Chromatography Forum

Hello,

We are analyzing VOC's in water using GC/MS. We have noticed a pattern in which the response of our internal standard fluorobenzene is increasing as the concentration of our calibration standards increases. We did notice that the response of the flourobenzene was constant between replicates of the same calibration standards so we assume that the correct amount of internal standard is being injected by the autosampler.

We realized this problem when trying to run quality control samples from an outside source and all the values were consistently higher than the actual known values. When we took out the internal standard and re-evaluated the same run, the quality control results then passed.

Does anybody know what could be causing the internal standard to increase. We did notice that between two different sets of quality control samples containing different compounds, that the flourobenzene response was consistently higher in one and lower in the other. This leads us to suspect there are interfering ions or we also read about active sites. Could either of these be our problem?

Instrument Conditions:
Column: RtxVMS
30m 0.25mm 1.4um
1ppb internal standard per 25mL sample purge
Helium carrier gas
split ratio 30:1


Thank you so much,

Hi Jackie

There is another current thread on this subject also.

You probably have absorption of all the compounds in the mix, including the i.s., or you have carryover.

Look in detail at the plot of peak area vs concentration for the analytes. Do not divide their areas by the area of the i.s.. If the curve is steeper at the top than at the bottom it confirms absorption or carryover.

To check for carryover inject clean solvent after the highest standard and look for small peaks at the analyte retention times.

In either case the cure is nearly always to put in a clean inlet liner, assuming that you are doing liquid injections. If you are doing headspace you probably need to increase the temperature of traps, loops or transfer lines.

Peter

Hi Jackie

There is another current thread on this subject also.

You probably have absorption of all the compounds in the mix, including the i.s., or you have carryover.

Look in detail at the plot of peak area vs concentration for the analytes. Do not divide their areas by the area of the i.s.. If the curve is steeper at the top than at the bottom it confirms absorption or carryover.

To check for carryover inject clean solvent after the highest standard and look for small peaks at the analyte retention times.

In either case the cure is nearly always to put in a clean inlet liner, assuming that you are doing liquid injections. If you are doing headspace you probably need to increase the temperature of traps, loops or transfer lines.

Peter

What column flow rate, what kind of vacuum system, and what ions are you using for fluorobenzene?

Column flow rate: 1.3 ml/min
Vacuum system: Oil diffusion pump
Ions for Fluorobenzene: We are running with primary ion 96. Also tried using 70 as primary ion, then tried using 96 as primary in conjunction with 70 as secondary, but none of these settings made a difference.


Thanks for your help

I had this same problem. Fluorobenzene response would be higher on the high point of the curve than the low point, but it was much worse as the response would jump during analytical runs. For a short period of time the response of fluorobenzene would double for a few runs then drop back down. For some reason the only way I found to solve this problem was to slow down my temperature ramp. I was using an RTX-VMS column with the same ramp from the Restek calalouge and getting all compounds to elute in less than 9 minutes. I spent a lot of time chasing active sites in the the GC, purge and trap and autosampler with no luck, also tried changing traps and replacing the injection port, also tried a new column. You may want to try slowing down the temperature ramp, particularly the period of time when fluorobenzene elutes and see what happens. FYI a good person to contact is Mark Ferry with ECS-MDL at www.ecsmdl.com, he helped me out quite a bit when I had that problem.