We are working to help further develop a low cost spectrometer design that was published online by B. Hickman. To support this work, we are developing simple methods to make reflective optics.
The figure to the left shows a screenshot of a parabolic reflector designed using Sketchup. We wrote a Ruby script that generates the 3D mesh for the reflector (let me know if you would like the script for this). We exported this model as a STL file and printed copies in ABS and PLA. The plastic copies are wet sanded with 180 grit sandpaper and a couple of different reflective surfaces have been applied. Our first attempts used aluminum duct tape, which goes on easily and readily conforms to the curved surface. We have, however, had limited success at achieving a high quality polished surface using this material. I think the problem is that duct tape is made from the cheapest possible metal. My guess is that they just melt down anything in the recycle bin and that there is stuff in there that is not ever going to give a good shine. For our next attempt, we are using aluminum flashing and securing it to the 3D printed plastic model using contact cement. This seems like it will give a higher quality reflector but this material has limitations as well. To start off, it is much thicker and thus is harder to work with. While you can cut duct tape with razors and scissors, flashing requires metal cutting shears. It will also not accept a compound curve (such as the one in the figure) and is prone to leaving a sharp edge. For the spectrometer, it is not entirely necessary to use a compound curve since we can incorporate separate mirrors to focus in different dimensions. To achieve a reflective surface, we first wet sanded the surface with 1000 grit sand paper, then 1500 grit. After that, we polish with toothpaste, then with an automotive polish. We have not refined this method yet and have a number of automotive finish products to try.
Of course, having a compound curve has advantages and we have other options to explore. We have some reflective mylar on order and I will let you know how that fares. I have a feeling that it will work well for visible light, but that it will not be a good reflector for either UV or IR.