In-situ particle analysis with heterogeneous background: a machine learning approach.

We propose a novel framework that combines state-of-the-art deep learning approaches with pre- and post-processing algorithms for particle detection in complex/heterogeneous backgrounds common in the manufacturing domain. Traditional methods, like size analyzers and those based on dilution, image processing, or deep learning, typically excel with homogeneous backgrounds. Yet, they often fall short in accurately detecting particles against the intricate and varied backgrounds characteristic of heterogeneous particle-substrate (HPS) interfaces in manufacturing. To address this, we've developed a flexible framework designed to detect particles in diverse environments and input types. Our modular framework hinges on model selection and AI-guided particle detection as its core, with preprocessing and postprocessing as integral components, creating a four-step process. This system is versatile, allowing for various preprocessing, AI model selections, and post-processing strategies. We demonstrate this with an entrainment-based particle delivery method, transferring various particles onto substrates that mimic the HPS interface. By altering particle and substrate properties (e.g., material type, size, roughness, shape) and process parameters (e.g., capillary number) during particle entrainment, we capture images under different ambient lighting conditions, introducing a range of HPS background complexities. In the preprocessing phase, we apply image enhancement and sharpening techniques to improve detection accuracy. Specifically, image enhancement adjusts the dynamic range and histogram, while sharpening increases contrast by combining the high pass filter output with the base image. We introduce an image classifier model (based on the type of heterogeneity), employing Transfer Learning with MobileNet as a Model Selector, to identify the most appropriate AI model (i.e., YOLO model) for analyzing each specific image, thereby enhancing detection accuracy across particle-substrate variations. Following image classification based on heterogeneity, the relevant YOLO model is employed for particle identification, with a distinct YOLO model generated for each heterogeneity type, improving overall classification performance. In the post-processing phase, domain knowledge is used to minimize false positives. Our analysis indicates that the AI-guided framework maintains consistent precision and recall across various HPS conditions, with the harmonic mean of these metrics comparable to those of individual AI model outcomes. This tool shows potential for advancing in-situ process monitoring across multiple manufacturing operations, including high-density powder-based 3D printing, powder metallurgy, extreme environment coatings, particle categorization, and semiconductor manufacturing.

Mechanical Performance of Three-Dimensional Printed Lattice Structures: Assembled Versus Direct Print.

Lattice structures are composed of a collection of struts with different orientations. During slicing, the inclined struts generate multiple disjoint contours along the build direction in additive manufacturing (AM). These contours are substantially smaller in size due to the narrow cross-section of the individual lattice struts, and they can lead to contour plurality in AM processes. Contour plurality reduces the amount of continuous contact region between two successive layers, thus resulting in poor interlayer adhesion, structural integrity, and mechanical properties of the printed lattice structure. A new interlocking and assemble-based lattice structure building approach is investigated by increasing continuity in layers and avoiding support structure to minimize contour plurality. Two lattice configurations in the form of cubic and octet lattice structures are examined. The compressive performance of the designed lattice structures is compared with the traditional single-build direct three-dimensional printed lattice structures. The mechanical performance (e.g., peak stress, specific energy absorption) of the assembled structures is found to be generally better than their direct print counterparts. The empirical constants of Ashby-Gibson power law are found to be larger than their suggested values in both direct print and assembly techniques. However, their values are more compliant for octet assembled structures, which are less susceptible to manufacturing imperfections.

Micro-particle entrainment from density mismatched liquid carrier system.

Micro-scale inorganic particles (d > 1 µm) have reduced surface area and higher density, making them negatively buoyant in most dip-coating mixtures. Their controlled delivery in hard-to-reach places through entrainment is possible but challenging due to the density mismatch between them and the liquid matrix called liquid carrier system (LCS). In this work, the particle transfer mechanism from the complex density mismatching mixture was investigated. The LCS solution was prepared and optimized using a polymer binder and an evaporating solvent. The inorganic particles were dispersed in the LCS by stirring at the just suspending speed to maintain the pseudo suspension characteristics for the heterogeneous mixture. The effect of solid loading and the binder volume fraction on solid transfer has been reported at room temperature. Two coating regimes are observed (i) heterogeneous coating where particle clusters are formed at a low capillary number and (ii) effective viscous regime, where full coverage can be observed on the substrate. 'Zero' particle entrainment was not observed even at a low capillary number of the mixture, which can be attributed to the presence of the binder and hydrodynamic flow of the particles due to the stirring of the mixture. The critical film thickness for particle entrainment is [Formula: see text] for 6.5% binder and [Formula: see text] for 10.5% binder, which are smaller than previously reported in literature. Furthermore, the transferred particle matrices closely follow the analytical expression (modified LLD) of density matching suspension which demonstrate that the density mismatch effect can be neutralized with the stirring energy. The findings of this research will help to understand this high-volume solid transfer technique and develop novel manufacturing processes.

Rheological Analysis of Bio-ink for 3D Bio-printing Processes.

