Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.arcetri.astro.it/~cesa/jd3all.ps.gz Äàòà èçìåíåíèÿ: Fri Feb 21 18:36:04 2003 Äàòà èíäåêñèðîâàíèÿ: Fri Dec 21 21:52:03 2007 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: ionization front

Massive Star Birth - JD3 IAU GA
ASP Conference Series, Vol. 000, 2001
P. S. Conti & E. B. Churchwell, eds.
Rationale for the Joint Discussion
Peter S. Conti
JILA and APS Department, University of Colorado, Boulder CO 80309
Edwin B. Churchwell
Department of Astronomy, University of Wisconsin, Madison, WI 53706
Abstract. This Joint Discussion (JD) will consider the birth processes
of massive stars. While similar phenomena (e.g., accretion discs, out ows,
etc.) are found in low mass star formation, additional physics must be
considered given the ionization of the interstellar environment by Lyman
continuum photons, stellar winds from the hot star(s), and their deeper
gravitational potentials. This JD will bring together experts from sev-
eral disparate astronomical communities: stellar astrophysics, interstellar
medium, radio astronomy, and stellar dynamics. The concept is to con-
trast observations of very young stars and star formation regions over
various wavelengths with theoretical expectations.
The birth places of massive stars are in molecular cloud cores, but stars
newly born within these regions are initially optically shrouded by the dust in
the natal cloud. Massive stars, those of types O and B, are typically formed
together in loose, or tight, groupings of associations or clusters. These hot
and luminous stars have a profound eect on their local environments from
their extensive Lyman continuum luminosity and strong stellar winds. Due to
the large gravitational potentials of the central stars, all dynamical processes
occur on shorter time scales than those near low mass proto stars. This has the
consequence that the neighborhood of massive protostars is a very dynamic place
in which gas velocities, densities and temperatures are expected to change by
orders of magnitude within a radius less than 10 16 cm. Changes of the ionization
state of the gas and the temperature and properties of the dust are also expected
to change dramatically in the neighborhood of massive protostars.
In an oversimplied early evolution scenario one would imagine that the
photons dissociate, excite, and ionize the local material and the stellar winds
will blow this away from the formation sites. Thus the initial birth processes
are highly time dependent, and dynamical eects from the ensemble of hot
stars probably play a major role in the overall formation processes. Addition-
ally, individual protostars are following their own initial evolution from collapse,
contraction, disc accretion, and possibly growth from mergers of lower mass ob-
jects, towards necular ignition. It is desireous to investigate the properties of
giant molecular clouds that are now forming stars along with the resultant HII
regions. These give estimates of the total numbers of stars potentially or already
present.
1

Observations of the earliest, most deeply embedded stages of massive star
formation are only just becoming feasible with (sub)millimeter and mid-infrared
telescopes. These objects are still so young that little or no radio continuum
is detected. They are characterized by strong water masers and a rich, time-
dependent chemistry in their surrounding envelopes. The formation of these
"hot cores" and their evolution to the HII region stage is still poorly understood,
and a thorough discussion of the latest observational and theoretical results is
warrented. For molecular core clouds one would like to evaluate the importance
of oblateness and if the distribution of water and methanol masers, which are
often observed to have linear distributions on the sky, are indicative of \edge-on"
geometry.
Ultra compact (UC) HII regions represent a well known early phase in the
evolution of massive stars. Statistically, there are far too many of them to
be consistent with the expectations from classical Stromgren theory, thus this
phase lasts on average about 100 times longer than expected from the sound
crossing times. Numerous postulates for the lifetime of UCHII regions have
been proposed but no general consensus has yet emerged. The limited number
of morphologies of UCHII regions and their dynamics (expansion rates, rotation,
turbulent motions, etc.) still require further eorts.
A very general issue to be considered is the similarities and dierences
between massive and low mass star formation; for the latter, there has been
considerable observational and theoretical progress in the last decade. So far,
the evidence for the presence of discs around massive young stars has been
controversal. Are they a necessary phase of massive star birth, and if so, how
can we detect them? Are collimated polar outlows (jets) a necessary consequence
of star birth and star formation and, if so, what is the physics of the out ow
formation and collimation? Of the known out ow sources, can we distinguish
between well collimated jets with entrained ISM or simply out ows with wide
opening-angles in which most of the out ow material is diverted from in-falling
gas? Could mergers of lower mass objects be an important formation mechanism
for massive stars?
The role of the ionizing photons from massive hot stars is poorly understood
in terms of star formation. On what time scale would they photoevaporate stellar
discs? How do they aect the surrounding envelopes, and how many photons are
\leaking out"? During the UCHII region phase there are still many unresolved
issues among which are: the lifetime problem; the origin of possibly extended
hard X-ray emission; the tight correlation of mass outlow rates with bolometric
luminosity; the origin of the mass in the bipolar out ows; the identicaion of
the sources of massive out ows (i.e., a massive protostar or a low mass star close
by?); etc.
While many of the overall properties of the dust and gas in giant HII regions
are well known from radio and IR wavelengths it is only recently that indvidual
exciting stars have been identied and classied through near IR photometry and
spectroscopy. There is strong evidence from observations of M17 that the earliest
type O stars are free of their natal dust clouds while somewhat less luminous
objects still have disc geometry. It is expected that more extensive near IR
photometry and spectroscopy of additional Galactic HII regions will become

available shortly. These observations, and others, can be used to confront the
current paridigms of massive star formation.
During the past few years, observational and theoretical progress in star
formation has been substantial, especially for low mass objects. We feel the
time is now ripe to consider the specic topic of massive star birth. By bringing
together experts in various subdisciplines for a one day Joint Discussion, we
believe that: 1) a better understanding of the problems of massive star formation
can be achieved; 2) an assement of where we are in solving those problems will
result; and 3) ideas for future programs to attach remaining problems will follow.

Massive Star Birth - JD3 IAU GA
ASP Conference Series, Vol. 000, 2001
P. S. Conti & E. B. Churchwell, eds.
Structure and Conditions in Massive Star Forming Giant
Molecular Clouds
Jonathan Williams
Astronomy Department, University of Florida, 211 Bryant Space
Science Center, Gainesville FL, USA; williams@astro.u .edu
Abstract. Massive stars form in clusters within self-gravitating molec-
ular clouds. The size scale of these clusters is suôciently large that non-
thermal, or turbulent, motions of the gas must be taken into account
when considering their formation. Millimeter wavelength radio observa-
tions of the gas and dust in these clouds reveal a complex, self-similar
structure that re ects the turbulent nature of the gas. Dierences are
seen, however, towards dense bound cores in proto-clusters. Examina-
tion of the kinematics of gas around such cores suggests that dissipation
of turbulence may be the rst step in the star formation process. Newly
formed stars, on the other hand, replenish turbulence through their winds
and out ows. In this way, star formation may be self-regulated. Obser-
vations and simulations are beginning to demonstrate the key role that
cloud turbulence plays in the formation and evolution of stellar groups.
Molecular cloud structure
Massive stars form in Giant Molecular Clouds (GMCs) that have masses M >
10 5 M , sizes R ' 20 50 pc, linewidths
vFWHM  2 5 km s 1 and kinetic
temperatures T ' 10 20 K. Since the observed linewidths are a factor 5 10
greater than the thermal linewidth for such cool objects, the gas motions within
GMCs are highly turbulent. Such turbulent conditions manifest themselves
not only through the high spectral linewidths but also in the complex internal
structure of these clouds (Williams, Blitz, & McKee 2000).
Molecular cloud structures have been studied over a wide range of scales
and environments and scaling laws such as the size-linewidth relation and mass
spectrum are found to be amazingly similar from cloud to cloud independent of
their star forming nature. Such a high degree of self-similarity lends itself to
fractal models of clouds but cannot be directly related to star formation since
the same structures are observed in clouds with stars and those without.
Recent observations of thermal emission from dense, dusty condensations
in star forming regions, however, have shown departures from self-similarity.
Motte, Andre, & Neri (1998) and Testi & Sargent (1998) found that the mass
spectrum, dN=dM  M  , of cores within cluster forming regions was steeper
( > 2) than typically measured for structures within clouds ( = 1:5 1:8) and
closer to the stellar IMF ( = 2:35).
The dierence between these dust continuum observations and the molecu-
lar spectral line observations are that the former are of dense, bound, individual
4

star forming cores while the latter are of lower (average) density structures.
Moreover, the linewidths of the dust continuum cores are small, near the ther-
mal value, while the spectral line clumps are predominantly non-thermal. If we
equate the structural similarities of the clumps with the universal nature of cloud
turbulence then is the departure from self-similarity in the thermally supported
cores related to the loss of this turbulent support? Examining turbulence in
molecular clouds may lead to a physical understanding of the relation between
cloud structure and star formation.
The role of turbulence in star formation
Most stars, and particularly all massive stars, form in clusters over size scales,
> 0:2 pc, where the pre-star forming material is supported by turbulent motions
of the gas. However, numerical simulations of hydrodynamic and magnetohy-
drodynamic turbulence in GMCs show that such motions cannot be maintained
over more than a few free-fall times (Mac Low 1999), yet the observed low star
formation eôciency of molecular clouds requires that cloud support be quasi-
static. We are led, therefore, to a dynamic picture of molecular clouds in which
turbulent motions are in a continual state of dissipation and replenishment.
Observations of the velocity eld in the Serpens cloud show both these eects.
The NE region of the Serpens molecular cloud contains a deeply embed-
ded cluster of very young (Class 0) protostars. It is one of the nearest known
examples of cluster formation and although it only contains low mass stars,
it provides an opportunity to study how stars form in groups and is therefore
an important stepping stone for understanding massive star formation. Here
I brie y summarize BIMA 3 mm inteferometer observations of the dense gas
toward the cluster. Full details can be found in Williams & Myers (2000).
Observations were made in two molecular lines; the optically thin N 2 H + (1{
0) tracing the turbulent velocity eld of the gas, and the optically thick CS(2{1)
tracing outer core envelopes and used as a diagnostic of infall and out ow mo-
tions. Several \quiescent cores" were found in a map of non-thermal N 2 H + veloc-
ity dispersion. These represent localized regions of turbulent dissipation. Con-
versely, the non-thermal N 2 H + velocity dispersion reached a maximum around
the two brightest embedded protostars indicating local stirring of the velocity
eld (this applies to the dense gas in the cores since N 2 H + is not seen in pro-
tostellar out ow wings). The CS line was generally self-absorbed across the
cluster and, by modeling the asymmetry in the double-peaked prole, it was
possible to determine the relative inward/outward motions of the cores. Inward
motions were greatest toward the quiescent cores and reversed around the re-
gions of greatest N 2 H + linewidth. This correlation suggests that the inward and
outward motions are turbulent ows from high to low pressure (linewidth).
The initial growth and contraction of star forming cores may occur, there-
fore, through the loss of turbulent pressure support. This process is more dy-
namic than the quasi-static slippage of neutral particles through magnetic eld
lines (although ambipolar diusion may still characterize the last stages of core
collapse). Newly formed stars stir up the gas through their powerful winds and
out ows and can reverse the pressure gradient and subsequent ow. Perhaps
star formation becomes self-regulated in this manner (c.f. Norman & Silk 1980).

Figure 1. Core-core velocity dispersion in two protoclusters in the
Serpens molecular cloud. The shaded histogram represents the distri-
bution of core velocities about the mean for a deeply embedded group
of Class 0 sources in the NW region of the cloud, and the solid line
shows the distribution for a less embedded cluster in the SE. As a clus-
ter emerges from a molecular cloud the loss of surrounding mass causes
the stellar group to become less bound.
New observations of a second, less embedded and hence more evolved, clus-
ter in the SE region of the Serpens cloud show evolutionary dierences in com-
parison with the NW cluster. Several quiescent cores were again seen in the
N 2 H + line but the core-core velocity dispersion, measured over a similar area,
was higher in the more evolved cluster (Figure 1) and similar to that mea-
sured for small stellar groups (Jones 1971). The evolution from a tightly bound
group to a more loosely bound, and ultimately unbound, state is expected as
a protocluster emerges from its natal molecular cloud (Hills 1980). Detailed
observations of individual star forming cores within embedded clusters, such as
described here, should reveal the processes involved.
References
Hills, J.G. 1980, ApJ, 235, 986
Jones, B.F. 1971, AJ, 76, 470
Mac Low, M.-M. 1999, ApJ, 524, 169
Motte, F., Andre, Ph., & Neri, R. 1998, A&A, 336, 150
Norman, C., & Silk, J. 1980, ApJ, 238, 158
Testi, L., & Sargent, A.I. 1998, ApJ, 508, L91
Williams, J.P., Blitz, L., & McKee, C.F., 2000, in Protostars and Planets IV,
eds. V. Mannings, A.P. Boss & S.S. Russell, 97
Williams, J.P., & Myers, P.C. 2000, ApJ, 537, 891

