Ametek 682T-HP, User Manual

Page: 52 / 118

Chapter 4: Calibration, Page 45
M o d e l 6 8 2 T - H P U S E R M A N U A L A T E X Z o n e 2
4.8 Counting Statistics
A peak is detected for each Element Of Interest (EOI) and the spectrum could be displayed as a plot of
X-ray Intensity (
count rate ) versus X-ray Energy (keV). Such a peak is a random distribution about a
mean value, and is termed a Gaussian, or bell-shaped, curve. The error involved in counting the events
of such a distribution is the square root of the number of events counted.
The calculation of the error in the concentration due to the error involved in the counting statistics of
the peak is termed the calculated one-sigma Standard Deviation of the peak, and is the theoretically
best precision to be expected, as the error involved in the counting statistics cannot be removed from
the system. Thus at best, it can be expected that 68% of the measurements of a peak will come within
this calculated one-sigma deviation from a mean value.
AMETEK ASOMA instrumentation can perform to this theoretical best precision limit for many
applications. For most applications, however, some error due to instrument precision and sample
presentation is added to the error due to the counting statistics. In general, actual
observed
precision
can be expected to be 1.5 times worse than the theoretical best precision limit set by the counting
statistics of a peak. Counting statistics may be improved by obtaining more counts in a peak by
counting for a longer Count Time during analysis. This increases the total number of counts collected,
thus decreasing the relative error due to counting statistics, which then may improve precision results.