I think you are on to something. For non-concentrated solar, what a low-cost application would use, the limit is given as 43%. The numbers given trigger my bull-detector hard and I've had some insight into the difficulties that you face if you actually want to build a rectifying diode. Those things are hard. Really hard. You're getting nowhere near where you'd have to be to achieve the limit. To satisfy the claims given, these guys have to be geniuses beyond imagination.

If something sounds too good to be true, it probably is.