The background count with the empty container is not the background of the sample. The sample background is altered by the sample matrix and contributions to the K-40 ROI from other radionuclides in the sample. The K-40 background of the sample must be determined from the sample spectrum. The variance of the net count is the sum of the variance of the determined net and the variance of the determined background. The uncertainty is the square root of the sum.

This article may help you..see section 3.2

Article A methodology for expanding the use of NaI(Tl) based spectro...

Thank you very much Joseph L Alvarez sir. Could you please explain more how to determined background from sample spectrum

Thank you very much mam Mirofora Pilakouta for your valuable suggestion. we are using HPGe for gamma spectrometry and in this case is it required to include Compton Background in the formula for evaluation of uncertainty regarding the net count.

It is necessary to include Compton correction with HPGe. Nevertheless, a good peak fitting program will include a background fitting algorithm. Review the MCA documentation used by your system. The background fitting algorithm should be included. You should be able to extract the peak net and ROI total. Some MCA programs will propagate the error and provide an estimate of total error.

Joseph L Alvarez sir for your kind information........Our system is from CANBERRA and we are using GENIE 2000 software. I am not sure whether it is included background fitting algorithm.

GENIE 2000 has an algorithm to fit the background under each peak. Search the documentation for peak fitting and quantification.

Mirofora geia sou,

tin uncertainty to net count rate mporeis na tin ipologiseis apo ti sxesi

uncertainty (net) = SQRT(uncertainty(gross) ^2 + uncertainty (bg)^2)

apo ta dedomena sou vgainei oti exeis net count rate 1590 ± 165 cps

sto K-40.

kali douleia me tin fetini intercalibration...(to deigma tis garidas einai poli tricky)

nasai kala

Stelios Stoulos

I understand you got 1.03E+04 +/- 122 counts for your sample as the output of your measurement program. So this program obviously calculates uncertainty taking into account background. In the same measurement time without sample you get 8710+/-111 counts as the lab background for K-40.

So your net count, corrected for lab background would be 10300-8710=1290 counts of K-40 in your shrimp sample and the uncertainty in first approximation adds up as 122+111=233.

So you get 1290+/-233 counts for your shrimp sample.

Erik Strub

You must add the uncertainties in quadrature.

Subtracting the background value will give a first approximation of the counts if 10300 is the gross count.

It is best to use the software net and uncertainty.

Joseph L Alvarez

I am still not sure what we are talking about. Talking about gross counts, net counts, background counts, uncertainties in 1 spectrum, you are right. But I think Abu Haydar is talking about substracting a second spectrum also containing K.

There are simple and accurate method if you have calibrated program in ur setup using good fit algorism if it give you net count and the error on it you can directly divided error by net count it uncertainty in net count rate its a type A of uncertinty ( random error) part of the total uncertainty . U should test it by classical method as sugesseted for one time taking the net count channel by channel with back ground and make sum then calculate the uncert.

You can calculate the net counts and error as follows.

Suppose G the gross counts and σG the error of the sample measurement and B and σB the error of background measurements. We also suppose that the mesurement time is the same for sample and backgroud (90000 sec) otherwise you should find the count rates cps and the errors.

Then the net counts N are given buy the equation N=G-B and the net error σN is given by the equation: σΝ= sqrt(σG^2+σB^2).

We can calculate Net count rate uncertainty in gamma spectrometry Technique by using square limit equation, you should calculate the total error ( systematic and theoritical) before calculate Activity of samples