non-repeateble signals GC/MS

[Krashan]

Dear all,
I use Thermo GC/MS system included DSQ MS with single quadrupole. I work in NCI mode. I use column 15 m Agilent DB-5 ms connected to 1.5 m precolumn with the same phase. Injections are in PTV injector in large volume mode with increasing temperature from 50 to 325 degrees. Column is heating from 80 to 320 degrees. I use liners with glass wool. This configuration is used for 2 years without any changing for pesticides determination.
Lately I noted that I have non-repeatable peaks areas for some compounds. It happens that I make 2 injections of standards mixtures and for the second one I observe much lower signals (300-500% difference). It refers to parameters as follows: chlorthalonil, cypermethrin, captan. In addition for some pesticides (e.g. iprodione, fenhexamid) I cannot observe any peaks although previously I had quite good signals. It is interesting that for these anaytes I have very good signals for prepared samples whilst for standards they are poor. I can observe very good resolution for the rest of pesticides without any drifting of retention times.
MS seems not to be origin of my problem. All the 3 modes (EI, NCI, CI) give good spectra during the tuning without any discrimination. I suspected problems with column but after changing for the new one the problem is still the same. Cleaning the injector with isopropanol did not help too.
Have you any ideas what could be the reason of such situation and how to ressolve this problem?

[bhuvfe]

Just few thoughts..

Have you tried a new liner?

Are the standards freshly prepared (or purchased)?

[Don_Hilton]

Have you changed supplers or type of inlet liner? And are your standards made up in matrix. The good injecton flowed by the bad of standard but typically good inejctions of samles (which would have matrix) leaves me wondering if the masking of active sites by matrix is being lost on the injecion of standards with little or no matrix.

[Krashan]

Your advice seems to be good, Don. I injected few times standard mix with addition matrix (approx. 15%). I have good repetition now.
Although it is strange that for standards in solvent chromatography is bad. It should not be that.

[Don_Hilton]

What you have encountered is not uncommon when analyzing compunds like pesticides. Because standards wihtout matrix do not have polar gunk to protect the analytes, calibrations are often made up in matrix. Without the matrix present in standards, recoveries show to be too high - as spikes in the matrix come out without significant degredation while the standards actually suffer degredation.

[Krashan]

Thank you Don, some recoveries were in fact too high. Now they are much real.
Best regards
Robert