3D Printing of Earth-Based Mixtures: Linking Material Design, Printability, and Structural Performance

Abstract

The advancement of sustainable construction requires the development of earthen materials compatible with 3D printing (additive manufacturing), along with specified engineering standards. Many existing studies improve workability and early strength using chemical stabilizers such as cement; however, these additives increase embodied carbon and undermine sustainability objectives. Challenges remain in the formulation of an earthen mixture that satisfies both printability and structural requirements for large-scale construction. Previous earth-based mixes have reported excessive shrinkage and inadequate compressive strength. This study presents the systematic optimization of a low-carbon, 3D-printable earthen mixture using locally sourced clay-loam soil from Belén, New Mexico (NM). The soil was modified with graded concrete sand and rice hull fiber to improve printing parameters such as buildability, extrudability, and printability while meeting the NM Earthen Building Code requirements for compressive and flexural strength. Soil characterization tests (particle size distribution, consistency, optimal water content) guided iterative refinement to enhance dimensional stability and mechanical performance. A baseline 2:1 soil-to-sand ratio (max aggregate size No. 4) was established. Incorporating $2\%$ rice hull fiber and reducing max aggregate size to No. 16 (S67F2) early-age shrinkage was reduced from $12.33\%$ to $3.48\%$ ($72\%$ reduction) while maintaining a 28-day compressive strength exceeding 660 psi, more than twice the code minimum. The optimized mixture supported 24 printed layers without deformation, achieved 189 psi flexural strength (three times the code minimum), and produced a stable 2-ft-diameter dome with minimal cracking.

1. Introduction

3D printing (3DP) involves the additive deposition of materials, layer by layer, through a computer-controlled process. It has become a pivotal part of the fourth industrial revolution, influencing various sectors of manufacturing [1,2]. In recent years, 3DP has successfully expanded into the construction industry, producing structures such as houses, bridges, and bus stops [3,4,5,6]. This technology provides greater flexibility in architectural design, accelerates construction speed, and offers a more affordable, sustainable, and safer construction method by reducing human involvement, material transportation, and waste [5,7,8,9]. While concrete is commonly used in 3DP due to its flowability and mechanical strength, the environmental impact of cement-based products has raised concerns, leading to a growing interest in alternative sustainable materials such as local soil [4,6,10,11,12,13,14,15]. The production of concrete significantly contributes to global carbon emissions, with the cement industry responsible for approximately $8\%$ of total emissions. Additionally, 3D-printed concrete often requires a higher concentration of cement to ensure printability and extrudability, which further exacerbates its environmental footprint [12,16,17,18]. Given the increasing global population, engineers and architects are being called upon to adopt environmentally responsible design approaches that ensure resources for future generations [19,20,21]. This study investigates the use of local soil gathered in Belén, New Mexico (NM), for 3D soil printing (3DSP), aiming to develop a mechanically viable material that meets the requirements of the NM Earthen Building Code [22]. By utilizing local resources, the research aims to enhance the sustainability of construction practices while maintaining necessary structural integrity and durability.

Earth has served as a fundamental building material for millennia, and in NM, this tradition is deeply woven into its architectural heritage. From Indigenous Pueblo structures to historic churches and traditional homes, local soil has shaped the built environment in ways that reflect both practicality and cultural identity. Among these earth-based techniques, sun-dried bricks, commonly referred to as adobe, made from soil, water, and fibrous materials such as straw, have played a central role for centuries, supported by the high-quality soils of the Rio Grande Valley [19,23,24,25,26,27]. Today, New Mexico is home to nearly one-third of the nation’s earthen dwellings, and Taos Pueblo (recognized as one of the oldest continuously inhabited communities in the United States) stands as a living testament to this enduring legacy [23,28].

The success of these traditional methods lies in their remarkable properties: excellent thermal mass and insulation, affordability, and sustainability. Structures built with these materials are fireproof, termite-resistant, and durable, while remaining easy to repair [23,24]. The bricks themselves rely on a careful balance of sand, silt, and clay, with straw added to minimize shrinkage and cracking during drying [23,24]. Mesoamerican Indigenous peoples and Spanish influence in the 1600s introduced wooden forms, which streamlined production and allowed greater flexibility in size and weight [29]. The sandy loam soils of the Rio Grande Valley, long recognized for their suitability in earthen construction, provide valuable insights for developing soil mixes optimized for modern techniques such as 3DP.

Although earth-based construction materials have a long history, their application in 3DP is relatively new. Recent research has explored using materials similar to adobe and cob in 3DSP to enhance sustainability [6,12,13,15,30,31]. Some notable large-scale 3D-printed structures using local soil include projects by IAAC, WASP, and Emerging Objects, which have demonstrated the potential of this technology [32,33,34,35]. However, challenges remain in developing a soil mix that is both printable and mechanically strong enough for large-scale production, limiting the broader application of 3DSP in construction [13,15,36].

Unlike adobe construction, where individual bricks are molded and allowed to dry and shrink under ambient conditions before assembly, 3D printing with earthen materials requires the in-situ fabrication of a continuous monolithic structure, creating significant challenges in controlling shrinkage while achieving adequate strength. Adobe’s staged drying process confines volumetric changes to discrete units, reducing internal stresses and cracking, whereas 3DP must contend with global shrinkage across an entire wall system without the benefit of a pre-drying phase. This demands careful optimization of rheology, water content, and reinforcement to ensure extrudability, buildability, and dimensional stability while limiting differential drying and interlayer weaknesses. Current literature on 3DSP rarely addresses volumetric stability, despite its critical role in structural performance. Clay, a naturally variable geomaterial, exhibits pronounced mineralogical diversity that governs its behavior under moisture fluctuations. High clay content in earthen mixes leads to significant swelling and shrinkage during wetting and drying cycles, introducing differential movements that compromise dimensional accuracy, durability, and serviceability. These deformations pose a major challenge for 3DP applications, where precision and reliability are essential. Moreover, clay influences strength development; expanding clay minerals such as montmorillonite and nontronite generally provide higher strength compared to non-expanding clays like kaolinite and illite of similar composition [23].

