Don K.'s Laser Page

Meanwhile, you might want to consider some messy limitations about minimum theoretical divergence of laser beams. This is some messy stuff resulting from the wave nature of light.
The minimum divergence rate of a visible red laser beam in milliradians is approximately equal to 1 divided by its initial width in millimeters. Given the very tiny size of the light emitting part of a diode laser, a very considerable beam divergence is expected. This does indeed occur.
This can be largely fixed by placing a convex lens in front of the diode laser, with a distance nearly equal to the lens's focal length. This reforms the beam, giving it a new, wider initial width with a correspondingly lower divergence. You will have to adjust the focus (or lens distance) yourself for best results.
Once you do this, you might wonder what happens with the beam, since the beam from a typical laser diode is not round, but oblong. This occurs because the light emitting part of the laser diode is oblong. At best, the wider dimension of the beam will diverge less than the narrower dimension. The best to be expected from compact lenses around a centimeter in diameter is a beam with initial dimensions (upon leaving the lens) of nearly 1 centimeter in the wider dimension, by a couple or a few millimeters in the narrower dimension. The wider dimension will expand by a millimeter every ten meters or so, while the narrower dimension of the beam will expand by a millimeter every couple or few meters. The divergence may even be greater than this if the lens is of poor quality or not exactly at the optimum distance from the laser diode, or if the beam exits the lens with smaller dimensions than just mentioned above.
The divergence may not be apparent within a few meters of the lens, if the "waist" of the beam occurs at that point. The beam "waist" is a region that sometimes occurs if the lens is trying to make the beam converge at the same rate that the wave nature of light is trying to make the beam diverge. At long distances, the beam *will* diverge, at best, at a rate in milliradians roughly inverse to its initial width in millimeters.
To get less divergence, you need more complex optics or a He-Ne laser, which has a very close-to-ideal round beam. Although a He-Ne laser's beam is fairly narrow and would diverge roughly by a millimeter every meter, this can easily be "fixed". Simply fire the He-Ne through a telescope, into the eyepiece and out the other end. If the telescope is optimally focused, the beam will exit the front of the telescope with a diameter magnified by the telescope's magnification. (Try not to magnify the beam beyond the diameter of any lens it has to go through.) With luck, you could get the beam to have an extremely low divergence of around a millimeter every several 10's or even roughly every 100 meters.
Diode lasers generally don't have ideal beam characteristics, but they are fairly easy to focus to the degree that the beam does not widen by more than a millimeter or two per meter. Most diode laser "pointers" and collimated diode laser modules should achieve this.