A Suzaku Observation of MCG-2-58-22: Constraining the Geometry of the Circumnuclear Material

Abstract

We have analyzed a Suzaku long-look of the active galactic nucleus
MCG-2-58-22, a type 1.5 Seyfert with very little X-ray absorption in the line
of sight and prominent features arising from reflection off circumnuclear
material: the Fe line and Compton reflection hump. We place tight constraints
on the power law photon index (Gamma=1.80+/-0.02), the Compton reflection
strength (R=0.69+/-0.05), and the Fe K emission line energy centroid and width
(E=6.40+/-0.02 keV, v_FWHM < 7100 km/s). We find no significant evidence for
emission from strongly ionized Fe, nor for a strong, relativistically broadened
Fe line, indicating that perhaps there is no radiatively efficient accretion
disk very close in to the central black hole. In addition we test a new
self-consistent physical model from Murphy & Yaqoob, the "MYTorus" model,
consisting of a donut-shaped torus of material surrounding the central
illuminating source and producing both the Compton hump and the Fe K line
emission. From the application of this model we find that the observed spectrum
is consistent with a Compton-thick torus of material (column density
NH=3.6(+1.3/-0.8) x 10^24 cm^-2) lying outside of the line of sight to the
nucleus, leaving it bare of X-ray absorption in excess of the Galactic column.
We calculate that this material is sufficient to produce all of the Fe line
flux without the need for any flux contribution from additional Compton-thin
circumnuclear material.

