Beam expanders: Keplerian and Galilean
Two ways to make a 4 mm laser beam 12 mm wide with two lenses: one focuses the beam between its lenses and is 200 mm long, the other never focuses it and is half that length.
Open in the canvas →Background
Two lenses that share a focal point turn a parallel beam into a parallel beam of another width. The first lens sends the beam towards (or away from) the common focal point; the second, one focal length from that point, makes it parallel again. Similar triangles on either side of the focal point give the ratio of the widths: f₂/f₁. This is a telescope used backwards: a telescope's magnification is the focal length of its objective divided by that of its eyepiece[1], and a beam sent in through the eyepiece comes out wider by the same factor.
The two classical telescopes give the two classical expanders. Kepler's, of 1611, has two converging lenses[1]: the beam comes to a real focus between them, and the lenses are f₁ + f₂ apart. Galileo's, of 1610, has a diverging lens and a converging one, and no focus in between[1]: the diverging lens has a negative focal length, so the same sum, f₁ + f₂, is now a difference and the pair is shorter.
The focus is what decides between them. A Galilean expander never focuses a collimated beam to a point, so it avoids the effects of high power density, such as dielectric breakdown, that a focusing design can meet with a powerful laser[2]. A Keplerian expander offers the focus itself: a pinhole placed there cleans the beam, as in the pinhole spatial filter.
What this setup demonstrates
Two 633 nm lasers each send a 4 mm beam into a pair of thin lenses. The upper pair is Keplerian: f₁ = 50 mm and f₂ = 150 mm, 200 mm apart, with the focus drawn 50 mm after the first lens. The lower pair is Galilean: a concave lens of f₁ = −50 mm and the same f₂, 100 mm apart. Both widen the beam by 150/50 = 3.
Each beam ends on a wavefront detector, whose screen reads COLLIMATED, divergence 0.00°. Select a detector and its inspector gives the width of the beam on it: a spot span of 12.0 mm, for both.
Things to try. Move the second lens of either pair 10 mm further from the first: the screen reads CONVERGING, 0.31°. Move it to 10 mm closer than where it started instead, and the screen reads DIVERGING by the same angle. This is how a beam expander is focused on the bench. Put the lens back where it started. Then, in the upper pair, set f₁ to 25 mm and bring the second lens 25 mm closer, so that the lenses are again f₁ + f₂ apart, now 175 mm: the beam leaves collimated and 24 mm wide, six times its width.
The second panel of the Imaging with lenses calculator works out the same pairs, thin and with real, thick lenses, and what they do to an object at a finite distance.
The lenses are ideal thin lenses, so both expanders are free of aberrations at any beam width, and the two designs differ only in length and in the focus. With real singlets they differ in aberrations too, and the best orientation of each lens matters; neither is shown here.
The beam is a bundle of 25 parallel rays with sharp edges. There is no Gaussian profile, so the beam does not spread by diffraction after the expander, and no account of how expansion lowers a real beam's divergence. The focus of the Keplerian pair is a point with no size: nothing is computed about the intensity there.