Hello all,

Was just curious if I might be able to receive some advice from someone who has encountered similar issues on this forum. We are working with anthracyclines on an HPLC-FLD, using a C-18 RP column. A formic acid buffer and acetonitrile are mixed in ratios of 90:10 and 10:90 for mobile phases A and B respectively, and a binary gradient is used.

This method has been previously validated by my supervisor (2 years ago approximately), but over the course of the last 5 months has lost repeatability and accuracy in calibration curves. For example, the same samples ran in two 7 point calibration curve from 1μg/L to 500μg/L consecutively from lowest to highest concentration and one solvent run between the highest of one curve and lowest of next had a relative standard deviation in accuracy levels of 16.4% (This is with respect to the first curve prepared, the second curve data is not used as calibration points but as a reference to accuracy of the two curves).

However, when precision assessments are carried out (same concentration sample, ran 10 times) the precision is assessed as fine. Moreover, if a calibration curve is built using varying injection volumes instead of concentration levels, it results in an acceptable curve. The detector has been checked by using a second FLD detector inline and comparing signal intensity/resolution, the needle has been replaced, column swapped for a different C-18 RP, and all tubing cleaned and assessed for leaks. I am not sure what next step to take in the troubleshooting process, so any ideas would help significantly.