Accelerating the design of lattice structures using machine learning.

Lattices remain an attractive class of structures due to their design versatility; however, rapidly designing lattice structures with tailored or optimal mechanical properties remains a significant challenge. With each added design variable, the design space quickly becomes intractable. To address this challenge, research efforts have sought to combine computational approaches with machine learning (ML)-based approaches to reduce the computational cost of the design process and accelerate mechanical design. While these efforts have made substantial progress, significant challenges remain in (1) building and interpreting the ML-based surrogate models and (2) iteratively and efficiently curating training datasets for optimization tasks. Here, we address the first challenge by combining ML-based surrogate modeling and Shapley additive explanation (SHAP) analysis to interpret the impact of each design variable. We find that our ML-based surrogate models achieve excellent prediction capabilities (R2 > 0.95) and SHAP values aid in uncovering design variables influencing performance. We address the second challenge by utilizing active learning-based methods, such as Bayesian optimization, to explore the design space and report a 5 × reduction in simulations relative to grid-based search. Collectively, these results underscore the value of building intelligent design systems that leverage ML-based methods for uncovering key design variables and accelerating design.

3D-Printing of Meso-structurally Ordered Carbon Fiber/Polymer Composites with Unprecedented Orthotropic Physical Properties.

Here we report the first example of a class of additively manufactured carbon fiber reinforced composite (AMCFRC) materials which have been achieved through the use of a latent thermal cured aromatic thermoset resin system, through an adaptation of direct ink writing (DIW) 3D-printing technology. We have developed a means of printing high performance thermoset carbon fiber composites, which allow the fiber component of a resin and carbon fiber fluid to be aligned in three dimensions via controlled micro-extrusion and subsequently cured into complex geometries. Characterization of our composite systems clearly show that we achieved a high order of fiber alignment within the composite microstructure, which in turn allows these materials to outperform equivalently filled randomly oriented carbon fiber and polymer composites. Furthermore, our AM carbon fiber composite systems exhibit highly orthotropic mechanical and electrical responses as a direct result of the alignment of carbon fiber bundles in the microscale which we predict will ultimately lead to the design of truly tailorable carbon fiber/polymer hybrid materials having locally programmable complex electrical, thermal and mechanical response.

Shape-morphing composites with designed micro-architectures.

Shape memory polymers (SMPs) are attractive materials due to their unique mechanical properties, including high deformation capacity and shape recovery. SMPs are easier to process, lightweight, and inexpensive compared to their metallic counterparts, shape memory alloys. However, SMPs are limited to relatively small form factors due to their low recovery stresses. Lightweight, micro-architected composite SMPs may overcome these size limitations and offer the ability to combine functional properties (e.g., electrical conductivity) with shape memory behavior. Fabrication of 3D SMP thermoset structures via traditional manufacturing methods is challenging, especially for designs that are composed of multiple materials within porous microarchitectures designed for specific shape change strategies, e.g. sequential shape recovery. We report thermoset SMP composite inks containing some materials from renewable resources that can be 3D printed into complex, multi-material architectures that exhibit programmable shape changes with temperature and time. Through addition of fiber-based fillers, we demonstrate printing of electrically conductive SMPs where multiple shape states may induce functional changes in a device and that shape changes can be actuated via heating of printed composites. The ability of SMPs to recover their original shapes will be advantageous for a broad range of applications, including medical, aerospace, and robotic devices.

Design and biocompatibility of endovascular aneurysm filling devices.

The rupture of an intracranial aneurysm, which can result in severe mental disabilities or death, affects approximately 30,000 people in the United States annually. The traditional surgical method of treating these arterial malformations involves a full craniotomy procedure, wherein a clip is placed around the aneurysm neck. In recent decades, research and device development have focused on new endovascular treatment methods to occlude the aneurysm void space. These methods, some of which are currently in clinical use, utilize metal, polymeric, or hybrid devices delivered via catheter to the aneurysm site. In this review, we present several such devices, including those that have been approved for clinical use, and some that are currently in development. We present several design requirements for a successful aneurysm filling device and discuss the success or failure of current and past technologies. We also present novel polymeric-based aneurysm filling methods that are currently being tested in animal models that could result in superior healing.

Reticulation of low density shape memory polymer foam with an in vivo demonstration of vascular occlusion.

