Light Bulb
behind the Supernova Remnant SN 1006
L140, 1997 September 10
Light Bulb behind the Supernova Remnant SN 1006 A point X-ray source located
9
NE
of the center of SN 1006 has been spectroscopically identified as a
background QSO, with a redshift of 0.335. The object is moderately bright,
with magnitude V = 18.3. If its ultraviolet spectrum is typical
of low-z quasars, this object will be a second source (after the
Schweizer-Middleditch star) to use for absorption spectroscopy of material
within SN 1006. Absorption spectra provide a unique probe for unshocked
ejecta within this supernova remnant and can possibly solve the
long-standing problem
of missing
iron in the remnants of Type Ia supernovae.
Subject headings: ISM: individual:
(SN 1006)
quasars: generalsupernova
remnantssupernovae: individual
(SN
1006)ultraviolet: ISM
1 Visiting Astronomer, Cerro Tololo Inter-American Observatory (CTIO), which is operated by the Association of Universities for Research in Astronomy, Inc., under contract to the National Science Foundation.
As the remnant of the brightest supernova
(SN) to have been witnessed in recorded history, SN 1006 has long held
interest for astronomers working at a variety of wavelengths and using
varied techniques. Among the most powerful of these has been the
measurement of absorption lines in the ultraviolet spectrum of a hot
subdwarf star (commonly referred to as
the SM
star)
discovered and classified as sdOB by Schweizer
& Middleditch (1980). The fortuitous location of this star behind
SN 1006, only
2
5
from the projected geometric center of the remnant, provides an unusual
opportunity to study the distribution of cold SN ejecta through absorption
spectroscopy. Very broad UV resonance absorption features due to Fe
II were first detected in IUE spectra of the
SM star by Wu et al. (1983; see
also Fesen et al. 1988). Subsequent spectra
from the Faint Object Spectrograph (FOS) on the Hubble
Space Telescope (HST) led to positive identification
of additional broad absorption lines of Si II and Si
IV, in addition to greatly improved measurements of
the Fe II lines (Wu et al.
1993, 1997).
Iron in SN 1006 is of special interest,
since it is the suspected remnant of a Type Ia event
(Minkowski
1966; Schaefer 1996), and conventional
models for Type Ia SNs require the eventual production of several tenths of
a solar mass of
ironformed
from the decay of the 56Ni that powers the Type Ia SN light
curve (Colgate & McKee 1969;
Arnett
1979; Nomoto, Thielemann, & Yokoi
1984; an extensive set of models and references to more recent
literature is given in Höflich &
Khokhlov 1996). However, the Fe II absorption-line
measurements indicate only 0.014
M
of Fe+, which is short of the expected total Fe mass by a factor
of 20 or more. A new analysis of the FOS data using a different method of
continuum fitting by Hamilton et al. (1997)
leads to
0.029 M
of Fe+ but is still far short of predictions for the total mass
of Fe. In other young remnants of probable Type Ia SNs, Tycho and Kepler,
strong X-ray lines from highly ionized iron result in large part from
ejecta that have been excited by a reverse shock, but the absence of such
lines in SN 1006 indicates that iron in the ejecta cannot yet have been
shocked and so must still be relatively cool.
Hamilton & Fesen (1988) considered
various states in which iron could be hidden: the absence of
Fe I absorption features excludes the possibility of
significant gaseous neutral iron, while the absence of strong IR emission
provides evidence against the formation of grains (iron rich or otherwise)
in the ejecta. They suggested that a significant amount of Fe might be more
highly ionized, but far-UV spectra of the SM star obtained
by Blair, Long, & Raymond (1996)
from the Hopkins Ultraviolet Telescope (HUT) show only weak Fe
III absorption and severely limit the amount of
Fe++. Where several tenths of a solar mass of iron might be
hiding in SN 1006 remains a mystery.
Further UV absorption spectroscopy of SN
1006, especially probing different lines of sight through the remnant
shell, would certainly be important, but one is, of course, limited by the
availability of background UV
light
bulbs
that are appropriately placed and bright enough for observation. We report
here the discovery of a second, moderately bright UV source within SN 1006:
a QSO located
9 NE
from the projected center of the remnant shell.
In an observation of SN 1006 from the
ROSAT HRI, we noted an unresolved X-ray source, located in the
northeastern region of the remnant shell, with no obvious optical
counterpart (Winkler & Long 1997;
hereafter Paper I). (In addition, a second,
brighter, X-ray source near the center of the remnant shell and three
fainter sources outside all coincide with bright, V
910, foreground
stars.) The unidentified northeastern source is also apparent in the PSPC
images of Willingale et al.
