CV% area/height

[greta]

Hello,
I'm making an headspace analysis using an autosampler and a GC system (FID detector) .
Can someone of you explain me a possible reason why I obtain a good repeatability (CV%: 0,3%) as concern the peak height and not goood as concern peak area (1-2%)?
Can it be a flow problem?
Thank you very much

Greta

[mbicking]

First, carefully examine your integration baselines. You may have to expand the scale to look closely. Are the start/stop points similar for all peaks? Are there any small peaks that are partially resolved?

[greta]

The peak integration seems correct and similar for all the peaks.
I don't think it's an integration problem.
Thanks
Greta

[kahmark]

Can you post a chromatogram? Is your baseline noisy or flow inconsistent?

[BZK]

The answer is simple. With peak height you measure only one dimension (the height) and with area you measure n time the peak height (integration during the time). The error of measure is depending from signal to noise ratio. This ratio is not the same at top of peak (the most favourable value) and beginnig or end peak (the most un-favourable value). As result the CV % is normally better with heigh than area. There are a lot of further consideration such as detector saturation, overload and broadening that at the end say that the true measure is better represented by area.
Use area.

Ciao.

Robertino Barcarolo, Italy

[greta]

Hi BZK,
The flow and the baseline are stable. Can you explain me how can I attach the chromatogram images to my post?

[BZK]

Hi BZK,
The flow and the baseline are stable. Can you explain me how can I attach the chromatogram images to my post?

Try to make a jpeg file of chromatogram and use Img button to upload.
[/img]

[tom jupille]

The instructions for displaying graphics have been on the LC board for years, but I just realized I had never copied them to the other boards

Now fixed. Instructions are here:
viewtopic.php?t=12659

[djpegram]

** I agree with what BZK said.
To elaborate a little on peak hieght and peak area...
Peak height can be measured with great precision, but it is inversely proportional to peak width (so you must be sure that your peak widths are remaining constant throughout your analysis). To do this you can control your columns temperature, eluent flow rate, or your rate of sample injection. Important to reiterate, do not overload your column.

Peak area doesn't vary with peak width, however it is affected by adjacent peaks. Peak area is also proportional to concentration; if the width changes then the height will change therefore peak height is not as good for analysis (peak area should remain the same).

Look up the Van Deemter equation too, that will help for ways of reducing zone-broadening and you can see the H vs u curve to give you an idea of optimum velocity at which H is at the minimum. Using Van Deemter, the five areas of zone broadening are Eddy Diffusion (A-term), longitudinal diffusion (B-term), mass transfer (stationary and mobile, cm and cs terms) and stagnant pools. If anyone would like these individually explained or how to reduce the band broadening using these equations, let me know.