Previous investigations conducted by our research group at the University of New Mexico (UNM) [25] involved characterizing locally sourced soils from multiple sites and evaluating their suitability for 3DP applications. The study assessed key geotechnical and rheological properties relevant to 3DP; however, significant limitations were identified, including excessive volumetric shrinkage measured at up to $7.7\%$ and inadequate compressive strength, which fell below the 300 psi threshold stipulated by the New Mexico Earthen Building Code [22]. The code does not specify a shrinkage maximum or minimum; it states that shrinkage cracks are allowed, provided that these cracks do not jeopardize the structural integrity of the blocks. The present work addresses these deficiencies by focusing on assessing material composition to mitigate shrinkage and enhance mechanical performance. Through systematic refinement of soil mix design, the objective is to achieve a balance between printability, structural integrity, and dimensional stability. This approach aims to advance the development of sustainable 3D-printed earthen construction systems that meet regulatory standards and are viable for community-scale implementation. The research concluded with the successful printing of a hemispherical dome, 2 ft (610 mm) in diameter, using the final optimized mix.

2. Materials and Methods

In our preliminary study on 3DP with local earthen materials, soil from Belén, NM, showed the most promise in terms of plasticity [26,27]. The soil for this study was sourced from a family property in Belén, where native vegetation was scraped off, and the top 6–12 inches of soil were removed. The local plain Belén soil (S) was collected below this layer. An on-site field test determined the approximate clay content by adding water to a walnut-sized soil sample and assessing its texture, cohesion, and clay content by hand, evaluating its ability to form a ball, snake, or ribbon shape [24,37,38]. The soil was transported to the lab, cleaned of organic debris, and manually crushed with a steel hand tamper to a fine powder for further characterization. The following section presents soil characterization, soil mix design, and test methods for casting and 3DP of designed soil mixes.

2.1. Soil Characterization

Particle size distribution is a critical factor in 3DSP because nozzle dimensions limit the maximum particle size, and the overall gradation directly affects material cohesion and performance. Soils rich in fine particles, particularly clays, exhibit unique strength and bonding characteristics, while also responding strongly to variations in printability in fresh mix and moisture content, which can compromise dimensional stability after hardening. For this reason, identifying the clay mineral content in local soils is essential, as it informs the development of proper mix designs for 3DSP. A well-graded particle size distribution not only enhances mechanical strength and stability but also ensures consistent extrusion and layer bonding, making it a key requirement for high-quality earthen materials in 3DP.

In our previous study [39], various grain size distribution tests were compared to determine the most effective method for characterizing local soils for 3DP. Detailed procedures are described in Zozaya et al. [39]. Results showed that a wet sieve test (Test 3) performed in the laboratory and a shake jar test (Test 4) conducted in the field provided practical and reliable assessments, correlating well with the properties of the designed soil mixes. These methods enable the rapid determination of sand, silt, and clay content in high-clay soils, facilitating adjustments to achieve a well-graded distribution for improved printability and mechanical strength. Both tests indicated approximately $61.4\%$ clay/silt content, demonstrating strong agreement, as summarized in

Table 1. Particle size distribution and classification of S and CS from different tests.

Gradation Test	Material	Gravel (%)	Medium to Coarse Sand (%)	Fine-Grain Sand (%)	Silt (%)	Clay (%)	Soil Classification
Wet Sieve	S	0.00	0.42	38.19	61.39		CL*
Shake Jar	S	0.00	38.55		30.12	31.33	CL*
Dry Sieve	CS**	14.86	38.59	46.20	0.36		SP***

Note: CL^* = Clay Loam; CS^** = Concrete Sand; SP^*** = Poorly Graded Sand.

. Based on the USDA classification chart, the soil (S) was identified as clay loam (CL) [40] and presented in Table 1 and Figure 1.

X-Ray Diffraction (XRD) analysis on the fine clay fraction of soil samples (passing through a No. 200 sieve) identified montmorillonite, kaolinite, illite, and quartz. A 1989 NM Bureau of Mines and Mineral Resources study [23] found that adobe soils in NM commonly contain a mixture of kaolinite, illite, smectite, and mixed-layer illite/smectite. The study noted that kaolinite and illite (nonexpandable characteristics) made up $50\%$ of the clay minerals, while smectite and mixed-layer illite/smectite (expandable characteristics) accounted for the other $50\%$. This mineralogy correlates with the XRD results of the current study, indicating suitability for 3DSP [26,27]. The study also indicated that clay-size fractions high in expandable clay minerals relative to nonexpandable clay minerals have been shown to increase the compressive strength of adobe soils, which can also correlate to 3DSP mix design and development [23].

2.2. Soil Mix Design

In our previous work [26,27], local soils were evaluated for 3DSP with an emphasis on printability and material performance. The study explored regional admixtures, including type-S hydraulic lime, wheat fiber, and pozzolana, which helped reduce shrinkage but resulted in decreased compressive strength. Building on those findings, the current mix design seeks to address these limitations by improving structural stability and printability. The approach involves reducing clay content through controlled particle-size distribution, achieved by incorporating non-clay aggregates and introducing new natural fibers. Specifically, a portion of the original soil (S) was replaced with concrete sand (CS) to lower clay content while maintaining compressive strength above the 300-psi minimum required by the NM Earthen Building Code [22]. Lowering clay content is also expected to enhance extrusion consistency and overall printability for 3DP applications. This section presents two mix sets (Set 1 and Set 2) developed for this purpose: the first combines plain soil (S) with CS (max aggregate size No. 4 sieve), and the second builds on the previously identified printable mix by lowering the max aggregate size of CS to a No. 16 sieve and integrating rice hull fibers to improve performance further.

2.2.1. Set 1 Mixes: Enhancing Soil Mix Design by Substituting Local Soil with Concrete Sand

To improve the grain size distribution of local soil (i.e., S), CS was obtained from Vulcan Materials Company, a local producer and distributor of construction aggregates located in Placitas, NM, USA. CS is used in concrete mixes, so it should be free of fine particles according to ASTM standards [41]. Accordingly, performing a dry sieve test (ASTM C136 [42]) is sufficient to determine the size distribution of the CS. The CS was washed at the plant, and a standard dry sieve analysis (ASTM C136 [42]) was conducted. Due to the size of the chosen printer nozzle, 20 mm in diameter, the CS was further sieved in the lab to achieve a maximum aggregate size of sieve No. 4 (i.e., 0.187 in., or 4.75 mm), the smallest sieve through which $100\%$ of the sample must pass. The grain size distribution results of the sieved CS are shown in Table 1 [39,40,43,44] and Figure 1.

Five soil mixes were formulated to control the replacement of S with CS. Each mix replaced varying amounts of soil with CS, as detailed in

Table 2. Soil mix design for S replacement with CS by weight percentage in Set 1.

