Examining the Evolutionary Sequence of Massive Stars

1
Examining the Evolutionary Sequence of Massive
Stars

• Tracey Hill
• (UNSW/ATNF)
• Collaborators
• Michael Burton (UNSW), Maria Hunt (UNSW), Jim
  Caswell (ATNF), Vincent Minier (CEA), Andrew
  Walsh (UNSW), Mark Thompson (UHerts).

2
The study of Massive Stars

• Why?
• We are made of star dust
• The death and birth of stars may be linked.
• Complex molecules can form on dust grains
• Young stars stir up clouds of gas
• Stars often harbour planets
• Evolution not well understood
• CEDAR
• Optically obscured by circumstellar dust prior to
  MS
• HOW?
• Associated with IRAS pt sources, UC HII, maser
  emission, mid-IR (MSX), sub-mm, mm

3
Massive Star Formation

• Evolution not well understood.
• Evolve quite quickly compared to low mass stars.
• Difficult to detect,
• Occur at large distances,
• Deeply embedded
• Optically obscured by circumstellar dust
• Not optically visible prior to MS phase.
• Form only in clustered mode (?)- difficult to
  distinguish between cores.
• Associated with maser emission, IRAS point
  sources, UC HII regions, molecular outflows.

4
Evolution of a Massive Star

• CH3OH OH
  H20
• Cold Hot
  
• GMC Core Molecular
  UC HII HII
• 
  Core
• SED mm Mid IR
  Near IR
• Molecular lines evolving (?)

5
Why mm-wavelengths?

• Gas cold need to go to longer ? to see.
• At mm and sub-mm wavelengths, both the continuum
  and molecular line emission is strong for stages
  prior to the formation of the UC HII region.
• mm and sub-mm data provide information about the
  dust emissivity exponent (?).

Need multiple wavelength studies for the full
picture of MSF.
6
Aims of my thesis

• To undertake a multi-wavelength study of star
  formation from the mid-IR to mm.
• From this determine the SED from mid-IR to mm for
  a range of sources.
• To propose an evolutionary sequence for MSF.
• To determine the dust grain properties of MSF
  regions. (mm and sub-mm)
• To determine the significance in fluctuations of
  the dust grain emissivity exponent (?).

7
Results (SEST)

• Targeted known positions of methanol masers and
  UC HII regions (129) using SIMBA.
• 404 sources (3-? detection limit).
• 100 of sources targeted have mm continuum
  emission. 20 others in fields.
• In the majority of sources, the position of the
  tracer targeted correlates with the peak
  millimetre emission.
• Evidence of methanol masers and UC HII regions
  devoid of millimetre continuum emission.
  Implications?
• Evidence of star formation devoid of methanol
  masers and/or UC HII regions (mm-only cores)
  Implications?
• Other Data
• ATCA (3mm) on few cores. Resolving bright SIMBA
• JCMT, fallback observations (Sept03), CS line
  data (Aug04) collaborators data, time allocated
  in Semester 05 A.

8
Clustered SIMBA
9
SCUBA data

• Sources targeted- mm-continuum sources with no
  tracers.
• Are these sources at an earlier evolutionary
  sequence prior to the onset of methanol masers?
• To be tested with SEDs and dust emissivity
  exponent.

10
ATCAResolving bright SIMBA sources

• Data from Aug02 run suggests - YES!
• Re-reduction of complimentary config. req.

11
More ATCA results

• ATCA Contour levels 20, 40, 60, 80

12
Introducing the mm-only core

• 60 sources (253/404) detected have no maser
  /or UC HII (mm-emission only). mm-only core
• Readily detectable at sub-mm wavelengths.
• Sept 03 JCMT data, showed 100 detection.
• Time allocated to observe JCMT 05A (April?)
• 50 do not have mid-IR MSX emission. lower
  limit or are devoid of a mid-IR source.
• What is their story?
• Younger? Deeply embedded? Intermediate mass?
  combination?

13
A Range of MSF Cores1.2mm Continuum, SEST/SIMBA
CH3CN
HCN
hanol
HCO
CH3OH
(Reverse) Evolutionary Order??
Minier, Purcell, Hill et al, 2004
14
Data Analysis

• 4 classes of source (diff. evolutions?)
• mm-only (mm), maser (mas), maser and radio (mr),
  radio (rad).
• Assumed near distance to all with an ambiguity
• 12 have no known distance.
• Parameter analysis
• Mass, radius, H2 number density (nH2)
• Kolmogorov-Smirnov testing (K-S testing).
• Cumulative plots of mass for each distribution.
• Histogram plots of each parameter
• Correlation plots of parameters

15
Kolmogorov-Smirnov Testing

• To test whether two distributions are drawn from
  the same distribution function.
• Disproving the null-hypothesis proves that the
  data sets are from different distributions.
• Failing to disprove the null-hypothesis shows
  that the data sets are consistent with a single
  distribution function.
• Null hypothesis that the two groups are the
  same.

16
How does K-S testing work?

• Cumulative plot of each distribution.
• Measures the maximum value of the absolute
  difference between two cumulative distribution
  functions. (it is the behaviour between the
  largest and smallest values that distinguishes a
  distribution).
• This is called the k-s statistic D.
• The Prob (P) of D gt obs is calculated.
• The null hypothesis (that the two groups are the
  same) should be rejected if P is "small".

