrete Institute (ACI) building code specifications for masonry structures [42] primarily rely on empirical design standards when addressing earthen buildings. Earthen building codes lag regarding updates and development compared to codes for more conventional materials such as concrete and steel.

Previous publications have extensively explored soil printability and its applications in the construction industry from various perspectives [22,43,44,45,46,47]. Barnes et al. [43] evaluated different soil textures, including sand, silt, and clay, without additives regarding printability and water retention capacity. Although all the soil mixes were printed successfully, it was found that the 3D-printed samples had a lower water retention capability compared to the molded one. In order to put into practice the plant-supported earthen building, researchers at the University of Virginia [45] printed a small wall prototype with the soil–seed combination that can sprout greenery. To overcome the high cost and improve the workability of cement composites for 3D printing, Iubin and Zakrevskaya [22] used clay soil as an additive to the 3D printable cement composite. They observed the positive impact of soil in increasing the workability time from 60–80 min to 160–200 min; however, adding clay soil decreased the compressive strength between 0.5–25%. Ferretti et al. [26] developed a 3D-printable earthen material modified with a rice-husk–lime biocomposite, and evaluated its mechanical properties. The results showed that shredded rice husk had higher compressive strength compared to the mix with the non-shredded rice husk, which can be attributed to the higher surface contact of rice husk with lime in the first mix compared to the latter one; however, it had lower ductility, which may be related to the shorter length of the rice husk in the specimens with shredded rice husk.

In the realm of earthen-based 3D printing, there is a notable research gap concerning our understanding of the fresh and hardened properties of 3D-printed components. Shrinkage, a key factor leading to cracks in hardened earthen mixes, has not received adequate attention in prior studies. Moreover, there are significant discrepancies in mechanical performance between soil mixes and conventional materials, necessitating specific strategies in soil mix design to bolster mechanical strength.

This paper builds upon a preliminary study conducted with locally available materials in New Mexico [27]. New Mexico has a rich history of adobe structures that have been a prominent feature for many years in construction. Given the widespread use of this material, this study aims to assess its suitability for 3D printing applications. The objective was to assess the fresh properties of earthen mixes, which included conducting flow table and rheometer tests for parameters such as static yield stress, plasticity yield stress, dynamic yield stress, and thixotropy. In the fresh stage, printability was evaluated through extrudability and buildability tests using a gantry-type 3D printer. The paper investigated the properties of the hardened mixes, encompassing factors like shrinkage, compressive strength, and flexural strength. Furthermore, it assesses the influence of soil stabilizers, fibers, and pozzolana on both fresh and hardened states of soil mixes.

2. Materials and Methods

This section briefly reviews the material collection, soil mix design and test methods used to evaluate the fresh and hardened properties of developed mixes.

2.1. Preparation of Local Soil

In the initial phase of our research [27], soil samples were collected from six distinct locations in Albuquerque, New Mexico, listed in

Table 1. Soil site locations and their Atterberg limit and soil type.

Soil ID	Soil Location	Liquid Limit	Plastic Limit	Plasticity	Soil Type
S1	Camino Don Thomas Street, Abrazos Family Service Yard	13.30%	19.30%	No Plasticity	Poorly Graded Sand (SP)
S2	Coronado Little League Park, Rotary Park	18.20%	20%	No Plasticity	Well-Graded Sand (SW)
S3	Ranchitos Drain-1	18.30%	20.70%	No Plasticity	Well-Graded Sand (SW)
S4	Ranchitos Drain-2	Too sandy	Too sandy	No Plasticity	Poorly Graded Sand (SP)
S5	Molino Rail Runner, Bernalillo Downtown Train Station	20.60%	19.30%	1.30%	Well-Graded Sand (SW)
S6	Belen	33.6%	19.9%	13.7%	Well-Graded Sand (SW)

