Follow-on Science Instrument

Contract NAS8-01129

Monthly Status Report Numbers 023 & 024

January & February 2004

Science Theme: Source Models

Prepared in accordance with DR 972MA-002; DPD #972

Prepared for

National Aeronautics and Space Administration

Marshall Space Flight Center, Alabama 35812


Center for Space Research; Massachusetts Institute of Technology; Cambridge, MA 02139

Source Models Created and Used with HETG GTO Research


Summary of Source Model Activities


One way of demonstrating our understanding of the physics involved in an astronomical source is through the creation of a model of the source that reproduces the essential spatial, spectral, and temporal features that we observe. With Chandra there have been large improvements in the spatial resolution and spectral resolution (and the combined ability to have spatial-spectral resolution) of our data. (I leave temporal-spectral, etc. considerations out of this brief summary.) Consequently the amount of detail needed in our source models has increased as well. A rough summary of the details seen at low and high spatial and spectral resolution is given here:



Spatial resolution

Low: Just a single or perhaps a few resolution elements per main source feature. Requires simple basic and generally symmetric shapes to model, e.g., "beta-model" for galaxy cluster or a ring+shell model of a supernova remnant.

High: Many resolution elements per feature. Complexity in spatial structure is seen, e.g., knots, wisps, filaments, "holes", "bubbles", non-uniformities, proper motions, small faint features, etc.

Note: that the "spatial resolution" here, "Low" or "High", depends on the source angular size as compared to the telescope resolution; thus there have been "High" spatial resolution observations pre-Chandra and there still remain "Low" spatial resolution data even with Chandra!



Spectral resolution

Low, Medium: Of order CCD resolution. Allows the overall spectral shape, identification of the brightest emission lines/regions from different elements and some ions, detection of absorption edges, detection of line broadening/distortion effects and bulk Doppler shifts at the thousands of km/s level for high S/N cases.

High: Of order Chandra and XMM-Newton gratings resolution. Can clearly separate emission lines from different ionic species, separate He-like triplets, resolve RRCs, show structure in absorption edges, see weak lines, measure Doppler widths and velocities of hundreds of km/s.


Of course improved spatial resolution requires a higher geometric fidelity in the source model. But even for a source seen at "Low" spatial resolution, e.g., the case of an AGN or an XRB observed by Chandra, if it is seen at "High" spectral resolution there may be many signatures of the detailed source geometry imprinted and visible in the observed spectrum. Hence, the actual 3D geometry and dynamics of the source become more important for both higher spatial and higher spectral resolution.


Higher spectral resolution of course additionally puts greater demands on the accuracy of the atomic physics used to predict the emission, scattering, and absorption of X-rays. Combined with a more complex geometry, there is also a need to ensure self-consistency between local plasma properties, e.g. electron temperature and ionization state, and the local radiation environment.


XSPEC models (as also imported into or recreated in other packages like ISIS, Sherpa, etc.) are an off-the-shelf community set of models. For these, however, the underlying geometry is generally simple (a volume of plasma) or hardwired (a continuum-illuminated cone) and the model output is at "Low" spatial resolution, typically a single spatial bin. Spectrally the APED database of lines and continuum emissivities is used (high-resolution) and Doppler broadening and shifts can be included and multiple components summed; e.g., modeling of the SS 433 jets in partial eclipse was done in this way. But because of the simplicity of the database interface and source geometry assumed, the model spectra do not encode the full complexity of a realistic source, e.g., continuous parameter variation along a jet.


"Community" codes of greater spectral and geometric complexity are also growing in use, e.g., XSTAR and PHOTOION and are beginning to address some of these issues.


In the course of our HETG GTO research program, driven by high-resolution HETG spectra, we have also worked on advanced X-ray source models to better model the data. In the following pages some of these modeling efforts are briefly described.

Direct APED Database-access Plasma Model


Use: Determine temperature distributions and elemental abundances.


Geometry: DEM method assumes a sum of regions of different temp.s and a common abundance.


Spectral: APED database: lines plus continuum combined with more flexibility than XSPEC models currently allow. Line-based modeling.


Output: Synthetic continuua, DEM distribution and abundances


Coded in: ISIS, S-Lang.


