Download Comparison of low- and high-mass star formation

3
on the
cate the
of stars.
d movie
ages.
Comparison of low- and high-mass
star formation — theory
No. 1, 2010
C
Krumholz
D
Jonathan Tan
(University of Florida)
Michael Butler (Zurich/MPIA)
Audra Hernandez (Wisconsin)
Karl Jaehnig
Wanggi Lim
Shuo Kong
Bo Ma
EBen Wu
Yichen Zhang (Yale/U.Chile)
OUTFLOW FEEDBACK REGULATED MASSIVE STAR FORMATION
the massive star formation is simply part of the rapid “core”
formation process. The use of sink particles in our simulation
enables us to go beyond the calculations of Banerjee et al. (which
stopped before significant mass accretion onto the stellar object)
and follow the mass accretion for a long time, and to reveal a
second phase of even more rapid accretion, driven by global
collapse.
The global collapse at the end of the pure hydro simulation
(t = 0.63 Myr) was already shown in Figure 2. It is further
illustrated in Figure 5, which shows the distribution of mass,
velocity field, and contours of gravitation potential in a subregion centered on the most massive object, with the positions
of stars superposed. It is clear that the global collapse has
supplied the central region near the bottom of the gravitational
potential well with plenty of dense material, which fuels both
the enormous rate of total stellar mass accretion (close to the
characteristic free-fall rate, see Figure 1) and the high accretion
rate of the most massive object, which is but one of many stars
in the region. In this crowded region, competitive rather than
core accretion is likely at work. The reason is that the stars in
are not enveloped by permanent host cores, because
et this
al.region
(2009)
Bonnell
etonto
al.stars
(2001)
the dense gas in the region is constantly
drained
and
constantly replenished by the global collapse. The most massive
object is located near the minimum of the global potential well
and accretes the global collapse-fed material at a rate higher than
any other object. Nevertheless, its accretion rate is only ∼10%
of the total rate. In other words, the vast majority of the globally
collapsing flow is diverted to stars other than the most massive
object. Its accretion rate is large (∼2.5 × 10−4 M⊙ yr−1 ) during
this late phase only because of an even larger global collapse
rate.
A common theme of the early and late phases of rapid
accretion is that the dense gas that feeds the most massive object
at a high rate is gathered by an agent external to the object. It is
the converging flow set up by the initial turbulence in the early
phase and the globally collapsing flow in the late phase. In this
aspect, the formation mechanism is closer to core collapse than
to competitive accretion. One may plausibly identify the dense
filament in the early phase and the dense region at the bottom
of the global gravitational potential well in the late phase as a
McKee–Tan core (McKee & Tan 2003). However, the “cores” so
identified are transient objects that are not in equilibrium. They
are evolving constantly, with mass growing (from converging
or collapsing flow) and depleting (into one or more collapsed
objects) at the same time. The replenishment of dense gas in
the “core” is the key to the long duration of high mass accretion
rate for the most massive object, which is many times the freefall time of the small dense “core” and comparable to the global
collapse time of the clump as a whole. In other words, there is no
hint of any drop in the mass accretion rate after a finite reservoir
of core material gets depleted over a (short) core free-fall time.
Indeed, the nearly constant or increasing accretion rate for the
most massive object highlights the important issue of when and
how the mass accretion is terminated; there is an urgent need
to tackle this issue if the final stellar mass is to be determined.
We will return to a more detailed discussion of the nature of
Sourav Chatterjee (NW)
Nicola Da Rio
Jan Staff
Kei Tanaka
Sven Van Loo (Leeds)
35
Wang et al. (2010)
Figure 6. Slice of the density through the most massive object (denoted by an
asterisk), showing a large magnetic bubble driven from near the central object
at t = 3.75tff,c = 0.79 Myr in the MHD model. Overplotted are velocity arrows
(white) and contours of the gravitational potential, which start at −10cs2 and
decrease inward in steps of 20cs2 .
plateau phases. The first starts somewhat later than that in
Maite
Beltrán
(Arcetri)
the HD
case, indicating
that the initiation of star formation
in general and massive star formation in particular is somewhat
Paola
Caselli (MPE)
delayed, as expected, because of magnetic cushion of turbulent
compression.
Once started, (SOFIA)
the accretion rate onto the most
James
De Buizer
massive object is, in fact, slightly higher in the MHD case
than in the HD case,
indicating that
the magnetic field does not
Francesco
Fontani
(Arcetri)
significantly impede the initial phase of rapid mass accretion,
which is enabled
by the dense (UNC)
structures formed by turbulent
Matthew
Goodson
compression. Indeed, dense structure formation is aided by a
strong magnetic
field, which forces
the turbulent flows to collide
Jonathan
Henshaw
(LJMU)
along the field lines (see, e.g., also Tilley & Pudritz 2007; Price
Izaskun
Jiménez-Serra
(ESO) the dense
& Bate 2008).
