Where Galaxies Point: First Measurement of the Large-Scale Axial Intrinsic Alignment

TL;DR

This work tests the cosmological principle by searching for large-scale axial intrinsic alignments (LAIA) in DES Y3 data, developing an orientation-field estimator that maps a sky-direction of maximal alignment. They report the first detection of LAIA on horizon-scale dipolar angular patterns, with bulge-dominated semi-major axes and disk-dominated semi-minor axes pointing toward a common direction and exhibiting a hierarchical amplitude consistent with tidal-torquing. The signal persists across redshift and sky-splits, passes extensive systematic checks, and is unlikely to originate from survey artifacts, indicating a potential imprint of a horizon-scale tidal field seeded by primordial inhomogeneities. This opens a new avenue—galactic cosmology—where galaxy orientation fields complement number counts and shear to probe isotropy, primordial physics, and the large-scale structure of the Universe.

Abstract

Applying a new estimator to the Dark Energy Survey (DES) Y3 weak lensing shape catalog, we map the galaxies' orientation field and report the first detection of a large-scale axial intrinsic alignment (LAIA) on dipolar angular scales. Ellipticals' semi-major axes and spirals' semi-minor axes coherently point toward a common direction, (RA,Dec)=$(306^{+6}_{-4},52^{+3}_{-3})^\circ$ and $(296^{+10}_{-18},50^{+6}_{-2})^\circ$, respectively, with amplitudes in the expected tidal-torquing hierarchy. This pattern could be produced by a horizon-scale tidal field imprinted by primordial inhomogeneities during galaxy assembly and, if confirmed, would signal a statistically significant preferred direction in the Universe, thereby probing deviations from statistical isotropy. The signal persists across spatial and redshift splits and is difficult to attribute to survey systematics. LAIA offers a new sky-wide compass linking galaxy evolution and cosmology.

Figures (6)

• Figure 1: Left) Spiral galaxies tend to align their angular-momentum vector $\mathbf{L}$ with the tidal field. The projection of $\mathbf{L}$ onto the image plane corresponds to the observed minor axis (b) direction. Right) By contrast, elliptical galaxies typically align their major axes (a). Thus, under the same tidal field, spirals and ellipticals become orthogonal on average.
• Figure 2: Left) Mollweide projection map in equatorial coordinates of the disk-dominated sample (DD), i.e. late-type (spiral), density and footprint considering HEALPixNSIDE=128 resolution. Right) The same for the bulge-dominated sample (BD), i.e. early-type (elliptical). Footprint differences reflect the different sky coverage of the classification catalogs DES:2020tkt2021MNRAS.507.4425C.
• Figure 3: Photometric redshift distribution of each sample.
• Figure 4: Final results. Highest-density interval (HDI) contours (68.27% and 95.45%) for the LAIA direction $\hat{\mathbf d}_{\rm IA}$ in equatorial coordinates $(\alpha,\delta)$ for the full bulge-dominated (BD; red) and disk-dominated (DD; blue) samples. Markers show the HDI peaks for splits by redshift $z$, right ascension $\alpha$, and declination $\delta$. The legend lists the measured $\theta_{\rm IA}$ (arcminutes, ${}^{\prime}$) for the full samples and each split. Both morphologies indicate a common preferred direction, bulge major axes and disk minor axes, consistent with a large-scale tidal field, with the disk signal smaller than the bulge signal as expected from tidal torquing. The panel is restricted to the quadrant where the contours concentrate to improve readability. The antipodal direction $(\alpha+180^\circ,\,-\delta)$ is physically equivalent for this axial estimator.
• Figure 5: Distribution of position-angle differences, $\delta\theta$, for a sample of 10,000 galaxies, comparing Morfometryka (MFMTK), metacalibration (MC), and visual inspection. Vertical lines indicate the 68.27% ($1\sigma$) upper bound for each distribution. MFMTK and visual inspection use only the $I$ band, whereas metacalibration combines the $R$, $I$, and $Z$ bands to estimate the PA.