The noise in systems processed by a Multi-Channel Amplifier (MCA) manifests itself as the width of the measured peak in energy when the input is known to be a single impulsive line. There is a lot of engineering that goes into low noise sensors for X-ray spectroscopy inside SEMs (in a vacuum). The best sensors today are high purity germanium cooled to liquid nitrogen temperatures which have an energy detection FWHM of about 120eV. Lithium Drifted Silicon sensors cooled to LN2 temps get to FWHM of about 130eV and Silicon Drifted Detectors get to about 130eV at about -30C (thermoelectrically cooled). If you want to detect and discriminate low keV photons, you need a sensor with narrow FWHM.
Of course, you also need to get your photons to the sensor, which is one of the hard parts in setting up to measure these photons in a high pressure Ni-H reactor. Basically the cold sensor needs to be exposed without a window or a good high pressure window needs to be found that will pass this low energy range of photons that can withstand high H2 pressure. Bob Higgins On Sat, Aug 16, 2014 at 2:23 PM, Eric Walker <[email protected]> wrote: > On Sat, Aug 16, 2014 at 8:17 AM, Bob Higgins <[email protected]> > wrote: > > The problem is the noise. Noise affects the FWHM of the system and >> normally getting this noise low enough so that the FWHM is smaller than >> 1keV (to get some resolution of low keV photons) requires cooling the >> sensor to liquid nitrogen temperatures. >> > > I imagine the noise obscuring the lower energy signals is stochastic. I > wonder whether a filter could be developed to do a fourier analysis and > then partially subtract out predominant frequencies seen during calibration > runs. Perhaps something like that could be effective enough to avoid the > need for liquid nitrogen cooling. >
