-
-
Notifications
You must be signed in to change notification settings - Fork 405
Paper plan
Section 1: introduction
- why model supernovae
- need RT
- various considerations: accuracy vs speed
- atomic physics vs detailed explosions models (dimensionality)
- this paper: want to compare various approximations that are made and establish how good/bad some of the common simplifying assumptions are
- nebular approximation, LTE excitation assumption, choice of atomic data, treatment of non-resonance scattering
Section 2:
- methods
- introduce "radiation field model" - J = B [LTE], J = W B [nebular approximation], J = ? [NLTE]
- introduce the code
- keep brief and to the point (refer lots to Lucy papers)
- idea of a simple but flexible and modular approach - will be public and available to build on
Section 3
- atomic input
- source of atomic data (how it's arranged/stored?)
- processes treated
- treatment of ionization: LTE, nebular ionization (detailed to follow): this paper is about excitation
Section 4:
- simple code tests
- pure Si atmosphere - compared to Synow
- pure Fe?
- other tests - parameters similar to one of the abundance tomography cases?
- Lucy 99 parameters?
- Luminosity density is wrong
- iteration happens with scattering rather than with downbranch or macroatom.
- convergence test, iteration information, level population results blah blah
Section 5: results
-
5.1 reference calculation
-
5.2 treatment of non-resonance scattering - macro atom vs down brach vs pure scattering (mostly a duplication of the Lucy 99 discussion except we have the full macro atom to compare too)
-
5.3 full NLTE for Si, Ca, S? [combinations thereof?] versus, LTE excitation and versus NLTE with Jblues from rad. field model. [the above is giving us for 1 model, with 1 set of atomic data: a grid of 3 x 3 sets of assumptions: [macro atom, down branch, pure scattering] x [full NLTE, NLTE+from rad. field model, LTE] ]
-
do this for one model? for a couple? [e.g. W7? or pure detonation?]
-
atomic data test? - Chianti versus Kurucz? Do the test for a full 3 x 3.
-
do we want to do separate NLTE calculations for pure Si, Ca, etc.? Or just jump into the real models?
Section 6: conclusions
- EITHER: NLTE doesn't matter! Hurray! Don't need to worry about it in multi-D calculations.
- OR IT DOES MATTER: then should use this new code rather than previous approximate codes (minor hurray)
- [CAVEAT: what about NLTE in Fe?]