Well, every molecule vibrates at a unique frequency, determined by its mass, configuration, etc. Microwaves work by emitting electromagnetic waves tuned to the unique harmonic frequency of water. The water molecules hit by such waves begin to move about wildly. Basically, it's as if you shouted "WATER!" and every water molecule in range started bouncing up and down going "HEY! THAT'S ME!!"
My glucometer works on the same basis. It sends a small voltage tuned to the frequency of glucose through a sample of my blood, and, based on the response, can tell how sugar there is in the sample.
So, if Fred were to have a tracer which sent signals around tuned to the frequency of a particular tracer molecule, forgetting that that molecule may exist elsewhere... It could work. I'm not sure how he'd manage to find the response. His equipment would have to be (a) touching the sample it was scanning, (b) extremely sensitive to tiny changes, or (c) using a high (perhaps dangerous) level of signal output.
If you're going with IR, you might want to look at
this page (which turned up on a quick web search). It explains, in detail, how to find a given molecule by analyzing the exact patterns of its response to IR exposue. (You shoot a beam of IR at it, and see what gets reflected back.) Plenty of nice mumbo-jumbo phrases to choose from, too.
Basically... you shoot a beam of IR light at a molecule, and there will be peaks at certain frequencies from each of its components. Different atoms bonded in different ways and configurations, will reflect at different peak frequencies. By analyzing the full pattern you get from a given molecule, you can figure out its composition. You could therefore program a sensor to look for specific frequency response profiles, and thereby detect the presence of one or more specific molecules. At least, that's what I get out skimming that page.
Does that help?
Paul
