Measurement of intrinsic alignments in galaxy ellipticities

TL;DR

This study analyzes intrinsic alignments of galaxy ellipticities in the local universe using $2$–$3$ million galaxies from the SuperCOSMOS Sky Survey, spanning 1′–100′ scales. By correcting for PSF anisotropy and seeing and applying a minimum-variance estimator, the authors separate intrinsic alignment contributions from lensing, finding an intrinsic signal that is about $10^2$ times larger than the weak-lensing expectation at $z\approx0.1$. The cross-band ($b_J$ and $R$) and cross-field consistency, plus external APM cross-checks, indicate the signal is real and dominated by intrinsic alignments rather than systematics. The results align with theoretical predictions for intrinsic alignments and have important implications for shallow cosmic shear surveys, offering potential constraints on galaxy formation models and local mass distribution, and guiding future high-redshift lensing analyses where the intrinsic term is less dominant.

Abstract

We measure the alignment of galaxy ellipticities in the local universe over a range of scales using digitized photographic data from the SuperCOSMOS Sky Survey. We find for a magnitude cut of b_J < 20.5, corresponding to a median galaxy redshift of z = 0.1, and 2x10^6 galaxies, that the galaxy ellipticities exhibit a non-zero correlation over a range of scales between 1 and 100 arcminutes. In particular, we measure the variance of mean galaxy ellipticities, sg^2(theta), in square angular cells on the sky as a function of cell size and find it lies in the range, 2 x 10^{-4} > sg^2(theta) > 1 x 10^{-5} for cell side lengths between 15 < theta < 100 arcminutes. Considering the low median redshift of the galaxies in the sample and hence the relatively low effective cross-section for lensing of these galaxies by the large-scale structure of the Universe, we propose that we have detected an intrinsic alignment of galaxy ellipticities. We compare our results to recent analytical and numerical predictions made for the intrinsic galaxy alignment and find good agreement. We discuss the importance of these results for measuring cosmic shear from upcoming shallow surveys (e.g. Sloan Digital Sky Survey) and we outline how these measurements could possibly be used to constrain models of galaxy formation and/or measure the mass distribution in the local universe.

Figures (16)

• Figure 1: Ellipticity fields for raw (LHS) and corrected (RHS) stars for three of the b$_{\rm J}$-band SuperCOSMOS fields. Stars are binned into cells of a side $10'$ and smoothed with a Gaussian with a smoothing scale $15'$. For each plot, the length of each vector drawn is $25$ cm times the measured cell ellipticity. The average ellipticity in a cell in the raw fields is $\bar{e} \approx 10^{-2}$, while the average ellipticity in a corrected cell is $\bar{e}\approx 10^{-4}$. These plots are typical for the PSF anisotropy distributions on a plate and the corresponding R-band plots are generally very similar.
• Figure 2: Ellipticity distribution per cell (stars on the left, galaxies on the right) for the same three b$_{\rm J}$-band fields shown in figure 1. The dots are the raw stars/galaxies, the crosses the corrected values. Each cell has an ellipticity estimated from equation (1) and has an uncertainty given by equation (\ref{['error']}). Average stellar distortions in the raw fields of $\bar{e}_1 \sim \bar{e}_2 \approx 10^{-2}$ are corrected to produce flattened stellar ellipticity distributions with residual ellipticities, $\bar{e}_1 \sim \bar{e}_2 \approx 10^{-4}$. Similar "before" and "after" distributions are found for the R plates.
• Figure 3: The linear correlation coefficients, $C_1$ and $C_2$ of the cell-averaged stellar and galaxy ellipticity components as described in the text, after correction for PSF anisotropy. The shaded region is the $3\sigma$ significance level.
• Figure 4: The observed anti-correlation of corrected galaxy cell ellipticity components ($e_1$ on top, $e_2$ on bottom) with the corresponding observed stellar ellipticities.
• Figure 5: $C_1$, $C_2$ and the $3\sigma$ significance level after excluding all galaxies with sizes, $\theta_g < 0.8 \, \langle \theta_g \rangle$ from the catalogue.