Towards Big Molecules with RealQM

Here is an application of the reduced models of the previous post to linear 3-atom molecules with 1 or 4 valence electrons showing good agreement to experimental data:

• H2O (or HOH)     (code)  (O 2 valence)
• CH2 (or HCH)     (code)   (C 2 valence)
• CO2 (or OCO)     (code)
• BeH2 (or HBeH)  (code)  (Be 2 valence)
• LiOH                    (code)  (alt code)
• NaOH                   (code)  (alt code)
• N2                         (code)  (N 3 valence)
• NH3                      (code)  
• CO                        (code) (alt code) (C 4 valence)
• H2CO                   (code)  (C 4 valence)
• OH                        (code)  (O 2 valence)

It thus seems possible to represent an atom in RealQM in a reduced model defined by an effective radius R and number of valence electrons by fitting to experimental atomisation energy data, and then build molecule models from such reduced atomic models as in the above examples. 

The computational complexity for a system with $N$ kernels will then scale with at most $N^2$ like a particle system in classical mechanics with all particles connected (or even $N$ with only local connections). RealQM modelling of big molecules thus seems to be possible with a window to ab initio protein folding.  

  

It may be that the number of valence electrons effectively is at most 4, and not up to 8 (as in the octet rule of stdQM) with O having 6,  and so that the above examples in fact covers a wide range of molecules.   

Also recall Real Atom Simulator allowing you to explore atoms in spherical symmetry.