The effective dispersion of titanium dioxide (TiO2) in waterborne coatings is essential for optimizing optical properties such as opacity. Traditional dispersing agents often fail to fully stabilize inorganic pigment particles and may increase the water sensitivity of the final film due to their hydrophilic nature. This study explores the use of functionalized polymer particles as an alternative to conventional dispersants. Polymer particles of varying size and functionalities were evaluated to assess their influence on TiO2 disaggregation and dispersion. This effect was directly analyzed in a conventional waterborne coating formulation. The polymer particles used as dispersants also form the polymer matrix during film formation. Results indicate that a functionalization of the particles is needed to enhance pigment dispersion when no conventional pigment dispersant is used. Interestingly, the obtained optical performance of the coating depends on the binder composition. Regardless of the functionality of the polymer particles used, the resulting films exhibited reduced water sensitivity.
Large Language Models (LLMs) were applied to automate the interpretation of electrochemical impedance spectroscopy (EIS) data, enabling classification and parameter estimation without the need for a task specific machine learning training. The approach achieved classification accuracies up to 96% and produced fitting results comparable to those obtained with specifically trained neural networks. The methodology reduces reliance on labelled data and manual intervention. While demonstrated in the context of organic coatings, the framework provides a scalable AI-based workflow that could, in principle, be extended to conceptually similar tasks in materials and corrosion research, subject to dedicated validation.
Atomistic models of crosslinked polymers provide invaluable insights into the nanoscale properties otherwise inaccessible to experiments. These amorphous, highly interconnected structures can only be produced in simulations using artificially accelerated crosslinking reactions, completing within the nanosecond timescale available. Atomistic reaction modeling techniques density functional theory (DFT) and ReaxFF are often too expensive for the large systems required for polymer modeling, and cure reactions too slow for simulation timescales. Heuristics have been designed to identify reactive functionalities and instantaneously perform crosslinking reactions during classical molecular dynamics simulations. We compare the networks produced by two major approaches used in literature for a model epoxy-amine polymer of diglycidyl ether of bisphenol A (DGEBA) and m-xylylenediamine (MXDA). In one method, the reactants are modified into 'activated' intermediate forms, while the other performs all bond formations and cleavages of a reaction simultaneously. We find minor differences in kinetics during the cure, but fully crosslinked bulk polymers are indistinguishable. However, the presence of a solid surface during the cure can result in significant differences in interphase structures due to the activation of reactants. We conclude that while both methods may be used equally in producing bulk networks, activated reactants may not produce accurate structures at interfaces.
Active protective coatings have been used and studied for decades. It is well known that the coating composition and microstructure determine the leaching behavior of the corrosion inhibiting species from the coating matrix. However, the leaching process on the microscale is a complex phenomenon important details of which remain obscured till today. Non-destructive spatial observation of the leaching process by nano-computed tomography using synchrotron radiation can contribute to a deeper understanding. Here, we report on the first truly in-situ 3D observation of microscale leaching. 3D images were generated while individual inhibitor particles dissolve from the coating matrix due to exposure to flowing water. The development and growth of pores and pore clusters was observed with a sequence of 3D images as a function of time demonstrating that leaching progresses by successive dissolution of inter-connected soluble particles.
A Finite Element Model (FEM) based approach describing particle leaching from an organic coating is presented. Since the dissolution of pigments loaded in a primer coating severely enhances the formation of water channels and consequently the leaching effect, an empirical formula describing this percolation effect is implemented in a pre-existing 2D FEM that solves for the diffusion and migration transport of all relevant ions in the primer and adjacent water phases. This FEM also accounts for homogeneous reactions of the species as occurring in the electrolyte and the primer. The model is applied to an aluminum alloy (AA2024-T3) coated with a primer loaded with lithium carbonate pigment, topped with a topcoat, playing the role of an insulating protective layer. The oxidation and reduction reactions at the AA2024-T3 surface leading to the corrosive effect are implemented as boundary conditions with kinetic parameters obtained from PotentioDynamic Polarization (PDP) measurements. Parameters of the empirical percolation formula were retro-fitted through comparison between simulated results and leaching measurements from a primer-topcoat coating system with three different initial pigment volume concentrations (PVC) of lithium carbonate. Model results include the 2D concentration profiles of all the model species for different PVC values.
