For the purposes of deorbiting junk it may be sufficient to just impart some delta-v, but you neither want that to be uncontrolled collisions sending stuff into even more erratic orbits, nor would you want to risk generating even more tiny pieces.
There is almost certainly a better material to use as well, but I don't think either of these practical concerns invalidates the idea.
Of course, big enough Space Sponges to make a practical difference would be very visible to the naked eye.
The problem in the end is launch costs. Yes, we're getting better, but that doesn't mean we can shoot up a "visible to the naked eye"-sized impactor. Not to mention what difficulties an object of that size would bring for non-garbage that also occupies those orbits.
Lastly, I have no idea how much research there has been in containing the fragmentation of hypervelocity impacts. But presumably it would be an important part of the mission to not generate more garbage.
Even if launching costs ever makes this feasible, any Space Sponge big enough to be efficient would be a big collision problem in itself.
But passive objects are only as likely to provide value as the ratio of the number of the total volume of the sacrificial objects to the total volume of assets, so you'd want a bunch of them.
Since the debris of interest is really just the material traveling in orbit, a cylinder might be more efficient, if some primitive attitude control could be included efficiently.
The exception to this would be if the object were in an elliptical orbit, which would allow for contact to be made with the scaffold at its perigee, which be impossibly rare.