X-Ray Imaging of the Jet in J2310-43

Preparing a Spectrum File for MARX

MARX includes an internal FLAT spectrum model which produces a uniform flux at all energies. More complicated spectral models require the use of the FILE option in which the user provides a spectrum file. This is an ASCII file with two columns; the first in the energy in keV, the second the flux in units of photons/s/cm^2/keV. The overall normalization of the spectrum (in photons/s/cm^2) can either be set by the SourceFlux parameter, or if SourceFlux is set to"-1" the intrinsic normalization of the file will be used. Here we show how to create a spectrum file within Sherpa, and estimate the source flux in photons/s/cm^2. If you want to use XSPEC to create the spectrum file, please refer to the MARX manual for details.

First run CIAO, then Sherpa. The following commands

• define a spectral model of a simple powerlaw convolved with absorption (using XSPEC "wabs" and "powerlaw" functions)
• set the powerlaw photon index to be 2.43 and the nH 0.016 (values appropriate for J2310-43
• create an energy grid for the model -- in this case from 0.1keV to 14 keV in steps of 0.01 keV
• write the source model to an ASCII file named jet.dat

sherpa> source = xswabs*xspowlaw
xswabs.nH parameter value [0.1] 0.016
xspowlaw.PhoIndx parameter value [1] 2.43
xspowlaw.norm parameter value [1] 
sherpa> dataspace(0.1:14:0.01) histogram
sherpa> fakeit
sherpa> write model jet.dat ascii
Write X-Axis: Bin  Y-Axis: Flux (Counts)

We can use the file jet.dat as input to MARX. Note that X and Y Axis units are given incorrectly as Bin and Counts. They are in fact keV and photons/s/cm^2/keV. The normalization of the spectral file is not scaled to the flux of J2310-43, but set to 1. As explained above, we will use the SourceFlux parameter to scale the overall flux.

From section 3, we found that the 0.2-10 keV flux for J2310-43 is 3.14E-12 ergs/cm^2/s. We can use Sherpa to convert this to photons/s/cm^2 (units for the SourceFlux parameter in MARX). First use the show command to view the current model parameters:

sherpa> show 

Optimization Method: Levenberg-Marquardt
Statistic:           Chi-Squared Gehrels

-----------------
Input data files:
-----------------

Data 1: fake.
Total Size: 1390 bins (or pixels)
Dimensions: 1
Total counts (or values): 6

------------------------------
Defined analysis model stacks:
------------------------------

source 1 = (xswabs * xspowlaw)

-------------------------------------------
Defined source/background model components:
-------------------------------------------

xswabs[xswabs]  (XSPEC model name: absw)  (integrate: off)
    Param   Type      Value        Min        Max                 Units
    -----   ----      -----        ---        ---                 -----
 1     nH thawed    1.6e-02      1e-07         10            10^22/cm^2

xspowlaw[xspowlaw]  (XSPEC model name: pwlw)  (integrate: on)
    Param   Type      Value        Min        Max                 Units
    -----   ----      -----        ---        ---                 -----
 1PhoIndx thawed       2.43         -3         10                      
 2   norm thawed          1          0      1e+24photons/keV/cm**2/s at 1 keV

Now the eflux command to get the flux of the model in ergs/cm^2/s with the normalization set to 1:

sherpa> eflux
Flux for source dataset 1: 5.31921e-09 counts

Rescale to the flux of J2310-43 (set normalization to 3.146e-12/5.31921e-9=0.0005915) and use the flux command to get the flux in photons/s/cm^2.

sherpa> xspowlaw.norm=0.0005915  
sherpa> fakeit
sherpa> eflux                  
Flux for source dataset 1: 3.14631e-12 counts
sherpa> flux  
Flux for source dataset 1: 0.00304953 counts

Once again, the units are given incorrectly on the command line. The flux in the model is 3.146e-12 ergs/s/cm^2 or 0.00305 photons/s/cm^2.

In the following simulations we make the simplifying assumption that the nuclear source and jet have the same spectral energy distribution.

MARX Simulations with Radio Image as Input

Our observational goal is to map the X-ray emission in J2310-43. We have no a priori way of determining the X-ray structure, but the radio map (J2310_radio.fits) shows a jet. For our first simulation we therefore assume that the X-ray structure follows the radio structure and use the radio image as an input to MARX.

First copy the default MARX parameter file to a file called radio.par. The default marx.par file resides in the directory MARX was installed in ($MARX_DIR).

unix% cp $MARX_DIR/marx.par radio.par

Now set up the simulation

unix% pset radio.par OutputDir=radio 
unix% pset radio.par ExposureTime=40000 
unix% pset radio.par DetectorType=ACIS-S
unix% pset radio.par SpectrumType=FILE 
unix% pset radio.par SpectrumFile=jet.dat 
unix% pset radio.par DitherModel=INTERNAL
unix% pset radio.par SourceFlux=0.00305 
unix% pset radio.par GratingType=NONE
unix% pset radio.par  SourceType=IMAGE
unix% pset radio.par S-ImageFile=J2310_radio.fits
unix% pset radio.par SourceRA=347.674167
unix% pset radio.par SourceDEC=-43.792861
unix% pset radio.par RA_Nom=347.674167 
unix% pset radio.par Dec_Nom=-43.792861

Full details of all these parameters are given in the MARX manual. The crucial parameters for this simulation are:

• The output directory is called radio
• the Exposure time is 40ks
• The detector is ACIS-S
• The source spectrum is read from a file
• The name of the spectrum file is jet.dat. Please note that in this simulation the parameter SourceFlux gives the overall flux normalization. The file jet.dat is used for the shape of the input spectrum.
• Use MARX default dither parameters
• The source flux is set to 0.00305 photons/sec/cm^2 (derived from the count rate)
• No grating is used. Note that the parameter file has HETG as the default, so this must be changed for an imaging simulation
• The spatial distribution of the source is read from a FITS file
• The name of the file is J2310_radio.fits
• Source RA and DEC
• Pointing position (in this case the same as the source)
• To use the contamination model in MARX, you need to provide a TSTART when you run marx. TSTART=264038464.18 will produce a MARX simulation with an amount of contaminant absorption projected for the date 2006-05-15T00:00:00, equivalent to the date used to create the ACIS Cycle 07 Responses available on the Chandra proposal planning web page.

Now run the simulation:

unix% marx @@radio.par TSTART=264038464.18
MARX version 4.2.0, Copyright (C) 2002-2005 Massachusetts Institute of Technology

        $MARX_DIR/marx/data/hrma/EKCHDOS06.rdb
Reading binary HRMA optical constants:
        $MARX_DIR/marx/data/hrma/iridium.dat
        $MARX_DIR/marx/data/hrma/corr_1.dat
        $MARX_DIR/marx/data/hrma/corr_3.dat
        $MARX_DIR/marx/data/hrma/corr_4.dat
        $MARX_DIR/marx/data/hrma/corr_6.dat
Reading scattering tables

(output truncated)

MARX does not output a FITS file. To create a FITS event file called radio.fits from the native MARX files in directory radio:

unix% marx2fits radio radio.fits
Examining radio/time.dat
Examining radio/detector.dat
Examining radio/energy.dat
Examining radio/b_energy.dat
Examining radio/xpos.dat
Examining radio/ypos.dat
Examining radio/zpos.dat
Examining radio/xcos.dat
Examining radio/ycos.dat
Examining radio/zcos.dat
Examining radio/xpixel.dat
Examining radio/ypixel.dat
Examining radio/pha.dat
Examining radio/mirror.dat
Examining radio/sky_ra.dat
Examining radio/sky_dec.dat

The resulting simulated FITS event file can be found here. It can be viewed using DS9 or other FITS image viewer. From this image we see that the bright point source and jet area clearly resolved. However, we can't be sure that the ratio of flux in the jet to flux in the nuclear point source is the same for the X-ray as it is for the radio. Therefore in the next set of simulations we assess the feasibility of a jet with reduced flux compared to the nuclear source.

MARX Simulations with Point Source Inputs

In this set of simulations we assume that the flux in the jet is only 1% of the nuclear source. We construct an image consisting of a nuclear point source and three point sources representing the knots in the jet. The overall flux in the jet is 0.0000305 photons/sec/cm^2. From the radio image, we see that the approximate peak fluxes in the knots scale as 14:9:9 for knot1:knot2:knot3. Here knot1 is the brightest knot closest to the nuclear source, knot3 the furthest. Assuming that the total flux scales as the peak flux we find knot1 has a flux of 1.31e-5 photons/sec/cm^2, knot2 and knot3 have a flux of 8.84e-6 photons/sec/cm^2. We use the radio image to measure the coordinates of the jet knots. Here is a summary of our composite source:

• Nuclear source, assumed to be point, 99% of total flux=0.003020 photons/sec/cm^2 RA=347.674 DEC=-43.7927
• Knot1, knot closest to nuclear source, flux=1.31e-5 photons/sec/cm^2, RA=347.6743 DEC=-43.7933
• Knot2, middle of jet, flux= 8.84e-6 photons/sec/cm^2, RA=347.6746 DEC=-43.79387
• Knot3, knot furthest from source, flux=8.84e-6 photons/sec/cm^2, RA=347.67489 DEC=-43.79452

To create this composite source we need to run MARX 4 times, once to create the nuclear source and 3 times for each of the knots. We then run the program marxcat to combine the 4 separate images.