Full text

Page 1

arXiv:1107.0070v1 [astro-ph.CO] 30 Jun 2011
Draft version July 4, 2011
Preprint typeset using LATEX style emulateapj v. 11/10/09
A SUZAKU OBSERVATION OF MCG–2-58-22:
CONSTRAINING THE GEOMETRY OF THE CIRCUMNUCLEAR MATERIAL
Elizabeth Rivers1, Alex Markowitz1, Richard Rothschild1
Draft version July 4, 2011
ABSTRACT
We have analyzed a Suzaku long-look of the active galactic nucleus MCG–2-58-22, a type 1.5 Seyfert
with very little X-ray absorption in the line of sight and prominent features arising from reflection
off circumnuclear material: the Fe line and Compton reflection hump. We place tight constraints on
the power law photon index (Γ=1.80±0.02), the Compton reflection strength (R=0.69±0.05), and
the Fe K emission line energy centroid and width (E=6.40±0.02 keV, vFWHM<7100 kms−1). We
find no significant evidence for emission from strongly ionized Fe, nor for a strong, relativistically
broadened Fe line, indicating that perhaps there is no radiatively efficient accretion disk very close
in to the central black hole. In addition we test a new self-consistent physical model from Murphy
& Yaqoob, the “MYTorus” model, consisting of a donut-shaped torus of material surrounding the
central illuminating source and producing both the Compton hump and the Fe K line emission. From
the application of this model we find that the observed spectrum is consistent with a Compton-thick
torus of material (column density NH=3.6+1.3
nucleus, leaving it bare of X-ray absorption in excess of the Galactic column. We calculate that this
material is sufficient to produce all of the Fe line flux without the need for any flux contribution from
additional Compton-thin circumnuclear material.
Subject headings: Galaxies: active – X-rays: galaxies – Galaxies: Individual: MCG–2-58-22
−0.8×1024cm−2) lying outside of the line of sight to the
1. INTRODUCTION
MCG–2-58-22 is an X-ray bright Seyfert 1.5 active
galactic nucleus (AGN) located at a redshift of z =
0.04686. Past X-ray observations of this source per-
formed with EXOSAT, ASCA, XMM-Newton and Bep-
poSAX have revealed the following spectral components
in addition to the primary X-ray power law: a soft excess,
Fe emission lines, and a Compton reflection hump. Im-
portantly, there has been no evidence for X-ray absorp-
tion by gas along the line of sight in excess of the Galactic
column, indicating that in the X-ray band this AGN is a
“bare nucleus”. This combination makes MCG–2-58-22
an interesting target of study, since the lack of signifi-
cant X-ray absorption provides a clean view of the nu-
cleus with a relatively simple spectum to model, while
the presence of strong reflection components allows us to
place constraints on the physical geometry of the circum-
nuclear material surrounding the AGN.
Ghosh & Soundararajaperumal (1992) analyzed EX-
OSAT data obtained in 1984 that revealed the highly
variable soft excess below about 2 keV. They modeled
this component with a steep power law in addition to
their continuum power law. ASCA data covering 2.5–10
keV with good CCD resolution were analyzed by Weaver
et al.(1995). They modeled the spectrum using a hard
X-ray power law (Γ = 1.75±0.05) with Galactic absorp-
tion and confirmed the need for a soft excess, as well as
an Fe Kα emission line which was unresolved. Weaver,
Gelbord & Yaqoob (2001) analyzed two additional ASCA
observations of MCG–2-58-22, tracking the Fe line flux
over a time scale of years and showing large variation
erivers@ucsd.edu
1University of California, San Diego, Center for Astrophysics
and Space Sciences, 9500 Gilman Dr., La Jolla, CA 92093-0424,
USA
in the flux of the underlying continuum. However large
uncertainties precluded definitive conclusions about the
variation of the Fe line parameters.
A more recent analysis by Bianchi et al.(2004) using
simultaneous data from XMM-Newton and BeppoSAX
covered a much broader energy range than previous ob-
servations (0.5–200 keV). Unfortunately only 7 ks of good
EPIC-pn data were obtained for the source, providing
only loose constraints in the Fe K bandpass (the Fe Kα
line was unresolved with σ < 340 eV and an equivalent
width, EW, of 45+85
−24eV). They were able to loosely con-
strain the Compton reflection hump with R = 0.4 ± 0.3
and Γ = 1.72+0.08
−0.06.
In this paper we present an in-depth analysis of a single
140 ks long-look Suzaku observation of MCG–2-58-22.
Suzaku is a Japanese observatory that provides broad-
band X-ray spectra from ∼0.5 keV to above 500 keV
with good energy resolution and effective area around 5–
7 keV for detailed analysis of the Fe K complex (Mitsuda
et al.2007). Our goals for this observation were to study
the Fe K emission, constrain the Compton hump, con-
firm the bare nucleus, and study the soft excess, all of
which Suzaku is capable of doing well. This information
can then be used to explore possibilities for the geome-
try of the circumnuclear material including the Fe line
emitting gas and we improve upon parameter values in
the literature for the Compton hump and Fe K emission
lines. This paper is structured as follows: Section 2 de-
tails data reduction, Section 3 describes spectral fitting
and analysis, and Section 4 contains a discussion of the
results.
2. DATA REDUCTION AND ANALYSIS
Suzaku observed MCG–2-58-22 with the X-ray Imag-
ing Spectrometer (XIS; Koyama et al.2007) and the Hard

