salycilic acid

[vicentexxx]

hi everyone
i have this problem:

i have to analyse a tincture with salycilic acid and a NSAID.
column is an Alltima C18 250 *4.6
MP CH3CN/H2O 70/30
flow 0.6 ml/min

NSAID sample is prepared simply diluting tincture while salycilic acid sample is prepared diluting NSAID sample and adding 1 ml of H3PO4 10%.
the two analysis are consecutive but NSAID assay is always OK while salycilic acid assay is very unpredictable. It seems that autosampler can't inject the same amount from each vial, i say that becouse also the NSAID has a poor recovery and his behaviour is similar to an internal standard respect to salycilic acid. But it's impossible becouse NSAID analysis is OK so autosampler works.
Usually the second or third analysis seem to be OK while the first trial is a disaster...
What the problem?is the critical step tha addition of H3PO4? if it's so..where did salycilic go?an d why the second and third trial go better..

Help me to solve this problem...
thank you
vincenzo

[Mark Tracy]

I'm going to take a guess and suggest conditioning the column with a few injections of H3PO4 blank. Salicylic acid is a strong chelating agent, and any trace of metal ions in the system will cause trouble, especially if you are using an unbuffered mobile phase. I think you might get better results by adding 0.05% H3PO4 to the mobile phase, but that may not be permissible in your lab.

[Consumer Products Guy]

I ALWAYS add acid to my aqueous phase as well when assaying salicylic acid. So does USP. I've used phosphoric acid and I've used acetic acid, both work fine. So I think Mark is correct.