Calibration problems

[orcicdejan]

Hello!
I've got Agilent HP6890N GC with HP5973N MS, ChemStation software (G1701DA ver D.00.00.38), and a big problem, usually with BTEX calibration.
For small concentrations (less than ~1 ug/mL), I see the peak, he integrates it, but when it's time to include it into calibration table, he simple ignores it and doesn't calculate it's response. He exibits the same behaviour for quantitation (both samples and standards) - just says N.D.

Parameters (for example, benzene):
RT 3.01min +-0.1min (default is 0,5min, but then this problem appears for even greater concentrations)
QuantSignal: Target ion
Target: 78
Qual1: 77

Ident by: Qual & RT (tried also with RT only, doesn't help)
Max. number of hits: 9
Subtraction: extend. area quant

Can anyone help me?
Thanks in advance!

[theavenger]

Hello:

At less than 1 ug/mL you're having problems? My highest 5 level BTEX calibration on M.S. is a fifth of that, at 200 ug/L. I assume you're using a purge and trap concentrator of some kind? I have an older M.S. with Chemstation, but it seems to me that if you set your global detection limit too high, it might ignore a particular level of calibration below the global set point. I would calibrate in parts-per-billion (ug/L) if you're not, and see if it works better.
Your retention time window seems fine at 0.1 min (make sure it's not on %) The benzene retention time seems quite fast for a typical volatiles unit here, but we have to get the MTBE that elutes before benzene does, plus other low-boiling volatiles. Thus on my M.S. with a 0.2 mm column, 30 meters long, I get benzene at about 12 minutes. Your quant ion and qualifier are perfect. We use average response factor to quantify results in most cases.

I hope this is of some help. Surely the others here will have some great suggestions. I have been here a short time, but have been very impressed by the knowledge of the members.


Regards,

Glenn

[orcicdejan]

Nope, I use direct injection of sample solution (in MeCl2) or eluate of charoal tube. Global detection limit is set to 0.00.
Benzene separates perfectly from other VOC's (except cyclohexane, but anyway I work in SIM), our temperature program is optimized for VOC's. 12min for benzene would mean too long analysis for complete BTEX + ISTD.
My coleague partially solved the problem by using only target ion, no qualifier. Now it doesn't show N.D., but error is still high for lower part of calibration. Any suggestions?
Thanks anyway!

[theavenger]

That 1 ug/L seems really low for direct inject work. What kind of area count are you getting on such a standard? Is the resolution good? Column type, size and length, split ot splitless? injection port temperature?


Regards,

Glenn

[orcicdejan]

No, 1 ug/[b]mL[/b]! Peaks are completely separated and 'as in books'.
Column is DB-5MS, 30m.
We inject 3 uL splitless.
Injector is at, I think, 220"C

We solved problem, I hope so, by splitting callibrations in two - one for >2ug/mL and one for <2 ug/mL, and not using the qualifier (we can do it, RT of BTEX is very stable). Now it quantitates concetrations at least 10x smaller.