PI:  Karl Gordon, kgordon@stsci.edu, webpage
Group Website: ISM*@ST
Project Duration: Summer/1 year rotation with potential for thesis project

Project Abstract: I have a number of possible projects that center around the subject of dust radiative transfer.   These calculations would use the DIRTY 3D dust radiative transfer model (Gordon et al. 2001, Misselt et al. 2001, Law et al. 2018, https://github.com/karllark/DIRTY_dustrt).  The projects are:

  1. Importance of Scattering in Extinction Curves: Simple full RT modeling of slab in front of a star.  Quantify the importance of scattering into the beam for extinction curves.  This is important for understanding to what distance can accurate dust extinction be measured using stars (Local Group? 2 Mpc?  Further?).  How does this depend on the telescope resolution (diameter, JWST?  Roman?  LUVOIR?)?   Do this as a function of distance given constant apparent resolution of the star (e.g., how far out can we measure dust extinction curves, Local Group?  Beyond?).  What about distance between star and slab (e.g., external galaxies)?
  2. Dust near Supernovae: Often measurements of dust extinction towards supernovae (SN) are seen to have quite low R(V) values (below 2 - the minimum measured in the MW).  Is this due to the impact of dust nearby SN that results in significant dust scattering into the beam?  Dirty RT models with circumstellar (or just nearby) would allow this to be tested quantitatively.  This project is related to the previous, but is different as the S/N dust is likely much nearer than that for a star one would use for ISM extinction measurements.   Could be expanded to include calculation and analysis of infrared emission from SN.
  3. DirtyGrid Expansion: The DirtyGrid (Law et al. 2018) was computed using a specific stellar population synthesis (SPS) code for the stellar SED (PEGASE.2).  DirtyGrid used approximately 5 million CPU hours (approx 10 hours per grid point).  There are a number of other SPS codes.  Is it possible to use the DirtyGrid with other SPS codes?  The dust RT really cares about the overall stellar SED shape, not the fine details.  So, as long as the SED shape for a different SPS model matches one of the stellar SED shapes used in the DirtyGrid, then the dust attenuation and IR emission can be layered on top of the other SPS model SED.   This can/should be tested by re-running some DirtyGrid points using the different SPS SED and confirming that the layering idea works.
  4. DirtyGrid in SED fitters: Can the DirtyGrid (Law et al. 2018) be used in popular SED fitters?  CIGALE?  MAGPHYS?  One of the other *many* existing SED fitters?  Or is a separate SED fitter needed?  Having the DirtyGrid in a SED fitter usable by the community would provide higher fidelity (more physics) models for investigating dust in galaxies.
  5. Adaptive Dust RT Grids: Larger project.  Investigate using adaptive grids in dust RT to improve the fidelity of the calculations - optimize the size of the cells in a hierarchical grid.  Use on-the-fly estimates of the radiation field density as the prime criteria (other criteria have been tried w/o much success).  Use the slab case to start with (extensively tested to find the converged solution in benchmark paper - Gordon et al. 2017).  Likely requires an initial start based on the dust density (optical depth).  1st on-the-fly estimate should be done by absorption from stellar photons alone.  Then update including scattered photons.  Mean field first, then wavelength dependent?
  6. Circumstellar Dust: What is the impact of non-homogenous dust distributions (3D clumpy) around stars?  This is motivated by AGB stars and the measurement of dust masses from their winds.  But it could also be related to the SN dust issue/project (see above).  And probably other circumstellar environments (YSOs, WR stars, ...).  Global geometries (raidial, disk, envelope).  Local geometries (simple clumps, hierarchical clumps).  What are the best literature models?  How important is dust scattering?  Clumps?  Type of dust grains?   The goal here would be to mainly focus on the theory side - like Witt & Gordon (1996).

Student work:
The work would vary by project with the main focus on running DIRTY models with different parameters, analyzing the results, and writing a paper.  All the projects should result in a student lead refereed paper and presentation at a meeting.  Most can be expanded to a larger (thesis) project included analysis of observations, answering multiple science questions, and writing multiple papers.


DirtyGrid SEDs (from Law et al. 2018, Fig. 9), projects 3 & 4

 


TRUST slab benchmark geometry (from Gordon et al. 2017, Fig. 1), projects 1 & 5

  • No labels