The atomic numbers and the masses of fragments formed in quasi-fission reactions have been simultaneously measured at scission in 48 Ti + 238 U reactions at a laboratory energy of 286 MeV. The atomic numbers were determined from measured characteristic fluorescence X-rays whereas the masses were obtained from the emission angles and times of flight of the two emerging fragments. For the first time, thanks to this full identification of the quasi-fission fragments on a broad angular range, the important role of the proton shell closure at Z = 82 is evidenced by the associated maximum production yield, a maximum predicted by time dependent Hartree-Fock calculations. This new experimental approach gives now access to precise studies of the time dependence of the N/Z (neutron over proton ratios of the fragments) evolution in quasi-fission reactions.PACS numbers: 25.70. Jj, 25.70.Gh, 32.50.+d, 24.10.Cn Since the mid-70s, it has been known that the formation of super-heavy nuclei by fusion is hindered by out-of-equilibrium mechanisms [1][2][3]. In these mechanisms, the available kinetic energy can be totally dissipated and large mass transfers between the projectile and the target can occur, leading to emerging fragments quite difficult to distinguish from fragments arising from fusion followed by fission (that might be mass symmetric or asymmetric) [4][5][6][7]. Due to these characteristics, the generic name quasi-fission (QF) is nowadays often used for all these mechanisms. Since the pioneering works, many experimental aspects of QF have been explored [8][9][10][11][12][13][14][15][16][17] and dynamical models, macroscopic or microscopic, have been developed in order to reproduce cross-sections, distributions of mass, angle, kinetic or excitation energy and some of the correlations between these observables [15,[18][19][20][21][22][23][24][25]. Considering the huge experimental difficulties to extract in a non-arbitrary way small cross-sections of fusion followed by fission from dominant quasi-fission cross-sections, a key issue for super-heavy nucleus formation studies, it is now essential to get a very good understanding of the QF mechanisms and to confront and improve the models with unambiguous exclusive data in order to reach reliable predictive capacities.A simultaneous determination of the fragment atomic number (Z) and mass (A) formed in QF or in fission processes remains nowadays a challenge [26][27][28][29][30], especially difficult because these quantities are most of the time measured after particle evaporation. In this letter, an experimental approach giving access for QF fragments to A and Z at scission will be presented and the data compared with predictions of a microscopic time dependent HatreeFock (TDHF) model [22]. The atomic number was determined from the coincident characteristic fluorescence X-rays, as already attempted for fission fragments [31], whereas the mass was determined from the velocities of the emerging fragments.A 48 Ti 19+ beam was accelerated at 5.75 MeV/nucleon by the Australi...