. For material collection at each site, the initial step involved the removal of the topsoil, typically the uppermost layer of soil (about 5–10 inches in depth). Subsequently, the necessary amount of soil was collected from each location. Then, any remaining organic materials, vegetation, and roots were removed for the material preparation, and the soil was sun-dried. Thereafter, the larger clumsy clay particles were crushed into fine particles and mixed homogeneously into a concrete drum mixer to ensure the homogeneity of the soil. The particle size distribution of soil samples was performed, and particles that passed through sieve No. 3/8” (9.5 mm) but were retained on sieve No. 8 (2.36 mm) were collected for plasticity testing. Then, Atterberg limit and plasticity were determined according to ASTM D4318 to classify the soil according to the ASTM D2487 [48] (i.e., Unified soil classification system (USCS)). The plasticity and soil classification results are shown in Table 1.

After conducting preliminary tests for gradation and plasticity evaluation, it was determined that the local soil in Belen, New Mexico, as indicated in Table 1, was the most suitable choice, which showed larger plasticity and was used for the rest of this study. The size distribution of three different soil samples of S6 is depicted in Figure 1.

2.2. Soil Mix Design

Four soil mixes were designed using S6 soil and studied for this paper, as shown in

Table 2. Mixed proportions of soil mixes (weight percentage).

Mix ID	Clay-Rich Soil (S)	Hydrated Lime (L)	Wheat Straw Fiber (F)	Pozzolana (P)	Water to Dry Materials Ratio (W/D)
S	100%	-	-	-	0.380
SL	98%	2%	-	-	0.415
SLF	96.5%	2%	1.5%	-	0.427
SLFP	92.5%	2%	1.5%	4%	0.430

. Besides S6 soil, the designed mixes incorporated locally available additives, including stabilizers (i.e., hydrated lime), pozzolana, and biodegradable natural fibers, advancing toward carbon-neutral construction, as illustrated in Figure 2. The ratios in Table 2 indicate the weight proportions (weight percent) of the different components. The type S hydrated lime was used in the research as a chemical soil stabilizer to control the plasticity of natural soil. The incorporation of 2% lime was determined from preliminary phase results, where the effect of lime proportion on the strength and shrinkage properties was studied [27]. Wheat straw was chosen as a fiber to prevent shrinkage and cracks, and 1.5% of fiber was adapted as having better shrinkage and crack control in the previous phase of this study [27]. The fibers were chopped and ground into smaller pieces using a leaf blower, as shown in Figure 3, and then they were sieved to a size less than 20 mm in length [27]. Wheat straw was chosen as a fiber to prevent shrinkage and cracks, and 1.5% of fiber was adapted as having better shrinkage and crack control in the previous phase of this study [27]. Additionally, the addition of a naturally occurring additive that could react with lime and exhibit cementitious properties was evaluated. Powdered tephra natural pozzolana supplied by CR Minerals company in NM was added in this study as a pozzolanic additive. Previous researchers have used a lime-to-natural-pozzolana ratio of about 1:(2–3) times to stabilize and increase the strength performance of earthen mixes [49,50,51]. Therefore, a ratio of 1:2 of hydrated lime to pozzolana was taken in this research. Additionally, S, L, F, and P in the Mix ID shown in Table 2 represent soil, lime, fiber, and pozzolana, respectively. The water content for the S mix was adjusted according to the plasticity range measured in the soil characterization step. Because of the addition of L, F and P, the water content required to reach acceptable flow properties for each mixture is different, and adjusted according to initial printability tests to achieve a flowable and high-quality mix.

2.3. Mixing Procedure

To mix all the components shown in Figure 2, the dry components were initially blended in a Kitchen Aid mixer at a low speed (140 ± 5 rpm) for 2 min. Water was added gradually to the mixer while it continued to operate at a low speed, and the mixture was scraped off the mixer walls using a spatula after 1 min of water addition. Subsequently, the mixer was allowed to continue blending for an additional 2 min at a low speed, followed by another scraping of the walls. During the final minute of mixing, the speed was increased to high speed (285 ± 10 rpm).

2.4. Test Methods

A summary of tests conducted for this paper is presented in this section: material characterization, evaluation of fresh properties, evaluation of hardened properties, and printing tests, which are all illustrated in Figure 4.

