Toward Predictive Dimensional Accuracy in DLP-Fabricated Alumina Lattices: Volumetric Error and Shrinkage Control

Abstract

This work introduces a quantitative framework that couples slurry formulation, curing dynamics, and lattice architecture to predict and control defects in DLP-fabricated alumina ceramics. While prior studies have primarily focused on slurry optimization, this study advances the field by establishing predictive correlations between solid loading, shrinkage, staircase error, and mechanical performance. Increasing alumina content from 60–80 wt % reduced volumetric shrinkage from 29.2% to 9.7% and enhanced density up to 90.1%. The staircase effect in ceramic lattices was quantified using cusp height (∼35 μm) and volumetric error (1.77 × 10^–4 mm³), providing a direct measure of surface fidelity. Mechanical testing of the lattice architectures revealed that the hexagonal design exhibited a compressive strength of 8.3 MPa, approximately six times higher than that of the truss–plate lattice (1.32 MPa), despite both possessing a similar porosity of 90%. This demonstrates that the lattice architecture plays a more dominant role than porosity in determining the overall strength. By integrating rheology–optics–mechanics into a shrinkage-controlled, architecture-optimized strategy, this study establishes generalizable design guidelines for high-fidelity, load-bearing ceramic lattices manufactured via DLP.

1. Introduction

Ceramics are a class of material known for their high thermal stability, exceptional hardness, and remarkable durability, thus playing a vital role in human history. From ancient pottery and architectural marvels to modern-day aerospace components and biomedical implants, ceramics adapt to the ever-changing world. Injection molding, dry pressing, gel casting, and tape casting are conventional ceramic shaping methods that have been widely used for decades. All four methods are still limited to fabricating ceramic parts with complex shapes. In the above-mentioned method, the first three can be easily controlled and processed, but theories are highly dependent on models of high cost. This takes a long lead time for the mass production of ceramic components. Tape casting manufactures products by stacking film layer by layer without molds. This final product will not have precision.

Additive manufacturing, also known as 3-D printing, provides a promising solution to the issues in ceramic forming. Fused Deposition Modeling (FDM), Selective Laser Sintering (SLS)/Selective Laser Melting (SLM), Stereolithography (SLA), Digital Light Processing (DLP), Three-dimensional printing, Laminated Object Manufacturing, and Direct Ink Writing (also known as robocasting) are the various additive manufacturing techniques. Among them, SLA and DLP are the most prospective technologies for manufacturing ceramics. − Due to the final product’s high precision and surface quality, photopolymerization has attracted extensive attention. DLP is a technology used in digital imaging and 3D printing, where a digital light source projects high-resolution images layer by layer onto a surface. It works through three key steps: digital projection of layer images, image processing to define printing patterns, and high-resolution exposure to ensure precise and detailed prints.

A 40 vol % alumina suspension was prepared using (1,6)-hexanediol diacrylate (HDDA) as the curable monomer, with oleic acid (OA), stearic acid, and poly­(acrylate ammonium) (PAA-NH₄) as surfactants to improve dispersion, as demonstrated by Li and Zhao. Using the SLA method, they obtained alumina parts with a relative density of 95%. A ZrO₂–Al₂O₃ composite ceramic part was fabricated using the SLA method and sintered at 1500 °C for 60 min, as reported by Liu et al. They obtained ceramic samples with a density of fracture toughness and hardness of 3.75 g/cm³, 4.05 MPa/m², and 14.10 GPa, respectively. The effect of oleic acid (OA), alkane acrylic phosphate ester (PM1590), and dispersant (BYK-111) on the properties of UV-curable suspensions was investigated in a study by Xu et al. They developed a stable colloidal suspension with a relatively high solid loading of 55 vol % and excellent self-leveling behavior using Disperbyk-111 as a dispersing agent. Zirconia green parts were successfully fabricated using DLP technology, and crack-free components were obtained after debinding in argon at a controlled heating rate of 5 °C/min, as reported by Sun et al. During ceramic fabrications, the key issue is the optimization of slurries and the printing parameters. Debinding and sintering are also very important parts of ceramic forming. Poor densification, cracking, and poor mechanical properties are due to sintering defects and formation failures. This could be avoided by a reasonable design of experiments. ,

The staircase effect, a layer-wise surface deviation that reveals on inclined or curved geometries, is known to reduce part fidelity and influence mechanical properties by producing surface notches and stress concentrators. Early approaches that were used to evaluate staircase morphologies are metrics like R_a (average roughness), R_z (maximum height), and specific cusp height calculations that relate directly to layer thickness and build angle. Advanced quantification methods in stereolithographic AM, using image analysis tools to assess the geometric quality and defect morphology of ceramic and polymer components. Recent work in ceramic AM expands upon these methods. ZrO₂ and silica scaffolds implementing a combination of morphometric image analysis and surface profilometry done by Dadkhah et al. to assess the staircase effect and dimensional errors. A process-stage evaluation using extrusion-based ceramic AM has been presented by Rane et al., quantitatively tracking porosity and staircase defect evolution with statistical and image-based methods. Yang and Miyanaji further highlight the impact of staircase defects on the mechanical performance and postprocessing requirements for advanced ceramics, detailing mitigation strategies such as optimized slicing, layer reduction, and laser-assisted finishing. Cusp height and volumetric error metrics have also been standardized in polymer and metal research and have seen direct translation to ceramic systems. Sivasankar et al. discuss these metrics in the context of zirconia-based scaffold design for biomedical applications, proving their relevance and broad application in ceramic AM. Sa et al. and Liu et al. use direct measurement and microstructural characterization to show how staircase-related defects evolve during postprocessing and sintering, confirming the reproducibility of quantitative analysis across different fabrication environments. The field continues to refine these approaches for advanced ceramics, underlining the importance of both process optimization and defect mitigation.

In the present study, alumina ceramic slurry of various solid loading is optimized by planetary ball milling technique, that is, 60 wt % to 80 wt % are achieved by adding dispersant. To develop a quality ceramic product, a homogeneous slurry prepared is utilized for printing the samples via DLP technique. Meanwhile, the printing parameters and postprocessing were optimized to achieve quality alumina ceramic parts without any delamination. The diverse application areas of alumina lattice structures is shown in Figure .

1.

Impact of lattice-structured ceramics.

2. Experimental Section

This section details the materials, processing methods, and characterization techniques used for the fabrication of alumina-based ceramic lattice structures via DLP. The experimental workflow begins with the preparation and optimization of high solid-loaded ceramic slurries, followed by 3D printing using DLP. Different characterization techniques were also employed at each stage to evaluate the structural, thermal, rheological, and mechanical properties of the material system.

