How does one create a polarized beam of particles? I want to know how experimental physicists create spin-polarised beams of particles, say electrons, muons or quarks. My first guess is that one would polarise such a beam in a magnetic field. The two spin-states will have slightly different energies and the particles will spontaneously decay into the lower energy spin state. However, I see two problems with that. The energy gap should be minuscule and the lifetime of the higher energy state therefore quite high, such that unstable particles might decay first. Also, the magnetic field will cause charged particles to be deviated by the Lorenz force.
 A: While synchrotron radiation have an effect of self-polarization (and is actually the only way to polarize a positron beam) often, when a highly polarized beam is required, it is produced straight from the source and kept such to the best extent through all the accelerator chain.
For ring-shaped machines special sections are introduced which rotate the spin in such a way that the bending magnets do not affect it any further. These sections often include solenoid magnets and/or a combination of horizontal and vertical bends, therefore they are friendly called Snakes. The first family of them was invented in Novosibirsk, and got the adjective Siberian.
These can be avoided entirely by going for a figure-8 design such as the JLab EIC. In this case, the effect of dipoles cancel out and the spin is automatically and effectively realigned after a full turn.
These are of course the cases in which you want high beam current/energy and/or you are operating a collider. If you can afford to waste a significant fraction of the beam and have a fixed target experiment, then you can do any kind of filtering.
