An introduction to p-adic L-functions — Lean blueprint

14.3. The analytic p-adic L-function🔗

The analytic p-adic L-function L_p^{\mathrm{an}}(f) is a p-adic distribution on \Zpx that interpolates the critical values of L(f,s). Its existence is governed by the slope v_p(\alpha_p) of the U_p-eigenvalue \alpha_p of f.

Theorem14.3.1
Statement uses 3
Statement dependency previews
Preview
Definition 3.7.3
Loading preview
Statement dependency preview content is loaded from the Blueprint HTML cache.
Used by 2
Reverse dependency previews
Preview
Proposition 14.3.2
Loading preview
Reverse dependency preview content is loaded from the Blueprint HTML cache.
XL∃∀N

Let \alpha_p be the U_p-eigenvalue of f, and suppose v_p(\alpha_p) < k+1. Then there is a unique locally analytic distribution L_p^{\mathrm{an}}(f) \in \sD^{\mathrm{la}}(\Zpx) on \Zpx such that:

  • L_p^{\mathrm{an}}(f) has growth of order v_p(\alpha_p); and

  • for every Dirichlet character \chi of conductor p^n and every 0 \le j \le k, L_p^{\mathrm{an}}(f,\overline{\chi},j+1) = \int_{\Zpx} \chi(x)\,x^j \cdot L_p^{\mathrm{an}}(f) = -\alpha_p^{-n}\left(1 - \chi(p)\tfrac{p^j}{\alpha_p}\right) \frac{G(\chi)\,j!\,p^{nj}}{(2\pi i)^{j+1}} \cdot \frac{L(f,\overline{\chi},j+1)}{\Omega_f^{\pm}}, where G(\chi) is the Gauss sum and \Omega_f^{\pm} the canonical complex periods.

This generalises Theorem 5.1.1 and lives in the space of Definition 3.7.3; it depends on Definition 14.2.1.

Proof for Theorem 14.3.1
uses 0

This is the theorem of Mazur–Swinnerton-Dyer, Amice–Vélu and Višik. One forms the modular symbol attached to f, an H-valued additive functional on \Q-paths in the upper half-plane whose values along \{i\infty \to a/p^n\} encode the twisted L-values L(f,\overline{\chi},j+1). Applying the U_p-operator and dividing by \alpha_p makes these symbols compatible under the maps in the cyclotomic tower; the slope bound v_p(\alpha_p) < k+1 guarantees that the resulting compatible system has bounded growth of order v_p(\alpha_p), so it defines an admissible (locally analytic) distribution on \Zpx rather than merely a sequence of values. Uniqueness follows because a distribution of growth < k+1 is determined by its values against the characters x \mapsto \chi(x)x^j with 0 \le j \le k, which are exactly the prescribed critical L-values. The Euler factor 1 - \chi(p)p^j/\alpha_p and the power \alpha_p^{-n} arise from the U_p-stabilisation, exactly as the factor 1-p^{-s} arose for \zeta_p.

Proposition14.3.2
Statement uses 2
Statement dependency previews
Preview
Definition 3.3.3
Loading preview
Statement dependency preview content is loaded from the Blueprint HTML cache.
used by 1XL∃∀N

If f is p-ordinary, i.e. v_p(\alpha_p) = 0, then the growth condition forces L_p^{\mathrm{an}}(f) to lie in the subspace of p-adic measures on \Zpx, that is in the Iwasawa algebra \Lam(\Zpx) \subset \sD(\Zpx). This uses Theorem 14.3.1 and Definition 3.3.3.

Proof for Proposition 14.3.2
uses 0

Growth of order 0 is precisely boundedness, and a bounded locally analytic distribution is a measure; this is the same dichotomy seen for \zeta_p, where p-ordinarity (here v_p(\alpha_p) = 0) places the distribution in the Iwasawa algebra rather than the larger distribution space. The measure then sits in \Lam(\Zpx) under the Mahler/Mellin identification of measures with the Iwasawa algebra.

Boundary and infinite slope. If \alpha_p \neq 0 then v_p(\alpha_p) \le k+1, but the theorem above excludes the boundary case v_p(\alpha_p) = k+1; this is handled by the overconvergent modular-symbol methods of Pollack–Stevens and Bellaïche, with a slightly modified statement. The infinite slope case \alpha_p = 0 is harder still: partial p-adic L-functions with good interpolation are constructed from Kato's Euler system via Perrin-Riou's big logarithm maps. We do not develop these refinements here.