Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.arcetri.astro.it/~lt/preprints/hbe_disks/hbe_disks.ps.gz Дата изменения: Tue Sep 11 16:57:05 2007 Дата индексирования: Sat Dec 22 08:19:10 2007 Кодировка: Поисковые слова: massive stars

Not to appear in Nonlearned J., 45.
Preprint typeset using L A T E X style emulateapj v. 7/15/03
FIRST EVIDENCE FOR DUSTY DISKS AROUND HERBIG BE STARS
A. Fuente 1 ,A. Rodr guez-Franco 2 , L. Testi 3 , A. Natta 3 , R. Bachiller 1 , R. Neri 4
Not to appear in Nonlearned J., 45.
ABSTRACT
We have carried out a high-sensitivity search for circumstellar disks around Herbig Be stars in the
continuum at 1.4mm and 2.7mm using the IRAM interferometer at the Plateau de Bure (PdBI) . In
this letter, we report data on three well studied B0 stars, MWC 1080, MWC 137 and R Mon. The two
latter have also been observed in the continuum at 0.7 cm and 1.3 cm using the NRAO Very Large
Array (VLA) . We report the detection of circumstellar disks around MWC 1080 and R Mon with
masses of M d  0.003 and 0.01 M , respectively, while for MWC 137 we estimate a disk mass upper
limit of 0.007 M . Our results show that the ratio M d /M  is at least an order of magnitude lower in
Herbig Be stars than in Herbig Ae and T Tauri stars.
Subject headings: Radio continuum: stars { Circumstellar matter { Stars: individual (MWC 1080,
MWC 137, R Mon) { stars:pre-main sequence
1. INTRODUCTION
The existence of accretion disks around massive stars
(M> 5 M ) remains a matter of debate. There is in-
creasing evidence for the existence of attened structures
(disks) around high-mass protostars. However there is no
clear evidence of disks in later phases, namely in Herbig
Be (HBe) stars. Natta et al. (2000) compiled the inter-
ferometric observations in mm continuum around HBe
stars and found that the occurrence of disks in HBe is 
0. This contrasts with the case of the lower mass HAe
stars (2M  M   5 M ) in which the occurrence of cir-
cumstellar disks is similar to that in T Tauri stars. They
propose that the lack of disks around HBe stars is due
to the rapid evolution of these objects, which disperse
the surrounding dust and gas in about 10 6 yrs (Fuente
et al. 2002). However, this lack of detection can be a sen-
sitivity eect. HBe stars are usually further away than
HAe stars, and higher sensitivity interferometric obser-
vations are required to detect circumstellar disks around
these objects. In this Letter, we present the rst results
of a high-sensitivity search for circumstellar disks around
three B0 stars, MWC 1080, MWC 137 and R Mon, which
are the best studied of the ve Herbig B0 stars in the
northern sky listed by The et al. (1994).
2. OBSERVATIONS
Interferometric observations in the continuum at
1.4mm and 2.7mm have been carried out towards the
Herbig Be stars MWC 1080, MWC 137 and R Mon, us-
ing the IRAM 5 array at Plateau de Bure, France, in the
CD set of congurations. MWC 137 and R Mon, were
also observed with the NRAO 6 Very Large Array (VLA)
1 Observatorio Astronomico Nacional, Apdo. 1143, E-28800 Al-
cala de Henares,Spain; a.fuente@oan.es
2 Dpto Matematica Aplicada, Universidad Complutense de
Madrid, Av. Arcos de Jalon s/n, E-28037 Madrid, Spain
3 Osservatorio Astrosico di Arcetri, Largo Enrico Fermi, 5, I-
50125 Firenze, Italy
4 Institute de Radioastronomie Millimetrique, 300 rue de la
Piscine, 38406 St Martin d'Heres Cedex, France
5 IRAM is supported by INSU/CNRS (France), MPG (Ger-
many) and IGN (Spain).