2.4.1. Characterization through XRD and XRF Analysis

An X-ray fluorescence (XRF) test was performed to assess soil and pozzolana oxide compositions. The XRF technique enabled the determination of the soil’s elemental composition, including identifying and quantifying clay minerals. Additionally, X-ray diffraction (XRD) was conducted for the clay portion of the soil samples (passing sieve No. 200), providing valuable insights into the mineralogical composition of the soil and the presence of a specific type of clay.

2.4.2. Fresh Properties

Flow table tests were used to determine the fresh properties of the soil mixes according to ASTM C230/C230M-20 [52], as shown in Figure 4a. These tests helped gauge the soil mix’s consistency, enabling us to adjust its water content to achieve the required workability for meeting extrudability and buildability criteria.

Similarly, the rheology characteristics of the soil mixes were studied using a Brookfield rheometer, as shown in Figure 4b. The test was performed for a hysteresis method. This method has been previously used for cementitious mixes [53,54,55] and was accurate for soil used in this study. Once the materials were thoroughly mixed, the fresh mix was immediately poured into a cylindrical plastic cup and subjected to shear using a rheometer with a shear van spindle. The shear rate was applied from 0 S⁻¹ to 100 S⁻¹ in 60 s, and gradually decreased from 100 S⁻¹ to 0 S⁻¹ in the other 60 s.

2.4.3. Hardened Properties

The mechanical properties of the 3D-printed soil mixes were investigated through a compression and three-point bending test to evaluate the compressive strength and modulus of rupture.

The compressive strength of the cube specimens of size 50 mm × 50 mm × 50 mm was studied according to ASTM C109 [56], as shown in Figure 4c. The cube specimens were cast in two layers and compacted, and their strengths were measured after 7 and 28 days of curing. The materials were air-dried after casting and subsequently cured to enhance the reactivity of lime. The curing was carried out in a humidity-controlled room wherein the specimens were not directly in contact with the water vapor. It was covered in a polythene sheet, proving small pores to have a hydration process. The load compressed the samples at 1 mm/minute in the compression tests. The compressive strength was calculated as the axial force per unit area of the contact surface of the cube specimen. Efforts to 3D-print cubic structures of this size or larger, with the intention of later cutting them into smaller pieces, were unsuccessful in this study. All the samples experienced cracking before testing; therefore, 3D-printed specimens were unsuitable for conducting compressive tests.

The modulus of rupture (MoR) for the printed prisms of size 160 mm × 40 mm × 40 mm was tested following ASTM C293 [57] after 28 days of 3D printing after curing in the same condition stated for compressive strength. The prisms were initially printed with a size of 200 mm × 40 mm × 40 mm, but the prism was cut to make support to support a distance of 160 mm. This examination was conducted using a universal testing machine equipped with a three-point bend fixture, as shown in Figure 4d. The loading was applied in the middle span at a 0.5 mm/min strain rate. The modulus of the rupture was calculated according to the equation below.

$$\mathsf{\sigma }=\frac{3\mathrm{P}\mathrm{l}}{2\mathrm{b}{\mathrm{d}}^2}$$

(1)

where

P = maximum axial load applied on the prism specimen (N)

l = beam length (mm)

b = beam width (mm)

d = beam depth (height) (mm)

σ = modulus of rupture (MPa)

An important consideration when using soil in construction pertains to the volume instability of these materials when subjected to wetting or drying conditions. Hence, this study focused on evaluating the shrinkage and crack development in the studied soil mixes. Three prisms of 285 mm × 25 mm × 25 mm were cast for the axial shrinkage measurement using a caliper according to ASTM C490 [58] for each soil mix. These specimens were compacted and cast in two layers, as displayed in Figure 4f, and their shrinkage and crack development were monitored after a full drying period of 7 days. Shrinkage in the soil specimens could occur in the x, y, or z directions in the printed cast specimens. However, due to the brittleness of the soil prisms, which may crack under external forces by the comparator or any movement, this study was focused solely on measuring crack width and shrinkage in the axial direction. Additionally, for a 3D-printed single-layered hollow square of 500 mm × 500 mm, crack formations were monitored following a 7-day drying period.

