A cumulative search for hard X/$\gamma$-Ray emission associated with fast radio bursts in Fermi/GBM data

Context. Fast Radio Bursts (FRBs) are millisecond-long bursts uniquely detected at radio frequencies. FRB 131104 is the only case for which a $\gamma$-ray transient positionally and temporally consistent was claimed. This high-energy transient had a duration of $\sim 400$~s and a 15-150~keV fluence $S_{\gamma }\sim 4\times {10}^{-6}$ erg ${\mathrm{c}\mathrm{m}}^{-2}$. However, the association with the FRB is still debated. Aims. We aim at testing the systematic presence of an associated transient high-energy counterpart throughout a sample of the FRB population. Methods. We used an approach like that used in machine learning methodologies to accurately model the highly-variable Fermi/GBM instrumental background on a time interval comparable to the duration of the proposed $\gamma$-ray counterpart of FRB 131104. A possible $\gamma$-ray signal is then constrained considering sample average lightcurves. Results. We constrain the fluence of the possible $\gamma$-ray signal in the 8-1000 keV band down to $6.4 \times 10^{-7}$($7.1 \times 10^{-8}$)ergcm$^{-2}$ for a 200-s (1-s) integration time. Furthermore, we found the radio-to-gamma fluence ratio to be $\eta >{10}^8$ Jy ms erg${}^{-1}$ cm${}^2$. Conclusions. Our fluence limits exclude $\sim 94\mathrm{\%}$ of Fermi/GBM detected long gamma-ray bursts and $\sim 96\mathrm{\%}$ of Fermi/GBM detected short gamma-ray bursts. In addition, our limits on the radio-to-gamma fluence ratio point to a different emission mechanism from that of magnetar giant flares. Finally, we exclude a $\gamma$-ray counterpart as fluent as the one possibly associated with FRB 131104 to be a common feature of FRBs.