Thin sample experiments help predict moisture sorption in larger samples

Over time, moisture can change a material’s physical and chemical properties, altering its performance and the performance of other materials in close proximity. This change in performance is especially a concern when considering materials used in the food, packaging, medicine, electronics, and construction industries. Understanding the material moisture capacity and transport behavior of each material in a component is essential to assessing its long-term stability and service-life. Testing materials extensively is not always feasible or cost effective, so models are sometimes used to predict material performance. However, existing models are based on thin samples (one dimensional or 1D type) and were not specifically tested for scalability to the larger dimensions (bulk material or 3D type) used in many applications.

To bridge the gap between 1D and 3D scales, Lawrence Livermore researchers investigated the moisture sorption and diffusion behavior of materials over a range of sizes, humidities, and temperatures using a combined experimental and modeling approach. The team first conducted dynamic vapor sorption (DVS) isotherm experiments over a wide range of water activities to quantify the transient moisture–material interactions in quasi-1D samples on three different materials (filled polydimethylsiloxane, unfilled PDMS, and ceramic inorganic composite). Using parameters derived from the 1D experiments, they used a previously-developed dynamic triple-mode sorption model to simulate scaled-up samples and verified the simulations using step-experiments on 3D samples.

The results show that the full triple-mode sorption model is robust enough to predict the dynamic uptake and outgassing of 3D samples when using parameters derived from quasi-1D samples. However, optimization of the model was difficult, as small, subtle features overlooked in the 1D sample data had the potential to cause significant errors in the 3D samples. Therefore, multi-scale experiments are important when developing robust predictive capabilities for long-term bulk material behavior in real applications.

At LLNL, a validated model for moisture diffusion and sorption will be used to predict system-level moisture transport and concentrations in several weapon systems.  This validation step is a major programmatic milestone and gives the team confidence in the model and the capability to predict complex geometries with data from simple 1D samples.

[H.N. Sharma, Y. Sun, and E.A. Glascoe, Predicting 3D moisture sorption behavior of materials from 1D investigations, Sci Rep 10, 17852 (2020), doi: 10.1038/s41598-020-74898-w.]