Handedness after a mirror
A mirror reverses the handedness of circular light and swaps the two diagonals, while p and s light come back as they went in: two mirrors give the state back, one does not.
Open in the canvas →Background
Circular light is right-handed when its field turns clockwise as seen by someone looking towards the source, and left-handed when it turns the other way[1]. A quarter-wave plate makes it from linear light at 45° to the plate's fast axis[2]: in that convention, linear light along 0° through a plate whose fast axis is at 45° comes out right circular.
A mirror reverses the handedness[1]. The reason is a matter of directions. Of the two directions across the ray, the one perpendicular to the plane of incidence, s, is the same before and after the reflection; the one in the plane, p, turns over with the ray. The field still turns the same way in space, but the light now travels the other way, and seen looking towards its new source the rotation runs backwards. A second mirror reverses it again.
The same happens to linear light at any angle other than along p or s. Light polarized at +45° leaves the mirror at −45°, which is 135°: the mirror image. An analyzer at 45° that would pass all of it before the mirror passes none of it after. Light along p or along s is its own mirror image and leaves unchanged.
What this setup demonstrates
In the upper setup a laser polarized at 0° passes a quarter-wave plate with its fast axis at 45°, then two mirrors. Three beam probes show the polarization on the way: right circular after the plate, left circular after the first mirror, right circular after the second. The polarimeter at the end reads right circular, with s₃ = +1: its screen puts the state at the north pole of the Poincaré sphere.
In the lower setup a laser polarized at 45° meets one mirror. The probe after it reads linear 135°, and the power meter behind an analyzer at 45° reads nothing.
Things to try, each a single change from the setup as it opens:
| Change | Reading |
|---|---|
| Fast axis of the plate at 135° | left, right, left circular; the polarimeter reads s₃ = −1, the south pole |
| Analyzer at 135° | the power meter reads 1.00 mW |
| Lower laser polarized at 0° | the probe reads linear 0° and the power meter 500 µW |
| Lower laser polarized at 90° | the probe reads linear 90° and the power meter 500 µW |
An element's axis angle is fixed to the element. Light crossing a plate or a polarizer in the other direction, after a mirror for instance, sees the same axis at the opposite angle in its own frame, and the inspector says which direction the angle refers to.
The mirrors are ideal: they add no phase between s and p. A real metal or dielectric mirror adds some, which turns linear light at 45° slightly elliptical and circular light slightly elliptical, by an amount that depends on the coating and the angle of incidence. Every ray and mirror normal lies in the drawing plane, so s is out of the drawing at every mirror; a fold out of the plane, as in a periscope, is not drawn.