3D bio-printing is an emerging technology to fabricate tissue scaffold in-vitro through the controlled allocation of biomaterial and cells, which can mimic the in-vivo counterpart of living tissue. Live cells are often encapsulated into the biomaterials (i.e., bio-ink) and extruded by controlling the printing parameters. The functionality of the bioink depends upon three factors: (a) printability, (b) shape fidelity, and (c) bio-compatibility. Increasing viscosity will improve the printability and the shape fidelity but require higher applied extrusion pressure, which is detrimental to the living cell dwelling in the bio-ink, which is often ignored in the bio-ink optimization process. This paper demonstrates a roadmap to develop and optimize bio-inks, ensuring printability, shape fidelity, and cell survivability. The pressure exerted on the bio-ink during extrusion processes is measured analytically, and the information is incorporated in the bio-ink's rheology design. Cell-laden filaments are fabricated with multiple cell lines, i.e., Human Embryonic Kidney (HEK 293), BxPC3, and prostate cancer cells which are analyzed for cell viability. The cross-sectional live-dead assay of the extruded filament demonstrates a spatial pattern for HEK 293 cell viability, which correlates with our analytical finding of the shear stress at the nozzle tip. All three cell lines were able to sustain a transient shear stress of 3.7 kPa and demonstrate 90% viability with our designed bio-ink after 15 days of incubation. Simultaneously, the shape fidelity and printability matrices show its suitability for 3D bio-printing process.

3D metal lattice structure manufacturing with continuous rods.

In this paper, a new possibility of fabricating a metal lattice structure with a continuous rod is demonstrated. A multi-layer, periodic, and aperiodic lattice structure can be manufactured with a continuous thin rod by bending it with a repetitive pattern. However, joining their nodes are challenging and an important problem to solve. This paper is investigating the joining of nodes in a loose lattice structure by delivering materials through the dipping process. Both liquid state (epoxy) and solid-state (inorganic particles) joining agents are considered for polymer-metal and metal-metal bonding, respectively. Liquid Carrier Systems (LCS) are designed considering their rheological behavior. We found 40% solid loading with the liquid carrier system provides sufficient solid particles transfer at dipping and join the lattice node using transient liquid phase bonding (TLP). 3D metal lattice structures are constructed, and their mechanical properties are investigated. The lattice structure shows comparable strength even with smaller relative density (< 10%). The strength and elastic modulus of all the fabricated samples decreases with the increase in cell size, which is consistent with the traditional wisdom.

A Novel Rapid Manufacturing Process for Metal Lattice Structure.

A novel lattice structure manufacturing process is proposed in this article, which has the potential to overcome the manufacturing shortcomings of small-scale metal lattice structure. The proposed hierarchical process has four segments: Design, Bending, Dip, and Join (DBDJ). The proposed research use one-dimensional metallic wires/rods instead of powder, two-dimensional sheet, or liquid metal, which is highly transformative than the status quo. The topology-based design technique is focused to construct the lattice structure using a continuous thin rod. The layers are stacked in an additive manner to construct the three-dimensional lattice structure. The dip-coating meditate material transfer facilitates the node joining using transient liquid phase diffusion bonding, and hence, the manufacturing of the complex lattice structure is performed. The research framework provides a unique and holistic approach from design to manufacturing for realizing small-scale metallic lattice structure. A range of multiscale lattice structure is manufactured with the proposed DBDJ process. Very low relative density (∼3.8%) unit cell is achieved, and compressive tests demonstrate no failure at the joining node, which is reported in this article.

3D Printability of Alginate-Carboxymethyl Cellulose Hydrogel.

Three-dimensional (3D) bio-printing is a revolutionary technology to reproduce a 3D functional living tissue scaffold in-vitro through controlled layer-by-layer deposition of biomaterials along with high precision positioning of cells. Due to its bio-compatibility, natural hydrogels are commonly considered as the scaffold material. However, the mechanical integrity of a hydrogel material, especially in 3D scaffold architecture, is an issue. In this research, a novel hybrid hydrogel, that is, sodium alginate with carboxymethyl cellulose (CMC) is developed and systematic quantitative characterization tests are conducted to validate its printability, shape fidelity and cell viability. The outcome of the rheological and mechanical test, filament collapse and fusion test demonstrate the favorable shape fidelity. Three-dimensional scaffold structures are fabricated with the pancreatic cancer cell, BxPC3 and the 86% cell viability is recorded after 23 days. This hybrid hydrogel can be a potential biomaterial in 3D bioprinting process and the outlined characterization techniques open an avenue directing reproducible printability and shape fidelity.

Direct Bio-printing with Heterogeneous Topology Design.

Bio-additive manufacturing is a promising tool to fabricate porous scaffold structures for expediting the tissue regeneration processes. Unlike the most traditional bulk material objects, the microstructures of tissue and organs are mostly highly anisotropic, heterogeneous, and porous in nature. However, modelling the internal heterogeneity of tissues/organs structures in the traditional CAD environment is difficult and oftentimes inaccurate. Besides, the de facto STL conversion of bio-models introduces loss of information and piles up more errors in each subsequent step (build orientation, slicing, tool-path planning) of the bio-printing process plan. We are proposing a topology based scaffold design methodology to accurately represent the heterogeneous internal architecture of tissues/organs. An image analysis technique is used that digitizes the topology information contained in medical images of tissues/organs. A weighted topology reconstruction algorithm is implemented to represent the heterogeneity with parametric functions. The parametric functions are then used to map the spatial material distribution. The generated information is directly transferred to the 3D bio-printer and heterogeneous porous tissue scaffold structure is manufactured without STL file. The proposed methodology is implemented to verify the effectiveness of the approach and the designed example structure is bio-fabricated with a deposition based bio-additive manufacturing system.