The Ted Turner-ization of Drexlerian (and other) Nanotechnology

In the physical sciences, as is reportedly the case in other fields, a picture is worth 1000 words (within error bars). The goal of any good science image is to have the first 5 of those words not be "What the hell is that?" The description of a protein binding pocket without visual aids is about as painful to listen to as the description of a car engine without diagrams. These discussions are taxing even to people IN the field encompassing the details of the thing being described. In the case of science fields that receive lots of public attention (and nanotechnology is most definitely one of those), the need for visuals unencumbered by the assumptions of one's barricaded discourse community are all the more important.

It is with the above in mind that I post a nice little trick for improving the… readability of an atomically-precise image. The structure below is a low friction bearing assembly first posted off-somewhereville way back by Rocky Rawstern at the nanotech-now.com gallery. In preparation for a guest appearance in an upcoming Wiley-VCH book by Rolf Frobose, I wanted to do something beyond the simple "exploded view" of the two parts to show all of the interactions occurring between the fixed and rotating rings and the carbon nanotube shaft. As default-rendered by most chemistry programs, the structure would appear as the four atom-colored images at left. The dominance of carbon and hydrogen is obvious by the grey/white in the figure, while the remaining atom types pop out when you stare long enough. Among the many, many useful features in NanoEngineer-1 is the ability to colorize individual parts or pieces of parts, as well as to hide atoms to make the cut-away views (top right of the four-image sets). Using colors from the default palette, the atom-colored assembly can be turned into RGB images (center) that very clearly differentiate the rotating ring (red) from the diamondoid (blue) and nanotube (green) components (the blue and green pieces are covalently bound to one another, not separate). The clear benefit is identification of distinct pieces. Another benefit, to the editor, is the presence of color in the image (grey and white certainly looks technical, but it isn't very eye-catching). This trick was used in the nanotube junction image in the MTSU manufacturing article from earlier this year (which, although useful for showing all of the bound components, is a bit of an eye-sore). The loss in such a globally-colorized image is the atomic detail (there is no division beyond atomic radius for C, N, O, H, etc.) that may be unimportant to the casual observer but that is of some significance to a researcher trying to establish atomic structure and connectivity. Click on the image below for the full version.

The obvious compromise between component identification and atomic detail is the blending of the two views (right). In the assembly, this is performed in Photoshop by layering the colorized image above the atom-color image and changing the transparency of the colorized image (here, to 30%). There are some obvious benefits to this approach that go beyond the ease of viewing. First, superimposing the semi-transparent color on the atom-color image immediately provides the global detail in the colorized image with the atomic detail of the atom-color. Consider the time it would take to define the color of each hydrogen atom in each part, making the rotating ring some light shade of red, then defining some light shade of blue for one piece, some shade of green for the other, then doing the same thing for ALL of the unique atom types. Second, you haven't altered the original images (which seems like a simpleton point until you attempt to revisit the design a year later and can no longer provide the typical chemical (atom-colored) or instanta-fragment (colorized) views without re-rendering everything again (provided you saved the source files).

Quick and easy and a nice compromise of the two extremes. In the event anyone wants to try it out (or, please, suggest better approaches. I'm good with what POV-Ray I know, which is remarkably little considering what people can do with it), the two POV-Ray files for the cut-away view in grey and color are provided below, along with the final images and close-up overlay for each (there are no additional .inc or .ini files needed here; everything is in the .pov file). Again, click on the individual images for the full version.