Massive Star Birth - JD3 IAU GA
ASP Conference Series, Vol. 000, 2001
P. S. Conti & E. B. Churchwell, eds.
The Circumstellar Environment of Embedded Massive
Stars
Lee G. Mundy, Friedrich Wyrowski, and Sarah Watt
Astronomy Department, University of Maryland, College Park, MD
20742
Introduction
Millimeter and submillimeter wavelength images of massive star-forming regions
are uncovering the natal material distribution and revealing the complexities of
their circumstellar environments on size scales from parsecs to 100's of AU.
Progress in these areas has been slower than for low-mass stars because massive
stars are more distant, and because they are gregarious siblings with dierent
evolutionary stages that can co-exist even within a core. Nevertheless, observa-
tional goals for the near future include the characterization of an early evolu-
tionary sequence for massive stars, determination if the accretion process and
formation sequence for massive stars is similar to that of low-mass stars, and
understanding of the role of triggering events in massive star formation.
Two of the most intriguing results from studies of high-mass star formation
regions have been the discovery of hot molecular cores (e.g. Olmi et al. 1996,
Cesaroni et al. 1994, Walmsley 1995) and the clustering of massive stars. Hot
molecular cores are identied by their emission in unusual molecules such as
CH 3 CN, and by the high brightness temperatures of lines (typically 150 - 250 K)
associated with abundant molecules. It is argued that these cores are themselves
the sites of new massive star formation (c.f. Walmsley 1995). As such, they
may represent an important stage in the formation of massive stars. The second
result, the clustering of massive stars, suggests that massive stars form in a
locally special environment; the structure of their parent cores provides insights
into this clustering. The following two sections look at these points more closely.
Hot Cores: Chemistry or Cores?
Are all regions which show \hot core chemistry" and \hot core emission" massive
cores in the act of forming stars? It has often been found that hot core chemistry
(Millar et al. 1997, Charnley et al. 1992) is associated with massive, dense
cores: masses of 100's to 1000's of solar masses and densities of 10 7 to 10 8 cm 3
(Walmsley 1995). However, the original prototype for hot cores, the Orion Hot
Core, is estimated to be under ten solar masses and the individual hot core
tracer molecules have dierent morphologies when mapped at high resolution
(Blake et al. 1966). Is the presence of \hot core emission" truly an indicator of
a massive internally heated core or is it simply an indicator of high temperature
gas-phase chemistry?
We have recently acquired sub-arcsecond resolution observations of two well-
known hot core sources, G34.26+0.15 and G31.41+0.31, with the objective of
7

Figure 2. =2.7 mm Continuum, C 18 O and CH 3 CN images for
G31.41+0.31 and G34.26+0.15. Panels a and c show the C 18 O emis-
sion (grey contours) overlayed on the continuum emission. The dark
lines are 4 contours of the K=3 CH 3 CN emission. Panels b and d
show the J=6-5 K=3 CH 3 CN (contours) and continuum (greyscale)
emission. The black/white central contour in G31.41+0.31 is a dip.
answering this question. Figure 6 shows maps of the =2.7 mm continuum,
C 18 O J=1-0, and CH 3 CN emission from these two regions.
These images are discussed in more detail by Watt et al.(2000). The primary
point here is that the CH 3 CN emission, a traditional hot core tracer, is not
closely associated with dust continuum emission in G34.26+0.15 (the observed
=2.7 mm continuum is consistent with arising entirely from ionized gas in the
UC HII region). In G31.41+0.31, the CH 3 CN emission is roughly centered on
the continuum emission, and, in this source, there is no compact ionized gas
component so the =2.7 mm emission is tracing dust.
Thus, G34.26+0.15 is a hot core without a massive compact core while
G31.41+0.31 is a hot core with a massive compact core. We argued previously
(Watt and Mundy 1999) that G34.26+0.15 is an example of externally driven
hot core chemistry; the current higher resolution images support that picture.
The presence of traditional hot core molecules and emission is a statement about
the chemical state of the gas: the dust grain mantles have been evaporated and
the released molecules are being chemically processed in a hot (150 - 300 K) gas
environment. This can occur within a dense massive core, or it can occur along
the surface of a core which is strongly heated by nearby massive stars.

The Origin and Structure of Cores: Insights into Cluster Star Forma-
tion?
Recent improvements in the sensitivity and resolution of continuum observations
are providing new insights into the structure of massive cores and have stim-
ulated renewed searches for the earliest stages of massive star formation. The
Holy Grail is the establishment of an evolutionary sequence for massive stars
with observational prototypes. Works by Hunter et al. (2000), Sridharan et
al. (1999), and Wyrowski et al. (2000) have provided new candidate high-mass
protostellar regions and the work by Hatchell et al. (2000) illustrates the type
of structure studies that can be done once these regions are identied. It is clear
that forming massive stars are associated with massive cores which typically
extend out to scales of order a parsec. The cores contain much more mass than
needed for a single star, typically 10 3 to 10 4 solar masses. This is in agreement
with the historical wisdom and with the observations that visible massive stars
are normally associated with a cluster of lower mass stars.
Interferometric observations able to follow the structure of these cores down
to the 10 3 AU scale. In two sources that we have recently studied, G31.41+0.31
and G10.47+0.31, we nd that the power-law density structure of the cores
extends from the parsec scale down to around 3000 AU. Within roughly 3000
AU, the core structure, as traced by the continuum emission, attens. This
attening of the density structure in the central core may trace the size scale of
the forming cluster. This size scale would likely be set by the angular momentum
content of the original cloud, by strength of the original magnetic eld, or a
combination of the two. This type of scenario where the core feeds material into
a central stellar nursery naturally produces star clusters due to gravitational
interactions decoupling the protostars from the natal envelope as they grow in
mass.
Acknowledgments. We thank the Hat Creek sta for their eorts. We
thank R. Cesaroni, M. Walmsley, and E. Churchwell for enjoyable and informa-
tive conversations.
References
Blake et al. 1987, ApJ, 472, L49.
Cesaroni, R. Churchwell, E., Hofner, P., Walmsley, C.M., & Kurtz, S. 1994,
A&A, 288, 903.
Charnley, S.B., Tielens, A.G.G.M., & Millar, T.J. 1992, ApJ, 399, 71.
Hatchell et al. 2000, A&A, 357, 637.
Hunter et al. 2000, AJ, 119, 2711.
Olmi, L., Cesaroni, R., Neri, R., & Walmsley, C.M. 1996, A&A, 315, 565.
Sridharan et al. 1999, Star Formation 1999, Nagoya, Japan, Editor: T. Nakamoto,
183.
Walmsley, C.M. 1995, Rev. Mex. Astron. Astro., 1, 137.
Watt, S. Mundy, L.G., & Wyrowski, F. 2000, ApJ, submitted.

Massive Star Birth - JD3 IAU GA
ASP Conference Series, Vol. 000, 2001
P. S. Conti & E. B. Churchwell, eds.
Chemistry in the Envelopes around Massive Young Stars
Ewine F. van Dishoeck and Floris F.S. van der Tak
Leiden Observatory, P.O. Box 9513, 2300 RA Leiden, The Netherlands
Abstract. Recent chemical studies of high-mass star-forming regions
at submillimeter and infrared wavelengths reveal large variations in the
abundances depending on evolutionary state. Such variations can be ex-
plained by freezing out of molecules onto grains in the cold collapse phase,
followed by evaporation and high-temperature chemical reactions when
the young star heats the envelope. Thus, the chemical composition can
be a powerful diagnostic tool. A detailed study of a set of infrared-bright
massive young stars reveals systematic increases in the gas/solid ratios
and abundances of evaporated molecules with temperature. This `global
heating' plausibly results from the gradual dispersion of the envelopes.
We argue that these objects form the earliest phase of massive star for-
mation, before the `hot core' and ultracompact H II region phase.
Introduction
Massive star-forming regions such as Orion-KL and SgrB2 (L=10 4 10 6 L )
have traditionally been prime targets for astrochemistry owing to their bright
molecular lines (e.g., Blake et al. 1987, Nummelin et al. 1998). Thousands
of lines belonging to 100 dierent molecules have been detected. In recent
years, submillimeter telescopes combined with new ground- and space-based
(ISO) infrared observations have allowed systematic studies of both the gas and
the dust in a large sample of massive young stars (see van Dishoeck & Blake
1998, van Dishoeck & van der Tak 2000 for reviews). Dramatic variations in
the chemical composition are found, even between dierent objects in the same
cloud: some massive young stars (called `hot cores') show strong submillimeter
lines of saturated organic molecules like CH 3 OCH 3 , other objects are rich in
SO and SO 2 lines, whereas yet other sources have only simple molecules such
as radicals and ions (e.g., Helmich & van Dishoeck 1997, Hatchell et al. 1998).
Striking dierences are also seen in the infrared spectra: the coldest, deeply
embedded sources show strong absorption by silicates and ices, whereas the
more evolved objects have prominent PAH emission features and strong atomic
and ionic emission lines (van den Ancker et al. 2000).
The most successful models for explaining these dierent chemical char-
acteristics invoke accretion of species onto grains in an icy mantle during the
cold (pre-)collapse phase, followed by grain-surface chemistry and evaporation
of ices once the young star has started to heat its surroundings. The evap-
orated molecules subsequently drive a rapid high-temperature `hot core' gas-
phase chemistry for a period of  10 4 10 5 yr, resulting in complex, saturated
10

Table 1. Chemical characteristics of massive star-forming regions
Component Chemical Submillimeter Infrared Examples
characteristics diagnostics diagnostics
Dense cloud Low-T chemistry Ions, long-chains Simple ices SgrB2 (NW)
(HC 5 N, ...) (H 2 O, CO 2 )
Cold envelope Low-T chemistry, Simple species Ices N7538 IRS9,
Heavy depletions (CS, H 2 CO) (H 2 O, CO 2 , CH 3 OH) W 33A
Inner warm Evaporation High Tex High gas/solid, High GL 2591,
envelope (CH 3 OH) Tex , Heated ices GL 2136
(C 2 H2 , H 2 O, CO 2 )
Hot core High-T chemistry Complex organics Hydrides Orion hot core,
(CH 3 OCH 3 , CH 3 CN, (OH, H 2 O) SgrB2(N),G34.3
vib. excited mol.) W 3(H 2 O)
Out ow: Shock chemistry, Si- and S-species Atomic lines, Hydrides W 3 IRS5,
Direct impact Sputtering (SiO, SO 2 ) ([S I], H 2 O) SgrB2(M)
PDR, Compact Photodissociation, Ions, radicals Ionic lines, PAHs S 140,
H II regions Photoionization (CN/HCN, CO + ) ([NeII], [CII]) W 3 IRS4
organic molecules (e.g., Charnley et al. 1992). The abundance ratios of species
such as CH 3 OCH 3 /CH 3 OH may serve as `chemical clocks' in this period. Once
most of the envelope has cleared, the ultraviolet radiation can escape and forms
a photon-dominated region (PDR) at the surrounding cloud material, in which
molecules are dissociated into radicals (e.g., HCN ! CN) and PAH molecules
excited to produce infrared emission. Finally, the (ultra-)compact H II region
gives rise to strong ionic lines due to photoionization. Table 1 summarizes the
chemical characteristics of the dierent physical components present in mas-
sive YSOs. In single dish submillimeter data, most of these components are
blurred together, and sophisticated radiative transfer techniques are required to
disentangle them and derive reliable abundances.
A Sample of Embedded, Infrared-Bright Massive Young Stars
The availability of complete ISO-SWS spectra from 2.5{45 m provides a unique
opportunity to study massive young stars through a combination of infrared and
submillimeter spectroscopy. Van der Tak et al. (2000a) have selected a set of 9
massive young stars which are bright at mid-infrared wavelengths (12 m ux >
100 Jy) and relatively nearby d 2 kpc. The sources are all in an early, deeply
embedded evolutionary state (comparable to the `Class 0 or I' stage of low-mass
stars), as indicated by their weak radio continuum emission and absence of ionic
lines and PAH features. Complementary JCMT and OVRO data have been
obtained. For comparison, 5 infrared-weak sources with similar luminosities are
studied at submillimeter only, including hot cores and ultracompact H II regions.
The density structure of the sources has been derived from submillimeter
continuum data and from CS lines with a large range of critical density, whereas
the temperature structure is calculated self-consistently from the luminosity of
the sources. Assuming a power-law density prole n(r) = n o (r=r o )  , best-t

values of  = 1:0 1:5 are found for the infrared-bright sample, whereas the hot
core/compact H II region sample requires higher values,   2.
The ISO-SWS spectra show absorption by various gas-phase molecules, in
addition to strong features by ices. Molecules such as CO 2 , H 2 O, CH 4 , HCN and
C 2 H 2 (see van Dishoeck & van der Tak 2000 for refs.) have been detected. Both
the gas-phase and ice results can be plotted as functions of the `average' tem-
perature of the envelope, measured either from the 45/100 m continuum ux
ratio or the excitation temperature of CO and other molecules. The abundances
of H 2 O, HCN and C 2 H 2 increase by factors of >
 10 with increasing temperature
whereas the H 2 O and CO 2 ice abundances show a decrease by an order of mag-
nitude, consistent with evaporation of the ices. Moreover, the ices show evidence
for systematic heating with the same temperature indicators. These systematic
trends are also found in the JCMT submillimeter data of the sources. In par-
ticular, a careful analysis of the CH 3 OH data reveals a `jump' in its abundance
from  10 9 to  10 7 at T  100 K in the inner region of the warmer sources,
consistent with evaporation of CH 3 OH-ices (van der Tak et al. 2000b). These
warm sources however lack the typical crowded spectra of `hot cores' like Orion-
KL. Either their `hot cores' are still too small to be picked up by the single dish
beams or the hot chemistry has not yet had suôcient time to develop, or both.
A signicant result of this study is that even within the narrow evolution-
ary range of 10 4 10 5 yr, clear physical and chemical dierentiation of the
sources can be found. Since the indicators involve temperatures ranging from
<50 K (evaporation of ices) to 1000 K (T ex of gas-phase molecules), this sug-
gests that the heating is not a local eect, but that it occurs throughout the
envelope. This `global warming' correlates with the ratio of envelope mass over
stellar mass, suggesting that with time, the envelope is dispersed by the star,
resulting in a higher temperature throughout the envelope. We therefore argue
that the deeply-embedded infrared-bright objects form the earliest evolutionary
stage, followed by the `hot core' stage and nally the ultracompact H II region
stage. Future instrumentation such as ALMA, SOFIA, FIRST and NGST will
be essential to further develop these evolutionary diagnostics.
References
Blake, G.A., Sutton, E.C., Masson, C.R., & Phillips, T.G. 1987, ApJ, 315, 621
Charnley, S.B., Tielens, A.G.G.M., & Millar, T.J. 1992, ApJ, 399, L71
Hatchell, J. et al. 1998, A&AS, 133, 29
Helmich, F.P. & van Dishoeck, E.F. 1997, A&AS, 124, 205
Nummelin, A. et al. 1998, ApJS, 117, 427
van den Ancker, M. et al. 2000, A&A, 358, 1035
van der Tak, F.F.S., van Dishoeck, E.F., Evans, N.J., & Blake, G.A. 2000a, ApJ,
537, 283
van der Tak, F.F.S., van Dishoeck, E.F. & Caselli, P. 2000b, A&A, 361, 327
van Dishoeck, E.F. & Blake, G.A. 1998, ARAA, 36, 317
van Dishoeck, E.F. & van der Tak, F.F.S. 2000, in `Astrochemistry: From Molec-
ular Clouds to Planetary Systems', IAU Symposium 197, p. 97

