Revealing minijet dynamics via centrality dependence of double parton interactions in proton-nucleus collisions

One of the main challenges hampering an accurate measurement of the double parton scattering (DPS) cross sections is the difficulty in separating the DPS from the leading twist (LT) contributions. We argue that such a separation can be achieved, and cross section of DPS measured, in proton--nucleus scattering by exploiting the different centrality dependence of DPS and LT processes. We developed a Monte Carlo implementation of the DPS processes which includes realistic nucleon--nucleon (NN) correlations in nuclei, and an accurate description of transverse geometry of both hard and soft NN collisions. Our method allows the calculation of probability distributions of single and double dijet events as a function of centrality, also distinguishing double hard scatterings originating from a single target nucleon and from two different nucleons. We present numerical results for the rate of DPS as a function of centrality, which relates the number of wounded nucleons and the sum of transverse energy of hadrons produced at large negative (along the nucleus direction) rapidities, which is experimentally measurable. We suggest a new quantity which allows to test the geometry of DPS and we argue that it is a universal function of centrality for different DPS processes. This quantity can be tested by analyzing existing LHC data. The method developed in this work can be extended to the search for triple parton interactions.