When viewed in three-dimensional,
structure resembles a warped, fragmented disk, with transient
Jouni
Kainulainen
spirals
that come and go. The (MPIA)
spirals indicate significant levels
of rotation on relatively small scales, which may hinder the
Mark
Krumholz
stellar
mass accretion due(UCSB)
to centrifugal barriers.
The centrifugal barrier may be weakened or even removed by
Christopher
magnetic braking. McKee
Evidence for the(UCB)
braking is shown in Figure 6,
which shows a magnetic bubble driven from the central region,
Romain
Teyssier (Zurich)
where active mass accretion is occurring. Magnetic braking
driven bubbles provide a form of energy feedback that should
accompany any star formed in a magnetized cloud (e.g., Draine
1980; Tomisaka 1998; Banerjee & Pudritz 2006; Mellon & Li
2008; Hennebelle & Fromang 2008) but is absent from purely
Comparison of low- and high-mass
star formation — theory & observation
IRDC G028.37+00.07
NASA/Spitzer/IRAC+MIPS (Butler et al. 2014)
IRDC G028.37+00.07 C1
NRAO/ALMA/N2D+(3-2) Tan et al. (2013)
Orion Nebula Cluster
(VLT; JHK) (McCaughrean)
Jonathan Tan
(University of Florida)
Michael Butler (Zurich)
Audra Hernandez (Wisconsin)
Karl Jaehnig
Wanggi Lim
Shuo Kong
Bo Ma
Ben Wu
Yichen Zhang (Yale/U.Chile)
Sourav Chatterjee (NW)
Nicola Da Rio
Jan Staff
Kei Tanaka
Sven Van Loo (Leeds)
Maite Beltrán (Arcetri)
Paola Caselli (MPE)
James De Buizer (SOFIA)
Francesco Fontani (Arcetri)
Matthew Goodson (UNC)
Jonathan Henshaw (LJMU)
Izaskun Jiménez-Serra (ESO)
Jouni Kainulainen (MPIA)
Mark Krumholz (UCSB)
Christopher McKee (UCB)
Romain Teyssier (Zurich)
The Importance of
Massive Stars
Abel+
Vogelsberger+
Whitmore+
Gillessen+
McCaughrean+
O’Dell+
Massive Star Feedback and
Regulation of SFRs
10 kpc
1 kpc
Tasker & Tan (2009)
~1 pc
Clump -> star cluster
nH (cm-3): 102(green) 105 (red)
~100 pc
GMC
Van Loo, Butler, Tan (2013)
Van Loo, Tan, Falle (2015)
Butler, Tan, Van Loo (2015)
The Physics of Massive Star Formation
A complicated, nonlinear process:
- Wide range of scales (~12 dex in
space, time) and multidimensional.
- Uncertain/unconstrained initial
conditions/boundary conditions.
Notation for the gas structures:
Core -> star or close binary
Clump -> star cluster
Numerical models
Observations
Complete theory of star formation
Analytic theory
- Gravity vs pressure (thermal, magnetic,
turbulence, radiation, cosmic rays) and
shear.
- Heating and cooling, generation and
decay of turbulence, generation (dynamo)
and diffusion of B-fields.
- Chemical evolution of dust and gas.
- Stellar structure and evolution
- Feedback
•
•
•
•
•
Massive Star Formation: Open Questions
Causation: external triggering or spontaneous
gravitational instability?
Initial conditions: how close to equilibrium?
Accretion mechanism: [turbulent/magnetic/thermalpressure]-regulated fragmentation to form cores vs
competitive clump-fed accretion vs mergers
Timescale: fast or slow (# of dynamical times)?
Salpeter (1955)
End result
-2.35
– Initial mass function (IMF)
– Binary fraction and properties
N*
dN*/dm* = A m*
m*max?
m*
How do these properties vary with environment?
Can we scale-up models of
low-mass star formation to protocluster environments with Σ ~ 1 g cm-2?
198
“Standard” model of isolated low-mass star formation
Shu, Adams, Lizano (1987)
Many observed examples of this for low-mass star formation.
Starless Core
CFHT/Cuillandre
Protostellar Core
NASA/JPL/Tobin
Protoplanetary Disk
NASA/McCaughrean/OʼDell
Can we find the equivalent stages for massive star formation?
Or do the crowded cluster environments of massive star
formation preclude this model?
Massive Star Formation Theories
Core Accretion:
wide range of dm*/dt ~10-5 - 10-2 M! yr-1
(e.g. Myers & Fuller 1992; Caselli & Myers 1995; McLaughlin & Pudritz
1997; Osorio+ 1999; Nakano+ 2000; Behrend & Maeder 2001)
Turbulent Core Model:
(McKee & Tan 2002, 2003)
Stars form from “cores” that fragment from
the “clump”.