2.1. Preparation of Alumina Photopolymeric Resin

The alumina ceramic slurries used in this study are prepared using alumina powder (Rohini Industries, Pune) of particle size 492 nm and premixed in a photopolymeric resin (JG Maker, China) which is ready with Bisphenol A and Epichlorohydrin (38%), Propylidynetrimethanol, ethoxylated ester with acrylic acid (47%), (5-ethyl-1,3-dioxan-5-yl) methyl acrylate (10%) and Phenylbis­(2,4,6-trimethylbenzoyl)-phosphine oxide (5%). Epoxy resin is formed by Bisphenol A and Epichlorohydrin, a hybrid epoxy-acrylic resin formed by the reaction of Propylenetrimethanol and acrylic acid. Dioxane acrylate has also been included to improve the mechanical stability of the cured material. Phenylbis­(2,4,6-trimethylbenzoyl)-phosphine oxide is the photoinitiator that initiates the reaction by forming a free radical.

The ceramic powder alumina (Al₂O₃) is mixed with the resin and checked for compatibility. The powder has dispersed homogeneously within the resin, and thus the resin acts as a binder while doing 3D printing. The photopolymeric resin helps to print the ceramic resin layer by layer in the DLP printer. The ceramic slurries are optimized from 60 wt % to 80 wt % of solid loading of alumina powder. The slurries are prepared via a planetary ball milling process at 300 to 350 rpm for 6 h, which enhances the dispersion of the alumina powder. 0.2 vol % to 0.5 vol % of dispersant (Disperbyk-111 (BYK111)) is added to the slurry to improve the solid loading. The effect of dispersant in ceramic resin at different weight percentages is studied by Jeehwan Kim et al. and their coworkers. It was found that the viscosity decreased as BYK111 content increased from 1 wt % to 3 wt %, while the viscosity increased as BYK111 content was after the 4 wt %. The flowability of the resin can be improved by improving the steric stabilization of the nonaqueous solution. However, if the concentration of the dispersant increases, it may lead to flocculation, and thus, the slurry deterioration can happen. Therefore, the optimal dispersant content is to be added to the ceramic slurry, which must be examined. A low-viscosity slurry can thus be optimized by adding the proper amount of dispersant. In addition, a higher solid loading can also be achieved by this, so that this will, in turn, give a better density for the material. Bubbles were formed during the milling process, which were then removed by ultrasonic defoaming. The prepared slurry is then used for DLP printing by adjusting the printing parameters such as exposure time, layer thickness and, bed speed.

2.2. 3D Printing of Alumina Photopolymeric Slurry via DLP Technique

The DLP printer used in this study is the Creality LD-002H, which operates using a Digital Micromirror Device (DMD) and a movable build platform. The DMD functions as a dynamic optical mask, projecting UV light at a wavelength of 405 nm, which matches the absorption peak of the photoinitiator. The system has an output power of 1 W and a light intensity of 8 mW/cm².

Prior to printing, 3D models of the desired lattice structures (corrugated lattice, cuboidal gyroid, hexagonal lattice, truss-plate hybrid lattice, and cylindrical gyroid) are designed and exported in STL format. These models are digitally sliced using the Chitubox software. The sliced files are then uploaded to the printer. The lattices which are intended to print are shown in Figure .

2.

CAD models of lattice patterns (a) corrugated lattice, (b) cuboidal gyroid, (c) hexagonal lattice, (d) truss-plate hybrid lattice, (e) cylindrical gyroid.

During printing, the software projects the corresponding UV light pattern for each layer onto the surface of the ceramic slurry. The photopolymer resin in the slurry undergoes localized curing through photopolymerization, solidifying one layer at a time. This process is repeated layer-by-layer until the complete ceramic green body is formed. As a result, the selected lattice structures are successfully printed using the DLP technique.

One of the major challenges during postprocessing was the removal of uncured resin from the intricate, small-scale lattice structures. Due to the complexity and confined spaces within the printed geometries, complete removal of residual resin proved difficult during the washing stage. Since the resin used is acrylate-based, solvents containing hydroxyl groups such as ethanol, glycerol, and isopropanol are typically effective for cleaning. Among these, ethanol was selected for this work owing to its low viscosity and superior solubility in nonpolar systems, which facilitates efficient penetration and resin removal from fine structural features.

Following the cleaning process, the printed green parts were subjected to a controlled sintering cycle. The sintering was carried out at a peak temperature of 1550 °C for 4 h using a slow, stepwise heating profile. Intermediate temperature holds were incorporated to ensure gradual binder burnout, thereby minimizing internal stresses and preventing structural defects such as cracking. The final high-temperature dwell enabled complete densification and the development of strong ceramic bonding. This sintering strategy ensured both structural integrity and high density in the final ceramic components. The detailed sintering profile is illustrated in Figure .

3.

Sintering profile of the DLP printed samples.

The entire process, from slurry preparation to postprocessing, involves several critical steps that influence the final part quality. The stage-wise processing steps involved in ceramic 3D printing using the DLP technique are depicted in Figure .

4.

Flowchart of ceramic printing using DLP technology.

2.3. Material Characterization

A series of characterization techniques was systematically employed, starting from the raw alumina powder through to the green and sintered alumina lattice structures fabricated using DLP. The particle size distribution of the alumina powder was first determined using Dynamic Light Scattering (DLS), where the alumina powder was dispersed in photopolymeric resin and milled for 6 h. This analysis provided insights into the average particle size and homogeneity of the raw material that is needed to be printed.

The rheological behavior of the alumina photopolymeric slurry was analyzed by preparing formulations with varying dispersant contents at solid loadings of 60, 70, and 80 wt %. A rotational rheometer (Rheoplus/32 V3.61 21005995-33024 (MCR102 SN81077601)) was used to evaluate the viscosity and shear-thinning behavior under a range of shear rates, helping to identify optimal flow properties suitable for high-resolution printing. In parallel, the cure depth of the resin was assessed for each solid loading by exposing slurry samples to different UV exposure times and measuring the cured thickness using a digital micrometer. The UV–visible absorption spectroscopy (UH5300 spectrophotometer) helps to study how the light interacts with the particles and how it affects the curing depth. To assess the elemental composition of the green bodies, Energy Dispersive X-ray Spectroscopy (EDS) was performed using a scanning electron microscope (SEM) equipped with an EDS detector. This allowed confirmation of the purity and distribution of elements within the printed green structures.

Thermogravimetric Analysis (TGA) and Differential Thermogravimetric (DTG) analysis were performed using a Thermogravimetry Analyzer (TA Instruments) to study the thermal decomposition and binder burnout behavior. Green samples were subjected to controlled heating (2 °C per minute) in an oxidizing atmosphere, and the weight loss profiles were analyzed to determine degradation stages of the binder system, thereby developing the debinding and sintering profile.

X-ray Diffraction (XRD) analysis was carried out on the raw alumina powder and printed samples in different processing stages, such as green, debinded, and sintered. Disc-shaped samples with dimensions of 5 mm × 20 mm were printed and are used for this purpose. The diffraction patterns were used to confirm phase purity, detect any phase transformation, and assess crystallinity.

The SEM observations were carried out using a High-Resolution Scanning Electron Microscope (EVO 10, Carl Zeiss, Germany). The instrument employs a LaB₆ (Lanthanum hexaboride) filament as the electron source. In addition, optical microscopy was employed using a LEICA Microscope and Digital Camera (Germany), to examine surface features and printing-induced defects such as the staircase effect. Surface images were captured at various magnifications, and image analysis software was used for quantifying layer lines and roughness.

Mechanical performance was evaluated by conducting compression tests on the printed lattice structures, specifically the hexagonal and truss-plate designs. The tests were carried out on a Universal Testing Machine (UTM) at room temperature under a controlled cross-head speed of 0.5 mm min^–1. The sintered lattice specimens used for testing had dimensions of 20 × 20 × 20 mm. The compressive strength, load-bearing capacity, and failure behavior were recorded and analyzed to understand the influence of geometry and processing parameters on mechanical integrity.

3. Results and Discussion

3.1. Particle Size Distribution of Alumina

The raw alumina powder exhibited an average particle size of ∼492 nm with irregular morphology (Figure ). Such particles hinder tight packing and increase viscosity, necessitating planetary ball milling for agglomerate breakdown and homogeneous dispersion. Optimized milling combined with dispersant addition enabled stable slurries up to 80 wt % solid loading, which served as the basis for high-density DLP-printed structures.

5.

(a) Particle size distribution of alumina powder used. (b) Morphology of the alumina powder.

3.2. Cure Depth of Various Solid-Loaded Alumina Slurry

The light scattering and transmittance behavior in the ceramic photopolymeric resin is illustrated in Figure a. The degree of light attenuation directly correlates with the material’s absorbance and scattering behavior. Higher absorbance results in less transmitted light, thereby reducing the effective cure depth, necessitating the optimization of both the resin formulation and the exposure parameters to ensure adequate curing performance, especially at higher ceramic loadings.

6.

(a) Light scattering and transmittance in ceramic photopolymeric resin. (b) The measured UV–visible spectrum of pure photopolymeric resin.

From the graph (Figure b), it is clear that the photopolymeric resin absorbs light at a wavelength of 380 to 403 nm range with an absorbance value of 2.5. Whereas a significant amount of alumina particles in the mixture likely leads to increased light scattering and adsorption. The alumina ceramics scatter light, which also prevents the effective transmission of UV light through the mixture, resulting in a minimal or undetectable spectrum in the 300 to 405 nm range. Thereby, we can correlate this with the curing effect of the ceramic polymer resin. Since the absorption is high due to the interaction of particles with light, the light will not penetrate in depth, and thus, the curing may not happen effectively, which will affect the curing dynamics. Thus, to improve the curing while printing, it is necessary to optimize the exposure time required for the polymer to polymerize to a certain thickness. The factor called depth of cure also plays a role in achieving the resolution of the printed parts.

From the measured UV–visible spectrum shown in Figure b, it is evident that the pure photopolymeric resin exhibits strong absorption in the wavelength range of 380 nm to 403 nm, with a peak absorbance of approximately 2.5. This absorption range corresponds well with the emission profile of the UV light source used in the DLP printing process. However, upon incorporation of alumina particles into the resin, the optical behavior undergoes a marked transformation.

The presence of a significant volume of alumina particles suppresses the effective transmission of UV light through the suspension. As a result, the resin-alumina mixture exhibits minimal or undetectable absorbance within the same spectral window (300–405 nm), indicating that the light is either scattered away or attenuated before reaching sufficient depth.

This optical attenuation directly influences the curing dynamics of the ceramic photopolymer resin. High scattering reduces the penetration depth of light, limiting the extent to which the resin can be polymerized during exposure. Consequently, the depth of cure is reduced, potentially affecting both the mechanical strength and dimensional fidelity of the printed part.

To compensate for this, it becomes essential to optimize the UV exposure time to ensure that sufficient energy is delivered to initiate curing throughout the desired layer thickness. The parameter known as depth of cure becomes critical in this context, as it determines the resolution, accuracy, and layer adhesion of the printed component. Tailoring exposure parameters to account for changes in optical properties at higher ceramic loadings is therefore vital for achieving consistent and defect-free prints.

Cure depth is the maximum thickness of material that can be effectively cured by light exposure. The factors that directly influence the cure depth are the solid loading of the ceramic particle, light intensity, and curing time or exposure time. Cure depth can be calculated by the following equation derived from the Beer–Lambert law: ,

$${\mathrm{C}}_{\mathrm{d}}={\mathrm{D}}_{\mathrm{p}}\mathrm{l}\mathrm{n}({\mathrm{E}}_{\mathrm{m}\mathrm{a}\mathrm{x}}/{\mathrm{E}}_{\mathrm{c}})$$

where C_d is the cure depth, D_p is the depth of penetration, E_max is the maximum energy dosage per area (J/cm²), and E_c is the critical energy dosage (J/cm²). To experimentally determine the cure depth at various solid loadings, a series of single-layer exposure tests was conducted using the DLP printer. For each slurry composition (60, 70, and 80 wt % alumina), layers were exposed to UV light at incrementally increasing exposure times. The resulting cured layer thicknesses were measured and assumed to directly correspond to the cure depth under each condition.

For the 60 wt % slurry, no visible cured layer was observed for exposure times below 4 s, indicating that the energy delivered was insufficient to exceed the critical exposure threshold (Ec) required for polymerization. A discernible layer began to form at 4 s and progressively increased in thickness with longer exposure durations, reaching a maximum cure depth of 470 μm at 6 s.

A similar trend was observed for the 70 and 80 wt % solid-loaded slurries, though they required longer exposure times to initiate curing due to their higher ceramic content and associated light scattering. For 70 wt %, the first detectable cured layer appeared around 8 s, and the cure depth ranged between 500 μm and 720 μm for exposures up to 10 s. The 80 wt % slurry exhibited the most delayed onset of cure, with visible polymerization starting only at approximately 10 s, and achieving a cure depth of up to 1020 μm by 13 s. The optimized parameters are summarized in Table .

1. Cure Depth for Various Solid Loading at Different Exposure Times.

S No.	Solid Loading (wt %)	Exposure Time (s)	Cure Depth (μm)
1	60	4–6	182–470
2	70	8–0	500–720
3	80	10–13	750–1020

For each increase in solid loading, the exposure time (and therefore exposure energy) was increased to compensate for greater light attenuation. Because cure depth depends logarithmically on exposure energy, increasing the exposure energy produces a larger measured cure depth even when particle scattering is higher. Thus, the observed rise in cure depth with solid loading is explained by the intentional increase in exposure energy rather than an intrinsic reduction in optical attenuation. The graph of the cure depth of each solid loading is illustrated in Figure .

7.

Graph of cure depth for different solid loadings.

