Transcript Disk

Disks, toroids and the formation
of massive stars
Riccardo Cesaroni
O-B star  >103 LO  >8 MO  high-mass
1)
2)
3)
4)
5)
Observations: where do (massive) stars form?
Theory: how do (massive) stars form?
Search for disks in high-mass (proto)stars
Results: disks in B stars, toroids in O stars
Implications: different formation scenarios for B
and O stars?
High-mass star forming regions:
Observational problems
 IMF  high-mass stars are rare
 large distance: >400 pc, typically a few kpc
 formation in clusters  confusion
 rapid evolution: tacc=20 MO /10-3MOyr-1=2 104yr
 parental environment profoundly altered
• Advantage:
 very luminous (cont. & line) and rich (molecules)!
Where do massive stars form?
Clump
UC HII
HMC
Core
Clump
UC HII
HMC
Clump
HMC
High-mass star
forming region
0.5 pc
Clumps and hot molecular cores
D (pc)
Clump
HMC
1
0.1
Rclump = 10 RHMC
Mclump= 10 MHMC
nclump = 0.01 nHMC
M (MO) nH2 (cm-3) T (K)
1000
100
105
107
30
100
• nR-p with p=1.5-2.5  no break at HMC
 unstable density profile
• Mclump > Mvirial  clumps unstable
• Vclump = VHMC  HMCs at rest wrt clumps
• TR-q with q=0.4-0.5  clumps centrally heated
 Clumps might be collapsing
 HMCs are density peaks in clumps
 HMCs are T peaks “enlightened’’ by embedded stars
HMC
Clump
nH2  R-2.6
Fontani et al. (2002)
• nR-p with p=1.5-2.5  no break at HMC
 unstable density profile
• Mclump > Mvirial  clumps unstable
• Vclump = VHMC  HMCs at rest wrt clumps
• TR-q with q=0.4-0.5  clumps centrally heated
 Clumps might be collapsing
 HMCs are density peaks in clumps
 HMCs are T peaks “enlightened’’ by embedded stars
Fontani et al. (2002)
sample of 12 Clumps
• nH2R-p with p=1.5-2.5  no break at HMC
 unstable density profile
• Mclump > Mvirial  clumps unstable
• Vclump = VHMC  HMCs at rest wrt clumps
• TR-q with q=0.4-0.5  clumps centrally heated
 Clumps might be collapsing
 HMCs are density peaks in clumps
 HMCs are T peaks “enlightened’’ by embedded stars
 HMCs are deeply related to clumps
• nH2R-p with p=1.5-2.5  no break at HMC
 unstable density profile
• Mclump > Mvirial  clumps unstable
• Vclump = VHMC  HMCs at rest wrt clumps
• TR-q with q=0.4-0.5  clumps centrally heated
 Clumps might be collapsing
 HMCs are density peaks in clumps
 HMCs are T peaks “enlightened’’ by embedded stars
 HMCs are deeply related to clumps
How do massive stars form?
Low-mass vs High-mass
Shu et al. (1987): star formation from
inside-out collapse onto protostar
Two relevant timescales:
accretion  tacc = M*/(dM/dt)
contraction  tKH = GM*/R*L*
• Lowmass (< 8 MO): tacc < tKH  “birthline’’
• Highmass (> 8 MO): tacc > tKH  accretion on
ZAMS
Palla & Stahler (1990)
tKH=tacc
dM/dt=10-5 MO/yr
Sun
PROBLEM:
High-mass stars “switch on” still accreting 
 radiation pressure stops accretion (Kahn 1976)
 stars > 8 MO cannot form!?
SOLUTIONS
Yorke (2003): Kdust< Kcrit  M*/L*
1) “Increase’’ M*: large accretion rates
2) “Reduce’’ L*: non-spherical accretion
3) Reduce Kdust: large grains (coalescence of lower
mass stars)
Possible models
1) Large accretion rates: competitive accretion
(Bonnell et al. 2004); turbulent cores (McKee
& Tan 2002)
2) Non-spherical accretion: disk+outflow focus
ram pressure and dilute radiation pressure
(Yorke & Sonnhalter 2002; Krumholz et al.
