4
$\begingroup$

What are some general results describing the Krull dimension of the completion of a non-Noetherian ring with a "nice" topology?

An example of the sort of "nice" topological ring I'm looking for is a perfection of an affinoid algebra over a perfectoid (probably just complete non-archimedean with a non-discrete valuation is good enough) field $K$ in characteristic $p > 0$.

So let $R$ be a quotient of a Tate algebra $K\langle T_1, \ldots, T_n \rangle$ of power series which converge on the adic closed (i.e. the coefficients tend to $0$). The topology is induced from the non-archimedean valuation topology on $K$ and is given by the Gauss norm $\|\sum_\alpha a_\alpha T^\alpha\| = \sup_\alpha |a_\alpha|$. So in particular $R$ is a noetherian topological ring which is "Tate" (meaning its topology is defined by a topologically nilpotent unit $\varpi$, and there's an open subring on which the topology is $\varpi$-adic).

We can form its perfection $R^{\mathrm{perf}} = \varinjlim_m R^{1/p^m}$, which is a quotient of the perfection of the Tate algebra $\cup_m K \langle T_1^{1/p^m}, \ldots, T_n^{1/p^m} \rangle$. The perfection comes equipped with the direct limit topology, also compatible with the Gauss norm and "induced from the topology on $K$" in an appropriate sense. The completion of $R^{\mathrm{perf}}$ with respect to this topology is what is sometimes called the "completed perfection" of $R$.

I'd like to show that the Krull dimension of this ring agrees with that of $R^{\mathrm{perf}}$, and therefore with $R$.

$\endgroup$

2 Answers 2

3
$\begingroup$

Unfortunately, the Krull dimension often behaves in strange ways for such non-noetherian rings. My impression is that the results of Lang--Ludwig should adapt to the present situation, and show that this ring has infinite Krull dimension.

$\endgroup$
0
$\begingroup$

It turns out that the rings you mention, often called "perfectoid Tate algebras" in the literature, indeed have infinite Krull dimension. In fact, the Krull dimension is uncountable in the sense of Du, see arXiv:2002.10358.

I posted a preprint arXiv:2212.12215 which gives a proof very much inspired by Du's paper, which is in turn inspired by the Lang--Ludwig paper mentioned above.

$\endgroup$

You must log in to answer this question.

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