Finding cluster members

From CoolWiki
Revision as of 19:56, 28 February 2017 by Rebull (talk | contribs)
Jump to navigationJump to search

This document is also known as "Luisa’s Table of Characteristics of Young Stars for Determining Cluster Members".

Introduction

Anatomy.gif

Whenever we study stellar clusters the question is: Which objects are the cluster members? This is easier with young clusters than old because the young stars are noticeably different than older stars, so it is easier to distinguish the young cluster members from the surrounding interloper stars (foreground and background populations). This process has a nice analogy with people too... when the IC 2118 teacher team came to visit the SSC, we all went out to lunch at a local Mexican place. If someone who didn't know any of us walked into the restaurant while we were eating lunch, as a group of astronomers, we are (for the most part! ;) ) not distinctly different than the rest of the adults in there, so we’d be difficult to pick out as a distinct ‘cluster’ of people, especially while we weren’t all physically co-located -- some of us were in line, getting salsa, and/or at the table. But, if a group from a day care center had been there, it would have been immediately clearly obvious that the children were a group that was different than the rest of the people in the restaurant. Moreover, the amount of time a human spends as a child is short compared to their entire lifetime, and so it is with stars. You have to seek out the group of young stars/humans in order to study their development.

Astronomers use as many of the following characteristics of young stars as possible to determine cluster membership, and we will do the same.

After reading this table, if you now go back and look at Maria Kun’s original IC2118 papers, see how many of these items she’s listing in making her case that she’s found young stars in IC 2118. I haven’t done this. Have I missed any in the list below?

Anatomy of a young star system (for reference) is to the right.

The Table

Characteristics Pros Cons
IR Excess

(IR is emitted by circumstellar matter)

  • Need Spitzer or WISE or Herschel (that is, if we didn’t already have the data, as it would be in the general case of cluster membership, not specifically in IC 2118)
  • Will only find those stars which still have enough disk left to make an IR excess – will be unable to distinguish young stars without disks (Class IIIs) from the interlopers.
  • Background galaxies (many of which are forming stars) can have the same IR colors as stars with disks, so need additional data to distinguish stars from galaxies.
(Flaring) X-rays

(young stars emit lots of X-rays because they are completely convective and fast-rotating, so they have lots of starspots and therefore lots of flares, big and small)

  • Need space-based mission to see X-rays - can't do from the ground.
  • Need a large field of view to efficiently study large parts of the sky at once; all missions (now, anyway) have small FOV (due to methodology for detection)
  • Takes a long time (like 25,000 seconds for one 5x5 arcminute field), and even then you will still be able to count the number of individual photons that you see.
  • Not all will be detectable on a reasonable timescale.
  • Stars might not be flaring at the time you look.
  • Will only find those stars that are X-ray active enough (might miss those that are deeply embedded or have big enough thick disks to block out the X-rays).
  • Background galaxies can also be bright in X-rays, as can active foreground M dwarfs.
(Flaring) Radio

(young stars emit in radio when they flare; see above entry for X-rays)

  • Need something that can detect radio (ground-based)
  • Can find all of the stars that are bright in radio pretty straightforwardly - you just look, and see the ones that are bright.
  • I don't know of very many people using this method as a primary method for finding young stars. I invite you to find the ADS references and link them in!
  • Field M stars can also be active, and thus just being bright in radio is not enough.
  • Spatial resolution of radio telescopes usually means either you have low-resolution over a large area (making it problematic to match to specific stars) or high-resolution over a small area (but we have a big map).
  • Background galaxies can also be bright in radio.
Outflows

(only present for the very youngest objects, Class Os and Is)

