Not yet reviewed by Pith; the record is open.

We show results from the Coherent CAPTAIN Mills (CCM) 2019 engineering run which begin to constrain regions of parameter space for axion-like particles (ALPs) produced in electromagnetic particle showers in an 800 MeV proton beam dump, and further investigate the sensitivity of ongoing data-taking campaigns for the CCM200 upgraded detector. Based on beam-on background estimates from the engineering run, we make realistic extrapolations for background reduction based on expected shielding improvements, reduced beam width, and analysis-based techniques for background rejection. We obtain reach projections for two classes of signatures; ALPs coupled primarily to photons can be produced in the tungsten target via the Primakoff process, and then produce a gamma-ray signal in the Liquid Argon (LAr) CCM detector either via inverse Primakoff scattering or decay to a photon pair. ALPs with significant electron couplings have several additional production mechanisms (Compton scattering, $e^+e^-$ annihilation, ALP-bremsstrahlung) and detection modes (inverse Compton scattering, external $e^+e^-$ pair conversion, and decay to $e^+e^-$). In some regions, the constraint is marginally better than both astrophysical and terrestrial constraints. With the beginning of a three year run, CCM will be more sensitive to this parameter space by up to an order of magnitude for both ALP-photon and ALP-electron couplings. The CCM experiment will also have sensitivity to well-motivated parameter space of QCD axion models. It is only a recent realization that accelerator-based large volume liquid argon detectors designed for low energy coherent neutrino and dark matter scattering searches are also ideal for probing ALPs in the unexplored $\sim$MeV mass scale.