next up previous contents
Next: 5. ``README FIRST" for Up: Suzaku ABC Guide Previous: 3. Suzaku Data Specifics   Contents

Subsections


4. Suzaku Data Analysis Overview

This chapter provides a brief outline of the standard analysis steps. Details are explained in subsequent chapters 5-7. Analysis topics covered in this chapter are:

  1. Spectral analysis of the XIS and the HXD data.
  2. Timing analysis of the XIS and the HXD data.
  3. Imaging analysis of the XIS data.

We assume that the user has downloaded and decrypted the Suzaku data, and has access to the latest versions of the Suzaku FTOOLS and CALDB.

4.1 Important Events and Processing Version

Users should check

http://www.astro.isas.jaxa.jp/suzaku/log/operation/,
http://www.astro.isas.jaxa.jp/suzaku/log/hxd/, and
http://www.astro.isas.jaxa.jp/suzaku/log/xis/ for any important events (operational or instrumental issues) that may affect your specific observation.

Users should also check the processing version of the data, as recorded in the PROCVER keyword in any of the FITS files. Users should then consult
http://heasarc.gsfc.nasa.gov/docs/suzaku/aehp_proc.html to see if there are any issues for data processed using that version of the pipeline.


4.2 Checking for Updates

Users should check for any software or calibration updates that may affect the data in question. The Suzaku GOF disseminates information in several ways.

  1. By targeted e-mails to the PIs of proprietary data. This method is used to communicate important updates that affect a specific subset of observations. We expect the PIs to pass the information to any collaborators who may be analyzing the data.

  2. By e-mails to the suzakunews exploder. All types of Suzaku related news items are included in occasional messages through this exploder. All who are interested in the Suzaku mission should subscribe to this list at
    http://heasarc.gsfc.nasa.gov/docs/heasarc/news.html. if they are not already on the list. Note that users who are registered under a defunct address will get dropped if the exploder receives error messages several times in a row. On the other hand, after every proposal review cycle, we add the e-mail addresses of all successful US PIs who appear not to be on the suzakunews list.

  3. Through this guide. Chapter 5 provides a list of important caveats that users should be aware of.

  4. Via the Suzaku GOF web site. The Suzaku Data Analysis page:
    http://heasarc.gsfc.nasa.gov/docs/suzaku/aehp_data_analysis.html and links therein are updated more frequently than this guide. Of particular importance is the ``Things to Watch Out For'' page:
    http://heasarc.gsfc.nasa.gov/docs/suzaku/analysis/watchout.html.

These updates may simply note a newly discovered calibration problem, instruct users on how to obtain a software patch or updated calibration file, or provide workarounds such as an ad-hoc procedure that should be run on specific datasets.

4.3 Unfiltered or Cleaned?

The pre-extracted spectra and light curves in the products subdirectories, both for the XIS and the HXD, are provided for quick-look purposes only. We recommend against using these files for actual data analysis.

Event FITS files are provided to the users in two flavors (chapter 3), unfiltered (event_uf) and cleaned ( event_cl). The latter were obtained using the pre-determined screening criteria given in Table 6.2 (for the XIS) and Table 7.2 (for the HXD). Although this procedure can never be perfect for each individual observation it generally produces reliable event lists.

If stronger screening criteria are required, the cleaned events can be further screened by the user to produce the desired new cleaned event files, see sections 6.3, 6.4, and 6.11 for the XIS and sections 7.3, 7.4, 7.6.3, and 7.12.1 for the HXD.

For less strong screening criteria, one has to start with reprocessing the unfiltered data. This is also the case if one wants to ensure that the newest calibration is used for producing the event lists. If modeled Non X-ray Background (NXB) events provided by the HXD team for the GSO are used in the analysis, it has to be kept in mind, however, that reprocessing of the source events has to be performed with the same gain calibration as used for the creation of those background events (see next paragraph; for the PIN this is not relevant since only one realization of the gain file has been in use). The reprocessing procedures are described in sections 6.3, 6.4, and 6.10 for the XIS and in sections 7.3, 7.4, 7.6.3, and 7.11 for the HXD.

