Bias vs different internal standards

Chromatography Forum: GC Archives: Bias vs different internal standards
Top of pagePrevious messageNext messageBottom of pageLink to this message  By Anonymous on Tuesday, January 21, 2003 - 09:58 am:

All- Has anyone experienced a problem similar to mine? I am running a residual solvent method and my analytes are THF, toluene, DMF, and DMSO. I'm evaluating 2-butanone and dodecane as internal std.
I'm using a Restek RTX-200, 30m x .32mm column at 3mL/min He.

Inj = 280C, Det = FID at 280C.

Oven program: Init = 60, hold 4 min, ramp 15C/min to 180C.

THF, toluene, and 2-butanone elute between 3-6 min then there is a lag and DMF, dodecane, and DMSO elute between 9 and 11 min.

Here's my problem: When I calculate spike recoveries using 2-butanone as ISTD, I get 100% for THF and toluene, 110-115% for DMF and DMSO (biased high). When I calculate recoveries using dodecane I get 100% for DMF and DMSO but only 90% for THF and toluene (biased low). Can anyone explaine to me please why this may be happening? All other aspects of my chromatography (tailing, resolution, reproducibility) are excellent.

Thank you all in advance for your advice.

-Jeff


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Anonymous on Tuesday, January 21, 2003 - 12:25 pm:

I can't explain the difference in the recoveries but I can give you some advice. It is generally accepted that if you are using multiple internal standards, you should use each one to quantitate compounds with similar retention times.

You said that THF, toluene, and 2-butanone come out within a few minutes of each other and that DMF, dodecane, and DMSO come out within a few minutes of each other with a bit of a gap between the two groups of analytes. I would use 2-butanone as the internal standard for THF and toluene and use dodecane as the internal standard for DMF and DMSO.

Hope this helps


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Anonymous on Tuesday, January 21, 2003 - 12:34 pm:

Anonymous-
That was my plan; quanititate THF and toluene via 2-butanone, and DMF/DMSO via dodecane. However, I was asked by my supervisor to justify the need for the two different ISTD's (and in effect, the reason for the bias). Thanks for the reply.

-Jeff


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Rodney on Tuesday, January 21, 2003 - 03:32 pm:

Jeff,

I suspect the problem may lie in the integration of your peaks, in particular, the peak width doubling which the integrator does without direct imput from you.

If you would increase the initial temperature of the oven and allow the integration of DMF - C12 - DMSO to be done earlier, ie, have the peaks elute more quickly with less but enough resolution, you might see the peak area bias decrease or even disappear.

This is not intended as a permanent fix but to diagnose the problem.

Try it and see.

best wishes,

Rodney George
Supelco


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Anonymous on Wednesday, January 22, 2003 - 05:29 am:

Rodney-
Thanks for the input! I was not able to increase to initial temperature but I was able to increase my first ramp to 30C/min, allowing DMF-C12-DMSO to elute 2.5 minutes sooner with no significant loss in resolution. When I calculated recoveries that way the bias was significantly less (4-5% vs 10-15% before). Could you please explain to me what "peak width doubling" is, why the software may do it and anything I can do to "fix" that? We use Turbochrom C/S v6.1 in our lab. Thanks again for your advice.
-Jeff


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Anonymous on Thursday, January 23, 2003 - 08:05 am:

Peak width doubling is when the software adjusts the expected peak width of a peak in the chromatogram. In an isothermal run the peaks get progressively wider during the run, and to help assign the baselines better software has a feature where after a set time the expected peak width is doubled, so if the initial peak width is expected to be 15 sec. then at 15 min the software expects peaks to be about 30 sec wide and then at 30 min it expects peaks to be about 45 to 60 sec wide. Generally the software has an option to turn off peak doubling and this is generally what should be done for programmed GC runs. Hope this rambling helps somewhat.
Regards,
Mark


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Rodney on Thursday, January 23, 2003 - 12:13 pm:

just to let the forum know, I communicated to Jeff privately about the problem; containing essentially what Mark wrote above and commented about the used of timed events in the TurboChrom software to control the doubling of peak width of the integrator.

Rodney George
Supelco


Add a Message


This is a private posting area. A valid username and password combination is required to post messages to this discussion.
Username:  
Password: