Stellar Enrichment and Feedback¶
CosmicGrain couples dust creation to a time-resolved stellar enrichment and feedback model. Dust yields cannot be interpreted independently of the gas ejecta, element reservoirs, and thermal-energy deposition that create the local conditions in which grains form and subsequently evolve.
Delayed enrichment¶
Stellar particles retain their birth mass and age in an expanding universe. Feedback is therefore delayed rather than deposited entirely when the stellar particle forms.
- SNII and hypernova ejecta are released in eight time tranches spanning approximately 3–40 Myr.
- AGB enrichment follows the delayed mass return from the stellar population and uses IMF-integrated MESA yield tables.
- The LRN dust channel is injected once with the first SNII tranche rather than repeated in all eight tranches.
- SNII dust sampling is distributed across the full tranche sequence so that delayed feedback does not accidentally multiply the requested number of dust superparticles by eight.
StellarBirthMass is stored for star particles so the original SSP mass is
available for diagnostics and future normalization. Do not assume from the
field's presence alone that every yield path uses birth mass; the active
feedback implementation must be checked when changing normalization.
Element-resolved enrichment¶
The gas tracks C, N, O, Ne, Mg, Si, and Fe separately. SNII/hypernova yields depend on progenitor mass and metallicity, while AGB yields are integrated from the adopted stellar-yield grid. Total metallicity and the individual element reservoirs are updated together.
Dust formation is element limited. The code cannot condense more carbon or silicate material than is physically available in the relevant ejecta and gas reservoir. Gas-to-dust and dust-to-gas transfers operate on absolute tracked element masses and then resynchronize gas mass and metallicity fields.
Stochastic thermal feedback¶
SNII/hypernova energy is deposited stochastically rather than divided as a small temperature increment among every neighbor. The implementation:
- obtains an adaptive gas-neighbor kernel, requiring at least 16 receivers;
- computes the available event energy;
- evaluates the heating energy of each receiver using its own mass and mean molecular weight;
- stochastically selects receivers; and
- heats selected gas by a fixed \(\Delta T=3.0\times10^6\,\mathrm{K}\).
FeedbackFlag and EnergyReservoir support feedback-state tracking and
diagnostics. Debug-only temperature and internal-energy limiters should not be
silently enabled in production configurations.
Relationship to dust creation¶
The enrichment calculation first determines the ejecta and element budget. The dust module then condenses the allowed fraction into explicitly tagged PartType6 particles:
| Source | Current role |
|---|---|
| SNII/hypernova | Prompt, tranche-resolved dust and metal production |
| AGB | Delayed enrichment and carbon/silicate dust production |
| LRN | One-time early dust channel; normalization remains under calibration |
| SNIa | Delayed enrichment channel; not a current stellar dust source |
The dust source tag is retained after injection, allowing analysis of how much information about stellar origin survives later grain growth, erosion, coagulation, shattering, transport, and astration.
Validation expectations¶
After changing yields, feedback timing, dust sampling, or mass-transfer code, rerun:
- a short feedback timeline diagnostic;
- the global mass and element-conservation audit;
- dust-source count and mass summaries;
- the carbon-plus-silicate closure check; and
- a restart-continuity test spanning at least one feedback event.