Page 2

2Rivers et al.
1.00
1.20
1.40
1.60
1.80
2.00
050100150200250
0.00
0.05
0.10
0.15
0.20
XIS0
XIS1
XIS3
PIN
Count Rate (cts s−1)
Observation Time (ks)
Fig. 1.— The lightcurve over the duration of the observation.
Bins are 800s. The XIS count rates are in the 2–10 keV range
while the PIN count rates are in the 20–50 keV range. The in-
creases in flux seen in the XIS around 30–50 ks and 110–140 ks
are small (only about 20%) and flux-resolved spectroscopy did not
reveal significant change in the shape of the spectrum during these
episodes.
X-ray Detector (HXD; Takahashi et al.2007) on 2009
November 27 beginning at 22:49 UT (Observation ID
704032010). Data were processed with version 2.4.12.27
of the Suzaku pipeline and typical screening criteria were
applied (as per the Suzaku Data Reduction Guide2). All
extractions were done using HEASOFT v.6.9.
2.1. XIS Reduction
The XIS is comprised of 3 CCD cameras3each placed
in the focal plan of an X-ray Telescope module. Two
of these (XIS0 and XIS3) are front-illuminated (FI),
maximizing the effective area of the detectors in the Fe
K bandpass, while the third (XIS1) is back-illuminated
(BI), increasing its effective area in the soft X-ray band
(?2 keV). Two corners of each XIS CCD are illuminated
by an55Fe calibration source, which can be used to cal-
ibrate the gain and test the spectral resolution of data
taken using this instrument (see the Suzaku Data Reduc-
tion Guide for details).
After screening, the good exposure time per XIS was
138.9 ks. The XIS events data were in 3×3 and 5×5
editing modes which were cleaned and summed to create
image files for each XIS. From these we extracted source
and background lightcurves and spectra, using XISRM-
FGEN and XISSIMARFGEN to create the response ma-
trix (RMF) and ancillary response (ARF) files. Data
from the two front-illuminated CCDs were summed to
create a single co-added FI spectrum after it was con-
firmed that the two spectra were consistent.
Data were ignored above 12 keV (10 keV for BI) where
the effective area of the XIS begins to drop dramat-
ically.Data were ignored below 1.0 keV (0.7 for BI
which has a larger effective area at low energies) due
to time-dependent calibration issues of the instumental
2http://heasarc.gsfc.nasa.gov/docs/suzaku/analysis/abc/abc.html
3The fourth CCD camera, XIS2, is inoperative as of 2006
November. See the Suzaku ABC Guide for details.
O K edge at 0.5 keV, and between 1.5 and 2.4 keV due
to large calibration uncertainties for the Si K complex
and Au M edge arising from the detector mirror system.
These issues are not fully understood at the time of this
writing. Average 2–10 keV rates were 1.410±0.002 and
1.521±0.003 countss−1per XIS for FI and BI respec-
tively. Figure 1 shows the XIS lightcurves for the dura-
tion of the observation.
Fitting the55Fe calibration source spectra in XSPEC
v.12.6.0 (Arnaud et al.1996) with a model comprised
of three Gaussian components (Mn Kα1, Kα2and Kβ)
yielded the following results for the Mn Kα1line energy
(expected value of 5.899 keV): 5.886 keV (FI) and 5.890
(BI), showing that the energy calibration has a system-
atic untertainty of ∼10 eV for both the FI the BI. Addi-
tionally, these lines had an average width of 30 eV, which
we will take as instrumental broadening in excess of that
modeled by the response matrix, and have subtracted
this value in quadrature from all measured line widths.
2.2. PIN Reduction
The HXD gathered data with both its detectors, the
PIN diodes and the GSO scintillators, however we did
not use the GSO data because of the faintness of the
source relative to the non-X-ray background in the GSO
band. The HXD/PIN is a non-imaging instrument with
a 34′square (FWHM) field of view.
strument team provides non-X-ray background model
event files using the calibrated GSO data for the par-
ticle background monitor (“background D” or “tuned
background” with METHOD=LCFITDT). This yields
instrument background estimates with ?1.5% system-
atic uncertainty at the 1σ level (Fukuzawa et al.2009).
As suggested in the Suzaku ABC Guide, the Cosmic X-
ray Background was simulated in XSPEC v.12.6 using
the form of Boldt (1987).
Net spectra were extracted and deadtime-corrected for
a net exposure time of 98.0 ks. We excluded PIN data
below 13 keV due to thermal noise and above 60 keV
where the effective area of the detector falls significantly.
The average 13–60 keV rate was 0.202±0.002 countss−1.
Figure 1 shows the PIN lightcurve for the duration of the
observation.
The HXD in-
3. SPECTRAL FITTING
All spectral fitting was done in XSPEC, utiliz-
ing solar abundances of Anders & Grevesse (1989)
and cross-sections from Verner et al.(1996).
fits included absorption by the Galactic column with
NHGal=2.70×1020cm−2(Kalberla et al.2005). Uncer-
tainties are listed at the 90% confidence level (∆χ2=
2.71 for one interesting parameter).
All
3.1. The Fe K Bandpass
We began our analysis with a preliminary study fo-
cused on the Fe K bandpass. We used data from 4.5–8.5
keV from the FI spectrum only because of its excellent
response and effective area in this energy range. We an-
alyzed the Fe K complex, including the Fe Kα and Kβ
lines and the Fe K edge, and investigated the possiblity
of emission from ionized Fe, namely Fe XXV or XXVI
(the latter was reported by Bianchi et al.2004 with a 2σ
detection).

