Chromatography Forum

Have encontured something I have never seen before. Has anyone had problems with the Thermo GCMS ISQ 7000 instrument and uneven fragmentation in EI?
I get problems with upward curvature in my calibration of PCBs since the fragmentation seems to decrease with increased concentration. If I quantitate using TIC in Chromeleon the curve becomes linear but if a use a the molecular ion as quantification ion the curve is quadratic (upward curvature). Has anyone encountered this before with this instrument. I have only used Agilent instruments (several versions of the 597x series) before and I have never seen this kind of problem before.

in EI it is normal to observe an effect of auto-chemical ionization when the concentration of your analyte is sufficiently high so it acts as a reagent gas to itself, and this is non-linear

in CI we always get higher abundance of the molecular ion, so this explains why increase in concentration increases the relative abundance of the molecular ion in EI

the ion volume in ISQ is much smaller than that in the Agilent ion sources; this also contributes to higher "spacial" concentration of your analyte within the ISQ source and hence amplifies the ion-molecular reactions leading to auto-CI

that said, increasing the split flow or decreasing the injected amount should make the curve linear; if your repeller voltage is set to default 0.5V, you may also try increasing it to 3V and see if it changes the picture

Thanks for the help Trozen!

Okay! Auto-EI. Sounds very reasonable. Never encountered this on a Agilent GCMS. Must be due to the much smaller ion volume in the ISQ instrument. One thing that I can't understand is why auto-CI alters in intensity. I can have a period were the calibration curves are linear and a couple of days later I get an upward curvature with the exact settings on the instrument. Are there any factors that can decrease/increase the auto-CI?

I have spoken to Thermo about my problem and they told me to decrease the emission current from the default value 50 µA to 25-20 µA (producing less ions) or raising the repeller. In the EI tune the minimum repeller voltage is 3 V.
I Think 0.5 V is only used in the "Initial EI tune". If decreasing the emission current doesn't help I can raise the minimum voltage of the repeller to 5 V in the tune file.

I have been so focused on the msd that I have totally forgotten the parameters for the GC. For the moment I'm injecting 1 µl and splitless. I will absolutely try with split injection (split ratio 4). Have never performed a split injection for PCB analysis. Hope I don't encounter any problems.

/Stefan L

Stefan,

this indeed is weird that occasionally the calibration curve is linear; can you say that in those instances (linear curve) you have had lower absolute counts (peak area) of your analyte(s)?

also, you may want to look into the tune reports and see if there is any apparent correlation between linearity and voltages in the source (lens 1,2,3), ion guide and Q1

can you tell me what is the absolute amount (ng) you inject? lowest and highest
and the ion source / transfer line temperatures

for the split ratio, it is generally not recommended to set values below 1:10, as the GC may have difficulties controlling the split flow

you have an EI source, not AEI, right?

Sorry for my delayed reply Trozen and thanks again.

I am injecting between 0.1 ng-20 ng for all seven PCBs (#28, 52,101,118, 138, 153 and 180). Both the source and transfer line are set at 330 °C. Note also that I use non-labelled internal standards (PCB 30 and 143).

For the moment all my curves are linear. A small tendency for upward curvature for PCB 180 but within acceptable levels (residual <10 % for the highest conc).
This is after replacing the filament, lowering the emission current to 20 µA from 50 and changing the electron lens voltage in the tuning to 5 V (previously 15 V).

I got some interesting results when I changed the repeller voltage to 5 V (usually 3 V). This seems to increase the auto CI?! Much more abundance of the molecular ion and aspecially for PCB 52. The response ratio (=PCB 52/PCB 30 (ISTD)) for the highest concentration (20 ng) changed from 2.87 to 4.71. The ratio for my qualifer (m/z 220/292) for PCB 52 changed from 88 % to 40 %. When I changed back to the original repeller voltage I got my original values again.
Very weird!!

I will look into if I can see a apparent correlation between linearity and voltages in the source (lens 1,2,3), ion guide and Q1 as you mentioned.