Atomistic simulation of stress evolution during thin film growth.

摘要

The ability to understand and control stress development in thin films during growth is one of the most important keys in developing future generations of reliable electronic and opto-electronic devices. In this dissertation, I employed atomistic simulation and theoretical modeling techniques to investigate film stress evolution during the growth of single crystal and polycrystalline films. Since these stresses are often related to surface stress, I examined the effects of surface defects on the surface stress of Cu (001) and (111) surfaces. These simulations show that these surface defects have only a mild effect on surface stresses and thus cannot be responsible for the large compressive film stresses observed in the initial and late stages of polycrystalline film growth. A recently discovered surface defect, surface dislocations, is more stable than adatoms on Au (001). Surface dislocations make the surface stress of Au (001) more tensile and hence cannot explain the experimentally observed compressive stress during polycrystalline film growth. Interestingly, minimum energy path calculations shows that surface dislocation migration provides an efficient surface mass transport mechanism. The surface/interface effects at the early stages of polycrystalline and heteroepitaxial film growth were also studied using a hybrid molecular dynamics simulation. In the early stages of polycrystalline film growth, the film stress-thickness product is proportional to the substrate surface coverage with slope equal to minus the surface stress of the substrate. For early stage heteroepitaxial film growth, compressive film stress were observed even when the lattice misfit was tensile. These surprising results suggest that at early stages of film growth, the commonly employed wafer curvature method for measuring film stress is not reliable. Stress evolution during polycrystalline film growth was also studied using molecular dynamics. These simulations, combined with theoretical modeling, demonstrate that adatoms, incorporated at grain boundaries near the surface, is the major source of the compressive film stress observed during the growth of continuous polycrystalline films.