Separating intrinsic alignment and galaxy-galaxy lensing

TL;DR

This paper introduces a new, self-consistent method to measure and constrain intrinsic galaxy alignments in galaxy-galaxy lensing studies, showing that IA contamination is likely below 10%, thus not significantly affecting future surveys.

Contribution

The authors develop an improved, self-consistent technique for measuring intrinsic alignment contamination in galaxy-galaxy lensing, applicable with photometric redshifts, and provide tight constraints on IA effects.

Findings

Intrinsic alignment signal is consistent with zero.

IA contamination is constrained below 10% for certain scales.

More stringent photo-z cuts reduce IA contamination to 1-2%.

Abstract

The coherent physical alignment of galaxies is an important systematic for gravitational lensing studies as well as a probe of the physical mechanisms involved in galaxy formation and evolution. We develop a formalism for treating this intrinsic alignment (IA) in the context of galaxy-galaxy lensing and present an improved method for measuring IA contamination, which can arise when sources physically associated with the lens are placed behind the lens due to photometric redshift scatter. We apply the technique to recent Sloan Digital Sky Survey (SDSS) measurements of Luminous Red Galaxy lenses and typical (L*) source galaxies with photometric redshifts selected from the SDSS imaging data. Compared to previous measurements, this method has the advantage of being fully self-consistent in its treatment of the IA and lensing signals, solving for the two simultaneously. We find an IA signal consistent with zero, placing tight constraints on both the magnitude of the IA effect and its potential contamination to the lensing signal. While these constraints depend on source selection and redshift quality, the method can be applied to any measurement that uses photometric redshifts. We obtain a model-independent upper-limit of roughly 10% IA contamination for projected separations of approximately 0.1-100 Mpc/h. With more stringent photo-z cuts and reasonable assumptions about the physics of intrinsic alignments, this upper limit is reduced to 1-2%. These limits are well below the statistical error of the current lensing measurements. Our results suggest that IA will not present intractable challenges to the next generation of galaxy-galaxy lensing experiments, and the methods presented here should continue to aid in our understanding of alignment processes and in the removal of IA from the lensing signal.