6 The NRAO is a facility of the National Science Foundation
operated under cooperative agreement by Associated Universities,
at 0.7 cm and 1.3 cm in its D conguration. Flux cal-
ibration is accurate within 10% at 2.7mm and 20% at
1.4mm in the PdBI images and within 20% in the VLA
images. No correction for primary beam attenuation has
been applied. PdBI and VLA images are shown in Fig 1.
Flux densities at the star positions are shown in Table 1.
3. MWC 1080
Although, we have not observed MWC 1080 at cm
wavelengths. some information about the cm emission
can be found in the literature. Curiel et al. (1989), from
10 00 resolution observations at 6cm, detected a weak ex-
tended component which includes the star and the re ec-
tion nebula. Skinner et al. (1993) with higher angular
resolution ( 1 00 ) tentatively detected at 3.6cm a com-
pact source at the star position. Recently, Girart et al.
(2002) detected three 6cm sources using the VLA with
an angular resolution of  5 00 . One of these sources is
coincident with the star position. They did not detect
a counterpart of this emission at 2.0cm with an upper
limit of  0.32 mJy.
Intense emission is detected in the millimeter contin-
uum images observed with the PdBI (see Fig 1ab). In-
terferometric observations at 2.7mm in MWC 1080 were
previously reported by Di Francesco et al. (1997) with
a non-detection. Our image improves by a factor of 10
the sensitivity and by a factor of 2 the angular resolu-
tion of their data. The millimeter continuum emission in
MWC 1080 is extended. Two intense components con-
nected by a bridge of weak emission are observed in the
2.7mm image. These components are much closer to the
star than the VLA 6cm sources detected by Girart et al.
(2002). We will refer to these components as \NW" and
\SE". The better angular resolution of the 1.4mm image
resolve the \NW" and \SE" clumps in several compo-
nents, \NW1",\SE1", \SE2" and \Star". The total ux
in the 1.4mm image is  24 mJy. In our single-dish data
we measured a 1.4mm ux of 73.8 mJy at the star posi-
tion with an angular resolution of  10: 00 5 (Fuente et al.
1998). Thus the interferometer has resolved the 1.4mm
emission and missed  70% of the ux. The "Star" com-
Inc.

2
ponent is coincident with the optical star position and
remains unresolved in the 1.4mm image. This implies
that its size is < 2" which corresponds to a spatial extent
of <2000 UA at the distance of MWC 1080 (d=1000pc).
This size is compatible with the emission arising in a
circumstellar disk.
4. MWC 137
MWC 137 is associated with the one-arcminute size HII
region S266 (Fich 1993). Observations of this region
at 3.6cm were reported by Skinner et al. (1993). Our
1.3cm image of MWC 137 consists of a compact source
coincident with the HBe star surrounded by a thin shell
with a radius of 30"-40" from the point source. Weak
emission is found towards the thin shell and the point
source. In the 0.7cm image, the compact source appear
more intense while the emission of the thin shell becomes
fainter. Since our aim is the detection of a circumstellar
disk, we are only interested in the compact component
(see Table 1).
The continuum millimeter images show intense emis-
sion centered at the star position. The total ux in the
1.4mm image is  10 mJy, i.e.  30% of the ux mea-
sured in the single-dish observations (Fuente et al. 1998).
From our 1.4mm image we derive a source size of 1.8 00
 0.8 00 , which corresponds to  2300 AU  1000 AU
assuming a distance of 1300 pc and a position angle of
7515deg. Note that the accuracy of the position angle
estimate is heavily limited by the S/N ratio of the data.
5. R MON
R Mon was observed at 6cm, 3.6cm and 2cm using the
VLA by Skinner et al. (1993). They detected continuum
emission at a position slightly shifted from the optical
one. A compact source is detected in R Mon at 1.3cm
and 0.7cm. The position of this source is oset (0.74",-
0.45") from the optical position but coincident with the
radio source. A jet-like feature is observed in the 1.3cm
image which is surely driven by the radio source. This
jet was already detected by Brugel et al. (1984) in optical
lines. Since the uncertainty of the optical position is not
known, the identication of the radio source with R Mon
is uncertain.
