At the end of this talk, I got 2 questions.

First, M.Guidal asked why we fitted the double ratio by the function A * sin(phi - phi_S) and not by the function A * sin(phi - phi_S) + B * cos(phi - phi_S).
The answer is that there is no cos(phi - phi_S) term if we consider an unpolarized beam which scatters on a transversely polarized nucleon. If the beam is longitudinally polarized, as COMPASS one, a cos(phi - phi_S) appears with a factor sqrt(1-epsilon*epsilon) which is close to zero. Then we neglected it.

Second, M.Dielh asked about the maximum pt2 cut (which is pt2<0.5) which is applied in order to reduce background. He asked why we choose this value since we do not see any modification of the slope of the pt2 distribution. The answer is that we looked at the Emiss distribution to estimate the fraction of background. if we look at events with pt2 > 0.5, the background fraction is very high (>50%), so we reject them. This study has been done on 2004 data, the same cut is applied on 2007 since we want our proton and deuteron results in the same cinematic domain (for comparison).