Download UV, IR, and mm Studies of CO Surrounding the Red Supergiant α

UV, IR, and mm Studies of CO Surrounding the
Red Supergiant α Orionis (M2 Iab)
Graham M. Harper∗, Kenneth G. Carpenter†, Nils Ryde∗∗ , Nathan Smith‡ ,
Joanna Brown§ , Alexander Brown∗ and Kenneth H. Hinkle¶
∗
CASA, University of Colorado, Boulder, Colorado, USA
†
NASA-Goddard Space Flight Center, USA
∗∗
Lund Observatory, Sweden
‡
University of California, Berkley, USA
§
Max-Planck-Institut für Extraterrestrische Physik, Germany
¶
National Optical Astronomy Observatory, Tucson, Arizona, USA
Abstract. Carbon monoxide has been detected in Betelgeuse’s ultraviolet (electronic), infrared
(vibrational), and mm-radio (rotational) spectra, but its spatial distribution has remained elusive.
We have obtained new mm-radio interferometry (CARMA), infrared spatially-resolved spectra
(Phoenix/Gemini-South), and we have performed non-LTE simulations to help establish the spatial
location of the S1 (V = 10 km/s) and S2 (V = 17 km/s) CO shells. The strong CO Fourth-Positive
ultraviolet scattering signature is apparent in the HST GHRS G140L spectrum and the S1 and S2
shells account for much of the UV opacity. Phoenix spectra reveal that the low velocity S1 shell
is present at 1.5 arcsec West of the star and extends out to ∼ 4 arcsec, and its mm-radio emission
appears to originate within an 6 arcsec radius. The interpretation of the S2 shell is less clear. The
Phoenix spectra tentatively suggest that the S2 shell extends to ∼ 7 arcsec. The CARMA channel
maps show an additional strong narrow emission component 5 arcsec from the star, but the spectra
do not show all the expected S1 and S2 signatures. Future CARMA observations should help to
disentangle the signature of the S1 and S2 shells.
Keywords: α Orionis, Betelgeuse, mass loss, radiative transfer
PACS: 97.10.Ex, 97.10.Fy, 97.10.Me
A STUDY IN COMPLEXITY
We have probed the CO surrounding Betelgeuse (α Ori, M2 Iab) in its three characteristic wavebands on smaller spatial scales than has previously been attempted. The spatial
scales of the two shells (S1 and S2) observed in absorption in 12 C16 O and 13 C16 O at
4.6 µ m by Bernat et al. (1979, ApJ, 233, L135) have yet to be determined. Single dish
radio observations appear not to have fully resolved the larger S2 shell with 12 arcsec
FWHP beams and ionization calculations by Huggins and collaborators suggest that
both shells are within ∼ 10 arcsec radius of the star (Huggins et al. 1994, ApJ, 424,
L127). Below we present new observations that sample the region 1-6 arcsec from the
star in scattered 12 C16 O 4.6µ m R(1), R(2), and R(3) vibrational lines, and 230 GHz
CARMA interferometry with resolutions of 2 arcsec and 1.3 km s−1 in the J=2-1 transition. We also present a comprehensive non-LTE model of CO (uv, IR, and mm-radio)
that is shown in Figure 1.
14
HST G140L
10
O
Flux (10-13 erg cm-2 s-1 A-1)
12
8
Combined
6
4
Off Star
2
On Star
0
1300
1400
1500
1600
O
Wavelength (A )
1700
1800
FIGURE 1. Non-LTE CO simulation showing the ultraviolet Fourth-Positive system and the HST
GHRS G140L spectrum (top). The model molecules, 12 C16 O and 13 C16 O, include X 1 Σ+ : v = 0 − 8, j =
0, 60 and A 1 Π : v = 0−23, j = 0, 60. Collisional data has been taken from the literature (e.g., Balakrishnan
et al. 2002, ApJ, 568, 443) and these have also been used to estimate data that are not available, i.e., most
of the collisional matrix. Electronic Einstein A-values are from Kurucz (1976 SAO Spec. Report 374),
and IR and rotational A-values are taken from Goorvitch and collaborators (e.g., 1994, ApJS, 91, 483).
CO FOURTH-POSITIVE UV SIGNATURE
Figure 1 shows a simulated uv spectrum with just the S1 and S2 shells. The spectrum of
the stellar disk (red) shows the deep absorption first described by Wahlgren et al. (1992,
ASP Conf Ser. 26, p. 37) and analyzed by Carpenter et al. (1994, ApJ 428, 329). The
scattered emission (blue) is an important component of CO spectrum and its intensity is
controlled in part by the size of the S1 shell, so the spatially unresolved GHRS spectrum
provides spatial information on the inner arcsecond scales of the S1 shell. Some of the
spectral features in the synthetic spectrum (green) nicely reproduce the observations,
e.g., 1400-1500Å, while some do not, e.g., 1550-1620Å. These models will be fine tuned
using the spatial information gleaned from the Phoenix/Gemini-S long-slit spectra and
CARMA spectral-interferometry. The thermodynamic and spatial parameter space to be
explored is vast so we are using escape probability techniques, in spherical geometry, to
reduce the non-linearity of the large system of non-LTE rate equations to be solved. The
radiative transfer is solved on fine frequency and spatial grids for both molecules with
full line-overlap. Each model provides uv, IR and, mm-radio spatially resolved spectra.