  • Again, need to cover large areas (outflows can extend over many parsecs).
  • Easily detectable in IRAC or optical emission line studies from the ground (search in ADS on “John Bally” to find lots such optical surveys)
  • Signpost to star formation – really big, obvious literal pointer saying “there is a very young star right HERE”
  • Real life examples of people using this method as a primary method for finding young stars: Walawender et al., “Multiple Outflows and Protostars near IC348 and the Flying Ghost Nebula,” 2006, AJ, 132, 467, Bally et al., “Irradiated and Bent Jets in the Orion Nebula,” 2006, AJ, 131, 473
  • Orientation might not be good – if it’s pointing right at us, we’ll miss it.
  • Not all stars have jets - only the very youngest, and stars don't spend much of their lives in that particular phase, so it's hard to catch them "in the act."
  • Sometimes hard to connect the maze of jets back to their source -- 2 main reasons: (a) central object often very embedded, and may be missed in optical and/or shallow surveys; (b) object precesses and moves, so jets twist and turn and don’t always point straight back to their source. In complicated regions (e.g., NGC 1333, see Spitzer image in press release archive), this is particularly tough.
Emission lines and other line shapes

(emitted/absorbed by accreting matter and technically disks too, though I wasn’t thinking of that at the time)

  • Photometry: Often easy to cover large areas with ground-based telescopes and a narrow-band filter such as Halpha or Neon II.
  • Spectroscopy: fast enough sequence of Halpha spectra can literally allow you to see blobs of matter as they fall into the star (!), which is pretty incontrovertible evidence you have a young star.
  • If you have a single spectroscopic observation of something with a P Cygni profile, this can also indicate accretion (emission line slightly redshifted from absorption line because matter is falling into the star).
  • Spectroscopy of the disk: need IR spectroscopy to see emission lines from molecules in disk
  • Real life examples of people using this method as a primary method for finding young stars: Ogura et al., “Halpha Emission Stars and Herbig-Haro Objects in the Vicinity of Bright-Rimmed Clouds,” 2002, AJ, 123, 2597, Edwards et al., “Probing T Tauri Accretion and Outflow with 1 Micron Spectroscopy,” 2006, ApJ, 646, 319 (ok, this is not blind searching, but it is really using line shapes to learn more about the stars in question.)
  • For a more precise measurement of Halpha, need to take spectra, which take longer to acquire than photometry.
  • The nebula itself can emit in Halpha (especially true in Orion Nebula, M41/42), so it can be hard to distinguish the young star emission from the nebular emission (photom or spec).
  • Older stars which are simply chromospherically active can emit in Halpha, so it can be hard to distinguish young stars from older stars on Halpha alone.
  • Spectroscopy of the disk – usually too expensive in terms of observing time to just go hunting blindly – usually need to have some reason to suspect a star is already young before embarking on such a project.
Variability

(because so much is happening in and around young stars, they are highly variable. In all cases here, I’m thinking of photometry, but as mentioned above, temporal studies using spectroscopy are also possible.)

  • Most frequently done in V, I, and/or J bands; variability in young stars has been seen in nearly all possible wavelengths
  • Can do from the ground, so can cover large areas of sky if you have a large FOV camera
  • With a large FOV, can do many stars at once.
  • Young stars highly variable, so relatively easy to do (need ~week or two rather than ~month or two of telescope time, and need only to go to 0.1 mag accuracy, not 0.001 mag accuracy, though that would help)
  • Can do relative photometry (photometry with respect to the other stars in the frame rather than with respect to photometric standards) so don’t really need calibrators, and you can keep observing if the night is strictly not photometric conditions.
  • Can be done (often best done) using small (<1 m) telescopes
  • Can look for periods at the same time (see below)
  • Real life examples of people using this method as a primary method for finding young stars: Carpenter et al., “Near-Infrared Photometric Variability of Stars toward the Orion A Molecular Cloud,” 2001, AJ, 121, 3160
  • Takes time, need many observations per night over many nights
  • Need to see photosphere (or close to it), so deeply embedded stars are harder to do, or at least harder to make the case to our colleagues that we’re not seeing variation in the nebula or outer disk
  • Need to do both short and long integrations to be able to get valid data on the bright and faint stars, respectively.
  • Older stars can vary too, but generally not at the rate or amplitude
Rotation rate