Mix ID	Plain (Belén) Soil, S (%)	Concrete Sand, CS (%)	Clay/Silt Content (%)	Liquid Limit (%)	Water/(S+CS) (%)
S100	100.00	0.00	61.39	36.12	36.00
S75	75.00	25.00	46.67	35.50	32.00
S67	67.00	33.00	41.89	33.21	25.33
S50	50.00	50.00	31.61	32.56	23.00
S25	25.00	75.00	16.17	25.03	16.00

. The mix ID denotes the percentage, by weight, of S in each formulation. For instance, S75 indicates that the soil mix weight comprises $75\%$ S and $25\%$ CS. The literature supports this substitution [15,23,37,45,46], suggesting that adobe mixes typically require between $15\%$ and $45\%$ clay. Referring to Table 1, which indicates S contains a clay/silt content of approximately $61.4\%$, efforts were made to achieve a clay quantity comparable to that used in traditional adobe soil mixes. The clay/silt content by weight ranges between $16.2$–$61.4\%$ in S25 to S100, respectively.

To better understand the particle size distribution of each designed soil mix, we combined results from the wet sieve test for S [44] and the ASTM C136 dry sieve test for CS [42]. Figure 1 illustrates the particle size distribution of the five soil mixes, and Table 2 summarizes the weight percentage and clay/silt content for each mix. In Figure 1 particles finer than the No. 200 sieve (0.075 mm) were not further analyzed, representing the clay fraction in each mix. As shown in Figure 1 and Table 2, coarser soil mixes were achieved by increasing the CS content, reducing the silt/clay content from $61.39\%$ in S100 to $16.17\%$ in S25. Adding CS also improved soil gradation, transitioning from uniformly graded to well-graded soil.

Beyond particle size distribution, accurately determining the optimal water content for each mix is critical. While previous studies [27] explored flow table tests and rheology measurements for this purpose, flow tables have limitations in the context of soil mixes for 3DP. This paper tested the Liquid Limit (LL) of each new soil mix according to ASTM D4318 [47]. This method determined the approximate baseline water content needed for each mix, as displayed in Table 2. Determining the optimal water content for each soil mix was crucial for successful 3DP. Two methods were employed depending on the batch size and the ambient conditions. For small batches in a controlled lab environment (i.e., 73 °F (23 °C) ± 2° and an average indoor humidity of 16 ± $2\%$), the LL value served as a starting point. This value reflects the maximum water content where the soil behaves like a liquid. However, adjustments were made based on the results of the traditional consistency test, reported in the last column (“Water”/“S+CS” ) of Table 2. Figure 2 displays the workflow diagram of the consistency testing method. This test allowed for a more delicate evaluation of the desired consistency for 3DP by considering a balance between smooth extrusion and adequate structural integrity. The chosen method for determining the water content of each soil mix is based on a traditional texture-by-feel consistency test established from the author’s experience and literature for testing soil and adobe material in the field as a guide [37,39,48]. This test offers a straightforward approach to finding the required water content. In this test, some amount of water below the LL was added to the soil, and a sample of the soil mix was kneaded and rolled into a ball approximately 1.5–2 in (37.5–50 mm) in diameter and a sausage shape approximately 4 in (100 mm) in length and 0.78 in (20 mm) in diameter, as displayed in Figure 3a,b. The water content was determined by examining the consistency and texture of the mix in the ball and sausage form according to the following three states:

• If the material crumbled and could not be rolled into a cohesive ball, the water content was insufficient. In this case, water was added incrementally in 5 mL steps, followed by re-kneading and re-testing;
• If the material cohered and could be rolled into a stable ball without cracking, the corresponding water amount was recorded as the optimal water content (Figure 3a);
• If the material became excessively sticky or exhibited slumping behavior, the water content was considered excessive (Figure 3b).

For instance, the LL measured by the ASTM standard for mix S75 was $35.5\%$, when the material was tested in a small batch; at this water content, the material was too watery and lacked stability, as displayed in Figure 3b. Therefore, to determine the optimum water content needed for 3DSP of a small batch under the controlled conditions of the lab, 500 g of the S+SC mix was weighed, and an amount below the base line liquid limit, $20\%$ (100 mL) of water was initially added. 5 mL of water was incrementally added to the mix until the material could be rolled into a ball and a sausage easily, which occurred at $32\%$ (160 mL), as displayed in Figure 3a and reported for the various soil mixes in the last column (“Water”/“S+CS” ) of Table 2. It was found that the more CS was added to S, the lower the optimum water content was needed compared to the baseline LL determined, most likely due to the decrease in clay content.

For larger soil mix batches intended for 3DP, a different approach was necessary to determine optimal water content. The values established under controlled lab conditions proved insufficient during large-scale printing, where materials were exposed to ambient temperature and environmental variations. Time needed to prepare and mix the large batches could also play a factor. An initial estimate, based on the traditional consistency test (Table 2), was refined during mixing using a penetration technique (Figure 3c) with a portable digital penetrometer hardness tester (VTSYIQI, Hefei, China) [49]. This method enabled precise adjustments to achieve the optimal water content required for printability under real-world conditions and varying nozzle sizes.

2.2.2. Set 2 Mixes: Enhancing Soil Mix Design by Including Natural Fibers

In this phase, Set 2 soil mixes were developed to reduce the shrinkage observed in previous trials while considering nozzle size limitations. The S67 mix (containing CS with a maximum aggregate size of No. 4 sieve) was selected for further evaluation based on successful preliminary printability tests. To enhance printability, the CS in Set 2 of mixes was sieved to a finer maximum aggregate size of No. 16 (0.046 in./1.18 mm). Figure 4 shows the grain size distribution of the improved S, CS, and S67 mixes used in the new formulations to assess the addition of natural rice hull fiber. Set 2 of the mixes incorporated rice hull fiber sourced from Victor’s Home Brew Supply in Albuquerque, NM. Rice hulls, a byproduct of rice grain processing, constitute approximately $20\%$ of the rice grain weight and are considered agricultural waste due to their poor nutritional value, low bulk density, and high ash content, which contributes to environmental pollution [50,51].

For this study, rice hull fibers with an average length of 8–10 mm were selected for their uniformity, potential to enhance sustainability by recycling and reutilizing hull waste for 3DSP, and ability to reduce shrinkage using a natural product. Fibers were added to the soil mix by weight at four fractions ($0.5\%$, $1.0\%$, $1.5\%$, and $2.0\%$), following previous research [12,25,52,53]. This second mix set consisted of five formulations based on mix S67 (maximum aggregate size No. 16) with increasing fiber content, as shown in

Table 3. Soil mix composition (by weight) incorporating rice hull fibers as a partial replacement.

