The rest of the legend says "100v test reject". These are a batch of cells that failed one of the tests at Sunpower and made it into a reject bin, and yet are probably completely functional, especially if not strung together into a very long series.

Here's the Sunpower page on their CPV cell line. The cells I have, HEDA312, are the successor to the HEDA303 listed on that page. They're available for about $10 each in quantity 10,000 (last I heard ... which is too expensive in my opinion by a factor of two or so). 1.5 cm^2 area. Generates 8.3 watts at 250 suns according to their web page. Back contact. I got 60 of them from a colleague to get me started. If you need one or two for experimenting, please contact me ($15 each).

Below, you can see the first cheesy heatsink I made, late one night. One CPU heat sink, plus one layer of clear packing tape for insulation, and copper mesh tape to put 6 cells in series. I don't think I'll even try to attach cells to this piece of junk.

In the background, you can see some nice sky-blue highlights reflecting off the active surface of one of the cells, which looked black in the previous picture. The rest of them are face down; the entire back surface is a contact area, with half positive and half negative.

Update: here's version two of the cheesy heatsink, more or less ready to test. I glued a thermistor on top to try to measure temperature; it's totally the wrong tool, but I had it around. That's typical of this whole heatsink, as you can see.

For future heat sinks, I've found an almost magical keyword online: "waterblock". That's what people call their water-cooled CPU heatsinks. For instance, this page of great looking waterblocks from the guy in Ohio who makes them (at this email ).