Three-dee, continued
May. 25th, 2017 08:12 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
mmcirvinSeven and a half years ago I attempted to reverse-engineer some filched pairs of 3D glasses. I successfully figured out that they were using circular polarization to separate the left- and right-eye images. What I couldn't figure out was why the backs of the lenses seemed to have a linear polarizer on them.
I just belatedly learned why: the way you filter circularly polarized light is to use a quarter-wave plate with a linear polarizer behind it. The quarter-wave plate is made of birefringent material in which the speed of light is different for linear polarizations along perpendicular axes. If it's the right thickness, it will alter the relative phase of the perpendicular components such that right and left circularly polarized light turn into different linear polarizations. Then you filter out the component you don't want with a linear polarizer. It works in the other direction too, to make circularly polarized light, which is how I could figure out much of what was going on by looking through the glasses in a mirror.
I'd vaguely imagined a layered construction with a circular polarizer in front of a linear polarizer, but in fact the linear polarizer is itself a vital component of the filter for circularly polarized light. The "fast" and "slow" axes of the quarter-wave plate (which, alone, would leave the linear polarization of light unchanged) are at 45 degrees to the linear polarizer, so there's no way to send linearly polarized light in through the front so it all gets filtered out. But you can do that by sending it through the back, just as I saw in my experiments.
It's interesting that this can be made to work well for all the colors of visible light--the base wavelength over the visible spectrum varies by about a factor of two. I guess the wavelength itself doesn't matter as much as the difference in the speed of light (or index of refraction) between the linear polarizations--all that's needed is for that to be relatively stable over the desired range.
I just belatedly learned why: the way you filter circularly polarized light is to use a quarter-wave plate with a linear polarizer behind it. The quarter-wave plate is made of birefringent material in which the speed of light is different for linear polarizations along perpendicular axes. If it's the right thickness, it will alter the relative phase of the perpendicular components such that right and left circularly polarized light turn into different linear polarizations. Then you filter out the component you don't want with a linear polarizer. It works in the other direction too, to make circularly polarized light, which is how I could figure out much of what was going on by looking through the glasses in a mirror.
I'd vaguely imagined a layered construction with a circular polarizer in front of a linear polarizer, but in fact the linear polarizer is itself a vital component of the filter for circularly polarized light. The "fast" and "slow" axes of the quarter-wave plate (which, alone, would leave the linear polarization of light unchanged) are at 45 degrees to the linear polarizer, so there's no way to send linearly polarized light in through the front so it all gets filtered out. But you can do that by sending it through the back, just as I saw in my experiments.
It's interesting that this can be made to work well for all the colors of visible light--the base wavelength over the visible spectrum varies by about a factor of two. I guess the wavelength itself doesn't matter as much as the difference in the speed of light (or index of refraction) between the linear polarizations--all that's needed is for that to be relatively stable over the desired range.