ECD linearity Problem

Chromatography Forum: GC Archives: ECD linearity Problem
Top of pagePrevious messageNext messageBottom of pageLink to this message  By Donna Ruszaj on Wednesday, February 2, 2000 - 06:37 am:

I am using a HP5890 Series II GC for PCB analysis. Our standard range is 0.05-2.0ng/ul. We see a continual rise in responce as the concentration decreases. Our responce factor for the 0.05 standard is quite a bit higher than our 2.0 standard. I'm not sure how to explain this increase. We are using a DB-5 megabore column for the analysis. The chromatograhy looks good but we are unable to get a curve at this concentration range. What is the range of standards for ECD 100x? We always prime the instrument prior to use. If anyone has any suggestions or explanations we would appreciate the help.
Thank you in advance.


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Steve B on Wednesday, February 2, 2000 - 09:59 am:

Assuming sample prep etc has been ruled out, here are a couple of useful things to know about the response of the ECD. First, ECDs inherently have a limited linearity (typically 10^3). It is important that the baseline signal be at the low end of response needed to match that linearity. On HP 5890s, that is usually somewhere around 20-50 HZ. If your signal is lower than 10 HZ (which can happen with newly cleaned ECD cells) or above 60 HZ (which indicates contamination) this might be your problem.
How to proceed? First check your carrier and auxillary gas flows. Has either gas cylinder been changed recently? The total of carrier and aux gas should be 30-60 mL/min and should be either N2 or P5 (5%Ar/Methane) Next check your baseline signal. If it is too high, a thermal cleaning or exchange may be in order (Please see 5890 reference manual). If is too low, then it is possible to artificially raise the baseline signal into your linear range. This can be done directly on the ECD board. For those details, you can follow the refence manual or call for service.
If you have more questions, respond back and include info on flows, gas type and baseline signal.


Top of pagePrevious messageNext messageBottom of pageLink to this message  By Donna Ruszaj on Thursday, February 3, 2000 - 06:22 am:

Hi Steve,

Thank you for your responce. Our signal is 54hz and the total flow is 60ml/min ( make-up is P5).
All our instruments run off the same cylinders and therefore we are always changing. I still can not theoretically explain why the low standards give a higher responce. I wouldn't don't that the detector is contaminated given the wide range of samples that have been run. We will try the thermal cleaning.

We start at 150 hold for 1min then ramp at 20deg/min to 210 then ramp at 10deg/min to 285 then hold for 10min. Column flow is 4ml/min.

Thanks again Steve!


Top of pagePrevious messageNext messageBottom of pageLink to this message  By John Hinshaw on Friday, February 4, 2000 - 09:37 am:

Assuming that you use splitless or on-column injection, 2 ng into an ECD might be on the high side of its linear range. You could make a graph like this:

Click to view graph

In this example, the detector response drops off at higher amounts because its linear range has been exceeded. Cleaning the detector may restore the range. If not, consider injecting a smaller sample volume.

Also, if you decide to live with the non-linearity, you could apply polynomial curve fitting to calibrate the data handling system and attempt to correct the drop off at the higher end.


Posting is currently disabled in this topic. Contact your discussion moderator for more information.