The following set of commands copies over the default MARX parameter file to a file point.par, sets the critical parameters for the point source and runs the simulation. Note that unlike the "radio" simulation, we do not input a file for the source spatial distribution. We use the MARX internal POINT source model.

unix% cp $MARX_DIR/marx.par point.par
unix% pset point.par  OutputDir=point
unix% pset point.par   ExposureTime=40000
unix% pset point.par  DetectorType=ACIS-S
unix% pset point.par  SpectrumType=FILE
unix% pset point.par  SpectrumFile=jet.dat
unix% pset point.par   DitherModel=INTERNAL
unix% pset point.par  SourceFlux=0.0030789 
unix% pset point.par  GratingType=NONE
unix% pset point.par  SourceType=POINT 
unix% pset point.par  SourceRA=347.674
unix% pset point.par  SourceDEC=-43.7927 
unix% pset point.par  RA_Nom=347.674 
unix% pset point.par  Dec_Nom=-43.7927
unix% marx @@point.par TSTART=264038464.18

The following set of commands runs the simulation for knot1. Note that here the SourceRA and SourceDEC parameters have changed from the nuclear source simulation to reflect the location of the knot. The pointing direction is the same (RA_Nom and Dec_Nom unchanged).

unix% cp $MARX_DIR/marx.par knot1.par
unix% pset knot1.par OutputDir=knot1 
unix% pset knot1.par ExposureTime=40000 
unix% pset knot1.par DetectorType=ACIS-S
unix% pset knot1.par SpectrumType=FILE 
unix% pset knot1.par SpectrumFile=jet.dat 
unix% pset knot1.par DitherModel=INTERNAL
unix% pset knot1.par SourceFlux=1.31e-5
unix% pset knot1.par GratingType=NONE 
unix% pset knot1.par SourceType=POINT
unix% pset knot1.par SourceRA=347.6743 
unix% pset knot1.par SourceDEC=-43.7933 
unix% pset knot1.par RA_Nom=347.674 
unix% pset knot1.par Dec_Nom=-43.7927
unix% marx @@knot1.par TSTART=264038464.18

The parameter files for knot2 and knot3 are very similar to that for knot1. All the parameter files can be found here: knot1, knot2, knot3.

The composite image can now be created using the following commands:

unix% marxcat point knot1 knot2 knot3 jet
unix% marx2fits jet jet.fits

This combines the simulations in the directories point, knot1, knot2, knot3 into a directory jet. The command marx2fits then creates a FITS events file jet.fits from the native MARX files in the jet directory.

Inspection of this simulated image shows that the nuclear source and knots are still well resolved. Note that in order to view the knots, you will most likely need to use a log scale. You may notice other features in the image, such as streaks or bands. These are residual features of the input radio image and are not X-Ray events.

To obtain a more quantitative measure of the significance of the detection of the knots the user might want to run a source detection program such as celldetect. In addition, the jet might have some extended emission, and there might be extended emission from hot gas in the cluster. It might be useful to add an additional model component to simulate extended emission and show that it is still possible to detect the jet.

It is important to note that we have not included instrumental background in our simulations. To demonstrate why, we start by stating that in this thread, we have chosen to use a 1/8 subarray on the S3 chip (these choices are dicussed in the next section). Estimates of the background as a function of energy band are provided in the ACIS total background section of the POG. A table gives the total background rate for the S3 BI chip to be 0.74 counts/sec/chip, for the 0.3-10keV energy range, multiplied by the exposure time (40000 s), we find there will be 29600 cts over the entire S3 chip. We divide this by 8, to get that there will be 3700 cts over the 1/8 subarray. We scale this by a ratio of the jet are to subarray area to find the number of counts of background covering the area of the jet.

The subarray will contain 1024 pixels in each of 128 rows; converting this area to arcseconds results in an area of 30033.36 arcsec^2 (there are .492 arcsec/pixel). Since the jet only covers an area of about 10 arcsec by 3 arcsec, or 30 arcsec^2, we want to scale the background counts to this area. Consequently, we multiply the 3700 counts in the entire subarray by the ratio 30 [arcsec^2]/30033.36 [arcsec^2] to find there are 3.70 background counts over the same area as the jet. The jet contains in excess of 200 counts. If the background is greater than about 10% of the jet flux, the background counts should be included. In this case, the background is only about 2% of the jet flux, and therefore does not need to be included in our simulations.