Abstract

Flexible displays form complex structures joining together multiple layers of very various materials: metal or plastic flexible substrate, barrier and encapsulation layers, thin-films of transistor and optoelectronic devices. Residual stresses are inherent to these structures. In the case of inorganic layers deposited from a vapour phase onto a polymer substrate, these stresses comprise a so-called intrinsic contribution, which arises during the deposition, and thermal and hygroscopic contributions, which arise during temperature or relative humidity variations due to the different physical properties of materials. These residual stresses hinder to control the geometrical dimensions of the produced structures and are also at the origin of cracking and delaminating phenomena. Thus it is crucial to predict their development and to follow their evolution in order to optimize the production and the lifespan of these structures. For that purpose, our objective is to propose a numerical modelling framework of finite element type that makes possible to follow the geometrical evolution of multilayer composite structures with respect to the deposition conditions and the external conditions of temperature and relative humidity. In this study, we used bilayer samples formed with a nanometric silicon nitride layer, developed to protect organic light-emitting devices from ingression of oxygen and moisture, deposited on a polyimide foil. First experimental results obtained for the thermomechanical behaviours of the different materials, and for the diffusion and sorption properties for water of the polyimide substrate, are presented. Then these data are implemented in a non-linear numerical model coupling a quasi-static mechanical balance equation with some transient equations for the water concentration and for the temperature of the structure. Simulation results are finally compared with some experimental results, where the evolution of the curvature of bilayer samples subjected to thermal ramps, and/or to relative humidity jumps, was recorded. This study concludes that an accurate simulation needs to take into account the dependence of the water diffusion coefficient on the concentration, as well as on the transient evolutions of temperature and water concentration, whose time-scales are very different.

Details

Actions