Evaluating Directional Dependency of Selective Laser Sintered Patient Specific Biodegradable Devices to Improve Predictive Modeling and Design Verification.

Abstract

Additive manufacturing, or 3D printing, of the bioresorbable polymer [Formula: see text]-polycaprolactone (PCL) is an emerging tissue engineering solution addressing patient specific anatomies. Predictively modeling the mechanical behavior of 3D printed parts comprised of PCL improves the ability to develop patient specific devices that meet design requirements while reducing the testing of extraneous design variants and development time for emergency devices. Predicting mechanical behavior of 3D-printed devices is limited by the variability of effective material moduli that are determined in part by the 3D printing manufacturing process. Powder fusion methods, specifically laser sintering, are known to produce parts with internal porosity ultimately impacting the mechanical performance of printed devices. This study investigates the role of print direction and part size on the material and structural properties of laser sintered PCL parts. Solid PCL cylinders were printed in the XY (perpendicular to laser) and Z direction (parallel to laser), scanned using microcomputed tomography, and mechanically tested under compression. Compositional, structural, and functional properties of the printed parts were evaluated with differential scanning calorimetry, gel permeation chromatography, microcomputed tomography, and mechanical testing. Computational models of printed and scanned cylinders were fit to experimental data to derive effective moduli. Effective moduli were used to predict the mechanical behavior of splints used for emergency repair of severe tracheobronchomalacia. Laser sintering did not cause significant differences in polymer material properties compared to unmanufactured powder. Effective moduli (Eeff) were greater for larger part sizes (p < 0.01) and for parts oriented in the XY direction compared to the Z direction (p < 0.001). These dependencies were congruent with the differences in void volumes associated with the print direction (p < 0.01) and part size (p < 0.01). Finite element models of splint parallel compression tests utilizing the Eeff dependent on print direction and size agreed with experimental closed compression tests of splints. Evaluating the microstructural properties of printed parts and selecting effective moduli for finite element models based on manufacturing parameters allows accurate prediction of device performance. These findings allow testing of a greater number of device design variants in silico to accomodate patient specific anatomies towards providing higher quality parts while lowering overall time and costs of manufacturing and testing.