2.4.4. Printability Tests

The printability of earthen mixes refers to the concrete’s capacity to maintain its shape and texture during extrusion and when subjected to external weight loads from other layers [59]. Essentially, it signifies the mix’s ability to be extruded or deposited in a controlled manner, forming the intended shape or structure without collapsing or compromising its structural integrity. Several factors can impact printability, including mix design, soil mix rheology, setting time, and curing conditions. A well-designed and properly formulated earthen mix with good printability ensures that the printed structures meet the required mechanical properties, durability standards, and surface finish. This study assesses extrudability and buildability, two vital aspects of fresh earthen materials’ printing properties.

Printing System

A gantry-type printer with a steel frame of 2 m × 2 m × 2 m was used for printing, as shown in Figure 5. The printer used G-code, a programming language that converts the 3D model into slices and prints them layer by layer. Throughout the research, printing parameters such as extrusion speed, printing speed, soil matrix mixing speed, nozzle size, layer height, and layer width were kept consistent, as shown in

Table 3. Printing parameters for the extrudability and buildability tests.

Extrusion speed	0.15 rounds/s
Printing speed	20 mm/s
Soil matrixes mix speed	30 mm/s
Nozzle size	20 mm
Filament height	10 mm
Filament width	10 mm

. However, a nozzle with a diameter of 20 mm was used in this study. This size was chosen based on the fiber length, as determined during preliminary investigations. The printing temperature was maintained at 23 ± 2 °C.

Extrudability Test

The extrudability refers to its capability to produce continuous filaments of desired dimensions from the fresh plastic mix through the nozzle The extrudability of a printer refers to its capability to produce continuous filaments of desired dimensions from the fresh paste through the nozzle [22,59]. Extrudability evaluation was conducted manually via the 3D-printing of a single-layered filament with a width of 20 mm and a layer height of 10 mm in a square path with an area of 500 mm × 500 mm, as illustrated in Figure 4f and Figure 6a. The other printing parameters were followed according to. The extrusion process was tested for all the mixed designs, and the layer width was measured at 20 different locations (5 on each side of a square) along the square path. The average value of the measured filament width was compared to the required filament width, and any deviations from the required width were measured. Similarly, during the printing process, the printing path and quality were observed to ensure continuous and accurate printing from the nozzle.

Buildability Test

The buildability test can evaluate how well a printed filament can support the deposited layers above it while retaining its dimensions [22,59]. A filament layer with a width of 20 mm and a height of 10 mm was considered to print the wall and cone geometries, as shown in Figure 6b. Consequently, the buildability properties were assessed in three ways:

(1)

Determining the maximum number of stackable layers before total collapse for the single filament layer, as illustrated in Figure 4g and Figure 6b. The wall was 300 mm in length and 20 mm in width, and the count of stackable layers was recorded.

(2)

Assessing the deformation in the bottommost layer filament of the same single-filament wall printed up to a maximum of 12 layers. Deformations in the bottommost layer were recorded after 7 days of thorough drying (to ensure accurate measurement). Ten readings were collected, and their average deformations, along with standard deviations, were analyzed.

(3)

3D-printing a circular cone with the mix SLFP for geometrical exploration purposes, as shown in Figure 6b.

Figure 6. Printability tests (a) extrudability evaluation (b) buildability assessment.

Figure 6. Printability tests (a) extrudability evaluation (b) buildability assessment.

3. Results and Discussion

3.1. Characterization through XRD and XRF Analysis

The results of XRF analysis indicated that almost two-thirds of the chemical composition of the soil was made up of siliceous material, as illustrated in

Table 4. Elemental composition of soil and natural pozzolana.

