Source-linked AI summary
Critical Influences of Particle Size and Adhesion on the Powder Layer Uniformity in Metal Additive Manufacturing
Christoph Meier, Reimbar Weissbach, Johannes Weinberg, Wolfgang A. Wall, A. John Hart
TL;DR
Powder-layer packing and surface uniformity are important to powder-bed metal AM, but the influence of powder cohesiveness during recoating requires quantitative study. The paper uses a cohesive DEM model and layer-quality metrics to examine Ti-6Al-4V powders and recoating parameters, finding poorer layers for finer, more cohesive powders and identifying parameter-specific mitigation measures.
Problem
The study asks how powder cohesiveness and recoating conditions affect packing density and surface uniformity in powder-bed metal additive manufacturing.
Method
A cohesive discrete element model simulates Ti-6Al-4V powder recoating with particle and wall interactions, evaluated using packing fraction and surface profile statistics.
Results
Decreased particle size or increased cohesiveness considerably reduces powder-layer quality, producing low, strongly varying packing fractions and highly non-uniform surface profiles.
Takeaways & Limitations
Layer quality can be improved for very fine cohesive powders by reducing particle-to-blade adhesion, using suitable layer thicknesses, and adjusting recoating conditions.
Takeaways & Limitations
The simulations assume identical particle-to-particle and particle-to-wall interaction parameters, although this is unlikely in practical metal AM processes.
Abstract
from arXiv · showhide
The quality of powder layers, specifically their packing density and surface uniformity, is a critical factor influencing the quality of components produced by powder bed metal additive manufacturing (AM) processes, including selective laser melting, electron beam melting and binder jetting. The present work employs a computational model to study the critical influence of powder cohesiveness on the powder recoating process in AM. The model is based on the discrete element method (DEM) with particle-to-particle and particle-to-wall interactions involving frictional contact, rolling resistance and cohesive forces. Quantitative metrics, namely the spatial mean values and standard deviations of the packing fraction and surface profile field, are defined in order to evaluate powder layer quality. Based on these metrics, the size-dependent behavior of exemplary plasma-atomized Ti-6Al-4V powders during the recoating process is studied. It is found that decreased particle size / increased cohesiveness leads to considerably decreased powder layer quality in terms of low, strongly varying packing fractions and highly non-uniform surface profiles. For relatively fine-grained powders (mean particle diameter $17 μm$), it is shown that cohesive forces dominate gravity forces by two orders of magnitude leading to low quality powder layers not suitable for subsequent laser melting without additional layer / surface finishing steps. Besides particle-to-particle adhesion, this contribution quantifies the influence of mechanical bulk powder material parameters, nominal layer thickness, blade velocity as well as particle-to-wall adhesion. Finally, the implications of the resulting powder layer characteristics on the subsequent melting process are discussed and practical recommendations are given for the choice of powder recoating process parameters.
1. Introduction
Metal additive manufacturing quality depends on complex, competing physical mechanisms, including powder feedstock and its resulting layer characteristics. This work addresses cohesive powder recoating using metrics for packing fraction and surface uniformity.
- Motivation: Selective laser melting offers near-net-shape production of near-limitless geometries but involves competing physical mechanisms.Sub-optimal process parameters can deteriorate material properties or cause part failure during manufacturing.
- Related work: Existing AM experiments and models are commonly classified as macroscopic, mesoscopic, or microscopic by considered length scale.These categories address phenomena ranging from part-level physical fields to smaller-scale process behavior.
- Related work: Powder feedstock properties influence particle behavior, bulk flowability, packing density, surface uniformity, and effective thermal and mechanical properties.The feedstock is characterized by particle-surface properties, morphology, granulometry, and resulting bulk behavior.
- Research gap: Earlier DEM recoating models included elastic contact, sliding friction, and rolling friction, but generally omitted cohesive effects within the powder.This omission is notable because cohesiveness increases as particle size decreases.
- Contribution: Surface energy can vary by orders of magnitude with roughness and contamination or oxidation, motivating experimental calibration of cohesive powder models.The study defines spatial packing fraction and surface profile metrics, using their means and standard deviations to analyze powder-layer characteristics.
2. Cohesive powder model and choice of parameters
The study models spherical plasma-atomized Ti-6Al-4V particles with a cohesive DEM formulation and calibrated interaction parameters. The formulation includes particle dynamics, contact and adhesive interactions, fitted particle-size distributions, and recoating simulation settings.
- Model formulation: The cohesive powder model uses a soft-sphere discrete element method to represent individual, approximately spherical Ti-6Al-4V particles in a Lagrangian framework.The model is applied to plasma-atomized powders with different size distributions.
- Model formulation: Particle equations include gravity, normal and tangential contact forces, adhesive forces, rolling-resistance torques, and tangential-force torques.The force and torque contributions arise from interactions with neighboring particles.
