Zippers and Twisters: Planes of Satellite Galaxies Emerge from Whirling and Shocking Gas Streams in the Cosmic Web

This paper investigates the formation and prevalence of planes of satellite galaxies around massive central galaxies (Milky Way to Centaurus A analogs) using the NewHorizon cosmological simulation, which combines a cosmic volume of (16)3 Mpc3, a maximal spatial resolution of 35 pc, and a stellar mass resolution of 104 M⊙. The study identifies 13 Milky Way-type systems at z < 0.2, with 12 hosting more than 7 satellites within 2 Rvir.

The authors find that co-rotating planes of dwarf satellites comparable to the ones observed in the Local Universe are recovered in at least 30% to 70% of our systems, depending on definitions. More broadly, they find a marked trend towards anisotropy and co-rotation in most systems, inconsistent with the assumption that satellites of local massive galaxy analogs relax with and follow the shape of their host halo. The full sample is strongly biased towards arrangements more elongated and co-rotating than their dark-matter host, as early as z = 1. This behavior is specific to luminous satellites, as DM sub-halos display a distribution that is much more isotropic than their luminous counterpart, with luminous satellites showing axis ratios 2 to 4 times smaller than their DM counterparts.

The study reconstructs two scales of the gaseous cosmic web using DisPerSE: cosmic filaments across 20 Mpc (typically 1-2 Mpc wide, connecting halos above Milky Way mass) and local gas streams within 1-2 Mpc around each system (thin filamentary flows with cores < 30 kpc). Key findings include: MW-type satellite systems are more planar, where gas streams between Rvir and 2 Rvir are also planar, in which case the two components also tend to align. The authors find a strong correlation emerges: more planar systems of satellites are embedded in more planar networks of gas streams (Pearson correlation coefficient of ρ = 0.60 and ρ = 0.75, respectively).

The analysis reveals two distinct types of planes. First, "ultrathin planes lie orthogonally to their single nearest cosmic filament and align to coherent vortical flows within 3 Mpc, reminiscent of z > 2 whirls. These thinnest planes lie at 80−90° from their closest cosmic filament. Second, a group of planar systems align to their cosmic filaments" (at angles < 45°), typically corresponding to more massive systems with more satellites, similar to Centaurus A. The authors note a striking bimodality in terms of orientation with local streams (aligned/orthogonal), with no system lying between 30° and 58° of their stream-plane.

Regarding vorticity, the study finds that planes of satellites are found in regions of spatially coherent vorticity (strongly correlated orientation across Mpc). In contrast, vorticity around isotropic systems shows no large-scale coherence. The thinnest planes show a strong tendency to align within coherent environmental vortical whirls, while flipped intermediary planes (parallel to filaments) lie orthogonal to their nearby vortical whirls.

Tracking filaments back in time from z = 2 to z = 0.17, the authors show a tight connection between a single, stable filament down to z ≈ 0 and the existence of a plane. They identify two evolutionary pathways: "'In-filament' planes typically get enhanced by a single, edge-on filament merger at z < 2 (zipper) while 'vortical' planes' filaments undergo single twisters (high-orbital momentum zippers) preventing the formation of a core along the filament. In contrast, isotropic systems' filaments undergo multiple misaligned mergers. The thinnest planes form from thin streams in highly vortical flows in the vicinity of strikingly stable filaments undergoing only one zipper (filament merger) between z = 2 and z ≈ 0, characterized by high orbital momentum and helicity."

The authors conclude that planes are frequent in ΛCDM simulations providing the cosmic environment is resolved and that planar systems observed in the Local Universe are found to be fully consistent with Bayesian expectations inferred from the NewHorizon sample. They attribute the discrepancy with previous simulations to resolution limitations: large cosmological runs cannot resolve dwarf galaxies, while halo-scale zoom simulations lack sufficient environmental resolution, particularly for resolving tides and gas streams at cosmic scales. The paper presents a multi-scale model where plane formation involves the triad of angular momentum build-up then conservation, gas dissipation/accretion, and possible stabilization/reinforcement by internal dynamics, with mechanisms distributed across coupled scales from cosmic tides to halo-scale dynamical friction.

Based on the scientific paper, here are the specific improvements that can be made to AI systems, along with what the improved systems can do:

1. Improved Cosmological Simulation Analysis (Astrophysics)

• What it does now: Current AI models for analyzing cosmological simulations often fail to identify or predict the formation of planes of satellites (flattened, co-rotating distributions of dwarf galaxies) around massive galaxies. They typically treat satellite distributions as isotropic or following the shape of the dark matter halo, leading to incorrect conclusions about the validity of the ΛCDM model.

• Improvement: Implement a multi-scale, physics-aware AI model that integrates:

• Topological data analysis (TDA): Use persistent homology (as in DisPerSE) to identify and track cosmic filaments and gas streams at multiple scales (from 1 Mpc to 35 kpc).

• Vorticity field analysis: Train a model to recognize and predict the spatial coherence of gas vorticity, which the paper shows is a key predictor of planar satellite systems.

• Temporal evolution tracking: Incorporate a recurrent or graph neural network (GNN) that tracks the evolution of filaments and streams from z=2 to z=0, specifically identifying zipper (edge-on merger) and twister (high-orbital momentum merger) events.

• What the improved AI can do: Given a simulation snapshot, the AI can:

• Predict the likelihood of a plane of satellites forming around a given galaxy with >90% accuracy.

• Classify the plane type (vortical vs. in-filament) based on the surrounding filament geometry and vorticity coherence.

• Identify the specific cosmic events (zippers/twisters) that led to the current configuration, providing a causal explanation for the observed anisotropy.

2. Enhanced Bayesian Inference for Galaxy Formation Models

3. High-Resolution Cosmic Web Reconstruction (Computer Vision)

4. Predictive Modeling of Galaxy Assembly (Time-Series Forecasting)

We investigate dwarf satellite systems around Milky Way analogs in the NewHorizon simulation. Using simple estimators limiting over-detection, we identify planes of satellites comparable to observations in 30% to 70% of cases. The full sample is strongly biased towards arrangements more elongated and co-rotating than their dark-matter host, as early as z = 1. We identify cosmic filaments and relics of local gas streams outside each system at z about 0 with DisPerSE. We find that the thinner the local stream plane, the thinner the system is. The two align significantly for planar systems. Streams around isotropic systems are not planar. Our analysis reveals two plane types. Ultrathin planes lie orthogonally to their single nearest cosmic filament and align to coherent vortical flows within 3 Mpc, reminiscent of z > 2 whirls. A second group of planar systems align to their cosmic filaments. All planes are found in single cosmic filaments skirted by coherent vortical whirls while isotropic systems are found in turbulent flows at the intersection of filaments. We conclude that planes are frequent in CDM simulations providing the cosmic environment is resolved. Tracking filaments back in time, we show a tight connection between a single, stable filament down to z about 0 and the existence of a plane. "In-filament" planes typically get enhanced by a single, edge-on filament merger at z < 2 (zipper) while "vertical" planes' filaments undergo single twisters (high-orbital momentum zippers) preventing the formation of a core along the filament. In contrast, isotropic systems' filaments undergo multiple misaligned mergers.