Massive Star Birth - JD3 IAU GA
ASP Conference Series, Vol. 000, 2001
P. S. Conti & E. B. Churchwell, eds.
The New Generation of Ionization and Recombination
Fronts
J.E. Dyson & T.W. Hartquist
Department of Physics and Astronomy, University of Leeds, Leeds LS2
9JT, UK
R.J.R. Williams
Department of Physics and Astronomy, University of Cardi, PO Box
913, Cardi CR24 3YB, UK
M.P. Redman
Department of Physics and Astronomy, University College London,
London WC1E 6BT, UK
Abstract. We summarize previous work on hydrodynamic ionization
fronts and new work where magnetic elds are incorporated into their
structures. We describe recent work on recombination front structures
and outline renements which need to be made to both them and ioniza-
tion fronts.
Introduction
Expanding H ii regions are bounded by ionization fronts (IF). IFs divide into R-
types, which move supersonically into upstream neutral gas, and D-types, which
move subsonically (Kahn 1954). At the edges of the velocity regions allowed by
the jump conditions, the ionized ow out of these `critical' IFs takes place at the
isothermal ionized sound speed. IFs in which the gas both enters and leaves the
front supersonically are called weak R-type, while in weak D-types the gas enters
and leaves subsonically. Strong R-type IFs, in which the gas would experience a
supersonic-subsonic transition, are overdetermined and do not exist. In strong
D-type IFs, overheating allows a subsonic-supersonic transition within the IF.
Sudden ionization produces an initially weak R-type IF. Once its velocity
drops to about twice the sound speed in the ionized gas, a shock moves out of
the IF, which becomes weak D-type (Axford, 1961; Goldsworthy 1961; Mathews,
1965; Lasker 1966a). This sequence is relevant when the ionized gas is conned
to a region bounded by an IF. The IF incident on an isolated neutral clump
can stall and the ionized gas ows from its surface as a wind. The IF is strong
D-type and the gas accelerates through the isothermal sonic point inside the IF.
Axford's (1961) study solved the internal structure equations using a simple
cooling function and gave the full range of hydrodynamic structures and the
evolution from weak R through strong D- and weak D-types. He showed that in
strong-D IFs, the specic enthalpy of the gas in the front reached a maximum at
13

the adiabatic sound speed. Most other steady state studies concentrate on the
D and weak R solutions. Various authors included numerous ions, molecular
hydrogen and radiative transfer (Hjellming 1966; Mendis 1969; Briscoe 1972;
Harrington 1977; Hill 1977; Mason 1980; Bertoldi 1989).
The stability and evolution of IFs are better addressed by the calculation
of time-dependent global ows. Early calculations were by Mathews (1965) and
Lasker (1966a), while Yorke (1986) reviews subsequent work. The nite spatial
and temporal dynamic range of these calculations can lead to articial features
within the IF structures themselves but the global evolution of the regions is
generally still well described (cf. Bedijn & Tenorio-Tagle 1984).
There have been many signicant recent developments in the study of IFs.
Strong-D type IFs around clumps have been directly observed at the surface
of photoionized clumps and disks in H ii regions such as Orion (e.g. Henney
& O'Dell 1999), and planetary nebulae (e.g. Dyson et al 1993), and are found
to be of great importance in the structure of ultra-compact H ii regions (e.g.
Dyson et al 1995). Sysoev (1997) found long-wavelength instabilities to occur in
D-type IFs, while Williams (1999) described a shadowing instability of R-type
IFs. Franco et al. (2000) review the most recent studies of global structures.
MHD Ionization Fronts
All diuse astronomical sources are clumpy (e.g. Hartquist & Dyson 1993). To
date, studies of clump photoevaporation (e.g. Dyson 1968; Bertoldi 1989; Arthur
& Lizano 1997) have treated non-magnetized clumps. However, evidence (e.g.
Myers & Khersonsky 1995) suggests that the ratio of magnetic pressure to ther-
mal pressure in diuse clouds may be up to ten or so, and even higher in the
CO emitting clumps in the Rosette Molecular Cloud.
Lasker (1966b) rst studied the jump conditions across MHD IFs. Redman
et al. (1998) extended this work to regions of much higher magnetic-to-thermal
pressure ratios (). The fast mode sound speed plays the role of the ordinary
sound speed in the hydrodynamic case. Importantly, the forbidden velocity
gap decreases as  increases. D-critical IFs are aected signicantly if   > 1.
R-critical IFs are aected only at relatively high (  > 50) (Redman et al. 1998).
Across obliquely magnetized IFs, there are R and D-type IFs about both
the fast- and the slow-mode speeds (Williams et al. 2000). If v z;o and v z;i are
the gas velocity components perpendicular to the IF in the neutral and ionized
gas respectively, v f and v s as the fast and slow mode velocities respectively and
v A the Alfven velocity, the possible IFs become fast R-type (v z;o ; v z;i > v f ); slow
R-type (v A > v z;i ; v z;o > v s ); fast D-type (v f > v z;i ; v z;o > v A ); slow D-type
(v s > v z;o ; v z;i ). R-type IFs are compressive while D-type IFs are rarefying, as
in the hydrodynamic case.
In obliquely-magnetized IFs, we specify two parameters in the upstream
neutral gas, the ratios  = B 2
x =8c 2 and  = B 2
z =8c 2 . B x and B z are the
magnetic eld strength components respectively parallel to and perpendicular
to the plane of the IF,  is the gas density and c is the isothermal sound speed.
Slow D-critical velocities are strongly aected for even small values of  and
. High enough transverse magnetic elds allow fast-D and slow-R IFs with
velocities forbidden to hydrodynamic IFs. Since fast mode shocks increase the

transverse eld, the transition from a fast R-type front to a fast D-type front
can take place by the generation of a fast shock ahead of the IF, analogous to
the transition from a weak R-type to a weak D-type by the generation of a shock
in the hydrodynamic case. Slow mode shocks switch o the transverse magnetic
eld and allow a transition sequence such as fast D to slow R to slow D. The
interpretation of magnetic eld strengths associated with expanding H ii regions
(e.g. Roberts, Crutcher & Troland 1995) requires considerable caution (Williams
et al 2000).
The internal structures of MHD IFs are complex. Williams & Dyson (in
preparation) have generated the internal structures of a wide variety of MHD
IFs using Axford's (1961) approximations.
Recombination Fronts
Ionized gas recombines if the ionization source is removed (e.g. gas cooling behind
supernova driven shock waves). Radiatively ionized regions recombine if the
ionizing source is suôciently dimmed. Generally, the recombining regions are
extended but there are astrophysically interesting situations where there is a
relatively sharp ionized-neutral interface, a recombination front (RF), across
which there is a steady ux of ionized gas. The velocity of the ionized gas through
the RF denes the interface velocity. The thicknesses of RFs are determined by
the recombination of ionized gas advected through them and can be broader
than IFs (although their thicknesses are comparable to the thermal relaxation
distance behind an IF, for similar densities of ionized gas).
RFs are important in, e.g., accretion ows onto stars with strong UV radi-
ation elds (Mestel 1953), ultra-compact H ii regions (Dyson et al. 1995 et seq;
Lizano et al. 1996) and recombination zones in mass-loaded jets (Redman &
Dyson 1999).
Newman & Axford (1968) examined the properties of RFs and found there
existed the analogies of the weak R, strong R and weak D IFs but there was
no analogy to strong D IFs. Williams & Dyson (1996) re-examined this work
and found an additional small, but important, class of transonic solutions. They
also found extra supersonic-subsonic solutions generated partly by the insertion
of shocks in transonic solutions.
Transonic solutions become important when initially subsonic global ows
approach the sonic point as a result of geometrical divergence and/or mass load-
ing, (e.g. if gas `leaks' through the surface of an H ii region). Exiting supersonic
ows can match up to a range of external pressures simply by the insertion of a
shock at some radius. Williams (unpublished) has calculated the structures of a
series of mass-loaded ows which start subsonically and emerge supersonically
through an RF.
Conclusions
We have described our recent work on the structure of MHD IFs and on RFs.
Much more realistic models are still required: to accurately discriminate between
MHD and purely hydrodynamic IFs; to determine the eects of magnetic elds
on global models of H ii regions and photoionized clumps; to study the eects of

clumpy media and the transition from IFs to RFs as a result of mass injection.
A subject area which seemed to be reasonably well understood forty years ago
has become dynamically active again.
References
Arthur S.J., Lizano S. 1997, ApJ., 484
Axford W.I. 1961, Phil. Trans. R. Soc. Lon. A, 253, 301
Bedijn P.J., Tenorio-Tagle C. 1984, A&A, 135, 81
Bertoldi F. 1989, ApJ., 346, 735
Briscoe F. 1972, unpublished thesis, University of Leeds, UK
Dyson J.E. 1968, Ap&SS, 1, 388
Dyson J.E., Hartquist T.W., Biro S. 1993, MNRAS, 261, 430
Dyson J.E., Williams R.J.R., Redman M.P. 1995, MNRAS, 277, 700
Goldsworthy F.A. 1961, Phil. Trans. R. Soc. Lond. A, 253, 277
Harrington J.P. 1977, MNRAS, 179, 63
Hartquist T.W., Dyson J.E. 1993, Quart. J. R. Astr. Soc., 34, 57
Henney W.J., O'Dell C.R. 1999, AJ, 118, 235
Hill, J.K. 1977, ApJ., 212, 685
Hjellming R.M. 1966, ApJ., 143, 420
Kahn F.D. 1954, Bull. Astr. Inst. Netherlands, 12, 187
Lasker B.M. 1966a, ApJ., 143, 700
Lasker B.M. 1966b, ApJ., 146, 471
Mason D.J. 1980, A&A, 92, 117
Mathews W.G. 1965, ApJ., 142, 1120
Mendis D.A. 1969, MNRAS, 142, 241
Mestel L. 1953, MNRAS, 114, 437
Myers P.C., Khersonsky V.K. 1995, ApJ., 442, 186
Newman R.C., Axford W.I. 1968, ApJ. 151, 1145
Redman M.P., Dyson J.E. 1999, MNRAS, 302, L17
Redman M.P., Williams R.J.R., Dyson J.E., Hartquist T.W., Fernandez B.R.
1998, A&A, 331, 1099
Roberts D.A., Crutcher R.M., Troland T.H. 1995, ApJ., 442, 208
Sysoev N.E. 1997, Astron. Lett., 23, 409
Williams R.J.R. 1999, MNRAS, 310, 789
Williams R.J.R., Dyson J.E. 1996, MNRAS, 279, 987
Williams R.J.R., Dyson J.E., Hartquist T.W. 2000, MNRAS, 314, 315
Yorke H.W. 1986, ARAA, 24, 246

