We have reported the first high-resolution, moderate signal-to-noise ratio X-ray observations of the Broad Line Radio Galaxy 3C445 obtained with ASCA. Our principal results are:
ergs s 
;
line is confirmed;
), and a soft excess is present below
1.5 keV;
ergs s ) and its spectral shape are most likely
inconsistent with an origin in the host galaxy, and are related
instead to the AGN activity;
) intrinsic slope and two absorbers, a nonuniform
one with N 
cm 
and covering fraction
C 
80%, and a uniform one with N 
cm . The Fe line has EW 270 eV;
The amount of X-ray absorption inferred from the ASCA data is one to
two orders of magnitude larger than the extinction derived from the
IR/optical data (assuming Galactic gas-to-dust ratios), N 
cm (see § 1). This immediately suggests that
the X-ray absorbers must lie in the inner region of the AGN, inside
the BLR radius, or the latter would be obscured. For example, the
nonuniform absorber with N 
cm could be
interpreted as a distribution of clouds in thermal equilibrium at
distances 
cm, as in the model of Ferland & Rees (1988),
with a second shell of uniform absorption at 
cm to account for the second ASCA absorber. Alternatively, the
X-ray absorber has a lower than Galactic dust content.
The presence of cold gas clouds in the inner regions of 3C445 is in
agreement with the results of Granato, Danese, & Franceschini (1997),
who compared the column densities derived from IR and X-ray
observations for a sample of Seyfert 2s, and concluded that the amount
of X-ray absorption always exceeds that inferred at longer wavelengths
by a factor 2-10, indicating that the gas responsible for the X-ray
absorption is located inside the dust sublimation radius. In
particular, for column densities 10 
cm , they
find that the IR opacity is approximately a factor 10 lower than at
X-ray wavelengths, as in the case of 3C445. We also note in passing
that marginal evidence for column variability is suggested in
3C445. It is difficult to compare the EXOSAT and GINGA results to ASCA
because of the different bandpasses (and because of the contamination
from the nearby cluster in the GINGA data). However, the EXOSAT ME
data were consistent with a column density N 
cm (Turner & Pounds 1989), which is one order of magnitude
lower than ASCA and GINGA. If confirmed, the column density
variability in 3C445 could support a scenario where the absorbing gas
is in the form of clouds intercepting the line of sight.
With the continuum described as a dual absorber the Fe line is clearly
detected. The observed EW ( 270 eV), is suggestive of an origin
in material with high column densities, N 
a few 
cm (Leahy & Creighton 1993). The line is probably
broad, 
keV, implying velocities 
35,000 km
s for the emitting gas.
The presence of complex absorption by cold gas in 3C445 is in conflict
with the unification scenario (extended to radio-loud objects,
e.g. Urry & Padovani 1995), where the nucleus of a type 1 AGN is seen
directly, with little or no neutral absorption. A first possibility is
that 3C445 switched its classification from type 1 to type 2 at the
time of the ASCA and GINGA observations. This is a concrete
possibility in the refined version of the unification model proposed
by Turner et al. (1997a). In order to account for the similar Fe
K profiles of a sample of Seyfert 1s and 2s, and to explain
their intrinsically different accretion properties, the latter authors
proposed that the absorbing material in these objects is fragmented
into a uniform and isotropic distribution of clouds at the distance of
the putative dusty torus, with density and composition of the gas
varying from cloud to cloud. Thus an object could easily appear as a
type 1 or 2 depending on whether a cloud of sufficient column density
intercepts the line of sight. A change of classification in 3C445
would imply strong variability for the broad emission lines on
timescales of years due to variable reddening. Indeed, the H
line profile has been reported to vary dramatically in 12 years
(Crenshaw et al. 1988), although in a way which does not suggest
immediately a change of classification for the object: a strong
H blue wing seen at the time of the observations by Osterbrock
et al (1976) is no longer seen by Crenshaw et al. (1988).
