After its discovery in 1947 by Willis Eugene Lamb and Robert C. Retherford the Lamb shift was used to create Lamb shift polarimeter to separate the $2S_{1/2}$ $\alpha_{1}$ and $\alpha_{2}$ hyperfine
substates of hydrogen as well as the $\alpha_{3}$ substate of deuterium. But for a new project at the
Technical University of Munich, investigation of the bound-beta decay of a neutron into a hydrogen atom and
a neutrino, a Lamb shift polarimeter is needed that is also capable of separating the $\beta_{3}$
substate of hydrogen. Unfortunately, our first attempt to use a Sona transition unit to
exchange the occupation numbers between $\alpha_{1}$ and $\beta_{3}$ failed, because of the unexpected
complexity of the transitions in this unit [1]. The second idea of using a new kind of spinfilter
which uses two radio frequencies to separate all four hyperfine substates of hydrogen also
failed.
Our third attempt is now to build a transition unit that can induce magnetic dipole transitions
between $\alpha_{2}$ and $\beta_{3}$ as well as between $\alpha_{2}$ and $\beta_{4}$ ($\pi$ transitions, i.e. an exchange of the occupation numbers of these states). This is a similar
transition to what is used in atomic beam sources, in this case not for ground state atoms but for metastable
atoms, which requires a much lower radio frequency. Another difference of this new idea is
the smaller interaction time of the atoms with the photons inside the transition unit due to
their much higher velocity of roughly $2\cdot10^5$ m/s compared to velocities of about $10^3$ m/s in an atomic beam source.