Re: Newtonian simulations of dark matter only sphere

From: ABINASH KUMAR SHAW <abinashkumarshaw_at_gmail.com>
Date: Thu, 10 Sep 2020 22:28:12 +0530

Hello Sten,

I have been using a sphere of dark matter particles with NFW radial density
profile. We expect that, due to self-gravity, the particle should collapse
to the halo centre with time unless it crosses the centre and goes to the
other side. There is also an issue with the softening length. For some
small values of softening length we have seen that the dark matter halo
oscillates in size around its centre. Regarding the radius, I have used a
core radius of length 4 kpc and the cutoff radius to be 40-60 kpc.
There are 1 million particles in my simulation.

Abinash

On Thu, Sep 10, 2020 at 1:30 AM Sten Delos <delos_at_unc.edu> wrote:

> While I'm not quite clear on the initial conditions here, I would propose
> that the cuspless outcome may be real. Cusps appear in simulations of
> overdense regions surrounded by a dark matter background. I've found that
> if I surround the region with vacuum instead, I obtain a much less
> pronounced cusp.
>
> In other words, if you're just simulating a uniformly dense sphere, try
> surrounding it with a less uniformly dense background. You can still have
> vacuum boundary conditions at some radius, of course---that radius just has
> to be large enough to not be collapsing at the same time as the halo is
> first forming.
>
> - Sten Delos
>
>
> On Wed, Sep 9, 2020, 2:38 PM Cameron McBride <cameron.mcbride_at_gmail.com>
> wrote:
>
>> My guess here is that you need to adapt your gravitational softening in
>> Gadget-2. I've been out of actively following the literature for the last
>> ~5ish years, but there was a host of papers that talk about cusp - core
>> connection to gravitational softening that can change the cusp vs core.
>> It's a little bit of a numerical quagmire, as you can tune softening to act
>> differently (and there are tradeoffs even if you force a strict N^2
>> calculation, which Gadget-2 typically doesn't). I also suggest you don't
>> use grid, but tree only configurations.
>>
>> I think Zavala et al 2019 or Kuhlen et al 2012 might discuss aspects of
>> this. There are some older articles, like Boylan-Kolchin et al 2005 that
>> might also discuss some of it.
>>
>> Others might have more insight, but hopefully that might help you start
>> digging deeper.
>>
>> Cameron
>>
>> On Wed, Sep 9, 2020 at 8:32 AM ABINASH KUMAR SHAW <
>> abinashkumarshaw_at_gmail.com> wrote:
>>
>>> Hello,
>>>
>>> I am trying to study time evolution of the density profile of a single
>>> dark matter only halo in Newtonian space (without boundary) using Gadget-2.
>>> My initial condition file contains a spherical distribution of only the
>>> cold dark matter particle which is written in gadget format. I am following
>>> the example of the gassphere (provided with the code) to evolve the CDM
>>> halo. I expect the density profile to show the cusp at the halo centre
>>> whereas I am getting a core. I understand that a core forms when there are
>>> baryons in addition to the CDM. There are no baryons here, however I expect
>>> that somehow baryonic interaction is still taking place. So, do I need to
>>> modify the code of my simulations; or am I doing something wrong?
>>>
>>> Thank you in advance.
>>>
>>> Sincerely,
>>> Abinash Kumar Shaw
>>>
>>> *Senior Research Fellow*
>>> *Department of Physics*
>>> *Indian Institute of Technology Kharagpur*
>>> *Kharagpur, India -- 721302.*
>>>
>>
Received on 2020-09-10 18:58:56