Disentangling Drivers of Disk Warps in Tilted and Tumbling TNG50 Halos

Dark matter (DM) halos in $Λ$ Cold DM cosmological simulations are triaxial. Most exhibit figure rotation. We study 40 isolated halos with stellar disks from the TNG50 simulation suite across $\sim 4$~Gyr to understand whether and how a triaxial halo’s tumbling and orientation relative to the disk can drive warps. We measure a warp angle $ψ$ and find even our isolated disks are all at least slightly warped, with each galaxy’s maximum $ψ> 1.8^{\circ}$. We perform a modified cross-correlation analysis between $ψ$ and the figure rotation pattern speed, as well as the misalignment between the disk spin axis and (a) the figure rotation axis, (b) the halo minor axis, and (c) the gas angular momentum axis. We use snapshots spanning a lookback time $t_{lb} ~4$ Gyr with 25 linearly-spaced lags from $ 0 - 2.33$ Gyr. We do not find evidence for a consistent lag between the onset of a warp and any of the aforementioned factors on the population level. However, we find significant correlations between individual time-series at various lags. These maximum correlation coefficients were significantly offset from random chance at the population level, suggesting that several of these factors do correlate with disk warping in specific situations. By examining four case studies whose maximum correlation coefficients were significantly higher than random chance, we establish clear qualitative relationships between these factors and warps. While a non-warped galaxy typically shows minimal halo tilt and figure rotation, warped galaxies can have strong/weak tilts and/or strong/weak figure rotation. Keywords: Disk galaxies(391), Galaxy dynamics(591), Hydrodynamical simulations(767), Galaxy DM halos(1880)

💡 Research Summary

In this paper the authors investigate whether the triaxial shape, figure rotation, and orientation of dark‑matter (DM) halos can drive warps in the stellar disks of isolated galaxies. Using the highest‑resolution TNG50 cosmological magnetohydrodynamic simulation, they select 40 galaxies that host well‑defined stellar disks and have remained relatively isolated over the past ~4 Gyr. The sample spans halo virial masses from 10¹⁰ to 10¹³ M⊙ and stellar masses from 10⁸ to 10¹⁰ M⊙.

The methodology is built on three pillars. First, halo shapes are quantified with the standard shape‑tensor approach, yielding axis ratios (c/a, b/a) and derived ellipticity e and prolateness p. The authors adopt semi‑analytic Poisson‑noise‑driven uncertainty estimates for the major and minor axis orientations (σₓ, σ_z) as calibrated in Ash & Valluri (2023). Second, figure rotation is measured by tracking the orientation of the halo’s principal axes between successive snapshots using the quaternion method of Bailin & Steinmetz (2004). This yields a pattern speed Ωₚ and a rotation axis for each halo at a temporal resolution of ~0.1 Gyr, allowing the construction of cross‑correlation functions with up to 25 linearly spaced lags (0–2.33 Gyr). Third, the warp of each stellar disk is quantified with a novel azimuthal harmonic analysis. For each radial bin the authors compute the m = 1 Fourier amplitude z(R, 1) of the vertical displacement of star particles relative to the inner disk plane. In the radial range 0.5 r₁/₂ ≤ R ≤ 2 r₁/₂ the amplitude is approximately linear; a straight‑line fit provides the warp angle ψ. Uncertainties on ψ and all angular measurements are obtained via 100 bootstrap realizations (sampling 90 % of the particles with replacement).

The authors then perform a modified cross‑correlation analysis between ψ(t) and four candidate drivers: (a) the figure‑rotation pattern speed Ωₚ(t), (b) the misalignment between the disk spin axis and the halo’s minor axis (the “halo tilt”), (c) the misalignment between the disk spin axis and the halo’s major axis, and (d) the misalignment between the disk spin axis and the angular momentum of the gas within 10 r₁/₂. For each galaxy they compute the correlation coefficient r(τ) as a function of lag τ and identify the maximum absolute correlation and its associated lag. To assess statistical significance, they generate 1 000 randomised ψ time‑series (preserving the autocorrelation structure) and compare the distribution of maximum r values to those obtained from the real data.

At the population level, no single lag emerges as a universal predictor of warp onset; the distribution of peak lags is broad and consistent with random chance. However, when examined on a galaxy‑by‑galaxy basis, many systems exhibit a statistically significant peak correlation (p < 0.05) at a specific lag. The magnitude of these peaks is systematically higher than those from the randomised control, indicating that at least one of the examined factors is physically linked to warp evolution in individual cases.

To illustrate the diversity of behaviours, the paper presents four detailed case studies selected for having the strongest, most significant correlations.

1. Strong rotation & strong tilt – Halo 1292 shows Ωₚ ≈ 0.5 Gyr⁻¹ and a halo‑disk tilt of ≈ 30°, with ψ rising sharply after a lag of ~0.8 Gyr, suggesting a direct torque from the tumbling halo.
2. Weak rotation & strong tilt – Halo 1456 has a modest Ωₚ ≈ 0.1 Gyr⁻¹ but a tilt that grows from 10° to 25° over 1 Gyr; ψ follows the tilt increase with a lag of ~0.5 Gyr, implying that the changing halo orientation alone can excite a warp.
3. Strong rotation & weak tilt – Halo 1123 exhibits a high Ωₚ but a nearly aligned halo (tilt ≈ 5°); ψ remains low, indicating that rotation without a significant misalignment exerts only a weak torque on the disk.
4. Weak rotation & weak tilt, but strong gas misalignment – Halo 1589 shows little halo tumbling and a small tilt, yet the gas angular momentum vector swings by ≈ 20° over 0.6 Gyr; ψ responds with a comparable lag, highlighting that misaligned gas accretion can also drive warps even when the halo is dynamically quiet.

These examples demonstrate that disk warps can arise from a range of dynamical configurations: (i) combined strong figure rotation and large halo tilt, (ii) dominant halo tilt with modest rotation, (iii) strong rotation but near‑alignment, and (iv) gas‑accretion‑induced torques in otherwise quiescent halos. The authors therefore conclude that warps in isolated galaxies are not attributable to a single universal mechanism; instead, they reflect the interplay of halo tumbling, halo‑disk misalignment, and the angular momentum of newly accreted gas.

The broader implication is two‑fold. First, the ubiquity of modest warps (all 40 galaxies have ψ > 1.8°) suggests that even low‑level torques from a tumbling triaxial halo are sufficient to maintain an S‑type warp over gigayear timescales. Second, because the presence and amplitude of a warp encode information about the underlying halo dynamics, observational measurements of warp geometry in truly isolated galaxies could serve as indirect probes of dark‑matter halo figure rotation—a property otherwise inaccessible to direct observation. The paper thus opens a pathway for using disk morphology as a diagnostic of the hidden dynamical state of DM halos in the ΛCDM universe.