Fig. 2

a ¹⁶⁶Ho spectrum acquired for dead times equal to 42.25% (solid line), corresponding to a total phantom activity of 2.7 GBq and 9.25% (dashed line), corresponding to 447 MBq. The curves were scaled for comparison by matching the sum of the counts in the scatter window centered at 118 keV (12% width). The main ¹⁶⁶Ho photopeak window (80.6 keV, 15% width) and the scatter window (118 keV, 12% width) are represented by the gray areas. The effect of pulse pile-up due to the high-count rate is visible as the shifting and broadening of the photopeak. b The ratio between the mean of counts in the 81 keV photopeak window and the 118 keV upper scatter window within the recorded spectra, plotted as function of total activity in the phantom at time of recording. Linear regression is represented by the black line, while gray dashed lines represent prediction intervals. Data in the blue area are within the confidence intervals