These results highlight that cure depth is not solely a function of solid loading but rather a combined outcome of exposure energy and slurry composition. Therefore, careful co-optimization of these parameters is essential to ensure dimensional accuracy, layer bonding, and defect-free DLP-printed ceramic structures. These findings underscore the importance of tailoring exposure parameters to specific slurry formulations to ensure adequate layer bonding and dimensional accuracy during DLP-based ceramic fabrication. Optimization of both exposure time and solid loading is important to achieve the desired mechanical properties and structural integrity. Thus, to avoid several defects like delamination and dimensional irregularities in the printing parts, optimization of cure depth is important.

3.3. Rheological Study of Alumina Photopolymeric Slurry

The viscosity of ceramic slurries plays a crucial role in additive manufacturing processes, including DLP, slip casting, and robocasting. Excessive viscosity hinders uniform recoating, leading to poor layer adhesion and print defects, whereas insufficient viscosity promotes particle sedimentation, compromising dimensional fidelity and mechanical integrity. Therefore, achieving an optimal viscosity is essential to balance flowability with particle stability. Improved dispersion not only stabilizes the slurry but also promotes uniform densification, minimizes porosity, and facilitates homogeneous grain growth, ultimately enhancing the performance of sintered ceramics (Figure ).

8.

Critical role of optimal resin viscosity for improved dispersion in ceramic processing.

All slurries exhibited non-Newtonian, shear-thinning behavior (Figure ), typical of ceramic suspensions. At low shear rates, high viscosity arose from particle flocculation, which progressively decreased with increasing shear rate as agglomerates were disrupted. Dispersant concentration had a pronounced effect: at 0.2 vol % the viscosity was highest (∼120 Pa·s), reflecting inadequate stabilization and particle agglomeration. Increasing dispersant to 0.3–0.5 vol % markedly reduced viscosity, indicating improved particle separation and reduced interparticle forces. At 0.5 vol %, the slurry became noticeably thinner at 60 wt % solid loading, suggesting excess dispersant and possible depletion stabilization effects that reduce interparticle friction beyond the optimal level. Whereas, at higher solid loading (80 wt %), 0.5 vol % provided the most balanced flow behavior, indicating that the optimal dispersant level shifts upward with increasing particle concentration to ensure complete surface coverage and prevent agglomeration.

9.

(a) Viscosity vs shear rate graph and (b) shear stress vs shear rate graph of 60 wt % (a, b), 70 wt % (c, d) and 80 wt % (e, f) alumina at varied vol % of dispersant.

Shear stress versus shear rate plots further confirmed this trend. Poorly dispersed slurries (0.2 vol %) exhibited high resistance to flow, while 0.3–0.5 vol % dispersant provided stable, well-dispersed suspensions with stronger structural networks. At 0.5 vol %, stress decreased, and the slurry was consistent with maximum particle separation and cohesive flow behavior. These results were consistent across all solid loadings (60–80 wt %), with 0.5 vol % dispersant consistently yielding the best dispersion and flowability, thereby selected as the optimal condition for DLP printing.

The 0.5 vol % dispersant is optimal for dispersion across all solid loadings (as shown in the previous graphs). Increasing the solid loading beyond 70 wt % significantly increases the viscosity and shear stress, potentially limiting processability. Therefore, a balance between solid content (for mechanical strength) and flow behavior (for processing) is crucial, with 80 wt % as a practical upper threshold for manageable viscosity under optimal dispersion of the ceramic photopolymer slurry. An increasing dispersant content from 0.2 vol % to 0.5 vol % consistently improves dispersion, as seen by decreasing viscosity, indicating reduced interparticle attractions and better flowability. The addition of 0.5 vol % dispersant provided sufficient surface coverage of alumina particles, enhancing steric and electrostatic stabilization while avoiding issues related to overdosing, such as bridging flocculation or increased viscosity due to excess free dispersant. Therefore, 0.5 vol % dispersant was selected as the optimized concentration for subsequent DLP printing processes in this study to ensure highly processable, homogeneous, and defect-free ceramic slurries suitable for the fabrication of complex lattice structures. Hence, 0.5 vol % dispersant is used for the DLP processes in the current study for achieving well-dispersed, processable ceramic suspensions without overdosing the dispersant, as shown in Figure .

10.

(a) Viscosity vs shear rate graph and (b) shear stress vs shear rate graph for standard resin , 60, 70, and 80 wt % solid loading of alumina ceramics using 0.5 vol % dispersant.

3.4. Material Evaluation of Additively Manufactured Ceramics

The lattice structures, fabricated using optimized printing parameters, exhibited excellent build quality and structural integrity, demonstrating the effectiveness of the DLP process in producing complex alumina ceramic geometries (Figure ).

11.

DLP printed alumina green sample lattices.

The printed lattice structures not only showcase the high precision and repeatability of the DLP process but also reflect successful control over critical parameters such as resin formulation, layer curing, and solid loading optimization. The structures exhibit minimal warping and sharp feature resolution even in complex geometries. This is crucial for maintaining mechanical performance. Their open-cell architecture offers significant advantages in terms of mass reduction, thermal management, and fluid flow, making them suitable for multifunctional applications such as scaffold materials in tissue engineering, catalyst supports, lightweight load-bearing components, and dielectric frameworks in electronic packaging. Furthermore, the successful fabrication of both fine and coarse lattice designs confirms the versatility of the DLP printing across a wide range of complex geometries.

The EDS analysis conducted on the green samples confirms that alumina is the primary constituent, as evidenced by the prominent peaks for aluminum and oxygen in the spectrum (Figure ). In the EDS spectra, traces of chlorine (Cl) and phosphorus (P) signals are detected, which are most likely originating from the dispersant or residual binder components. Incomplete removal of these additives during sintering can lead to the formation of low-melting chloride or phosphate compounds. Such species may accumulate at grain boundaries and form thin glassy intergranular films, which can locally alter diffusion behavior, and this can hinder neck growth between particles. As a result, densification may be uneven, and the final microstructure can contain weak grain boundaries or minor glassy phases. These features can degrade mechanical strength and fracture resistance and may also influence dielectric or ionic properties depending on the ceramic composition. Therefore, ensuring complete burnout of organics through optimized thermal debinding is essential for achieving stable microstructures and reliable properties.

12.

EDS analysis of the DLP printed alumina green sample.

The thermal events associated with the organic resin and dispersant used for the preparation of the ceramic slurry for DLP printing are depicted in Figure . The exothermic peak is detected in the range of 200–250 °C, which is attributed to the oxidation of low molecular weight organic compounds or additives such as dispersants or plasticizers. This early thermal event likely represents the initial breakdown or combustion of more volatile components prior to the main decomposition step. Organic binders generally start decomposing around 200 °C–300 °C, releasing heat as they break down or oxidize, causing an exothermic event. Further, the endothermic peak that is observed indicates thermal decomposition of higher molecular weight organic components, i.e., the photoresin in this work. Any remaining volatile organics are eliminated at this point, resulting in further mass loss. Beyond 450 °C, the sample primarily consists of ceramic material (alumina). The TG curve indicates that most organics have been burned off and shows a stable thermal behavior.

