One thing I noticed with a few eggcrate benthic zones, is that there was a greater amount of organisms in the initial (sediment settling) section. This are gets stronger flow, and is nutrient rich, so opportunistic inverts set-up camp there first.
I find that overflows are quite rich in many inverts, as it has the highest concentration of nutrients. Overflows are also dark, so there are no photosynthetic organisms (such as algae) to compete with.
Seeding is another aspect that needs to be fine tuned. If you have "clean" macro algae to start your refugium, or a very small starter culture, it will not have sufficient hitchhikers to seed a well populated benthic zone. Using small rocks at the base of the refugium will help anchor macro algae down until it attaches and provide adequate hitchhikers to populate all regions of the eggcrate below.
Time will eventually allow the entire eggcrate structure to be populated, but it will slowly grow, starting at the beginning of the path (settling zone near skimmer). Just like a city is populated, benthic inverts will slowly move out to the "suburbs" (deeper into the eggcrate structure), as the "downtown core" (settling zone) becomes overpopulated, and food becomes more scarce.
The continual breakdown of macro algae provides detrital nutrients for benthic inverts below. You could use rock in place of the eggcrate structure with success. A good experiment would be to use a few rocks in the benthic zone and see how populated they are compared to the neighbouring eggcrate. You would have to exclude any inverts that were likely to have started out on the rocks before being added to the benthic zone.
It's also possible that another media that is somewhere between the parameters of rock and eggcrate, such as (submerged) bioballs would be more efficient (ie. more surface area). The downside is that you would lose the ability to siphon away detritus, and the flow/nutrient supply would be greatly diminished in some areas. Live rock would only be efficient if it was small pieces, as the core serves no purpose in this zone. The problem with small pieces is they stack tightly, so you would lose most of the surface area.
There doesn't appear to be any benefit to a calcareous (calcium-based) media in the benthic zone. The system has ample sand and rock for chemical requirements such as buffering, and the benthic inverts seem to be just as happy on plastic.
Another worthwhile experiment would be to use a more significant number of rows of eggcrate. This would not effect flow or access for cleaning that much. The only thing to keep in mind is the amount of room sponges and tunicates (sea squirts) need to avoid competition from neighbours.
I'm looking forward to hearing more reports good, bad or indifferent.