If in equilibrium,
then self-gravity
is balanced by
internal pressure:
B-field, turbulence,
radiation pressure
(thermal P is small)
Cores form from this
turbulent/magnetized medium: at any instant
there is a small mass fraction in cores.
These cores collapse (quickly) to feed a central
disk to form individual stars or binaries.
Competitive (Clump-fed) Accretion:
(Bonnell et al. 2001; Wang et al. 2010)
Massive stars gain most mass by
Bondi-Hoyle accretion of ambient
clump gas.
Massive
stars form
on the
timescale
of the star
cluster.
Violent interactions? Mergers?
(Bonnell et al. 1998; Bally & Zinnecker 2005)
Schematic Differences Between
Massive Star Formation Theories
massive prestellar core
Rare evolution from
magnetically
subcritical state?
massive-star-forming core [protostar+gravitationally-bound gas]
massive-protostar (MP)
massive
Kunz & Mouschovias (2009)
star
m*f>8M!
Outflowconfined
HII Region
Turbulent core model
(MT02, 03)
time
Isolated massive
star formation?
(Bressert+2012; Oey+13)
Competitive Bondi-Hoyle accretion model
(Bonnell ea. 2001; Bonnell & Bate 2006; Dobbs+, R. Smith+, P. Clark+)
Prestellar core
mass function?
(e.g. Motte et al. 1998;
Testi & Sargent 1998;
Alves et al. 2007)
t=0
protostar
formation
m*=8M!
Radiation pressure likely
to prevent accretion of
dusty, unbound gas
(Edgar & Clarke 2004)
What are the appropriate initial
conditions for theoretical models?
Do massive starless cores exist?
How do we find them?
Mid-IR Extinction Mapping of Infrared Dark Clouds
(Butler & Tan 2009, 2012; see also Peretto & Fuller 2009; Ragan et al. 2009; Battersby et al. 2010)
G28.37+00.07
Spitzer IRAC 8µm
16’
(GLIMPSE)
(Churchwell et al. 2009)
MJy sr-1
Mid-IR Extinction Mapping of Infrared Dark Clouds
(Butler & Tan 2009, 2012; see also Peretto & Fuller 2009; Ragan et al. 2009; Battersby et al. 2010)
G28.37+00.07
Spitzer IRAC 8µm
16’
(GLIMPSE)
Median filter for background
around IRDC; interpolate for
region behind the IRDC
g cm-2
Correct for foreground using
“saturated” cores
~Arcsecond scale maps of
regions up to Σ ~0.5 g cm-2;
independent of dust temp.
Distance from molecular line
velocities -> M(Σ)
MJy sr-1
V
deep
8
µm
Spitzer-IRAC
imaging
of
massive
Infrared
DarkCloud
Cloud
(I
se
deep
8
µm
Spitzer-IRAC
imaging
of
massive
Infrared
Dark
(IRD
y extinction mapping. Merging with an NIR extinction map of the region
nfrared
(MIR)
extinction
map
that
probes
mass
surface
densities
-infrared
(MIR)
extinction
mass
densities
upup
to
veals
structures
down
to AV ∼ map
1 mag.that
Weprobes
utilize the
mapsurface
to: (1) measure
stgst
yet
by extinction
extinction
mapping.
Merging
with
Sample
ofkinematics
~50
massive
“starless”
thehighest
highest
values
yet probed
probed
by
mapping.
Merging
with
an a
iusthe
of
∼8
pc. 13values
CO
indicate
that
the cloud
iscore/clumps
gravitationally
(Butler
& Tan 2012;
Butler et al.
2014)
dynamic
range
reveals
structures
down
1 magW
V∼∼
of the
most massive
youngthat
starreveals
clustersstructures
known indown
the
Galaxy.
esaone
ahigh
high
dynamic
range map
map
that
totoAA
1 mag.
V (2)
4 densities (M=60M!)
Mass
surface
1313 COpolytropic
10
M
radius
of
∼8
pc.
kinematics
indic
emass
cores
within
findinga
can of
be∼8
fit by
singular
ud
mass∼7
∼7××the
104IRDC,
M⊙⊙ within
within
athey
radius
pc.
CO kinematics
indicate
−2 form one of the most massive young star
It They
has
potential
3.
have
≃ 0.1–0.4
g cmto
—relatively
along young
with star c
d.