Page 3

Suzaku observation of MCG–2-58-223
TABLE 1
Model Parameters
ModelΓ
Aa
EIb
σ
EW
τR
Γsoft
Asofta
(10−3)
χ2/dof
(10−2)(keV)(10−5)(eV)(eV)
Fe K Band1.68±0.051.15 ±0.086.40±0.022.8 ±0.660 ±3050±100.05±0.02103/95
Broadband 1
1.80±0.021.30±0.026.40±0.022.4 ±0.3
< 6540±100.69±0.053.0±0.60.14+0.16
−0.07
604/464
kBT (keV)
0.18 ±0.02
Abbodyc
2.2 ±0.5 Broadband 21.83±0.011.37±0.01 6.40±0.022.4 ±0.3
< 6041±50.76±0.07611/464
Note. — Best fit parameters for models in the Fe K bandpass and broad band spectrum. “Fe K Band 1” is a model fit over the energy
range 4.5–8 keV using only XIS1 (as described in Section 3.1) including the primary power law, Fe Kα and Kβ lines as Gaussians and the
Fe K edge (“zedge” in XSPEC). The Broadband models were fit over the energy range 0.7–50 keV using FI and BI XIS data as well as
the PIN. Broadband 1 includes the continuum, Gaussian Fe emission lines, the Compton hump modeled with pexrav, and the soft excess
modeled with a power law as described in Section 3.2. Broadband 2 differs from Broadband 1 in the use of a blackbody (“bbody”) to
model the soft excess. We adopt the parameters of the best fit to the Broadband 1 model for the discussion in Section 4.
aPower law normalization (phkeV−1cm−2s−1at 1 keV)
bFe Kα line normalization (phcm−2s−1)
cBlackbody normalization (10−5phkeV−1cm−2s−1)
−4
−2
0
2
4
6
8
χ
Energy (keV)
5 6 7 8
Fe Kα
Fe XXVI
and/or Fe Kβ
Fe K edge
Fig. 2.— Data–model ratios to a simple power law fit for the Fe
K band from 4.5–8.5 keV using the FI XIS data only. Dashed lines
show expected locations (from left to right) of the energy centroids
of the Fe Kα, XXVI, and Kβlines. Note that these are the observed
energies (z = 0.04686).
As a first step we fit a simple power law with Galactic
absorption. Model–data residuals for the simple power
law are shown in Figure 2. This yielded a poor fit with
χ2/dof = 322/100 and obvious residuals around 6.4 keV
(rest frame energy), the location of the Fe Kα line. Fit-
ting the line with a Gaussian component provided a much
better fit with χ2/dof = 117/97. Visual inspection then
revealed additional residuals around 7.1 keV (the loca-
tion of the Kβ line and Fe K edge).
We then added an edge component to model additional
Fe K shell absorption in excess of the Galactic column
and/or the Fe K edge associated with Compton reflec-
tion. We fixed the edge energy at 7.11 keV and left the
optical depth (τ) free. χ2/dof dropped to 108/96, indi-
cating a significant detection of the edge at a confidence
level of ∼99.4% according to an F-test4.
Assuming an origin in neutral or lowly-ionized gas, an
Fe Kβ line should be present in addition to the Fe Kα
line. We added a Gaussian emission line with its energy
4Note that an F-test is inappropriate to perform in this case (see
Protassov et al.2002), however it can give a rough approximation
of the significance.
centroid frozen at 7.056 keV (degeneracy with the Fe K
edge at 7.11 keV and lack of sufficient line strength to
provide good constraints led to our freezing the param-
eter at its expected value), its width tied to that of the
Kα line and its normalization left free. The fit improved,
with χ2/dof dropping to 103/95 with a normalization of
15±12% of the Kα normalization, consistent with that
expected for cold/neutral gas. An F-test3indicates that
this is a 2σ detection at the ∼96.4% confidence level.
In some AGN, contributions to the total observed Fe K
emission profile can arise from material which is ionized,
either by collisional- or photo-ionization. Using XMM-
Newton-EPIC data, Bianchi et al.(2004) found a degen-
eracy between the parameters of the Kβ line and those
from a possible Fe XXVI emission line. They reported
a 2σ detection of the Fe XXVI line when the Kβ line
was not included and with all the parameters of the line
left free. When we allowed the energy of the Kβ line to
be free to vary we found an energy centroid of 7.0±0.1,
consistent with both the Kβ and Fe XXVI line energies.
χ2/dof was 102/95 (not a significant improvement) and
the normalization was 18±14% of the Kα normalization.
Freezing the line energy at 6.966 keV, the weighted av-
erage of the Fe XXVI doublet, provided a fit virtually
identical to the one presented above. Fitting both lines
simultaneously with energies frozen at their expected val-
ues and widths tied to that of the Kα line gave χ2/dof
=102/94 with a normalization of 13+18
and an upper limit of 26% for Kβ (both percentages are
with respect to the Kα normalization).
cannot rule out the that the source may contain both
emission lines and that we are simply unable to deblend
them. For simplicity, in all further models described in
this paper we have included only the Kβ line with frozen
or tied parameters.
We also tested for the presence of Fe XXV (using a
Gaussian component with energy centroid fixed at 6.70
keV), however χ2/dof did not improve and only an up-
per limit to the normalization was obtained (?6×10−6
phcm−2s−1). Final parameters for the Fe K complex
model fit including the Kα line, Fe K edge and Kβ line
−11% for Fe XXVI
We therefore

