See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/265949118

Primitives of 𝑝-adic meromorphic functions

Article · April 2011

DOI: 10.1090/conm/551/10884

CITATIONS READS

8 85

3 authors, including:

Alain Escassut Jacqueline Ojeda

Université Clermont Auvergne University of Concepción
184 PUBLICATIONS 1,043 CITATIONS 28 PUBLICATIONS 217 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Alain Escassut on 17 October 2014.

The user has requested enhancement of the downloaded file.

 Primitives of p-adic meromorphic functions.

Kamal Boussaf, Alain Escassut and Jacqueline Ojeda

Abstract. Let K be an algebraically closed field of characteristic 0, complete

with respect to an ultrametric absolute value. We show that a meromorphic
function in K or in an open disk admits primitives if and only if all its residues
are null.

1. Introduction and results

Notation and definitions: Let K be an algebraically closed field of characteris-
tic 0, complete with respect to an ultrametric absolute value | . |. Given α ∈ K and
R ∈ IR∗+ , we denote by d(α, R) the disk {x ∈ K | |x − α| ≤ R} and by d(α, R− )
the disk {x ∈ K | |x − α| < R}, by A(K) the K-algebra of analytic functions in K
(i.e. the set of power series with an infinite radius of convergence) and by M(K)
the field of meromorphic functions in K [2],[4].
In the same way, given α ∈ K, r > 0 we denote by A(d(α, r− )) the K-algebra
of analytic functions in d(α, r− ) (i.e. the set of power series with a radius of
convergence ≥ r) and by M(d(α, r− )) the field of fractions of A(d(α, r− )). Here we
mean to characterize meromorphic functions in K or inside an open disk that admit
primitives. In complex analysis, a meromorphic function having all its residues
equal to 0 admits primitives. On our field K, such a conclusion seems obvious but
apparently has not been published in any paper. Actually, the proof requires the
use of analytic elements [1], [3].
Notation and definitions: Let f ∈ M(d(a, r− )) have a pole of order q at a: f
X∞
then admits an expansion in a Laurent series of the form f (x) = an (x − a)n .
n=−q
As usual, a−1 is called residue of f at a and will denoted by resa (f ).

Theorem 1: Let f ∈ M(K) (resp. f ∈ M(d(a, R− ))). Then f admits primitives

in M(K) (resp. in M(d(a, R− ))) if and only if f has no residue different from
zero.

1991 Mathematics Subject Classification. Primary 12J25 Secondary 46S10.

Key words and phrases. p-adic meromorphic functions, primitives, residues.
Partially supported by CONICYT N ◦ 9090014 (Inserción de Capital Humano a la
Academia)”.
1
2 K. Boussaf, A. Escassut and J. Ojeda

2. The proofs
Let D be a subset of K. If D is bounded of diameter R we denote by D e the
disk d(a, R) for any a ∈ D. If D is not bounded we put D = K.
e
Given a point a ∈ K we put δ(a, D) = inf{|x − a| |x ∈ D}. Then δ(a, D) is named
the distance of a to D.

Recall that given any subset D of K, D e \ D admits a unique partition of the

form (Ti )i∈I , whereas each Ti is a disk of the form d(ai , ri− ) with ri = δ(ai , D).
Such disks d(ai , ri− ) are called the holes of D [1].
Given a ∈ K and r0 , r00 ∈ IR∗+ with r0 < r00 , we denote by Γ(a, r0 , r00 ) the
annulus {x ∈ K | r0 < |x − a| < r00 }.
A set D is said to be infraconnected if for every a ∈ D, the mapping Ia from D
to IR+ defined by Ia (x) = |x − a| has an image whose closure in IR+ is an interval.
In other words, a set D is not infraconnected if and only if there exist a and b ∈ D
and an annulus Γ(a, r1 , r2 ) with 0 < r1 < r2 < |a − b| such that Γ(a, r1 , r2 ) ∩ D = ∅.

Given a subset E of K, we denote by R(E) the algebra of rational functions

with no pole in E and the functions that are uniform limits of a sequence of R(E)
are called analytic elements on E. We denote by H(E) the set of analytic elements
on E. Moreover, we denote by Rb (E) the sub-K-algebra of R(E) consisting of
the h ∈ R(E) that are bounded in E. So, Rb (E) admits the norm of uniform
convergence as a K-algebra norm and we denote by Hb (E) its completion that
consists of all h ∈ H(E) that are bounded in E. The algebra H(E) is then a
K-Banach algebra. Moreover, when E is unbounded, we denote by H0 (E) the set
of f ∈ H(E) such that lim f (x) = 0. Particularly, when E is of the form
|x|→+∞,x∈E
K \ d(a, R− ), then H0 (E) is K-Banach algebra with respect to the norm of uniform
convergence on E denoted by k . kE .

