# Predicting melting points of substances based on their molecular/atomic structure

Wouldn't it be great to be able to predict the melting point of a metal based on the atomic number and not by looking in some table for the value that was measured experimentally by someone else? Or to predict the same melting point for a simple salt like NaCl or KI by applying some formula to the atomic numbers of the 2 atoms involved?

Can you help me understand why is it so difficult to build a theoretical model for this particular physical property? Were there attempts made in the scientific world for such a model?

--EDIT--

Yes, this is part of my question: I want to understand if there is a connection between the theoretical explanations of properties like melting point and the current theoretical model of atomic structure. Can the electronic structure of $\ce{Fe 1s^2 2s^2 2p^6 3s^2 3p^6 3d^6 4s^2}$ tell us anything about it's probable melting point without measuring that point in the first place?

With regard to organic compounds, you might be interested in group contribution methods, e.g. the Joback method.

These methods divide a molecule into functional groups or similar small building blocks and estimate the melting point (and other thermodynamic properties) based on the contributions of the individual groups.

• do you know how this model was built? Was it statistically inferred from a smaller database and and then verified on a bigger data set? Or did they propose some theoretical model (of quantum level interactions, let's say) and then calculated the individual group contributions and built the formula? May 12, 2015 at 8:29
• @Liviu-AurelianRau-Neacsu These are statistical methods. It is way more easy to do that way than create universal theoretical models from ab initio or dft theories.
– Greg
Dec 28, 2017 at 17:52

Melting point depends not only on the structure of the molecule itself, but also on the structure of the solid (crystalline) phase and molten/liquid state of the substance. That is a much bigger computational problem than calculating energetics for a small molecule or ensemble. It is not easy to predict how a molecule (especially a complex one) will crystallize, and even if you do know that, a big computation to calculate the energy of the crystalline state.

• I suspected that much, but can you help me understand why? "It's difficult" does not enlighten me one bit. What does "calculating energetics for a small molecule or ensemble" mean? May 12, 2015 at 8:14
• One way to think about melting point is in terms of the energy difference between the solid and liquid phases of a substance. There are a range of ways to calculate the energy of a chemical system, but for a bulk material, even the simplest of these requires an energy calculation for each of a long series of geometrical configurations of the atoms and molecules in your system. The complexity of the computation increases, the bigger the molecule and the more of them that are present in the system you are calculating the energy of. You might want to check out "molecular mechanics" on Wikipedia. May 12, 2015 at 14:07
• Thanks. "Molecular mechanics" was what I needed to know :) . May 25, 2015 at 9:04