Materials	SiO2	Al2O3	CaO	Fe2O3	MgO	TiO2	Na2O	Others
Soil-S	63.00%	15.50%	8.10%	5.70%	1.70%	0.96%	0.53%	4.58%
Tephra P	71.80%	13.30%	0.70%	0.70%	-	-	-	13.50%

. In addition, the chemical composition of the pozzolana was determined by the CR minerals company, and showed a high content of silica. XRD analysis was used to identify the clay mineral in the soil with the flow diagram as recognized by the US Geological Survey [60]. The clay under examination using the drying method in XRD was initially identified as montmorillonite, as illustrated in Figure 7. However, following treatment with ethylene glycol, the clay reduced to around 10 Å, accompanied by a faint peak at 5 Å, suggesting a potential nontronite composition.

Further XRD analysis revealed that the collected soil comprised kaolinite, nontronite, and illite clay minerals. Kaolinite was distinguished from dickite or nacrite by a peak at 19.74 Å/4.5 Å, and was distinguished from the chlorite (001) peak after glycolation, as shown in Figure 8. Nontronite was distinguished from montmorillonite by a peak at 1.52 Å in a randomly oriented sample scan and a 5 Å peak after heating at 400 °C, as shown in Figure 9. The illite peak became more visible in the glycolate mount as the nontronite (001) peak shifted to the left.

The quantitative analysis showed that clay minerals had the composition shown in

Table 5. Clay types, their mineral composition, and weight proportion.

Clay Mineral	Chemical Formula	Composition by %
Kaolinite	Al2Si2O5(OH)4	49.3
Nontronite	(CaO0.5,Na)0.3Fe3+ 2(Si,Al)4O10(OH)2·nH2O	14.6
Illite	(K,H3O)(Al,Mg,Fe)2(Si,Al)4O10[(OH)2,(H2O)]	36.1

. The presence of the nontronite clay mineral in the soil sample justifies the observed shrinkages during the preliminary tests for 3D printing, as detailed in the preliminary study [27]. Despite being present in relatively low proportion, nontronite, a member of the smectite clay mineral group [61], possesses unique swelling and sinking properties that contribute significantly to the overall behavior of the soil mix. In contrast, other clay minerals, such as kaolinite, have no swelling potential, and illite has less pronounced expanding properties than nontronite when exposed to water [61]. The structure of kaolinite is fixed by the hydrogen bonding between the layers, causing it to have low shrinking or swelling properties when wet [61]. Similarly, the illite mineral, categorized in the mica group, has K⁺ ions in the interlayer space, balancing the charge deficiency in the layers and preventing its expansion [61]. This distinctive characteristic of nontronite makes the soil mix more susceptible to shrinkage and increases its water holding capacity [61,62]. To address these issues, adding hydrated lime can be considered to reduce the plasticity properties of the soil and mitigate its swelling and shrinking behavior. As discussed in our preliminary study [27], this soil stabilization process is crucial for achieving the desired structural integrity and printability of the soil mix for 3D printing applications [61]. This distinctive characteristic of nontronite makes the soil mix more susceptible to shrinkage and increases its water holding capacity [61,62].

3.2. Fresh Properties

For the flow table test results within the range of 14 cm to 19 cm, as shown in

Table 6. Flow table test result.

Mix ID	Water-to-Dry-Mix Ratio	Flow Value (%)
S	0.380	39.76
SL	0.415	79.13
SLF	0.427	68.31
SLFP	0.430	79.13

, SL and SLFP had the maximum flow table results, indicating more flowability and greater workability than the other two mixes. The results also showed that flow values increased with the water-to-dry-mix ratio in Mix S. Adding a stabilizing agent (i.e., hydrated lime) increased the water demand to form a suitable printable mixture. Similarly, adding fibers and NP required more water for proper hydration and flowability. Additionally, adding water to the SL mix might have compensated for using dry straw fiber for flowability. Similarly, it was observed that adding water to the dry mix ratio in SLFP led to a higher flow value. Overall, the results showed that an increment in the water-to-binder ratio increased the flow value, reaching a higher flowability.

Figure 7. XRD results from the air-drying method.

