Our galaxy has been described as roughly the shape of a pizza and the Andromeda galaxy, though I've never heard it described that way, appears similarly disk shaped. Spiral galaxies make up (roughly) about 60% of the observed galaxies in the Universe, so there must be a standard process that makes this happen.

I don't believe that collisions, and flattening around the axis of rotation would work, which is how the solar-system flattened mostly into an orbital plane. Space is too empty for that, unless the galaxy begins as a lot of dust and gas, then . . . maybe?

Is it gravitational interactions? Is it the massive black hole/quasar in the center that drives the flattening?

I'm curious if this question has been answered by observation of young galaxies, because it seems unlikely to me that it would be collisions like it was with the solar system or with accretion disks. Am I wrong in that assumption?


1 Answer 1


The Galaxy is disc shaped because most of the stars formed in a disc. They formed in a disc because most of the gas was in that disc.

If you have a collapsing cloud of gas then as it collapses it becomes turbulent and heats up. As a result, it loses energy through radiation. However, whilst that radiation carries away energy, it carries little angular momentum. Thus, as the cloud collapses its energy decreases but its angular momentum stays the same.

This process leads to a disc-like geometry - a geometry that (approximately) minimises energy for a given angular momentum. Most of the star formation then takes place in this compressed gas disc and we end up with a disc-like galaxy of stars. A similar process is the reason that the Solar System has a planar geometry.

Some people are happier thinking about this in terms of forces. If you go to the frame of reference of a parcel of gas$^{*}$, then the vertical structure (here, vertical means parallel to the spin axis defined by the angular momentum) and height are governed by the balance between gravitational force and gas pressure. However, in the direction perpendicular to the spin axis, there is gravitational force and opposing gas pressure as before, but also a centrifugal force acting outwards. It is this that ensures the gas cloud is extended into a disc-like structure.

The above explanation describes a simple, monolithic collapse. Many galaxies, including the Milky Way, are built up through mergers. However, the principle is the same; if the merger is "gas-rich", then the gas will settle into a disc, and then subsequently stars are born in that disc.

Not all stars are distributed in a disc. The Galactic halo stars are more spherically distributed. These are older stars and mostly metal-poor. They are the initial stars formed before the gas had collapsed into a disc-like geometry or stars that already existed in preceding, smaller galaxies that merged with the Milky Way. They stay in a roughly spherical distribution because the dissipative processes that act in a gas cloud are ineffective in a "collisionless" system of stars.

$^*$ NB You consider a parcel of gas as a whole, not the individual "orbits" of gas molecules because the mean free path of a gas molecule (before it collides with another) is very short compared with an orbit. The opposite is true for stars.

  • 1
    $\begingroup$ It feels like there's parts missing here. So you have this random sphere of warm gas, and it starts radiating energy, ... and then what? Dr. Becky throws up a "several eternities later" slide and suddenly everything's magically in a flat disk? How did the gas molecule that was orbiting over the top realize it needed to orbit around the middle, and figure out where the middle is? -- Is it collisions? I'm unsure, because the question doubts it's collisions and you never explicitly discuss whether that belief is right or wrong. $\endgroup$
    – R.M.
    Commented Oct 18, 2023 at 11:56
  • 4
    $\begingroup$ @R.M. Cloud radiates energy, cloud contracts, cloud gets hotter, gas molecules move faster and collide more, more energy gets lost. Gas molecules do not simply orbit in the gravitational potential whilst never colliding. If they did they would behave like collisionless halo stars. The timescale for this process is nowhere near "eternity", it is rather quick on cosmological timescales - tens of millions of years I would say. $\endgroup$
    – ProfRob
    Commented Oct 18, 2023 at 12:46
  • 2
    $\begingroup$ Worth saying that a great number of stars behave as a gas as well. While they rarely collide, they do scatter gravitationally and from a distance this looks like an ellastic collision. $\endgroup$
    – fraxinus
    Commented Oct 18, 2023 at 18:54
  • $\begingroup$ It would perhaps help if an estimate of the extend of the initial cloud (one particle per cubic meter counts as "dense"?) vs. the resulting galaxy could be given. As I read it, the galaxy should approximately be a dot when seen from the initial gas cloud, that is why random imbalances in angular momentum can become a substantial spin in the condensed state. $\endgroup$ Commented Oct 19, 2023 at 7:47

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .