The polarization structure of p+p->d+π and of spinwise similar reactions

The polarization structure of the reaction 1/2+1/2->1+0 is discussed in all conceivable optimal formalisms, with an application to p+p->d+π in mind. First, the relationship between the observables and the bilinear products of amplitudes is given when only Lorentz or rotation invariance holds. Then parity conservation is imposed (with the product of the four intrinsic parities taken to be -1) and the resulting modified relationship between observables and amplitudes is displaced in all four possible optimal types of formalisms. This relationship is also constructed in the special case when the reaction angle is 900. A comparison of the four types is made from the point of view of the phenomenological determination of amplitudes from observables, and it is found that the pure transversity formalism has the largest number of advantages. In this pure transversity formalism, a complete experimental procedure is specified for the determination of amplitudes. The procedure is also outlined for the special case of thetasub(cm)=900. It is shown that while a complete determination of all amplitudes involves some measurement of the tensor polarization of the deuteron and also tends to involve observables in which all three particles are polarized, nine out of the eleven amplitude parameters can in fact be determined by measurements which involve at most two polarized particles at a time and which avoid the measurement of tensor polarization for the deuteron. (orig.)