Bars disrupt normal galaxy rotation

Typical disk galaxies have well-ordered rotation and are axially symmetric since most of their stellar motion is in the tangential direction. Their rotation can be easily modeled with a parametric model using any number of common rotation curve models, or with a nonparametric model that describes the specific shape of the galaxy's rotation curve.

Bars disrupt this orderly rotation. Bars are linear features at the center of a galaxy with bilateral symmetry. According to dynamical theory and models, the stars in galactic bars move radially in and out near the galactic center. The galaxy velocity field on the right has a significant distortion in its normal rotation due to the motions in its bar.

Because of this irregular motion, traditional axisymmetric velocity field models often do not describe the velocity fields of barred galaxies, so different models should be applied to them.

Nirvana fits bisymmetric distortions

I developed a velocity field fitting code called Nirvana (Nonaxisymmetric Irregular Rotational Velocity Analysis) that can more accurately describe bisymmetric motions in galaxies. The Nirvana model of the above barred galaxy is shown on the left.

Based on a second-order Fourier model from Spekkens & Sellwood (2007), Nirvana uses the nested sampling package dynesty to fit a nonparametric Bayesian velocity model to galactic rotational data. The code is optimized to be galaxy agnostic despite the highly degenerate and multimodal likelihood space. It also accounts for observation quality by taking into account velocity dispersion, PSF smearing, and observational errors.

Picking apart velocity modes

Nirvana models galaxy rotation by splitting the galaxy's rotation into three components: the first-order tangential component that represents the dominant rotational component and the second-order radial and tangential components that come from the bar. These three components are shown separately here.

I also don't assume any functional form for the galaxy's rotation, allowing it to describe the data in a more flexible model agnostic manner. Nirvana reproduces typical rotation curves on its own and generates consistent bar rotation curves as well.

A catalog of >1000 barred galaxies

I ran Nirvana on all of the barred galaxies in the MaNGA survey, resulting in models of over 1000 barred galaxies, as well as a matched control sample of similar but non-barred galaxies. Nirvana finds systematically higher second order rotational modes in the barred galaxies, and elevated bisymmetric motions are preferentially aligned with visually identified bars, verifying theoretical models on the origins of bars.

Future Projects

  • What do the stellar populations of bars look like? Previous studies find that bar stellar populations are older than the surrounding disk, but that the gradients in stellar populations are not as steep due to the increased radial mixing within the bar. Is radial motion correlated with stellar population gradients? I began investigating this question in my dissertation and found early evidence of a correlation
  • What is a good model for bar rotation curves? At present, there is no accepted kinematic model for the rotation curves of radial flows within bars. With Nirvana, we have a large enough sample of nonparametric bar rotation curves that we could statistically test different rotation curve models. Can future fitting be improved using simpler models?
  • How do bars change as galaxies evolve? JWST is obtaining resolved spectroscopy for galaxies at high redshifts, showing snapshots of galaxy evolution over the life of the Universe. Prior studies suggest that bar fraction has continually increased over time, so how were early bars similar to/different from modern bars?
  • Can we detect outflows with Nirvana? In its present form, Nirvana assumes no inflows or outflows, but the model could be modified to look for AGNs or other galaxies that violate the continuity equation. Can Nirvana detect outflows that are otherwise obscured?
  • Is Tully-Fisher Relation scatter dependent on second-order motions? Past studies have found some correlation between galaxy asymmetry and scatter off of the Tully-Fisher Relation. Is this scatter due to suboptimal rotational models poorly describing asymptotic velocity? Or do nonaxisymmetric galaxies follow a different velocity-size relation? Colby undergraduate Nico Flota-Sanchez is investigating this question.