×

Local \(L\)-factors of motives and regularized determinants. (English) Zbl 0762.14015

In a previous paper by the author [Invent. Math. 104, No. 2, 245–261 (1991; Zbl 0739.14010)] it was shown, among other things, that the local Euler factors of the \(L\)-function of a motive over a number field at the archimedean (infinite) places can be written as regularized determinants. Here the non-archimedean places are treated, thus giving a unified approach to all \(L\)-factors.
Let \(K\) be a non-archimedean local field with inertia group \(I\) and residue field \(\kappa\simeq\mathbb F_q\), \(q=p^f\) for some prime \(p\). Write \(W_K\) and \(W_K'\) for the associated Weil group, resp. Weil-Deligne group, and \(\text{Rep}(W_K')\) for the Tannakian category of finite dimensional complex representations of \(W_K'\). The group ring \(\mathbb C [\mathbb C]\) consists of elements of the form \(\sum re^\alpha\), where \(r,\alpha\in\mathbb C\) and the symbols \(e^\alpha\) obey the rule \(e^{\alpha+\alpha'}=e^\alpha e^{\alpha'}\). \(B\) denotes the \(\mathbb C\)-algebra \(\mathbb C[\mathbb C]\) equipped with two structures:
(i) an unramified representation of \(W_K'\) such that \(\rho(w)(e^\alpha)=\| w\|^\alpha\cdot e^\alpha\) for all \(w\in W_K\), where \(\rho\) defines the representation of \(W_K'\);
(ii) a \(\mathbb C\)-linear derivation \(\Theta\) defined by \(\Theta(e^\alpha)=\alpha e^\alpha\).
Finally, let \(\mathbb L=B^{W_K'}\) denote the \(\mathbb C\)-algebra of Laurent polynomials in \(e^{\alpha_ q}\) with \(\alpha_q=2\pi i/\log q\), and let \(\Delta=\mathbb L(\Theta)\). The additive category of left \(\Delta\)-modules which are free of finite rank over \(\mathbb L\) is written \(\text{Der}_ \kappa\).
One defines an additive functor \(\mathbb D:\text{Rep}(W_K')\to\text{Der}_ \kappa\) by \(\mathbb D(H)=(H\otimes B)^{W_K'}=(H_ N\otimes B)^{\rho(W_K)}\) with \(H_ N\) the kernel of the nilpotent endomorphism \(N\) occurring in the definition of the representation \(H\) of \(W_K'\). In general one has \(\text{rk}_\mathbb L\mathbb D(H)\leq\dim H^I_N\), and in case of equality, the representation \(H\) is called admissible. One obtains a full semisimple Tannakian subcategory of admissible representations \(\text{Rep}^{\text{ad}}(W_K')\).
On the other hand, one defines an additive functor \(\mathbb H: \text{Der}_\kappa\to\text{Rep}^{\text{ad}}(W_K')\) by \(\mathbb H(D)=(D\otimes B)^{\Theta=0}\) with \(W_K'\)-action induced by the one on \(B\). In general \(\dim\mathbb H(D)\leq\text{rk}_\mathbb L D\). \(D\) is called admissible if there is equality. \(\text{Der}_ \kappa^{\text{ad}}\) will denote the full subcategory in \(\text{Der}_\kappa\) of admissible objects. Several characterizations of admissibility can be given. The first result now says:
The functors \(\mathbb D\) and \(\mathbb H\) provide quasi- inverse equivalence of the Tannakian categories \(\text{Rep}^{\text{ad}}(W_K')\) and \(\text{Der}_\kappa^{\text{ad}}\), commuting with tensor products, twists and duals.
As usual, for a representation \(H\) in \(\text{Rep}(W_ K')\), the local \(L\)-factor is given by
\[ L_K(H,s) = \det(1-\rho(\Phi)q^{-s}| H^I_N)^{-1}, \]
where \(\Phi\) is the geometric Frobenius. For the maximal admissible submodule \(H^{\text{ad}}\) of \(H\), the following result can be shown:
\[ L_K(H^{\text{ad}},s) = \det_\infty\left(\frac{\log q}{2\pi i}(s-\Theta)|\mathbb D(H)\right)^{-1}\]
where \(\det_\infty\) denotes the regularized determinant. Writing \(\mathcal M_K\) for the category of (pure) Deligne motives, i.e. motives for absolute Hodge cycles over \(K\), and fixing an embedding \(\sigma: \mathbb Q_ \ell\hookrightarrow\mathbb C\), one has the realization functor, believed to be independent of \(\ell\) and \(\sigma\), \(H_{\ell,\sigma}: \mathcal M_K\to\text{Rep}(W_ K')\), \(\ell\neq p\), given by \(H^\bullet_{\ell,\sigma}(M)=H^ \bullet_ \ell(M)\otimes_{\mathbb Q_ \ell,\sigma}\mathbb C\). Denote by \(\mathcal M_K^{\text{ad}}\) the subcategory of \(\mathcal M_K\) with objects \(M\) such that \(H^\bullet_{\ell,\sigma}(M)\in\text{Rep}^{\text{ad}}(W_ K')\). For \(M\) in \({\mathcal M}_K^{\text{ad}}\) one sets \(H^\bullet(M/\mathbb L)=\mathbb D H^\bullet_{\ell,\sigma}(M)\). For a smooth projective variety over \(K\) such that the associated motive \(h(X)\) is in \({\mathcal M}_K^{\text{ad}}\), one writes \(H^\bullet(X/\mathbb L)\) for \(H^\bullet(h(X)/\mathbb L)\). Also, for any \(M\) in \({\mathcal M}_K\), one writes \(L_ K(H^w(M),s)=L_K(H^w_{\ell,\sigma}(M),s)\). For \(M=h(X)\) where \(X\) has good reduction this local factor is known to be independent of \(\ell\) and \(\sigma\). The following theorem is proven: For \(X\) smooth projective over \(K\), and \(w=0,1\) one has
\[ \det_ \infty\left(\frac{\log q}{2\pi i}(s-\Theta)|\mathbb D H^ w_{\ell,\sigma}(X)\right)^{-1}=L_ K(H^ w(X),s), \]
in particular, if \(X\) has good reduction the \(\mathbb D H^ w_{\ell,\sigma}(X)\) may be replaced by \(H^ w(X/\mathbb L)\). The theorem is expected to hold for all \(w\).
The theorem suggests a global cohomological approach to \(L\)-functions of varieties over number fields. Concretely, there should exist a big site containing \(\overline{\text{Spec}(\mathbb Z)}=\text{Spec}(\mathbb Z)\cup\{\infty\}\) with suitable properties. This formalism should give rise to some remarkable expressions involving the Riemann zeta-function. One result can indeed be proven by analytic means: For \(\operatorname{Re} z>1\) consider the Dirichlet series \(\xi(s,z)=\sum_ \rho\frac{1}{\bigl[\frac{1}{2\pi}(z-\rho)\bigr]^ s}\), where \(\rho\) runs over the nontrivial zeros of the Riemann zeta-function and such that \(\arg(z-\rho)\in\bigl(-\frac{\pi}{2},\frac{\pi}{2}\bigr)\). Then \(\xi(s,z)\) converges absolutely for \(\operatorname{Re} s>1\) and for fixed \(x\) it admits an analytic continuation to a holomorphic function in \(\mathbb C\backslash\{1\}\). One has the formula:
\[ 2^{- 1/2}(2\pi)^{-2}\pi^{-z/2}\Gamma\left(\frac{z}{2}\right)\zeta(z)z(z-1)=\exp(-(\partial_ s\xi)(0,z)). \]

