Subleading contributions to the decay width of the $T_{cc}^+$ tetraquark

Recently the LHCb collaboration has announced the discovery of the $T_{cc}^+$ tetraquark. Being merely a few hundred ${\rm keV}$ below the $D^{*+} D^0$ threshold, the $T_{cc}^+$ is expected to have a molecular component, for which there is a good separation of scales that can be exploited to make reasonably accurate theoretical predictions about this tetraquark. Independently of its nature, the most important decay channels will be $D^+ D^0 \pi^0$, $D^0 D^0 \pi^+$ and $D^+ D^0 \gamma$. Its closeness to threshold suggests that the mass and particularly the width of the $T_{cc}^+$ tetraquark depend on the resonance profile. While the standard Breit-Wigner parametrization generates a $T_{cc}^+$ that is too broad for current theoretical calculations to reproduce, a three-body unitarized Breit-Wigner shape reveals instead a decay width ($\Gamma_{\rm pole} = 48\pm 2\,{}^{+0}_{-12}\,{\rm keV}$) consistent with theoretical expectations. Here we consider subleading order contributions to the decay amplitude, which though having at most a moderate impact in the width still indicate potentially significant differences with the experimental width that can be exploited to disentangle the nature of the $T_{cc}^+$. Concrete calculations yield $\Gamma^{\rm LO} = 49 \pm 16\,{\rm keV}$ and $\Gamma^{\rm NLO} = 58^{+7}_{-6}\,{\rm keV}$, though we expect further corrections to the ${\rm NLO}$ decay widths from asymptotic normalization effects. We find that a detailed comparison of the ${\rm NLO}$ total and partial decay widths with experiment suggests the existence of a small (but distinguishable from zero) non-molecular component of the $T_{cc}^+$.