Inconsistent presentation of thermophysical properties

[ccokerlanl]

Description

The mendeleev Element object returns the melting and boiling temperatures, if they exist, at what appears to be the minimum pressure at which that phase transition may occur. As a result, the boiling point for carbon returned is less than the melting point for carbon. It is not clear to the user that this is the convention.

Current Behavior

Minimum Reproducible Example

Script

#!/usr/bin/env python
import mendeleev

# create element
carbon = mendeleev.element('C')

# print melting and boiling points
print(carbon.melting_point)
print(carbon.boiling_point)

Output

>>>{'graphite': 4762.15, 'diamond': 4713.15}
>>>{'graphite': 4098.15, 'diamond': None}

Discussion

Data from this Mendeleev carbon object present a boiling temperature (4762 K) for graphite that is lower than the its melting temperature for graphite (4098 K). According to the carbon phase diagram, Fig. 3 in [1], this melting temperature corresponds to the triple point (~10 GPa or ~100,000 atm), whereas the boiling temperature from this Mendeleev object appears to correspond to standard pressure (1 atm).

[1] Wang, X., Scandolo, S., & Car, R. (2005). Carbon Phase Diagram from Ab Initio Molecular Dynamics. Phys. Rev. Lett., 95, 185701.

Expected behavior

If the entire pressure-temperature phase relationship is not implemented, thermophysical properties (e.g. melt temperature, latent heat of fusion) are expected to be presented for standard pressure (1 atm) unless otherwise specified. If such a phase transition does not exist at standard pressure (e.g. carbon melt point), the thermophysical property should return None (or perhaps NaN or Inf if that is more useful).

Compute Environment Specification

• mendleev version 0.17
• python version 3.12.7
• RHEL

[ccokerlanl]

Edit: Fixed a typo.

[ccokerlanl]

Edit: added a discussion and reference to the example.

[ccokerlanl]

Edit: fixed markdown heading levels.

[lmmentel]

Thanks, I'll have a look.

[lmmentel]

I've verified that indeed the melting and boiling point might be misleading. The values are taken straight from the PhaseTransition object but the crucial context is missing. It'll be safer to return None in those cases and leave users to dig through PhaseTransition data for further details.

[lmmentel]

Have a look at the PR #214 for the fix. I ended up dropping data for diamond since there was nothing available at the source at std conditions. There is a small example illustrating the changes, appreciate any feedback if that works.