We have detected continuum emission towards the
star in the 2.7mm image with a peak emission of 4.1
mJy/beam. The total integrated ux is 6.4 mJy re-
vealing that the emission is extended. Deconvolving the
2.7mm image with the beam, we estimate an emission
size of 3"{4", which corresponds to  3000 AU at the
distance of 800 pc. Natta et al. (2000) reported a total
2.7mm ux of 13.01.3 mJy towards R Mon based on un-
published data by Mannings (1998). This ux exceeds by
a factor of 2 our present result. This reinforces our con-
clusion that the emission is extended. Then, the dierent
beams and calibration uncertainity can account for this
discrepancy.The 2.7mm continuum emission is elongated
in the direction perpendicular to the bipolar nebula axis
and the centroid of the emission lies to the northeast of
the position of the radio-source. The shape and position
of the 2.7mm emission suggest that the 2.7mm emission is
tracing a dierent emission component from that traced
by the emission at cm wavelengths, and argues in favor
of the existence of a attened structure (disk) around the
star.
6. SPECTRAL ENERGY DISTRIBUTIONS (SEDS)
The SEDs in the cm-mm range of MWC 1080,
MWC 137 and R Mon are shown in Fig. 2. In or-
der to avoid the contribution of the envelopes of these
sources to the continuum emission, only interferometric
measurements have been included in these SEDs. Even
though, some problems still remain. The emission at
cm wavelengths in MWC 1080 is extended. The neg-
ative index derived by comparing the ux at 6cm and
3.6cm is very likely due to the larger beam of the 6cm
ux ( 5.5 00 5.03 00 ) compared with that of the 3.6cm
one (1.24 00 0.83 00 ). The disk emission should be un-
resolved at the distance of our source. Therefore, we
are not interested in the extended component and con-
sider the 6cm ux as an upper limit. Unfortunately, the
free-free emission arising in the stellar wind is also ex-
pected to be unresolved with our beam and could make
an important contribution to the continuum ux even
at millimeter wavelengths. In order to derive the disk
mass in MWC 1080, we have assumed that the 3.6cm
ux is due to free-free wind emission and subtracted it
from the observed mm uxes. These corrected values
are then interpreted as optically thin disk emission, ac-
cording to the expression F  = d 2 M d   B  (T d ) with
 = 0:01((mm)=1:3)  cm 2 g 1 . We have assumed a
constant value of T d = 215 K following the recommen-
dation of Natta et al. (2000) for B0 stars. With these
assumptions, we obtain from the 1.4mm ux a disk mass
of 0.003 M . Since we have only an upper limit to the
2.7mm ux, the value of  cannot be determined. We
have compared the disk mass derived with this simpli-
ed expression to those of more elaborated models (see
Dullemond et al. 2002), varying the disk outer radius
and the surface density prole over a large range of pos-
sible values, and conrmed the conclusion of Natta et al.
(2000) that the values of M d derived in this way are ac-
curate within a factor 2 { 3. We have also considered the
possibility that the 3.6cm emission arise in an optically
thin HII region instead of free-free emission. With only
two points in the SED, we cannot discern between these
two possibilities. In this case, we obtain a slightly larger
value of the disk mass of, M d  0.004 M . Therefore we
estimate that the mass of the disk around MWC 1080 is
0.0030.001 M .
The complete SED of MWC 137 can be tted with a
single component with spectral index  = +0.760.01.