Itthus
thus
hasΣthe
the
potential
one oflow
thevalues
mostthat,
massive
t that magnetic
fields,
rather
thanmassive
accretion-powered
radiative
heating,
are
terize
the
of
16
cores within
within
theIRDC,
IRDC,finding
finding
the
acterize
thestructures
structures
of concentration
16
massive
cores
the
they
Cores
show
central
of
these
(3)
the
Σ (equivalently
column
density
orgAcm
) −2−2
ρ ρ Determine
1.5±0.3
andkkρρ =
= 1.3
ΣΣ≃≃
0.1–0.4
—rela
eswith
withρcores.
ρ∝∝r r−k−k
and
± 0.3.
0.3. They
Theyhave
have
0.1–0.4
gV cm
—r
r a region that is nearly complete for AV > 3 mag. The PDF is well fit by
measured
coldtemperatures,
temperatures,
suggest that
than
accreti
easured
cold
suggest
thatmagnetic
magneticfields,
fields,rather
rather
than
accr
Contain
many
mag, high
compared
toJeans
other masses.
known clouds. It does not exhibit a separate
rtant
for
controlling fragmentation
fragmentation
of
(3)
the Σ (eq
ant
for
controlling
of these
thesecores.
cores.
(3)Determine
Determine
Fragmentation
not
suppressed
claimed
to indicate the
importance
of self-gravity.
However,
we suggest the Σ (
bility
distribution
function
(PDF)
for
complete
AV
by radiative function
heating
(c.f.,
lity
distribution
(PDF)
for aa region
region
that
nearly
complete
lar, self-gravitating
hierarchy
of structures
presentthat
overisis
anearly
wide
range
of forfor
Krumholz
& McKee
gle
log-normal
with
mean2008).
AVV ≃
≃ 99 mag,
clou
e log-normal
with
mean
A
mag, high
highcompared
comparedtotoother
otherknown
known
c
end
power
lawtail,
tail,which
which has
has been
been claimed
of os
d –power
law
claimedtotoindicate
indicatethe
theimportance
importance
ds
stars:
formation
hePDF
PDFdoes
doesresult
resultfrom
from aa self-similar,
self-similar, self-gravitating
structu
self-gravitatinghierarchy
hierarchyofof
stru
sninthe
thecloud.
cloud.
Fragmentation
suppressed by B-fields?
Critical– Mass
McKee
1992)
words: Magnetic
dust, extinction
ISM: (Bertoldi
clouds – &stars:
formation
rds: dust, extinction – ISM: clouds – stars: formation
e-only material: color figures ′′
telescope;
∼22 for the 250 µm observations of Peretto et al.
only material: color
figures
-3, B~200μG -> M ~100 M
nH~105cm
B
2010
with Herschel-SPIRE).
!
clusters from
Interferometric observations are
possible at submillimeter and longer wavelengths, but on their
own provide poor constraints on dust temperature.
Dynamics of Massive Starless Cores:
Are they close to Virial Equilibrium?
The Astrophysical
Journal, 754:5 (22pp), 2012 July 20
Butler & Tan
Four IRDC core/clumps
selected
to be dark at 8, 24, 70 μm
C1
F1
G2
F2
So use high angular resolution observations of N2D+(3-2) to
1. Identify exact location of (massive) starless cores
2. Measure core velocity dispersion, σ.
High Deuterium Fraction [N2D+]/[N2H+]
(Fontani et al. 2011)
n
CO freeze-out
e.g. Hernandez et. al (2011)
H3+ + CO → HCO+ + H2
(T<20K)
H3 + HD → H2D+ + H2
+
H2D+ + N2 → H2 + N2D+
aged radial profile figures (notation as in Figure 5(b)).
ournal.)
