Let $C$ be a smooth projective curve over $\mathbb{Q}$. Its associated L-function is defined by

$$ L(C, s)=\prod_{p \text{ prime}} L_p(C, s), $$

where, if $p$ is a prime of good reduction, $L_p(C, s)=Q_p(p^{-s})^{-1}$ for $Q_p$ the polynomial appearing in the numerator of the zeta function of the reduction of $C$ over $\mathbb{F}_p$.

I would like to prove the statement

$L(C, s)$ converges absolutely for $\mathrm{Re}(s)>3/2$

I guess one should combine two ingredients:

1) the control of the zeros of $Q_p(p^{-s})$ which, by the Riemann hypothesis, have real part $\mathrm{Re}(s)=1/2$

2) some knowledge about the bad factors, i.e. that they do not introduce poles in the region $\mathrm{Re}(s)>3/2$

I have two questions:

1) can somebody isolate the analytic statement behind the convergence of the product over good primes (this should be something like I have a Dirichlet series with an Euler product, I know the real part of the zeros, then the convergence is...)

2) is it known or conjectural that the bad factors do not introduced poles?


  • 1
    $\begingroup$ If I'm not mistaken, the bad factors are L-functions of singular curves over finite fields, so the real parts of their zeroes and poles satisfy the usual cohomological bound. Also, see section 3.3 of maths.bris.ac.uk/~madjmdc/hindry-zeta.pdf $\endgroup$ – S. Carnahan Sep 24 '14 at 13:30
  • $\begingroup$ @S.Carnahan your link is broken. Maybe section 2.4 of webusers.imj-prg.fr/~marc.hindry/Notes_rev_Brasilia.pdf would be a substitute? $\endgroup$ – KConrad Jun 20 '18 at 20:25
  • $\begingroup$ @KConrad Thank you - that seems to be more or less the same document. Section 3.3 still has the same estimates. $\endgroup$ – S. Carnahan Jun 21 '18 at 4:47

You can find a discussion of the convergence of Hasse-Weil zeta and L functions in Chapter 16 of Husemöller's Elliptic Curves. In particular, the fact that the product over good primes satisfies normal convergence on $\Re(s) > 3/2$ is Proposition 2.3 in Chapter 16. You can find a general discussion of convergence of Euler products in Chapter 11 section 6 - one simply estimates the growth of the terms in the Dirichlet series.

Bad fibers of curves have no high-weight pieces, so their local factors don't have zeroes or poles in $\Re(s) > 3/2$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.