Large Scale Structure at 24 Microns in the SWIRE Survey

We present initial results of galaxy clustering at 24um by analyzing statistics of the projected galaxy distribution from counts-in-cells. This study focuses on the ELAIS-N1 SWIRE field. The sample covers ~5.9 square-degrees and contains 24,715 sources detected at 24um to a 5.6-sigma limit of 250 micro-Jy (in the lowest coverage regions). We have explored clustering as a function of 3.6 - 24um color and 24um flux density using angular-averaged two-point correlation functions derived from the variance of counts-in-cells on scales 0.05-0.7 degrees. Using a power-law parameterization, $w_{2}(\theta)=A(\theta/deg)^{1-\gamma}$, we find [A,\gamma] = [(5.43\pm0.20)\times10^{-4},2.01\pm0.02] for the full sample (1-sigma errors throughout). We have inverted Limber's equation and estimated a spatial correlation length of $r_{0}=3.32\pm0.19 h^{-1}$Mpc for the full sample, assuming stable clustering and a redshift model consistent with observed 24um counts. We also find that blue [f(24)/f(3.6)<=5.5] and red [f(24)/f(3.6)>=6.5] galaxies have the lowest and highest $r_{0}$ values respectively, implying that redder galaxies are more clustered (by a factor of ~3 on scales >0.2 degree). Overall, the clustering estimates are smaller than those derived from optical surveys, but in agreement with results from IRAS and ISO in the mid-infrared. This extends the notion to higher redshifts that infrared selected surveys show weaker clustering than optical surveys.