IRAM 30m
Astrochemical
Figure 10. Core E2, E3, F1, F2, F3, and F4 Σindicator
maps (notation asthat
in Figurethese
5(a)) and azimuthally av
(A color
version of this cores
figure is available in the onlin
are
starless
(Caselli et al. 2002)
13
C1
F2
G2
0.91
0.250
0.070
C1-N
Tan, Kong et al. (2013)
0.246
-0.564
0.244
-0.566
C1-S
G2-N
0.248
b (degrees)
“Massive”
“Starless”
“Cores”
with ALMA
F1
G2-S
0.74
0.068
0.246
0.242
-0.568
0.240
-0.570
0.238
-0.572
0.066
0.244
0.58
0.064
28.328
28.326
28.324
28.322
28.32 4.424
34.422
34.420
34.418
34.416
0.250
440
34.438
34.436
34.434
34.432
34.784
0.246
-0.564
0.244
-0.566
0.242
-0.568
0.240
-0.570
34.782
34.780
34.778
34.776
0.070
C1-S
0.248
G2-N
0.068
G2-S
0.246
0.33
0.066
0.244
0.064
-0.572
0.238
28.328
28.326
28.324
28.322
28.32 4.424
34.422
34.420
34.418
34.416
440
34.438
34.436
34.434
34.432
34.784
34.782
34.780
34.778
34.776
0.23
0.250
0.246
-0.564
0.244
-0.566
0.242
-0.568
0.240
-0.570
0.238
-0.572
0.070
DCO+(3-2)
b (degrees)
DCO+(3-2)
C1-N
C1-S
0.248
G2-N
G2-S
0.068
0.246
0.15
0.066
0.244
0.082
0.064
28.328
28.326
28.324
28.322
28.32 4.424
34.422
34.420
34.418
34.416
0.250
440
34.438
34.436
34.434
34.432
34.784
0.246
-0.564
0.244
-0.566
0.242
-0.568
0.240
-0.570
34.782
34.780
34.778
34.776
1.3mm
1.3mm continuum
b (degrees)
0.070
0.036
C1-N
C1-S
0.248
G2-N
0.068
0.246
G2-S
0.009
0.066
0.244
0.064
-0.572
0.238
28.328
28.326
28.324
l (degrees)
28.322
28.32 4.424
34.422
34.420
34.418
l (degrees)
34.416
440
34.438
34.436
34.434
l (degrees)
34.432
34.784
34.782
34.780
l (degrees)
34.778
34.776
0
Mass Surface Density (g cm-2)
N2D+(3-2)
b (degrees)
N2D+(3-2)
0.45
C1-N
Comparison to Turbulent Core Model
core
C1, ΣMIREX, N2D+(3-2)
contours
Σcl
Mc
clump
3.6”
0.09pc
Core masses inside 3σ
N2D+ contour:
Σcl = 0.36 g cm-2
Mc,MIREX = 55.2 ± 25 M!
Mc,mm = 62.5 129 26.9 M!
ALMA beam
Spitzer beam
Predictions from Virial Equilibrium
• 1D velocity dispersion if virialized:
(
= 1)
Core
C1-N
C1-S
F1
F2
G2-N
G2-S
Σcl (g cm-2)
0.48
0.40
0.22
0.32
0.21
0.19
Mc 4
(M!)
16
6.5
4.7
2.4
0.83
Tan et al.
63
σvir (km/s)For0.66±0.22
0.88±0.30
the sample of
6 cores the0.43±0.15
ratio of the 0.44±0.15
observed to 0.33±0.11
the predicted0.25±0.09
virial equilib-
rium velocity
dispersion
is 0.810.97
, while the ratio
of the observed
to predicted
size is
0.71
σobs (km/s)
0.41±0.03
0.41±0.02
0.25±0.02
0.42±0.04
0.34±0.02
0.30±0.02
1.541.97
1.26 . Thus the cores appear to be close to the predictions of the model. For the most
< σobs
= 0.81
0.13 than
massive core, C1-S, the observed velocity dispersion
is /σ
about
a factor
of 2±smaller
vir >
the predictions of the fiducial model. For it to be in virial equilibrium would require
stronger magnetic fields of ∼ 1.0 mG, implying mA ≃ 0.3. In fact, given the core den5 cm=
−30.28
-> Bvir=0.9mG
sity of nH ≃ 6 ×m
10A,vir
, the predicted
median B-field strength using Crutcher et al.’s
0.65
(2010) relation Bmed ≃ 0.12nH µG (for nH > 300 cm−3 ) is(Crutcher
0.7 mG. et al. 2010)
nH,c=6.4x105cm-3 -> Bmed=0.7mG (see also Pillai+ 2015)
4. The
Chemical “Deuteration”
Ages
of the
Cores
Tentative
Conclusion:
Cores appear
to be
near
virial equilibrium, after
accounting for clump envelope. Possibly slightly sub-virial; or have stronger
The Turbulent Core Model does not make any prediction about the timescale for the
B-fields. But we need a larger sample - massive starless cores are rare!
cores to assemble, except that at least one dynamical (i.e. or ≃ 2 free-fall times, tff )
is needed to reach an equilibrium state. If strong magnetic fields are regulating core
formation, then the timescale for core formation could be considerably longer than this.
CO freeze-out
e.g. Hernandez et. al (2011)
H3+ + CO → HCO+ + H2
(T<20-30K for small OPRH2)
H3+ + HD → H2D+ + H2
H2D+ + N2 → H2 + N2D+
Observed
Dfrac of C1-S
Collapse rate αff ≲ 0.3
# =1
αff
Model deuteration of N2H+ with “complete,”
gas-phase, spin-state astrochemical network
to constrain age and/or rate of collapse.
