Follow-on Science Instrument

Contract NAS8-01129


Monthly Status Report No. 012

February 2003

HETG Science Theme: the Inter-Stellar Medium


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


Inter-Stellar Medium Research Progress


X-ray Absorption in the Interstellar Medium


The space between stars is an ultrahigh vacuum by our laboratory standards but none-the-less contains gas, molecules and even dust -- the interstellar medium. The source of this stuff is most likely the expulsion of matter into space by supernova explosions (wisps in the figure here) and possibly other phenomena, e.g., jets, stellar winds and iteractions, etc. The following discussion and summary of ISM work is taken from A. Juett et al. “X-ray Absorption in the Interstellar Medium”, a poster presented recently at the AAS/HEAD ’03 meeting.



Among the most prominent features in the high-resolution spectra of bright continuum sources (e.g., X-ray binaries) are the photoelectric absorption edges and lines due to metals in the interstellar medium (ISM). The K shell absorption edges for Z = 8--26 (O through Fe) are all located in the 0.5--8~keV (1.5—25) energy range accessible to Chandra and XMM-Newton. The depths of these edges allow us to measure the column densities and relative abundances of these elements. These measurements provide independent estimates of ISM abundances that can be compared with measurements in the optical and UV (see e.g., Sofia & Meyer 2001, and references therein).


The high-resolution capabilities of Chandra and XMM also allow for studies of the structure of the ISM edges. X-ray absorption by the ISM is due to gas, molecules, and dust. Wilms, Allen, & McCray (2000) provided an updated absorption model for use with CCD resolution spectra, but the authors stress that their model, and similar models, are not appropriate for high-resolution grating spectra. The main cause for this is the lack of high-resolution cross-section measurements for atomic and molecular (dust) edges. Photoelectric absorption of an X-ray photon leads to inner shell

excitations, leaving the atom in an excited state. When the element is in atomic form, there are various X-ray resonance complexes which can be resolved, specifically resonances from the 1s-2p transitions of various ions. Once the element is bound in a molecule, energy levels, and therefore the line energies, may shift due to the change in binding energy. In addition, elements bound into dust grains show modification of their absorption edge energy and structure due to elastic scattering off the molecular structure or crystal lattice of solids. This modification is the cause of X-ray absorption fine structure (XAFS; see Woo 1995 and references therein) and particularly the so-called XANES (X-ray Absorption Near Edge Structure; Kossel 1920) which introduces significant structure to the edge. Many of the cosmically abundant elements, e.g., C, O, Si, and Fe, have large measured depletions (the fractional amount of the element thought to be in dust grains) and should therefore show XAFS in high-resolution spectra.



Behar, E. & Netzer, H. 2002, ApJ, 570, 165

Henke, B. L., et al. 1993, Atomic Data & Nucl. Data Tables, 54, 181

Kortright, J. B., & Kim, S.-K. 2000, Phys. Rev. B, 62, 12216

Kossel, W. 1920, Z. Phys., 1, 119

Pradhan, A. K. et al. 2003, MNRAS, in press (astro-ph/0302238)

Schulz, N.S., Juett, A.M., Chakrabarty, D., and Canizares, C.R., "X-ray Absorption in the ISM", 2003, AN, vol. 324, p. 166

Sofia, U. J. & Meyer, D. M. 2001, ApJ, 554, L221

Wilms, J. Allen, A., \& McCray, R. 2000, ApJ, 542, 914

Woo, J. W. 1995, ApJ, 447, L129

Summary of Inter-Stellar Medium GTO Observations and Activities


Recently we have used photoelectric absorption features in the Chandra/HETGS spectra of seven bright X-ray binaries to study the relative abundances of O, Fe, and Ne in the interstellar medium (ISM). Three of these sources are in our GTO program: 4U 1636-53, Cyg X-2, and GX 349+2. We find Ne/O abundance ratios up to twice as large as predicted by standard ISM models, and Fe/O ratios of 0.6--1 times the predicted ratio. The shape and wavelength of the photoelectric edges can be used to discriminate between atomic and molecular forms of these elements. For example, we find that the O edge contains complex structure and measured edge wavelengths range from 22.89--23.12 . Through these measurements, we hope to better understand the line of sight variations in the abundances and to produce a universal absorption edge model for use by the community.



Oxygen Absorption


At right is the counts spectrum of the O-K absorption in Cyg X-2. The absorption line at 23.5 is due to the O I 1s-2p transition in the ISM. The line at 23.36 is due in part to the instrument, and in part to the O II 1s-2p transition.


Neon Absorption


The Ne-K edge shows variations in its shape, with some sources having a well defined edge and others not; the edge from GX 349+2 is shown at left. Ne/O abundance ratios range from 0.7--2.2 times the ISM abundances given by Wilms et al. (2000). Such variations suggest line-of-sight variations in abundances, although there are large uncertainties in these measurements We note that for many of these sources, the Ne edge in the MEG minus first order lies in a chip gap where the effective area is poorly determined. This effect is most important for the lowest signal-to-noise observations. Better observations without chip gaps at the Ne edge are needed to reconcile real Ne/O ratio variations from instrumental effects.


Most sources show at least one accompanying absorption line at 14.61 (as shown here), possibly from the 1s-2p transition of Ne ions, comparable to that seen in O. Behar & Netzer (2002) recently calculated wavelengths and oscillator strengths for the 1s-2p transitions of Li-like to F-like ions of Ne and other elements. Comparing our lines to those of Behar & Netzer (2002), we find that if the 14.61 line is the F-like transition (predicted at 14.631 ), then there is a wavelength offset of 0.02 . Given this shift, the O-like transition would be expected at 14.506 , and there is evidence in some of the sources for a line-like feature near this wavelength.


Iron (Fe-L) Absorption




At right is the flux spectrum of Fe-L absorption edge from 4U 1636-53. The Fe-L III edge at 17.5 is due to transitions from the 2p_{3/2} level, while the Fe-L II edge at 17.2 is due to transitions from the 2p_{1/2} level and varies from source to source. A third transition is also possible, the Fe-L I edge at 14.6 (weak and not in the range shown), due to transitions from the 2s level, but this is the weakest edge.

Because this structure is not accounted for in the commonly used optical constants and absorption models (Henke et al. 1993; Wilms et al. 2000), we fit the Fe edges here using a custom table model for XSPEC and ISIS that uses the optical constant measurements of Kortright & Kim (2000). This model will be made publicly available upon publication of our initial results (Juett, et al. 2003, in prep).


Future Work in ISM


      Continue analysis of O K, Fe L, Ne K with respect to determining the atomic fine structure.

      Extend sample of observations towards lower column densities.

      Observe Ne edges out of the chip gap.

      Correlate NH parameters with optical, UV and radio data.

      Analyze higher energy edges, Mg K, Si K, S K ..., at high column densities.

      Generate a generic model spectrum for high resolution spectral analysis use by the community.