Massive Star Birth - JD3 IAU GA
ASP Conference Series, Vol. 000, 2001
P. S. Conti & E. B. Churchwell, eds.
Jets and Out ows from Massive Protostars
Karl M. Menten
Max-Planck-Institut fur Radioastronomie,
Auf dem Hugel 69, D-53121 Bonn, Germany
Abstract. Molecular out ows are found in most regions of massive star
formation. However, evidence for highly collimated jets from massive
protostars is still elusive. This paper brie y summarizes the observational
status.
Introduction { Out ows and High Mass Star Formation
Mass out ow from young stellar objects is a ubiquitous phenomenon in star-
forming regions (SFRs) and seems to commence in the earliest phases of pro-
tostellar development; see Bachiller (1996) and Churchwell (1999) (1996a) for
a survey of 122 high mass SFRs, 90% of which exhibit evidence for out ow
activity.
Many low mass protostars show highly collimated, jet-driven bipolar ows
observable in atomic and ionic optical lines and in infrared- and millimeter-
wavelength molecular emission. Out ows in high mass SFRs frequently have a
bipolar morphology as well (Shepherd & Churchwell 1996b). However, evidence
for the existence of highly collimated jets originating from bona de massive
protostars is presently sparse at best. Whether or not such jets exist may provide
important information on the high mass star formation process itself, since in
a variety of astrophysical settings the presence of a collimated jet implies an
accretion disk at its origin (Livio 1999) and it is presently unclear whether
massive stars form by accretion (involving a disk or not) or in other ways, such
as coalescence of lower mass stars (Stahler et al. 2000; Bonnell et al. 1998). In
this context it is important to mention the observation that in many cases the
mass in the out owing material is greater (by up to factors of a few) than any
plausible value for the mass of the protostar driving it. This lead Churchwell
(1997) to conclude that infalling matter has to be diverted into the bipolar
out ow. While the mechanism for this remains unclear, these observations seem
to support an accretion scenario (Norberg & Maeder 2000). In the following
we summarize the observational information on out ows and jets in high-mass
SFRs.
Molecular Line Observations
The rst evidence for high velocity molecular gas in high-mass SFRs came from
observations of high velocity components in spectra of the 22 GHz water vapor
(H 2 O) maser transition, which in some regions, such as Orion-KL or W49 cover
17

velocity ranges up to 200 km s 1 around the systemic velocity. Strelnitskii
& Sunyaev (1973) were the rst to propose mass out ow to explain the high
velocity H 2 O and this interpretation was conrmed by VLBI proper motion
studies, which generally show expansion, although not necessarily in a highly
collimated way (e.g., Gwinn et al. 1992).
With the proliferation of molecular line observations at millimeter- and
infrared-wavelengths, rotational lines from CO and vibrational-rotational emis-
sion from shock-excited H 2 became major tracers of out owing gas. Indeed,
wide-eld H 2 surveys have become an eective tool for nding out ows (e.g.
Stanke et al. 2000) in addition to high velocity CO and H 2 O maser searches.
In particular, observations of the CO lines, which trace the bulk of the high
velocity material, allow estimates of the physical properties of the out owing gas.
Churchwell (1999) compiles observational data and derived physical quantities
for known massive bipolar out ows. In summary, he nds out ow masses, M f ,
between a few and 5000 M , mass out ow rates, _
M f , ranging from 3  10 5 to
10 2 M yr 1 , kinetic energies from 10 46 to 6  10 48 ergs, and out ow luminosi-
ties, L f , between 0.2 and 1300 L . Thus, mean out ow masses and luminosities
are 100 times greater than the values found in low mass SFRs. Notably, _
M f in-
creases monotonically with L bol for L bol ranging from 1 to 10 6 L (Shepherd &
Churchwell 1996b). Churchwell (1999) nds _
M f / L 0:7
bol ; see also Henning et al.
(2000). This may provide important information on the out ow mechanism(s)
and, as discussed by Norberg & Maeder (2000), on the star-forming process(es)
itself.
Jets from Massive Protostars and Young Stellar Objects?
The interpretation of the H 2 O proper motion data mentioned above calls for a
jet to explain the maser acceleration (Mac Low et al. 1994). Direct evidence
for this scenario is provided by the VLA detection by Reid et al. (1995, see
also Wilner et al. 1999) of a non-thermal radio source with double jet mor-
phology whose centroid is coincident with the center of expansion determined
from VLBI proper motion measurements of the H 2 O maser out ow in this re-
gion (Alcolea et al. 1993). The jet/H 2 O source is coincident with a molecular
hot core in the vicinity of the archetypical ultracompact HII region W3(OH).
Using millimeter interferometry, Wyrowski et al. (1999) resolve this core in at
least three hot ( 200 K) condensations, one of which is exactly centered on
the non-thermal jet/H 2 O source, for which a luminosity of a few times 10 4 L
is derived, indicating that the embedded object is of intermediate, rather than
high mass.
Sensitive radio continuum observations have the potential to reveal jets from
and, even more fundamentally, the exact locations of embedded massive stars.
The apparent paradox that the by far the most luminous object in a given region
may be diôcult to locate is well illustrated by the example of Orion-KL. Here,
VLA observations by Menten & Reid (1995) show that the weak thermal radio
source I is coincident with the unique SiO maser source found in this region.
Given that the SiO masers almost certainly require an exciting source with a
luminosity in excess of 10 4 L strongly suggests that the radio source indeed

is coincident with the massive protostar powering the region, which due to its
extreme extinction is impossible to locate even at infrared wavelengths.
Another example for a radio source coincident with the center of activity is
found in the Cepheus A region, where Garay et al. (1996) interpret the spatially
elongated thermal emission from their source 2 as arising from a collimated jet
of ionized gas (see also Hoare & Garrington 1995).
In summary, the evidence for collimated jets in high-mass SFRs is inconclu-
sive at present. One lesson learned from the examples described above is that
jet emission from protostellar objects, thermal or non-thermal, might in general
be quite weak. With present technology, all of the radio sources discussed above
would be impossible to detect in more distant regions like W49. Nevertheless,
given the extreme extinctions in the regions in questions, radio imaging with
high resolution, high sensitivity, and high dynamic range may be the only way
to nd such jets.
References
Alcolea, J., Menten, K. M., Moran, J. M., Reid, M. J. 1992, in Astrophysical
Masers, eds. A. W. Clegg & G. E. Nedoluha (Heidelberg: Springer), 225
Bachiller, R. 1996, ARA&A, 34, 111
Bonnell, I. A., Bate, M. R., Zinnecker, H. 1998, MNRAS, 298, 93
Churchwell, E. 1997, ApJ, 479, L59
Churchwell, E. 1999, in Unsolved Problems in Stellar Evolution, ed. M. Livio
(Cambridge University Press), 41
Garay, G., Ramrez, S., Rodrguez, L. F., Curiel, S., Torrelles, J. M. 1996, ApJ,
459, 193
Gwinn, C. R., Moran, J. M., Reid, M. J. 1992, ApJ, 393, 149
Henning, T., Schreyer, K., Launhardt, R., Burkert, A. 2000, A&A, 353, 211
Hoare, M. G., Garrington, S. T. 1995, ApJ, 449, 874
Livio, M. 1999, Phys. Rep., 311, 225
Mac Low, M.-M., Elitzur, M., Stone, J. M., Konigl, A. 1994, ApJ, 427, 914
Menten, K. M., Reid, M. J. 1995, ApJ, 445, L157
Norberg, P., Maeder, A. 2000, A&A, 359, 1025
Reid, M. J., Argon, A., Masson, C. R., Menten, K. M., Moran, J. M. 1995, ApJ,
443, 238
Shepherd, D. S., Churchwell, E. 1996a, ApJ, 457, 267
Shepherd, D. S., Churchwell, E. 1996b, ApJ, 472, 225
Stahler, S. W., Palla, F., Ho, P. T. P. 2000, in Protostars and Planets IV, eds.
V. Mannings, A. P. Boss, & S. S. Russell (Tucson: Univ. Arizona Press),
327
Stanke, T., McCaughrean, M. J., Zinnecker, H. 2000, A&A, 355, 639
Strelnitskii, V. S., Sunyaev, R. A. 1973, Soviet Astron., 16, 579
Wilner, D. J., Reid, M. J., Menten, K. M. 1999, ApJ, 513, 775
Wyrowski, F., Schilke, P., Walmsley, C. M., Menten, K. M. 1999, ApJ, 514, L43

Massive Star Birth - JD3 IAU GA
ASP Conference Series, Vol. 000, 2001
P. S. Conti & E. B. Churchwell, eds.
Disks and Jets in High-Mass Young Stellar Objects
Riccardo Cesaroni
Osservatorio Astrosico di Arcetri, Firenze, Italy
Introduction
The formation of stars is a complex process which is poorly understood at
present, although recently important progress has been made both on a the-
oretical and observational ground. It is clear that star formation must proceed
through contraction of a large molecular clump into a dense optically thick proto-
stellar core: the obvious consequence is that conservation of angular momentum
must force the material to spin up and atten. Thus, formation of disks around
newly formed stars is a very sensible expectation. Indeed, the recent develop-
ment of instruments like the Hubble Space Telescope (HST) and the millimeter
interferometers has allowed detection of several disks around low-mass young
stellar objects (YSOs), such as the Keplerian disk in GGTau (Guilloteau et al.
1999) and that seen with the HST in HH 30 (Burrows et al. 1996). The situa-
tion is quite dierent for high-mass YSOs. In this case, the evidence for disks is
scarce, although a priori one would expect these to be more massive than those
in low-mass YSOs and hence easier to detect. Various eects may complicate
this simple-minded picture: for instance, magnetic eld is likely to play an im-
portant role coupling the inner part of the collapsing cloud to its outer layers,
thus making angular momentum conservation diôcult to apply to any single
\portion" of the cloud; depending on the ratio between disk and stellar mass,
the disk may be unstable and hence short-lived; the eects of the stellar wind
and radiation have to be taken into account; nally, the mass and size of the
disk depend on the accretion process, which is not well understood. All these
caveats probably explain why disks around massive YSOs are diôcult to detect.
It is also important to stress that the formation of the disk and the ejection
of a jet/out ow from the newly formed star might be strictly related. Therefore,
one should investigate both phenomena at the same time when studying the
formation of stars. In the following, we will summarise the current evidence for
disks in massive YSOs and the relevance of jets as a tool to prove their existence.
Evidence for disks in massive YSOs
Unlike the case of low-mass stars, a complete theory for the formation of massive
stars has not been developed yet and it is hence diôcult to guess the expected
size of disks in these objects. However, it seems unlikely that they can be much
greater than 1000 AU, which correspond to 1 00 at a distance of 1 kpc. Since the
majority of massive stars are located at much larger distances, the disk angular
diameters are expected to be  <1 00 , namely comparable to the angular resolution
of the millimeter interferometers currently available. Furthermore, unlike low-
20

mass stars, massive ones reach the zero age main sequence still deeply embedded
in their parental cloud, with visual extinctions up to 1000 mag, which makes
impossible to detect the disks even at NIR wavelengths. Maser lines are instead
an excellent probe for these dense regions: in fact, they are emitted in the
radio regime, where extinction is negligible, and trace the velocity eld of the
molecular gas on scales as small as a few AU, where they can be observed thanks
to VLBI measurements at centimeter wavelengths. However, masers can sample
the gas only in a limited number of points, whereas thermal lines { although
observable with much worse angular resolution { have the advantage to sample
the bulk of the molecular gas. It is hence clear that both thermal and maser
lines are necessary to investigate the global structure of disks.
Table 2. Possible disks in high-mass YSOs
Tracer Source M M Disk R Disk Reference
(M) (M) (AU)
CH3OH UC Hii 0.5{73 | 95{3700 Norris et al. 1998
regions Walsh et al. 1998
SiO Orion-KL 4.4 | 80 Barvainis 1984
Plambeck 1990
H2O Ceph A 20 7040 300 Torrelles et al. 1996,1998
W75N(B) ?  <8 200 Torrelles et al. 1997
C 18 O, G192.16{3.82 5{10 7{32 500 Shepherd et al. 1998
H2O Shepherd & Kurtz 1999
CS G92.67+3.07 4{7 10{20 7000 Bernard et al. 1999
CH3CN, Hot Cores: O{B 10{10 3 10 3 {10 4 Keto et al. 1987
NH3 W3OH, G10.62, stars Wink et al. 1994
G10.47, G19.61, Cesaroni et al. 1994
G29.96, G31.41, Olmi et al. 1996
W51e2, W51e8, Cesaroni et al. 1997,1999
IRAS 20126 Zhang et al. 1998a,1998b
In Table 1 we summarise the current evidence for disks associated with high-
mass YSOs. With \high-mass" we will refer to (proto)stars with luminosities
above 10 3 L . Such an evidence basically consists of the detection of attened
cores or chains of maser spots, with velocity gradients along their major axis.
In the table we give the observed tracer, the source name, the mass of the star,
the mass and radius of the disk, and the corresponding reference.
Problems and conclusions
Although several of the examples listed in Table 1 look quite convincing, still
an iron-clad detection of a disk around a massive YSO has to be found. Various
problems can hinder a fully convincing detection. Sometimes the \disk" is just an
eect of poor angular resolution, and it \disappears" once better observations are
available. In other cases, dierent tracers reveal velocity gradients in completely
dierent directions. However, the problem which aects most of the cases is
the lack of an independent check on the direction of the disk axis: under this
respect, the detection of a jet perpendicular to the plane of the disk is a crucial

test on the disk hypothesis. Only few of the sources in Table 1 are associated
with a jet/out ow, coincident with the axis of the disk: the best examples are
G192.16{3.82 and IRAS 20126+4104. In particular, the latter has been studied
in great detail demonstrating that the jet does not bend going from a scale of
0.1 pc (Cesaroni et al. 1999) to 100 AU (Moscadelli et al. 2000) and is
perpendicular to the Keplerian disk seen in the NH 3 (Zhang et al. 1998b) and
CH 3 CN lines on 2000 AU. The advent of new (sub)millimeter interferometers
such as ALMA will dramatically improve the angular resolution of molecular
line observations, thus making possible to conrm the existence of disks around
massive YSOs by resolving their structure and deriving their rotation curve.
For the time being, we can only conclude that very likely disks exist also in
high-mass YSOs, with masses of 10{10 4 M and radii of 10 2 {10 4 AU. Despite
these are several orders of magnitude larger than in low-mass YSOs, disks in
massive YSOs remain elusive and this might be due to their shorter life-time,
for which at least two explanations are possible: the fact that the disk is much
more massive than the associated (proto)star, which makes it highly unstable;
and the large luminosity, UV ux, and stellar wind of the O-B (proto)star, which
accelerate the destruction of the disk.
References
Barvainis R. 1984, ApJ, 279, 358
Bernard J.P., Dobashi K., Momose M. 1999, A&A, 350, 197
Burrows C.J. et al. 1996, ApJ, 473, 437
Cesaroni R. et al. 1994, ApJ, 435, L137
Cesaroni R., Felli M., Testi L., Walmsley C.M., Olmi L. 1997, A&A, 325, 725
Cesaroni R. et al. 1999, A&A, 345, 949
Guilloteau S., Dutrey A., Simon M. 1999, A&A, 348, 570
Keto E.R., Ho P.T.P., Haschick A.D. 1987, ApJ, 318, 712
Moscadelli L., Cesaroni R., Rioja M.J. 2000, A&A, in press
Norris R.P. et al. 1998, ApJ, 508, 275
Olmi L., Cesaroni R., Walmsley C.M. 1996, A&A, 307, 599
Plambeck R.L., Wright M.C.H., Carlstrom J.E. 1990, ApJ, 348, 65
Shepherd D.S., Watson A.M., Sargent A.I., Churchwell E. 1998, ApJ, 507, 861
Shepherd D.S., Kurtz S.E. 1999, ApJ, 523, 690
Torrelles J.M. et al. 1996, ApJ, 457, 107
Torrelles J.M. et al. 1997, ApJ, 489, 744
Torrelles J.M. et al. 1998, ApJ, 509, 262
Walsh A.J., Burton M.G., Hyland A.R., Robinson G. 1998, MNRAS, 301, 640
Wink J.E. et al. 1994, A&A, 281, 505
Zhang Q., Ho P.T.P., Ohashi N. 1998a, ApJ, 494, 636
Zhang Q., Hunter T.R., Sridharan T.K. 1998b, ApJ, 505, L151

