AGES OF YOUNG STARS
D. Soderblom (Space Telescope Science Institute, Baltimore, United States), L. Hillenbrand (Caltech, Pasadena, United States), R. Jeffries (Astrophysics Group, Keele University, Staffordshire, United Kingdom), E. Mamajek (Dept. of Physics & Astronomy, University of Rochester, United States), T. Naylor (School of Physics, University of Exeter, United Kingdom)
Determining the sequence of events in the formation of stars and planetary systems and their timescales is essential for understanding these processes, yet establishing ages is fundamentally difficult because we lack direct indicators. In this review we discuss the problem of ages for young stars, specifically for those less than ~ 100 Myr old. We start by establishing a reliable scale scale for young stars using the measurement of the Lithium Depletion Boundary (LDB) in young clusters, a method that involves fairly simple and wellestablished physics and observations. We show that LDB ages are consistent with those from the upper main sequence and the main sequence turn-off if modest core convection is included in the models of higher-mass stars. The LDB method can be used to set limits on the age of individual objects but is primarily applicable to groups of stars. We then review the available methods for age estimation, which include kinematic traceback of young groups, placing stars in HRDs, pulsations and seismology, gravity differences, rotation, activity, and lithium abundance as age indicators. We list the known strengths and weaknesses of each of these. Some of these techniques are useful within certain limits (such as age and mass ranges) and others are entirely problematic. We look at the issue of age spreads within star-forming regions and how well they can be quantified. Finally, we offer suggestions and recommendations to guide interpretation of observations, and we show how the current situation can be improved over the next few years.
Talk
Can measure radii, luminosity, but never age directly.
Ages matter quite a lot: use them to age-date everything else
Ages during the first Gyr of stellar lifetimes.
- Steps:
- Categorise stars
- Put them in age-order
- Estimate age ratios
- Estimate absolute ages?
- What methods can we use?
- No fundamental methods available
- Nearly fundamental = lithium depletion boundary, kinematic traceback
- model-depndent: isochrone fitting, asteroseismology, surface gravity, radii, radiative-convective
- not useful: metallicity
Lithium depletion
- fully convective star drags lithium to core, processes lithium
- Lithium burns at 1e6 K
- Why semi-fundamental? Very good agreement at which luminosity Li dep. occurs
- 5-10% error
- easy to measure at 6701 AA
- can't go past 200 Myr
- models agree on luminosity vs age, but disagree strongly on temperature
- Measure lithium in whole cluster, look for boundary * 8 measurements: 22-122 Myr old * uncertainty from unresolved binaries
Below 20 Myr? Kinematic ages, independent of stellar physics
- traceback ages - trace orbits back to closest location
- expansion age: position vs. velocity.... hubble expansion
- flyby age: find when object was closest to group [how does this help?]
- runaway age: to age-date ejected sources. Assume that ejection occurred at birth (sn ejection)
- Example: Beta pic, trace back to a common smaller volume (40 pc across)
- problems: some stars don't fit
- is 40 pc across really a reasonable coeval group?
- actually, expansion is inconsistent with 12 Myr age using revised Hipparcos
- Runaways: what if they ran away from a different cluster?
- ONC example - must be absolutely sure you know where runaway came from
Model-Dependent Methods
- Fitting isochrones to data
- main-sequence turnoff
- Hayashi->ZAMS
- requires distance
- Choice: Fit in HR diagram or CMD?
- HR: can account for extinction, etc.
- CMD: Statistics are very badly correlated on HR diagram
- Physical uncertainties increase high-mass stellar lifetimes:
- Rotation
- Convective overshooting
- errors are ~30%
- low masses unaffected
- ZAMS NGC 6530
- model including uncertainties and binaries
- requires age 5.5 Myr
- statistical uncertainties often dominant
Models without convective overshooting fail to reproduce Lithium ages
- Pre-main-sequence isochrones
- best to use HR diagram for differential extinction
- need spectroscopy to get Teff
- luminosity spread: binary, variability, age spread, accretion
- luminosity spreads ubiquitous * are they real? * do they imply real age spread * is there one real age?
- ONC: Cannot explain age spread as observational uncertainty
- observed age spread 0.4 dex (factor of 3)
- uncertainties can't explain!
- HR diagram may be set by Accretion history instead of age of star
HR scatter shrinks after 10 Myr
models do not agree on age/teff relation
observations do not agree with models
- Bell: compare PMS and upper main sequence ages.
- for some models, there is agreement
- But, still uncertain: What if high mass stars form significantly after low-mass?
- Order is OK, but spacing is uncertain by factor ~2
Asteroseismology
Kepler can't see "p-modes" because of variability
delta-scuti instability strip... delta-scuti age more precise than HR fit
spectroscopic measurements of gravity can get rank-ordered measurements
- R-C gap:
- distance & extinction independent
- good to ~15 Myr
- Gyrochronology: good for old stars
- >100 Myr
- Slowing period for low-mass stars. Age-dependent "Slope"
- Disk-presence clock?
- 50% uncertainty
- dangerous until we understand what disperses disks and when?
- Lithium abundance:
- extremely strong dependence on efficiency of convection
- useful relative but not absolute age indicator
- M and K dwarves have very large EW[Li] change
- ~factor of 2 age indicator
Conclusions
- Lithium depletion boundary model-independent, precise (10%), accuracte (10%)
- Expensive in terms of telescope time
- Kinematics: currently problematic
- Runaways: Origin
- Model-dependent ages:
- Large statistical uncertainties, modest systematic (main seq)
- pre-main-sequence: absolute entirely unclear
R-C gap distance independent Empirical ages
Final thought: Deuterium depletion boundary should be a <10 Myr age indicator. Requires measuring deuteration in BD atmosphere
Questions
- Q: Gunther? from COROT. Deuterium driven: model independence stems from a concensus on the initial conditions. Concensus disappears if you consider colors instead of Hayashi track. What is the zero point?
- A: Zero point uncertain. Star forgets about initial (cloud) conditions after 20 Myr. Helps keep LDB secure.
- Q Mamacek: Takeaway - Rescaled ages. Half-life of T-Tauri is 3-4 Myr.
- Q: What observing proposal for D-burning boundary?
- A: Deuterated water in atmosphere? CrH->CrD? No idea whether it's even remotely feasible.
- Q Ian Crossfield: Gyrochronology useless for young stars. Why does it work for old stars?
- A: The models, and the data, converge
- Q: Self-consistency of ages for Beta-Pic (isochronal, LiDep, etc. agree). Too pessimistic to disregard kinematics. Need self-consistency between methods.
- A: self-consistency can lead you along the wrong path. Subjectivity in excluding objects.
- Q John Tobin: Reggiani work on Orion - subregion, or whole central region?
- A: Not sure, HST treasury...
- Q: Multiple bursts?
- A: Foreground population. Take out low-extinction stars, still looks the same. There is an age spread.
- Q Neal Evans: Ages of Class I, Class 0, all depend on half-life of 2 Myr for T-Tauri (Class II) stars. 2 Myr is the goal.
- A: 3 Myr is a more conventional half-life
- Q: Deuterium depletion. Deuterated methane line in the thermal IR. CRIRES maybe?
- A: Too hot for methane
- Q Krumholz: Age spread in Orion. Must have age spread at <1 Myr because crossing time is ~0.4 Myr. Even most efficient SF, things happen in a crossing time, not instantaneously.
- A: Note the log-scale. 10 Myr is pretty broad
- Q: Advertise poster....
- Q: 10 Myr is much longer than the crossing of Orion
- Krumholz: I disagree