Alignment of spiral and elliptical galaxies from Siena Galaxy Atlas with filaments

TL;DR

This study tests how galaxy spins align with the cosmic web filaments using the Siena Galaxy Atlas and Bisous-identified filaments from SDSS DR12 (z ≤ 0.2). Elliptical galaxies show a strong perpendicular alignment of their short axes with filament spines, while spiral galaxies display only a weak, otherwise marginal signal that becomes detectable under targeted luminosity and distance cuts. The analysis employs a projection-based estimator of 3D spin normals, deprojection informed by morphology, and a robust nonparametric significance framework with randomized controls, including a Bayesian optimization to identify the most signal-rich subsamples. The findings reinforce tidal torque theory predictions for morphologically distinct systems and highlight the importance of selection effects in uncovering subtle spin–web correlations, with implications for future wide-area surveys mapping filaments at $z<0.2$.

Abstract

The properties of galaxies are known to have been influenced by the large-scale structures that they inhabit. Theory suggests that galaxies acquire angular momentum during the linear stage of structure formation, and hence predict alignments between the spin of halos and the nearby structures of the cosmic web. In this study, we use the largest catalog of galaxies publicly available - the Siena Galaxy Atlas - to study the alignment of the spin normals of elliptical and spiral galaxies with filaments constructed by applying the Bisous process on galaxies ($z \le$ 0.2) from SDSS - DR12. Our sample comprises 32517 disk and 18955 elliptical galaxies that are within 2 Mpc of any filament spine. We find that the spin normals of elliptical galaxies exhibit a strong perpendicular alignment with respect to the orientation of the host filaments, inconsistent with random distributions by up to $\approx$ 13 $σ$. The spin axis of spiral galaxies shows a much weaker but nonzero alignment signal with their host filaments of $ \approx$ 2.8$σ$ when compared with random. These numbers depend on exactly how the significance is measured, as elucidated in the text. Furthermore, the significance of the alignment signal is examined as a function of distance from the filament spine. Spiral galaxies reach a maximum signal between 0.5 and 1 Mpc. elliptical galaxies reach their maximum significance between 0.2 and 0.5 Mpc. We also note that with a tailored selection of galaxies, as a function of both i) distance from the filaments \& as a ii) function of absolute luminosity, the alignment significance can be maximized.

Figures (8)

• Figure 1: Physical properties of the SGA galaxies used in this study. There are a total of 51472 galaxies (32517 spiral galaxies and 18955 elliptical galaxies). The left panel illustrates the (differential) redshift distribution for all galaxies (orange), spiral galaxies (green dotted) and elliptical galaxies (blue dotted). The central panel depicts the absolute magnitude (R-Band) distribution of these galaxies and the right panel illustrates the distribution of galaxy distances from the nearest filament axis (in Mpc).
• Figure 2: Distribution of the absolute magnitude of the cosine of the angle between the line of sight (l.o.s.) and various objects. In dashed green we show the angle between the l.o.s. and spiral galaxy spin normal. In dot dashed blue with show the angle between the l.o.s and the filaments spines used (namely those filaments with galaxies in the SGA). In solid red we show the angle between the l.o.s. and all the filament points from the Bisous process 2016AC....16...17T.
• Figure 3: Normalized histogram of cosines of the angles between spines of the filaments and the spin axes of spiral and elliptical galaxies within 2 Mpc ($\cos \theta = |\hat{c}_{\text{galaxy}} \cdot \hat{L}_{\text{filament}}|$). The upper panels show the raw alignment signal. The dark cyan band represents the 1 $\sigma$ null hypothesis corridor, beyond which the probability density function is deemed significant, while the light cyan band provides a 2$\sigma$ corridor as a visual comparison. The significance of the alignment signal is expressed in terms of the standard deviations (1$\sigma$) from the random distribution of the null hypothesis. The p-value from the KS test is included to quantify the probability of observing the alignment signal from the given null hypothesis. The lower panels present the alignment signal normalized by the mean random signal such that a random distribution would appear uniform in cos $\theta$. An inset plot depicting the Mean of alignment signal (red line) along with the distribution of the mean from the null hypothesis cases is inserted to visualize how much is the mean of the alignment signal is further from the median of the distribution of the mean from the null hypothesis cases (black line), in terms of standard deviation of the distribution ($\sigma_{\langle \cos \theta \rangle}$). The inset plots also includes the $\sigma_{\langle \cos \theta \rangle}$ values for both distribution, denoted in red text
• Figure 4: Alignment signal - Probability density distribution (PDF) of the cosines of the angle between spines of the filaments and the spin axes of spiral galaxies at different intervals of (< 0.2 Mpc), (0.2 - 0.5 Mpc), (0.5 - 1 Mpc) & (1 - 2 Mpc, in red). Dark cyan band shows 1 $\sigma$ null hypothesis corridor beyond which the PDF is considered significant. 2 $\sigma$ band (light cyan) is also provided as a visual comparison for the significance of the PDF. For comparison, the cumulative version for the distance bins (i.e, PDF for all the spiral galaxies that are within the distance range of the upper bound of the distance bin) is superimposed (in blue dashed lines) over the PDF for a given distance interval bin along with their error corridors (in blue dash-dotted lines). Inset plots depicting the Mean of alignment signal (red line: for differential subsets and blue line: for cumulative subsets) along with the distribution of the mean from the null hypothesis cases is inserted to visualize how much is the mean of the alignment signal is further from the median of the distribution of the mean from the null hypothesis cases (black line), in terms of standard deviation of the distribution. The statistical significance of alignment in each subset is quantified and presented in Table \ref{['TABLE1']} (Note: The alignment signal are normalized with the mean of the random signal that the random is based around 1.0 for better inference of the alignment signal, and not under the assumption of null hypothesis to be an uniform distribution (refer Fig. \ref{['Alignment signal - Ellipticals and Spirals']}))
• Figure 5: Alignment signal - Probability density distribution (PDF) of the cosines of the angle between spines of the filaments and the spin axes of elliptical galaxies at different intervals of (< 0.2 Mpc), (0.2 - 0.5 Mpc), (0.5 - 1 Mpc) & (1 - 2 Mpc). Dark cyan band shows 1 $\sigma$ null hypothesis corridor beyond which the PDF is considered significant. 2 $\sigma$ band (light cyan) is also provided as a visual comparison for the significance of the PDF. For comparison, the cumulative version for the distance bins (i.e, PDF for all the elliptical galaxies that are within the distance range of the upper bound of the distance bin) is superimposed (in blue dashed lines) over the PDF for a given distance interval bin along with their error corridors (in blue dash-dotted lines). Inset plots depicting the Mean of alignment signal (red line: for differential subsets and blue line: for cumulative subsets) along with the distribution of the mean from the null hypothesis cases is inserted to visualize how much is the mean of the alignment signal is further from the median of the distribution of the mean from the null hypothesis cases (black line), in terms of standard deviation of the distribution. The statistical significance of alignment in each subset is quantified and presented in Table \ref{['TABLE1']} (Note: The alignment signal are normalized with the mean of the random signal that the random is based around 1.0 for better inference of the alignment signal, and not under the assumption of null hypothesis to be an uniform distribution (refer Fig. \ref{['Alignment signal - Ellipticals and Spirals']}))