Massive Star Birth - JD3 IAU GA
ASP Conference Series, Vol. 000, 2001
P. S. Conti & E. B. Churchwell, eds.
Massive Star Formation: New Results
Ed Churchwell
Washburn Observatory, University of Wisconsin, 475 North Charter
Street, Madison, WI 53706
Introduction
In the past year several new observations with important implications for massive
star formation (MSF) have been obtained. Among these were the detection of
UC HII region precursor candidates at 350m and the discovery of many hard
X-ray point sources in the Orion and W3 MSF regions. These observations are
summarized below.
Precursors of UC HII Regions
Hunter et al. (1998, 2000) imaged 25 MSF regions at 350m to search for
candidate precursors of UC HII regions (i.e. luminous submm/FIR emission,
maser sources, and no HII emission). Of the 28 sources detected, 10 appear to
be UC HII precursor candidates.
Figure 1 shows an example of one (IRAS18317-0757 SMM2).
Cont peak flux = 2.3892E+00 JY/BEAM
Levs = 1.000E+00 (0.800, 1, 1.500, 2)
DECLINATION
(B1950)
RIGHT ASCENSION (B1950)
18 31 46 44 42 40 38
­07 56 30
45
57 00
15
30
45
58 00
12a. IRAS18317­0757
SMM1
SMM2
Figure 3.
IRAS18317-0757 SMM1 is
centered on the UC HII
region G23.96+0.15, SMM2
has an H 2 O maser and no
HII region. The mean phys-
ical properties (derived from
grey body models) of the 28
350m \hot cores" found by
Hunter et al. (2000) are
given in Table 1.
Table 3.
Range Mean
log L/L 4.3 - 6.1 5.0
log M/M 2.4 - 5.0 3.5
log NH 22.2 - 24.9 23.8
log (L/M) 0.9 - 2.7 1.5
The main conclusions are that MSF requires core masses >10 3 M and
densities 10 5 cm 3 .
23

X-Ray Emission from MSF Regions
Hofner, Churchwell, and Whitney (2000), using the high spatial resolution and
sensitivity of Chandra, found that the \extended, hard X-ray emission" ob-
served previously with ASCA by Hofner and Churchwell (1997), resolves into
several hundred X-ray point sources distributed over the entire W3 complex.
The sources at the core of W3 are very energetic with most of their energy emit-
ted in the 3-7.5 Kev range. Their hardness is enhanced by extinction of softer
X-rays by over-lying hydrogen.
X-ray sources lie at the peak of most of the UC HII regions in W3. The
locations of the X-ray sources suggests that they may be the ionizing star of the
UC HII components, although this still remains to be established. There are
also many X-ray sources that do not coincide with radio sources. The nature
of these sources are not yet known. They could be T Tau stars (i.e. low-mass
pre-main sequence stars), interacting close binaries, stars with strong coronal
activity, etc. Unfortunately, not enough data on these objects exist to establish
their nature.
The NIR sources of Megeath et al. (1996) and the X-ray sources in the core
of W3 show very few coincidences. The NIR sources are uniformly distributed in
the observed eld with little increase in density toward the UC HII regions. The
NIR sources appear to be objects located on the near side of the W3 molecular
cloud and may have little to do with the deeply embedded UC HII regions. This
is not the case for the hard X-ray sources. This leads one to the interesting
speculation that perhaps the best way to locate the ionizing stars of UC HII
regions is in the 2-8 Kev X-ray band, not in the NIR.
In summary, MSF regions are rich in hard X-ray sources, which probably
represent a variety of stellar types. The W3 study is also consistent with the
earlier detection of many X-ray sources in the Orion MSF region by Yamauchi
and Koyama (1993) and spectacularly conrmed by Garmire et al. (2000) using
Chandra.
References
Garmire, G., Feigelson, E. D., Broos, P., Hillenbrand, L. A., Pravdo, S. H.,
Townsley, L., Tsuboi, Y. 2000, AJ, in press.
Hofner, P., Churchwell, E. 1997, Ap. J. Lett., 486, L39
Hofner, P., Churchwell, E., Whitney, B. 2000, in preparation
Hunter, T. R., Churchwell, E., Watson, C., Cox, P., Benford, D.J., Roelfsema,
P.R. 2000, AJ, 119, 2711
Hunter, T. R., Neugebauer, G., Benford, D. J., Matthews, K., Lis, D. C., Ser-
abyn, E., Phillips, T. G. 1998, ApJ, 493, 97
Megeath, S. T., Herter, T., Beichman, C., Gautier, N., Hester, J. J., Rayner, J.,
Shupe, D. 1996, A& A, 307, 775
Yamauchi, S., Koyama, K. 1993, ApJ, 405, 268

Massive Star Birth - JD3 IAU GA
ASP Conference Series, Vol. 000, 2001
P. S. Conti & E. B. Churchwell, eds.
Environs and Formation of Massive stars
Susana Lizano
Instituto de Astronoma, UNAM, Apdo. Postal 72-3 (Xangari), 58089
Morelia, Michoacan, Mexico
Guido Garay
Departamento de Astronoma, Universidad de Chile, Casilla 36-D,
Santiago, Chile
INTRODUCTION
Massive stars in our Galaxy are born predominantly within the dense cores of
giant molecular clouds. They aect their environment very soon after a stellar
core has formed through their large rate of ionizing photons and their strong
stellar winds. Massive stars are formed in clusters with stellar densities n  
10 4 stars=pc 3 and sizes 0:2 0:4 pc, and seem to preferentially form near the
central region of the cores. Disks have been found in several sources (see review of
R. Cesaroni). Associated with massive stars also are bipolar molecular out ows,
which have masses, mass loss rates, and energies that are factors of  100 larger
than those of low mass stars (see review of Churchwell 1999).
An understanding of the physical processes that dominate during the early
stages of formation of massive stars and their in uence back on the molecular gas
from which they formed requires a detailed knowledge of the physical conditions
of the environment prior to and after the formation of the star. Here we present
an abstract of an extensive review on this subject by Garay & Lizano (1999).
For compactness, most of the references have been omitted.
Compact HII Regions
Regions of ionized gas around recently formed massive stars have sizes L 
0:005 0:5 pc, emission measures EM  210 6 10 9 pc cm 6 , electron densities
n e  2  10 3 3  10 5 cm 3 , and are often found in clusters. The observations
show that the width of the recombination lines from HII regions decreases as
size increase,
v obs (compact)>>
v obs (extended); the non-thermal part of the
line width being the dominant component.
Wood & Churchwell (1989) noted that the short dynamical ages of UCHII re-
gions, derived assuming a classical expansion into a constant density medium,
implies a too high present star formation rate. Dierent models have been pro-
posed to extend the time scale in the compact phase: i) in the champage ow
model, the HII region is conned in the direction of the high density region; ii)
in the bow-shock model, the ram pressure of a stellar wind balances the hydro-
dynamic pressure of the cloud material entering a static bow shock structure as
the star moves through a dense core; iii) in the photoevaporating disk model,
25

the ionizing photons of the star evaporate the surface of a circumstellar disk
producing a slow and dense photoevaporated disk wind which lasts as long as
the reservoir of gas in the disk is avalaible; iv) HII regions can achieve pressure
equilibrium at small radii if they are expanding in a warm and high density
environment; v) mass loading of stellar winds via hydrodynamical ablation or
photoevaporation of clumps inside the HII region can increase the density of
stellar winds producing a recombination front at a small radius.
Molecular environment
The presence of dense, hot, molecular gas toward regions of newly formed stars is
quite common. High resolution observations yield information of the structure,
physical properties and kinematics of hot molecular cores (HMCs) (see review
of Kurtz et al. 2000). HMCs have sizes L < 0:1 pc, densities n(H 2 )  10 4
7  10 7 cm 3 , masses M  10 3  10 2 M , temperatures T > 100 K, and
velocity dispersions
v  4 10 km s 1 : HMCs are found in the vicinity of
UCHII regions, thus they can be internally or externally heated: the densest
and more massive cores could be the cradles of massive stars while the less
dense and less massive hot cores would just be remnant molecular core material
that has survived the powerful eects of the formation of massive stars.
HMCs and compact HII regions are immersed in massive molecular cores
with sizes L  0:3 1 pc, densities n(H 2 )  2  10 4 3  10 6 cm 3 , masses
M  10 3 3  10 4 M , temperatures T  25 50K, and velocity dispersions,

v  2 3 km s 1 . Massive cores are highly inhomogeneous and contain  20%
of the total gas in Giant Molecular Clouds.
HMCs and the Formation of Massive Stars
Some luminous HMCs are not associated with an UCHII region. Walmsley
(1995) proposed that these cores host a young massive OB-type star (or stars)
undergoing an intense accretion, which quenches the development of a detectable
UCHII region. Kaufman, Hollenbach, & Tielens (1998) calculated the temper-
ature structure of dense hot cores with either internal or external illumination.
They found that internally heated cores can more easily produce large column
densities (NH 2  10 23 cm 2 ) of hot (T > 100 K) gas than externally heated
cores.
Recently, Osorio et al. (1999) modeled the dust thermal spectrum of several
HMCs as massive envelopes accreting onto young massive central B-type stars.
This model can reproduce the observed spectra of several HMCs. In order to t
the available data, one requires central early B-type stars, and mass accretion
rates _
M >  6  10 4 M yr 1 . These objects are young, t age  6  10 4 yr, and
the accretion luminosity is larger than the stellar luminosity, L acc > L  . Also,
the mass weighted velocity dispersion is large, compatible with the line widths
observed in NH 3 inversion transition lines.
What is the fate of these objects undergoing this intense accretion phase?
In a short timescale,
t <
 0:5 t age , the increasing stellar luminosity will stop the
accretion of matter onto the central star. As the accretion is shut o, so is the
accretion luminosity, an important contributor to the total luminosity. Thus,

accretion will start again, and this cycle will repeat until a stellar wind can also
turn on, which can help to clear out the accreting ow. An HII region can then
be produced, although in this scenario, its evolution would be governed by the
reversal of the accretion ow, a problem we are currently studying.
The total number of this type of collapsing HMCs is unknown, so we are
unable to determine if the statistics of these type of cores are consistent with the
birth rate of massive stars. All of the luminous hot cores without radio emission
reported so far have been found in the neighbourhood of HII regions, and is not
clear how much of their heating is due to an internal or an external source of
energy.
Another mechanism that has been recently proposed to form massive stars is
the coalescence of already existing stars of lower mass in dense clusters. Since the
cross section for these type of encounters is too small for naked stars, Stahler et
al. (2000) suggested that the coagulation occurs when the low-mass stars are still
surrounded by dense molecular cores which increases the eective cross sections.
It is unclear if in such a collision the core will be disrupted, or will merge, or
if the central low mass protostars will be ejected from the cores. Another issue
is how do these objects lose energy and angular momentum to nally coalesce.
Models of this complex hydrodynamical problem have yet to be constructed.
Conclusions
The observational evidence gathered during the last decade suggests that the
formation of stellar clusters and OB associations becomes favored in molecular
cores with large densities and masses. The new observations show the presence
of circumstellar disks around young massive stars, collimated bipolar ows, both
in molecular gas and ionized gas, and hot and dense molecular structures un-
dergoing high mass accretion rates. This evidence suggests that the formation
of massive stars has similar characteristics than that of low-mass stars.
References
Churchwell, E. 1999, The Origin of Stars and Planetary Systems, eds. Charles
J. Lada & Nikolaos D. Kylas, Kluwer Academic Publishers, 515
Garay, G. & Lizano, S. 1999, PASP, 111, 1049
Kaufman, M.J., Hollenbach, D.J., & Tielens, A.G.G. 1998, ApJ, 497, 276
Kurtz, S., Cesaroni, R., Churchwell, E., Hofner, P., & Walmsley, M. 2000, Proto-
stars & Planets IV, ed. V. Mannings, A.P. Boss, & S.S. Russell (Tucson:
Univ. Arizona), 229
Osorio, M., Lizano, S., & D'Alessio, P. 1999, ApJ, 525, 808
Stahler, S.W., Palla, F., & Ho, P.T.P., 2000, Protostars and Planets IV, ed. V.
Mannings, A.P. Boss, & S.S. Russell (Tucson: Univ. Arizona), 327
Walmsley, C.M. 1995, RMA&A Conf. Ser., 1, 137
Wood, D.O.S., & Churchwell, E. 1989, ApJ, 340, 265

