Exploring the diffuse interstellar bands with the Sloan Digital Sky Survey

TL;DR

This study utilizes SDSS spectra to map diffuse interstellar bands across 5000 square degrees, revealing their varied correlations with interstellar medium components and providing a unified parametrization of their behaviors.

Contribution

It introduces a method to measure and map DIBs using SDSS data, uncovering their diverse dependencies on hydrogen and dust, and proposing a simple model to describe their complex relationships.

Findings

Detected over 20 DIBs in SDSS spectra.

Revealed different DIB responses to atomic and molecular hydrogen.

Established a unified parametrization for DIB behaviors.

Abstract

We use star, galaxy and quasar spectra taken by the Sloan Digital Sky Survey to map out the distribution of diffuse interstellar bands (DIBs) induced by the Milky Way. We show that, after carefully removing the intrinsic spectral energy distribution of each source, it is possible to measure statistical flux fluctuations at the 1e-3 level, detect more than 20 DIBs and measure their strength as a function of position on the sky. We create a map of DIB absorption covering about 5000 square degrees and measure correlations with various tracers of the interstellar medium: atomic and molecular hydrogen, dust and polycyclic aromatic hydrocarbons (PAHs). After recovering known correlations, we show that each DIB has a different dependence on atomic and molecular hydrogen: while they are all positively correlated with N(HI), they exhibit a range of behaviours with N(H2) showing positive, negative or no correlation. We show that a simple parametrization involving only N(HI) and N(H2) applied to all the DIBs is sufficient to reproduce a large collection of observational results reported in the literature: it allows us to naturally describe the relations between DIB strength and dust reddening (including the so-called skin effect), the related scatter, DIB pair-wise correlations & families, the affinity for $\sigma/\zeta$-type environments and other correlations related to molecules. Our approach allows us to characterize DIB dependencies in a simple manner and provides us with a metric to characterize the similarity between different DIBs.