2003)
3) Coalescence: many low-mass stars merge into
one massive star (Bonnell et al. 2004)
Disk + outflow may be the solution (Yorke &
Sonnhalter, Kruhmolz et al.):
Outflow  channels stellar photons 
 lowers radiation pressure
Disk  focuses accretion 
 boosts ram pressure
 Detection of accretion disks would support
O-B star formation by accretion, otherwise
other mechanisms are needed
The search for disks
in high-mass YSOs
• Disks seem natural outcome of star formation:
collapse + angular momentum conservation 
 flattening + rotation speed up  disk
• Disks are likely associated with outflows:
outflow detection rate = 40-90% in massive YSOs
(luminous IRAS sources, UC HIIs, H2O masers,…)
(Osterloh et al., Beuther et al., Zhang et al., …)
 disks should be widespread!
BUT…
Where and what to search for…?
Where to search for?
disk?
0.5 pc
What to search for?
Theorist’s definition:
Disk = long-lived, flat, rotating structure in
centrifugal equilibrium
Observer’s definition:
Disk = elongated structure with velocity gradient
perpendicular to outflow axis
outflow
core
disk
outflow
TRACER
PROs
CONTRAs
Maser lines
High angular &
spectral resolution
Unclear geometry &
kinematics
Continuum
Sensitivity (and
resolution)
No velocity info
Confusion with freefree and/or envelope
Limited angular
resolution and
sensitivity (but see
SMA and ALMA)
Thermal lines Kinematics and
geometry of outflow
and disk
Results of disk search
Two types of objects found:
Toroids
Disks
• M > 100 MO
• R ~ 10000 AU
• L > 105 LO (O stars)
• (dM/dt)star > 10-3 MO/yr
• trot ~ 105 yr
• tacc ~ M/(dM/dt)star ~ 104 yr
 tacc << trot
 non-equilibrium, circumcluster structures
• M < 10 MO
• R ~ 1000 AU
• L ~ 104 LO (B stars)
• (dM/dt)star ~ 10-4 MO/yr
• trot ~ 104 yr
• tacc ~ M/(dM/dt)star ~ 105 yr
 tacc >> trot
 equilibrium, circumstellar
structures
Examples of rotating toroids:
Furuya et al. (2002)
Beltran et al. (2004)
Beltran et al. (2005)
Furuya et al. (2002)
Beltran et al. (2004)
Beltran et al. (2005)
Furuya et al. (2002)
Beltran et al. (2004)
Beltran et al. (2005)
Furuya et al. (2002)
Beltran et al. (2004)
Beltran et al. (2005)
Furuya et al. (2002)
Beltran et al. (2004,2005)
Mdyn= 19 MO
Moscadelli et al. (2007)
Mdyn= 55 MO
First result:
• velocity gradient perpendicular to bipolar
outflow  rotating toroid
• conservation of angular momentum from 2”
(15000 AU) to 0.5” (4000 AU)  possible
formation of circumstellar disk?
absorption
HC HII
hypercompact HII + dust
O9.5 (20 MO) + 130 MO
Beltran et al.
(2006, Nature)
Second result:
• Red-shifted absorption in molecular line
towards HII region  infall towards star
 accretion onto star?
Hypercompact HII region
Moscadelli et al. (2007)
Beltran et al. (2007)
7mm free-free & H2O masers
500 AU
Hypercompact HII region
Moscadelli et al. (2007)
Beltran et al. (2007)
7mm free-free & H2O masers
30 km/s
Third result:
• H2O masers along HII region border have
proper motions away from star  expansion of
shell HII region with
tHII = 500 AU/50 km/s = 50 yr !!!
note that this is distance independent
 hyperyoung HII region
Final scenario:
• G24 A1 is a massive toroid, rotating about a
bipolar outflow and infalling towards an O star
with very young expanding HII region
 a 20 MO star has been formed through
accretion (now finished…?)