Page 4

4Rivers et al.
are listed in Table 1 as the “Fe K Band” model, including
the Fe Kα energy centroid (E), intensity (I), width (σ),
and EW.
We also tested the “diskline” model for the emission
lines in place of the more phenomenological Gaussian
model. diskline models the Doppler broadening of an
emission line associated with the inner region of an accre-
tion disk (Fabian et al.1989). The diskline parameters
were not well constrained, giving a very large inner ra-
dius (Rin?20 RS, where RSis the Schwarzchild radius)
and a narrow profile. It did not improve the fit over a
simple Gaussian. Next we tried the addition of a broad
diskline (Rin constrained to around 3 RS and energy
fixed at 6.4 keV) to a narrow Gaussian line with energy
and width fixed at the values found with the Gaussian
fit (see Table 1). We obtained an upper limit to the
diskline normalization of ?2×10−5phcm−2s−1and an
EW of ?35 eV. It is thus possible that a weak broad line
exists in this source and that we are simply unable to de-
tect it; however combining the weakness of this feature
with the lack of ionized emission indicates that most if
not all of the Fe line flux comes from material that is not
close in to the central black hole.
It is possible that such a weak broad line would be
degenerate with the Compton shoulder, a feature which
could arise if there is a significant amount of Compton-
thick material. Additionally, by visual inspection we see
a shallow shelf-like shape in the residuals on the low-
energy side of the emission line when σ is set to ?30 eV.
We tested this by modeling a moderately broad Gaussian
in addition to the narrow Fe Kα line in the “Fe K Band
2” model with an energy centroid fixed at 6.34 keV (Matt
2002). This resulted in an improvement in χ2/dof of only
2/1, which is not a significant detection but indicates
that the Compton shoulder should be tested for in future
observations.
3.2. Broadband Fitting
Next we fit the broadband spectrum of MCG–2-58-
22 covering the range from 0.7–60 keV. We used the XIS
FI, XIS BI, and PIN spectra. We included an instrumen-
tal cross-normalization constant in our fits with the PIN
constant set to 1.16 (this is the expected value for XIS-
nominal pointing and leaving the parameter free caused
degeneracy with the Compton reflection component in
our models) and the BI constant left free relative to the
FI spectrum (values were typically around ∼1.05). The
broadband data are shown in Figure 3a. Figure 3b shows
residuals to a simple power law fit in which we can clearly
see the need for modeling the Fe K complex, Compton
reflection peaking around 20–30 keV, and a soft excess
below ∼1 keV.
We began by modeling the soft excess with a simple
power law in addition to the continuum power law and
modeling the Compton reflection hump using pexrav
(Magdziarz & Zdziarski, 1995) which assumes a disk-like
geometry for the reflecting material and where the value
of R is the proportion of the primary power law that is
reflected off Compton-thick material. Best fit parameters
for this model are listed under “Broadband 1” in Table
1 and data–model residuals are shown in Figure 3c. We
did not find the need for additional cold absorption with
an upper limit to the column density of 2.5×1020cm−2
10−3
10−2
10−1
10 0
Normalized cts s−1 keV−1
−5
0
5
10
χ
b)
−4
−2
0
2
4
χ
c)
110
Energy [keV]
−4
−2
0
2
4
χ
d)
XIS−FI
XIS−BI
PIN
a)
Fig. 3.— Spectral fitting for MCG–2-58-22 from 0.7–50 keV.
Panel a) shows the data from the XIS FI and BI, and the PIN.
Panel b) shows the data–model residuals for a simple absorbed
power law. Panel c) shows the data–model residuals for our best-fit
disk/slab geometry model including the iron lines, soft excess and
Compton hump in addition to the continuum. Panel d) shows the
data–model residuals for our best fit torus geometry model utilizing
MYTorus. It is obvious that the fits shown in panels c) and d)
are virtually identical though the modelling is quite different.
0.0
2.0
4.0
6.0
8.0
10.0
12.0
1975198019851990
Year
1995200020052010
Uhuru
Ariel V
HEAO−1
Einstein
EXOSAT
Ginga
ASCA
RXTE
XMM−Newton/Beppo−SAX
Suzaku
F2−10 (10−11 erg cm−2 s−1)
Fig. 4.— Historical values for the 2–10 keV flux. From left to
right this source has been observed by Uhuru (Cooke et al.1978),
Ariel-V (Marshall, Warwick & Pounds 1981), HEAO-1 (Grif-
fiths et al.1979), Einstein (Turner et al.1991), EXOSAT (Ghosh
& Soundararajaperumal 1992), Ginga (Nandra & Pounds 1994),
ASCA (Weaver et al.1995), RXTE, simultaneously by XMM-
Newton/BeppoSAX (Bianchi et al.2004), and most recently by
Suzaku (this paper).
in excess of the Galactic. We also tested for the pres-
ence of warm absorption using an xstar table but found
it unnecessary for a good fit. The Compton reflection
hump and Fe K edge were well fit by the pexrav model
and Fe line parameters were very similar to those found
from narrow band fitting, including the upper limit to

