Diffraction is purely a function of photosite size, not count.
That says that a larger sensor, with the same pixel count, will result in a smaller aperture yielding the same diffraction performance characteristics. In other words, 36mpx FX will be superior to 36 mpx DX, in terms of "diffraction limited" resolution at the smallest possible aperture. And Medium Format would be better still.
But larger sensors require smaller apertures to maintain the same depth of field so in the end it can be a net push . That is the conundrum of ever increasing sensor density and sensor size for macro and landscape photographers, and even wildlife photographers shooting TCs.
Assuming you are asking about shooting the 1:2 format vs full sensor, 1:2 will not improve things because the photosite size is unchanged. You are probably better off shooting full frame and getting a bit closer or using a longer focal length to similarly fill the frame. The only argument for 1:2 here is that in principle you can shoot a slightly smaller aperture and get the same depth of field, nullifying the difference. But most people would probably not use that much precision in selecting an aperture for a given shot, and the computations are a bit fuzzier than most people consider (keep reading).
The issue is not if diffraction will destroy the image from denser/higher resolution sensors. The issue is simply that the additional cost of those sensors results in very diminishing (or no) returns. That cost is not only the camera but all the storage, and the more limited shooting frame rate if that is important.
Just as an example, at f/22 or F/32 a D800 likely results in little additional resolution over a D700. It might be a mistake to upgrade a D700 to a D800 just to go shoot images at f/22 or smaller.
Due to the lower true (color) resolution of the Bayer filter, and some other issues, there may be more real world benefits to shooting A D800 at small apertures, and even at F/22, than the simple diffraction formulas suggest. I think it is a complex issue that is not well modeled by standard diffraction formulas.
The incremental cost of higher resolution and sensor densities (and digital storage) is declining. At this point in time a D800 might make sense over a D600 simply for the added feature set- those features may make it worth the ~$1000 msrp price difference. I think that is an important consideration in the context of the poor modeling of the diffraction calculations. Same cab be said for the decision to shoot 1:2 to try to solve any of these problems.