Hello at all,
for the analysis of the micotoxyns fumonisin B1 (FB1) and fumonisin B2 (FB2) I have worked with an old column (Hewlett Packard Hypersil ODS 5 µm - 100 * 2,1 mm).
The eluents used are acetonitrile (A) and ammonium acetate 5 mM + 1,1 mL/L formic acid [pH=2.7÷2.8] (B) at 0.5 mL/min with the following gradient:
Time %B
0.00 35.0
10.00 60.0
12.00 90.0
14.00 90.0
16.00 35.0
Detection is performed with electrospray MS SIM in positive mode. The separation is good with FB1 at 4.8 min, FB2 at 8.4 min (FB3 is also separated in real samples at 7.3 min).
Problems arises for linearity and reproducibility that aren't good enough. During a construction of a calibration curve I observe a decrease in signal (10-20%) within 5-6 injections of standards.
In attempting to resolve this problem I have used eluent B without ammonium acetate. In fact a FIA experiment shows good results with acetonitrile:formic acid 0,1% 40:60 (elevated response for pseudomolecular ion) but, when passing to chromatography with Hypersil ODS and the same conditions listed above, the chromatogram shows no peaks!
I have also tried other, more recent type, columns with and without ammonium acetate with very poor results.
Someone have any explanation for this behavior?
Sorry for my english. Thanks in advance.

If we make the assumtion that the first few injections are OK, and things fall apart after that, then we need to determine why the loss of signal. My first question would be, is it the the LC side (pump/column/injector/sample) or the MS side. You did do the FIA experiment, but I dont belive this tells you the answer to the question. I think you need try your standards using some other detector (UV/FLD/ELSD) and see if the response holds steady using the exact columns/sample/MP/gradient. The results of this experiment will allow you to focus on the problem.

Droid

Thanks for your suggest,
unfortunately is very difficult for me to try other type of detector because I have only MSD for this LC; in adjunct the analytes (fumonisin B1 and B2)do not present any absortion in UV

Does the pump work reliably?

Thanks to Droid and Anonimous,
the HPLC pump work well and this fact is confirmed by the regularity of pressure profile during a chromatographic run.
In my laboratory an ELS detector is not available: to fix the problem clearly outlined by Droid I have tried to refine my FIA experiment.
I have injected in MS 10 ng of fumonisin B1 (and 5 of FB2) every 12 minutes at constant fragmentor voltage (110 V). Response results (area of ions 722, 723, 706, 707) are the following:
min area
0.16 331454
12.18 298230
24.19 478748
36.21 463294
48.22 434309
60.23 336249
First and second injections are perhaps affected by insufficient system conditioning. Others injections clearly evidentiate a tendency of signal decrease.
In definitive the response is not reproducible and my opinion, based on this data, is that the problem is on MS side.
Have you any ideas to solve this problem?
Thanks.

Based on this data from your latest experiment, how do you know if the injector is delivering the correct amount of sample to your system?

If you picked another compound and MP combination, would you get the same result?

If you really think the MS is at fault, you need to focus in on the things that would cause an ESI signal to vary, gas flows, heater temps, dirty esi probes, vacuum system issues, variable liquid flows, power supply issues causing variable voltages are a few things off the top of my head.