Mix ID	Plain (Belén) Soil, S (%)	Concrete Sand, CS (%)	Fiber Content (%)	Clay/Silt Content (%)	Water/(S+CS) (%)
S67–F0.0	67.00	33.00	0.00	45.87	22.00
S67–F0.5	66.75	32.75	0.50	45.64	22.00
S67–F1.0	66.50	32.50	1.00	45.41	22.00
S67–F1.5	66.25	32.25	1.50	45.19	22.00
S67–F2.0	66.00	32.00	2.00	44.95	22.50

. Mix IDs indicate the base mix (S67) and the fiber percentage; for example, S67-F1.0 represents a dry mix comprising $66.5\%$ S, $32.5\%$ CS, and $1.0\%$ rice hull fiber. Optimum water content was estimated using the traditional texture-by-feel consistency test described earlier and is reported in the last column (“Water”/“S+CS” ) of (Table 3).

2.2.3. Mixing Procedure for Soil Mixes

For small-scale soil batches, each soil mix was blended in a small KitchenAid mixer. First, the dry materials were measured by weight and added to the mixing bowl. The dry ingredients were mixed at the lowest mixing speed (60 ± 5 rpm), and the specified amount of water was gradually added while mixing for 2 min. After mixing at low speed, the mixer was paused, and the soil mix material was scraped off the sides of the mixing bowl with a spatula. The mixing continued at medium speed (140 ± 5 rpm) for 2 min. After mixing at medium speed, the mixer was paused, and the material was scraped off the walls. The mixer was allowed to continue mixing at high speed (280 ± 5 rpm) for the final 2 min. The total combined mixing time was 6 min.

In contemporary adobe construction, large stockpiles of suitable soil are mixed using front-end loaders, water is added, and the mixture is typically left to soak overnight in a mud pit before being transferred to molding forms via loaders or conveyor pugmills [23]. While this approach could serve as an alternative for large-scale 3DSP, it requires further evaluation and testing. Due to the lack of access to such equipment, the dome-scale 3DSP in this study relied on manual hand-mixing utilizing a hoe in a wheelbarrow to mix the soil and water to achieve the desired soil consistency. Although effective for the dome scale, hand mixing becomes highly labor-intensive and impractical as project scale increases, underscoring the need for automated mixing solutions for large-scale 3DP.

2.3. Test Methods

After formulating the soil mixes and establishing a consistent mixing process, a series of tests was conducted to evaluate their suitability for 3DSP. These tests focused on printability, mechanical strength, and dimensional stability (linear shrinkage measurement). Initially, established methods for assessing cast soil specimens were reviewed, including compressive strength tests on soil blocks and shrinkage tests on soil bars. Based on these evaluations, the most promising soil mix for 3DP was identified. This mix then underwent printability assessments, including extrudability (the smooth flow of material through the printer nozzle) and buildability (the ability to form and maintain desired shapes during printing). These evaluations ensured the successful printing of a hemispherical dome with a diameter of 2 ft (610 mm) in the final stage of the study. Detailed descriptions of all test methodologies are provided in the following sections.

2.3.1. Measurement of Compressive Strength in Cast Soil Cubes

The compressive strength of earthen material must meet the minimum requirement of 300 psi, as specified by the NM Earthen Building Code [22]. To evaluate this, the compressive strength of two sets of mixes was assessed following ASTM C109 guidelines [54], used here as a reference since no standard exists for soil materials intended for 3DP. Each mix was prepared using the previously described small batch mixing method and placed into 50 mm × 50 mm × 50 mm (2 in × 2 in × 2 in) plastic molds in three equal layers. Each layer was gently compacted with a rectangular plastic probe, and the mold was tapped after each lift to release air bubbles. Three specimen cubes were produced for each mix for testing at 7 and 28 days, with the 7-day interval selected based on the required drying time for soil mixes. For curing, molds were left uncovered, and side walls were removed on the third day to facilitate drying. Specimens were cured under ambient laboratory conditions (average temperature: 73 °F [23 °C]; indoor humidity: 16 ± $2\%$). Compressive strength testing was performed using a Forney high-strength concrete testing machine (Zelienople, PA, USA). Each cube was weighed and measured prior to testing. The load was applied at a rate of 0.00066 in/s (1 mm/min) with a ramp rate of 50 psi/s.

2.3.2. Measurement of Shrinkage in Cast Soil Samples

Shrinkage characteristics of each mix were evaluated following ASTM C490 [55], by measuring length change. The mixed material was placed into steel prism molds measuring 10 in × 1 in × 1 in (254 mm × 25 mm × 25 mm). To remove air bubbles, the material was compacted in two equal lifts using a rectangular plastic probe with a tamping motion. After three days, the mold walls were removed to facilitate drying. Specimens were cured under ambient laboratory conditions (average temperature: 73 °F [23 °C] ± 2°; relative humidity: 16 ± $2\%$). For each mix, three specimens were prepared, and shrinkage was assessed as the average length change across the replicates. Length measurements were taken daily for seven days, as specimens typically reach complete drying within this period, minimizing further volume changes. Measurements were recorded on opposite sides of each specimen and averaged to account for variation. In cases of cracking, the total length was determined by summing the lengths of all intact pieces, excluding crack gaps. Final shrinkage values were calculated according to Equation (1).

$${\epsilon }_{\mathrm{sh}–\mathrm{l}}\left(t\right)={\displaystyle \frac{l_t-l_0}{l_0}}\times 100,$$

(1)

where,

${\epsilon }_{\mathrm{sh}–\mathrm{l}}\left(t\right)$ = linear shrinkage of the soil prism in % at the age of t;

$l_t$ = length of the prism at the age of t;

$l_0$ = length of the prism after casting.

