Cross-correlating WMAP5 with 1.5 million LRGs: a new test for the ISW effect

We present the cross-correlation of the density map of LRGs and the temperature fluctuation in the CMB as measured by the WMAP5 observations. The LRG samples were extracted from imaging data of the SDSS based on two previous spectroscopic redshift surveys, the SDSS-LRG and the 2SLAQ surveys at average redshifts z~0.35 and z~0.55. In addition we have added a higher-redshift photometric LRG sample based on the selection of the AAOmega LRG redshift survey at z~0.7. The total LRG sample thus comprises 1.5 million galaxies, sampling a redshift range of 0.2 < z < 0.9 over ~7600 square degrees of the sky, probing a total cosmic volume of ~5.5 h^{-3} Gpc^3. We find that the new LRG sample at z~0.7 shows very little positive evidence for the ISW effect. Indeed, the cross-correlation is negative out to ~1 deg. The standard LCDM model is rejected at ~2-3% significance by the new LRG data. We then performed a new test on the robustness of the LRG ISW detections at z~0.35 and 0.55. We made 8 rotations through 360deg of the CMB maps with respect to the LRG samples around the galactic pole. We find that in both cases there are stronger effects at angles other than zero. This implies that the z~0.35 and 0.55 ISW detections may still be subject to systematic errors which combined with the known sizeable statistical errors may leave these ISW detections looking unreliable. We have further made the rotation test on several other samples where ISW detections have been claimed and find that they also show peaks when rotated. We conclude that in the samples we have tested the ISW effect may be absent and we argue that this result may not be in contradiction with previous results.