13.

TG–DTG profile of the DLP printed green alumina sample.

The XRD pattern of alumina-printed DLP is indexed and presented in Figure . In the XRD pattern of the green, debinded, and sintered samples, prominent peaks observed throughout the scan range are characteristic of α-alumina (corundum phase), which is the stable and dominant crystalline phase of aluminum oxide at high temperatures. The consistent presence of the peaks across all stages (green, debinded, and sintered) confirms the retention of the α-alumina phase during processing. Further, no new or secondary phase formation or undesirable transformation is observed, which agrees with the previous research reports. , This suggests that there is a consistent mechanical property and thus is expected to retain desirable properties like thermal stability. ,

14.

XRD analysis confirming α-alumina phase in both raw powder and DLP-printed sample.

The polymer decomposed cleanly during debinding, without leaving behind any detectable crystalline residue or reaction products. Additionally, the absence of shifting or broadening of the main alumina peaks suggests that the crystallite size and strain are not significantly altered, although slight sharpening of peaks in the sintered sample may indicate grain growth and improved crystallinity due to high-temperature densification (as seen in the green-to-sintered transition).

3.5. Structural Evolution of DLP Printed Alumina

In the green-state SEM micrographs (Figure ), the alumina particles are, for the most part, uniformly embedded within the cured photopolymer matrix (Figure a–b), reflecting the effectiveness of our dispersant system and mixing protocol. At higher magnification (Figure c–d), however, a small number of residual agglomerates (on the order of 2–5 μm) and interparticle voids become apparent, indicating localized regions where dispersion was not fully achieved. The characteristic “staircase” striations arising from the layer-by-layer DLP process are also clearly visible in Figure a, underscoring the need to balance layer thickness and cure depth to minimize surface stepping.

15.

SEM micrographs showing the microstructure of alumina particles dispersed in photopolymeric resin. Images captured at 15 kV accelerating voltage using a secondary electron (SE) detector with a magnification of (a) 360X, (b) 3KX, (c) 25KX, (d) 75KX, respectively.

16.

SEM micrographs of sintered alumina (60 wt % solids) fired at 1550 °C for 4 h: (a) low-magnification view showing uniform grain distribution; (b) higher-magnification detail highlighting angular/subrounded grain morphology and isolated pore sites.

In this green body, alumina particles are held in place by both mechanical interlocking and a network of interfacial interactions with the photocured resin. Physical adsorption, principally van der Waals forces and hydrogen bonding between resin polar groups (e.g., hydroxyl or ester functionalities) and surface −OH groups on alumina, promotes initial particle stabilization and helps suppress extensive agglomeration. Complementing these weaker interactions, specific functional moieties in the resin can form stronger chemical bonds (e.g., esterification or coordination at Lewis-acidic sites on the alumina surface), further enhancing green strength and ensuring adequate particle–binder adhesion during debinding. Together, these features control both the microstructural integrity of the printed green body and the pathways for binder removal, with direct implications for the densification and defect evolution in the sintered ceramic.

Hydrogen bonds between hydroxylated alumina surfaces and electronegative moieties in the resin oligomers further reinforce interfacial anchorage, thereby improving mechanical interlocking at the particle–binder boundary. Complementarily, radical-mediated covalent linkages formed during photopolymerization build a chemically cross-linked network whose elevated cohesive strength translates into greater durability and mechanical performance in the sintered ceramic. Where agglomeration is minimal, surface-active dispersants adsorb onto alumina particles to form a steric shell that exploits entropic repulsion, maintaining particle separation and slurry stability. During UV cure, these adsorbed macromolecules become incorporated into the growing polymer network, effectively locking each coated particle in place and preserving dimensional fidelity in the green layers. Finally, any residual charges on ceramic grains or ionizable resin components give rise to long-range electrostatic forces, either repulsive or attractive, that further modulate dispersion quality and help prevent sedimentation in the slurry.

In the sintered state (Figure ), the 60 wt % solid-loaded alumina exhibits a homogeneous, fine-grained microstructure with grains ranging from angular to subrounded shapes (Figure a–b). Well-defined grain boundaries are evident, and the limited coalescence between neighboring grains indicates that grain growth was effectively suppressed under the 1550 °C/4 h sintering regime. Although the body approaches full density, discrete, submicron pores persist at triple junctions and along some grain boundaries, revealing that ideal densification was not fully realized. This residual porosity can be traced back to the lower initial packing density of the green part, which constrained the driving force for mass transport during sintering. Overall, the dominance of densification mechanisms over coarsening processes in this composition underscores the importance of optimizing both solid loading and thermal profile to achieve truly pore-free alumina ceramics.

The SEM micrographs of the 80 wt % solid-loaded alumina sintered at 1550 °C for 4 h are presented in Figure . In contrast to the 60 wt % sample, the microstructure is markedly more compact, with minimal visible porosity and well-developed grain boundaries. The grains exhibit a more equiaxed and smooth morphology, with clear evidence of particle fusion and grain boundary advancement. Some grain coarsening is evident, as particles appear slightly larger and more rounded compared to the 60 wt % sample. This suggests that the higher initial solid loading provided a more densely packed green structure, which facilitated effective mass transport during sintering, promoting both densification and moderate grain growth. The reduced pore volume and improved connectivity between grains are indicative of enhanced sintering kinetics at higher solid loading.

17.

SEM micrographs of sintered alumina (80 wt % solids) fired at 1550 °C for 4 h: (a) low-magnification image showing dense grain distribution; (b) high-magnification detail highlighting equiaxed grain morphology and absence of residual porosity.

Despite being processed under identical sintering conditions, the microstructures of the 60 and 80 wt % alumina samples show distinct differences due to their initial solid loadings. The 60 wt % sample retains finer grains and exhibits some residual porosity, indicating that densification was incomplete and grain growth was minimal. In contrast, the 80 wt % sample shows signs of more advanced sintering: the structure is denser, pores are nearly absent, and moderate grain growth has occurred. This contrast highlights the critical role of solid loading in influencing sintering behavior. A higher solid loading leads to a denser green body, which promotes faster densification and allows the microstructure to enter the grain growth stage within the same thermal profile.

The grain size distribution of alumina sintered from 60 wt % alumina is skewed toward finer grains, mostly below 0.5 μm illustrated in Figure a. The corresponding SEM image (also shown in Figures and ) confirms a relatively irregular and dense grain morphology with limited grain growth. This indicates that lower solid loading constrains particle packing during photopolymerization and sintering, leading to restricted coalescence and smaller grain sizes. Whereas, a broader and more symmetric distribution extending beyond 1 μm, with the SEM image revealing more spherical and uniformly grown grains for the 80 wt % alumina loaded sample presented in Figure b. The higher solid loading facilitates improved particle–particle contact and enhanced diffusion during sintering, promoting grain coarsening and densification. Thus, increasing the solid loading from 60 wt % to 80 wt % results in a distinct shift from fine, irregular grains to larger, rounded grains, highlighting the influence of alumina powder packing on sintered microstructure.

