The making of a CIGS solar cell

Finally the nitty gritty of how to make thin film CIGS solar cells. Let's see, where to start?
The beginning comes to mind. "Substrate" in thin film lingo means the base upon which the solar cell will be formed. I'm using glass, ordinary window glass. Actually, I'm already cheating because the glass I use has a low iron content so it's not quite ordinary. Iron is a no-no for CIGS.

Step 1: Wash the glass. A hot ultrasonic soap bath and lots of rinses in purified water does it for me. Ends with a spin dry.

Step 2 is to coat the glass with a metal. Not just any metal. Molybdenum. In the pic you can see 4 pieces of the glass I use, 12.5 x 12.5 centimeters which has just been coated with molybdenum by a method called "sputtering". Since I'm going to focus on the CIGS layer, I won't describe what that means at this time (ok, briefly it means that highly energetic argon gas bombards a molybdenum "target" to knock off atomic chunks of the metal which conveniently land and stick to the glass substrate that's been placed nearby).

Step 3 is the fun part. CIGS, that is. Copper-indium-gallium-diselenide. As you may guess from the name, it's made up of four elements: three metals, copper, indium and gallium and one semi-metal, selenium. In a vacuum chamber where all the air has been pumped out, these metals get "evaporated". Literally. I have three pots for copper, indium and gallium. I heat these up way beyond melting. They get heated so much that they vaporize. Just for curiosity, the temperatures required are over 1000 degrees centigrade and about 1500C just for the copper. There's also a pyrex container with selenium that's vaporized, too, at a relatively measly 300C. These four vapors get mixed together in the vacuum. I then send the molybdenum-coated glass on a trip through this metal-selenide vapor and out it comes with a fine grey-black coating. If it happens that I get just the right amounts of all the elements and they coalesce at just the right temperature to form just the right crystal structure then I can get amazingly efficient solar cells, almost 20% efficient. But, really, it often seems like black magic to get it all just right like that. And therein lies the make-it or break-it future of CIGS. I, myself, have been able to make world record CIGS solar cells, but have been hardly reliable. Mostly I'm fixing corroded heating elements and trying to get this CIGS stuff to form in the right way. I think of it as making the world's best pizza with the tricky part being to make the world's best every single time I make a pizza. By the way, for those of you into character analysis, a lot of my analogies have to do with food.

Step 4 is the buffer layer. To refresh your memory on all the coatings involved I'll refer you to an earlier post about a CIGS solar cell. The standard buffer is a cadmium sulfide layer which involves a wet chemical dip in a mixture that turns intensely yellow and leaves a thin coating on the glass. I consider this process to be entirely disgusting and also a nuisance in that the used solutions contain cadmium and need to be treated as toxic waste. Good thing that alternatives to CdS exist. Even better that I use a non-toxic alternative myself, a zinc oxide - zinc sulfide mix which I deposit using another vacuum process. The process this time is called atomic layer deposition. It involves letting various gases flow over the substrate in turn, each leaving an atomic coating on the glass.

Step 5 is actually made up of two different zinc oxide layers. It's another sputtering process and is, in fact, carried out in the same vacuum chamber where the initial molybdenum was deposited. The first layer is pure zinc oxide. It's actually an insulator, but the layer is so extremely thin that charge carriers are able to cross it anyway by means of what physicists call "tunneling". The second layer is also zinc oxide, but it's mixed with 2% aluminum to make it conductive. It's a relatively thick layer because it needs to carry all the solar cell's current while minimizing resistive losses.

The last step is to deposit a metal grid on top. This layer is deposited by evaporating aluminum through a template in the pattern that you can see in the pic with the finished cells. It forms a convenient contact pad and the metal fingers help collect the generated current, thus providing some current carrying relief to the underlying ZnO:Al.

I typically carry out all of these steps on my own. From start to finish it could go as fast as a couple of days, but I'll usually spread it out over a week. Once the cells are made, they need to be characterized under sunlight and otherwise evaluated. Very often the CdS buffer will be replaced by some experimental layer. The CIGS evaporation system can make 24 substrates at a time, so it's quite time consuming to process and evaluate them all. I end up making a CIGS "run" only 2 or 3 times a month. Which is good, since it takes 12 or 13 hours to run the CIGS machine.

And in between, there are the endless repairs.