Chemical segregation analysed with unsupervised clustering

TL;DR

This study applies unsupervised clustering algorithms to molecular emission data in dense cores, revealing chemical segregation patterns and physical structures, and demonstrating the effectiveness of small datasets in understanding core chemistry.

Contribution

It introduces a novel application of density-based clustering to molecular emission data, uncovering chemical differentiation and physical structures in starless cores.

Findings

Successfully reproduces known molecular segregation patterns.

Identifies new segregation between c-C3H2 and CH3CCH.

Shows small datasets can yield meaningful structural insights.

Abstract

Molecular emission is a powerful tool for studying the physical and chemical structures of dense cores. The distribution and abundance of different molecules provide information on the chemical composition and physical properties in these cores. We study the chemical segregation of three molecules (c-C$_3$H$_2$, CH$_3$OH, CH$_3$CCH) in the starless cores B68 and L1521E, and the prestellar core L1544. We applied the density-based clustering algorithms DBSCAN and HDBSCAN to identify chemical and physical structures within these cores. To enable cross-core comparisons, the input samples were characterised based on their physical environment, discarding the 2D spatial information. The clustering analysis showed significant chemical differentiation across the cores, successfully reproducing the known molecular segregation of c-C$_3$H$_2$ and CH$_3$OH in all three cores. Furthermore, it identifies a segregation between c-C$_3$H$_2$ and CH$_3$CCH, which is not apparent from the emission maps. Key features driving the clustering are integrated intensity, velocity offset, H$_2$ column density, and H$_2$ column density gradient. Different environmental conditions are reflected in the variations in the feature relevance across the cores. This study shows that density-based clustering provides valuable insights into chemical and physical structures of starless cores. It demonstrates that already small datasets of two or three molecules can yield meaningful results. This new approach revealed similarities in the clustering patterns of CH$_3$OH and CH$_3$CCH relative to c-C$_3$H$_2$, suggesting that c-C$_3$H$_2$ traces regions of lower density than to the other two molecules. This allowed for insight into the CH$_3$CCH peak in L1544, which appears to trace a landing point of chemically fresh gas that is accreted to the core, highlighting the impact of accretion processes on molecular distributions.