Distribution of Radio Spectral Slopes of Galaxies in Optical Diagnostic Diagrams

TL;DR

This study analyzes the distribution of radio spectral slopes in galaxies within optical diagnostic diagrams, revealing three distinct groups linked to different AGN activity modes and merger stages.

Contribution

It introduces a novel classification of radio spectral index groups in galaxies using combined observational and machine learning techniques.

Findings

Identified three distinct spectral index groups with different radio loudness and emission line ratios.

Linked spectral groups to different AGN activity cycles and merger stages.

Demonstrated the effectiveness of machine learning in classifying galaxy radio properties.

Abstract

For about 500 intermediate-redshift sources ($0.04<z<0.4$), whose radio flux densities at $1.4\,{\rm GHz}$ are larger than 10 mJy, we performed additional observations at 4.85 GHz and 10.45 GHz using 100-m Effelsberg telescope. Our radio-optical galaxies are located preferentially in the composite and AGN spectral classes in the narrow line optical diagnostic diagrams (ODD). In the analysis, we focused on the distribution of radio spectral indices of radio synchrotron power-law profiles, $S_{\nu}\propto \nu^{+\alpha}$, in the ODDs. Using different analysis techniques, both observationally motivated and machine-learning based, we found three distinct groups--clusters in the radio loudness, [OIII]/H$\beta$ ratio, and spectral index volume: (1) sources with steep radio spectral index, large radio loudness and large [OIII]/H$\beta$ ratio; (2) sources with flat radio spectral index, intermediate radio loudness and lower [OIII]/H$\beta$ ratio; (3) sources with inverted radio spectral index, low radio loudness and low [OIII]/H$\beta$. The groups (1), (2), and (3) are located along the Seyfert-LINER spectral classes towards lower ionization ratios in the ODDs and hence can represent different activity cycles/accretion modes of AGNs, which could be in some cases associated with different merger stages.