Search for: All records

Ultra-high-molecular-weight polyethylene (UHMWPE) components for orthopedic implants have historically been integrated into metal backings by direct-compression molding (DCM). However, metal backings are costly, stiffer than cortical bone, and may be associated with medical imaging distortion and metal release. Hybrid-manufactured DCM UHMWPE overmolded additively manufactured polyetheretherketone (PEEK) structural components could offer an alternative solution, but are yet to be explored. In this study, five different porous topologies (grid, triangular, honeycomb, octahedral, and gyroid) and three surface feature sizes (low, medium, and high) were implemented into the top surface of digital cylindrical specimens prior to being 3D printed in PEEK and then overmolded with UHMWPE. Separation forces were recorded as 1.97–3.86 kN, therefore matching and bettering the historical industry values (2–3 kN) recorded for DCM UHMWPE metal components. Infill topology affected failure mechanism (Type 1 or 2) and obtained separation forces, with shapes having greater sidewall numbers (honeycomb-60%) and interconnectivity (gyroid-30%) through their builds, tolerating higher transmitted forces. Surface feature size also had an impact on applied load, whereby those with low infill-%s generally recorded lower levels of performance vs. medium and high infill strategies. These preliminary findings suggest that hybrid-manufactured structural composites could replace metal backings and produce orthopedic implants with high-performing polymer–polymer interfaces.

 

(1) Introduction. Although new particle formation (NPF) constitutes an important process in air, there are large uncertainties regarding which species participate in the formation of the first nanoclusters. Acid-base reactions are known generate new particles, with methanesulfonic acid (MSA) from the photooxidation of biogenic organosulfur compounds becoming more important with time relative to sulfuric acid as fossil-fuel related sources of the latter decline. Simultaneously, the use of alkanolamines in carbon capture and storage (CCS) is expected to result in increased atmospheric concentrations of these bases. This study applied a unique mass spectrometry method to examine the chemical composition of 2-10 nm particles from the MSA reaction with monoethanolamine and 4-aminobutanol, the most efficient system for NPF from MSA examined to date. (2) Methods. Thermal desorption chemical ionization mass spectrometry (TDCIMS, HToF mass analyzer, Tofwerk AG) was used to measure the size and acid-to-base molar ratios of nanoparticles formed from the reaction of MSA with multifunctional amines. A high-flow differential mobility analyzer (half-mini DMA, SEADM) was interfaced with the TDCIMS, which provides a high mobility resolution and high particle transmission in the diameter range 2-10 nm, where chemical composition measurements are the most challenging due to the very small amount of mass. With this novel combination of techniques we were able to examine MSA-amine systems either from nanoparticles exiting the outlet of a flow reactor or nanoclusters generated via electrospray. (3) Preliminary Data. These experiments show that MSA-driven acid-base reactions with monethanolamine or 4-aminobutanol are even more efficient in NPF than that of simple alkylamines, exhibiting to date the highest nanoparticle formation rates measured in laboratory flow tube studies. The observed enhancement is rooted in the presence of an -OH group on the parent molecules, which initiates a H-bond network throughout the nanoclusters. In these systems, water had only a minimal enhancing effect. We demonstrated that the nanoparticles formed in both systems are neutral (i.e. contain as much acid as base molecules) in the range 2-10 nm. This contrasts with MSA reactions from previous studies on the smallest alkylamine, methylamine, where particles smaller than 9 nm were more acidic. Investigations of reactions of MSA with a diamine (1,4-diaminobutane) showed a similar pattern of neutral particles across the diameter range studied and experiments with larger alkylamine, butylamine, are underway to probe the relationship between structure- and NPF potential from MSA. These findings highlight that there is no “one size-fits-all” regarding NPF from MSA reactions with amines and illustrates the need for studies of more complex amines to fully characterize the NPF potential of this atmospherically relevant strong acid. (4) Novel Aspect. The combination of TDCIMS with a novel particle sizing system provided the chemical composition of 2-10 nm particles. 