Massive Star Birth - JD3 IAU GA
ASP Conference Series, Vol. 000, 2001
P. S. Conti & E. B. Churchwell, eds.
Studies of compact H II regions with ISO
Pierre Cox
Institut d'Astrophysique Spatiale, F-91405 Orsay, France
Abstract. We summarize results of a study of combined ISO SWS-LWS
grating spectra of compact H II regions in our Galaxy.
Compact H II regions, which are the result of the ionizing radiation of newly
formed massive stars, are either completely embedded in their natal molecular
clouds or in the process of breaking out. These regions consist of ionized material
from the local interstellar medium (ISM) and therefore, are excellent probes of
the ISM and can be used to trace the spatial distribution of chemical elements
in galaxies. Abundance determination methods based on infrared observations
present clear advantages over optical studies as the infrared ne-structure lines
are not very sensitive to large changes in the electron temperature and do not
suer from high extinction.
Using the Infrared Space Observatory (ISO), we observed with the SWS
and LWS in the grating mode 45 compact H II regions sampling the galactic
plane from the center to the outer regions. The spectra cover a range from 2.5
to 196 m with resolving powers =
 between 150 and 1500. The complete
spectral coverage gives access to nearly all the atomic ne-structure lines in the
infrared range. For some elements, dierent ionization stages are measured,
which alleviates the problem of applying ionization correction factors for unseen
ions. In addition to the atomic emission lines, the spectra are dominated by
strong dust continua and a series of dust bands, as well as absorption bands
from molecular ice species (see Fig. 1).
The data have been reduced using up-to-date SWS and LWS reduction
software. A description of the reduction techniques as well as a catalogue of the
line intensities will be given in Peeters et al. (in preparation) and the rst results
of this study will be presented in Martn-Hernandez et al. (in preparation) - see
also a preliminary report in Martn-Hernandez et al. (2000). This report will
summarize the problems related to the determination of the N/O abundance
ratio and give some results obtained on the dust properties.
The abundance ratio N/O has a special signicance for galactic nucleosyn-
thesis models because the main stellar source of N is still unclear. Simple
chemical evolution models predicts that its dominant isotope ( 14 N) originates
principally from primary nucleosynthesis (through H-burning of carbon newly-
generated within the star), but, how much 14 N is produced from secondary
nucleosyntheis (i.e. through the CNO cycle where the carbon is already present
in the star) is still an open question.
Fig. 2a shows the ionic abundance ratio N ++ /O ++ versus the galactocentric
distance D gal (the ratio is derived from the measured strengths of the nearby
[OIII] 51.8 m and [NIII] 57.3 m lines, adopting the electron density from the
28

Figure 4. Combined SWS and LWS spectrum of K3-50A. The posi-
tions of the atomic ne-structure lines and H I recombination lines are
labeled, as well as the main dust bands.
Figure 5. (a) Plot of the ionic abundance ratio N ++ =O ++ versus
D gal . The arrows indicate upper limits. The solar abundance ratio
N/O is indicated by . (b) Plot of the ionic abundance N ++ =O ++
versus the intensity ratio between lines of dierent ionization stages of
neon and argon.
O ++ region). A clear trend is observed between 3 and 12 kpc, in agreement with
previous results based on Kuiper Airborne Observatory (KAO) observations of
galactic H II regions (e.g., Aerbach et al. 1997). Beyond 14 kpc, the scatter
of the data is too large to make yet a rm statement. In order to derive the
abundance ratio N/O, ionization correction factors must be applied. Assuming
that most of the nitrogen and oxygen is found in the form of N ++ and O ++ ,
this ionic abundance ratio can be interpreted as a measure of the N/O in the
nebulae. However, this assumption depends on the degree of ionization of the
central star(s). Fig 2b shows N ++ /O ++ against the intensity ratio between
lines of dierent ionization stages of Ne and Ar, which are very sensitive to the
stellar eective temperature. The clear trends which are observed indicate that
the dependence of N ++ /O ++ on galactocentric distance also includes an ioniza-

tion dependence. As mentioned by Stasinska & Schaerer (1997), who presented
photoionization models combined with CoStar stellar models (Schaerer 1997),
the determination of N/O using N ++ /O ++ gives rise to important biases. At
high stellar eective temperatures, T eff  4:5  10 4 K), N ++ partly transforms
into N 3+ , and at low eective temperatures N + and O + are the dominant ionic
species. First estimates using these models suggest that we are in the latter
case, i.e. N/O  N ++ /O ++ - see Martn-Hernandez et al. (in preparation) and
Morisset et al. (in preparation) for detailed discussions.
In addition to the study of the ionized gas and the derivation of the abun-
dance gradient, the ISO data also provide a wealth of information on the dust
around massive young stars. There are indications that the dust properties
change in the surroundings of compact H II regions. For instance, the large
number of H I recombination lines in K3-50A has been used to derive the extinc-
tion towards this object which has been found to deviate from the `standard'
extinction curve (Martn-Hernandez et al. 2000). The attenuation is higher
than expected and clearly lacks the deep minimum near 7 m, indicating that,
either the properties of the dust surrounding K3-50A are very dierent from
those of the typical interstellar dust, or that there are geometrical eects. All
the spectra of the H II regions show strong emission bands and dust continuum.
The spectral appearance of the dust spectrum varies from source to source and
are in some cases dierent from the spectra observed in less excited regions of
the ISM (Roelfsema et al. 1996). The changes are thought to be the result of
variations in the physical conditions of the emitting region such as the radiation
eld or the density (Cox & Roelfsema 1999; Peeters et al. in preparation). A
study of these spectra and of the properties of the sources, together with de-
tailed laboratory measurements, will provide constraints on the models of the
dust (PAHs) which emit this family of emission bands.
This work is presented on behalf of the ISO H II consortium whose members are:
L. Martn-Hernandez, E. Peeters, P. Roelfsema, A.G.G.M. Tielens, and R. Vermeij
(SRON and Kapteijn Laboratory, Groningen, The Netherlands); C. Morisset, J.-P. Ba-
luteau, and F. Damour (Laboratoire d'Astronomie Spatiale), J. Mathis and E. Church-
well (Madison, Wisconsin, USA); M.F. Kessler (Vilspa, Spain) and A. Jones (IAS, Orsay,
France)
References
Aerbach, A., Churchwell, E. & Werner, M.W., 1997, ApJ, 478, 190
Cox, P. & Roelfsema, P.R., 1999, in proc. \Solid Interstellar Matter: The ISO Rev-
olution". Eds.: L. d'Hendecourt, C. Joblin, and A. Jones, EDP Sciences and
Springer-Verlag, p. 151
Martn-Hernandez, L., Peeters, E., Damour, F. et al. 2000, in ISO beyond the Peaks,
ESA-SP 456, in press
Roelfsema, P.R., Cox, P., Tielens, A.G.G.M. et al., 1996, A&A, 315, L289
Schaerer, D. & de Koter,A., 1997, A&A, 322, 598
Stasinska, G. & Schaerer, D., 1997, A&A, 322, 615

Massive Star Birth - JD3 IAU GA
ASP Conference Series, Vol. 000, 2001
P. S. Conti & E. B. Churchwell, eds.
Spectroscopy of the ionizing sources of UC Hii regions
Margaret M. Hanson 1
Dept. of Physics, Univ. of Cincinnati, Cincinnati, OH 45221 USA
Abstract. The utility of near-infrared, spectroscopic studies of central
ionizing sources of UC Hii regions is presented, in conjunction with a
recently available, sophisticated atmospheric code, to constrain the phys-
ical conditions and environment of very massive stars at extremely early
stages of evolution.
Introduction
The greatest stumbling block for developing a theory for the formation of massive
stars is the lack of direct observations of massive stars at very early evolutionary
stages. This is due to the very fast contraction time for massive stellar cores.
While radio and millimeter observations can directly observe conditions of the
collapsing material, only shorter wavelength observations, which become possible
later in the formation of massive stars, are capable of directly detecting the star.
Unfortunately, by the time a massive star can be observed in the optical,
critical signatures present in the spectrum and environment of the star, which
yield clues to the formation process, have disappeared. We must observe mas-
sive stars before their strong ionizing winds dissipate their accretion disks and
out ows, before their stellar atmospheres erase chemical and physical evidence
tracing their formation mechanism (due to mixing and spin-down from mass-
losing winds), and before the central cluster disperses from destruction of the
local molecular cloud. To investigate massive star formation, young massive star
systems must be studied while they are still deeply embedded in an environment
of gas and dust. The most important observations for this purpose are taken
at near-infrared wavelengths because OB stars show few photospheric features
at wavelengths greater than 2.2 m. The H- and K-band spectral regions, from
1.5 2.2 m, are thus the most eective wavelengths which can reasonably
penetrate the shrouded environment of the ultra-compact Hii (UC Hii) region
and allow us to derive the characteristics of massive stars at their earliest evo-
lutionary phase.
The strategy of our work is to obtain high-quality near-infrared spectra of
extremely young and heavily shrouded OB-stars, and to explore quantitative
analyses capable of deriving the physical conditions of these very young massive
stars while they are still heavily buried in their birth clouds.
1 Visiting Astronomer, Very Large Telescope, European Southern Observatory.
31

Figure 6. The vertical iso-contours of equivalent width for high-
gravity (dwarf) stars indicate that the Hei 2.112 m line is an excellent
temperature indicator (from Puls 2000, in prep). For supergiants with
extreme winds (log q represents a parameterization related to mass loss,
see Puls et al. 1997), we will need to rst constrain the wind strength
in conjunction with measurements of Br .
Near-Infrared Quantitative Spectroscopic Analyses
A critical component of our program is the concurrent development of sophis-
ticated model atmospheres which can be used with our near-infrared spectra.
Puls (2000, in prep) has recently extended the non-LTE, Unied Model Atmo-
sphere code of Santolaya-Rey, Puls & Herrero (1997) to include near-infrared
line formation and the in uence of metallic background continua. These models
are spherically extended, and include the important eects of winds. They yield
the entire sub- and supersonic atmospheric structure. Presently this code is
among the only one in existence which is able to reproduce exactly hydrostatic,
non-LTE atmospheres in the limit of very small mass-loss rates, and able to
model atmospheres with intermediate to very strong winds. The rst step in
the analysis is to constrain the stellar and wind conditions by measuring the
equivalent widths (E.W.) of several strategic lines, and comparing them with
model grids of E.W. One such grid is shown in Figure 1. There are a total of
seven Helium and Hydrogen lines in the H- and K-band spectral regions. Our
initial results suggest that these lines are enough for us to consistently constrain
the stellar and wind parameters for O and early-B stars, T e , v sin i, L, log g,
and mass, and to search for binaries and possible disk or in-fall signatures in
very young, massive stars, relying on near-infrared spectroscopy alone.

Observations
To locate the rare and very young massive stars needed for our study, Lex
Kaper, Fernando Comeron and myself have undertaken a near-infrared imaging
survey using the NTT/SOFIA. Our survey comes from a selection of radio and
mid-infrared (IRAS) identied deeply embedded UC Hii regions in the south-
ern hemisphere. Nearly every eld we observed showed a star forming region of
interest. These images serve as guides for locating the central ionizing sources,
which we believe will be among the youngest of the most massive stars accessible
to near-infrared study and analyses. Furthermore, our images give us a glimpse
at the rarely measured, yet highly important, stellar density of very young clus-
ters. Based on the success of that imaging survey, our group was granted time
on the VLT using the near-infrared spectrometer, ISAAC, to obtain high reso-
lution, high signal-to-noise spectra of the central ionizing stars. These stars are
among the youngest, massive stars ever to be directly studied.
Final Remarks
The very fast contraction time of massive stars sets up an immense radiation
eld early in the stars formation, and may reverse the in-fall of additional mass
once the star reaches 10 20 M . This \radiation pressure problem" is so severe,
it has led some theorists to suggest that stars more massive than this cannot
form from accretion alone, but instead form from collisions of intermediate-
mass stars, which rst formed through accretion (Bonnell, Bate & Zinnecker
1998). Our goal is to make observations capable of dierentiating between these
two current theories for massive star formation: accretion and coalescence. The
earlier the phase of evolution we can directly observe, the better hope we have of
detecting the short-lived characteristics unique to each star formation scenario.
Near-infrared imaging studies have identied numerous potential young OB
stars, but only a few stars have been observed spectroscopically, and these stud-
ies have relied on a comparative analysis (Hanson et al. 1997, Watson & Han-
son 1997). However, it should soon be possible to derive more accurate stellar
characteristics as sophisticated atmospheric models extend their analysis to the
near-infrared regime. These characteristics make up a vital boundary condition
constraining theories on massive star formation.
Acknowledgments. I am grateful to Joachim Puls, Lex Kaper and Fer-
nando Cameron, without whom this work would not have been possible.
References
Bonnell, I. A., Bate, M.R., & Zinnecker, H., 1998, MNRAS, 298, 93
Hanson, M.M., Howarth, I.D., Conti, P.S., 1997, ApJ, 489, 698
Sontolaya-Ray, A.E., Puls, J., Herrero, A., 1997, A&A, 323, 488
Puls, J., Kudritzki, R.P, Herrero, A., et al., 1996, A&A 305, 171
Watson, A.M., & Hanson, M.M., 1997, ApJ, 490, L165

