Failure mechanisms of anodized aluminum parts used in chemical vapor deposition chambers

Abstract

Anodized aluminum parts used in chemical vapor deposition (CVD) reactor chambers for TEOS (tetra ethyl ortho silicate), passivation, and tungsten process clean chemistries were analyzed using scanning electron microscopy and energy-dispersive spectroscopy x-ray maps. Examination reveals that the primary cause of failure for all anodized Al parts is the same while individual mechanisms vary depending on the process chemistry used. The fundamental reason for failure of anodized aluminum part in CVD reactors is the formation and growth of aluminum fluoride below the anodized film at point defect sites. These point defect sites are imperfections in the anodized layer caused by precipitate particles such as Fe and Si, voids, sharp corners, and cracks due to differential expansion. In the 6061 alloy, the point defect sites act as high energy sites for the plasma fluorine attack. However, such defect sites can be capped off by the formation of a layer of MgF2 below the anodized film. The formation of MgF2 is dependent upon the availability of elemental Mg to diffuse to these point defect sites and also the number of defect sites. Some of the Mg is typically tied up with precipitates like Fe and Si. When free Mg is unavailable to form a barrier layer, subsequent fluorination of the substrate causes formation of AlF3. This results in stresses in the anodized layer that can ultimately lead to delamination of the anodized layer and failure of the part. The 1100 alloy does not contain Mg and thus fluorination results in the direct formation of AlF3 at point defect sites with no protective MgF2 layer present. An enhanced performance of anodized Al parts can be achieved by using pure aluminum alloys with controlled quantities of Mg to form the MgF2 barrier layer.