Figure 7. XRD results from the air-drying method.

Figure 8. XRD results after heating the materials.

Figure 8. XRD results after heating the materials.

Figure 9. XRD results of a randomly oriented slow scan.

Figure 9. XRD results of a randomly oriented slow scan.

The hysteresis loop of the rheology tests showed that the soil mixes followed the Bingham model [53,54,55], as shown in Figure 10. Therefore, static yield stress (${\mathsf{\tau }}_s$) was calculated from the peak value from the shear rate from 0 S⁻¹ to 100 S⁻¹. The shear stress at the zero rates in the loop descending part was calculated for the dynamic yield stress (${\mathsf{\tau }}_d$) [55]. Thixotropy was calculated as the areas between the curves between the increasing shear rates and decreasing shear rates integrated between shear rates 20 S⁻¹ to 80 S⁻¹, as described by previous studies [53,54,55]. The results of rheology tests (i.e., static yield stress, dynamic yield stress, thixotropy, and plastic viscosity) of four soil mixes are shown in

Table 7. Rheology results of soil mixes.

Mix	Static Yield Stress (Pa)	Plastic Viscosity (Pa/s)	Dynamic Yield Stress (Pa)	Thixotropy (Pa/s)
S	1485.0	4.1	354.9	11,132.5
SL	2582.9	8.3	84.1	32,766.0
SLF	1169.4	2.8	292.7	268.8
SLFP	1450.0	3.8	182.3	24,182.8

and Figure 10.

As reported in Table 7, the SL mix had the highest static yield stress, followed by S and SLFP mixes. This means the SL mix required the highest stress to change its rheology from static stage to flow ones. The plastic viscosity of SL increased with lime (SL) and had a maximum value of 8 Pa/S. This means that SL had the highest resistance to flow once it started moving. The S mix was found to have the highest dynamic yield stress, and SL was found to have the lowest. This means that mix S had the highest viscosity, causing it to require more stress to keep the soil matrix moving. In addition, it was found that the dynamic yield stress decreased by adding lime and more water to the dry mix ratio. This shows that lime and water content were important in stabilizing the mix S.

SLFP mix exhibited the highest thixotropy at 24,182.8 Pa/s, while SLF mix demonstrated the lowest thixotropy at 268.8 Pa/s. According to this result, the pozzolana could not provide enough binding or stabilization to prevent the soil mixture from flowing under stress. Furthermore, the high thixotropy for SLFP mixtures demonstrated that the materials take longer to recover their structure after shearing or disturbance.

3.3. Hardened Properties

3.3.1. Drying Shrinkage Results

Figure 11 and Figure 12 illustrate the shrinkage and cracking patterns observed in dried cast prisms and printed hollow squares. The substantial drying shrinkage in S and SL prisms led to their failure, characterized by extensive visible cracks, as shown in Figure 11. A similar trend was evident when these mixes were 3D-printed, as depicted in Figure 12. Hence, the significant cracking and drying shrinkage observed in S and SL mixes were deemed unsuitable for construction purposes. The substantial shrinkage observed can be attributed to the presence of 15% nontronite clay mineral in the local soil. Nontronite, characterized by its unstable structure, is susceptible to both swelling and shrinking phenomena. This mineral possesses a layered structure, facilitating water absorption and expansion. When water infiltrates between these layers, nontronite swells, causing an increase in volume. Conversely, when deprived of water, it undergoes shrinkage, contributing to the observed shrinkage in the soil [61].

Including fibers in SLF effectively mitigated excessive shrinkage, resulting in reduced cracking and drying shrinkage. The fibers played a crucial role in maintaining specimen cohesion and preventing the formation of wide cracks induced by substantial drying shrinkage for both cast and 3D-printed ones. Likewise, introducing fiber into SLFP led to shrinkage reduction and fewer cracks. The length shrinkage observed in cast specimens was 7.7% for SLF and 2.6% for SLFP. Additionally, the maximum crack widths on extruded specimens were 0.95 mm for SLF and 0.46 mm for SLFP.