2.3.3. 3D Printability Tests and Mechanical Properties of 3D-Printed Soil Samples

In this section, we will detail the 3D printer used for our soil sample tests. The 3D printer utilized at the Dana C. Wood Materials and Structures Lab at UNM is a gantry 3D printer with a 2 m × 2 m × 2 m frame with 3 degrees of freedom in the x, y, and z planes. The system is controlled by a computer control unit that interprets and converts STL files into G-code. The system uses Cartesian coordinates with no rotational movement. The nozzle system features a small hopper for feeding material and a rotating auger that extrudes the material downward from the nozzle head onto the printing bed. The circular nozzle head can be removed and replaced with different-sized nozzles. A 20 mm nozzle was used in this study, enabling precise extrusion of the soil mix. The nozzle system is guided by drive motors that move in horizontal and vertical directions along moving guide rails. The printer’s extrusion speed can be adjusted between 1 cm/s and 5 cm/s, and between 1 rad/s and 5 rad/s, respectively, providing flexibility for different material properties and printing requirements. Figure 5 is an image of the 3D printer setup, showcasing all its components, including the nozzle, guide rail, control unit, drive motor, and printing bed [56]. The following subsections will explain the procedures for extrudability and buildability tests. These tests are crucial for determining the feasibility of using soil mixes in 3DP applications.

Extrudability Tests

Extrudability is defined as the ability of the soil mix to be continuously extruded through the nozzle without clogging, flow interruption, or cracking, while maintaining dimensional accuracy. To evaluate this property, the soil mix was extruded using a 3D printer at varying print speeds to identify optimal parameters. A 20 mm nozzle was selected based on particle size considerations, and its printability was assessed using a zigzag pattern consisting of six segments. Printing was performed with a constant layer height of 10 mm and an extrusion rate of 0.15 rounds/s. Each segment was printed at different speeds (10, 15, 20, 25, 30, and 35 mm/s) to determine the speed that produced a filament width closest to the design specification of 20 mm. Figure 6a illustrates the schematic of the extrudability test setup and provides an example of the printed pattern.

Buildability Tests

Buildability is defined as the ability of the soil mix to maintain geometric stability when printed in successive layers without collapse or significant deformation. The assessment was conducted by printing a wall segment measuring 20 mm × 300 mm and recording the maximum achievable layer height prior to failure. Printing parameters, including nozzle diameter and speed, were consistent with those established during the extrudability test. Figure 6b illustrates the buildability test setup and provides an example of a 3D-printed wall.

2.3.4. Flexural Strength of 3D Printed Beam

Before scaling up to 3D printing a dome, a beam was fabricated using the final mix, S67F2, which demonstrated acceptable printability and reduced shrinkage compared to other mixes. This beam was tested for flexural strength using ASTM C293 [57] as a reference for soil-printed specimens. The initial printed dimensions were 200 mm × 40 mm × 40 mm; while still wet, the specimens were trimmed to 160 mm × 40 mm × 40 mm to correct any shape irregularities from the printing process and achieve the precise rectangular geometry specified in the standard. Flexural testing was performed using a three-point bending setup, with the load applied at mid-span at a displacement rate of 0.1 mm/min. The modulus of rupture (MoR) was then calculated according to Equation (2).

$$\sigma ={\displaystyle \frac{3Pl}{2b{d}^2}},$$

(2)

where,

P = Axial load applied on the beam midspan;

l = Beam length from support to support.

2.3.5. 3D Printability of Soil Samples on a Larger Scale

To evaluate the printability and volumetric stability of the optimal soil mix identified in the previous section, a hemispherical dome with a diameter of 2 ft (610 mm) was 3D printed and examined. Mix S67-F2.0 was selected for dome fabrication due to its favorable mechanical and shrinkage properties. To meet buildability requirements without formwork, an oscillatory toolpath (WeaveSlicer) [58] was implemented to maintain constant wall thickness while scaling earthen 3DP to dome geometries [59]. Printing was performed using a 20 mm nozzle, an extrusion rate of 0.15 rounds/s, and a printing speed of 15 mm/s, based on parameters established during the printability evaluation. Following printing, the dome was cured under laboratory ambient conditions, and changes in wall diameter were measured to assess volumetric stability and compare shrinkage behavior with previously tested shrinkage bars. During the drying phase of the 2-ft-diameter dome, the structure was covered with plastic sheeting to slow moisture loss and promote a more controlled and uniform drying rate across the dome volume. Despite these measures, some drying gradients might still develop throughout the drying process. Figure 7 presents the corresponding 3DSP analysis of the dome.

3. Results and Discussion

This section presents and summarizes the results and observations for the two sets of mixes introduced earlier: the first set (Set 1) composed of S and CS with a maximum aggregate size of No. 4, and the second set (Set 2) composed of S, CS (maximum aggregate size No. 16), and rice hull fiber.

3.1. Traditional Consistency Test for Determining Optimum Water Content

The results of traditional consistency tests performed on two sets of mixes are summarized here. The results indicated that the “optimum” consistency of each mix was highly sensitive to practical factors such as batch size, clay content, pre-soaking, mixing duration, ambient temperature, and the mixing method (mechanical versus hand mixing). As described in the previous section (Figure 3), the LL was measured on a 250 g sample, and 500 g batches were then tested by incrementally adding water from below the LL until a cohesive ball and sausage could be formed.

Table 4. Results of consistency tests and optimum water content for each soil mix in Set 1.

Mix ID	Liquid Limit (%)	Water/(S+CS) (%)	Traditional Consistency Test Observation	Picture of Consistency Test at Optimum Water Content
S100	36.12	36.00	Thick but smooth and creamy; sticky, slippery, similar to peanut butter.
S75	35.50	32.00	Very sticky, slurry-like, and slumpy; not able to roll into a ball or sausage.
S67	33.21	25.33	Thick but smooth; semi-sticky, semi-grainy, and semi-slippery texture.
S50	32.56	23.00	More manageable; medium sandy, less sticky, medium-grainy, and less slippery.
S25	25.03	16.00	Very crumbly and sandy; slumpy but still able to cohere into a ball.

compares the LL and optimum water content for Set 1 of mixes, showing that as the S content decreases, the optimum water content also decreases, while the gap between LL and optimum water content becomes larger. The consistency tests further demonstrate that mixes with higher S content tend to be stickier at their optimum water content, whereas increasing the CS content makes the mixes crumblier and sandier, illustrating the trade-off involved in balancing S and CS proportions. In Set 2, the optimum water content was evaluated using CS with a smaller particle size (passing sieve No. 16). Based on the consistency tests, the optimum water content for the S67 base mix with this finer CS was determined to be $22\%$, as shown in Table 3. Fiber was then added to the S67 mix in varying amounts from $0.5\%$ to $2\%$, while keeping the base proportions of S and CS constant. Within this range, fiber inclusion did not significantly affect the fresh consistency: four of the five fiber-reinforced mixes required the same optimum water content of $22\%$, and only the mix with $2\%$ fiber required a slightly higher value of $22.5\%$, as summarized in Table 3. At the optimum water content, all mixes in Set 2 exhibited a thick yet smooth texture with a semi-sticky, semi-grainy, and semi-slippery feel.

