Novel methods of gamma ray spectroscopy focusing on the determination of the vertical source distribution in a media.

Abstract

This dissertation presents novel techniques for performing gamma ray spectrometry. It presents methods applicable to in situ measurements and laboratory measurements. For in situ gamma ray spectrometry, in order to obtain accurate estimates of the radionuclide activity present at a particular location, the distribution of the source being measured must be determined. Typically, the horizontal distribution can be assumed to uniform; however, the vertical distribution is usually unknown. By using multiple gamma spectroscopy measurements different measurement geometries, the vertical distribution of a source may be estimated. The measurement geometry is varied by employing an adjustable collimator to vary the field of view of the detector. By employing an expression that is both analytically and empirically derived, the response of the detector can be calculated. Employing reasonable simplifying assumptions, a system of equations can be solved to estimate the vertical distribution of the source. Methods are presented for calculating the depth of a planar source, determining the distribution parameters of an exponentially distributed source, determined the activity of a combination of uniform volume sources, and for determining the analytical form of an unknown distribution. The analytical form of a distribution is determined by fitting a series of typical source distributions to the results of a set of measurements. The analytical form that results in the lowest residual, with some additional criteria applied, is selected as the source distribution present. The primary difficulty with all of these methods is conditioning. Small errors in measured count rates lead to large errors in calculated activities. The error proprieties of the methods are examined. Methods are presented for significantly reducing the condition number of the measurement system thus reducing the system's susceptibility to errors in measurements.