Alternatively, taking at face value the unification model, we would be
tempted to conclude that 3C445 is indeed intrinsically a type-2 AGN,
and that the broad emission lines are produced outside the compact
nuclear environs, such as e.g. in a jet.
The luminosity of the hard X-ray component in the ASCA spectrum,
L ergs s , is similar
to the EXOSAT luminosity (Turner & Pounds 1989) and a factor 
lower than GINGA (Pounds 1990) (after rescaling to the different
value of H 
used by the latter authors), indicating some
variability of the hard X-rays. This is in contrast to the 0.2-2.0
keV flux, which remained constant between the ROSAT and ASCA
observations taken 2 years apart (§ 3.1.3), and consistent with an
earlier marginal detection of the Einstein IPC (Wilkes et
al. 1994). This could support a scattering origin for the soft X-rays,
possibly off free electrons which could be also responsible for
producing the polarized H line (Kay et al. 1996). More
observations of 3C445 at X-rays and at longer wavelengths are needed
to confirm this suggestion and to determine the nature of the
scattering mirror.
As a final note we would like to compare briefly 3C445 to other radio
galaxies observed with ASCA. Data have been published so far for a
handful of objects, including the BLRGs 3C390.3, 3C120, 3C382, 3C109,
and Pictor A (Eracleous et al. 1996; Grandi et al. 1997; Reynolds
1997; Allen et al. 1997; Eracleous & Halpern 1997a,b), and the NLRGs
Centaurus A and Cygnus A (Arnaud 1996; Turner et al. 1997b). In both
BLRGs and NLRGs, a power law ( 
) continuum plus a
strong and broad Fe line were detected, with the notable exceptions of
Pictor A (Eracleous & Halpern 1997a,b) and of Cyg A (Arnaud 1996). To
facilitate the comparison we plot in Figure 4 the ASCA photon index,
the intrinsic column density (the difference of the fitted N 
and
the Galactic value), and the Fe line EW versus the 2-10 keV
luminosity for BLRGs (open circle) and NLRGs (filled triangles); 3C445
is shown as a filled circle.
Despite the low statistics, the photon index distribution of both
BLRGs and the two NLRGs is remarkably narrow around 
over more than 3 luminosity decades. This value of the photon index is
similar to the intrinsic slope of the Seyfert galaxies observed with
ASCA (Nandra et al. 1997), in contrast with earlier claims, based on
lower sensitivity data, that radio-loud AGN are flatter in the X-rays
than their radio-quiet counterparts (e.g., Wilkes & Elvis 1987). A
more detailed discussion awaits larger samples of radio-loud objects
observed with ASCA (Sambruna et al. 1997, in prep.). In Figure 4b,
little or no excess N over Galactic is measured for BLRGs,
contrary to the two NLRGs which are obscured by N
cm . 3C445 sticks out as a moderately luminous BLRG with the
more heavily absorbed X-ray spectrum, while its Fe line is not
particularly weak compared to other BLRGs (Figure 4c).
The only other BLRG where intrinsic X-ray absorption was detected so
far is 3C109, a luminous FRII at z=0.306 (Allen et al. 1996 and
references therein). In 3C109, like in 3C445, the optical continuum
and H line are intrinsically polarized (although with a higher
polarization degree than in 3C445), and the large Balmer decrement and
steep IR-to-optical slope indicates the presence of substantial
intrinsic absorption, most likely dust (Elvis et al. 1984). The
intrinsic X-ray column density in 3C109 is 
cm (Fig. 4b); however, in 3C109 the X-ray column density is in
good agreement with the value derived from the optical reddening
(Allen et al. 1996), and supports a picture where the line of sight to
this BLRG skims the outer edge of the molecular torus. This is in
contrast to 3C445, where the more complex absorption properties can
not be easily reconciled within the unified models, again underlining
the unique nature of this source among BLRGs.