This spectral index is consistent with that expected in
the free-free emission arising in the stellar wind. Al-
though a value of  =+0.6 is expected for a ionized
isotropic wind, small deviations of this value can be ex-
plained by a dierent geometry or a partially ionized
wind. However, we cannot discard the existence of a
very low-mass disk. In order to estimate an upper limit
to the disk mass, we have extrapolated the 3.6cm ux to
millimeter wavelengths with an spectral index of +0.6.
Then, we have estimated the disk mass from the excess
ux at 1.4mm. With the same values of  and T d than
in the case of MWC 1080, we obtained an upper limit to
the disk mass of 0.007 M .
The SED of R Mon is the most complete one. We have
been able to t all the centimeter data using a spectral
index, cm  0.64, which is characteristic of a spherical
stellar wind. In order to estimate the mass of the cir-

3
cumstellar disk, we have extrapolated the free-free emis-
sion to millimeter wavelengths and t the millimeter part
of the SED with a simple disk model. As in the case
of MWC 1080, we have assumed an unresolved disk at
a constant temperature, T d = 215 K and adjusted the
value of M d and . Only values of  = 0.25 { 0.5 t our
observational points. A value of  larger than 0.5 would
produce an excess of ux at 1.4mm. This low value of 
could be interpretred as an optically thick compact disk,
or an optically thin disk with large grains. The rst pos-
sibility would imply millimiter uxes larger than those
measured. A small disk of R = 100 AU should have a
mass of 0.35 M in order to have  1:4mm 1. But in this
case, the 1.4mm ux would be at least an order of mag-
nitude larger than our upper limit. The situation would
be worse for a larger disk. The second possibility seems
more plausible. Similarly low values of  have been found
in Herbig Ae and T Tauri disks and reveal highly pro-
cessed dust (Beckwith & Sargent 1991; Testi et al. 2003).
With the ducial value of the dust emissivity, we derive
a disk mass of 0.01 M . Note, however, that, if the low
value of  is interpreted as evidence of grain growth, one
needs to reduce the ducial 2.7mm dust emissivity by a
factor of about 4 (Testi et al. 2003), and to increase the
estimated disk mass by a similar amount.
7. DISK DETECTIONS IN HBE STARS
Our results prove the existence of circumstellar dust
around MWC 1080 and R Mon. However, it is very dif-
cult to interpret it beyond doubt as evidence of a cir-
cumstellar disk. In general, a compact and small 1.4mm
source (< 1 00 { 2 00 ) at the star position is considered as
a disk detection. This criterium seems suфcient in TTS
located at distances of  150 pc, in which an angular
resolution of 1 00 corresponds to a linear scale of 100 - 200
AU , typical size of circumstellar disks. Since HBe stars
are usually located at distances larger than 1 Kpc, even
the higher angular resolution provided by interferome-
ters,  1 00 , corresponds to sizes of a few thousands of
AU, and it is not possible to distinguish between \bona-
de" circumstellar disks and a attened structure of a
few 1000 AU.
Another diфculty comes from the fact that many HBe
stars are likely associated to lower mass companions that
will not be resolved in our beam. MWC 1080 is a close
binary with a separation of 0.7 00 (Leinert et al. 1997).
Mid-infrared observations by Polomski et al. (2002) us-
ing the Keck II telescope resolve the binary and found
that warm material ( 350 K), with a 3 { 10 m spec-
tral index consistent with the existence of a circumstellar
disk, is surrounding MWC 1080 A. A cooler dust compo-
nent is associated with MWC 1080 B. We have detected
an unresolved mm continuum source in MWC 1080. We
consider that this source is a good candidate for disk de-
tection around a HBe star. However, since the angular
resolution of our observations is not suфcient to resolve
the two binary components, we cannot exclude the pos-
sibility that the emission is associated with the younger
companion.
R Mon is also a close binary with a separation of 0.69".