PHOENIX/GEMINI-S OBSERVATIONS
Phoenix 4.6µ m long-slit spectra of the 12 C16 O R-branch v0 = 1 → v00 = 0 R(1), R(2),
and R(3) lines were obtained on the 8m Gemini-South on 2005 February 27. These have
a spectral resolution of R ∼ 75, 000 and the 2 pixel slit-width has a sky projection of
0.17 arcsec. The observations were obtained at ±0.5 and ±1 arcsec North, South, East,
and West of Betelgeuse. The off-star long-slit spectra were cleaned of contaminating
1•10-13
Scaled Fλ (erg cm-2 s-1 A-1)
8•10-14
-14
6•10
Impact Parameter
2.2 arcsec
3.1 arcsec
4.0 arcsec
5.0 arcsec
S1 Shell
4•10-14
2•10-14
0
-30
S2 Shell
-20
Edge of S2 ?
-10
0
∆v (km s-1)
10
20
30
FIGURE 2. Phoenix 4.6µ m long-slit spectra of the 12 C16 O R-branch v0 = 1 → v00 = 0 R(2) obtained at
different impact parameters and integrated over 0.5 arcsec in the spatial direction.
stellar “seeing” signal using a scaled on-star spectrum. In the best exposed 1 arcsec
North and West observations the cleaned spectra of all three vibrational lines show a
similar dependence of line profile shape with impact parameter.
Figure 2 shows four R(2) spectra obtained at different impact parameters. Beyond
an impact parameter of 2.5 arcsec the S1 component begins to narrow as the sightline
samples the tangential flow near the edge of the shell. The S2 shell then appears as a
broad pedestal of the scattered emission profiles. This suggests that the S1 shell has an
outer radius at ∼ 4.5 arcsec. At 4 arcsec the profile is dominated by the S2 shell, and
by 5 arcsec the scattered peak narrows once more suggesting that the S2 shell has an
outer radius of ∼ 7 arcsec. Similar Phoenix observations made at Kitt Peak at an offset
4 arcsec North with a 50 arcsec slit by Ryde et al. (1999, A&A, 347, L35) show an
excitation temperature of ' 40 K which is more characteristic of the S2 shell (70 +10
−10 K)
+50
than the S1 shell (200−10 K) (Bernat et al. 1979, ApJ, 233, L135).
CARMA MILLIMETER OBSERVATIONS
Observations of the 230 GHz 12 C16 O J = 2 − 1 emission line were made with the
Combined Array for Research in Millimeter-wave Astronomy (CARMA) on 2007 June
18, 21, and 24 UT. The frequency channels are 1 MHz wide and the synthesized beam
FWHM is 2.32 × 1.96 arcsec with a position angle of −13.2◦ .
Figure 3 shows two spectra: The inner 4 × 4 arcsec box (blue) captures the pointsource stellar continuum and the onset of both the S1 and S2 shells, and the larger
14 × 14 arcsec box (red) shows the flux contribution from the extended emission for
both S1 and S2 shells. A spatially unresolved source with narrow velocity dispersion
near the stellar rest frame was discovered 5 arcsec S-W of the star, which adds to the
complexity of the line profile. Curiously the CARMA data do not show the line profile
suggested by the Phoenix observations.
14x14 arcsec Box
8
S2 Shell
Flux (Jy)
6
S1 Shell
Discrete
Off-Source Emission
4
2
Inner 4x4 arcsec
0
-20
-10
0
vLSR (km s-1)
10
20
FIGURE 3. Left: CARMA 230 GHz 12 C16 O J = 2 − 1 interferometric emission spectra extracted with
4 × 4 and 14 × 14 arcsec boxes. Right: Discrete off-source emission discovered 5 arcsec S-W of star.
The absence of flux in the narrow range between 27 and 30 km s−1 and also between
32 − 48km s−1 is puzzling. While the former may result from a non-uniform shell, the
later may be an artifact of the interferometry. The onset of the S1 and S2 shells are seen
as point sources in the image-channel maps at the corresponding shell velocities which
reflects the material moving directly towards and away from us. At lower projected
velocities, in the rest frame of the star, the shells become low surface brightness rings.
These soon disappear into the noise of the image maps. This, in part, may explain the
lack of nested classic top-hat profiles even in the large 14 × 14 arcsec box.
SUMMARY
Observations and non-LTE synthetic spectra show scattering from extended CO
material. Most of the CO column is in the warm (Texc ∼ 200 K) low velocity
(V = 10km s−1 ) S1 shell.
• Phoenix high spectral and spatial resolution spectra obtained on Gemini-S clearly
resolve the S1, and probably the S2, shell. These spectra probe the S1 density
distribution.
• CARMA 12 C16 O J=2-1 interferometry reveals 4 emission components: unresolved
star, extended emission from S1 and S2 shells, and a narrow discrete source 5 arcsec
S-W of the star.
• The Phoenix and CARMA data are not fully reconciled - more CARMA data are
needed.
•
ACKNOWLEDGMENTS
Support for GMH was provided by STScI Program # HST-AR-10956.01-A.