3
$\begingroup$

Is completion of infinite degree extension of perfectoid fields perfectoid ? It is known that finite extension of perfectoid fields is also perftoid from tilting correspondence, but what about infinite cases ?

Infinite degree extension of perfectoid is not perfectoid because it is not always complete, but completion of it possibly be perfectoid. If you could find some counter example(infinite degree extension of perfectoid field $K$ whose completion is not perfectoid),I would be appreciated.

$\endgroup$

1 Answer 1

6
$\begingroup$

I'm assuming you mean infinite algebraic extensions, as otherwise there is no standard way of completing them.

Let $K$ be a perfectoid field, let $L$ be an infinite algebraic extension. Then $L$ admits a unique valuation extending that of $K$, and hence we can take the completion $\widehat L$. It is clearly complete and the valuation is nondiscrete (since the one on $K$ was). It remains to check the Frobenius is surjective on $O_{\widehat L}/p$. We have $O_{\widehat L}/p=O_L/p$. Write $L$ as a direct limit of finite extensions $L_i/K$. Then $O_L$ is the direct limit of $O_{L_i}$, and $O_L/p$ is the direct limit of $O_{L_i}/p$. Since each $L_i$ is perfectoid, the Frobenius is surjective on each $O_{L_i}/p$, and hence it is also so on the direct limit. We conclude $\widehat L$ is indeed perfectoid.

$\endgroup$
9
  • 3
    $\begingroup$ In addition, it does not seem to be reasonable to consider any non-algebraic extension. For example, I cannot imagine any way to complete $\mathbb F_p(T)/\mathbb F_p$ such that it becomes perfect. $\endgroup$
    – Z. M
    Commented Jun 5, 2022 at 12:05
  • $\begingroup$ Could you tell me why '$O_L$ is direct limit of $O_{L_i}$ ? I think that is the key point because another parts are trivial, and this part is much more difficult than other parts. $\endgroup$
    – Duality
    Commented Jun 6, 2022 at 14:47
  • $\begingroup$ @dandelion This follows from the fact $O_L\cap L_i=O_{L_i}$, which holds because both sides can be characterized as elements of $L_i$ whose valuation is nonnegative. $\endgroup$
    – Wojowu
    Commented Jun 6, 2022 at 18:28
  • $\begingroup$ Sorry to bother, but could you explain me more why from that , $O_L$ is direct limit of $O_{L_i}$ ? $\endgroup$
    – Duality
    Commented Jun 6, 2022 at 19:38
  • $\begingroup$ @dandelion This shows $O_L=\bigcup O_{L_i}$, and the union is directed - any two $O_{L_i},O_{L_j}$ are contained in $O_{L_iL_j}$. So $O_L$ will be the colimit of the diagram formed by the $O_{L_i}$ and inclusion maps between them. $\endgroup$
    – Wojowu
    Commented Jun 6, 2022 at 19:47

You must log in to answer this question.

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