(1996), especially the image at E > 1.5 keV, which indicates
that it is a relatively hard source. The HRI count rate was measured at 5
counts ks-1, equivalent to
2 ×
10-13 ergs cm-2 s-1
(0.52.5
keV) for a hard-spectrum source. In the
H
image
from Paper I, the only evident object within the
5
error radius for the northeastern X-ray source was what appeared to be a
star estimated at about 17 mag. Located at R.A. (2000) =
15h03m33
93, decl.
(2000) =
-41°52237,
this object is
only 3
7
from the position of the HRI X-ray source, well within the HRI error
circle.
On 1997 March 29, we obtained a spectrum
of this candidate object from the Cerro Tololo Inter-American Observatory
(CTIO) 1.5
m f
7.5
telescope and Ritchey-Chrétien spectrograph. The spectrograph was
configured with a 300 line mm-1 grating blazed at 4000 Å
and a Loral 1200 × 800 pixel CCD, to give wavelength coverage of
35007000
Å at a dispersion of 2.9 Å pixel-1. A spectrograph
slit width of
35 gave
a resolution of 9 Å. The candidate was observed under photometric but
brightly moonlit conditions for a total of 4000 s, split among four frames
of 1000 s each. All the data reduction has been carried out with
conventional IRAF
reduction techniques: 2 flat
fielding using a combination of dome flats and internal quartz flats,
secondary illumination
correction
using twilight sky flats, and wavelength calibration from an internal He-Ar
source. Flux calibration was achieved by observing several
spectrophotometric standards from Hamuy et al.
(1992).
The extracted spectrum is shown
in Figure 1. The relatively flat spectrum
and pattern of broad and narrow emission lines is typical of QSOs. The
strongest line is Mg II
2798,
redshifted into the blue end of the optical spectrum with velocity width of
7000
km s-1 (FWHM). Also evident are broad
H
and
narrow [O III]
4959,
5007 at the red end of the spectrum. A redshift of z = 0.335 ±
0.001 is consistent with all the identified features.
Fig. 1
We have also obtained broad-band images of
the northeastern region of SN 1006 from the CTIO 0.9 m telescope, equipped
with the Tektronix No. 5 2048 × 2048 pixel CCD, on 1997 February 10.
This combination gives a field of
137
at a scale of
040 pixel-1.
Exposures of 1200, 240, and 240 s, respectively, were obtained in the
U, B, and V bands under photometric, moonless
conditions. These were reduced with standard IRAF procedures for overscan
subtraction and flat fielding based on well-exposed twilight sky flats.
Photometric calibration was based on exposures of several
Landolt (1992) and
Graham (1982)
fields. Figure 2 shows a section of
the V-band image with the error circle for the HRI X-ray source
indicated. The QSO position given above was measured from this image, with
the use of 18 surrounding stars from the HST Guide Star Catalog to
define the reference frame.
Fig. 2
Photometry of the object that was
subsequently identified as a background QSO indicates unusual, very blue
colors: V = 18.32 ± 0.05, B - V = 0.17 ±
0.04, U - B = -0.78 ± 0.04. This is significantly
fainter than our earlier estimate (Paper I) of about 17
mag, based on a narrowband
H image.
The difference is most likely due to the fact that the QSO redshift of
0.335 shifts the wavelength of
H
emission into the bandpass of our
H filter.
2 IRAF is distributed by the National Optical Astronomy Observatories, operated by the Association of Universities for Research in Astronomy, Inc., under contract from the National Science Foundation.
The discovery of another quasar at low
redshift would be entirely unnoteworthy were it not for its location. But
through its felicitous placement behind the supernova remnant SN 1006, this
object becomes potentially quite valuable as an ultraviolet continuum
source against which to measure absorption lines. As we noted in
§ 1, the one line of sight through SN 1006 that has
so far been probed, i.e., that to the SM star, has yielded rich information
while deepening the mystery of where the iron in this supposed Type Ia
remnant may be hiding. The broad absorption lines of Fe
II and Fe III measured with the
HST FOS and HUT, respectively, demonstrate convincingly that a
significant amount of fast-moving, cold iron is present in the remnant
shell, but the inventory of Fe+ and Fe++ obtained
from spherically symmetric models based on this single line of sight falls
short of the
0.3
M
of iron predicted from models for Type Ia supernovae by a factor of at
least 10, and perhaps 20 or more.
The absorption-line measurements for the
SM star lead to another problem: how to keep ejecta expanding at the
velocity indicated by the widths of the absorption lines within the
confines of the remnant shell without placing SN 1006 too far away. As we
discussed more fully in Paper I, the velocity width for
the cold iron, together with the known age of the remnant, gives a minimum
extent for ejecta along the line of sight that is barely smaller than
the transverse dimensions of the outer remnant shell at a distance of
1.8 kpcthe
value inferred from proper motions of the optical filaments measured by
Long, Blair, & van den Bergh (1988),
together with the velocity of the shock as measured spectroscopically in
these same filaments by Laming et al.
