My (cursory) understanding from talking to @maxliboiron is that this microscope, which has a light source that passes through the sample (on the opposite side from the lens) is not as good for this kind of thing as a dissecting microscope, which would light the sample from the same side as the lens. There's a great activity page on how to do a visual analysis of a sample here: https://publiclab.org/notes/maxliboiron/07-19-2018/how-to-analyze-plastics-forensically.
Some of what I'm seeing (and what @kgradow1 points to) suggests that a relatively low magnification (but higher resolution?), is what's needed-- I am not sure about the Raspberry Pi microscope, but the Community Microscope (with the webcam) might actually be able to get close by NOT inverting the lens, and simply using it for very close-up macro photography, which was a thing we discovered it could do really nicely by accident.
I think a modified version of the Community Microscope, where we reorient the light source and provide some options for its positioning and intensity (and possibly explore how well an unmodified lens performs here) would be a great next-step for a prototyping project!
Yes-- the issue is the lighting, not the magnification. Most plastics are opaque, and the closer you zoom in, the harder it is to identify plastics from non-plastics, so a scope that has an oblique light with relatively low magnification (even x10-25) and high resolution is best from microplastics. (In a traditional lab, we use a dissecting scope rather than a compound scope). There is no lower size limits to microplastics (they go down to the nanoscale), but smaller than 1mm means that visual identification, even with a microscope, is no longer reliable as a source of identification. You'd have to use spectrometry.
So, yes to the web cam community scope route as a potential step for microplastics.