αff#=
Kong, Caselli, Tan, Wakelam, Sipilä (2015);
see also Pagani et al. (2009, 2013)
.01
The Deuteration Clock
Summary of Initial Conditions
• Massive starless cores exist
• Require strong B-fields for dynamic
stability and to prevent fragmentation
• Chemically old compared to free-fall time
Constraints for Initial Conditions of Numerical Simulations
Peters et al. (2011)
M = 100M!, R=0.5pc,
nH = 5400cm-3, B=10μG
Seifried et al. (2012)
M = 100M!, R=0.25pc,
nH = 4.4x104cm-3, B~1mG
Myers et al. (2013)
M = 300M!, R=0.1pc,
Disc formation
in turbulent mass
6cm-3, B>~1mG
nH = 2.4x10
8
disc 1
disc 2
disc 3
disc 4
7
6
µ
5
4
3
2
1
0
0
5
10
t / kyr
15 0
5
t/
Figure 3. Mass-to-flux ratio µ (left) and incli
magnetic field to the angular momentum vecto
in spheres with a radius of 500 AU around th
found in run 2.6-4-A.
purpose we calculate the volume-weighted
field hBi in a sphere with a radius of r =
the CoM of each disc. In combination with
M we obtain the mass-to-flux ratio
Figure 2. Column density in logarithmic scaling for the top-on
view of disc 1 (top left) and disc 2 (top right) of run 2.6-4-A and
of the disc in run 2.6-4 without turbulence (bottom). The figures
are 800 AU in size.
lines. This is a remarkable result since for previous simulations of low- and high-mass cores with mass-to-flux ratios
µ . 10 only sub-Keplerian discs were found (e.g. Allen et al.
2003; Price & Bate 2007; Mellon & Li 2008; Hennebelle &
Fromang 2008; Duffin & Pudritz 2009; Seifried et al. 2011).
We emphasise that for the other runs we find qualitatively similar results, i.e. discs with sizes of up to ⇠ 100 AU
and masses of the order of 0.1 M . The number of discs per
µ=
M
⇡r2 | hBi |
0.13
/p .
G
We plot the time variation of µ in the lef
for the same four discs as in Fig. 1. As can
around a mean of 2 - 3. Hence, the val
agree with the overall value of 2.6 and a
the value of ⇠ 2 found in run 2.6-4. More
the range where simulations without turbu
found sub-Keplerian discs only. We therefo
turbulent reconnection is not responsible fo
Keplerian discs in our runs.
Another way of reducing the magnetic
was investigated by Hennebelle & Ciardi (
What is the accretion mechanism of
massive protostars?
Outflowconfined
HII Region
Continuum Radiative Transfer Modeling
Temperature
Density
Robitaille, Whitney et al. (2006+); Indebetouw et al. (2006); Molinari et al. (2008); Zhang & Tan (2011), Zhang et al. (2013, 2014)
hydrostatic core
expansion wave
rotating infall
active accretion disk
disk wind
dust destruct. front
gas + dust opacities
protostellar evolution
Parameters:
Σclump = 1 g cm-2
Mcore = 60 M⦿
β
= 0.02
m* = 8 M⦿
Lbol = 6x103 L⦿
Continuum Radiative Transfer Modeling
Zhang, Tan & Hosokawa (2014)
The Astrophysical Journal,
100 788:166
AU (35pp), 2014 June 20
1000 AU
20,000Zhang,
AU Tan, & Hosokawa
Figure 5. Input density and converged temperature profiles for the fiducial model (Mc = 60 M⊙ , Σcl = 1 g cm−2 , βc = 0.02) at selected stages with m∗ = 1, 2, 4, 8,
density
temperature
temperature
density
temperature
12, 16, and
24 M⊙ . At each stage
(each row), these profiles aredensity
shown on three different
scales (from left to right, 100
AU, 1000 AU, and 20,000
AU). In each panel,
the protostar is at the (0,0) point, the x-axis lies on the disk midplane, and the y-axis along the outflow axis. The velocity fields are shown by the arrows. Note that the
black arrows for the outflow have a much larger scale than the blue arrows for the infalling envelope.
(A color version of this figure is available in the online journal.)
11
The Astrophysical Journal, 788:166 (35pp), 2014 June 20
Zhang, Tan, & Hosokawa
1M⦿
2M⦿
4M⦿
NIR to FIR
morphologies
Rotation and
outflow axis
inclined at 60˚
to line of sight.
8M⦿
12M⦿
16M⦿
24M⦿
Figure 21. Resolved images for the selected evolutionary stages (m∗ = 1, 2, 4, 8, 12, 16, and 24 M⊙ , from top to bottom) of the fiducial model in various bands
(columns) at the inclination of 60◦ between the line of sight and the axis. Each image is normalized to its maximum surface brightness, which is labeled in the bottom
left corner. The total fluxes are labeled in the top right corners. A distance of 1 kpc is assumed. Each image has a field of view of 40′′ × 40′′ . The dotted lines mark the
projected opening angle of the outflow cavity on the sky plane.
