Chromatography Forum

How to calculate limit of detection? Please kindly let me know, step by step.

Thanks
K.M.P

Kindly go to Google.com and do a search on LIMIT OF DETECTION. You will find references to the information you seek.

This is a place to discuss chromatography problems, not a place to receive free education courses.

Thanks.

Rod

Feel sorry a little bit na^^.
Actually in my mind , the most believable forum is this chromatography forum.
I should improve myself first na.

Thanks
K.M.P

If you have the following, you can do it:

1. Peak height of a peak with a known concentration

2. Chromatogram baseline "blown-up" so you can estimate the height of the noise.

You should be able to figure it out from there.

Many Thanks for your kind reply.
Last 2 months we tried to run OP with GC/FID.I lost std solutions and time.The sensitivity not good. Luckily , today we can tried with GC/PFPD, we can see beautiful chromatograms. Next dayss, we will try calibration ,LOD and % recovery. I have not too much knowledge. Still trying and learning na.


Regards,
K.M.P

Dear Khun K.M.P.,

There are limited resources on the Net in Thai. I believe that rather earlier than later you will figure out how to do calculation of LOD in your particular case. Dont stop there, please create a Thai version of Wikipedia page http://en.wikipedia.org/wiki/Detection_limit

All Thai speaking students/new starters will be grateful for that. Make your contribution to the World of chromatography!

Krap

Most people calculate LOD as (mean blank reading + (3 X standard deviation of blanks)). This is usually sufficient, but it is an oversimplification.

If you want to be precise, you need to decide on the confidence limits you want (i.e the degree of certainty with which you can distinguish a sample at LOD from a blank). Then you need to look at method variation, analyst variation, and blank variation, and factor them into a t test calculation. As the previous posters have said, you can find plenty of reading material on this using google. Just remember that 'blank' injections should be prepared in exactly the same manner as samples, minus the analyte/s of interest.

I found this paper, I think It is very useful.

http://www.chem.agilent.com/Library/tec ... 8341EN.pdf

Not to belabor cdm's response, but all methods of statistically estimating the LOD from a series of standards are oversimplifications .The only valid way to assess the LOD is on a per-sample basis - you can readily do this using the method of known addition. Outside of that, you are ignoring the matrix effects of the sample whenever you quote a LOD.