I've thought of salvaging a beam splitter from a CD/DVD drive, to provide a reference beam and a free-path beam. Differences in readings of the detectors would imply the presence of particles.
Not sure how sensitive it would be.
Current techniques use two types of "channels". On is called a Light Channel and the other is a Dark Channel.
What you describe with the CD/DVD would be an example of a Light Channel setup, which is the lower sensitivity of the two. The typical setup of a Light Channel would have a laser beam shoot into an attenuation block that would have a screen the size of the laser spot to block all laser light from the Photo Multiplying Tube (PMT) behind the screen. When the particle enters the laser-beam, it will deflect photons around the attenuation screen and hit the PMT to register. The photon(s) hits the charged plate of electrons and they scintillate to the anode of the vacuum tube and you now you have a sensitive signal 
You use the reference beam for setting up the Automatic Gain Controller circuit. You can control the software thresholds this way and call the machine voltage calibrated.
Well, that is the theory, until you introduce very low quality PMTs and/or lasers which reduce the price and you get a "dust detector".