Its companion is a very young TTS. Close et al. (1997)
found an extinction of Av = 13.1 mag towards the star
which they interpreted as due to an optically thick disk
of R = 100 AU. Our 2.7mm image reveals the existence
of a compact source of  3000 AU  2300 AU. This
size is larger than that expected for a circumstellar disk
and suggests the existence of a attened envelope sur-
rounding the disk. Our value of the disk mass may be
overestimated due to the contribution of the attened
envelope. However, the analysis of the mm spectral in-
dex shows that the grains responsible of the emission are
heavily processed, so that most of the emission is very
likely arising in a circumstellar disk. If this is true, it
is unlikely that the disk could be associated to the TTS
companion, since in this case one should use in deriving
the disk mass a value of T d  15 K (Natta et al. 2000)
and the estimated mass would increase to values in the
range 0.2{0.5 M , too large for a TTS disk.
We have completed the compilation of (interferomet-
ric) disk observations of HAeBe stars carried out by
Natta et al. (2000) with the observations in MWC 1080,
MWC 137 and R Mon reported in this paper, and the
observations in HD 200775 and LkH 234 reported by
Fuente et al. (2001). In Fig. 3, we plot the disk masses
as a function of the spectral type of the star and the stel-
lar age. We have plotted MWC 1080 and R Mon as disk
detections; however, as discussed in the previous section,
it is possible that the \disk" emission is either overesti-
mated or not associated to the Herbig Be star itself, but
to unresolved companions, thus reinforcing our conclu-
sions. In any case, disk masses in HBe stars are at least
an order of magnitude lower than in HAe stars. Plotting
M d /M  as a function of the spectral type, one nds that
while the value of M d /M  is roughly constant and equal
to 0.04 for stars with spectral type A0-M7, M d /M  <
0.001 in HBe stars. The same eect remains when one
plots the disk mass against the stellar age. The values of
disk masses in HBe stars are systematically lower than
in HAe stars for stars with ages between 10 5 and 10 7 yrs.
This lack of massive disks in HBe stars even for stars as
young as 10 5 years put several constraints to the possi-
bility of forming planetary systems around massive stars.
In a time-scale of < 10 6 years the disk mass around a HBe
star is similar to or even lower than the total mass of the
planets in the solar system ( 0.0013 M ). Thus planet
formation should occur very fast, for planetary systems
to exist at all around massive stars.
This paper has been partially funded by the Spanish
MCyT under projects DGES/AYA2000-927, ESP2001-
4519-PE and ESP2002-01693, and European FEDER
funds. A.N. and L.T. acknowledge support from ARS-
1/R/073/01 grant to the Osservatorio di Arcetri. A.R.
acknowledge support from ES P2002-01627 y AYA2002-
10113-E
REFERENCES
Beckwith, S. V. W. & Sargent, A. I. 1991, ApJ, 381, 250
Brugel E.W., Mundt, R., Buhrke, T. 1984, ApJ287, L73
Close, L.M., Roddier, F., Hora, J.L., Graves, J.E., Northcott, M.,
Roddier, C., Homan, W.F., Dayal, A., Fazio, G.G., Deutsch,
L.K., 1997, ApJ489, 210

4
Curiel, S., Rodrguez, L.F., Canto, J., Bohigas, J., Roth, M.,
Torrelles, J.M., Ap. Lett. Comm., 27, 299
Di Francesco, J., Evans, N.J.,II, Harvey, P.M., Mundy,
L.G.,Guilloteau, S., Chandler, C.J. 1997, ApJ 482, 433
Fich, M. 1993, Ap&SS86, 475
Fuente, A., Martn-Pintado, J., Bachiller, R., Neri, R., Palla, F.
