Last week started off mostly organisational, and I spent the first few days touring the factory working out how everything works and what gets done. The production line is manned about 16 hours a day (there are three work teams that rotate), but there are also teams of engineers, who are there to fix problems, manage, and I think also deal with clients and expriment and things.
So after a few days, I wanted to do something.There is a control room in the factory where all the cells get tested before being assembled onto the panels, and the room also has a few other testing devices. One of them tests reflectivity of a sample, and since it was available, I thought I'd use it. At each stage of the production there are a few breakages, so I used some of these from different stages to check how reflective they were. Most interestingly, one of the stages involves aving the cells sprayed with titanium dioxide to form an anti-reflective layer. If light hits bare silicon, a significant portion of the energy is reflected away because light travels so much more slowly inside it than it does in the air (this is roughly why you can see a faint reflection of yourself in glass windows, since not all of the light that bounces off your face passes straight through the window, a small amount bounces back. If you look at a window that doesn't have much light coming from the other side, eg. look outside at night, it's much easier to see this effect). It turns out that if you choose an intermediate layer perfectly and have the right (very thin) thickness, you can stop all light of a certain wavelength from being reflected, and significantly reduce reflection at other wavelengths. Glasses and camera lenses use this effect, camera lenses look a bit purpley because they block green reflectionmore-or-less totally, but only reduce the red and blue reflection (which are on either side of green). Red + blue = purple.
In the factory, this layer is currently applied by people with spray guns and they judge the thckness by the colour of the sample (because it changes the reflective properties, the colour also changes as the spray is applied. Some factories use a more advanced piece of kit - they're gettig one fitted here in december - that is more accurate and improves power output of the finished cells by about 10%). I wanted to see what the measured reflectivity was with a bit less spray than usual and a bit more than usual (and of course the normal amount as a control and no spray as another control). The other student engineers were keen to help - they seem not to have much to do in the factory and hadn't really thought about doing their own experiments.
I got some data last friday from one set of samples, and was planning to get lots more early this week. Unfortunatly, however, the machine has broken and I haven't been able to get more data. (It wasn't actually due to me! I think they were quite interested in my idea, last Thursday my professor told one of the students looking after me that he should do a similar experiment, 20 samples from each stage and measure the reflectivity, but being much more careful about getting samples from the same batch, etc, as a longer term experiment. So the machine was getting a lot of use and at somepoint seems to have given up! There's another component coming next week to fix it).
So this week I've been putting together a model of the reflective properties using excel (and a lot of notebook space! Both my pens ran out on the same day and it was surprisingly tough to obtain a new one!). It's a fairly simple model but without access to my notes from the last few years it's taken abit of time to get my brain back into gear, although quite fun to have a go. I've been able to find silicon's refractive index data on the internet (refractive index is hw fast light travels inside it compared to a vacuum, and has a lot of effect on its reflective/refractive properties. A pencil in a glass of water looks bent because water has a higher refractive index than air, so the light bends on entry). I've been trying to fit the model to the dat I got last week, although to be honest one set of samples is far too few to be considered scientific, so I'm going to collect more data as soon as the machine starts to work again.
The model with the anti-reflective layer is a bit more complicated because of various different reflections inside the layer itself that depend on the thickness in quite a complicated way and all interfere with one another. However, the model is quite good and I've now adjusted it to take account of silicon's low wavelength behaviour as well.
Yesterday I was also looking at the absorption of sunlight by solar cells. A good 95% or so of all solar panels are the basic silicon ones like the ones they make here. There is lots of innovation around the world about other types, but most are constrained by cost or still very young technology. Unfortunatly, the maximum possible efficiency (the proportion of incident solar radiation that gets turned into electricity) in a normal silicon cell is about 32%, and most real cells are lower - the ones they make here are about 16% efficient. This limit is due partly to the solar spectrum, much of which falls at wavelengths that are too long to excite the electrons inside the silicon, and also due to the way that electrons that are excited behave - a lot of their energy is converted into heat inside the cell. There are lots of ideas about how to improve this (such as using mirrors to concentrate the light, altering the spectrum using clever materials, or using thing other than silicon) and I've been studying a review paper on the topic, but at the moment these are all some way from being practical.
Next week I'll get more data and then see where to go from there, and then after that I'll be back in the lab as Shanghai University which will have finished its refurbishment and able to try out some new experiments!
ps. Whats happening with Russia? Getting snippets on the news but noone here knows how to translate it. I *would* check the news websites, but it seems anything with 'news' in the website is blocked here...
No comments:
Post a Comment