(a special case of ‘variability’ above)

  • Young stars rotate in general much faster than old stars, so fast rotation is also generally taken as evidence for youth.
  • Spectroscopy: only need one observation per star to get vsini.
  • Spectroscopy: high-res spectra can often also tell you if there is a nearby companion
  • Spectroscopy: high-res spectra can also tell you if the star still has lithum (Li burns so easily that only the youngest stars are thought to have any left)
  • Photometry: know the true value of the period (number is either really right, or wrong by a lot, as a result of observing method), no inclination (sin i) uncertainty
  • Photometry: Period is often something we know with more precision than anything else about these young stars.
  • Photometry: can use the same data you’re using for variability study above.
  • Real life examples of people using this method as a primary method for finding young stars: Rebull, “Rotation of Young Low-Mass Stars in the Orion Nebula Cluster Flanking Fields,” 2001, AJ, 121, 1676; Makidon et al., “Periodic Variability of Pre-Main Sequence Stars in the NGC 2264 OB Association,” 2004, AJ, 127, 2228
  • Spectroscopy: need high spectral resolution to get measurement of projected rotational velocity (v sin i)
  • Spectroscopy: can’t do anything about that inclination (sin i) uncertainty
  • Photometry: need many observations per night over many nights, and even then maybe only a fraction of your observed young stars will be detectably periodic.
  • Photometry; need stars to cooperate -- another observing campaign on the same stars a year later will only recover about half(!) of the periodic stars, presumably due to changes in the stars themselves (star spot shape and coverage, disk ‘puffiness’, etc)
  • Photometry: possible – though unlikely for fast rotation rates – to be fooled by binaries or disk occultations
UV

(due to shocks as accretion material hits star)

  • Long integration times needed because star faint at shorter wavelengths
  • Star needs to be accreting in order to be "brighter than you expect" at these wavelengths.
  • Subtle accretion rates look like coronal activity in older stars (similar to Halpha “cons” above
Spatial location

(localized in area of gas and dust)

  • Details of extinction not easy to measure
  • Chance superposition of foreground or background stars (and galaxies) can easily fool you, so usually you need at least one other indicator of youth before you can write a paper.
Similar brightness (similar age)

(can also think of this as placing them on a color-magnitude diagram [CMD] or HR diagram [HRD])

  • Need optical spectra to give us a spectral type (we had time to do this at Palomar for IC 2118) to help with placement in CMD/HRD (we need to get a handle on optical reddening, since reddening will make the stars appear fainter than they should, making it hard to see if they all have similar brightnesses)
Spatial motion

(Vradial = radial velocity, AND motion across the sky = proper motion, often abbreviated with the greek letter “mu”)

  • Takes a long time; have to wait for star to move (units of proper motion are commonly arcseconds per century). Old telescopes like Palomar or Yerkes are best for doing these kinds of studies because they have such a long baseline of observation. An ESA mission called Hipparcos was designed for determining proper motions of things all over the sky. They are planning a future mission called Gaia to do the same thing but for fainter stars.

Additional questions asked at the time

Can you have a disk without accretion? – yes, because the disk could just be sitting there, not actively dumping stuff onto the star; that’s how you get stars with an IR excess but no UV excess. (Cindy originally had: “yes, because you have Av extinction in the visible” .. the problem with that is that the Av could come from the general ISM, not just the circumstellar disk.

Can you have accretion without a disk? – seems awfully hard to imagine how this could happen, but we have a handful of stars that appear to be doing it. We don’t know what’s going on there. Since we are sensitive to DUSTY disks, maybe it is GAS that is still accreting onto the star.

Questions to think about and things to try

What happens if a star satisfies some but not all of the criteria above? What if it has only one of the properties? How many properties does it need to have before you could stand up and claim to have found a young star, and have no one argue with you about it?

Find a paper in the literature about finding young stars, not necessarily using Spitzer data, and see how many of these characteristics they use. Did we miss any?