On the role of defect charge state in the stability of point defects in silicon