Massive Star Birth - JD3 IAU GA
ASP Conference Series, Vol. 000, 2001
P. S. Conti & E. B. Churchwell, eds.
Formation of massive stars by growing accretion
Andre Maeder
Geneva Observatory, CH{1290 Sauverny, Switzerland
Abstract. We calculate pre{Main Sequence evolutionary tracks with
accretion rates growing with the actual stellar masses. We show that ac-
cretion rates growing at least as M 1:5 are necessary to t the constraints
on the lifetimes and HR diagram. Most interestingly, such accretion rates
growing with the stellar mass well correspond to those derived from obser-
vations of mass out ows (Churchwell 2000; Henning et al. 2000). These
rates also lie in the permitted region of the dynamical models.
The formation of massive stars is one of the last frontiers in stellar evo-
lution. There are three scenarios in literature. 1). The classical scenario of
pre-MS evolution at constant mass. 2). The coalescence scenario (cf. Bonnel et
al. 1998). It was proposed due to the alleged diôculty of forming massive stars
by accretion, because of radiation pressure eects. However, as discussed below,
there is a domain of dynamically possible accretion rates. Recently, Elmegreen
(2000) has shown that the star density in young clusters is probably not high
enough for the collision scenario to operate. He emphasizes that there is not
enough time for a protostar to move around in a young cluster and to coalesce
with other stars. 3). The scenario of massive star formation by accretion (Beech
and Mitalas 1994; Bernasconi and Maeder 1996). We show here that the accre-
tion scenario is possible, if the accretion rates _
M accr grow fast enough with the
already accreted stellar mass.
Firstly, we recall that there is a basic dierence between low and high mass
stars. For high mass stars (mass  8 M ), the accretion times t accr are longer
than the Kelvin-Helmholtz time t KH , thus the stars ignites their nuclear reactions
and set on the ZAMS, still accreting and hidden in their molecular clouds. For
low mass stars, t accr  t KH is the rule.
Models with constant accretion rates or slowly growing accretion rates lead
to much too long pre{MS lifetimes. Also the birth lines, which is the path in the
HR diagram followed by accreting stars, do not provide a good upper envelope
of the observations. Norberg and Maeder (2000) have calculated several grids
of pre{MS tracks with accretion rates growing like _
M accr = _
M ref

M
M
 '
, with
values of ' equal to 0.5, 1.0 and 1.5 and also for dierent values of _
M ref . The
best t to the short lifetimes required (t  10 6 yr, cf. Elmegreen, 2000) and to
the observations of PMS stars in the HR diagram is achieved for ' ' 1.5 and
for _
M ref of the order of 10 5 M yr 1 , (cf. Fig. 1).
Temperatures, luminosities and out ow rates of UC HII regions have been
studied by radio and IR observations by Churchwell (2000) and Henning et al.
(2000). Huge bipolar out ows have been observed. Remarkably, the out ow
34

4.5 4
2
4
6
Figure 7. The continuous lines represent birthlines obtained with '
= 1.5 and _
M ref being equal to 10 6 , 5
10 6 and 10 5 M yr 1 from
bottom to top respectively. The points represent observations from
various sources collected by Norberg and Maeder (2000). The broken
lines are post-MS tracks, and the dot{broken lines pre{MS tracks with
constant mass.

rates _
M out behave continuously like L 0:7
bol over 6 decades of luminosity. At solar
luminosity, the values of _
M out are about 10 5 M yr 1 , i.e. of the same order
as the currently estimated accretion rates. However for massive stars with L
between 10 4 and 10 6 L , the out ow rates _
M out are in the range of 10 3 to
10 2 M yr 1 . From the masses present in the out ows and the luminosity
of the central object, Churchwell estimates that the fraction f of the infalling
material incorporated into the star is about 15 %, while 85% is de ected in the
out ows. With a mass{luminosity relation appropriate for the interval of 2 to
85 M , we get
_
M accr = 1:5 10 5 f
1 f

M
M
 1:54
M yr 1 ; (1)
where f is the accreted fraction of the infalling material. It is remarkable
that the slope and constant factor derived in Fig. 1 are so close to that obtained
from UC HII regions.
Not all values of _
M accr are possible (cf. Wolre and Cassinelli 1987). For
too low _
M accr , the momentum in the radiation is larger than that of the wind, so
accretion is not possible, while for too high _
M accr , the shock luminosity created
by accretion becomes supra{Eddington. The rates given by Equ. 1 are in
the permitted domain, even when rotation and accretion in the midplane are
accounted for (cf. Nakano, 1989).
In the literature, an expression like _
M accr ' v 3
sound
G is often found. According
to it, high accretion rates as given by Equ. 1 for massive stars would imply a
relatively high temperature T of the infalling gas. However, high T are not
observed (Caselli and Myers, 1995), and strong turbulent motions, for which
there are many evidences, may well provide the necessary initial support of the
cloud. Moreover, turbulent motions fed by the stellar radiation may critically
participate to the regulation of the accretion rates of the matter falling from the
cloud onto the disk.
References
Beech, M, Mitalas, R. 1994, ApJS, 95, 517
Bernasconi, P.A., Maeder, A. 1996, A&A, 307, 829
Bonnell, I.A., Bate, M.R., Zinnecker, H. 1998, MNRAS, 298, 93
Churchwell, E. 2000 in Unsolved Problems in Stellar Evolution, Space Telescope.
Sci. Inst. Symp. Ser., vol 12, Ed. M. Livio, Cambridge Univ. Press, 41.
Elmegreen, B. 2000, astro{ph/0005189
Henning, Th., Schreyer, K., Launhardt, R., et al. 2000, A&A, 353, 211
Nakano, T. 1989, ApJ, 345, 464
Norberg, P., Maeder, A. 2000, A&A, 359, 1025
Wolre, M.G., Cassinelli, J.P. 1987, ApJ, 319, 850

Massive Star Birth - JD3 IAU GA
ASP Conference Series, Vol. 000, 2001
P. S. Conti & E. B. Churchwell, eds.
Obscured Galactic Giant HII Regions; Discovery of Young
Clusters
Peter S. Conti
JILA and APS Department, University of Colorado, Boulder CO 80309
Robert D. Blum
Cerro Tololo Interamerican Observatory, La Serena, Chile
Near IR Observations
Giant HII (GHII) regions in our Galaxy are typically initially found by radio ob-
servations of their optically thin free-free continuum emission. Most of them are
partially or totally obscured in the visible by the absorbing eect of intervening
and/or local interstellar dust. We (Blum et al 1999, 2000) have selected a list
of the brightest GHII regions in our Galaxy (from Smith et al 1978) and have
begun a program of JHK imaging and K band spectroscopy to identify and
classify the exciting stars. We have obtained near IR imaging of eight GHII re-
gions (and data is available for four others). All of these, aside from W49, show
the presence of a stellar cluster in the K band at the radio source position. The
K;H K diagrams are used to select the brightest stars. The J K vs: H K
diagrams distinguish those stars along the normal reddening line from those with
K band excesses. The former group ought to have normal OB star spectra; the
latter will have featurless continua in the K band due to emission from localized
warm dust arising in a natal disc (Hanson et al 1997). A disc geometry can also
produce CO emission (or absorption) band features.
Classications of the Brightest Cluster Stars
As a rst step in classication, we have obtained K band spectra of a few of the
brightest stars from the C-M diagram for most of the clusters. In Table 1, we
list the 12 GHII regions with remarks about the spectra of the brightest stars,
from ongoing work with A. Damineli (aside from M17 from Hanson et al 1997,
NGC3603 from Drissen et al 1995, and W51 from Hanson - private communica-
tion). This sample represents about half of the brightest GHII listed by Smith
et al 1978. Curiously, at least ve of the clusters (out of eight with spectra avail-
able) contain one or more stars with evidence of stellar discs from featureless K
band continuua, K band excesses, and/or CO emission lines. While this may
only be small (source) number statistics, it does suggest that either the natal
disc phase for massive stars lasts an appreciable fraction of the OB star lifetime
or radio selected GHII regions are not sampled uniformly over the massive star
main sequence lifetime.
37

Table 4. Bright Galactic GHII With Imaging Data
Catalog Radio Dist. # Lyc Remarks on Individual
Name Source Kpc 10 49 Stellar Spectra
NGC3576 G291.3-0.7 3.6 26 CO emiss + CO abs + no lines
NGC3603 G291.6-0.5 8.2 188 several Of/WN + several early O
- G333.1-0.4 13.4 169 Cluster; No spectra yet
- G333.6-0.2 14.1 1140 2 (?) Clusters; No spectra yet
- G351.6-1.3 4.4 11 LBV + CO abs + no lines
W31 G10.2-0.3 5.1: 30 1 no lines + 4 early O
W33 G12.8-0.2 4.6 5 1 disc (Br emission)
M17 G15.1-0.7 2.3 54 OB stars; some with discs
W42 G25.4-0.2 13.4 82 1 early O + 2 no lines
W43 G30.8-0.0 7.0 107 Of/WN + 2 early O
W49 G43.2-0.0 13.8 172 YSOs ?
W51 G49.5-0.4 6.6 154 1 early O (others S/N problem?)
The Special Case of W49
There are about 30 point-like radio sources visible in the core of W49 at cm
wavelengths (De Pree et al 1997). They interpret these optically thick free-free
sources to be individual ultra compact HII (UCHII) regions (Wood & Churchwell
1989). W49 is thus a massive OB star cluster. In most cases, there is no
stellar image in the K band corresponding to the radio point source position.
This implies that the exciting stars are deeply embedded in their natal clouds,
composed of extensive dust that absorbs in the near IR. This interpretation has
recently been strengthened by the detection of re-radiated dust emission from
many of the radio point sources in the thermal IR by Smith et al (2000). All the
other GHII regions in our Galaxy that we have observed have a stellar cluster
visible in the K band, whereas in W49 the cluster is buried in its birth material.
W49 is the youngest example of a cluster of newly born massive stars in our
Galaxy and intermediate in luminosity between UCHII regions and the buried
SSC in starburst galaxies that are the subject of the next two talks.
References
Blum, R.D., Conti, P.S., & Damineli, A. 1999, AJ, 117, 1392
Blum, R.D., Conti, P.S., & Damineli, A. 2000, AJ 119, 1860
De Pree, C.G., Mehringer, D.M., & Goss, W.M. 1997, ApJ, 482, 307
Drissen, L., Moat, A.F.J., Walborn, N., & Shara, M.M. 1995, AJ, 110, 2235
Hanson, M.M., Conti, P.S., & Rieke, M.J. 1996, ApJS. 107, 281
Hanson, M.M., Howarth, I.D., & Conti, P.S. 1997, ApJ, 489, 698
Smith, L.F., Biermann, P., & Mezger, P.G. 1978, AA, 66, 65
Smith, N., Jackson, J.M., Kraemer, K.E., Deutsch, L.K., Bolatto, A., Hora,
J.L., Fazio, G., Homan, W.F., & Dayal, A. 2000, ApJ, 540, 316

Massive Star Birth - JD3 IAU GA
ASP Conference Series, Vol. 000, 2001
P. S. Conti & E. B. Churchwell, eds.
Signatures of the Youngest Starbursts:
The Discovery of Ultradense H II Regions
Kelsey E. Johnson
JILA, Univ. Colorado, Boulder, CO 80309, USA
Introduction
Globular clusters are ubiquitous in the local universe, and their younger and
bluer siblings, \super star clusters" (SSCs), have been observed in a large num-
ber of starburst galaxies. Recently a handful of ultra-young SSCs have been
found still embedded in their birth material (e.g. Kobulnicky & Johnson 1999
(here after KJ99); Turner et al. 2000; Beck et al. 2000; Tarchi et al. 2000).
Because of their similarity (although on a vastly larger scale) to ultracompact
H II regions in the Galaxy, KJ99 dub these embedded massive star clusters \ul-
tradense H II regions" (UDH IIs). I will review the discovery of UDH IIs, their
modeled properties, and their connection to the more familiar Galactic ultra-
compact H II regions.
The Discovery of \Ultradense H II Regions"
VLA radio continuum images have revealed compact radio sources in the star-
burst galaxy Henize 2-10. While the global radio continuum spectrum is indica-
tive of non-thermal processes, the compact radio sources have positive spectral
indices | suggesting an optically thick thermal bremmsstralung origin. More-
over, these regions are not visible in optical images. After considering a variety
of physical scenarios (such as supernovae or supernovae remnants, see KJ99), the
most likely explanation is that these sources are dense H II regions surrounding
massive star clusters. Therefore, we propose that the radio data preferentially
reveal the youngest, densest, most highly obscured star-forming events.
Modeled Properties
Simple models of H II regions invoking free-free emission and self-absorption
allow us to extract more information about the physical properties of these
UDH IIs. As shown in gure 1, sizes of  a few parsecs and densities of 
5000cm 3 nicely reproduce the shape of the radio spectra. If we assume all
of the 2cm ux is thermal and a standard initial mass function, we can also
estimate N Lyc and the masses of the UDH IIs. The resulting physical properties
are truly remarkable (see KJ99 for more details); the estimated sizes (a few
parsecs), masses (a few 10 5 M ), excitations (N Lyc  10 51 52 erg), and ages
(less than a few 10 5 years) of the newly discovered UDH IIs imply that we may
be witnessing the birth process of the familiar globular clusters.
39

Figure 8. VLA 6 cm and 2 cm data for the ve UDH IIs discovered
by KJ99 in He 2-10. The data are best modeled by sources with radii
between 3 and 8 pc and electron densities of  5000cm 3 .
Connection to Ultracompact H II Regions
There are a number of connections between UDH IIs and the UCH IIs. We
already know that UCH IIs are often found in \complexes" (e.g. W49 or W51),
so clustering in massive star formation is not a revelation. It seems likely that
there is a continuum in the scale of massive star formation regions from single
stars to UDH IIs. Also, based on dynamical arguments as well as number counts,
we've estimated that UDH IIs spend  15% of their lifetimes enshrouded, which
is nearly identical to the result found for UCH IIs (Wood & Churchwell 1989).
Based on these similarities, there is perhaps a parallel scheme for the evolution
of single massive stars and clusters, and we can begin to observe UDH IIs with
the same tools, techniques, and questions that have been applied to UCH IIs.
Hot Cores ! UCH IIs ! OB
(Complexes) Associations
???? ! UDH IIs ! Super Star ! Globular
(not yet observed) Clusters Clusters
References
Beck, S.C., Turner, J.L., & Kovo, O. 2000, AJ, 120, 244
Kobulnicky, H.A. & Johnson, K.E. 1999, ApJ, 527, 154
Tarchi, A., Neininger, N., Greve, A., Klein, U. et al. 2000 A&A, 358, 95
Turner, J.L., Beck, S.C., & Ho, P.T.P 2000, ApJ, 532:L109
Wood, D.O. & Churchwell, E. 1989, ApJS, 69, 831

