This question was prompted by my answer to this question.

An exotic $\mathbb{R}^4$ is a smooth manifold homeomorphic to $\mathbb{R}^4$ which is not diffeomorphic to $\mathbb{R}^4$ with its standard smooth structure. An exotic $\mathbb{R}^4$ is said to be small if it can be embedded in the standard $\mathbb{R}^4$ as an open subset. An exotic $\mathbb{R}^4$ which is not small is called large.

Does every large $\mathbb{R}^4$ embed in $\mathbb{R}^5$?

Freedman and Taylor showed there is a maximal exotic $\mathbb{R}^4$, into which all other $\mathbb{R}^4$'s can be smoothly embedded as open subsets. So it would be enough to show that this one embeds in $\mathbb{R}^5$.

By the Whitney Embedding Theorem, every large $\mathbb{R}^4$ embeds in $\mathbb{R}^8$, but I suspect one can do better.

  • 2
    $\begingroup$ How about this: if $M$ is an exotic $\mathbb{R}^4$ then $M\times\mathbb{R}$ is homeomorphic to $\mathbb{R}^5$, but since there is a single differentiable structure on $\mathbb{R}^5$, it is diffeomorphic to it, and restricting such a diffeomorphism to the submanifold $M\times\{0\}$ gives a submanifold of $\mathbb{R}^5$ diffeomorphic to $M$. Did I say something stupid? (I suspect I did, but I can't see it.) $\endgroup$ – Gro-Tsen May 21 '19 at 21:14
  • $\begingroup$ I guess that is exactly the same argument mentioned by Ian Agol in the comment and a bit generalized result mentioned by Igor. $\endgroup$ – Anubhav Mukherjee May 22 '19 at 0:54

The product of any two smooth open contractible manifolds is diffeomorphic to a Euclidean space, see e.g. remark 5.3 in my paper; the result is of course not due to me but I don't know any other place where all references are collected.

In particular, this applies to the product of $\mathbb R$ and any exotic $\mathbb R^4$, so the latter embeds into $\mathbb R^5$. In this case the point is that the product is simply-connected at infinity, and hence by a result of Stallings it is PL homeomorphic to $\mathbb R^5$, but any PL structure on $\mathbb R^5$ is induced by a unique smooth structure (as proved by Munkres).

  • $\begingroup$ Wow, that first result is amazing. Thanks! $\endgroup$ – Michael Albanese May 21 '19 at 12:00
  • $\begingroup$ What can we say about the end of such embedding? $\endgroup$ – Anubhav Mukherjee May 21 '19 at 15:04
  • 2
    $\begingroup$ @AnubhavMukherjee: I do not know what you mean an "end of an embedding" but I think the diffeomorphism of the product onto $\mathbb R^5$ is fairly explicit. Take a look at Stallings' paper maths.ed.ac.uk/~v1ranick/papers/stallings2.pdf. $\endgroup$ – Igor Belegradek May 21 '19 at 15:21
  • 2
    $\begingroup$ This just follows from the fact that $\mathbb{R}^n$ has a unique smooth structure for any $n\neq 4$ from the Stallings paper you reference in your comment. In particular, there is a proper embedding en.wikipedia.org/wiki/Exotic_R4 $\endgroup$ – Ian Agol May 21 '19 at 16:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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