Oxygen Contamination

I have a new NETZSCH DSC 404 F3 Pegasus. I ran a pure zirconium sample but ran into oxygen contamination problem. The zirconium sample after a DSC run up to 1100 C showed tarnish. We have done everything possible to mitigate o-contamination. The sample shows less tarnish now, but the endothermic peak due to HCP-to-BCC allotropic transformation is not symmetric, but on the high-temperature side of the thermal peak there seems to be an overlap with a small thermal peak. Any comments? Also, how does one minimize oxygen contamination when working with a very reactive material like zirconium. Thank you,...............GA