Comments: This modeling is an example of using APED continuum emission models as an iterative baseline for fitting emission line fluxes. This is followed by the use of APED line emissivities to determine temperature distributions and elemental abundances. Finally, generation of synthetic spectra and observed counts using the multi-thermal, variable abundance plasma model (or even a pair for binary components at different radial velocities). Also some direct line-ratio fitting of the He-like triplets for density constraints.


Scientist: Dave Huenemoerder

Single-Zone Photoionization Model

Use: Applied to AGN, e.g., MR 2251, data modeling. Study the effect of (the unmeasured) EUV flux.


Geometry: Assumes a uniform plasma under illumination.


Spectral: Arbitrary ionization continua. Ionization rates, etc. from many sources.


Output: Heating rates, cooling rates, temperature, and ionization levels as a function of ionization parameter. Also ranges of ionization parameters and temperatures where a given ion is expected to be abundant.


Coded in: C++


Comments: A single-zone thin photo-ionized plasma model designed after the fashion of Krolik, McKee, and Tarter (1981). It includes the following heating processes: photo-ionization, Compton scattering, and Auger electrons. And the following cooling processes: recombination (including dielectronic), bremsstrahlung, and collisional line excitation (assumed from ground.) Collisional ionization is also modelled, as a contribution to ionization balances.


Scientist: Rob Gibson

Non-thermal Emission from Electron Distribution


Use: Model non-thermal emission see in SNe and SNR; SN 1006.


Geometry: Assumes a uniform volume of electrons.


Spectral: The spectral characteristics are calculated based on a given, arbitrary distribution of non- thermal electrons.


Output: Spectrum of non-thermal emission.







Coded in: C (comp. intensive) and S-Lang


Scientist: John Houck

Monte Carlo 3D Disk reflection model


Use: Accretion disk reflected spectra.


Geometry: General: array of grid cells with density, temp., and ionization state. Input from n-D hydro-code; e.g. 2D radiation-MHD calculations by Neal Turner. Input illumination given by position and direction distributions.


Spectral: Illuminated by power law. Detailed photon-tracing of through propagation, absorption, scattering, (re-)emission, and ionization effects.


Output: Reflected photons and their spectrum.




Coded in: C.


Comments: XSTAR pre-computed results are used to map the local ionization parameter to ionization fraction and temp. of each cell. Ultra-rough draft of paper in process...


Scientist: Andy Young.

Disk Atmosphere and Corona Emission

Use: Model disk/corona emission in LMXBs containing neutron stars; Model narrow-line Seyfert I galaxies containing, e.g., a super-massive Kerr Blackhole, MCG --6-30-15; NGC 4051; Mkn 766.


Geometry: 2D disk: height(r), to 3D: azimuthally symmetric. Orbital velocities are included. Centrally illuminated (Neutron star.) "Self illuminated": assume non-thermal X-rays are produced uniformly above the disk using the disk's own energy.

Spectral: Semi-Analytic or Monte Carlo spectral model used. Kerr metric relativistic effects and Doppler shifts are included for BH simulation.


Output: Continuum spectrum, line spectrum; 3D structure with charge state distribution, Temp., density.


Coded in: FORTRAN, IDL, S-Lang.


Comments: Using Raymond code for ionization balance, HULLAC recombination rates, and own structure and spectral calculations.


Scientist: Mario Jimenez-Garate

Recombination and Resonance Scattering Models


Use: Model Seyfert II ionization cone; wind of super-soft source, CAL 87.


Geometry: Single or multiple ionization zones; arbitrary fixed covering fraction.


Spectral: Illumination by a BB or PL continuum, e.g., white dwarf or Seyfert, respectively.


Output: Binned spectra of components: resonance scat., radiation recombination lines, and RRCs.


Coded in: IDL.


Comments: Uses XSTAR for ionization balance, and HULLAC recombination rates.


Scientist: Mario Jimenez-Garate

3D Model Projected to Sky w/Doppler Shifts



Use: SNR spatial-spectral models especially for E0102.


Geometry: General: given by 3D array of "norm". Routines allow easy input of simple shapes.


Spectral: Emission is spectrally colored by: i) single line w/Doppler shifts, or ii) a MARX-assigned spectrum.

Output: Events file for input to MARX ray-trace.


Coded in: IDL; uses S-Lang user interface to MARX .


Comments: Reproduces qualitative E0102 plus/minus order differences in Ne X line. Multiple components can be combined. This is an exploratory prototype...


Scientist: Dan Dewey