Lemma 1: Let f ∈ H(d(0, R)) and let r ∈]0, R[. Then f admits primitives in
H(d(0, r)).
∞
X
Proof: Since f lies in H(d(0, R)), f (x) has expansion of the form an xn with
n=0
lim |an |Rn = 0. Particularly, the radius of convergence
n→∞
1
ρ= p
lim supn→∞ n
|an |
is such that ρ ≥ R. But by Theorem 1.5.4 of p-adic value distribution [2] the
∞
X an n+1
primitive F (x) = x also has a radius of convergence equal to S. Con-
n=0
n +1
sequently, for all r < R, by Proposition 13.3 [1], F belongs to H(d(0, r)).

In the proof of Theorem 1, we need Krasner’s Mittag-Leffler Theorem. Here

for simplicity, we will only state it when the set D is closed and bounded.

Theorem A : (Krasner-Mittag-Leffler Theorem) [1], [3] Let D be a closed

and bounded infraconnected subset of K and let f ∈ H(D) . There exists a unique
 Primitives of p-adic meromorphic functions 3

sequence of holes (Tn )n∈IN∗ of D and a unique sequence (fn )n∈IN in H(D) such that
f0 ∈ H(D),
e fn ∈ H0 (K \ Tn ) (n > 0), lim fn = 0 and
n→∞
∞
X
(1) f= fn and kf kD = sup kfn kD .
n=0 n∈IN
Moreover for every hole Tn = d(an , rn− ) , we have
(2) kfn kD = kfn kK\Tn ≤ kf kD .
If D
e = d(a, r) we have
(3) kf0 kD = kf0 kDf
≤ kf kD .

Let us recall Theorems B and C [1].

Theorem B (Theorem 19.5 [1]): Let E be an open set in K such that E also
is open. Then E is infraconnected if and only if for every f ∈ H(E) such that
f 0 (x) = 0 whenever x ∈ E, we have f = ct.

Theorem C (Theorem 25.5 [1]): Let (aj )j∈IN be a sequence in d(0, r− ) such
that |an | ≤ |an+1 | for every n ∈ IN and lim |an | = r. Let (qn )n∈IN be a sequence in
n→∞
IN∗ and let B ∈]1, +∞[. There exists f ∈ A(d(0, r− )) satisfying
i) f (0) = 1
n
Y an
ii) kf kd(0,|an |) ≤ B | |qj whenever n ∈ IN
j=0
aj
iii) for each n ∈ IN, an is a zero of f of order zn ≥ qn .

Proof of Theorem 1: Let a be a pole of f . According to the Laurent series of f

at a, if f admits primitives then f has no residue different from zero at a because
1
has no primitive in H(d(a, r)) (whenever r > 0).
x−a
Now suppose that for every pole α of f we have resα (f ) = 0. Without loss
of generality, we may assume that f admits infinitely many poles and has no pole
at the origin. Moreover, when f is a meromorphic function in d(a, R− ), we may
assume that a = 0.
Let (an )n∈ IN be the sequence of poles of f in K (resp. in d(0, R− )), with |an | ≤
∞
Y x qn −1
|an+1 |, ∀n ∈ IN, each pole an being of order qn and let ψ(x) = (1 − ) .
n=0
an
For each S > 0 (resp. for each S ∈]0, R[), we denote by t(S) the greatest integer n
t(S)
[
such that |an | ≤ S. And for each r ∈]0, S[, we set D(r, S) = d(0, S − ) \ ( d(aj , r− )).
j=0
According to results on analytic elements [1], [2], f obviously defines an element of
H(D(r, S)).
Let us take r < min(|ai − aj |, 0 ≤ i < j ≤ t(S)). In each hole d(an , r− ) of
D(r, S), f has a unique pole that is an . It is of order qn and the Mittag-Leffler term
qn
−
X λn,i
fn of f in the hole d(an , r ) is of the form because, by hypothesis,
i=2
(x − an )i
the residue of f at an is null. On the other hand, as an element of H(D(r, S)),
by Theorem A, the Mittag-Lefler decomposition of f on the infraconnected set
4 K. Boussaf, A. Escassut and J. Ojeda

t(S)
X
D(r, S) is of the form fn + gS with gS ∈ H(d(0, S)). Moreover, we notice that
n=0
there exists S 0 > S (resp S 0 ∈]S, R[) such that f admits no pole b ∈ Γ(0, S, S 0 ).
Consequently, the holes of D(r, S) are the same as these of D(r, S 0 ) and therefore,
the Mittag-Leffler decomposition of f on D(r, S) also holds on D(r, S 0 ). Hence gS
belongs to H(d(0, S 0 )). But then by Lemma 1, gS admits primitives in H(d(0, S)).
Now, as an element of H(D(r, S)), each fn admits a primitive Fn of the form
qn
X λn,i
− .
i=2
(i − 1)(x − an )i−1

And as an element of H(D(r, S)) again, f admits a primitive LS satisfying LS (0) =

t(S))
X
0, equal to Fn + GS with GS ∈ H(d(0, S)) .
n=0
Now, fixing S we can take r arbitrary small: the Mittag-Leffler terms f1 , ..., ft(S)
and GS remain the same because if r0 < r < min(|ai − aj | 0 ≤ i < j ≤ t(S)), the
Mittag-Leffler expansions on D(r0 , S) holds on D(r, S) and the two expansions are
equal, due to the properties of the Mittag-Leffler expansions. Thus, the function
LS is defined in d(0, S) \ {a1 , ..., at(S) } and satisfies (LS )0 (x) = f (x), ∀x ∈ d(0, S) \
{a1 , ..., at(S) }.
Consider first the case when f belongs to M(K). The function
∞
Y x qn −1
ψ(x) = (1 − ) is an entire function, hence an element of H(d(0, S)).
n=0
an
Let φS (x) = LS (x)ψ(x). Then for any r > 0, φS is an element of H(D(r, S))
meromorphic on each hole of D(r, S) [1], (chap. 31). But actually, by construction,
φS has no pole in any hole of D(r, S). Consequently, φ belongs to H(d(0, S)).
Now, let S 0 > S. Similarly we can make a function φS 0 = LS 0 ψ. We will show
that the restriction of φS 0 to d(0, S) is φS . Indeed, by definition, both functions
LS and LS 0 are null at 0 and have a derivative equal to f in D(r, S), for any
r > 0. Consequently, LS and LS 0 coincide in a disk and therefore so do φS and φS 0 .
Hence the equality φS (x) = φS 0 (x) holds in all d(0, S). Thus, we can define the
function φ(x) = φS (x), ∀x ∈ d(0, S). Since φS belongs to H(d(0, S)), φ belongs
to H(d(0, S)) for all S > 0 and therefore is an entire function. Now, we can set
φ(x)
F (x) = and then F belongs to M(K). On the other hand, since (LS )0 = f
ψ(x)
and since F (x) = LS (x) in d(0, S)\{a1 , ..., an , ...}, we have F 0 (x) = f (x), ∀d(0, S)\
{a1 , ..., an , ...} and hence, by Theorem B we know that the equality F 0 (x) = f (x)
holds in each set D(r, S), for all r > 0, hence in all K \ {a1 , ..., an , ...}.
Similarly, consider the case when f belongs to M(d(0, R− )). By Theorem C,
we can find a function ψ(x) ∈ A(d(0, R− )) admitting each an as a zero of order
sn ≥ qn − 1. Then, as in the previous case, using the same notation, we can show
that LS (x)ψ(x) lies in H(d(0, S)) for every S < R, because it has no pole in d(0, S).
Now, for every S < S 0 < R, φS is the restriction of φS 0 to d(0, S). Let φ be defined
as φ(x) = φS (x) for all x ∈ d(0, S), ∀S < R. Then φ belongs to A(d(0, R− )) and
φ(x)
hence the function F (x) = belongs to M(d(0, R− )). And similarly, we have
ψ(x)
F 0 = f , which ends the proof.
 Primitives of p-adic meromorphic functions 5

References
[1] Escassut, A. Analytic Elements in p-adic Analysis. World Scientific Publishing Co. Pte.
Ltd. (Singapore, 1995).
[2] Escassut, A. p-adic Value Distribution. Some Topics on Value Distribution and Dif-
ferentability in Complex and P-adic Analysis, p.42- 138. Mathematics Monograph, Series
11. Science Press.(Beijing 2008)
[3] Krasner, M. Prolongement analytique uniforme et multiforme dans les corps valués com-
plets. Les tendances géométriques en algèbre et théorie des nombres, Clermont-Ferrand, p.94-
141 (1964). Centre National de la Recherche Scientifique (1966), (Colloques internationaux
du C.N.R.S. Paris, 143).
[4] Hu, P.C. and Yang, C.C. Meromorphic Functions over non-Archimedean Fields. Kluwer
Academic Publishers, (2000).

Current address: Laboratoire de Mathématiques, UMR 6620, Université Blaise Pascal &
CNRS, Complexe scientifique des Cézeaux, BP 80026, 63171 Aubière Cedex, France,
Current address: Laboratoire de Mathématiques, UMR 6620, Université Blaise Pascal &
CNRS, Complexe scientifique des Cézeaux, BP 80026, 63171 Aubière Cedex, France,
Current address: Departamento de Matemática Facultad de Ciencias Fı́sicas y Matemáticas
Universidad de Concepción Chile
E-mail address: kamal.boussaf@math.univ-bpclermont.fr,
E-mail address: alain.escassut@math.univ-bpclermont.fr,
E-mail address: jacqojeda@udec.cl

View publication stats