Subject: Jem-X detector resolution

Dear all,

Calibration of detector spatial resolution. A first test.

As you all know, the flux estimate for a source is based on
the difference in count rate on the source illuminated (open)
and shadowed (closed) part of the jem-X detector. In source
extraction I define the open area as the pixels for which
the illumination fraction (PIF) is larger than some cut value
(usually 0.4). Likewise the closed area is defined by pixels
with PIF < 0.1. The finite detector resolution reduces the 
contrast between the open and closed parts. This must be 
corrected for, which I do by using the instrument model to 
compute how much of the source flux remains in the area
defined as open and how much ends up in the area defined
as closed. 

The detector resolution depends on energy, in particular it
becomes worse towards lower energies.

There is a possibility to calibrate this leakage (in effect
the spatial resolution of the detector) from observations
of a pulsating source. Analysing one science window for
Vela X-1 (period = 284 sec) I have devided the detector
in two regions, PIF > PIFCUT and PIF < PIFCUT. I extract
the lightcurve for each and estimate the pulsation
amplitude, which is proportional to source flux.
I then compare this with the expected leakage which I
compute from the instrument model.

The attached plot shows the fraction of pulsed flux
leaking from the open to the closed area for different
PIF CUT values. Solid line is the measured leakage and the
dashed is the predicted. PIFCUT = 0 is with the full detector 
regarded as open. Larger PIFCUT selects only the most source 
illuminated regions of the detector.

The energy band was 4.5-7 keV.

Taken at face value the result suggests that the resolution
has a sharper core and larger wings than the model.
I am not sure that this is significant but remember that it
is just one scw. There is much more data so the dependence
on energy and off-axis angle can also be tested.
It just takes a bit of work...

Cheers,

Stefan