3.2. Properties of Soil Mixes

This section presents the evaluated properties of the Set 1 and Set 2 soil mixes and explains how certain mixes were excluded or revised based on their compressive strength, shrinkage behavior, and printability performance. Mixes that did not meet the required criteria (compressive strength below 300 psi, excessive shrinkage, or inadequate printability) were removed from further consideration. Mixes that satisfied the strength and printability requirements will undergo additional refinement to reduce shrinkage in Set 2. Overall, this work represents a sequential study aimed at identifying an appropriate soil mix suitable for large-scale 3DP from local soil.

3.2.1. Properties of Set 1 Mixes Meeting Strength and Printability Requirements

In Set 1 mix design in this study, S was substituted by CS to find out the optimum mixes that meet strength and printability requirements. Figure 8 shows the compressive strength of the Set 1 mixes with increasing replacement of S by CS. Mix S100, with the highest S content and traditional-consistency water content $(36\%)$, achieved the greatest 28-day strength (939.5 psi). As CS content increased, both water demand and strength decreased, with S25 ($75\%$ CS) exhibiting the lowest 28-day strength at 458.3 psi, an overall $51.2\%$ reduction relative to S100. Decreasing S content also reduced strength gain between 7 and 28 days: S100 gained $94.3\%$, whereas S25 showed a slight $2\%$ loss. These trends indicate that extensive replacement of high-clay local soil with CS can significantly impair strength development, as the clay matrix continues to gain bonding capacity with extended drying. As noted earlier, each mix’s water content was adjusted using the traditional consistency test. For S75, compressive strength was measured at both the LL water content $(35.5\%)$ and the lower traditional-consistency value $(32\%)$. Reducing the water content from $35.5\%$ to $32\%$ in the same mix increased compressive strength by $28\%$ in 7 days and $13.3\%$ in 28 days. This underscores the importance of using an appropriate procedure to establish optimum water content when engineering the mechanical behavior of earthen mixes. All mixes in this first set exceeded the minimum compressive strength requirement of 300 psi. Previous studies [27] have shown that additives such as lime or straw fiber can further reduce shrinkage at the expense of some compressive strength, suggesting similar trade-offs may arise in the present soil–sand system.

Figure 9 presents the linear shrinkage Set 1 mixes as S is progressively replaced by CS (max aggregate size No. 4). As expected, reducing the proportion of S decreases shrinkage. XRD results confirmed that the plain soil contains high-swelling clays (montmorillonite and nontronite), which contribute to volumetric instability: S100 exhibited $12.33\%$ linear shrinkage after 7 days. As CS content increased from S100 to S25, 7-day linear shrinkage dropped from $12.33\%$ to $1.12\%$, a $90.9\%$ reduction. However, this shrinkage reduction was accompanied by a $51.2\%$ loss in 28-day strength from S100 to S25, emphasizing the need to balance shrinkage control against strength retention. Moreover, although S25 shows the smallest shrinkage $(1.12\%)$, the traditional consistency tests revealed a tendency to slump, suggesting that its strength and stiffness are insufficient to resist the load from overlying filaments during 3DP. Low shrinkage alone is therefore not a sufficient criterion for choosing a 3DSP mix.

Shrinkage behavior in S75 was further examined at two water contents ($35.5\%$ LL and $32\%$ traditional consistency). After 7 days, the linear shrinkage curves intersected at $11.36\%$ for both conditions. Thus, while reducing water content from $35.5\%$ to $32\%$ in S75 improved compressive strength ($28\%$ at 7 days and $13.3\%$ at 28 days), it did not reduce linear shrinkage over the same period. Notably, the S25 mix dried much more quickly, reaching a shrinkage plateau after only one day due to its lower soil (and lower clay) content, whereas the other mixtures required roughly three days to stabilize. As shown in Figure 9, this behavior reflects the slower drying rates observed in mixes with higher soil (and higher clay) content.

The reduction in shrinkage observed with the addition of $2\%$ rice-hull fiber can be attributed to several reinforcing mechanisms characteristic of natural fiber–soil composites. As the soil matrix dries and undergoes volumetric contraction, the fibers function as bridging elements, spanning developing microcracks and restricting their propagation, thereby delaying the formation of larger shrinkage cracks. Beyond crack bridging, the fibers introduce localized stiffness enhancement by forming a distributed internal network that resists deformation within the fresh mix. Their rough, silica-rich surfaces facilitate mechanical interlocking at the fiber–soil interface, enabling more effective transfer of tensile stresses from the shrinking matrix into the fibers. This interaction contributes to early-age tensile resistance and helps reduce the strain gradients that typically drive shrinkage cracking. Further analysis is required to quantitatively assess the relative contribution of these mechanisms to shrinkage mitigation in fiber-reinforced earthen materials. Overall, the mechanical and dimensional stability results indicate that replacing S with CS must be carefully optimized. Higher CS contents substantially reduce shrinkage but can also compromise strength, stiffness, and inter-layer bonding in 3DP. Given that shrinkage levels remain relatively high in the more clay-rich mixes, additional mitigation strategies, e.g., introducing shrinkage-control additives or fibers, are likely required, as is common in adobe practice and, by extension, for a balanced 3DSP mix.

In the next phase of the study, the printability of the Set 1 mixes was evaluated in terms of both extrudability and buildability. For the extrudability assessment, a zigzag pattern was printed at various travel speeds. For each mix, the extrusion rate was first adjusted to achieve a steady and continuous material flow. The printhead speed was then increased incrementally from 10 mm/s to 35 mm/s along a zigzag path (Figure 6a). The width of each printed filament was measured shortly after deposition, and the speed that produced the smallest deviation from the nominal 20 mm filament width, while avoiding cracking in either the fresh or hardened state, was identified as the optimum printable condition (Figure 10). Buildability tests were subsequently performed at this selected speed, and filament widths were measured after one day to assess dimensional stability.

Mixes S50 and S25 exhibited low consistency and a very sandy texture due to their high CS content and larger particle sizes, preventing them from being extruded through the nozzle during the 3DP trials. As shown in Figure 10, the S100 mix demonstrated excellent extrudability, producing relatively consistent filament widths over most printing speeds, with only minor deviations from the nominal value. In contrast, S75 and S67 showed greater variability, displaying over-extrusion at slower speeds and noticeable inconsistency at higher speeds.