(1996). Hamilton et al. (1997) reanalyzed the
absorption data for the SM star and concluded that the nearside,
blueshifted velocity is smaller than that originally measured
by Wu et al. 1993. With this result and a geometry that
is elongated along the line of sight, they were able to achieve a
self-consistent model for SN 1006.
Both these
problemsthe
missing iron and the geometry-distance
puzzlemay
be elucidated by probing the distribution of ejecta along a second line of
sight through SN 1006. The QSO is well placed for this purpose, as it is
located
9 away
from the remnant center,
58% of
the shell radius in this direction, compared with
25
for the SM star (see Fig. 3). Absorption
would, of course, be observed only from material that lies outside the
projected radius to the line of sight, so an ejecta ion species
concentrated entirely within
58% of
the shell radius would produce no absorption. But for a species that is
smoothly distributed around a spherical shell with a radius of
60% of
the shell radius, the line of sight to the QSO may have a larger optical
depth and narrower velocity width than that to the SM star. Absorption
spectra and comparison of line profiles to the two sources for all species
that have been observed in the SM
starFe
II, Fe III, Si
II, Si III, and
Si IVwould
be important for constraining the geometry of the ejecta distribution and
the radius of the reverse shock in SN 1006.
Fig. 3
Is the QSO bright enough to measure
absorption profiles? We have redshifted the composite spectrum from 101
QSOs observed with the HST FOS (Zheng
et al. 1997) to z = 0.335 and scaled it to match the QSO
spectrum from Figure 1 in the region of overlap. The
extinction is unknown, but we have assumed a value E(B -
V) = 0.12, the same as that measured for the SM star
(Blair et al. 1996). The continuum is nearly flat
at F
3 ×
10-16 ergs cm-2 s-1 Å-1
over the range
22003500
Å, fainter by a factor of about 50 than that of the SM star at the
wavelengths of the strong Fe II lines,
2383 and
2600. Nevertheless,
the QSO is sufficiently bright that high-quality UV spectra can be obtained
with modest exposures by using the new imaging spectrograph on
HST.
There is an interesting application for
the QSO behind SN 1006 for X-ray astronomy as well. The brightest part of
the X-ray shell is in the northeast, passing only
6 away
from the QSO. Since the QSO is a point source of X-rays, it is potentially
an excellent fiducial point for measuring the position and proper motion of
the X-ray shell in this area. This is the same region in
which Koyama et al. (1995) argued for a
synchrotron origin for the X-ray emission on the basis of the absence
of X-ray lines in the ASCA spectrum. In Paper I,
we showed a detailed correlation between the X-ray morphology observed with
the ROSAT HRI and that observed in radio from the VLA
by Reynolds & Gilmore
(1986). Moffett, Goss, & Reynolds
(1993) have measured the radio expansion rate for the entire shell at
044 ±
013
yr-1. This differs somewhat from the proper motions for the
H
filaments of
030
± 004
yr-1 determined by Long et al. (1988), but
the optical value was specifically for the northwestern filaments, while
the radio value was for the overall expansion of the shell. It would be
interesting to determine the X-ray proper motion at a specific shell
location. Measurement of the expected differences of a few arcseconds over
a baseline of several years would be marginal at best on the basis of the
typical aspect uncertainty
of 5
for the ROSAT HRI, but the presence of a fiducial reference so close
to the X-ray shock front would make measurement of the proper motions
feasible.
We are grateful to Becky Walldroff for her assistance in carrying out the spectroscopic observations and to the staff of the Cerro Tololo Inter-American Observatory (CTIO) for their capable and efficient support. This work has been supported in part by NSF grant AST-9315967 and NASA grant NAG5-1668, with additional support from the W. M. Keck Foundation through the Keck Northeast Astronomy Consortium. P. F. W. gratefully acknowledges the hospitality of CTIO, where he has been in residence during the course of the work presented here, and in particular valuable discussions with J. Baldwin, M. Keane, and M. Phillips.
Full image (17kb) | Discussion in text
FIG.
1.Optical
spectrum of the counterpart to the X-ray source shown
in Fig. 2, now identified as a QSO at redshift
0.335.
Full image (97kb) | Discussion in text
FIG.
2.Section
of the V-band image of the northeastern region of SN 1006, showing
the 10
diameter error circle for the X-ray point source noted in
Paper I. The object in the circle is the QSO with the
spectrum shown in Fig. 1. Several background galaxies
may also be seen, which is indicative of the low extinction in
the direction of SN 1006. The illustrated section measures exactly
6 ×
6;
north is up and east is to the left.
Full image (70kb) | Discussion in text
FIG.
3.Smoothed
X-ray map (from Paper I), showing the locations of the
newly discovered QSO and the SM star in the SN 1006 shell. The plus
indicates the center of the remnant. The frame measures
40
×
40.