The research team, led by Israa Elfaki and Li Cao with colleagues at several institutions, dispersed three grades of hexagonal boron nitride microplatelets into a commercial UV-curable resin. The fillers, purchased from Momentive Technologies, had average lateral sizes of approximately 0.9, 12, and 45 micrometers, and the researchers labeled them BN0.9, BN12, and BN45 accordingly. They mixed the powders into the Anycubic Standard Resin+ photopolymer using a SpeedMixer at 2500 rpm for three minutes, targeting loadings of 5 to 20 percent by weight. Each formulation was then printed on an MSLA printer with a 405 nanometer light source, an XY resolution of 0.05 millimeters, and layer thicknesses of 10, 20, 30, and 50 micrometers. Printed parts were washed in isopropanol and post-cured under a 450-watt UV system to complete polymerization.

Before any specimens could be printed, the team had to establish what the fillers did to the feedstock itself. Rheological measurements at 23 degrees Celsius showed that all formulations exhibited pronounced shear-thinning, a desirable trait because viscosity drops under shear during resin recoating and build-plate movement. Adding boron nitride raised viscosity, and the effect grew stronger at higher loadings and with smaller platelets, which present a larger specific surface area. The 20 percent BN0.9 formulation proved so viscous that reliable measurement was impossible, and it was excluded from printing, marking a practical printability limit. The researchers also noted that boron nitride, at 2.25 grams per cubic centimeter, is denser than the 1.2 grams per cubic centimeter resin, making particle settling a concern during long prints, and that the platelets scatter or block ultraviolet light, forcing longer exposure times as loading rises.

Density measurements on printed 10-millimeter cubes suggested these processing hurdles did not compromise the parts. Using the Archimedes principle, the team found that density rose consistently with filler loading for all three platelet sizes, closely tracking theoretical values from a rule-of-mixtures calculation. Changing layer thickness from 10 to 50 micrometers had no significant effect on density or void formation, and deviations from theory reached only about 2 percent or less, seen mainly in the 20 percent BN12 and BN45 composites and attributed to viscosity-related effects on dispersion and packing. The results pointed to effective filler incorporation and low porosity across the printed parts.

The central question was orientation, and the team attacked it with X-ray diffraction on composite cubes measured along both the X-Y plane, parallel to the printing layers, and the X-Z plane, perpendicular to them. Because the layered crystal structure of hexagonal boron nitride produces distinctive basal-plane reflections, the ratio of the (002) and (100) peak intensities could be converted into an orientation function, where zero represents random orientation, one represents basal planes lying perfectly parallel to the measured surface, and negative one-half represents perpendicular orientation. For BN45 composites at 15 percent loading, decreasing the layer thickness from 50 to 10 micrometers steadily increased the alignment function measured from the printing plane, confirming that thinner layers force platelets to lie flat. Optical microscopy of 5 percent BN12 and BN45 composites provided direct visual confirmation, showing pronounced, directionally consistent platelet alignment at 10 micrometers that grew progressively more randomized as layer thickness increased to 50 micrometers. Cross-sectional scanning electron microscopy showed the platelets distributed without obvious large-scale agglomeration.

Platelet size mattered just as much as layer geometry. At a fixed 10-micrometer layer thickness, the XRD comparison showed the strongest basal-plane alignment for BN45, an intermediate degree for BN12, and the weakest for BN0.9. The explanation is geometric: a 45-micrometer platelet cannot rotate freely inside a 10-micrometer gap between the previously cured layer and the transparent vat film, so shear flow during resin recoating and geometric confinement jointly press it flat. A 0.9-micrometer platelet, by contrast, retains enough rotational freedom to remain largely random. The authors note that MSLA shear is weaker than the nozzle-induced shear of extrusion methods like fused deposition modeling or direct ink writing, so confinement plays an especially important role when the layer thickness is comparable to or smaller than the platelet’s lateral dimension.