Massive Star Birth - JD3 IAU GA
ASP Conference Series, Vol. 000, 2001
P. S. Conti & E. B. Churchwell, eds.
The Supernebula and Protoglobular Cluster in NGC 5253
Jean L. Turner
UCLA, Division of Astronomy & Astrophysics, Los Angeles CA 90095
Abstract. A compact and bright radio and infrared source within the
starburst center of the dwarf galaxy NGC 5253 appears to be an extremely
large compact HII region. The high density (n e  few 10 4 cm 3 ) and
size (1-2 pc) of the nebula require the ionization equivalent of at least
4000 O7 stars. We suggest that this optically obscured nebula is a forming
super star cluster, or protoglobular cluster, with an age of a few hundred
thousand years.
The Dwarf Galaxy NGC 5253 and its Luminous Starburst
NGC 5253 is a dwarf I0/S0 galaxy, located near the large spiral galaxy M83, at
a distance of 4 Mpc. It is well known for the anomalous dust lane extending
into the central part of the galaxy (e.g., van den Bergh 1980, Graham 1981).
NGC 5253 is presently undergoing a starburst of infrared luminosity L IR 
2  10 9 L (Beck et al. 1996). The starburst consists of an extended region
of young super star clusters (Gorjian 1996; Calzetti et al. 1997) and bright,
lamentary H emission (Bohuski et al. 1972; Calzetti et al. 1997). Surrounding
the galaxy are numerous intermediate age clusters (Caldwell & Phillips 1989),
indicating that the dwarf galaxy has had a complex history.
In the center of the starburst, there is an optically obscured, radio bright
HII region. This HII region has no obvious optical counterpart. The nebula was
rst discovered in arcsecond resolution VLA images (Beck et al. 1996; Turner et
al. 1998). The radio ux indicated a Lyman continuum rate requiring thousands
of OB stars within the 10 pc beam. Subsequent, higher resolution VLA images
revealed that this source is even more compact: 2 pc  1 pc in extent (Turner
et al. 2000.) The electron density is remarkably high for a nebula of this size:
n e  410 4 cm 3 , a density characteristic of the youngest \compact HII regions"
in our Galaxy, regions that are typically a few hundred thousand years old (Wood
& Churchwell 1989). However, in spite of the abundant circumstantial evidence
in favor of an HII region interpretation, other sources of radio emission for this
unusual source, albeit unlikely, were possible (Turner et al. 1998, 2000).
Using the Keck LWS in imaging mode, we have conrmed that the radio
source is an HII region (Gorjian et al. 2000). Diraction limited imaging at 12
and 18m shows a bright infrared source, with ux densities of 2.1 and 10.4 Jy,
respectively, and size < 0:5 00 . The ratio of mid-IR to radio ux, S IR =S 2cm  100
is typical of those observed in Galactic HII regions. The size of the IR nebula is
consistent with the radio size. The source is clearly not a supernova remnant,
and extremely unlikely to be an AGN. The overwhelming odds are that it is an
41

HII region. It is probably the source of the strong Brackett lines (Kawara et al.
1989) and radio recombination lines (K. Anantharamaiah, private commun.) in
NGC 5253, although the lower resolutions of these data do not allow a direct
spatial comparison with the radio and mid-IR source.
The Supernebula in NGC 5253
This nebula provides a remarkable demonstration of how compact starburst
emission can be. The radio and IR observations require a Lyman continuum
ux of 4  10 52 s 1 , the equivalent of 4000 O7 stars. If the IMF is Salpeter, as
seems to hold for R136 (Massey & Hunter 1998; H. Zinnecker, private commun.),
then there are over a million stars localized to this 1-2 pc region. If the stars
are more distributed, then the excitation requirements on the cluster are even
more extreme. This tiny 1 x 2 pc nebula is responsible for 80% of the total mid-
infrared IRAS ux of NGC 5253, and at least 30% of the total infrared luminosity
of the entire galaxy. Unless there is some unusual source of connement (there
is, surprisingly, no evidence of any molecular gas in the vicinity of the HII
region: Turner, Beck, & Hurt 1997), then the nebula, like Galactic compact
HII regions, is at most a few hundred thousand years old. It is therefore likely
that in this source we are witnessing the birth of a super star cluster, and a
potential protoglobular cluster. Similar nebulae, although not as large as the
one in NGC 5253, have been seen in He 2-10 (Kobulnicky & Johnson 1999) and
NGC 2146 (Tarchi et al. 2000).
References
Beck, S. C., Turner, J. L., Ho, P. T. P., Lacy, J. H., & Kelly, D. 1996, ApJ, 457,
610
Bohuski, T. J., Burbidge, E. M., Burbidge, G. R., & Smith, M. G. 1972, ApJ,
175, 329
Caldwell, N., & Phillips, M. M. 1989, ApJ, 338, 789
Calzetti, D., Meurer, G. R., Bohlin, R. C., Garnett, D. R., Kinney, A. L.,
Leitherer, C., & Storchi-Bergmann, T. 1997, AJ, 114, 1834
Gorjian, V. 1996, AJ, 112, 1886
Gorjian, V., Turner, J. L., & Beck, S. C. 2000, submitted
Graham, J. A. 1981, PASP, 93, 552
Kawara, K., Nishida, M., & Phillips, M. M. 1989, ApJ, 337, 230
Kobulnicky, H. A., & Johnson, K. E. 1999, AJ, 527, 154
Massey, P. & Hunter, D. A. 1998, ApJ, 493, 180
Tarchi, A. et al. 2000, A&A, 358, 95
Turner, J. L., Beck, S. C., & Ho, P. T. P. 2000, ApJ, 532, L109
Turner, J. L., Beck, S. C., & Hurt, R. L., ApJ, 474, L11
Turner, J. L., Ho, P. T. P., & Beck, S. C. 1998, AJ, 116, 1212
van den Bergh, S. 1980, PASP, 92, 122
Wood, D. O. S., & Churchwell, E. 1989, ApJS, 69, 831

Massive Star Birth - JD3 IAU GA
ASP Conference Series, Vol. 000, 2001
P. S. Conti & E. B. Churchwell, eds.
Denitions, Summary, and Some Issues
Peter S. Conti
JILA and APS Department, University of Colorado, Boulder CO 80309
Denitions
This Joint Discussion has been titled Massive Star Birth. Perhaps it is appro-
priate here to dene what we mean by a massive star. The very word massive
suggests we consider a minimum mass M below which one would speak of low
(or intermediate) mass evolution, and above which is the realm of massive stars.
It is natural to take this mass limit as that in which a (single) star will end its
life as a supernova: 8M . This corresponds to a (minimum) luminosity L of
a few x 10 3 L , a (minimum) T eff of 20000K, and a ZAMS spectral type of
about B1.5V. Note that this mass division refers to the nal evolution of a star,
and might well have nothing to do with dierence in physical processes between
massive and low mass star birth. For example, the minimum T eff for a star to
produce an UCHII region, a readily observable quantity, corresponds to a T eff
closer to 30000K and a mass of 15M . On the other hand, the UCHII phase
is time dependent during the stellar birth processes and its absence would not
necessarily be a good minimum mass indicator. Maeder has noted that an im-
portant distinction between high and low mass star formation is whether or not
the accretion time scale is larger or smaller than the \Kelvin-Helmholz" time,
respectively. He suggests that this point is near 8M . Finally, in a whimsical
spirit I suggest that massive stars throughout their birth processes be referred
to as Massive YSOs, or MYSOs, to be pronounced as \my sos".
Summary
Massive star birth, unlike nearly all other elds of stellar astrophysics, is nearly
entirely hidden from view in the visible but amenable to observation at IR,
mm, and cm wavelengths. This JD was roughly organized so that the earliest
phases of massive star birth were discussed rst, followed by later stages, end-
ing with the emergence of a hot star from its dust cocoon and optical visibility.
Williams stated that massive stars form in clusters within self-gravitating molec-
ular clouds. He considered the conditions needed for the formation of massive
stars and concluded that turbulence probably plays a leading role. As this quan-
tity dissipates the cloud(s) can begin to collapse. Mundy's talk was concerned
with the environment of embedded massive stars. This material radiates strongly
in mm and sum-mm regions. He pointed out that the initial masses involved
in massive star formation are much larger than what eventually ends up in the
star. This very substantial non-accreted material is somehow ejected in bipolar
out ows, eventually forming low mass stars which are typically associated with
massive stars in a cluster. Van Dishoeck & van der Tak discussed spectroscopy
43

of the complex molecules that are found during initial formation stages. There
are a wide range of physical conditions throughout the pre-stellar envelope and
various molecules sample the dierent spatial regions, ranging from hot cores to
regions containing evaporated ices. Dyson et al. summarize the hydrodynamics
of the ionization and recombination fronts. The issue of the role of magnetic
elds was again raised, but answers concerning its signicance are not yet in
hand. Observations of the jets and out ows are detailed by Menten. Unlike
low mass stars, this ejected material does not seem to be highly collimated in
jets, but rather the out ows have a relatively wide opening angle. The evidence
for the presence of discs in massive stars is considered and shown to be likely
by Cesaroni. Here the tracers involve high velocity resolution studies of various
molecular species, looking for Keplerian motion.
Churchwell pointed out the potential presence of precursor UCHII candi-
dates from sub mm observations of luminous sources. He also suggested that
hard X-ray point sources in star forming regions might be another indication of
the initial formation stages. Lizano & Garay gave a summary of their exquisite
and detailed published review of massive star formation. Cox summarized ISO
spectroscopic studies of the abundances that may be determined from the com-
pact HII regions surrounding massive stars. Hanson gave a progress report on
the theory which will be used to connect the observed properties of near IR
spectral lines with stellar parameters. This can be used for stars which have
emerged from their dust cocoons. More detailed observations are to be expected
soon as new telescopes and better detectors come on line. Maeder showed that
massive stars can grow by accretion of dust and gas in the time scales available if
that rate goes as a power law, say M 1:5 or so. He points out that this exponent
has been observed in the bipolar out ow mass loss surrounding massive stars.
Under the assumption that the accretion and out ow rates are correlated, this
means that massive stars can grow by accretion within the current framework.
Conti & Blum called attention to the presence of a (radio detected) cluster
of O star UCHII objects in W49, a GHII region in our Galaxy. These objects
are not visible in the K band, whereas in other GHII regions they have stud-
ied an O star cluster is clearly seen at this wavelength. In W49, each cluster
O star is still contained within its birth cocoon and thus in the earliest birth
stages. W49 is a less luminous, but nearby, example of buried super star clusters
discovered by Johnson and by Turner. These two authors described their inde-
pendent and innovative discoveries of luminous thermally excited radio regions
in starburst galaxies. These objects emit optically thick free-free radio emission
but are mostly invisible at optical and near IR wavelengths. These ultra-dense
HII regions, or buried super stars clusters, should have very strong dust emis-
sion signatures in the mm and sub-mm regions. These would be the youngest
phases of massive cluster formation, with properties similar to but scaled up
from individually excited UCHII.
Some Issues
A number of interesting questions occurred to me while sitting in on the JD con-
cerned with massive star birth. Given that I am assigned to write this summary,
I will take the opportunity to put them down (in no particular order).

1) What happens to the mass that is not accreted by the nal stellar object?
As we have seen, this is 90% or more of that involved in the intial infall processes
for a massive star. Exactly how does it turn into lower mass stars?
2) Are low mass stars already present as individual entities in the earliest
stages of formation of massive stars? Or do they only form from the ejecta of
massive objects?
3) Some O stars (and W-R types) seem to be found in our Galaxy, and
elsewhere, in isolation from any other stars, or clusters, and in between spiral
arms. Do these have a dierent formation origin? For example, not in GMCs?
4) How does one account for the formation of close binaries in the current
(or any) framework? Or, for that matter, wide binaries?
5) What are the time scales for dust and ice evaporation by the massive
hot star as compared to the dynamical times of the ionization front expansion
and the wind dissipation of the environment? All these processes are at work in
massive star formation.
It is nice to end the Proceedings of JD3 with notice of a brand new eld of
study, namely the advent of highly luminous counterparts to the UCHII regions,
the buried super star clusters. It appears that there is a luminosity sequence
in going from (more or less) individually excited UCHII regions, through W49
with some few dozen hot stars, to the buried super star clusters in starbrust
galaxies with hundreds to thousands of O type objects. We may be beginning
to be able to study the earliest stages of cluster formation. This work, just in its
infancy, is going to give an important connection to IR luminous galaxies and
to cosmology, with respect to the formation of the rst generations of stars.
Finally, it is clear from the enthusiasm expressed at this meeting that a
larger conference is called for. Ed and I have agreed to organize a Symposium
on the topic of this JD which we will propose to the IAU to be held in 2003 or
2004. In the meantime, we will also host a Workshop (Hot Stars III) in Boulder
in summer of 2001 concerning massive star birth from the prospective of stellar
astrophysics.
References
Names where included refer to papers presented at this Joint Discussion