1998, A&A 334, 253
Fuente, A., Neri, R., Martn-Pintado, J., Bachiller, R., Rodrguez-
Franco, A., Palla, F. 2001, A&A 366, 873
Fuente, A., Martn-Pintado, J., Bachiller, R., Rodrguez-Franco,
A., & Palla, F. 2002, A&A, 387, 977
Girart, J.M.,Curiel, S., Rodrguez, L.F., Canto, L. 2002, Revista
Mexicana de Astronoma y Astrofsica 38, 169
Leinert,C., Richichi, A., & Haas, M. 1997, A&A, 318, 472
Natta, A., Grinin, V. P., Mannings, V. 2000, in Protostars and
Planets IV, ed. V. Mannings, A. Boss, & S. S. Russell (Tucson:
University of Arizona Press), 559
Polomski, E.F., Telesco, C.M., Pia R., Schulz, B. 2002, ApJ124,
2207
Skinner, S.L., Brown, A., Stewart, R.T. 1993, Ap&SS87, 217
Testi, L., Natta, A., Shepherd, D. S., & Wilner, D. J. 2003, A&A,
403, 323
The, P. S.; de Winter, D.; Perez, M. R. 1994, A&AS 104, 315

5
Table 1
Observed fluxes.
Object S 1:3cm S 0:7cm S 2:7mm S 1:4mm
(mJy/beam) (mJy/beam) (mJy/beam) (mJy/beam)
MWC 1080 ... ... < 1.7 3.1 (0.2)
MWC 137 1.5(0.5) 1 2.35(0.13) 4.1(0.2) 7.1(0.6)
R Mon 0.78(0.10) 1.26(0.13) 4.1(0.5) < 15
1 Subtracting the emission of the extended component

6

7

8

9
Fig. 1.| Continuum images obtained with the PdB and VLA arrays. The star marks the optical position [MWC 1080 : 23:17:25.574
60:50:43.34; MWC 137: 06:18:45.504 +15:16:52.4; R Mon: 06:39:09.94 +08:44:10.0] and the lled triangle the position of the radio source
(Skinner et al. 1993). Contour levels are: a. -2.0 to -0.2 mJy/beam and 0.2 to 4.0 mJy/beam by 0.2 mJy/beam; b. -1.6, -0.7, 0.7 to 3.4
mJy/beam by 0.9 mJy/beam; c. -0.2 mJy/beam, 0.2 to 2.2 mJy/beam by 0.4 mJy/beam; d. -0.3 mJy/beam, 0.3 to 3.6 mJy/beam by
0.6 mJy/beam; e. -1.8 mJy/beam, -0.6 mJy/beam 0.6 to 6.6 mJy/beam by 1.2 mJy/beam; f. -0.2 mJy/beam, 0.2 to 2.4 mJy/beam by
0.2 mJy/beam; g. -0.5 mJy/beam, 0.5 to 4.0 mJy/beam by 0.5 mJy/beam; h. -0.2 mJy/beam, 0.2 to 0.8 mJy/beam by 0.2 mJy/beam; i.
-0.2 mJy/beam 0.2 to 1.4 mJy/beam by 0.2 mJy/beam.
Fig. 2.| Spectral energy distributions (SEDs) at cm and mm wavelengths of MWC 1080, MWC 137 and R Mon. The uxes at cm
wavelengths have been taken from Skinner et al. (1993) and Girart et al. (2002), and at mm wavelengths are from this paper. The
short-dash lines are model predictions for the free-free emission component. The long-dash lines are model predictions for circumstellar
disks of 0.003 M and =1 (MWC 1080), 0.007 M and =1 (MWC 137), and 0.01 M and =0.5 (R Mon). The solid line is the sum
of both components. The short-long straight line in central panel is the predicted uxes for the emission arising in a stellar wind with =
+0.76 in MWC 137.
Fig. 3.| Plot of the M d and M d /M versus the spectral type of the star (top panels) and M d versus the stellar age (bottom panel)
for a wide sample of HAEBE stars. Data have been taken from the compilation by Natta et al. (2000) (empty symbols) and the results
reported by Fuente et al. (2001) and this paper (lled symbols).Dots mean detections while triangles mean upper limits. Note that disk
masses in HBe stars are at least an order of magnitude lower than in HAe stars.