Japanese version is here
Abstract of Doctor Thesis (pdf)
"Iron Line Analysis of Clusters of Galaxies
and the Effect of Resonance Scattering"
Clusters of galaxies are the largest observed objects in the sky which
are gravitationally bound, and the brightest X-ray emitters. Observed
X-ray spectra of clusters of galaxies consist of continuum emission
components and lines of highly ionized heavy elements from a thin hot
plasma in collisional ionization equilibrium (CIE). The existence of
lines revealed that the intracluster medium (ICM) is a mixture of
metal-poor primordial gas and metal-rich processed gas. The
distributions of heavy elements from line analysis is a clue to
investigate the chemical evolution of galaxy clusters. There are such
schemes as which heavy elements are ejected into the interstellar
space through supernova or stellar mass loss and injected into the
intra-cluster space through ram-pressure stripping or galactic winds.
The ICM has a small optical depth, that is, optically thin for
continuum because the Thomson scattering optical depth
tauTh is much less than unity, typically 10-2.
However this is not the case for the stronger resonance lines. A
resonance-line photon is scattered by the ions in the same state of
ionization as an ion emitting the photon, and has relatively large
absorption oscillator strength (e.q. f12 > 0.1). Then, it
is re-emitted, generally in a different direction. Therefore, the
optical depth of the resonance lines is hundreds or thousands times
larger than the tauTh.
The resonance lines are scattered mainly in the center of a cluster
where the density is high, and are observed to be suppressed. The
diffusion of the photons with respect to frequency tends to form the
saddle-shaped line profiles. On the other hand, the suppressed
component is observed to be a significant enhancement in outer part of
cluster. Eventually the total intensity over the whole cluster is not
changed.
149 clusters of galaxies, Virgo Cluster and Perseus Cluster and so on
were observed with ASCA satellite, which is capable for imaging
spectroscopy. From the spectra integrated over annual rings whose
center is located at the cluster center, temperatures are confirmed to
be lower at the center of many nearby clusters (redshift < 0.1) with
factor < 2, for example, A496 and Perseus cluster. On the other hand,
some of distant clusters (redshift > 0.1) are suggested to have radial
temperature decline with factor less than 2, for example, A2204 and
A1689. Based on the line features, especially iron K alpha line,
abundances are enhanced at the center of A496, A3158, M87 and Perseus
cluster and so on.
Among the observed lines, iron K beta lines are found separately from
K alpha lines by the GIS detectors, and He-like and H-like iron K
alpha lines can be distinguished by at SIS detector. I called them K
beta-Blend, K alpha-Blend, He-K alpha-Blend and H-K alpha-Blend.
However K beta-Blend includes Ni K alpha lines. I estimated the
intensities and the ratios (K alpha-Blend / K beta-Blend = R1, He-K
alpha-Blend / H-K alpha-Blend = R2) of these lines. According to the
results of GIS analysis, 80 % of clusters suggest that R1 of cluster
center tends to be smaller than that expected from optically thin
plasma model and to increase as the radius. The results of SIS
analysis show that R2 of each cool clusters (kT < 6 keV) is smaller
than the model value. Furthermore, in order to get the sufficient
photon statistics, 83 bright clusters of galaxies are divided into 4
groups with respect to temperature (1-3 keV, 3-6 keV, 6-9 keV, > 9
keV), all spectra of each groups are summed up, after the correction
of the redshift. These spectra make it clear that these trends of
ratios, R1 and R2, are univeral for clusters of galaxies. These
results are consistent with the effect of resonance scattering.
On the other hand, He-K beta-Blend and H-K beta-Blend are separated in
the statistical line analysis of SIS spectra. As the three new ratios
(He-K beta-Blend / H-K beta-Blend = R3, He-K alpha-Blend / He-K
beta-Blend = R4 and H-K alpha-Blend / H-K beta-Blend = R5) were also
obtained, the line analysis could be done in detail.
In order to understand these observed ratios, I simulated 41 lines
(He-K alpha-Blend and H-K alpha-Blend include 7 and 16 lines, and K
beta-Blend consists of 18 lines.) in energy range between 6.4-8.2 keV
in consideration of the effect of the resonance scattering, using
Monte Carlo method. Then, I got the line intensity ratios R1 and R2,
and compared them with observed line intensity ratios.
As a consequence, observed R1 and R2 of hot clusters (kT > 6 keV) can
be explained by the resonance scattering. The obtained tau of He-like
iron K alpha line of cluster center is 2-4. This indicates abundances
estimated from K alpha lines without considering the resonance
scattering is about factor 2 smaller than real values at the cluster
center.
Observed R1 and R2 of the cool cluster (kT < 3 keV) are smaller than
the values obtained by simulation of the resonance scattering, though
the trend is consistent with the effect of the resonance scattering.
Moreover the R1 estimated from spectrum of the whole cluster is also
smaller than the model ratio. It is needed to consider other causes,
for example, some more characteristics of cluster such as size or to
reform the atomic data. Especially, as one of them, the contamination
fo Nickel K alpha lines for cool cluster are considered. The R3, R4
and R5 ratios suggest the existence of hot plasma or contamination of
K gamma and K delta.