Author: Eiko

Time: 2025-02-01 23:03:11 - 2025-02-01 23:03:11 (UTC)

References: p-adic Differential Equations by Kiran S. Kedlaya

Rings of functions on discs and annuli

Let K be a complete field of characteristic 0 for a norm || with residue field κ of characteristic p (which is allowed to be zero as well o.o). The norm is normalized |p|=p1 (if p0).

Various Power Series Rings

Recall in rigid analytic geometry of P1, the most important examples of affinoid subset FP1 are the ring domain {|z|=1} and the thick ring domain {r1|z|r2}. The ring of functions on these domains are given by power series rings with convergence conditions, similar to a Tate algebra.

  • The classical Tate algebra is the ring of power series n0antn with |an|n0, corresponding to O({|z|1}).

  • For F={|z||β|}, this is just a slight generalization of the Tate algebra, where you can also conveniently think of the Tate algebra of the variable z/β instead of z.

    O(F)=Kzβ={n0anβntn:|an|=o(1)}={n0cntn:|cnβn|=o(1)}={n0cntn:v(cn)nv(1/β)n}

  • F={|α||z|}. If we do a variable substitution and consider power series ring based at infinity, i.e. with variable αz, the Tate algebra of variable αz give the ring O(F).

    O(F)=Kαz={n0anαnzn:|an|=o(1)}={n0cnzn:|cnαn|=o(1)}={n0cnzn:v(cn)(n)v(1/α)n}

  • For F={|α||z||β|}, the thick ring domain. This O(F) is in fact the intersection of Kzβ and Kαz.

    O(F)=Kαz,zβ={nZantn:v(ak)kv(1/β)kv(ak)kv(1/α)k}

    or equivalently as

    Kαz,zβ={nZantn:|akαk|=o(1)|akβk|=o(1)}.

    Which pictorially is a V-shaped region, your function need to eventually go above the V-shape and the height to the V-shape needs to go to infinity on both sides. We also conveniently denote this ring as Kαz,zβ.

  • For F={|z|=1}, the ring domain, we have

    O(F)=Kz,z1={nZantn:|an||n|0}={nZantn:v(an)nv(1/1)|n|}

Some Other Intermediate Rings

  • We define the ring with finite Gauss norm as

    K[[t/β]]0:={i=0aiti:sup|ai|βi<}.

    Clearly we have for δ<β that

    Kt/βK[[t/β]]0Kt/δ

    Note that we can also write it as

    K[[t]]0=oK[[t]]oKK

    where oK is the ring of integers of K.

  • It is also possible to mix the two type of boundaries, producing rings like

    Kαz,zβ]]0={nZantn:sup|ak|βk<|ak|αk=o(1)}

    K[[αz,zβ]]0={nZantn:sup|ak|βk<sup|ak|αk<}

    For which we have the simple property, for δ[α,β)

    Kα/t,t/βKα/t,t/β]]0Kα/t,t/δ

Newton Polygon

Given an element x of Kα/t,t/β, we can associate a Newton polygon to it.

  1. The full Newton Polygon NP(x) is the lower convex hull of the set of points (i,v(ai)).

  2. The essential Newton Polygon ENP(x)=NP[log|α|,log|β|](x) is generated by the subset of slopes within [log|α|,log|β|].

  3. Denote by dλ(f)=inf(v(ai)λi) the distance of the Newton polygon to the line y=λx.

Properties

Note: all the log are respect to the base of the valuation. log|α| can also be more easily understood as v(1/α).

  • ENP(x) has finite length.

  • d()(x) is continuous and piecewise linear on [log|α|,log|β|].

  • |x|λ is continuous on [log|α|,log|β|] and log-convex, for t[0,1] we have

    |x|tλ+(1t)μ|x|λt|x|μ1t.

    Or equivalently, for xKα/t,t/β we have that d()(x) is concave, i.e.

    dtλ+(1t)μ(x)tdλ(x)+(1t)dμ(x).

    Proof:

    inf[v(ai)(tλ+(1t)μ)i]=inf[t(v(ai)λi)+(1t)(v(ai)μi)]tinf(v(ai)λi)+(1t)inf(v(ai)μi)

    Taking |x|λ=pdλ(x) we obtain the log-convexity.

  • Since |x|λ is log-convex, there exists a point λ0 (if we allow ±) such that |x|λ0 is minimal, |x|λ is decreasing for λ<λ0 and increasing for λ>λ0.

    If α=0, this λ0= and we have |x|λ is increasing on (,log|β|].

Frechet Completeness

The rings Kα/t,t/β and Kα/t,t/β]]0 are Frechet complete with respect to any Gauss norm ||λ with slope λ[log|α|,log|β|].

Unit Structure

  • A non-zero element fKt/β is a unit iff |fc|log|β|<|f|log|β| for some unit cK×.

  • A nonzero element fKα/t,t/β is a unit iff there exists a term cti such that |fcti|λ<|f|λ for λ{log|α|,log|β|}.

  • A nonzero element fKα/t,t/β]]0 is a unit iff there exists a term cti such that |fcti|λ<|f|λ for λ[log|α|,log|β|).

Factorization

In Kα/t,t/β or Kα/t,t/β]]0, if the series x has a unique term maximizing |xm|ρm for some ρ[α,β] or the supremum of |xi|ρi is achievable with least achieving index m, then by moving the term xmtm to center we can obtain the following factorization:

x=xmtmg(t)h(t1)

with h(t1)Kα/thKt/α1, |h(t1)1|<1 and g(t)Kt/β or Kt/β]]0 with |g|log|β|=1 (If you are getting unique index m, |g|log|β|<1).

Analytic Elements

Sitting between Kt/βK[[t/β]]0, the set of elements in K[[t/β]]0 that comes from limit of rational functions K(t) is called the set of analytic elements, denoted K[[t/β]]an.