17
Parameter Mass

• KS-test mm-only distributions are not from the
  same population as maser, mr, or radio.
• maser and radio populations produce D stat
  consistent with being related. confirms work of
  Walsh et al and evol. seq. of MSF

18

• m ? r2.2

mm-only dominate low-mass low-radius end
19
Results - Summary

• mm-only cores are smaller and less massive than
  cores with a maser and/or UC HII region.
• mm-only cores have a range of masses consistent
  with those cores with tracers.
• Mean mass mm-only 0.9 x 103 M?
• Mean mass masers UC HII 2.5 x 103 M?
• mm-only cores have radii lt 2.0 pc (bar one), with
  the majority (94) lt 1.0 pc.
• Mean radius 0.4 pc
• Mean radius of masers UC HII 0.7 pc

20
Interpretation of the mm-only core

• Precursor to the maser?
• New class of source that represents the earliest
  stage of massive star formation prior to onset of
  maser emission.
• Intermediate mass star formation?
• Harbour protoclusters which do not contain any
  high mass stars (below HII limit).
• Cross-section of sources supporting both
  arguments?
• More massive mm-only cores support 1.
• Less massive mm-only cores support 2.
• Hill et al. MNRAS 2005, submitted

21
Further Work
22
Spectral Energy Distributions

• The SED for each source is fitted using a two
  component grey-body function.
• Models the emission from a warm dust core
  embedded in a larger cold envelope.
• The SED is compiled using data spanning the MIR
  to the mm regime of the Electromagnetic spectrum.
  
• MSX data at 8.3?m, 12.1?m, 14.6?m, 21.3?m.
• Sub-mm data (SCUBA) at 450?m and 850?m.
• mm data (SIMBA) at 1.2mm.
• IRAS data (?) at 12-100?m.
• Reveals information relating to the source
• Temperature (leads to estimate of age, and hence
  ES)
• Luminosity, density and calculation of mass.

23
Compiling a SED

• From multi-Wavelength data to the SED

24
G12.86-0.27
25
G192.60-0.05
26
The Dust Emissivity Exponent?

• Tells us about the behaviour of the dust
  emissivity with wavelength (Dunne Eales 2001).
• Identifies the type of grain which makes up the
  central star forming core.
• Ex. ? 2 crystalline grains, while ?1
  amorphous carbon grains, ?2 for graphitic grains
  (Dunne Eales 2001).
• Ex. A value of ? 1.52.0 is indicative of class
  0 (collapsing protostar) sources (Furuya et al. )
• Ex. ? 1.5 infers composite grains, while 0.6 ? ?
  ? 1.4 infers fractal grains (Dunne Eales 2001
  and references within).
• Sub-mm/mm emissivities are particularly
  important
• Molecules known to deplete inside protostellar
  cores.
• Dust emissivity best tracer of gas density
  distribution just prior to onset of gravitational
  collapse. i.e define the initial conditions
  from which a core collapses to form a star.

27

• What does ? reveal?
• The dust mass the mass of the star forming
  cloud.
• The star formation efficiency.
• Can determine the dust-to-gas ratio (Hoare et al.
  1991).
• ? is complicated by
• Grain size, grain shape (assume spherical?),
  grain mixtures. (Hildebrand 1983).
• Temperature (? in come cases), emissivity, and
  extinction with redshift and metallicity (Dunne
  Eales 2001).
• Not much work done observationally.
• Most generally assume a value of ? when fitting
  the B/B fxn to data.
• Computationally Ossenkopf and Henning (1994).
• Hildebrand (1983), Dunne Eales (2001).
• Models and observations suggest that emissivities
  increase in dense cores (Ossenkopf Henning
  1994).

28
How do we determine ??

• F??c B? ?? where ?? (1-e-?)
• F??c 2hc2/?3 1/(ehc\?kT 1) ?0 (?o/?)?
• Let A ?c 2hc2?o? ?0 , then
• F? A/?3 1/(ehc\?kT 1) (1/?)?
  
• F? A/?3? 1/(ehc\?kT 1)
• Fit function and derive ?

29
Assumptions (preliminary)

• A 10 error assumed for all sub-mm and mm fluxes.
• RJ-approx to B/B fxn.
• Dealing with cold objects.
• B/B fxn begins to break down.
• However, 450?m may not satisfy the
  RJ-approximation in all cases.
• RJ-approx breaks down hence, results incorrect.

30
Preliminary Results ?
31
Interpreting the results

• Typically, values of ? fall between 1-2
• Some ? values are extreme ie. 0.4, 3.5
• This may be due to incorrect fluxes.
• The fits themselves may not be robust.
• Error contour plots show a high correlation
  between the gradient (?) and the y-intercept of
  the plots.
• What is the significance (if any) in
  fluctuations of the dust emissivity exponent - ??

32
Future Work

• SED compilation for all sources.
• Determination and examination of beta
• Compare with results determined theoretically
  (Ossenkopf Henning, Dunne Eales).
• Indication of sturdiness of our results.
• Interpretation of our results - what each beta
  value translates to in terms of grain properties.
• Finish project and write up thesis.