Simulations by Norbert S. Schulz
Page by Norbert S. Schulz
| CSR Home Page: space.mit.edu/ |
MIT Chandra Science Center space.mit.edu/ASC/ |
Chandra Science: asc.harvard.edu/ |
Simulations by Norbert S. Schulz
Page by Norbert S. Schulz
TITLE: Strong X-ray emission line complexes in the highly compact binary pulsar 4U1626-67.
4U1626-67 is 7.7s X-ray pulsar in a compact (40 min) binary system X-ray continuum consists of 0.6 keV blackbody and a variable power law of 0.8 - 1.2. ASCA discovered a 4.4 c/s (SIS0) strong emission-line complex at 1 keV (Angelini et al. 1995) accompanied by a broad feature at 1.42 keV, probably Ne(H) recombination edge (1.36 keV), Mg(H) at 1.47 keV and Mg(He) at 1.34 keV. The whole complex, however, remained unresolved. Blends with unresolved O-K and Fe-K line complexes seemed highly probable. The data indicated an overabundance of Ne of 6 times, O and Mg 2 times the solar abundance. Several questions arise from these indications: Why is there so much Ne, O, Mg? Would that point towards a Ne or He-white dwarf companion? Could we possibly constrain evolutionary scenarios such as an AIC from a CO or ONeMg white dwarf or a type II supernova of the progenitor star. If it was a type I, where is the Si and S?
Model Spectra
The basis for this simulation is the proposed model spectrum obtained from fitting the ASCA spectrum as presented in Angelini et al. 1995. Clearly because of the limited spectral resolution of the CCD devices on ASCA the best fit line emission there gives only be a rough indication of the underlying emission line structure. In an early attempt to simulate the ASCA result we used the precise fit obtained by Angelini et al. (see table 1 therein), which consisted of a two component blackbody plus a power law model and the fitted energy O- and Ne-emission lines plus the Ne- recombination edge. The resulting photon spectrum is shown in figure 1.
Figure 1: ASCA SIS0 model spectrum
This model spectrum was used to optimize the exposure time.
Figure 2 shows a result for an MEG combined first order spectrum.
With 30 ks all input lines for O- and Ne- are clearly detectable.
Figure 2: An MEG combined first order spectrum using the model
in figure 1 for 30 ks.
Figure 3: The photon spectum using SPEX bright lines and a blend of
a Ne-recombination and Mg-absorption edge
(figure 4:
same with in energy scale).
Simulation Set-up
The current set of simulations below were carried out with
the MARX simulator version 2.15, which is still under development and not yet
accessible to the public, but which offers several improved features such
as an upgraded HRMA psf and a more realistic vignetting function. It however
still utilizes a HRMA effective area model with 'old' optical constants
which especially around the Ir-M-edges will significantly differ
from reality. Please look at the
MIT MARX Simulator
page for
latest updates and versions. Note, that any version lower than
2.15 won't work for the set-up below.
For the simulations we use a template
marx.par file and set every other parameter
from the command line:
"marx ExposureTime=30000 OutputDir='MyDir' GratingType='HETG' DetectorType='ACIS-S',
SpectrumType='FILE' SourceType='POINT' SourceFlux=-1
SpectrumFile='table5'"
Table 5 is the
input spectral table as seen in figures 3/4. The same procedure has been applied
using a spectrum without edges and emission line features. The resulting spectrum
is in table 6.
Because we do not
yet have a working grating response matrix, we constructed model spectra
each for the MEG and HEG +/-1st order by folding the continuum components
of the source spectrum with the instrument areas and efficiencies.
Multiplied by exposure time and adjusted to the correct energy binning, the result
serves as a perfect continuum model fit to the simulated spectra.
Note, that the ccd gaps appear in the model and in the simulated data,
since for the time being, no dither motion has been applied. For the residual
spectra below we use a fit of the model spectra to the simulated spectra
of the continuum alone to correct for imperfections in the model (gap edges
etc..). Those model spectra are in
table 7
for the MEG combined 1st order, and in
table 8
for the HEG combined 1st order.
The final simulated eventlists are somewhat large in size (~20-50 Mb) and we will not
offer them on the public domain. However, we will make them available upon
request (send e-mail to: nss@space.mit.edu).
Simulated Spectra
Table 9 ( figure 5 left below) and
Table 10 (figure 7 right below) show the full
bandpass 1st order spectra of with
the MEG and HEG on a wavelength scale (note, in most most cases for those
who are not yet used to wavelengths in X-ray binaries, we added plots with
energy scale also, see
figure 6 and
figure 8 ). All figures shown below use one resolution cell
binning (0.048 mm), which is the highest possible for each instrument.
In order to ee how well each instrument detects and separates
the input line emission we constructed residual plots for several
wavelength regions for the MEG and HEG. Most line emission is expected
to be around 1 keV and below (we did not include the possibility of Si- and S-
emission). Therefore the MEG will be the instrument with the largest
bandpass for expected line features. Figure 9 below shows the
expected MEG residuals in between 5 and 25 Å.
Norbert's AXAF Science page
