MARX simulations

The Model of AXAF Response to X-rays MARX (Wise et al. 1997) is widely used to simulate AXAF data in detail. The MARX associated tool pileup allows simulations of the pileup effects (for a description of the pileup tool, see the MARX manual). We used MARX v.2.0 to simulate the pileup fraction for Phase H data taken with ACIS-I (Table 6.22a), since the code presently supports only FI chips. MARX provides a more flexible tool to predict pileup than the analytical model ofMcNamara (1997), since it incorporates a more realistic physical description of the CCDs, allowing for grade selection. Note that the currently available MARX version has incorrect values of the encircled energy at energies > 4 keV, whose effect is to broaden the PSF and thus affect the pileup analysis at those energies. We thus restrict ourselves here to using MARX at the lower energies.

We first simulated the source spectrum assuming a monochromatic emission line at 1.486 keV for the Al-K$\alpha$ tests and at 0.525 keV for the O-K$\alpha$ tests, normalized to the source flux as derived from the analysis of the BND data and integrated over the same net exposures (Table 6.21). The parameter ``DetIdeal'' in the marx.par parameter file was set equal to yes, as recommended. The staggered mode of Phase H data was simulated by activating only 38 rows centered on the nominal I3 aimpoint in the acis.par parameter file, and adjusting the frametime to the observed value (see above). The split and event thresholds were fixed at 13 and 38 adu, respectively, as appropriate for an FI detector. The ``MeanNoise'' and ``SigmaNoise'' parameters were fixed at their default values (2.0 and 1.0, respectively).

Table 6.23 reports the results of the MARX pileup simulations. Column 2 lists the total number of photons detected by the pileup tool, while column 3 gives the fraction of total counts that were collected. The following columns give the number of events detected in the primary and piled peaks, and the corresponding pileup fraction (according to the definition in eq. 6.14). The latter was derived in the same regions of interest as for the XRCF data. Note: No grade selection is considered here, since the order of the pixels in the 3x3 island in the FITS files produced by MARX is different than in the XRCF data.

The results are shown in Figure 6.38, where the percent difference of the MARX predicted pileup and the fraction measured for the Phase H data is plotted versus incident count/frame. Only data taken at the Al-K$\alpha$ energy are plotted. The Figure shows that for every event detection rate the predicted pileup by MARX is lower than the measured pileup.
 

Figure 6.38:  Comparison between the pileup fraction predicted with MARX and measured for phase H data for All-255 (see text). MARX underestimates the measured pileup at all observed event detection rates.

 

 

Table 6.23: MARX Pileup Simulations for Phase H Data 

Test Name		# Events for All - 255
	Detected Events	Frac. of Events	Peak1	Piled Peaks	Pileup Frac.
H-IAI-CR-1.001	13528	0.99	13089	419	0.031
H-IAI-CR-1.003	19052	0.99	16439	2572	0.135
H-IAI-CR-1.005	14601	0.99	10077	4491	0.308
H-IAI-CR-1.007	15721	0.99	9807	5852	0.374
H-IAI-CR-1.009	13199	1.00	12434	765	0.058
H-IAI-CR-1.011	9643	1.00	8291	1352	0.140