Thread by @ReadDark

There is a growing consensus in the astro simulation community that, once we begin to resolve the interstellar medium (ISM), gas cooling and heating causes gravitational potential fluctuations that kinematically "heat up" dark matter [a thread 1/N]

If we can get the gravitational potential to fluctuate "significantly" on a timescale comparable to the local dynamical time, then dark matter (DM) is "heated up":

adsabs.harvard.edu/abs/1996MNRAS.…
adsabs.harvard.edu/abs/2005MNRAS.…
adsabs.harvard.edu/abs/2012MNRAS.…
adsabs.harvard.edu/abs/2014Natur.…

[2/N]

A "significant" fluctuation here can actually be quite small. Fluctuating just ~25% of the total enclosed mass can be sufficient if repeated enough times:

adsabs.harvard.edu/abs/2016MNRAS.…

DM heating can occur even when stars and gas do not dominate the enclosed mass at any time.

[3/N]

If fluctuations are much faster than the local dynamical time, then DM heating is suppressed (consider infinitely fas gas flows!); if much slower, then the system approaches the adiabatic regime and, again, DM heating is suppressed. See e.g.:
online.kitp.ucsb.edu/online/cdm18/r…

[4/N]

Such gas flows occur due to gas cooling and then heating up again due to feedback from stellar winds and supernovae. To begin to resolve these physics, we must begin to resolve the multiphase ISM.

[5/N]

This means:

Spatial res: ~10-100pc
Mass res: ~100-1000Msun
Gas temp: ~<100K
Gas density: ~>100 atoms/cc

[At the lower end of these ranges, we have to worry about supernovae "overcooling", e.g:

adsabs.harvard.edu/abs/2008MNRAS.…
adsabs.harvard.edu/abs/2006MNRAS.…
adsabs.harvard.edu/abs/2016MNRAS.…]

[6/N]

If your simulation doesn't reach the above resolution, you won't see "DM heating".

So, if you have a low density threshold for star formation (e.g. 0.1 atoms/cc), you won't resolve the density of the ISM => no DM heating:

adsabs.harvard.edu/abs/2012MNRAS.…
adsabs.harvard.edu/abs/2018arXiv1…

[7/N]

Similarly, if you don't allow gas to cool below 8000K, you won't resolve the ISM and DM heating will be suppressed:

adsabs.harvard.edu/abs/2018arXiv1…

[8/N]

Note that raising the density threshold for SF *without* allowing gas to cool is an odd thing to do. To reach high density, the gas then has to clump together in very massive and unphysical clouds. These can move around => DM heating, but it cannot be physically correct.

[9/N]

Finally, it is important to emphasise that:

gas flows => DM heating AND, as an observational consequence, "bursty" star formation (SF):

adsabs.harvard.edu/abs/2013MNRAS.…

Bursty star formation *does not* => DM heating

[10/N]

There is mounting observational evidence for bursty SF in dwarf galaxies:

adsabs.harvard.edu/abs/2014MNRAS.…

This is great because such bursty SF was *predicted* by DM heating models. But it is *not* a "smoking gun" for DM heating, because bursty SF could happen without gas flows.

[11/N]

We have recently found "smoking gun" evidence for DM heating: an anti-correlation between the inner DM density of dwarf galaxies and the duration of their star formation:

adsabs.harvard.edu/abs/2018arXiv1…

This was also predicted by DM heating models:

adsabs.harvard.edu/abs/2014MNRAS.…

[12/N]

There could be other reasons for the above anti-correlation, or DM heating may proceed via a different dynamical mechanism, e.g.:

adsabs.harvard.edu/abs/2001ApJ...…

But it is great that simulations from different groups, with different codes/subgrid models, now agree on the above!

[End]