We examine the ways of extracting information from semi-invisible decays of (new) heavy particles at hadron colliders, i.e., such heavy particles are assumed to decay into visible/Standard Model (SM) particles and invisible particles. As a concrete realization, we employ the models with the stable weakly interacting massive particle (WIMP), a well-motivated dark matter (DM) candidate. By definition, dark matter is not seen by the detectors, i.e., invisible. Typically, stability of dark matter is ensured by introducing a new (unbroken) symmetry under which the DM is non-trivially charged while the SM particles are uncharged. Also, many new physics models contain other heavier particles which are charged under the same symmetry so that such heavier particles must decay into (invisible) DM particles along with the relevant visible/SM particles.

In particular, we study how to determine the masses of DM and heavy particles and the nature of the above-mentioned DM stabilization symmetries. For this purpose we take three kinematic variables as the main toolkits. We first discuss the distribution of the invariant mass formed by the visible part in the associated decays. As the second variable, we include the invisible part in forming the invariant mass. Because we are not aware of the longitudinal momentum of invisible particles, such a quantity is constructed in the plane transverse to the beam pipe, which is therefore called “transverse” mass. This is typically suitable for a singly produced heavy particle. Since the DM stabilization symmetries lead to pair-production of heavier particles, we here consider the “stransverse/M_T2” type variable, a simple generalization of the transverse mass. Finally, we consider the energy spectrum of visible particle(s), which is not Lorentz-invariant at all even under longitudinal boosts. The relevant strategy is predicated upon the new observations that we shall make about physical implications of the peak position in such an energy spectrum. We emphasize that the relevant methods using the three observables are complementary to one another.