Why is BJT scaling not as significant as MOSFET scaling?

Question

When reading about integrated circuits, MOSFET scaling comes up very often, however BJT scaling is never discusses. What's the reason for this?

Answer 1

This relates to your other question, so this may seem a little circular.

In the circuits that require the highest device density, MOSFETs are preferred for a number of reasons, mainly for their electrical characteristics such as minimal static power dissipation. But it turns out that MOSFETs are also easier to pack more tightly, because their physical structure is simpler than that of BJTs.

This is epitomized in NAND flash, in which you have basically a long stripe of diffusion that provides the "channel" for a large number of series-connected transistors. The only limitation on transistor density is how close together you can draw the patterns for the gate structures.