Radio--Far infrared correlation in "blue cloud" galaxies with 0<z<1.2

We study the radio--far infrared (FIR) correlation in "blue cloud" galaxies chosen from the PRism MUltiobject Survey (PRIMUS) up to redshift ($z$) of 1.2 in the XMM-LSS field. We use rest-frame emission at 1.4 GHz in the radio and both monochromatic (at 70$μ$m) and bolometric (between $8-1000~μ$m) emission in the FIR. To probe the nature of the correlation up to $z\sim1.2$, where direct detection of blue star-forming galaxies is impossible with current technology, we employ the technique of image stacking at 0.325 and 1.4 GHz in the radio and in six infrared bands, viz. 24, 70, 160, 250, 350 and $500~μ$m. For comparison, we also study the correlation for more luminous galaxies that are directly detected. The stacking analysis allows us to probe the radio--FIR correlation for galaxies that are up to 2 orders of magnitude fainter than the ones detected directly. The $k-$correction in the infrared wavebands is obtained by fitting the observed spectral energy distribution (SED) with a composite mid-IR power law and a single temperature greybody model. We find that the radio luminosity at 1.4 GHz ($L_{\rm 1.4GHz}$) is strongly correlated with monochromatic FIR luminosity at 70 $μ$m ($L_{\rm 70μm}$) having slope $1.09\pm0.05$ and with bolometric luminosity ($L_{\rm TIR}$) having slope $1.11\pm0.04$. The quantity $q_{\rm TIR} (=\log_{10}[L_{\rm TIR}/(3.75\times 10^{12} L_{\rm 1.4 GHz})])$ is observed to decrease with redshift as $q_{\rm TIR} \propto (1+z)^{-0.16\pm0.03}$ probably caused due to the non-linear slope of the radio--FIR correlation. Within the uncertainties of our measurement and the limitations of our flux-limited and color-selected sample, we do not find any evolution of the radio--FIR correlation with redshift.