I'd like to discuss some of the advanced design principles for semiconductor lasers. So, if you get a job where you're asked to design semiconductor lasers, by the end of this unit you would know how to do it. So, the simplest type of semiconductor laser is one where you have the P-N junction. The big problem with this is that the electrons and the holes can go wherever they want, and so they don't necessarily have to stay in close proximity to each other, which decreases the chances of radiative recombination. One way to fix this is to basically make the structure more complicated, and so the idea here is that we add in different materials, and different materials with different band gaps. So the idea is you make a sandwich and you basically keep the electrons and holes at the center of the sandwich so that they're in close proximity to each other. If the single heterostructure, heterojunction is not good enough we can go to a double heterojunction, which essentially just means making sure that we use several different materials, and the idea here again is to confine the electrons and the holes. Unfortunately, there's still a problem with carriers being able to escape and if they have enough energy we can figure out exactly how many carriers can escape. When they escape it essentially just contributes to loss and it lowers the gain that you can get from the structure. So, one potential alternate approach is to look at Quantum Well. So this is a really, really thin layer of semiconductor material. So now, all your electrons and holes will stay there. The other benefit is that because we have quantum confinement, we can actually tune the wavelength of the emission by the energy levels that are in this box. We really want in a semiconductor laser to have the electrons and the holes confine and then we want the optical mode to overlap with that region. So, here you can see the double heterostructure. So, the confinement is about one, but then as we said, the carriers are escaping. In the single quantum well, we can confine the carriers really nicely but then the optical mode is really not very well overlapped with the quantum wells. If we go to the separate confinement heterostructure, essentially here we can essentially get good overlap with the region that confines the electrons and the holes and basically get the best of every well. Another option is to go to multiple quantum wells. This also solves the problem and can provide increased gain, although it is at the expense of having to put a little bit more power into the laser in order to turn it on.