Identification of oxygen in the copper as the source of adhesion in Bridgman-grown CuInSe2 crystals

In synthesizing CuInSe2 from its starting elements for subsequent Bridgman crystal growth, a closed ampoule is used to prevent the loss of selenium. Frequently, the resulting ingot adheres strongly to the inner wall of the ampoule, unless a getter is included, such as graphite or boron nitride. The present work, shows that the ''ingot sticking'' problem is due to oxygen in the starting copper, even though this element is nominally of five-nines purity. Ingots of CuInSe2 prepared with oxygen-free copper, pumped-on-heated copper, or copper heated separately with boron nitride all show a reduction in adherence to the ampoule wall or exhibit no sticking at all. Gas analysis during the heat treatment has confirmed the presence of oxygen in the five-nines-pure copper and its reduction with continued pumping and heating.