add cooling documentation

[BenWibking]

This adds a Markdown file with documentation on the convention for CIE cooling curves that is expected by AthenaPK, as well as documenting the approximations used in calculating the mean molecular weight.

A script to plot two example cooling curves in the form expected by AthenaPK is provided.

[forrestglines]

This might belong in another PR, but we could make some functional changes to AthenaPK to reduce errors and bring cooling and treatment of temperatures more inline with the code:

• Use only 2 columns in the cooling table -- first is log10 temperature, second is log10 $\Lambda_{hd}$. This would avoid human errors in choice of metallicity.
• Use code units in the cooling table. This would force users to consider the generation of these tables. However, it would place extra work on the user and would require regeneration of cooling tables if simulation units changed.

Another larger change that I don't think belongs in this PR would be to address temperature units:

• Introduce a Temperature unit. Right now all other units are user-defined but all temperatures are fixed to Kelvin. This would take the form of a user defined value of $k_B$
• Or Use $k_B T$ instead of $T$ everywhere, since $k_B T$ already exists in our system of units as a unit of energy. This would affect parameters for cooling and potentially the temperature column in cooling tables.

[BenWibking]

This might belong in another PR, but we could make some functional changes to AthenaPK to reduce errors and bring cooling and treatment of temperatures more inline with the code:

• Use only 2 columns in the cooling table -- first is log10 temperature, second is log10 Λhd. This would avoid human errors in choice of metallicity.

I would be in favor of this change.

I do not agree with the other proposed changes. In particular, I think Kelvins are the only physical temperature unit that anyone would ever use. I think requiring input in code units would make human error more likely.

[forrestglines]

I do not agree with the other proposed changes. In particular, I think Kelvins are the only physical temperature unit that anyone would ever use. I think requiring input in code units would make human error more likely.

That's very fair, my programmer brain for consistency is usually wrong. There goes my dream of a code using Rankine though

[forrestglines]

If we include the updated Schure Cooling table (or tables for different metallicities) I think we should include this notebook in some form, perhaps reduced down to a small script without the plots: schure_cooling.ipynb.txt

[BenWibking]

I think a notebook with plots included would be good to include in the docs. Or maybe a docs/notebooks subfolder?