For XIS data produced with a processing pipeline version earlier than Version 2.1.6.16 reprocessing is mandatory, as well as for PIN data taken after 2007 July 28 and processed with pipeline Version 2.0.6.13 or earlier. For GSO data users should check the GSOGPT_F and GSOGPT_V keywords in the FITS headers of the cleaned event (observed) and background (simulated) files. These keywords define which HXD GSO gain parameter table (GPT) was used in the processing. If the values match then there is no need to reprocess the GSO data (note: the files ae_hxd_gsogpt_20091225.fits and ae_hxd_gsogpt_20100323.fits are identical in content). If the values do not match the data have to be reprocessed. If these keywords are not present the data were processed with pipeline Version V2.4.12.27 or earlier corresponding to a different way of performing GSO calibration (using GHT [gain history table] instead of GPT type gain files) and should be reprocessed, preferentially using the new setup. For more details on the two different GSO setups see http://heasarc.gsfc.nasa.gov/docs/suzaku/analysis/gso_newgain.html.

4.4 Barycentric Correction

Aebarycen is a Suzaku specific FTOOL to perform barycentric corrections for Suzaku event and housekeeping files. See Terada et al., 2008, PASJ 60, S25 for a brief description of the barycentric correction tool (and a detailed description of the timing calibration of the HXD).

Warning: In order for the barycentric correction tool to work correctly it is required to perform any good time interval filtering in xselect, i.e., removing times of telemetry saturation, before running aebarycen.

4.5 Spectral Analysis

The following are the steps, explained in detail in chapter 6, for the spectral analysis of XIS data:

  1. Extract source and background spectra for the desired extraction regions.
  2. Build the response files (RMF and ARF).
  3. Combine the spectra taken with XIS0, XIS2 (if available), and XIS3.

Chapter 6 contains pointers on the size and shape of the source extraction region, and on the typical background extraction region for a point source. The particle background is a strong function of the location of the Suzaku spacecraft within the geomagnetic field, and therefore is variable in time. The X-ray background is a function of the pointing direction. Therefore, ideally, the background spectrum should be extracted from the neighboring source-free region(s) of the same CCD chip from the same observation. However, for the analysis of extended sources, it is sometimes necessary to consult other observations, including the non X-ray background (NXB) database compiled by the XIS team.

We also explain the RMF and ARF generators. The latter is based on ray-tracing and can be extremely time-consuming. We therefore describe several options for speeding up ARF generation, as well as subsequent spectral fits.

Three of the XIS units (two, after the damage to XIS2 occurred), each with a frontside illuminated (FI) chip, are sufficiently similar that it is usually safe to sum the spectra from these units. However, we never recommend combining the event files (this will lead to the loss of information critical to downstream software). Also, XIS1, with a backside illuminated (BI) chip, has a distinctly different response and so XIS1 data should not be combined with those from other units.

The following are the steps, explained in detail in chapter 7, for the spectral analysis of HXD data:

  1. Obtain the appropriate background files.
  2. Create a joint good time interval (GTI) file, when both the data and background models are available.
  3. Extract spectra from the observation data and correct for deadtime.
  4. Extract spectra from the background files and correct for the factor of 10 oversampling (PIN) and for deadime (GSO).
  5. Select the appropriate response file and correct for off-axis location of the source if necessary.

Since the HXD is a non-imaging instrument, users need not/cannot consider ``extraction regions''. Instead, it is necessary to subtract the particle and X-ray background from the observation data. This is done using the background files generated by the HXD team, who model the particle background based on the orbital location and other information. In normal situations, the PIN background files are prepared within a few weeks of the distribution of the processed data to the PI; GSO background files are made about a month after distribution.

There is a noticeable deadtime even for faint sources because the particle background alone results in a high count rate. This must be corrected for in the data. The PIN background files, on the other hand, do not need a deadtime correction. However, in the case of the PIN, background files have an artificially inflated count rate to ensure sufficient statistical accuracy, and this has to be taken into account.

