A constellation of quarkonium-mass peaks has been reported in the last decade, opening what could be an entire new spectroscopy of nuclear-physics like complexity. Salient among these structures are the $Z_c$, much analyzed at BESIII as nicely summarised by J. Zhao at this conference. Their stature as new hadron states has often been challenged by proposals that the peaks arise from triangle singularities instead,
because they appear in three-body reactions and often satisfy the special kinematic conditions of the Coleman-Norton theorem.

To experimentally decide, we should like to resort to different reactions erasing the kinematic coincidence, and see whether each peak survives. We observe that the most universal "eraser" is the medium created in Heavy Ion Collisions, because its high temperature often affects the masses and widths of the particles participating in the reaction: if additionally the triangle reaction is completed during the lifetime of the fireball, our computations in thermal field theory show that the triangle singularities do not survive the hot hadron phase (much less the hotter quark-gluon plasma). On the contrary, ordinary quarkonia are affected but often survive the fireball and molecular-like states (such as the antinuclei $^2\overline{H}$, $^3\overline{H}$ or $^4\overline{He}$) are also routinely detected.

We have provided several examples, and here investigate the exotic $Z_{c}(4020)^\pm\to \pi^{\pm} h_c $ peak (produced against a $\pi^\mp$ in $e^+e^-$ collisions).
If, as has been proposed during the debates, it is produced or enhanced by a $D_{1} D^{*} D^*$ triangle loop, the estimate here reported suggests its disappearance
near the critical temperature of the confinement/deconfinement phase transition, so any future detection in heavy-ion collision data would suggest it to be a real hadron, with its absence leaning towards the singularity interpretation.