The energy landscape is changing worldwide, with a drastic reduction in sulfur dioxide (precursor to sulfuric acid, H2SO4) emitted from fossil fuel combustion. As a result, acid-base chemistry leading to new particle formation (NPF) from sulfuric acid is decreasing. At the same time, photooxidation of biogenic organosulfur compounds leading to the formation of H2SO4 and methanesulfonic acid (MSA) is expected to become more important. Aqueous solutions of alkanolamines have been proposed as carbon capture technology media to store carbon dioxide from stack plumes before release into the atmosphere. It is therefore expected that some of the alkanolamines will be released, making it critical to understand their atmospheric fates including their role in new particle formation and growth. We expanded our experimental studies of nucleation from the reaction of MSA with simple amines to the multifunctional alkanolamines, including mononethanolamine (HO(CH2)2NH2; MEA) and 4-aminobutanol (HO(CH2)4NH2; 4AB). Experiments were performed in a 1-m borosilicate flow reactor under dry conditions as well as in presence of water. These two systems were shown to produce sub-10 nm particles with MSA extremely efficiently. Surprisingly, the presence of water did not enhance NPF, in contrast to the drastic effect water had on small alkylamine reactions with MSA. This is likely due to the fact that MEA and 4AB have an -OH group that provides additional H-bond interactions within the cluster. Sampling of the chemical composition of these small nanoparticles with high resolution and high transmission was possible down to 3-4 nm using a novel high-flow differential mobility analyzer (half-mini DMA) interfaced to a thermal chemical ionization mass spectrometer (TDCIMS). There was no size dependence for the acid-to-base molar ratio (1:1) for either amine. Integration of these data with preliminary results obtained for a simple C4 alkylamine (butylamine) and a C4 diamine (putrescine) will be discussed in the context of developing a molecular structure-reactivity scheme for new particle formation from MSA and amines of varying structures. 

This study reports on the high yield of new particle formation (NPF) from the reaction of an alkanolamine commonly used in carbon capture and storage technology, monoethanolamine, with strong atmospherically relevant acid, methanesulfonic acid.

 

In this paper, we focus on the important yet understudied problem of Continual Federated Learning (CFL), where a server communicates with a set of clients to incrementally learn new concepts over time without sharing or storing any data. The complexity of this problem is compounded by challenges from both the Continual and Federated Learning perspectives. Specifically, models trained in a CFL setup suffer from catastrophic forgetting which is exacerbated by data heterogeneity across clients. Existing attempts at this problem tend to impose large overheads on clients and communication channels or require access to stored data which renders them unsuitable for real-world use due to privacy. We study this problem in the context of Foundation Models and showcase their effectiveness in mitigating forgetting while minimizing overhead costs and without requiring access to any stored data. We achieve this by leveraging a prompting based approach (such that only prompts and classifier heads have to be communicated) and proposing a novel and lightweight generation and distillation scheme to aggregate client models at the server. We formulate this problem for image classification and establish strong baselines for comparison, conduct experiments on CIFAR-100 as well as challenging, large-scale datasets like ImageNet-R and DomainNet. Our approach outperforms both existing methods and our own baselines by more than 7% while significantly reducing communication and client-level computation costs. 

The influence of additives on the detonation velocity of a polyethylene wax/RDX formulation was examined. Additives included species of various shock impedance: glass microballoons; glass microspheres; polymethyl methacrylate (PMMA) microspheres; thermally expandable microspheres (TEMs); and PMMA microencapsulated pentaerythritol tetranitrate (PETN). Performance of the insensitive explosive 2,4-dinitroanisole (DNAN) was enhanced by addition of PETN-either neat or encapsulated, but encapsulation did not increase the sensitivity of the formulation. The energy contribution of the encapsulated PETN to the detonation front of the insensitive explosive 2,4-dinitroanisole (DNAN) was also demonstrated. Present in 5 wt%, the encapsulated PETN allowed DNAN to sustain a reaction (5.36 km/s) at 13 mm, well below its critical diameter. 

Continual learning is a setting where machine learning models learn novel concepts from continuously shifting training data, while simultaneously avoiding degradation of knowledge on previously seen classes which may disappear from the training data for extended periods of time (a phenomenon known as the catastrophic forgetting problem). Current approaches for continual learning of a single expanding task (aka class-incremental continual learning) require extensive rehearsal of previously seen data to avoid this degradation of knowledge. Unfortunately, rehearsal comes at a cost to memory, and it may also violate data-privacy. Instead, we explore combining knowledge distillation and parameter regularization in new ways to achieve strong continual learning performance without rehearsal. Specifically, we take a deep dive into common continual learning techniques: prediction distillation, feature distillation, L2 parameter regularization, and EWC parameter regularization. We first disprove the common assumption that parameter regularization techniques fail for rehearsal-free continual learning of a single, expanding task. Next, we explore how to leverage knowledge from a pre-trained model in rehearsal-free continual learning and find that vanilla L2 parameter regularization outperforms EWC parameter regularization and feature distillation. Finally, we explore the recently popular ImageNet-R benchmark, and show that L2 parameter regularization implemented in self-attention blocks of a ViT transformer outperforms recent popular prompting for continual learning methods.