While searching at the Large Hadron Collider (LHC) for the production and decay of the CP-odd scalar (A0) in the 2-Higgs-Doublet Model (2HDM) with Natural Flavour Conservation (NFC) via the channels gg→A0 (through one-loop triangle diagrams) and A0→h0Z∗ (with mh0=125 GeV or mh0<125 GeV, with Z off-shell), respectively, a factorisation of the two processes is normally performed, with the A0 state being on-shell. While this approach is gauge-invariant, it is not capturing the presence of either of the following two channels: gg→Z∗→h0Z∗ (through one-loop triangle diagrams) or gg→h0Z∗ (through one-loop box diagrams). As the resolution of the A0 mass cannot be infinitely precise, we affirm that all such contributions should be computed simultaneously, whichever the h0(Z∗) decay(splitting) products, thereby including all possible interferences amongst themselves. The cross section of the ensuing complete process is significantly different from that obtained in the factorisation case, being of the order up to ten percent in either direction at the integrated level and larger (including changes in the shape of kinematical observables) at the differential level. We thus suggest that the complete calculation ought to be performed while searching for A0 in this channel. We illustrate this need for the case of a 2HDM of Type-I in the inverted hierarchy scenario with mh0<125 GeV.
Read full abstract