Resistion I think it is always possible to affix any or value definition you wish to hp (usually defined as half distance between minimum features). The self assembly (S-A) process) must be considered as independent of lithography. The distance between the S-A oxide discs will be defined by the chemical-oxidation process and added as a sub-lithographic overlay onto the best conventional lithographic process; in my example 20nm lithography. The S K Hynix 1Gb PCM that I covered in one of my earlier EETimes PCM papers was an example of a 42nm minimum feature size resulting in a 2F^2 cell with an hp of 42nm.
Again, in Part 2 above my “d” was the diameter of the oxide discs and that has nothing to do with lithography.
Maybe you can say hp is 20 nm and the cell is 4*hp^2, but if hp is 4x diameter, the cell is 4*16*d^2 or 64d^2. Some schools use hp as F some use d as F. Although hp is more in tune with lithographic requirements, d is more in tune with actual process requirements.
Resistion- Perhaps you can share with us the basis of your conclusion. In my figure 9 above, consider the example of 5nm oxide discs (your nano particles) and a fill factor of 50% for a 20 x 20nm (I.e F =20nm) aperture just over 10 discs are required. Even if you square-packed the discs in the same 20 x 20nm aperture so they were all touching each other a simple calculation shows 16 discs would be required.
The nano particles (discs) are sub-lithographic features so the number of them and the diameter should be independent of the lithographic feature size F. Assuming you are using F in the conventional manner as minimum lithographic feature size in my square aperture example the cell size will be dependent on the lithographic feature size as 4F^2.
The main questions that have to be answered for self-assembly (S-A) is can the discs be produced with diameters of less than 5nm and more importantly will the claimed low reset current density be observed.