- Particle-size parameters: The reference powder distribution uses D10 = 20µm, D50 = 34µm, and D90 = 44µm, with particles restricted to the D10–D90 range.The fitted log-normal distribution represents a medium-sized AM powder with mean particle diameter 34µm.
- Interaction parameters: The calibrated surface energy is γ0 = 0.1mJ/m2, and simulations vary adhesion through γ = γ0, 4γ0, 0.25γ0, and γ = 0.The calibration fitted experimental and numerical angle-of-repose values for the employed Ti-6Al-4V powder.
3. Recoating simulations
The simulations quantify how cohesiveness, layer thickness, and numerical choices affect powder-layer packing and surface uniformity. Greater cohesiveness generally worsens layer quality, while thicker layers mitigate packing and surface variations.
- Cohesiveness: Adhesion produces less uniform powder layers, with increased spatial variation in packing fraction and surface profile height.For γ = γ0 versus γ = 0, cohesive powder shows decreased layer quality in both spatial fields.
- Sensitivity: Ten powder-size-distribution realizations produce small deviations between curves, indicating reasonably low sensitivity to stochastic sampling and selected mechanical parameters.The uncertainty in these choices does not noticeably affect the general statements, although highest-cohesiveness deviations remain within stochastic uncertainty.
- Cohesiveness: Mean packing fraction decreases from almost 60% to below 40% as surface energy increases from γ = 0 to γ = 4γ0.Over the same range, the mean layer height decreases from 90% to approximately 70% of nominal thickness t0.
- Cohesiveness: 2.5% to 7.5%: the packing-fraction standard deviation increases as surface energy rises from γ = 0 to γ = 4γ0.The supplied passage also reports increasing surface-profile variation with increasing cohesiveness.
- Layer thickness: Thicker powder layers increase packing density and reduce packing-fraction variations because particles can arrange within a larger layer volume.Surface effects also have a decreasing influence as nominal layer thickness increases.
- Cohesiveness: Surface roughness increases with cohesiveness because larger agglomerates and particle-to-blade adhesion create surface asperities.Particle-to-blade adhesion can rip particles from the near-surface layer compound, contributing to roughness.
4. Discussion of results
The simulations show that powder cohesion, particle–substrate interaction, layer thickness, and blade velocity strongly affect recoated-layer quality and melting suitability. Practical improvements include reducing blade adhesion, increasing layer thickness, and compensating high-velocity post-flow.
- Powder cohesion: For 34 µm powder, adhesion substantially changes powder-layer metrics, making cohesive-force modeling essential for realistic recoating simulations.Layer metrics are less sensitive to friction, restitution, and stiffness variations than to surface-energy variations.
- Powder cohesion: Halving the mean particle diameter or quadrupling surface energy drastically reduces powder-layer quality relative to calibrated 34 µm powder.Lower packing fraction and surface height can increase energy inefficiency, thermal gradients, residual stresses, shrinkage, and local overheating.
- Melting implications: Fine cohesive powders spread with a simple blade are unsuitable for subsequent laser melting without additional powder compaction.Large thickness variations and asperities demand high energy density while exposing isolated particles to evaporation.
- Powder-to-substrate interaction: A fourfold reduction in powder-to-substrate adhesion can produce discontinuous layers with blank spots, partly compensated by increased particle-to-substrate friction.For medium-sized and coarse powders, substrate friction is expected to matter more than substrate adhesion.
- Recoating-blade interaction: Reducing powder–blade adhesion can improve very fine cohesive powders, especially when combined with subsequent powder compaction.The benefit is limited for coarse, less-adhesive powders.
- Layer thickness: Nominal thickness t0 = dmax,0 produces sparse discontinuous layers, whereas medium and coarse powders generally require about 2–3dmax,0 for near-optimal quality.Fine powders require a recommended minimum of roughly 3–4dmax,0.
- Blade velocity: Higher blade velocity causes post-flow and lower mean layer thickness; increased inter-particle friction or larger blade gaps can partly compensate.Less spherical particles reduce flowability but typically also reduce achievable packing fractions, creating a process trade-off.
5. Conclusion
The DEM study shows that cohesion, particle–wall interactions, and recoating parameters strongly affect powder-layer quality. Practical improvements include reducing blade adhesion, increasing nominal layer thickness, and compensating for velocity-driven post-flow.
- Increased cohesiveness or decreased particle size produces low, strongly varying packing fractions and rough, non-uniform powder layers.Packing fractions can fall below those of unspread bulk powder with comparable cohesiveness.
- For 17 μm powder, cohesive forces dominate gravity by almost two orders of magnitude, producing layers unsuitable for laser melting without additional finishing.
- Reduced adhesive interaction between the recoating blade and powder improves the surface uniformity of cohesive layers.
- Powder-layer quality increases with nominal layer thickness; two to three times the maximal particle diameter is recommended.
- Higher blade velocities cause dynamic powder post-flow and thinner mean layers, which can be offset by lower flowability or greater nominal thickness.
- The study derives practical implications for subsequent melting and identifies experimental verification and alternative recoating strategies as future work.