I noticed some halos have extreme values of potential energies that can cause virial ratios with extreme negative or positive values. For example, when calculating the potential as a sum of the individual particle potentials provided by TNG in TNG100-Dark, halo 636 has VR = 22233 and halo 262 has VR = -119. I've read other discussion posts that mention this ("Gravitational potential in halos without the contribution from particles external to the halo" and "DM Subhalo Potential") and recognize that the potential is a global quantity and therefore not representative of the local halo potential, however, the extent and behavior of these 'extreme' potentials is a bit puzzling to me.
I have calculated the energy distributions of many halos from TNG50-Dark and TNG100-Dark and have characterized them by a "shift" which I define as the difference between the maximum energy within R200 and 0 (the idealized energy expected for the least bound particles of isolated halos). This is, hopefully, made clearer in the figure below where I have an example of the energy distribution for one equilibrium halo from TNG50-Dark along with its shift. Additionally, I define E0 as the extent of the energy distribution, or the difference between the minimum energy and maximum energy within R200.
Upon investigating this for many halos, I found that the range of these shifts are greater in TNG100 than TNG50. I also found that the range of this shift (normalized by E0) aligns better for halos of similar number of particles than of similar masses between the simulations. Both of these are illustrated by the 4-panel plot below.
I am wondering if the values of extreme potentials in TNG are due solely to global potential effects, or if it is possible some numerical effects could be impacting these "shifts".
Dylan Nelson
20 Aug
TNG100 will sample different regions, i.e. rarer more extreme overdensities.
I would also encourage you to re-compute, from scratch, the potential (e.g. on a single particle), using whatever convention, tools, etc. One option is to use the AREPO public code and compute the potential in post-processing. You could review the relevant code to determine if the calculation is the one you are after. You could also use any other entirely different code/method, in order to verify correctness.
Hello,
I noticed some halos have extreme values of potential energies that can cause virial ratios with extreme negative or positive values. For example, when calculating the potential as a sum of the individual particle potentials provided by TNG in TNG100-Dark, halo 636 has VR = 22233 and halo 262 has VR = -119. I've read other discussion posts that mention this ("Gravitational potential in halos without the contribution from particles external to the halo" and "DM Subhalo Potential") and recognize that the potential is a global quantity and therefore not representative of the local halo potential, however, the extent and behavior of these 'extreme' potentials is a bit puzzling to me.
I have calculated the energy distributions of many halos from TNG50-Dark and TNG100-Dark and have characterized them by a "shift" which I define as the difference between the maximum energy within R200 and 0 (the idealized energy expected for the least bound particles of isolated halos). This is, hopefully, made clearer in the figure below where I have an example of the energy distribution for one equilibrium halo from TNG50-Dark along with its shift. Additionally, I define E0 as the extent of the energy distribution, or the difference between the minimum energy and maximum energy within R200.
Upon investigating this for many halos, I found that the range of these shifts are greater in TNG100 than TNG50. I also found that the range of this shift (normalized by E0) aligns better for halos of similar number of particles than of similar masses between the simulations. Both of these are illustrated by the 4-panel plot below.
I am wondering if the values of extreme potentials in TNG are due solely to global potential effects, or if it is possible some numerical effects could be impacting these "shifts".
TNG100 will sample different regions, i.e. rarer more extreme overdensities.
I would also encourage you to re-compute, from scratch, the potential (e.g. on a single particle), using whatever convention, tools, etc. One option is to use the AREPO public code and compute the potential in post-processing. You could review the relevant code to determine if the calculation is the one you are after. You could also use any other entirely different code/method, in order to verify correctness.