Page 5

Suzaku observation of MCG–2-58-225
a relativistically broadened diskline component. The
observed 2–10 keV flux was F2−10=5.0 ±0.1 × 10−11
ergscm−2s−1and the intrinsic luminosity, calculated us-
ing the cosmology-corrected luminosity distance given by
the NASA/IPAC Extragalactic Database5, and correct-
ing for Galactic absorption was L2−10=2.30±0.05×1044
ergss−1. Figure 4 shows historical values of the flux of
this source in the 2–10 keV range.
We also tried a (phenomenological) blackbody emis-
sion component to model the soft excess. Parameters
are listed in Table 1 under “Broadband 2”. Both the
blackbody and power law models fit the data reasonably
well, but in both cases parameters were difficult to con-
strain due to calibration issues with the O K edge below
about 1 keV. For simplicity we adopt Broadband 1 as
our best-fit model for the discussion in Section 4.
Bianchi et al.(2004) tested for high energy cut offs in
their sample finding model-dependent values of Ec∼ 200
keV with error bars of 50–800 keV. Extending up to only
60 keV, our data are not highly sensitive to a high energy
cut off or rollover. Utilizing the “cutoffpl” model in
XSPEC we found a lower limit to the rollover energy of
500 keV.
3.3. Applying a Self-Consistent Model
As our knowledge of AGN improves, so too should
the sophistication of our modeling. Self-consistent mod-
els should be able to simultaneously model absorption
and reflection by circumnuclear material, combining the
Fe line, Compton reflection hump and column density
along the line of sight.We have applied the model
“MYTorus” (Murphy & Yaqoob 2009) to our spectrum
of MCG–2-58-22 which assumes the circumnuclear ma-
terial is a donut shape of uniform density and includes
all three of the components listed above.
The MYTorus model was derived from Monte Carlo
simulations of a dusty torus of uniform density surround-
ing an illuminating supermassive black hole. Relevant
parameters to the model include the following: NHTor,
the column density of the material in the torus (not nec-
essarily in the line of sight); θincl, the inclination angle of
the torus, with 0◦corresponding to a face-on view, 90◦
corresponding to edge-on, and with the torus intersect-
ing the line of sight for angles larger than 60◦(the as-
sumed half-opening angle); the photon index (Γ) and the
normalization (APL) of the illuminating power law; the
width (σ) of the Fe Kα line (the material is assumed to
be cold and the energy is not a free parameter); and with
additional parameters ASand AL, the normalization fac-
tors of the Compton hump and Fe line respectively, to be
used when the amount observed for either is significantly
different from that expected by the model due to differ-
ences in covering factors, abundances, etc., from those
assumed. We also included an additional power law for
the soft excess.
We found a reasonably good fit (χ2/dof = 623/465,
very similar to the broadband fit χ2values given in Ta-
ble 1) with θinclfixed at 30◦(there was no significant im-
provement in fit with this parameter free). Since there
is no extra absorption in MCG–2-58-22, the value ob-
tained for NHToris driven primarily by the strength of