3.3.2. Compressive Strength

Figure 13 presents the compressive strength of various mixtures following 7 and 28 days of curing. Interestingly, the strength levels remained nearly consistent between both ages (7 days versus 28 days) for all mixes. Natural soil (S) displayed the highest strength among the studied mixtures (i.e., 6.17 MPa), while SLF exhibited the lowest (i.e., 0.51 MPa). These findings align well with previous research [27]. The addition of fibers increased the voids within the mixtures, reducing the strength of the soil cubes. Lime and pozzolana were not as effective additives, as expected; their inclusion reduced the compressive strength of the materials. Additional research is needed to identify the optimal combination of additives to effectively reduce significant shrinkage without compromising the material’s strength, according to compressive strength and drying shrinkage results. In contrast to expectations, incorporating hydrated lime or pozzolana as soil stabilizers did not result in enhanced strength or shrinkage prevention. This highlights the necessity of additional research to discover appropriate additives capable of improving soil mixes’ mechanical characteristics while successfully addressing the issue of substantial shrinkage and crack formation.

3.3.3. Flexural Strength

Figure 14 and Figure 15 depict the flexural strength of various soil mixes. As expected, all the soil prisms displayed a brittle behavior with limited ductility, as depicted in Figure 14 and Figure 16. Like the compressive strength results, the pure natural soil (S) exhibited the largest MoR (4.15 MPa) but the lowest deflection (0.64 mm). Additionally, adding fibers enhanced the flexibility of soil prisms and resulted in the greatest deflection for the beam test (1.57 mm at the peak load). As previously observed in the compressive strength tests, lime and pozzolana did not contribute to increased strength.

3.4. Printability Tests

3.4.1. Extrudability

Figure 17 illustrates the printed hollow squares for various soil mixes, assessing their extrudability by examining the absence of cracks/discontinuity immediately after extrusion, and the dimensional stability of the printed filament. All the studied mixes demonstrated successful filament extrusion, free of initial cracks, without any discontinuities immediately after extrusion from the nozzle head. Visual observation revealed that the highest print quality was achieved with the S mix, which contained no additives, while the lowest print quality was observed in mixes that included fibers.

Furthermore, measurements of filament width were taken at 20 different points along the filament length immediately after printing, and the results are presented in Figure 18. The width measurements revealed that mix SLF exhibited the highest standard deviation, indicating lower uniformity and greater variation in filament width size. This observation can be attributed to the introduction of fibers, which altered the consistency of the soil mix and resulted in this deviation in width measurements.

The intended filament width of 20 mm was not attained for any of the mixes due to the inherent flowability of the mixes during printing. Furthermore, the greatest deviation from the target values was noted in mixes containing fibers, which can be attributed to their higher water demand (SLF, SLFP), leading to over-extrusion, as clearly shown in Figure 18.

In summary, although all mixes achieved a consistent extrusion through the nozzle head, the inclusion of additives like lime, fibers, and pozzolana noticeably influenced the final printing width due to their heightened requirement for additional water content to ensure proper flowability, as indicated in Table 2. Moreover, the rheometer data support the notion that increased thixotropy enhanced the flowability of SLFP, resulting in a greater volume of extruded material and wider layer dimensions compared to other mix designs.

3.4.2. Buildability

As described in the test methods, the assessment of buildability involved three key evaluations:

(1)

Determining the maximum number of stackable layers before structural collapse for a single-filament wall, the results of which are displayed in

Table 8. Stackable layers of 3D-printed single walls for buildability evaluation.

Mix ID	Water-to-Dry-Mix Ratio	Maximum Stacking Layers
S	0.380	18
SL	0.415	12
SLF	0.427	14
SLFP	0.43	12

. The results clearly indicated that the S mix demonstrated the highest number of stackable printed layers before collapsing, aligning well with the water content in the soil mixes. Mixes with greater water requirements resulted in shorter single-filament walls.

Figure 18. The average width of the filaments for the soil mixes, along with their standard deviation.