It is also notable that the 3D-printed S100 zigzag pattern developed significant cracking after drying as represented in Figure 10 at laboratory temperature, which aligns with the high shrinkage previously reported for this mix. In comparison, S67 exhibited fewer cracks in the hardened state, reinforcing the finding that partial replacement of S with CS reduces shrinkage and improves dimensional stability relative to the pure soil mix (S100).

In addition to the extrudability assessment, the buildability of the Set 1 mixes was evaluated by determining the maximum number of layers each mix could support before collapse. The S100, S75, and S67 mixes were able to sustain 12, 10, and 13 printed layers, respectively. The buildability results illustrate the ability of each mix to withstand the weight of successive layers during the 3DP process.

Based on the combined extrudability and buildability outcomes, S67 emerged as the most suitable representative mix. It met the required strength threshold, achieving a 28-day compressive strength of 655 psi (4.51 MPa), and successfully passed both printability tests at a printing speed of 20 mm/s using a 20 mm nozzle while its shrinkage was the minimum compared to the other printable mixes. Consequently, S67 was selected for further investigation in the remainder of this study.

3.2.2. Properties of Set 2 Mixes Meeting Shrinkage and Printability Requirements

As noted earlier, the second part of this study focused on developing Set 2 mixes and evaluating their shrinkage performance with the addition of natural fiber. Because fibers influence extrudability, the CS aggregate was further sieved to a finer gradation, using only particles passing sieve No. 16 to maintain smooth extrusion through the 20 mm nozzle. Similar to Set 1, both compressive strength and linear shrinkage were measured for the fiber-reinforced mixes listed in Table 3, and a representative mix was selected for 3DP evaluation. Building on the results from Set 1 and initial printability trials, a 2:1 ratio of S to CS ($67\%$ S, $33\%$ CS; maximum aggregate size No. 16) was chosen as the baseline for Set 2, designated S67F0 (no fiber). The goal of this phase was to further reduce shrinkage through fiber inclusion while maintaining the minimum required compressive strength of 300 psi.

Figure 11 summarizes the compressive strength of the Set 2 mixes with fiber contents ranging from $0\%$ to $2\%$. S67F0 achieved the highest 28-day strength (881.67 psi), whereas S67F2 exhibited the lowest (663.0 psi). Increasing fiber content from $0\%$ to $2\%$ resulted in a $24.8\%$ reduction in 28-day strength; however, all mixes remained well above the 300-psi threshold. Strength gain between 7 and 28 days also decreased with higher fiber content: S67F0 increased by $19.35\%$, while S67F2 increased by only $7.5\%$. These trends indicate that fiber addition reduces ultimate strength but leads to earlier stabilization of the material, which is advantageous for buildability during printing. Importantly, all fiber-reinforced mixes in Set 2 met the target compressive strength, even at the highest fiber content $(2\%)$. It should also be noted that the compressive strength of S67 mix shown in Figure 8 and the S67F0 mix shown in Figure 11 differ in that the latter incorporates a finer CS gradation. The use of smaller CS particles contributes to slightly higher strength due to improved densification of the hardened soil matrix.

Figure 12 presents linear shrinkage results over 7 days of drying for the Set 2 mixes. The unreinforced baseline, S67F0, showed the highest shrinkage at $6.23\%$, whereas S67F2, containing $2\%$ fiber, shrank only $3.48\%$, corresponding to an approximate $44\%$ reduction. As in Set 1, most of the shrinkage occurred during the first three days, with only minor changes thereafter. Despite this significant shrinkage reduction, S67F2 maintained a 28-day compressive strength of 663 psi, well above the 300-psi minimum requirement. Because mixes with fiber contents greater than $2\%$ could not be pumped through the lab-scale 3DP system, additional fiber incorporation was not feasible in this study. Future work employing pumping systems designed for earthen 3DP may accommodate higher fiber dosages, potentially yielding further shrinkage reduction without compromising printability. Based on the balance of strength, reduced shrinkage, and compatibility with the printing system, S67F2 was selected for subsequent printability testing, including the fabrication of the 2-ft-diameter dome.

Because S67F2 exhibited the lowest shrinkage among the Set 2 mixes, it was selected for printability evaluation, including both extrudability and buildability tests. (Figure 13a) illustrates the extrudability performance of S67F2 when printed through a 20 mm nozzle at a constant extrusion rate of 0.15 rounds/s and varying printhead speeds. At 10 mm/s, the material was clearly over-extruded, whereas speeds of 30 and 35 mm/s caused skipping and filament thinning, indicating under-extrusion. Printing at 20 and 25 mm/s produced slightly narrow but still acceptable filaments. Based on visual inspection and filament-width measurements, a speed of 15 mm/s provided the most stable flow and the closest match to the nominal 20 mm filament width. The constant-speed test shown in (Figure 13b) further demonstrates that S67F2 can be extruded consistently, without clogging, interruptions, or cracking, while maintaining the dimensional accuracy required for 3DSP applications. Additionally, the S67F2 buildability test demonstrated that the mix could sustain up to 24 layers before collapse, with failure occurring at the 25th layer (Figure 13c,d). This improvement compared to Set 1 mixes indicates that S67F2 provides sufficient early-age stiffness and cohesion to support taller prints, making it a strong candidate for printing the 2-ft-diameter dome at a larger scale.

Figure 14 shows the flexural test specimens printed using the S67F2 soil mix. Five samples were cured for 28 days before being tested in accordance with ASTM C293 [57] using a three-point bending setup. The results indicate an average flexural strength of 189.24 psi (1.30 MPa), which is nearly three times the minimum requirement of 50 psi specified by the NM Earthen Building Code [22].

3.3. Large-Scale 3D Printing of the Selected Mix

This study developed a systematic approach to modify locally sourced natural soil to meet the mechanical and rheological requirements of gantry-based 3DP. Based on evaluations of compressive strength, shrinkage behavior, extrudability, and buildability, the S67F2 formulation was identified as the most suitable mix. It satisfied the minimum strength criteria, exhibited the lowest shrinkage among the Set 2 formulations, and demonstrated stable printing performance. Using optimized printing parameters, S67F2 was selected for large-scale validation.

