I don't think he was describing perpetual motion, but he gave no evidence that his method of photo-conversion is any more efficient than something like traditional solar cells. It is interesting that it captures infrared, though. He's also banking on something more efficient than lead-acid batteries for storage, which hasn't been invented yet. For his ideas to work, his technology has to prove that it can be a lot more efficient than the existing methods, both in capture and in storage. Lastly, it has to do it all at a price that is affordable.