Example of rotating disk:
IRAS 20126+4104
Cesaroni et al.
Hofner et al.
Keplerian
Moscadelli etrotation:
al.
M*=7 MO
Moscadelli et al. (2005)
Results of disk search
Two types of objects found:
Toroids
Disks
• M > 100 MO
• R ~ 10000 AU
• L > 105 LO (O stars)
• (dM/dt)star > 10-3 MO/yr
• trot ~ 105 yr
• tacc ~ M/(dM/dt)star ~ 104 yr
 tacc << trot
 non-equilibrium, circumcluster structures
• M < 10 MO
• R ~ 1000 AU
• L ~ 104 LO (B stars)
• (dM/dt)star ~ 10-4 MO/yr
• trot ~ 104 yr
• tacc ~ M/(dM/dt)star ~ 105 yr
 tacc >> trot
 equilibrium, circumstellar
structures
Are there disks in O stars?
• In Lstar ~ 104 LO (B stars) true disks found
• In Lstar > 105 LO (O stars) no true disk (only
toroids) found - but distance is large (few kpc)
• Orion I (450 pc) does have disk, but luminosity
is unclear (< 105 LO???)
 Difficult to detect disks in O (proto)stars.
Why?
Observational bias or physical explanation?
Observational bias?
Assumptions:
HPBW = Rdisk/4
FWHMline = Vrot(Rdisk)
Mdisk  Mstar
same <Ncol> in all disks
TB > 20 K
obs. freq. = 230 GHz
5 hours ON-source
spec. res. = 0.2 km/s
S/N = 20
Assumptions:
HPBW = Rdisk/4
FWHMline = Vrot(Rdisk)
Mdisk  Mstar
same <Ncol> in all disks
TB > 20 K
obs. freq. = 230 GHz
5 hours ON-source
spec. res. = 0.2 km/s
S/N = 20
Physical explanation?
• O-star disks “hidden” inside toroids
• O-star disk lifetime too short, i.e. less than
rotation period:
 photo-evaporation by O star (Hollenbach et al.
1994)
 tidal destruction by stellar companions (Hollenbach
et al. 2000)
In both cases we assume Mdisk=Mstar/2 and disk
surface density ~ R-1, i.e. Mdisk  Rdisk:
Cesaroni, Galli, Lodato, Walmsley, Zhang (2007)
photo-evaporation
rotational period
tidal destruction
• Photoionosation: inefficient disk destruction mechanism,
for all spectral types (if Mdisk comparable to Mstar)
• Tidal interaction with the stellar companions: more
effective to destroy outer regions of disks in O stars than
in B-stars
Disks in O (proto)stars might be shorter lived,
and/or more deeply embedded than those
detected in B (proto)stars
Conclusions
• Disks found in B (proto)stars  star formation by
accretion as in low-mass stars
• No disk found yet (only massive, rotating toroids)
in O (proto)stars 
– observational bias (confusion, distance, rarity,…)
– disks hidden inside toroids and/or truncated by tidal
interactions with stellar companions
– disks do not exist; alternative formation scenarios for
O stars needed: coalescence of lower mass stars, competitive
accretion (see Bonnell, Bate et al.)
Furuya et al. (2002)
Beltran et al. (2004)
Beltran et al. (2005)
NIR J+H+K
10 pc
G9.62+0.19
2 pc
G9.62+0.19
350 micron
0.5 pc
Hunter et al. (2000)
3.6cm
Testi et al.
Cesaroni et al.
Furuya et al. (2002)
Beltran et al. (2004)
Beltran et al. (2005)
1200 AU
Beltran et al. (2005); Hofner et al. (in prep.)
CH3CN(12-11)
NH3 red-shifted
NH3 bulk
NH3 blue-shifted
IRAS 20126+4104
Edris et al. (2005)
Sridharan et al. (2005)
NIR & OH masers
disk