18.

Grain size distribution and SEM micrographs of DLP-printed alumina parts prepared with (a) 60 wt % and (b) 80 wt % solid loading. The grain size analysis was performed using ImageJ to quantify microstructural differences between the two compositions.

In the as-printed (green) hexagonal lattices shown in Figure , the staircase-like layer lines are distinctly visible, and the thin struts retain sharp edges, highlighting the high fidelity of our optimized alumina-resin formulation. However, because these parts have not yet undergone sintering and thus lack ceramic densification, they remain mechanically fragile and prone to chipping along the fine features. Dimensional accuracy is further challenged by two intrinsic aspects of DLP printing: light scattering within the ceramic-laden resin and the finite pixel resolution of the projector. When UV photons diffuse beyond each targeted pixel, slight overcuring occurs at the periphery, enlarging designed holes and narrowing open apertures relative to the CAD model. Likewise, thicker layer settings exacerbate this “bleed,” as each cured slice overlaps more extensively with its neighbours. Although our high-resolution DLP system and tuned exposure parameters minimize these effects, achieving true-to-print lattice geometries in the green-state ultimately requires a careful compromise among resin opacity, layer thickness, and pixel pitch to control light dispersion and maintain feature fidelity.

19.

Optical microscopy of green-state hexagonal lattice patterns printed from the optimized alumina photopolymeric resin, illustrating layer lines, fine strut definition, and dimensional deviations due to light scattering.

Optical micrographs of the hexagonal lattice after sintering (Figure a–c) reveal a uniform dimensional shrinkage of the struts and apertures, on the order of 10–15% compared to the green-state, consistent with bulk densification. Grain coalescence and neck growth during the 1550 °C/4 h cycle smooth the originally sharp strut edges, producing a slight blurring of fine features (Figure b–c), yet the overall open-cell topology and mechanical integrity remain fully preserved. A pronounced color transition from the darker, resin-laden green body to a lighter, homogeneous alumina tone underscores the complete burnout of the photopolymer binder and the formation of a continuous ceramic network. The minimal warping or distortion of the lattice geometry further confirms that the chosen sintering profile promotes effective particle fusion without compromising the designed architecture.

20.

Optical images of the sintered hexagonal lattice structure (1550 °C, 4 h), showing (a) overall lattice compaction, (b) strut edge smoothing due to grain growth, and (c) reduced feature dimensions from sintering shrinkage.

The quantitative evaluation of the staircase effect is calculated using two key metrics: cusp height and volumetric error (VE). The cusp height represents the local vertical deviation between the nominal curved surface and the approximated layered surface, and is calculated for each surface facet as

$${\mathrm{h}}_{\mathrm{cusp}}={\mathrm{t}}_{\mathrm{layer}}\times |\cos \theta |$$

where t_layer is the layer thickness and θ is the angle between the build direction and the facet’s normal vector. To capture the cumulative effect of surface stepping over the entire geometry, the volumetric error (VE) is computed by integrating cusp contributions across all triangular surface elements. Thus, there are two ways to measure the staircase effect. One is the average weighted cusp height, and then the sum of volumetric differences. The illustration of the staircase effect is depicted in Figure .

21.

SEM micrograph of a DLP-printed alumina strut showing the layer-by-layer deposition. The corresponding schematic illustrates the staircase effect arising from the discrete layer stacking during printing, highlighting the step-like surface morphology at the edges of the structure.

This effect results from the layer-by-layer construction inherent to DLP printing, where each discrete cured layer introduces a small deviation from the ideal smooth geometry. With a layer thickness of 50 μm and a slope angle of 45°, the cusp height, representing the vertical deviation per step, was calculated to be about 35.4 μm. Using geometric analysis, the accumulated volume error caused by this staircase approximation was quantified to be around 8.85 × 10^–6 mm³. Over a 1 mm long inclined surface, this results in 20 discrete steps or layers. Thus, the total volumetric error is observed as 1.77 × 10^–4 mm³. This value represents the total material deviation between the ideal and printed surface over that 1 mm length. Such quantification is important when evaluating dimensional accuracy and surface fidelity, particularly in microfeatured ceramic components where even small volume errors can impact performance or aesthetics.

In Figure , the CAD mesh (Figure a) specifies a nominal gyroid pore diameter of 0.75 mm, yet optical measurements on the DLP-printed green body (Figure b) yield an average aperture of 0.853 mm, a nearly 14% overshoot. Such dimensional expansion can be traced to light overcuring from photon scattering in the ceramic-laden resin, compounded by the resin’s rheological properties, viscosity, solids loading, and particle size distribution, which govern layer recoating uniformity and local cure depth. Additionally, uncorrected manufacturing tolerances and insufficient shrinkage compensation during scale-up further magnify the deviation between the intended and actual geometry. To bring printed features into closer alignment with CAD targets, it is therefore necessary to apply iterative offset factors in the digital model (e.g., reducing aperture dimensions by the measured overshoot), refine resin formulation to optimize flow and light attenuation, and recalibrate exposure settings to curb peripheral cure bleed. ,

22.

(a) Gyroid CAD design with targeted 0.75 mm pore. (b) Optical microscopy of green-state DLP print showing an average 0.85 mm aperture.

The densification metrics for DLP-printed alumina ceramics across several investigations are compiled in Table . Early work by Pan et al. yielded only 2.4–2.5 g/cm³ (≈44–46% of theoretical density) with open porosities of 36.7–39.9%, reflecting suboptimal green packing and limited sintering. By 2021, improvements in slurry formulation and exposure control enabled densities up to 3.88 g/cm³ (≈97.5% of theoretical) with negligible residual porosity. Subsequent studies focused on fine-tuning sintering profiles to suppress voids, reporting considerable porosity reduction though often without explicit density values. More recently, the choice of photopolymer monomer has been shown to critically influence outcomes: DPGDA-based resins achieved 97.5 ± 0.5% relative density and ∼9.8% porosity, whereas TMPTA analogues exhibited higher void fractions. In comparison, our present work attains 90.1% relative density with 9.9% total porosity, highlighting the ongoing challenges of binder burnout, green-body compaction, and thermal regimen optimization under the applied 1550 °C/4 h protocol.