Predominantly closed-cell low density shape memory polymer (SMP) foam was recently reported to be an effective aneurysm filling device in a porcine model (Rodriguez et al., Journal of Biomedical Materials Research Part A 2013: (http://dx.doi.org/10.1002/jbm.a.34782)). Because healing involves blood clotting and cell migration throughout the foam volume, a more open-cell structure may further enhance the healing response. This research sought to develop a non-destructive reticulation process for this SMP foam to disrupt the membranes between pore cells. Non-destructive mechanical reticulation was achieved using a gravity-driven floating nitinol pin array coupled with vibratory agitation of the foam and supplemental chemical etching. Reticulation resulted in a reduced elastic modulus and increased permeability, but did not impede the shape memory behavior. Reticulated foams were capable of achieving rapid vascular occlusion in an in vivo porcine model.

A high-performance recycling solution for polystyrene achieved by the synthesis of renewable poly(thioether) networks derived from D-limonene.

Nanocomposite polymers are prepared using a new sustainable materials synthesis process in which d-Limonene functions simultaneously both as a solvent for recycling polystyrene (PS) waste and as a monomer that undergoes UV-catalyzed thiol-ene polymerization reactions with polythiol comonomers to afford polymeric products composed of precipitated PS phases dispersed throughout elastomeric poly(thioether) networks. These blended networks exhibit mechanical properties that greatly exceed those of either polystyrene or the poly(thioether) network homopolymers alone.

In vivo response to an implanted shape memory polyurethane foam in a porcine aneurysm model.

Cerebral aneurysms treated by traditional endovascular methods using platinum coils have a tendency to be unstable, either due to chronic inflammation, compaction of coils, or growth of the aneurysm. We propose to use alternate filling methods for the treatment of intracranial aneurysms using polyurethane-based shape memory polymer (SMP) foams. SMP polyurethane foams were surgically implanted in a porcine aneurysm model to determine biocompatibility, localized thrombogenicity, and their ability to serve as a stable filler material within an aneurysm. The degree of healing was evaluated via gross observation, histopathology, and low vacuum scanning electron microscopy imaging after 0, 30, and 90 days. Clotting was initiated within the SMP foam at time 0 (<1 h exposure to blood before euthanization), partial healing was observed at 30 days, and almost complete healing had occurred at 90 days in vivo, with minimal inflammatory response.

Ultra Low Density and Highly Crosslinked Biocompatible Shape Memory Polyurethane Foams.

We report the development of highly chemically crosslinked, ultra low density (~0.015 g/cc) polyurethane shape memory foams synthesized from symmetrical, low molecular weight and branched hydroxyl monomers. Sharp single glass transitions (Tg) customizable in the functional range of 45-70 °C were achieved. Thermomechanical testing confirmed shape memory behavior with 97-98% shape recovery over repeated cycles, a glassy storage modulus of 200-300 kPa and recovery stresses of 5-15 kPa. Shape holding tests under constrained storage above the Tg showed stable shape memory. A high volume expansion of up to 70 times was seen on actuation of these foams from a fully compressed state. Low in-vitro cell activation induced by the foam compared to controls demonstrates low acute bio-reactivity. We believe these porous polymeric scaffolds constitute an important class of novel smart biomaterials with multiple potential applications.

Opacification of shape memory polymer foam designed for treatment of intracranial aneurysms.

Shape memory polymer (SMP) foam possesses structural and mechanical characteristics that make them very promising as an alternative treatment for intracranial aneurysms. Our SMP foams have low densities, with porosities as high as 98.8%; favorable for catheter delivery and aneurysm filling, but unfavorable for attenuating X-rays. This lack of contrast impedes the progression of this material becoming a viable medical device. This paper reports on increasing radio-opacity by incorporating a high-Z element, tungsten particulate filler to attenuate X-rays, while conserving similar physical properties of the original non-opacified SMP foams. The minimal amount of tungsten for visibility was determined and subsequently incorporated into SMP foams, which were then fabricated into samples of increasing thicknesses. These samples were imaged through a pig's skull to demonstrate radio-opacity in situ. Quantification of the increase in image contrast was performed via image processing methods and standard curves were made for varying concentrations of tungsten doped solid and foam SMP. 4% by volume loading of tungsten incorporated into our SMP foams has proven to be an effective method for improving radio-opacity of this material while maintaining the mechanical, physical and chemical properties of the original formulation.