Thermal measurements translated the microstructure into performance. Thin films roughly 50 micrometers thick were tested for in-plane thermal diffusivity on a LaserPIT instrument under vacuum, with duplicate measurements averaging a relative difference of about 4.2 percent. Thermal conductivity was then calculated from diffusivity, Archimedes-measured density, and specific heat capacities estimated by the rule of mixtures and spot-checked against differential scanning calorimetry. In-plane thermal conductivity rose with filler loading for all three composites, but the layer-thickness effect depended strongly on platelet size. BN0.9 composites showed only modest changes as layer thickness shrank, consistent with their poor alignment, while BN12 and especially BN45 composites gained conductivity as layers thinned. The peak result was 1.035 watts per meter-kelvin for the 20 percent BN45 composite printed at 10 micrometers, roughly a 246 percent increase over the neat resin’s 0.299 watts per meter-kelvin, achieved at a moderate filler loading.

Infrared thermography offered a vivid demonstration of why orientation matters. The team printed one-centimeter cubes containing 15 percent BN45 at a 10-micrometer layer thickness in two configurations, one with platelets aligned parallel to the heat flow direction and one perpendicular, alongside a neat resin reference cube. Placed on thermal tape and heated from below, the vertically aligned cube, with platelets parallel to the heat flow, warmed fastest at its top surface, the horizontally aligned cube lagged, and the neat resin was slowest. The authors caution that their experiments spanned a limited loading range and were not designed to identify a percolation threshold, and because interfacial thermal resistance at the boron nitride-polymer boundary was not independently measured, no single theoretical model was quantitatively fitted to the data.

Mechanical testing on ASTM D638 Type V tensile dogbones told a more nuanced story. Young’s modulus generally increased with boron nitride addition for all three composites, reflecting load transfer to the stiff ceramic fillers, and the reinforcement was strongest for larger platelets printed at thinner layers, consistent with the alignment results. The BN45 composites reached the highest elastic modulus, though at 20 percent loading the modulus dropped at every layer thickness, possibly due to dispersion difficulties at that high loading. Tensile strength behaved less predictably: it held steady or improved slightly at 5 to 10 percent loading for BN0.9 but often declined at higher loadings, particularly at 20 percent for BN12 and BN45, since unfunctionalized platelets can interrupt the polymer matrix and act as stress concentration sites. Elongation at break fell below the neat resin across the board, indicating reduced ductility. The overall picture is that boron nitride reliably stiffens these printed composites while leaving ultimate tensile strength sensitive to composition, dispersion, and printing conditions.

The study’s practical message is that masked stereolithography offers a route to thermally tailored polymer parts without exotic chemistry. Because the fillers were used as-received, without surface functionalization, the gains came purely from processing-induced orientation, controlled by filler size and layer thickness. The authors position the approach for thermal management applications such as electronics packaging, where heat dissipation must coexist with electrical insulation, and note that the same layer-by-layer control could extend to complex geometries that conventional casting, stretching, or pressing methods cannot produce. With systematic comparisons of three platelet sizes, four layer thicknesses, and loadings up to 20 percent, quantified by XRD, microscopy, thermography, and mechanical testing, the work maps a clear parameter space for designers who need heat to flow in a chosen direction through a printed insulating part.

Subject of Research: MSLA 3D printing of hexagonal boron nitride platelet-filled polymer composites with tunable thermal and mechanical properties

Article Title: 3D printing of boron nitride polymer composites for tailored thermal transport and mechanical performance

Article References: Elfaki, I., Han, J., Cormier, R., Keune, M., Tran, V. D. H., Brown, K., Meziani, M. J., Sun, Y.-P., & Cao, L. (2026). 3D printing of boron nitride polymer composites for tailored thermal transport and mechanical performance. Journal of Materials Science. https://doi.org/10.1007/s10853-026-13938-8

Image Credits: AI Generated

DOI: 10.1007/s10853-026-13938-8

Keywords: 3D printing, boron nitride, polymer composites, thermal conductivity, masked stereolithography, platelet alignment, X-ray diffraction, vat photopolymerization, thermal management, mechanical properties, composite materials, additive manufacturing