2. Density and Porosity Obtained for Various References Based on DLP-Printed Alumina Ceramics.

S No.	Ref.	Year	Density (%) or Relative Density (g/cm3)	Porosity (%)	Mechanical Outcomes
1	Open porosity decreased slightly with increased holding time.	2020	2.4–2.5 g/cm3	36.7–39.9%	-
2	Achieved a density of 3.88 g/cm3, close to the theoretical density.	2022	1.36 g/cm3	-	Bending strength 8.58 MPa
3	Monomers like DPGDA showed minimal porosity, while TMPTA resulted in higher porosity.	2024	97.5 ± 0.5%	≈9.8%	-
4	Atmosphere debinding heat treatment of 3D printed alumina ceramics	2022	96.72%	-	Compressive strength ∼761.7 MPa
5	Digital light processing 3D-printing alumina-based ceramic core with high	2025	2.15 g/cm3	41.40%	Flexural strength ∼28.28 MPa
6	Digital light processing followed by pressureless sintering of metal-reinforced ceramics	2025	98.7%–99.6%	-	-
7	Present work – aimed to achieve complicated structures with good structural integrity.	-	90.10%	9.9%	Hexagonal Lattice- 8.3 MPa
					Truss-Plate Lattice- 1.3 MPa

While Zeng et al. required an argon atmosphere to achieve 96.7% density in DLP-printed alumina, the present study attains defect-free lattice structures under an air atmosphere using a higher solid loading (80 wt %) and controlled sintering (2 °C min^–1 with staged dwell holds), confirming the robustness of the developed formulation and process. Similarly, high-porosity alumina ceramic cores were achieved via DLP and achieved 41.4% porosity with a density of 2.15 g/cm³ and a flexural strength of 28.3 MPa at a sintering temperature of 1450 °C (heating rate 4 °C min^–1, hold time 120 min) under their optimized conditions by Zheng et al. Whereas the present study advances the field by employing a significantly higher solid-loading slurry (80 wt %), using a refined, defect-free lattice architecture, and implementing a controlled sintering protocol with a heating rate of 2 °C min^–1 under air atmosphere, yielding a relative density of 90.10% and a flexural strength of 8.3 MPa achieved for hexagonal lattice with 90% geometric porosity. Unlike Tanska et al., who utilized a controlled Ar/H₂ atmosphere during sintering to prevent defect formation, the present study achieves structurally intact, crack-free alumina lattices processed entirely under an air atmosphere, underscoring the effectiveness of the optimized slurry and sintering protocol.

The dimensional contraction of a 60 wt % alumina lattice subjected to our standard sintering profile (1550 °C, 4 h) is depicted in Figure . The green body measuring 11.63 mm in width and 11.18 mm in height reduces to 10.36 mm and 9.04 mm, respectively, corresponding to a planar shrinkage of ∼10.9% and an axial shrinkage of ∼19.2%. This pronounced size reduction arises from the volatilization of the photopolymer binder during debinding and the subsequent ceramic particle rearrangement and pore elimination during densification. Accurately predicting such anisotropic shrinkage is therefore essential: design models must incorporate scale-up compensation factors to offset the expected contraction and ensure that final ceramic components meet tight dimensional tolerances.

23.

Photographic comparison of a hexagonal lattice (60 wt % solids) before (left: 11.63 × 11.18 mm) and after sintering (right: 10.36 × 9.04 mm), highlighting the need for shrinkage allowances in DLP-printed ceramic designs.

As summarized in Table , increasing the alumina content from 60 to 80 wt % markedly reduces sintering shrinkage while boosting final density. At 60 wt % solids, the green bodies contract by approximately 15.6% in the X–Y plane and 20.8% along Z, yielding only 75.1% theoretical density. By contrast, the 80 wt % specimens exhibit just 10.5% X–Y and 11.3% Z shrinkage, and reach 90.1% density. The intermediate 70 wt % composition shows a proportional trend (12.5% X–Y, 15.3% Z, 81.4% density), underscoring that higher initial solids loading improves particle packing, reduces the magnitude of pore closure required, and drives more uniform densification. Moreover, the convergence of X–Y and Z shrinkage values at higher loadings indicates diminished anisotropy in sintering shrinkage, an important factor for preserving dimensional fidelity in complex ceramic geometries.

3. Shrinkage after Sintering of Samples at 1550°C/4hH for 60 wt %, 70 wt %, and 80 wt % Alumina Ceramics.

S No.	Solid Loading (wt %)	Dimensional Shrinkage (%)	Density (%)
1	60	X, Y = 15.60	75.06
		Z = 20.80
2	70	X, Y = 12.50	81.40
		Z = 15.30
3	80	X, Y = 10.50	90.10
		Z = 11.30

The overall volume shrinkage of our alumina ceramics is observed to drop sharply as the solid loading increases: from 29.2 ± 1.5% at 60 wt % to 20.1 ± 1.2% at 70 wt % and down to just 9.7 ± 1.1% at 80 wt % solids (present study, Figure ). This behavior arises because higher ceramic fractions entail less polymeric binder that must be burned out during debinding, leading to a more tightly packed green body and correspondingly lower volumetric contraction upon sintering. By minimizing shrinkage, elevated solid loadings also mitigate the buildup of internal stresses and reduce the risk of crack initiation or warping, thereby safeguarding dimensional accuracy and mechanical integrity in intricate lattice geometries. ,

24.

Variation of volume shrinkage with solid loading in ceramic slurry; error bars denote standard deviation of multiple measurements.

The anisotropic shrinkage evaluated from the sintered lattices can arise from nonuniform particle packing and the development of internal stresses during sintering. In addition, ceramic stereolithography specimens possess a characteristic layered architecture, leading to anisotropic sintering behavior and reduced material resistance at the interlayer regions, where higher porosity is often observed. − Therefore, any sintering model for printed lattices should account for both densification kinetics and the anisotropic behavior of porous specimens that exhibit greater compaction along the printing direction. , In general, printed ceramics display orthotropic anisotropy, characterized by similar shrinkage behavior within the layer plane (X–Y directions) but distinct shrinkage along the build (Z) direction. This orthotropic behavior has been experimentally confirmed through triaxial dilatometry, revealing shrinkage predominantly along the building direction, and by microscopic analyses showing a higher porosity in the interlayer zones. ,− Similar orthotropic sintering behavior has also been reported in other additive manufacturing methods such as robocasting, fused deposition modeling, , and binder jetting. ,

The anisotropic shrinkage in the X–Y and Z directions can significantly influence the mechanical performance and geometric fidelity of DLP-printed alumina lattices. Differences in shrinkage along different axes can lead to uneven load distribution and localized stress concentrations under compressive or bending loads, potentially initiating microcracks or early failure in regions with higher dimensional changes. Also, anisotropic densification can affect the preservation of lattice geometry, altering strut alignment and node connectivity, which in turn impacts the overall stiffness and strength of the structure. Therefore, accounting for these effects is essential for optimizing sintering protocols and predicting the mechanical response of ceramic lattices.

3.6. Mechanical Behavior of Hexagonal and Truss–Plate Triangular Alumina Lattices under Compression

The uniaxial compression tests, carried out on 20 × 20 × 20 mm green-state lattices, reveal a dramatic performance gap between the two architectures: the hexagonal lattice sustains a peak stress of 8.3 MPa before collapse, whereas the truss–plate triangular lattice fails at just 1.32 MPa. This disparity arises from fundamental differences in load transmission and failure modes. In the hexagonal topology, each strut is inclined at approximately 30° to the loading axis, so compressive forces resolve into a mix of axial and bending stresses; the closed-cell geometry and distributed three- and four-way nodes promote uniform stress sharing, delay local buckling, and produce a broad plateau of energy absorption. By contrast, the triangular lattice’s straight, slender members act like cantilevers under vertical load, concentrating bending stresses at the joints and causing early brittle fracture of the alumina struts. Moreover, according to Gibson–Ashby scaling for brittle foams, strength scales with relative density to a power >1; the slightly higher effective density of the hexagon network thus further amplifies its advantage. The stress maps (Figure ) confirm that the hexagon’s curved ligaments channel stresses smoothly around nodes with minimal peak intensities, whereas the triangular layout exhibits sharp stress concentrations and rapid strut buckling. These findings underscore that, for ceramic lattices, optimized cell orientation and nodal connectivity are as critical as material properties in maximizing compressive strength.

25.

Schematic representation of stress distribution in the hexagonal lattice under uniaxial compression, highlighting low peak stresses at curved strut–node junctions and minimal buckling.

In stark contrast to the robust performance of the hexagonal lattice, the truss-plate triangular network fails almost exclusively by local Euler-type buckling of its slender struts (Figure ). Here, the high slenderness ratio (length-to-thickness) of each inclined member under uniaxial compression precipitates bending at the nodal intersections, where axial loads convert into concentrated bending moments. Because the triangular topology lacks continuous cell walls, there is no lateral confinement to arrest this bending; instead, load transfer becomes highly anisotropic, funneling stresses into a few critical struts and triggering a cascade of hinge-like failures at only 1.32 MPa. The resulting load–displacement curve exhibits a sharp peak followed by an abrupt collapse, characteristic of brittle lattice collapse. By comparison, the honeycomb’s closed-cell architecture provides multidirectional support and enlists multiple load paths, allowing stresses to redistribute around curved ligaments and delaying buckling until much higher loads, hence the 8.3 MPa plateau observed in compression tests. These observations confirm that, for ceramic DLP lattices, achieving high compressive strength demands not only dense material but also a topology that combines low slenderness ratios with nodal geometries capable of mitigating local bending and stress concentrations.

26.

Buckling behavior of truss–plate lattice under compression showing localized deformation and corresponding load- displacement response.

The stress-deformation behavior of ceramic lattice structures with varying porosities and architectures is illustrated in Figure . The data represent mean values (n = 3), and error bars represent for the Hexagonal lattice and truss-plate lattice indicate standard deviation.

27.

Compression stress–deformation curves of 90% porous hexagonal and truss–plate lattices. Data represent mean values (n ≥ 3). Error bars for the hexagonal and truss–plate lattices indicate standard deviations.

Despite the increase in porosity, which typically leads to reduced mechanical integrity, the hexagonal lattice in our work demonstrates a compressive strength (8.3 MPa) significantly higher than literature cases F and G, and only slightly lower than H (11.7 MPa). This suggests that the hexagonal geometry contributes to superior load-bearing capacity due to efficient stress distribution and structural integrity, even at increased porosity levels. The stress-deformation curve for the hexagonal structure also shows a gradual and stable deformation profile, indicating ductile-like failure progression in the brittle ceramic system, likely influenced by the designed deformation pathways in the lattice.

On the other hand, the truss plate lattice, although also at 90% porosity, exhibits a lower compressive strength of 1.32 MPa. While this value is marginally higher than G from literature, it remains significantly lower than the hexagonal structure, highlighting the influence of architecture over porosity alone in determining mechanical performance. The truss plate structure shows an early onset of deformation and a more extended stroke, indicative of higher compliance and energy absorption, but at the cost of lower strength.

Thus, the present work’s results demonstrate that through architectural optimization, it is possible to achieve compression strengths in highly porous ceramic lattices (90%) that are comparable to or exceed those reported for lower porosity structures (83%). Future work could explore hybrid designs or material modifications to enhance performance further.

4. Conclusion

This study demonstrates a predictive, design-oriented approach to DLP fabrication of alumina ceramics:

• Shrinkage control: An increase in the solid loading of alumina from 60 wt % to 80 wt % significantly improved dimensional stability during sintering. At lower solid loadings, the higher resin content results in greater polymer removal during debinding, leaving behind a more porous green body. This undergoes substantial densification and volumetric shrinkage during sintering. By increasing the ceramic content to 80 wt %, the relative amount of polymeric binder with respect to alumina is less, reducing the volume of material lost during burnout. This leads to a more rigid and densely packed green structure, thereby minimizing overall shrinkage. Consequently, the volumetric shrinkage was reduced by nearly two-thirds compared to the 60 wt % sample, while achieving a high sintered density of 90.10% with minimal dimensional deviation between the printed and final sintered parts. This demonstrates effective shrinkage control through optimization of solid loading, which is crucial for maintaining geometric accuracy in DLP-based ceramic printing.

• Staircase effect quantification: The staircase effect was rigorously quantified through cusp height and volumetric error analysis, providing predictive tools for surface quality evaluation. The cumulative volume error arising from the staircase approximation was calculated to be approximately 8.85 × 10^–6 mm³ per layer. The total volumetric error accumulates to about 1.77 × 10^–4 mm³ for a 1 mm-long inclined surface comprising 20 discrete layers.

• Architecture–mechanics correlation: Mechanical testing revealed that the lattice architecture plays a more dominant role than porosity in determining load-bearing performance. The hexagonal lattice structure exhibited a compressive strength of 8.3 MPa, which is nearly six times higher than that of the truss–plate lattice with 1.32 MPa. This substantial difference highlights the importance of geometric configuration in effectively distributing stress and enhancing the structural integrity of ceramic lattices.

• Generalizable framework: By systematically correlating slurry rheology, optical curing behavior, sintering shrinkage, and mechanical performance, this study establishes a comprehensive and transferable framework for predicting defect formation and optimizing processing parameters in ceramic additive manufacturing. This integrated approach enables the identification of interdependent factors influencing print quality and final part integrity, making it applicable across various ceramic systems and printing platforms. Therefore, instead of focusing only on processing, this work establishes quantitative metrics and design guidelines for defect-controlled, high-fidelity ceramic lattices, offering valuable insights for biomedical, aerospace, and electronic applications.

The techniques and data are available within the paper.

A.A.: Conceptualization, Data curation, Formal analysis, Investigation, Writing – original draft; R.N.: Supervision, Validation, Visualization, Writing – review and editing.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

The authors declare no competing financial interest.