Description
Since low-mass galaxies have experienced little merger history, they retain the remnants of the initial state of galaxy formation. Then low-mass galaxies provide the best laboratory for us to study the formation and evolution of galaxies and its associated dark matter haloes. At the same time, exploring the supermassive black hole (SMBH) at the center of the dwarf galaxy provides new clues for us to study galaxy-SMBH coevolution, galaxy feedback and quenching mechanism. And the study of the circumgalactic medium of dwarf galaxies is essential to understand the star formation activities. Using the DESI Image Legacy Survey data (g, r, z) combined with the WISE near-infrared survey data (W1, W2), I will present our machine learning model to search the dwarf galaxies (10^6 < M_*<10^9 M_⨀) in the local Universe (z < 0.02). The precision of separating dwarf galaxies from the contaminators could be as high as 95% and the recall could reach 76%. The verification using the independent data set, including DESI-EDR, SAGA and ELVES, shows excellent precision of > 96%. The late-type, ultra-low mass dwarf galaxies with high star formation rate might be analogs of galaxies at early Universe. Using DESI-DR1 data, we study its spectral and structural features. Further investigation finds that more than 200 of those local dwarf galaxy candidates have high x-ray emission fluxes, indicating possible existence of SMBH or IMBH.