5http://nedwww.ipac.caltech.edu/
the Compton reflection hump and Fe line. When ALwas
free to vary there was an improvement in χ2of 9 for
1 less degree of freedom. We obtained a value for AL0f
0.75±0.14, that is the amount of material creating the Fe
line was about 75% of that creating the Compton hump,
possibly due to an underabundance of Fe or geometrical
effects not taken into account by the model (it should also
be noted that the upper uncertainty on ALis consistent
with the lower uncertainty on NHTor). Testing for an ad-
ditional relativistically broadened Fe line yielded an up-
per limit to the normalization of ?1.5×10−5phcm−2s−1
and an EW of ?30 eV. Our best fit parameters are listed
in Table 2 and data–model residuals are shown in Figure
3d.
4. DISCUSSION AND CONCLUSIONS
4.1. The Fe K Complex
Focusing on the Fe K band we found the need for both
Fe Kα and Kβ emission lines as well as an Fe K shell ab-
sorption edge (in broadband fits this edge was modeled
sufficiently by the edge associated with the Compton re-
flection hump in both pexrav and MYTorus). From
the value of the emission line width found in our best-fit
broadband model we calculated the velocity full width at
half maximum (vFWHM) of the emitting material to be
<7100 kms−1. This is consistent with values obtained
for the optical Hβ broad emission line of around 6400–
8500 kms−1(Osterbrock 1977; Kollatschny & Dietrich
2006; Winter et al.2010) and is a significant improve-
ment on previous upper limits set by Weaver et al.(1991)
and Bianchi et al.(2004) of ?30,000 kms−1. Using a
black hole mass estimated from optical luminosity and
line widths to have a value of 108.4M⊙ (Bian & Zhao
2003; Winter et al.2010) and assuming Keplerian motion
of the emitting material, we estimated the radius of the
emitting region to be ?45 lt-days or roughly 1200 RS.
We also tested for a broad line and Compton shoul-
der. According to de la Calle P´ erez et al.(2010), roughly
?1.5×105counts in the 2–10 keV band at CCD reso-
lution provide good enough statistical quality to signifi-
cantly detect a broad line. In the combined FI XIS we
have ∼4×105counts, and our upper limit on the EW of
a broad line places us in the lower part of the EW range
of detected broad lines in the FERO sample of Seyferts
observed with XMM-Newton, wherein significant detec-
tions of broad lines with EW’s in the range of 50–250
eV were reported. We conclude that a very strong broad
line (?50 eV) does not exist in MCG–2-58-22 or else our
observation would have been sufficient to significantly de-
tect it; if there does exist a broad line in this source, then
it must be very weak.
We would expect to see a Compton shoulder given the
presence of Compton-thick material, however we did not
obtain a significant detection. It should also be noted
that the Compton shoulder is included in the MYTorus
model automatically, based on the strength of the Fe line
and the column density of the torus.
4.2. Reflection and Geometry of the Circumnuclear
Material
This source also shows a very prominent Compton
hump around 20–30 keV (see Figure 3b). This feature,
arising from Compton scattering of high energy photons