Research highlights

Galaxies keep breaking their arms

Since the 1960s, the most accepted and widely reported explanation of spiral arms has been the spiral density wave theory (Lin & Shu, 1964, ApJ 140 646), which spiral arms rotate rigidly with constant angular rotation. This allows stars to move independently and pass through spiral arms, allowing them to be long-lived features.

However, we showed using N-body simulations that individual spiral arms co-rotate with the stars, and wind up with time (Fig. 1). Spiral arms constantly disrupt and form anew, ensuring that a spiral morphology exists at all times, consistent with their ubiquitousness in the Universe.


Fig 1: Snapshots of the face on density of the stars (left) and gas (right) showing the evolution of a spiral arm in an idealised simulation. The density peak of the bar is marked in black, and white crosses mark the position of the spiral arm at various radii at and rotate at the rotational velocity. The spiral arm begins to form at t = 0.994 Gyr, winds up and then is disrupted at t = 1.138 Gyr.

Radial migration – stars surf the spirals

Radial migration is the exchange of angular momentum between a star and a bar/spiral arm at the co-rotation radius. The movement of stars away from their birth places causes stars of different ages and metallicities to mix throughout the disc. Because stars rotate at the same speed as the spiral arm at all radii, not only at a single co-rotation radius, radial migration occurs everywhere spiral structure exists. Tracing the evolution of individual N-body star particles in detail reveals that migrating star particles have particular velocities and orbits that create streaming motions along the spiral arm. Star particles migrating radially outward always trail the spiral arm (Fig. 2), whereas stars that migrate radially inward always lead the spiral arm. This means that radial migration continues until the spiral arm disrupts and is more efficient than previously thought – a direct consequence of the co-rotating spiral arm.


Fig 2: Stellar surface over-density of an idealised disc simulation in cylindrical polar coordinates. The direction of rotation is from right to left, and spiral arms are visible as trailing white regions. Arrows depict the (rotation subtracted) peculiar velocity of stars (the arrow in the lower-left represents 10 km/s).

An interesting further prediction is that the migration of metal poor and metal rich stars along leading and trailing sides of the spirals, respectively, gives rise to azimuthal metallicity patterns – a unique prediction of co-rotating spiral arms. We confirmed these streaming motions and metallicity patterns in the external spiral galaxy NGC 6754 from MUSE data (Sanchez-Menguiano, …, Grand et al. 2016 ApJ, 830L, 40).

Disc formation in cosmological hydrodynamical simulations

In addition to idealised N-body simulations, I have worked  with cosmological zoom simulations of Milky Way analogues. Since 2014, I have lead the Auriga project: a large, high-resolution suite of Milky Way analogues simulated with world-leading numerical simulation techniques and a comprehensive galaxy formation model. A unique advantage of these simulations is that they self-consistently produce rotationally supported, star-forming late-type systems with bars and spiral arms (see Fig. 3) across the expected Milky Way mass range (0.5 – 2 x 1012 M⊙), thereby overcoming a historic difficulty for simulations in the Lambda Cold Dark Matter paradigm.


Fig 3: The face-on projected stellar density at present day of 5 Auriga simulations. The images are a composite projection of the K-, B- and U-band luminosity of stars, which are shown by the red, green and blue colour channels, respectively. Younger (older) star particles are represented by bluer (redder) colours. Bars and spirals are common. The plot dimensions are 50x50x25 kpc.

These simulations have been used to study a wide variety of aspects of galaxy formation, including: HI gas disc properties (Marinacci…Grand et al. 2017, MNRAS 466, 3859); magnetic field evolution (Pakmor… Grand et al. 2017, MNRAS 469, 3185); disc warps (Gomez… Grand et al. 2016 MNRAS 456, 2779, 2017 MNRAS 465, 3446) and the Milky Way stellar halo (Deason… Grand et al. 2017 MNRAS 470, 1259). 

Simulated Milky Way-analogues through the eyes of (Auri)Gaia

Galactic surveys such as Gaia are delivering an unprecedented amount of data for billions of stars in our Milky Way, highlighting out-of-equilibrium dynamics and a wealth of stellar streams and substructure in the stellar halo. Understanding the origin and evolutionary history of these fascinating features of the Galaxy requires the forward modelling of theoretical models to provide much needed interpretation. To this end, I developed the first-ever synthetic Gaia catalogues for the second data release (Aurigaia, Grand et al. 2018, MNRAS, 481, 1726), which provides a “mock” view of billions of stars in 6 different simulated Milky Way-mass galaxies (from 4 solar positions).


Fig. 4: Left: Mock stellar light for all stars within 20 kpc (top) and between 5 and 20 kpc (bottom) of a solar-like position in one of the Auriga galaxies. The x- and y-axis indicate the Right Ascension and Declination coordinates. Dust extinction is heaviest in the mid-plane, and removing stars within 5 kpc of the Sun emphasises the older, yellow bulge in favour of the bright blue nearby disc. Right: Colour-magnitude diagram of a subset of low-extinction stars.

These catalogues are created from a sophisticated method that takes into account dust extinction, the Gaia selection function, and the astrometric, photometric and spectroscopic errors of the survey. I demonstrated that these catalogues provide powerful assessments of the biases and limitations of Gaia: i) B-dwarf stars in the disc trace the vertical structure well out to a few kiloparsecs in the Galactic anticentre, whereas A-dwarf stars begin to overestimate the scale-height owing to extinction; and ii) the rotation of the stellar halo as traced by cepheid variable stars is robust out to large distances.