Mandelstam variables
It is clear that physical scattering amplitudes should better be independent
of the frame they're calculated or measured in. This translates into the fact
that they depend on covariant quantities only: Masses (or, better, their squares)
and products of four-momenta.
Basic scattering or production processes at collider experiments are typically
of the form 2→ 2:
where a, b, c, and d label different particles and the various p label
their respective momenta.
This leaves three independent scalar products of pairs of momenta, namely
The last identifications in each line are valid for massless particles in their centre-of
mass frame only. In this frame, θ is the scattering angle, i.e.\ the angle between
the momenta of particles a and c. For both massless and massive momenta, the Mandelstam
variables satisfy
which is obviously true for massless particles. This identity implies that only two of three
of these variables can be chosen at will. Usually, the choice is made to fix s, the squared
centre-of mass energy of the produced (in general: final state) particle pair, and to fix t,
the momentum transfer. From the identiy above, also
follows. Consider now the following diagrams, where the thick black blob of the process
scetch above has been decoded.
s-channel process:
the scattering process is realised through the exchange of a virtual particle
with momentum p=p1+p2
t-channel process:
the scattering process is realised through the exchange of a virtual particle
with momentum p=p1-p3
Of course, real processes may be more complicated, for instance, by mixing two or more different
ways in which they can proceed. In such a case, the total quantum mechanical transition amplitude
can be represented as the sum of individual contributions, i.e., as a sum of their respective
amplitudes. However, it is clear that specific processes may be dominated by amplitudes
like the left one in the figure above. In such a case, the diagram is usually coined an
"s-channel" diagram with corresponding "s-channel kinematics". Later on, this will be discussed in
more detail.