A hemispherical dome with a 2 ft diameter was successfully fabricated to assess constructability and time-dependent behavior at a representative scale as shown in (Figure 15a,b). The printed base circumference before drying measured 2044.7 mm, approximately $6.6\%$ larger than the designed value of 1918 mm, likely due to slight outward deformation during printing under the weight of subsequent layers. This deviation was minor and did not affect structural continuity, as the dome fully closed at the crown. To control early-age moisture loss, the structure was covered with plastic sheeting and monitored throughout drying. Dimensional changes were tracked along the base circumference and the meridional direction from crown to base.

The observed deformation behavior reflects the inherent stress distribution of a 3D-printed earth dome. Meridional stress remained predominantly compressive, allowing vertical shortening to accommodate drying shrinkage. In contrast, circumferential deformation at the base was strongly restrained by boundary conditions and by the presence of hoop tension, resulting in negligible measurable shrinkage in this direction. After three days, no shrinkage was detected along the base circumference, while a $1.5\%$ reduction was observed along the meridional path. Continued drying led to an accumulated meridional shrinkage of approximately $8.8\%$ after 33 days of drying, confirming the anisotropic nature of shrinkage in hemispherical earthen shells.

Cracking behavior was consistent with a restrained-shrinkage mechanism. Although the dome was wrapped in plastic sheeting to slow the drying rate and minimize uneven moisture loss, some drying gradients still developed across the 2-ft-diameter structure. These gradients arise when the surface dries and contracts more rapidly than the interior, generating internal moisture-transport–driven stresses that can contribute to crack formation even under geometric restraint. Nevertheless, because the laboratory environment was extremely arid, with a relative humidity of $16\pm 2\%$, the magnitude of these drying gradients was relatively small under the controlled conditions used in this study. Two primary cracks developed during drying, both initiating near the base where hoop tension and shrinkage restraint coincide. The first crack appeared on day 7 and propagated upward along the dome surface (Figure 15c), while a second crack formed by day 10 and advanced preferentially along interlayer filament interfaces (Figure 15d). Crack widths increased modestly with time, reaching maximum values of 4.37 mm and 11.74 mm by day 13. The alignment of cracks within the lower third of the dome and along print interfaces underscores the combined influence of tensile hoop stresses, restrained circumferential shrinkage, interlayer anisotropy, and the residual drying gradients that persisted despite controlled drying conditions.

Domes generally demonstrate favorable seismic behavior because their geometry enables efficient membrane action, reducing stress concentrations and minimizing reliance on bending resistance. However, earthen materials remain brittle with limited tensile capacity, so seismic vulnerability is often controlled by crack initiation and propagation along weak planes. In 3D-printed structures, interlayer interfaces may act as additional planes of weakness, and further work is needed to evaluate how layer adhesion influences cyclic loading response. The shrinkage-related cracks observed in the lower third of the dome could also serve as potential initiation points under seismic excitation. Although the 2-ft prototype cannot provide quantitative seismic performance data, the dome form is expected to distribute inertial forces effectively, and with improved interlayer bonding and reduced early-age cracking, 3D-printed earthen domes may offer promising seismic resilience. Future work will include structural modeling and cyclic or dynamic testing to assess this behavior more rigorously.

Overall, the successful construction of the 2-ft dome with limited cracking demonstrates the viability of the S67F2 mix for structural-scale 3D-printed earth applications. The observed damage patterns indicate that further crack mitigation can be achieved through targeted strategies focused on the lower third of the dome, including local material grading or confinement to reduce effective hoop tension. In addition, optimization of the printing path to vary filament orientations and disrupt continuous interlayer interfaces can reduce preferential crack propagation and enhance interlayer load transfer. Together, these measures provide a clear pathway for improving the durability and structural reliability of 3D-printed earth domes under drying-induced stresses.

4. Future Research

Future work should more comprehensively evaluate the stress–strain response and load-bearing behavior of 3D-printed earthen structures, including how these mechanisms interact with restrained circumferential shrinkage and hoop-tension cracking. Additional studies should investigate three-dimensional deformation and interlayer anisotropy, as well as the influence of printed geometry on compression and tension distribution. The effects of nozzle size and shape on layer deposition and interlayer bonding also warrant systematic evaluation.

Further mineralogical analysis is needed to quantify the proportions of expandable and non-expandable clays in the soil mix, beyond the preliminary XRD results obtained here. The reinforcing mechanisms of rice-hull fibers, such as crack bridging, stiffness modification, and interface strengthening, should also be assessed quantitatively. Mechanical tests should be extended to include compressive testing of printed specimens, allowing comparison with cast samples to evaluate printing-induced anisotropy.

Finally, future studies should examine drying behavior, moisture sensitivity, and erosion resistance under realistic indoor and outdoor conditions. Developing standardized testing methods specifically for 3D-printed earthen materials, as well as exploring printing with robotic systems offering more degrees of freedom, will further advance the structural and material performance of these systems.

5. Conclusions

• This study demonstrates the viability of adapting locally sourced clay-loam soil from Belén, NM, for gantry-based 3D printed earthen construction (without the use of harmful chemicals) through systematic modification with graded concrete sand and rice hull fiber. By integrating conventional soil characterization methods with additive manufacturing metrics such as extrudability, buildability, and strength, a multiscale framework was established to assess material behavior from laboratory specimens to structural components, paving the way for a potential standard to be developed.
• Partial replacement of soil with graded concrete sand (max aggregate size sieve No. 4), set 1 of the soil mixes, demonstrated decent results in shrinkage but not so well in strength. As the CS content increased from S100 to S25, 7 day linear shrinkage dropped from 12.33% to 1.12%; however, it was accompanied by a 51.2% reduction in compressive strength, emphasizing the need to balance shrinkage control against strength retention to meet code requirements. A 2:1 soil-to-sand ratio provided an effective baseline.
• In set 2 of the soil mixes, the max aggregate size of CS was reduced to sieve No. 16, and the incorporation of 2% rice hull fiber improved the performance of the earthen mixture. Linear shrinkage was reduced to 3.48%, a 72% reduction relative to set 1, and exceeded compressive (663 psi) and flexural (189 psi) strengths required by the NM Earthen Building Code by nearly double and triple, respectively. The mix demonstrated reliable extrudability and supported 24 printed layers without deformation.
• Successful fabrication of a 2-ft-diameter hemispherical dome confirmed the structural viability of the optimized mix and highlighted the vital interaction between material properties, printing parameters, and shell geometry. The results provide a scalable methodology for developing low-carbon earthen mixtures and offer insight into shrinkage-induced stress behavior in 3D-printed structures.