6
$\begingroup$

This question already has an answer here:

At some time in the future it is said our galaxy will collide with Andromeda. Yet at the same time the universe is expanding at a ever-accelerating pace. Using the old raisin-in-the-baking-loaf-of-bread analogy, the raisins grow farther apart and don’t “collide” with each other during the baking process. It seems to be the same way with galaxies and the expanding universe.

How are these two seemingly conflicting postulates reconciled? What am I missing here?

$\endgroup$

marked as duplicate by Mick, MBR, J. Chomel, Sir Cumference, Timtech Dec 23 '17 at 15:44

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

5
$\begingroup$

I'm not an expert, but I looked into this recently when toying with the idea of a scale model of the observable universe. To the best of my understanding, the following is approximately correct.

TLDR: The expansion happens on the very large scale of the universe. At the scale of the observable universe the Andromeda galaxy is a next-door neighbour, and at the local scale gravity is strong enough to override any effect of expansion.

The Milky Way is 100,000 light years across, 2.5 million light years brings us to the Andromeda galaxy.

Galaxies cluster in galaxy groups, and 10 million light years span about 50 galaxies and brings us to the edge of what's called the Local Group.

Galaxy groups cluster into superclusters. We're part of the Laniakea Supercluster, which spans about 100,000 galaxies and 500 million light years.

For any appreciable expansion (redshift 0.1) you need at least a billion light years. It's something that happens in the voids between superclusters, and doesn't become really pronounced until you're halfway across the observable universe (redshift 1, 7-8 billion light years as measured in light travel time).

For the "raisins in a loaf" analogy, I think it's not too wrong if you think of the raisins as superclusters rather than individual galaxies. Note that although the dough expands in a baking loaf, individual raisins do not, because there are other forces at work that hold them together.

$\endgroup$
  • 1
    $\begingroup$ +1. Just a remark on the opening. Expansion of space takes place everywhere. What happens at a very large scale is the holding of the cosmological principle (from observations and for the reasons mentioned in the A). $\endgroup$ – Alchimista Dec 16 '17 at 10:41
  • 2
    $\begingroup$ Expansion of space does not take place everywhere ('Brooklyn is not expanding!"). $\endgroup$ – Peter Erwin Dec 16 '17 at 19:37
  • 1
    $\begingroup$ @Peter Erwin yes of course. But space is expanding in brooklin. Just the building are bound :) $\endgroup$ – Alchimista Dec 17 '17 at 15:34
  • 1
    $\begingroup$ Space within superclusters is expanding (though not as fast as space within voids). If expansion only happened between superclusters, Edwin Hubble would never have been able to discover it by looking only at galaxies within a few tens of millions of light years. $\endgroup$ – Peter Erwin Dec 20 '17 at 13:08
4
$\begingroup$

In general, space in the Universe is expanding. However, this expansion can be locally retarded by high enough densities of mass-energy. On very large scales, the mean density is low enough that you can treat the whole universe as expanding. But if you look on smaller scales, you can find underdense regions (voids) where the expansion is actually faster than average, and overdense regions where the expansion is slower or has even ceased. Inside of clusters and galaxy groups (and of course inside galaxies themselves), the density is high enough that space is no longer expanding, and so it's perfectly possible for galaxies to move through local space towards each other and collide.

So in your raisin-loaf model, the trick is that regions right around each raisin -- and in local regions containing several raisins close to each other -- aren't expanding.

Note that j-g-faustus' argument is somewhat incorrect: if expansion only occurs in voids between superclusters, then Edwin Hubble would never have found it by looking at galaxies within the nearest 30 million light years. Space inside individual clusters and groups is not expanding, but space within superclusters is (just not as fast as space within voids).

$\endgroup$
  • 2
    $\begingroup$ This is correct. Essentially, any bound structure is not affected by the universal expansion. $\endgroup$ – Rob Jeffries Dec 20 '17 at 13:44
  • $\begingroup$ This seems inexact. Space expands everywhere, but the distances between and within bound objects don’t grow at the same rate and may even be reducing due to local gravity. $\endgroup$ – Chappo Says Reinstate Monica Dec 20 '17 at 20:49
2
$\begingroup$

Just as molecules in the bread have random motion due to heat, galaxies have random motion due to variations in the initial state of the universe. For galaxies that are close by, those variations are smaller than the Hubble expansion, and the Hubble expansion is simply a general trend; object that are farther away from us tend to, on average, be moving away from us at higher speed, but individual objects can still be moving towards us.

According to this site, the Hubble constant is "around 70 kilometers per second for every megaparsec in distance". According to wikipedia, Andromeda is "780 kiloparsecs" away. Multiplying those together, we get that Andromeda's movement due to the Hubble expansion should be about 5 km/sec. Wikipedia also says that "The Andromeda Galaxy is approaching the Milky Way at about 110 kilometres per second". That means that Andromeda's motion due to Hubble expansion is about 4% of its total movement. Andromeda has so much velocity relative to Earth that the Hubble expansion simply gets drowned out.

$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.