May 27, 2024, 01:22:38 AM
Forum Rules: Read This Before Posting


Topic: Calculating Properties of Elements and Compounds Analytically  (Read 1688 times)

0 Members and 1 Guest are viewing this topic.

Offline belfner

  • Very New Member
  • *
  • Posts: 1
  • Mole Snacks: +0/-0
Calculating Properties of Elements and Compounds Analytically
« on: November 29, 2017, 03:55:17 PM »
I was wondering if there was a way to calculate properties of elements and compounds such as boiling point, vapor pressure, density with out actually measuring these properties?

Offline Enthalpy

  • Chemist
  • Sr. Member
  • *
  • Posts: 4040
  • Mole Snacks: +304/-59
Re: Calculating Properties of Elements and Compounds Analytically
« Reply #1 on: November 29, 2017, 05:48:42 PM »
Hi belfner and the others,

Some properties of some compounds can be estimated if accepting a significant inaccuracy, others are completely off. Methods exist, they are semi-empirical, most supposing that the properties (like the boiling temperature) are additive and result from subgroups contained in the molecule. The heat of formation is an exception, which is computed by software using a bit more fundamental methods.

----------

The boiling point of organic compounds is accurate to +-15K for non-vicious compounds. Already useful. The vapour pressure has the corresponding accuracy (15K makes a heavy difference).

The heat of formation is insufficient for my needs. What's worse, most software evaluates (badly) a gas at zero K, while I need liquids at RT.

The density is nonsense. For cage molecules, amines... evaluation software can tell "1500kg/m3" when the measure tells 1150.

The melting point isn't usable. It's commonly 50K off, including for just branched alkanes.

I have no opinion about the dipole moment, the octanol/water, and so on. For instance Chemsketch and Mpbpvp try to make such estimates too.

----------

As for the melting point, I suppose that a decent estimate should determine in detail the organization of the crystal, by trying all possible orientations, positions and if needed conformations of dozens of molecules and their interaction energy. This is (1) hard to program (2) heavy to run. Then, the software should also evaluate the rotation and vibration energies of the molecules in liquid and solid state, which isn't easy neither.

I've heard of two research teams trying to do this. Success would be useful, more teams may be needed, so if you are interested, do it! My suggestion is to study many examples to debug all wrong assertions or prejudices and get an idea of the difficulty, before programming anything.

Sponsored Links