MSC:

14G10 Zeta functions and related questions in algebraic geometry (e.g., Birch-Swinnerton-Dyer conjecture)
14A20 Generalizations (algebraic spaces, stacks)
11M36 Selberg zeta functions and regularized determinants; applications to spectral theory, Dirichlet series, Eisenstein series, etc. (explicit formulas)
11M38 Zeta and \(L\)-functions in characteristic \(p\)

Citations:

Zbl 0739.14010

References:

[1] [A] Ahlfors, L.V.: Complex Analysis. New York: McGraw-Hill 1966 · Zbl 0154.31904
[2] [Ba] Barner, K.: On A. Weil’s explicit formula. J. Reine Angew. Math.323, 139-152 (1981) · Zbl 0446.12013 · doi:10.1515/crll.1981.323.139
[3] [C-V] Cartier, P., Voros, A.: Une nouvelle interpr?tation de la formule des traces de Selberg. (Grothendieck Festschrift II) Boston Basel Stuttgart: Birkh?user 1991
[4] [D] Deninger, C.: On the ?-factors attached to motives. Invent. Math.104, 245-261 (1991) · Zbl 0739.14010 · doi:10.1007/BF01245075
[5] [E] Erdelyi, A. et al.: Higher transcendental functions, vol. I. Bateman Manuscript project. New York: McGraw-Hill 1953
[6] [F] Fontaine, J.-M.: Modules galoisiens, modules filtr?s et anneaux de Barsotti-Tate. In: Journ?e de G?om?trie Alg?brique de Rennes (Ast?risque, vol. 65, pp. 3-80) Paris: Soc. Math. de France 1979
[7] [G] Grothendieck, A.: Mod?les de N?ron et Monodromie. Exp IX in SGA 7, I. (Lect. Notes Math., vol. 288) Berlin Heidelberg New York: Springer 1972
[8] [J1] Jannsen, U.: On thel-adic cohomology of varieties over number fields and its Galois cohomology. In: Y. Ihara, K. Ribet, J.-P. Serre (eds.) Galois Groups over ?, pp. 315-353. (Publ., Math. Sci. Res. Inst., vol. 16) Berlin Heidelberg New York: Springer 1989
[9] [J2] Jannsen, U.: Motivic cohomology,l-adic cohomology and vanishing orders ofL-functions. (Preprint 1990)
[10] [K1] Kurokawa, N.: Parabolic components of zeta functions. Proc. Japan Acad, Ser. A64, 21-24 (1988) · Zbl 0642.10028 · doi:10.3792/pjaa.64.21
[11] [K2] Kurokawa, N.: Analyticity of Dirichlet series over prime powers. (Lect. Notes Math., vol. 1434, pp. 168-177) Berlin Heidelberg New York: Springer 1990
[12] [K3] Kurokawa, N.: Multiple zeta functions: an example. In: Zeta functions in geometry. Adv. Stud. Pure Math. 1991 (to appear)
[13] [Se] Serre, J.P.: Facteurs locaux des fonctions z?ta des vari?t?s alg?briques (d?finitions et conjectures). S?minaire Delange-Pisot-Poitou, expos? 19, 1969/70
[14] [So] Soul?, Ch.: Letter to the author. February 13, 1991
[15] [Ta] Tate, J.: Number theoretical background. In: A. Borel, W. Casselman (eds.): Automorphic forms, representations andL-functions, pp. 3-26. Corvallis 1977. (Proc. Symp. Pure Math. XXXIII, 2) Providence: Am. Math. Soc. 1979
[16] [V] Voros, A.: Spectral functions and the Selberg zeta function. Commun. Math. Phys.111, 439-465 (1987) · Zbl 0631.10025 · doi:10.1007/BF01212422
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.