Research Interests

My research is primarily focussed on the application of stellar evolutionary models to low-mass stars, young clusters and star-forming regions (SFRs). My scientific interests can be broadly summed up as:

  • The origin of multiple populations in old globular clusters.
  • Robust age determinations for young clusters.
  • The effect of environment on circumstellar discs.
  • Stellar evolutionary models.
  • The initial mass function.
  • Data reduction/photometric techniques

Environmental effects on disc lifetimes

Circumstellar discs are the nurseries of planet-formation. It is thought that the lifetimes of these discs around pre-main-sequence stars will be shortened by erosion from the UV radiation and winds of nearby massive stars and by dynamical encounters with other stars. We may therefore expect that circumstellar discs survive longer in low-mass, low-stellar-density regions.
Understanding whether discs survive longer in low-density environments is crucial to our understanding of planet formation, and determining whether environment affects disc evolution necessitates accounting for the effect of age. We therefore need robust age determinations for a number of clusters of different environments.

Age determinations

The age of a star is a fundamental parameter and is key to deriving e.g. young-stellar-object lifetimes or disc dissipation timescales. Age is also hugely important for mass determinations which, in the absence of dynamical masses, are usually obtained by comparing observed stellar/planetary luminosities to (highly age-sensitive) theoretical models.
We can determine the age of a cluster by fitting the observed sequence in a CMD with theoretical model isochrones. The models themselves are constructed by taking sets of model atmospheres, reddening them as required and folding them through the photometric system responses to construct bolometric corrections, which are then applied to stellar interior models. This requires a good understanding of our observational system.
For young clusters the pre-main-sequence (PMS) is much more populated than the upper-main-sequence, and thus PMS ages have smaller statistical uncertainties. However at low temperatures (below ~4500K) there is a well known problem in which the models systematically overestimate the flux in the optical, leading to an error in the ages derived from the PMS of roughly a factor two.
Further errors can be introduced by the use of photometric transformations, and when dealing with extinction. If photometric transformations derived from main-sequence stars are applied to pre-main-sequence stars (to transform from the instrumental system to some standard system for example), you can introduce a further error on the derived flux again leading to an error on derived ages of factor two.
For extinction, we need to consider the effect of temperature. The flux distribution of a star shifts the effective wavelength of a filter. We see the same effect in reddening. i.e. a given amount of dust obscuration will produce a larger E(B-V) value for OB stars than for GK stars. Equivalently for a nominal E(B-V) the resultant extinction in a given filter will vary with Temperature. This effect is smaller than the previously discussed issues, accounting for roughly a 50% error in derived ages, but still important.
Figure 1
Figure 1 - Photometric transformation between the SDSS and INT-WFC systems in g (top) and g-i (bottom). Note the non-linearity, and the difference in transformation for stars aged 1 Myr, 10 Myr and those on the Zero-Age Main-Sequence (ZAMS). Source: Bell et al (2012)

We overcome these issues by adopting a careful approch to our data reduction, working entirely in the instrumental plane rather than a standard system (therefore avoiding the use of photometric transformations), and using semi-empirical model isochrones to account for the over-estimation of the flux (see below for further details).

My research so far has focussed heavily on Taurus, one of the nearest star-forming regions. Taurus is very low-density, and an ideal target for studying the effect of environment on star formation. I have a paper close to submission in which I have obtained robust evidence that discs survive longer in the low-density environment of Taurus.
I am currently working with data from the Dark Energy Camera (DECam) on the Blanco-4m telescope to extend my extinction and age-fitting work to other low-density star-forming regions, to test whether we see an abundance of discs in these low-density regions.

Extinction fitting

We have developed a method of extinction fitting using iZJH photometry. In an i-Z, J-H colour-colour diagram the reddening vectors are almost perpendicular to the majority of the sequence (see Figure 2) and so we can deredden each star individually. This filter choice also minimises the effect of discs.
We adopt a Bayesian method, which allows us to account for uncertainties due to age and binarity by marginalising over these parameters. Our derived extinctions show good agreement with other literature values, and importantly we are able to deredden our entire sample in a consistent manner.
Figure 1
Figure 2 - i-Z, J-H colour-colour diagram for Taurus members. Red asterisks are Class II Taurus members, blue open circles are Class III members. The black and green line are a 2Myr and 4Myr tuned BCAH98 isochrone respectively, showing the insensitivity to age along the majority of the sequence. The black dashed lines are the reddening vectors obtained from the Fitzpatrick (1999) reddening law, showing that stars will deredden to a single position on the isochrone.

Stellar Models

We correct for the additional flux in the stellar models by comparison to a fiducial cluster with a well established distance and age determined from methods independent of the PMS. We then introduce an empirical correction at low Teff. This method has the benefit of maintaining a consistent mass scale and the gravity dependence from the underlying theoretical models. Using these semi-empirical models brings the ages derived from the PMS into agreement with those derived from the upper-main-sequence fitting, as well as those derived using other techniques (e.g. the lithium depletion boundary technique). Further details are available in Bell et al. (2014). The models we use for age determinations are also available from our isochrone server, located here. Some information on the models are given on the website, with more detail given in the paper.