Behavior of Several Germanium Detector Full-Energy-Peak Efficiency Curve-Fitting Functions

sult in interpolated efficiencies differing by as much as 6% when using different curve-fitting functions in the 700-keV to 900-keV region? Second, since the coincidencefree standard only covered the energy range of 59.5 keY to 1115.5 keV,how much uncertainty would be introduced by extrapolating the curve fits to higher energy where there were no experimental points? Third, what effect would one inaccurate data point have on an efficiency curve? Fourth, if it is suspected that a particular point is less accurate than the Qthers in a data set, is it better to use it or delete it? Since most commercial software packages do not allow weighted fits, weighted fits will not be considered here. The coincidence-free, Marinelli-beaker standard in reference 2 emitted gamma rays at 59.5, 88, 122, 165.9,279.2,391.7,661.6,810,834.8,1099.3, and 1115.5 keY. Using a least-squares procedure, the efficiency data from one count was fitted to a fifth-order polynomial of the form: