The angular momentum - mass (j-M) relation has previously been studied for stars and neutral gas (HI). In this work, we derive the j-M relation for molecular gas (H2) for a large sample of nearby disk galaxies for the first time. We aim to obtain and quantify the shape and scatter of the relation, as well as its contribution to the overall baryonic j-M relation and its connection with the star formation rate. Moreover, we will set a comparison benchmark for studies at high-z aiming to measure the angular momentum of galaxies at earlier cosmic times, where molecular gas tracers are used. 

Chemical modelling is often an essential component of hydrodynamical simulations in astronomy. However, the inclusion of chemistry makes simulations prohibitively expensive in terms of time and computational resources; often, simplifying assumptions are made at the cost of accuracy. Deep learning is an effective tool to incorporate chemistry into hydrodynamical simulations while not giving up on accuracy, speed and computational resources. In this work, I use NeuralODEs to emulate 3DPDR - a photodissociation region (PDR) code - and train it on a dataset of initial chemical abundances and temperatures corresponding to physical conditions in the Orion Bar PDR. NeuralODEs prove to be useful for capturing variations in the large physical parameter space, and evaluate the underlying chemistry orders of magnitude faster than 3DPDR. This work demonstrates using NeuralODEs as a strong proof-of-concept for emulating PDR codes, especially 3DPDR.