Revisiting Currie - How low can you go?

Over the years many different formulae have been proposed to derive the 'limits of detection' for radiometric measurements. With the ever increasing employment of quality systems within the radioanalytical laboratory, it is vital that the term 'limit of detection' is well understood and defined in the documentation in use within the laboratory. This paper reviews the various formulae used to derive a limit of detection, most of which claim to be derived from Currie's landmark paper of 1968, which set out in detail the means by which the analyst can arrive at a limit of detection and limit of determination. However, with the vastly increased computing power available to the analyst, coupled with the off-the-shelf availability of low and zero background detectors, it is now more straightforward to arrive at exact solutions to define limits of detection and determination. Finally, we set out some examples of decision making at these limits and how this may inform the way in which detector time is allocated for the measurement of samples containing low levels of radioactivity. (C) 2000 The National Physical Laboratory.