Author: Eiko

Tags: affinoid, p-adic geometry, affinoid algebra, projective line, affinoid subspace

Time: 2024-09-12 14:48:05 - 2024-09-14 23:56:12 (UTC)

Affinoid Algebras and Affinoid Spaces with P1 as Example

Corresponding of Terminologies

There are some basic terminologies that resembles the ones used in algebraic geometry. Let K be a complete valued (non-archimedean) field with ring OK and prime p, k=OK/p.

AG terms Analytic terms
Polynomial algebra K[X1n] Tate algebra Tn=KX1n={cαxα||cα|0}
Gauss normcαxα=max|cα|
Affine algebra A=R/I Affinoid algebra (quotient of Tate)Tn/J
Affine Space Spec(A) Affinoid space (maximalideals) Sp(Tn/J)
Zariski Topology V(a),aA Zariski Topology V(a),aA
Canonical Topology whose topological base generated by (finite intersections and arbitrary unions) X(f,ε)={x||f(x)|ε}
Ringed space (X,OX) Ringed space with Grothendieck topology (X,OX)

The Example of Projective Line

Affinoid Subsets in P1

The projective line P1 over an algebraically closed non-archimedean field K for example Cp is intuitively depicted as K{}. It serves as a basic and useful example before heavy mechanisms come into place.

  • The linear group PGL2(K) acts on P1 by coordinates, and in affine coordinate it is just a fraction linear transform zaz+bcz+d.

  • The ’open’ disks are {|za|<r},{|za|>r} and closed disks where <,> are replaced by ,. They are both open and closed. The PGL2(K) group acts transitively on all open disks, and all closed disks.

  • A connected affinoid subset of P is the complement of a non-empty finite union of open disks. i.e. discarding finite many open disks, or equivalently, finite intersection of non-full closed disks, whose complement is of the form Fc=Bi where Bi are open disks. Thus PGL transforms connected affinoid subsets to connected affinoid subsets. Connected affinoid subset can be empty, but not full.

  • An affinoid subset of P is a finite union of connected affinoid subsets.

Example 1. Think through the following interesting examples to get a feeling of open disks in P1,

  • Let B1,B2 be two open disks s.t. B1B2P, then B1B2 is either empty or an open disk. Because we can use group action to move them into affine line, and the intersection is either empty or one of them is contained in another.

  • If B1B2=P, what is B1B2? They are strips, you can see this by moving their centers to 0 and .

Lemma 1. Let F1,F2 denote two connected affinoid subsets of P1

  • F1F2 is always a connected affinoid subset.

  • if F1 and F2 have both non-trivial intersection and union, F1F2 is a connected affinoid subset.

Proof.

  • Finite intersections of connected affinoid subsets are obviously connected affinoid subsets, since (F1F2)c=F1cF2c.

  • The condition F1F2 is crucial, it means P(F1cF2c)=BiBi. Therefore by lemmas BiBj is either empty or an open disk, so (F1F2)c=i,j(BiBj) is (the complement of) a connected affinoid subset.

  • F1F2P ensures at least one of the above BiBj is non-empty.

 ◻

  • Every affinoid subset F is in a unique way a finite union of connected affinoid subsets, called the connectedcomponents of F. This is similar to writing things in disjunctive normal form.

  • Affinoids are closed under finite union and intersection.

Affinoid Subsets on P1 given by Rational Functions

Let’s explore affinoids defined by a rational function. Let fK(z) be any rational function on P1 and r|K| we consider F:={aP1||f(a)|r} then F is either an affinoid subset or empty.

Proof. One may write f=i=1s(zai)ni with distinct aiK and integers niZ. We can prove by induction on the number of factors of f.

  1. The case of s=1 is trivial, regardless whether n10 or n1<0 (we are talking about affinoids so P1 is accepted as an affinoid subset).

  2. Assume now a1=0,a2=1. We know that P1=U1U2U3 where U1={aP1:|a|1,|a1|1} U2={aP1:|a|1,|a1|1} U3={aP1:|a|1,|a1|1}. (the other case degenerates to their intersections {|a|=1,|a1|=1})

  3. We can see from |a1|1,|a|1|a|=|a1|, this means F is locally Fzn1+n2g2(z), U1F={aU1:|zn1+n2g2(z)|r}=U1Fzn1+n2g2(z). Therefore by induction hypothesis, Fzn1+n2g2(z) is affinoid, U1 is also affinoid, so U1F is affinoid since finite intersection of affinoids is affinoid.

  4. In fact in U2 we also have |a|=|a1|. Since if |a|=1 we would have 1|a1|1, and when |a|>1 we must have |a1|=max(|a|,|a1|)=|a|. Therefore U2F is also an affinoid subset by the same arguments. Similarly for U3.

  5. We can conclude by using the fact that the union of affinoid subsets with non-trivial intersection and union is an affinoid subset.

 ◻

Proposition 1. A rational function viewed as f:P1P1 has the property that any preimage of affinoid subset is again an affinoid subset.

Proof. Consider first the case of a non-full closed disk DP1. Then its preimage is an affinoid subset as we have proved in previous lemma. This means the preimage of connected affinoid subsets are affinoid subsets, as a result the preimage of affinoid subsets are affinoid subsets. ◻

Connected affinoid subsets are characterized by the fact that the analytic function rings on it does not have zero divisors.