skip to content
 
HEASARC: Software

 

XSTAR Home Page

XSPEC fit of XSTAR model
Two component XSTAR model fit to the 800 ks Chandra HETG observation of NGC 3783. Model fit performed in xspec11 using the warmabs analytic model.


What is XSTAR?
XSTAR is a computer program for calculating the physical conditions and emission spectra of photoionized gases.

How Do I Get XSTAR?
Xstar is avaible in several different ways.

  • The XSTAR package is now integrated and distributed with the HEASARC's standard HEADAS package, so if your facility has this package installed then you already have access to xstar. Xstar Version 2.1kn7 is part of the HEADAS release, December 2008.

  • XSTAR is also available as a standalone package. The standalone XSTAR (or the full HEADAS), is available in either source or binary versions and it can be downloaded from here .

  • In addition, Xstar has been revised to more accurately calculate the profile functions for emission and absorption lines. The code is available here.

  • Version 2.1ln10 of xstar is available. It contains the iron K data described in 2004 Ap. J. Supp. 155, 675, along with the line data for iron from Chianti 5 and all the features from version 2.1kxx. An error was found in version 2.1ln9 in the book keeping for some inner shell photoionization cross sections, resulting in double-counting in the opacity for some inner shell bound-free transions. These affect primarily low ionization models, and do not affect Fe or O. This has been fixed in this version of the code. This does not affect version 2.1kxx.

  • The release version of xstar is now available on hera. To use it, you must install fv and set up your environment to use it (fv is part of the standard heasoft software distribution or it can be installed standalone), and type 'fv -r xstar'. Input parameters are specified by prompting or else on the command line. Currently the output is buffered in large blocks, rather than printing each line of output as processing progresses. Thus it may appear that nothing is happening for a while until the buffer fills and is sent to your screen. It's also possible to start up the xstar task in hera, and then retrieve the results later, without having to keep the command window open. One would need to specify all the parameters directly on the command line, and also add a "-nowait " flag. For example: "fv -r -nowait me@milkyway.gsfc.nasa.gov xstar cfrac=1. temperature=400. ..." You will then receive an email message when the job is finished, with instructions on how to retrieve the output.

  • A utility has been created for extracting atomic data and rates from the xstar database. This utility is a standalone fortran program, which calls many of the xstar routines, and returns the rate (forward and reverse) and raw atomic data for a single type of rate for a single ion. It can be found here

  • An analytic model for calculating warm absorber and emitter spectra within xspec is available. This has the advantage that it avoids the intrinsic approximations associated with use of tables for absorption with variable abundances treated as multiplicative parameters, and the intrinsic clumsiness of the use of tables. It also allows the use of arbitrary spectral resolution, not limited by the internal xstar spectral resolution. To use it, download the tarfile, unzip and untar in your user area, and follow the installation instructions contained in the README. This includes the analogous routine which simulates the absorption (and emission) by a gas in coronal equilibrium. This also contains the updates to the line data from Chianti version 5, and also a finer grid of level populations as a function of ionization parameter, and it allows for the use of level populations calculated by xstar (version 2.1ln and later) for other choices of ionizing spectrum. Scattered emission can be modeled using the 'windabs' model. These routines are still in the testing phase, so please send reports of bugs and problems. NB: It is important to maintain the distinction between the atomic data file associated with warmabs (which is the same as xstar version 2.1ln and later) from that associated with the release version of xstar (version 2.1kn7). This is described in more detail in the README for warmabs.

    Online Documentation
    XSTAR documentation is included as part of the standard source distribution and is also viewable online or can be downloaded as a PostScript file or HTML.

    Related Files
    Sample xstar output from optically thin photoionized model, alpha=1 power law, line emissivities
    Sample xstar output from optically thin photoionized model, alpha=1 power law, line opacities
    Sample xstar output from optically thin photoionized model, alpha=1 power law, RRC emissivities
    Sample xstar output from optically thin photoionized model, alpha=1 power law, RRC opacities
    Sample XSPEC table models
    Line list.

    Line finding list.

    Level list.

    Subroutine xstar
    Postscript plots of atomic rates(42 Mbytes).
    An ascii version of the xstar database (28 Mbytes).

    An ascii version of the xstar database, version 2.1ln2, containing up-to-date K shell cross sections (63 Mbytes).

    The xstar database sorted according to process. (16 Mbytes).

    Sample output: Lexington H II region.

    Sample inputs.

    Atomic data.

    Platforms Currently Supported
    XSTAR requires a standard Fortran 77 compiler. It has been tested on the following systems:
    • SunOS & Solaris
    • DEC OSF
    • Mac OS X
    • Linux (both x86 & PPC)

    Conferences
  • For questions and comments about XSTAR, contact Tim Kallman or use the Feedback form

    The xstar database has been updated to take into account the iron M shell UTA data of Gu et al., 2006, 641, 1227. A revised database file for use with xstar21kn7 is available here, and a revised file has been included in version 2.1ln and later available here. This has also been included in the latest version of the source code for warmabs/photemis (see below)
  • Recent updates include fixes to bugs in the routines associated with xstar2xspec. These caused numerical problems on 64 bit machines, and also resulted in errors when large grids of models were run. Others include the dielectronic recombination rates published by Badnell 2006 Ap. J. Lett. 651, 73, and a fix to an error on the database which led to the lines for the Ca UTA being assigned one ion stage too high.  A bug has been found in version 2.1ln2 which affects the results in the paper 2004 Ap.J.Supp. 155 675. This is a bookkeeping error resulting in multiple-counting of the iron L shell cross section when calculating the cross sections for the 'third row' ions, Fe I -- VIII. This makes a quantitative change to the results in figures 4a and 13a. That is, it affects opacity due to iron above approximately 1 keV, only for low ionization models (log(xi)<0). These errors have been repaired in the version 2.1ln (and later) on this site, and repaired versions of the figures can be found here