RealQM Configuration of Neon as 2+4+4

A basic pillar of moden chemistry (supposedly with some support from standard Quantum Mechanics) is the octet rule, which says in particular that the electrons of the Neon Ne atom are arranged in two spherical shells with an inner shell around the +10 charge kernel formed by 2 electrons and an outer (valence) shell with 8 electrons, with the message that the valence shell is "full" and so Neon does not want to interact with other atoms through a covalent or ionic bond.

On the other hand, Flourine F with +9 charge kernel with 7 electrons in the outer valence shell is very keen to attach another electron to "complete the octet" and so readily forms eg an HF molecule with an H atom in a covalent/ionic bond. 

RealQM gives a different picture with Ne in 2+4+4 configuration and F as 2+4+3 with a third shell harboring 4 or 3 electrons acting as the valence shell. This means that that F combines with H to form a HF molecule, while Ne like He does not. You can interact with simulations in p5js-code here:

• Fluorine
• Neon
• HF    

For Neon in a 2+8 configuration RealQM gives a much too small energy compared to observation, which thus appears to not be attained since it requires somehow squeezing 8 electrons into one shell.

The reason H can bond to F in a 2+4+3 configuration, rather than in 2+7, appears to be that the size of an H atom better fits with the size of electrons in a valence shell of 3 electrons instead of 7.

The from observations estimated radius of an F atom is about 1 atomic unit au (50 pm). The thickness of atomic shells scales with 2/Z with Z the reduced charge reaching a shell, with thus 2/10 + 2/8 + 2/4 about 1 au for a 3-shell 2+4+4 configuration, while a 2-shell 2+8 configuration would give a too small radius compared to observation.

What can the difference be between C as 2+4, which bonds with H and O in particular, and Ne as 2+4+4, which does not want to bond with anything, when they have the same 4 electron valence  shell? It is thinkable that this is a geometric effect with the C valence shell being more compatible with those of H and O, which may be uncovered by further explorations with RealQM including the validity of the octet rule.