The HXD team provides the response files for the PIN and GSO, rather than response generators. The PIN settings have been adjusted since the initial operation several times, including changes in the bias voltage used on-board, and in the low energy thresholds used in ground processing. Therefore it is necessary to select a response file appropriate for the epoch of the observation.

4.6 Timing Analysis

The following are the steps for the timing analysis of XIS data:

  1. Extract source and background light curve for the desired extraction region.
  2. Subtract the latter from the former with appropriate scaling.
  3. Combine the light curves taken with the XIS FI units.

Users who have become familiar with XIS spectral analysis should find little difficulties in performing XIS timing analysis.

The following are the steps for the timing analysis of HXD data:

  1. Obtain the appropriate background files.
  2. Create a joint good time interval (GTI) file, when both the data and background models are available.
  3. Extract light curves from the observation data, and correct for deadtime. Keep in mind that the minimum recommended time resolution for deadtime corrected HXD lightcurves is 128s.
  4. Extract light curves from the background files and correct for the factor of 10 oversampling (for PIN) and for deadtime (for GSO).

Again, this process parallels that of the spectral analysis, but requires the correction of time-variable deadtime, as explained in chapter 7.

4.7 Imaging Analysis

The following are the steps for the imaging analysis of XIS data:

  1. Extract XIS images for the desired extraction regions.
  2. Generate corresponding exposure maps.
  3. Create exposure-corrected XIS images, and apply further arithmetic as desired.

Chapter 6 contains a detailed description on how to generate exposure maps.

4.8 Recipe 1: XIS Spectral Analysis

  1. Make sure that you have access to the latest software and calibration files (chapter 2).

  2. Download the data (chapter 3).

  3. Check for any updates (this chapter).

  4. Perform reprocessing (update energy calibration) and rescreening using aepipeline, if required (section 6.4).

  5. Use xselect to make, display and save products from cleaned event files.
    Refer to the xselect User's Guide at
    http://heasarc.gsfc.nasa.gov/docs/software/lheasoft/ftools/xselect/xselect.html for available commands. For example: 

    xsel:SUZAKU-XIS-1-STANDARD > read event
  > Enter the Event file dir >[] .
  > Enter Event file list >[] ae101005070xi0_0_3x3n066z_cl.evt
xsel:SUZAKU-XIS-1-STANDARD > filter region
xsel:SUZAKU-XIS-1-STANDARD > extract all
xsel:SUZAKU-XIS-1-STANDARD > plot image
xsel:SUZAKU-XIS-1-STANDARD > plot curve
xsel:SUZAKU-XIS-1-STANDARD > plot spectrum
xsel:SUZAKU-XIS-1-STANDARD > save spectrum

  6. Build the response files, using xisresp (see chapter 6).

  7. Use addascaspec to combine FI spectra and responses. 

    example% addascaspec fi.add fi.pha fi.rsp fi_b.pha

    (see chapter 6).

4.9 Recipe 2: HXD Spectral Analysis

Note: The steps described below to produce PIN and GSO spectra can also be performed by applying the FTOOLS hxdpinxbpi (section 7.5.6) and hxdgsoxbpi (section 7.7.2) for PIN and GSO, respectively. Similar extraction tools exist to produce lightcurves: hxdpinxblc (section 7.9.1) and hxdgsoxblc (section 7.10.1).

  1. Make sure that you have access to the latest software and calibration files (chapter 2).

  2. Download the data (chapter 3).

  3. Check for any updates (this chapter).

  4. Perform reprocessing (update energy calibration) and rescreening using aepipeline, if required (section 7.4 and 7.6.3).

  5. Determine if the observation was performed using the XIS nominal or the HXD nominal pointing position (PIN & GSO). Determine the epoch of the observation (PIN only, chapter 7).

  6. Obtain the appropriate response files -- one for the source (PIN & GSO), and one for the Cosmic X-ray background (``flat'', PIN only) -- from the CALDB, as well as, for the GSO correction arf, from
    http://heasarc.gsfc.nasa.gov/docs/suzaku/analysis/gso_newarf.html.

