Producing high quality thin films of controlled thickness is a critical step for the development of ferroelectric nanophotonic devices. Developed recently, the process referred to as layer transfer has been shown to be very promising: ions are implanted in a plan parallel to the interface of a bilayer system that is then heated. The high temperature induces nucleation and propagation of cracks in the weakened plan of the specimen, resulting after coalescence in a full splitting of the upper part of the original sample (Fig. (a)). However, this process is often accompanied by uncontrolled film failure. Understanding these failure mechanisms and defining the limitations of the layer transfer process are the central points of this study.
The post-mortem analysis of cracked samples suggests two failure mechanisms that are then theoretically investigated: for some systems, the cracks confined initially in the weakened plan of the sample deviate from their trajectory, producing transverse cracks in the film. In other cases, layer transfer is accompanied by film buckling, resulting ultimately in its failure by bending (Fig. (b)).
Crack stability analysis shows that the first process can be avoided if the film is under compression during the process. The second failure mechanism is shown to occur above a critical compressive stress in the film. Thus, layer transfer can only be achieved in a narrow window of compressive stresses. This implies that only systems with positive thermal expansion coefficients mismatch between film and substrate can be produced by layer transfer and the temperature increase has to be less than a critical amount.
By providing a quantitative understanding of these failure mechanisms, this study has defined the conditions for which a film can be produced by layer transfer.
Researchers: L. Ponson and K. Diest