 Microscopes are one of the standard tools in almost every lab nowadays. It gives us the ultimate chance to look at living cell interaction, to look at sub-cell structures and actually learn a lot about biochemics there. One special research field in the microscopy section is the fluorescence-labeled microscopy. There you have the big challenge that you're depending on your fluorescence signal for gathering information. You have to be really sure that if you get an uneven response from your spasm, that this uneven response is actually resulting from an uneven or inhomogeneous response from your probe. And not coming from an uneven illumination and therefore interfering with your non-linear interaction with the fluorescence molecules. Therefore, our question, our research question is how can we create an illumination for such microscopic applications that are so homogeneous that you can be definitely sure that the illumination is not messing with your results and that you are looking at the pure interaction of your excitation and then your fluorescence signal that's coming back from your spasm. In a common microscope on this area in the fluorescence microscopy, usually lasers or laser diets are used for illumination. Therefore, either they are fiber-coupled or free beam lasers, but they have a Gaussian beam distribution. It's very bright in the center and the intensity declines to the edges. This is definitely not something we want. And additionally, because we're dealing with that special total internal reflection fluorescence microscopy, we have to address other attributes as well. So we need a high spatial coherence to make that microscopy method work. We have to have a long working distance, which means the distance between where the profile is originally generated and where the profile then is used has to be several centimeters. It has to be acrobatics, so it has to work for all the wavelengths in the visible region at the same time. And of course, it has to have a high throughput efficiency. We do not want to lose any light, any photon in between in that illumination path. For all those reasons, we choose to go for a setup that almost looks like a Galilean telescope. So it's made out of two lenses, two very special lenses. It's a spherical lens and those lenses are facing each other and they are calculated in a way that the rays that come from the first surface that are redistributed to achieve this homogeneous illumination are collected by the second lens and then recolimate this beam to go for this long working distance. By doing this approach, we actually were able to achieve all the requirements to work within such a microscope setup. Now with that optical design being done for the beam shaping device, we actually built that and then it was implemented into a microscope setup to actually figure out how much of the demands we could actually meet in experimental testing. So the first thing we looked at was the threshold. By illuminating your spasm either with the Gaussian distribution or with the top head distribution, it's essential to know how much light you give into the system to get your fluorescence signal at an optimum so you can actually work with the data. For the Gaussian distribution, it's extremely sensitive to not overdo it in the center and actually have enough light on the edges to get a signal as a response. For the top head distribution, we could find that there's a huge region actually where we could change the incoming intensity and still get a stable counting of labeled cells in that experimental work. The second thing we looked at was photopleaching. Of course, fluorescence dyes have a photopleaching effect. This is just physics at that point. You excite the dye, the dye is giving the light back and at a certain time it's losing that ability and then it's just going dark at that certain spot. That's not problematic because we're talking about seconds, so you have enough time to get your image, but it gets problematic if the bleaching is not homogeneous, especially if you have a spasm that has dynamics and you're looking into fluid interaction or something. What we could find is that we have an extremely homogeneous photopleaching time all over the illuminated area having that top head distribution. The third thing we looked into was the background noise. If you do total internal reflection fluorescence microscopy, it's all about just illuminating the top layer. You're actually just looking at the top cells of your spasm and everything you get from the volume behind is more or less a background noise messing up your fluorescence signal. For achieving that, you need that high spatial coherence and we could show that with our optics or the beam shaping device we could keep the high spatial coherence from our illumination source from the laser diet and get that through the complete system and reduce the background noise by a significant factor compared to having a multimode fiber or something like that for illumination. In the experiment I just described, it's giving you better results. You have the better reliability that your fluorescence signal is actually coming from regions that want to interact with you and if you don't see anything then probably there's nothing there. What's even more interesting or what new options you have for that is going into high throughput stitching. Since you only can see a small part of your spasm when you look through a microscope, something like maybe 30 microns in diameter then you have to put images together in order to see a whole picture. To do so you have to move your spasm and you have to take several frames. If you have an even illumination like we demonstrated with the top head profile you have the chance to do that without further data processing and not ending up with a checkerboard pattern where you don't really know if it's coming from the illumination or actually there is parts missing in your sample. So I think there's a huge chance of speeding up optimized high throughput biological probe imaging there. I think there are two main topics here. The first thing especially for the high throughput imaging would be to lose that rotational symmetry and illumination path and actually go for a square or rectangular shape because your CCD camera has a rectangular shape usually and also that stitching is more efficient and can be even speed up more if you add rectangles next to each other. The second thing is of course right now we were limiting ourselves just to one very specific field in microscopy. So we were just looking at this total internal reflection for a sense microscopy. Of course that approach to go for homogeneous illumination can be expanded to other fields in microscopy. Even common microscopy where you just look through and see an image of whatever you want to look at can benefit because you see a more homogeneous illuminated image there but also more advanced techniques like stat microscopy can benefit from this approach.