Answer to Question #13219 Submitted to "Ask the Experts"

I need to create an attenuation curve for our Gross Alpha/Beta Counter. Is there a source for specifics on number of samples, spike amount, etc.?

I don't recall any standard or standard method that specifies a requirement for the number of samples to generate a given self-absorption curve, but, in my experience, the typical number would range between 5 and 15 samples that cover the range of expected density thicknesses that you will be experiencing with your samples.

Regarding the amount of radioactivity to be used in the self-absorption standards, this is typically also not specified, but it is commonly inferred by specifying the level of uncertainty you are willing to accept in your counting results. The most common recommendation that I have observed is to have sufficient activity that the standard that you count will deliver 10,000 counts above background during the counting time (see, for example EPA Method 900 Section 10.6). The rationale for this, based on Poisson statistics, is a willingness to accept a relative one sigma uncertainty of + 1% in the result. This allows the user some flexibility in specifying the amount of radioactivity to be used. For example, if you project that the lowest beta counting efficiency that you might expect for the highest mass density thickness standard that you prepare will be 0.15 counts per beta particle emitted, then the number of beta particles emitted during the counting interval to obtain 10,000 counts would be about 6.7 x 10⁴. If you wanted to restrict your counting time to 10 minutes, and the standard emitted one beta particle per disintegration, the standard would have to contain about 112 Bq of activity. If you do not want to exceed the 10-minute counting time, and if you can justify a larger uncertainty, you may decide to use less activity; a more stringent uncertainty requirement might require more radioactivity or a longer counting time.

The amounts of inert material to be used in the standard preparations depend, at least in part on the energies of the particulate radiations involved and the expected mass loadings of the samples that you will be measuring. Most commonly, samples prepared for alpha analysis should have mass densities less than about 5 mg cm^-2 and samples for beta activity analysis less than about 15 mg cm^-2; samples prepared in a 5-cm planchette would then have respective masses of about 100 mg and 300 mg. For liquid samples that are very high in dissolved solids or for some contaminated solid samples, people sometimes elect to prepare samples in so-called "infinite thicknesses"; infinite here refers to a thickness greater than the expected range of the emitted radiations in the sample matrix. Samples and standards would all be prepared with the same approximate "infinite thickness" and the efficiency determined from counting the appropriate standard. The advantage of this technique is that the emission rate of radiations of interest does not change if sample thickness increases beyond the radiation range. Since counting efficiency may change because of closer proximity to the detector if the sample gets appreciably thicker, it is desirable to maintain similar thicknesses for samples and standards when using this technique. This is most true when the detector face is very close to the sample surface. Determined efficiencies are sometimes expressed as counts per unit per unit density thickness (cps per Bq per mg cm^-2) so that small changes in sample mass and associated density thickness may be accounted for. This assumes that the sample and standard matrices have comparable mass densities.

You can find numerous other references to the topic of alpha and beta self-absorption studies on the internet by searching alpha beta self-absorption.

George Chabot, PhD, CHP