Figure 18. The average width of the filaments for the soil mixes, along with their standard deviation.

(2)

Evaluating the deformation in the bottommost layer filament of the same single-filament wall, which was printed with a maximum of 12 layers after drying, the results of which are illustrated in Figure 19 and Figure 20. The S mix exhibited the least deformation, while the SLFP mix had the highest average deformation in the lowermost filament layer, 3.67 mm. This deformation was over 50% greater than that observed in mix S. Furthermore, the results obtained from the flow table tests indicated that the SL and SLFP mixes had significantly higher flow values than S and SLF mixes. The outcomes of this evaluation align with the number of stackable layers in Table 8, where larger deformations in the lowermost layer corresponded to shorter single walls and fewer stackable layers. Furthermore, Figure 19 illustrates the 3D printing quality after the fresh mix has undergone drying. As previously noted in the extrudability section, the S mix exhibits the highest printing quality following the drying process.

(3)

Printing a circular cone using the SLFP mixture for exploratory purposes related to geometrical considerations, as depicted in Figure 21. This cone was printed primarily to assess the feasibility of 3D printing various geometries, including the hollow cone. However, it is worth noting that SLFP resulted in the shortest 3D-printed single wall and the largest deformation in the bottom layer, even though it successfully printed the cone up to 15 layers. Further research is needed to systematically evaluate geometry’s impact on the printing process.

4. Conclusions

This study evaluates the feasibility of using local materials, including soil, in the construction, in terms of printability, fresh and hardened properties. One area of importance is the effect of different local and natural admixtures, including lime, natural fiber and pozzolana, on the material performance of developed mixes. The preliminary phase of this study determined the proportion of these additives, and four soil mixes were designed. The key findings from this study are as follows:

• XRD analysis identified the clay composition, which included 49% kaolinite, 15% nontronite, and 36% illite clay minerals. Identifying nontronite clay, known for its volumetric instability, contributes to the soil mix’s higher susceptibility to shrinkage and greater water retention capacity.
• The soil mixture composed solely of pure soil demonstrated the highest printing quality in both its fresh and hardened states. It successfully produced the tallest single-filament 3D-printed wall, consisting of 18 layers, with minimal deformation in the bottom layer. Furthermore, it exhibited superior mechanical performance in terms of compressive (6.17 MPa) and flexural strength (4.15 MPa). However, due to the high nontronite clay content in the soil, this mixture experienced cracking and significant drying shrinkage after the drying process.
• Incorporating various additives (lime, pozzolana, and fiber) raised the water demand necessary to achieve a printable mixture. Consequently, this led to a shorter single filament 3D-printed wall, increased filament width, and greater deformation in the bottommost layer of the printed wall due to the higher water content.
• The incorporation of fibers proved to be highly effective in mitigating excessive shrinkage and reducing both cracking and drying shrinkage. These fibers played a pivotal role in enhancing the cohesion of the specimens and preventing the formation of wide cracks induced by substantial drying shrinkage, whether in cast or 3D-printed specimens. Additionally, the introduction of fibers into the SLFP mix resulted in significant shrinkage reduction and fewer cracks compared to the SLF mix. Specifically, the length shrinkage observed in cast specimens was 7.7% for SLF, and notably reduced to 2.6% for SLFP. Furthermore, the maximum crack widths on extruded specimens decreased from 0.95 mm for SLF to 0.46 mm for SLFP. This demonstrates the beneficial impact of fiber incorporation on the structural integrity and reduced shrinkage of the studied soil mixes.
• Contrary to expectations, the inclusion of lime or pozzolana as a soil stabilizer did not lead to improvements in strength or the prevention of shrinkage. This suggests that further research is needed to identify suitable additives that can enhance the mechanical properties and printability of soil mixes while effectively mitigating large shrinkage and crack development.
• The authors of this paper suggested further research about the possibility of using eco-friendly polymers for shape retention and early age cracking should be carried out. Some potential options are as follows: xanthan gum, guar gum, and anionic polyacrylamide.