Prediction:
increasing
symmetry
from MIR-FIR
Massive Protostar G35.2N: d=2.3kpc; L~105L⦿
Gemini-T-ReCS
SOFIA-FORCAST
radio (cm)
continuum
(ionized gas)
De Buizer (2006)
Zhang, Tan, De Buizer et al. (2013)
Spectral energy distribution
Flux profiles along
outflow cavity axis
o
t
t
fi
d
o
o
Σ
=1g
g
s
e
.
M
= 240 M
d
d
i
g
e
l
v
i
n
i
o
a
m
= 34 M
t
r
m
e
p
r
l
d
o
f
e
:
s
d
s
i
e
o
l
r
fi
m
a
t
o
c
r
s
i
p
e
tΣrcore/clump
MIR SED requires high
y
v
e
i
t
i
s
m ens
s
a
m
y
t
m
s
n
i
,
a
e
le
w
g
p
o
a
m
h
m
i
Si
n
o
&
s
t
D
L ~ (0.66 - 2.2)×10 L
n
E
i
S stra
M ~ 240M
n
co
Ʃ ~ 0.4 - 1 g/cm
clump
cm-2
core
*
⦿
⦿
5 ⦿
bol
core
t
a
s
e
g
t
l
s
i
h
f
fi
l
1
t
P
i
c
o
W
T
fl
⦿
cl
2
θw ~ 35 - 51˚
θview ~ 43 - 58˚
m* ~ 20 - 34 M⦿
Fig. 4.— Intensity profiles along the outflow axis. The squares
are observational data sampled at intervals of the resolutions of
n
u
e
Z
s
w
m
Massive stars play a crucial role in the production of heavy elements and in the evolution of the
interstellar medium, yet how they form is still a matter of debate. We report high-angular-resolution
submillimeter observations toward the massive hot molecular core (HMC) in the high-mass
star-forming region G31.41+0.31. We find that the evolution of the gravitational collapse of the
HMC is controlled by the magnetic field. The HMC is simultaneously contracting and rotating,
and the magnetic field lines threading the HMC are deformed along its major axis, acquiring an
hourglass shape. The magnetic energy dominates over the centrifugal and turbulence energies,
and there is evidence of magnetic braking in the contracting core.
lecular core surrounding the protostars
figuration that was shown to be con
theoretical models for the formation o
stars, where well-ordered, large-scale
turbulent, magnetic fields control th
and collapse of the molecular cores
stars form (12).
We investigated the hot molecular
in G31.41+0.31 (G31.41), a massive
region [~500 to 1500 M◉ (13, 14)] l
parsecs (pc) away (15). G31.41 has
Continuities between Low- & High-Mass Star Formation
S
tars more massive than 8 M◉ (where M◉ is
the mass of the Sun) account for only 1%
of the stellar population in our Galaxy.
Nevertheless they dominate the appearance and
evolution of its interstellar medium and are responsible for the production of heavy elements.
The formation of massive stars is not completely understood. Stars form when dense molecular clouds collapse as a result of gravity. But
as the mass of a young star reaches 8 M◉, its own
radiation can exert enough outward pressure to
halt infall, inhibiting further stellar growth (1).
The presence of a flattened accretion disk surrounding the protostar (2) can alleviate this in-
Core Mass Function:
e.g., Nutter & Ward-Thompson (2007)
Protostellar Cores &
B-field morphologies:
hibition by shielding the infalling material from
stellar radiation and by creating a lower density
section along the rotation axis of the disk and a
molecular outflow, which helps by channeling the
radiation out, allowing the formation of stars more
massive than 40 M◉ (3–5). Massive stars may also
form through mergers of smaller stars (6).
The scenario whereby massive stars form
through disk-assisted accretion resembles the
way stars like the Sun form. Both processes
involve accretion through a flattened disk and
molecular outflows. The magnetic field is thought
to play an important role in the formation of Sunlike stars by shaping cloud collapse, removing ex-
1
Institut de Ciències de l’Espai [Consejo Sup
tigaciones Científicas (CSIC)–Institut d’Estudi
(IEEC)], Campus Universitat Autònoma de Ba
Facultat de Ciències, Torre C5 - parell 2a, 08
Catalunya, Spain. 2Departament d'Astronomia
(IEEC-UB). Institut de Ciencies del Cosmos y Un
CSIC, Universitat de Barcelona, Martí i Fran
Barcelona, Catalunya, Spain. 3Harvard-Smithso
Astrophysics, 60 Garden Street, Cambridge, M
4
Academia Sinica, Institute of Astronomy and A
North Aohoku Place, Hilo, HI 96720, USA.
*To whom correspondence should be addr
[email protected]
†Present address: Osservatorio Astrofisico d
Enrico Fermi 5, 50125 Firenze, Italy.
