The goal is to measure the elliptic flow $v_2$ in d+Au collisions using the multiparticle correlation method. The flow was reconstructed in Au+Au collisions - Phys. Rev. Lett. 118, 212301.

As D0 is a reconstructed particle (using hadron decay channel to K and pi with BR approx. 3.9%), we need to compute the reference flow (from all charged hadrons) and the differential flow (from D0 candidates - K pi pairs with right sign combination).

However, the detector acceptance is not uniform (as seen in figure of hadron phi distribution), therefore a correction has to be applied. The correction was implemented using the method described in Phys. Rev. C 83, 044913, which is the same method as in elliptic flow measurement of D0 meson in Au+Au collisions. 

Hadron phi distribution:  

Based on the paper, such a method is used for the non-uniform detector acceptance. In the case of a dead HFT sector (which is probably a cause of this hole in hadron phi distribution) it might not be enough (!! this is just our assumption with no other evidence as described here !!). Obtained cumulant has after this correction a negative value. As the reference flow is obtained as a square root of the cumulant another method will be used for fixinf the non-uniform acceptance - generic framework (Phys.Rev. C89 (2014) no.6, 064904).