CV% in liquid GC

[greta]

Hello,
I have two questions.
I'm making a GC analysis using an autosampler and I'm calculating the CV%. After 9 injection I obtain an area CV% under 1% .
My first question is: is it a good CV%?
My second question is: I obtain height CV% also of 2-3%, why this difference between area and height CV%?
Thanks

Greta

[JGK]

< 1% is what I would expect (usually from 5- 6 injections).

I would generally expect area RSDs to be less than peak height (but it's been so long since I used a method based on peak height i've forgotten why). How is the RSD for the retention time? small fluctuations in carrier gas flow could lead to wider peaks which may have a minmal effect on the area but affect height.

[HW Mueller]

Normally in chromatography, the peak hight is not effected by flow, but the area is directly dependent on it. The lower precision with peak hight is probably due to the lower intrinsic averaging of noise in peak hight as compared to area. (The integrators which I know take the highst point on the apex for determining peak hight, so no noise averaging there).

[mattj63]

To put HW Mueller's answer another way - the GC software only records a data point every so often. Maybe 20 times per second. For a fast, narrow peak, you may not always record a measurement exactly at the top of the peak, so there will be some error (noise) introduced. (If you greatly enlarge the time axis on your plot, you can see the peaks are actually a bit jagged and not nice and smooth.

For peak area, you are recording 20, 30, or more points across the entire peak and then integrating. This is less dependent on the exact location of the points across the peak (i.e. it is signal averaging).

Your %CV looks good to me - to get better can be difficult and depends on what you are analyzing.

[greta]

Thanks to all guys!
And so my question is: can I do not worry about the height and refer only to the area?

Daniela

[Ron]

The majority of GC quantitative work is done using area instead of height. Most regulated methods use area and not height. There should be no problem using area for your methods.