Probing the spectral shape of dust emission with the DustPedia galaxy sample

The objective of this paper is to understand the variance of the far-infrared (FIR) spectral energy distribution (SED) of the DustPedia galaxies, and its link with the stellar and dust properties. An interesting aspect of the dust emission is the inferred FIR colours which could inform us about the dust content of galaxies, and how it varies with the physical conditions within galaxies. However, the inherent complexity of dust grains as well as the variety of physical properties depending on dust, hinder our ability to utilize their maximum potential. We use principal component analysis (PCA) to explore new hidden correlations with many relevant physical properties such as the dust luminosity, dust temperature, dust mass, bolometric luminosity, star formation rate (SFR), stellar mass, specific SFR, dust-to-stellar mass ratio, the fraction of absorbed stellar luminosity by dust (fabs), and metallicity. We find that 95 per cent of the variance in our sample can be described by two principal components (PCs). The first component controls the wavelength of the peak of the SED, while the second characterizes the width. The physical quantities that correlate better with the coefficients of the first two PCs, and thus control the shape of the FIR SED are the dust temperature, the dust luminosity, the SFR, and fabs. Finally, we find a weak tendency for low-metallicity galaxies to have warmer and broader SEDs, while on the other hand high-metallicity galaxies have FIR SEDs that are colder and narrower.