I remembered reading about this idea, too. Sad to see that nobody has yet experimented with it. Seeing as a pound of TiO_2 is pretty cheap on ebay, I may give this a shot, myself. For anyone interested, here are some sources I found about it.
I might be missing something, but my impression is that you can basically have a panel coated with the TiO_2, exposed to sunlight, and trickle some water down the panel to collect the nitrogen. As explained in one of the links, the catalyst loses some efficiency while exposed to oxygen and other things in air, which could be a reason we haven't seen it used on any kind of industrial scale. It's likely just easier and cheaper (monetarily) to use methods and infrastructure already in place to produce on the scales that industry relies on.
I'm figuring smearing some sort of waterproof adhesive over a board, and liberally applying the TiO_2 before it dries. Then simply wait for a good sunny day, and run some water down the board, collect it in a bucket. Should be easy enough from there to see if it is functional, marginally or otherwise. And cheap and simple enough that it's no biggie if it fails. I'll post results once I get things together (It could be a few weeks to a few months before I get to it, but it IS boldly written down on my to-do board.)