A
B
Girart et al. (2009); Q. Zhang et al. (2014)
Outflow properties:
- collimated morphologies (Beuther et al.
2002; Duarte-Cabral et al. 2013)
[however, BN/KL - Bally & Zinnecker 2005]
- momentum flux vs. luminosity
- CO ladder (talk by van Dishoeck)
Astrochemical properties:
- Deuteration (Fontani et al. 2011, 2015)
- Hot Cores/Corinos
Fig. 1. (A) Contour map of the 879-mm dust emission superposed on the color
image of the polarized flux intensity in units of Jy per beam. Black thick bars
indicate the position angle of the magnetic field. These maps were obtained by
using a natural weighting to the visibility data, which yielded to a full width at
half maximum synthesized beam of 1.34″ × 0.83″ with a position angle of 67°
(shown in the bottom left corner). Contour levels are 0.8, 1.5, 2.5, 4, 6, 16,
26, 36…96% of the peak intensity, 9.13 Jy per beam. (B) Contour map of the
879 mm dust emission superposed on the color image of the flux weighted
velocity map of the CH3OH 147-156 A. Black thick bars indicate
of the magnetic field. These maps were obtained by using a robu
Girart et al. (2009)
of 0 to the visibility data, which yielded to a full width at ha
1408
12 JUNE 2009
VOL 324
synthesized beam of 1.04″ × 0.59″ with a position angle of 82° (
bottom left corner). Contour levels are the same as in the previou
a peak intensity of 6.55 Jy per beam. (C) Spectrum of the C34S 7position of the dust emission peak. The continuum has been sub
the line emission (this is valid for all the molecular line data pres
SCIENCE
www.sciencemag.org
Duarte-Cabral et al. (2013)
Conclusions
Core Accretion, Competitive (Clump-fed) Accretion, Protostellar Mergers are
potential formation mechanisms of massive stars.
Key questions: Does clump fragment into cores with range of masses,
including massive starless cores? If so, what prevents fragmentation?
Magnetization strongly affects the outcome of numerical simulations.
“Turbulent Core Model”:
- Core surface pressure set by surrounding clump pressure
- Core supported by nonthermal pressure (B-fields/
turbulence), but otherwise this is a scaled-up version of lowmass star formation
- mG B-fields are likely to be preventing fragmentation
- Massive protostars will ionize their outflows
No. 1, 2006
Observations & Modeling:
- Massive, deuterated, starless cores exist and may
be near virial equilibrium, with slow collapse
compared to free-fall
- Massive protostars can have a similar morphology
to low-mass protostars. B-field to regulate collapse &
fragmentation.
- Can sometimes be influenced by cluster
environment (e.g., BN/KL)
MID-INFRARED JET OF G35.20!0.74
Fig. 2.—The G35.20!0.74 jet as seen at different wavelengths. (a) The 11.7 mm image in false color overlaid with K
white contours) and the 8.5 GHz high-resolution radio continuum emission of Gibb et al. (2003, gray contours). (b) Th
with the low-resolution 15 GHz radio continuum image of Heaton & Little (1988, white contours) and L′ image from Full
in on the central region of the 11.7 mm image in false color, the L′ contours in white and the high-resolution radio conti
of Hutawarakorn & Cohen (1999) are shown as asterisks, water masers of Forster & Caswell (1989) as crosses, and meth
communication) as large plus signs. The bars at lower right show the !1 j relative astrometric uncertainty between the
infrared emission coincident with and immediately north of the
to ∼16,000 AU by a B2.6 sta
Figure 3: Cycle 2 observations of Cep A at 7.7µm (left) & 37.1µm (right). Similar to G35.2, at shorter wavelengths
position
of cavity
G35.2N
demonstrates
that wavelengths
the infraredthe
emission
here
couldDetailed
heat out to the distan
only the near-facing
outflow
is seen,
while at longer
far-facing
cavity is one
revealed.
by longer
emission.models
There-& test formation
a typical
size of 0.005 mm,
RT models of thisis&dominated
the other sources
will bewavelength
carried out tocontinuum
constrain protostellar
theories.
Mass surface density →
Σ - M Diagram
Giant Molecular Clouds
Star-Forming Clumps
Young Star Clusters
IRDC Clumps & Cores
Physical Properties of
Star-Forming Regions
AV=230
A8μm=8.1
NH=4.2x1023cm-2
Σ=4800 M! pc-2
AV=7.5
A8μm=0.30
NH=1.6x1022cm-2
Σ=180 M! pc-2
Tan et al. (2014, PPVI)
Mass →
Fig. 1.— The Environments of Massive Star Formation. Mass surface density, ⌃ ⌘ M/(⇡R2 ), is plotted versus mass, M . Dotted