  7. Obtain the non X-ray background (NXB) files -- see
    http://heasarc.gsfc.nasa.gov/docs/suzaku/analysis/pinbgd.html and
    ftp://heasarc.gsfc.nasa.gov/suzaku/data/background/gsonxb_ver2.4/.

  8. Generate the GTI file common to data and NXB using mgtime.

  9. Apply GTI and extract spectra in xselect.
    Refer to the xselect User's Guide at
    http://heasarc.gsfc.nasa.gov/docs/software/lheasoft/ftools/xselect/xselect.html for available commands.

    The name of the ``*_cl2.evt.gz'' files referred to in the following indicates that not the original cleaned files are used but cleaned files that were produced by reprocessing and rescreening the original ``*_uf.evt.gz'' files (see chapter 7). 

    xsel > read event
xsel > ./
xsel > ae101005040hxd_0_pinno_cl2.evt.gz
xsel > filter time file ae101005040hxd_wel_pin.gti
xsel > extract spec
xsel > save spec ae101005040hxd_0_pinno_cl2.pha
xsel > clear all
xsel > yes
xsel > read event
xsel > ./
xsel > ae101005040hxd_0_pinbgd.evt.gz
xsel > filter time file ae101005040hxd_wel_pin.gti
xsel > extract spec
xsel > save spec ae101005040hxd_wel_pin_bgd.pha
xsel > exit

    This creates the source spectrum (ae101005040hxd_0_pinno_cl2.pha) and background spectrum (ae101005040hxd_wel_pin_bgd.pha).

    The procedure for the GSO data is the same and furthermore the GSO spectrum should be rebinned to at least the 64 bins that are used to create the GSO background. This grouping is available at
    http://heasarc.gsfc.nasa.gov/docs/suzaku/analysis/gsobgd64bins.dat. 

    unix% grppha ae101005040hxd_gsono_cl2.pha ae101005040hxd_gsono_cl2bin.pha
GRPPHA[] group gsobgd64bins.dat
GRPPHA[] exit

  10. Correct for deadtime 
    unix% hxddtcor event_fname="ae101005040hxd_0_pse_cl.evt" \
           pi_fname="ae101005040hxd_0_pinno_cl2.pha"

    This is to be done for the PIN source spectrum (see above) as well as for the GSO source and background spectrum. Note that the deadtime effect in the PIN background spectra is already taken into account, please do not apply this procedure to background spectra.

  11. Correct the exposure of the PIN background (this step is not required for the GSO) by multiplying it by a factor of 10: 
    unix% cp ae101005040hxd_wel_pin_bgd.pha \
           ae101005040hxd_wel_pin_bgd_expcor.pha
unix% fkeyprint infile=ae101005040hxd_wel_pin_bgd_expcor.pha keynam=EXPOSURE

----- output  ---
# FILE: ae101005040hxd_wel_pin_bgd_expcor.pha
# KEYNAME: EXPOSURE

# EXTENSION:    0
EXPOSURE= 1.755875832736492E+03 / Exposure time
# EXTENSION:    1
EXPOSURE= 1.755875832736492E+03 / Exposure time
# EXTENSION:    2
EXPOSURE= 1.755875832736492E+03 / Exposure time
-----------------

unix% fparkey value=1.755875832736492E+04 \
        fitsfile="ae101005040hxd_wel_pin_bgd_expcor.pha+0" keyword=EXPOSURE
unix% fparkey value=1.755875832736492E+04 \
        fitsfile="ae101005040hxd_wel_pin_bgd_expcor.pha+1" keyword=EXPOSURE
unix% fparkey value=1.755875832736492E+04 \
        fitsfile="ae101005040hxd_wel_pin_bgd_expcor.pha+2" keyword=EXPOSURE

next up previous contents
Next: 5. ``README FIRST" for Up: Suzaku ABC Guide Previous: 3. Suzaku Data Specifics   Contents
Katja Pottschmidt 2013-09-04