For several months now, I have been experiencing difficulty in measuring low levels of Se in the ICP-OES 5800. In my calibration, my blank, 4 ug/L standard, and 10 ug/L standard all tend to be anywhere from 45% to 100+% RSD, but it isn't consistent. Sometimes they return less than 10%, other times more than 100%, even if the same preparation is run twice. I've ruled out incorrect prep, as other analytes in those standards come back normally, and I've ruled out contamination, as running those samples through an ICP-MS results in consistent data. I sometimes get issues with As, Ni, and Sb as well, but nowhere near the degree or frequency of issues I have with Se, and there's no correlation between when I have issues with any of these analytes.
Blanks, especially, cause problems. I can get a replicate measuring as high as 5 ug/L while other replicates in the same set can read 0 or negative.
In higher concentrations, such as the upper portion of the curve and ICV solutions, Se comes back consistently with <1% RSD. Thus, I'm starting to think there's interference somewhere, but I can't be sure.
So, my question is: What are common interferents to Selenium, and are there measures I could take to mitigate high RSD values?