Genus-zero and genus-one string amplitudes and special multiple zeta values (Q2308286): Difference between revisions

In the article, the authors address several open problems and provide proofs for conjectures regarding connections and relations between four-point open- and closed string amplitudes at genus zero (``tree-level'') and two-point open- and closed string-amplitudes at genus one. While relations between the former have been discussed constantly during the last three decades, the latter only got renewed interest rather recently. Starting from the known expansions in single zeta values for the four-point genus-zero amplitudes by Veneziano and Virasoro, the authors identify a particular subspace of multiple zeta values, which includes all Taylor coefficients in the symmetric expansion of the Virasoro amplitude. This particular representation in terms of special linear combinations of multiple zeta values \(Z(k,r)\) allows to bridge the gap to the genus-one two-point amplitudes (a.k.a. two-vertex modular graph functions): their Laurent-polynomials have been shown to exhibit Taylor coefficients in precisely the subspace of multiple zeta values spanned by \(Z(k,r)\). In the first theorem of the paper the authors state the explicit representation of the Taylor coefficients \(e_{p,q}\) in terms of the combinations \(Z(k,r)\), thereby proving equivalence of the two spaces spanned. Repeating similar considerations for open genus-zero and genus-one amplitudes allows, an approach using suitable generating functions allows to show that the ``single-valued-projection'' allows to infer the closed-string one-loop Laurent-polynomials form their open-string counterpart (for the amplitudes mentioned above, although it is conjectured to hold in more generality). On the way, many additional results are provided closing gaps in the literature: the special linear combinations of multiple zeta values in the expansion of the Veneziano amplitude \(H(k,r)\) appearing in the expansion in Mandelstam variables \(s\) and \(t\). In the second main theorem proven in this article the authors state that the coefficients of the Laurent polynomial for closed two-point one-loop modular graph functions belong to the ring over \(\mathbb{Q}\) generated by odd Riemann zeta values. The corresponding proof relies on the ability to relate these coefficients \(\gamma_{n,k}\) to the coefficients \(e_{p,q}\) from the first part using a (non-trivial) polynomial relation. Both proofs are spelt out in detail and it is -- sometimes an computer algebra system is helpful -- straightforward to follow the carefully drafted steps. Exposition is dense at times, however, well balanced given the complexity of the manipulations. Many different subspaces of multiple zeta values are defined and compared in due course. Still, the complexity of the notation is always justified. Most importantly, mathematical implications are clearly communicated and related to the appropriate references. In addition, where necessary, links to the physics literature dealing with string theory amplitudes are drawn. Several of the results have been conjectured or -- in case of [\textit{E. D'Hoker} and \textit{M. B. Green}, ``Absence of irreducible multiple zeta-values in melon modular graph functions'', Preprint, \url{arXiv:1904.06603}] -- proven on different ways. Overall the article provides a thorough and detailed numer-theoretic account with proofs of all parts of the procedures lined out in the graph on page 418, thereby complementing parts of the analytic account on the subject in the recent paper by \textit{F. Brown} and \textit{C. Dupont} [``Single-valued integration and superstring amplitudes in genus zero'', Preprint, \url{arXiv:1910.01107}] and extending it to a special class of open- and closed-string one-loop amplitudes. The article is at very good typographical level. Here is a list of some format recommendations: page 417, Theorem B: \(l \to \ell\) (that is important, actually) page 426, in the first line of the cases environment, better use \(\setminus\) page 432, correct superfluous space at the end of the first paragraph before reference [\textit{M. B. Green} et al., ``Low energy expansion of the four-particle genus-one amplitude intype II superstring theory'', JHEP 2008, Article Id 0802, 57 p. (2008; \url{doi:10.1088/1126-6708/2008/02/020})]