The crystallization of vacuum-evaporated Se can be greatly affected by impurities present either in the Se source or on the substrate surface [1-3]. The presence of oxygen (in the form of SeO2) in the Se source was found to enhance Se crystallization [1], due to formation of new nucleation sites on the substrate surface. In contrast, the addition of C1, As, and Te alloys in the Se source was observed to inhibit Se crystallization, due to deactivation of chain growth on the crystal surface [2, 3]. Crystallites were also observed in amorphous Se films after ageing in air, even at room temperature and without illumination (the activation energy for the crystallization of Se is about 1 eV) [4-7]. Along with reorganization of the Se molecules [6, 7], the adsorption of oxygen and water impurities on the Se surface is believed to play an important role. We have reported [8] that the crystallization of vacuum-evaporated high purity Se on glass, ITO (indium tin oxide), and PC (polycarbonate) substrates can be greatly enhanced and the structure of the crystallites modified if the substrates have been soaked in water prior to Se deposition. Crystallization was further increased by heating the Se films in air, the rate of crystal growth being faster than that in the vacuum chamber. These observations lead to the conclusion that water impurities, initially absorbed in the substrate during soaking, can 'enhance crystallization of Se by forming additional nucleation sites on the substrate surface. To further investigate this effect, and in particular, to effectively eliminate water molecules from the substrates, we examine the influence of several parameters affecting water content in the substrates on the crystallization of Se. These parameters include water soaking temperature and heating and pumping substrates in the vacuum chamber. Glas, ITO and PC substrates [8] were cleaned using reagent grade isopropanol and dried with a jet of dry pre-purified nitrogen. Some of the substrates were soaked in de-ionized water for 1 min and again