Non-arithmetic lattices and the Klein quartic (Q2272904)

Thanks to the superrigidity and arithmeticity theorems, the only possibilities for an abundance of non-arithmetic lattices are in the rank one groups \(\mathrm{SO}(n,1)\) or \(\mathrm{SU}(n,1)\). In 1988, Gromov and Piatetskii-Shapiro introduced a hybridization procedure to produce infinitely many commensurability classes of non-arithmetic lattices in \(\mathrm{SO}(n,1)\) for any \(n \geq 2\). It is not known how to carry over such a method in the case of \(\mathrm{SU}(n,1)\). However, already in 1980, Mostow constructed non-arithmetic lattices in \(\mathrm{SU}(2,1)\) defined as complex reflection groups. These are related to monodromy groups of hypergeometric functions studied by Picard. Using this interpretation, Deligne and Mostow extended this construction to produce some non-arithmetic lattices in \(\mathrm{SU}(2,1)\) as well as a single one in \(\mathrm{SU}(3,1)\). Till date, this is the only non-arithmetic latice known in any \(\mathrm{SU}(n,1)\) for \(n \geq 3\). The Deligne-Mostow non-arithmetic lattices in \(\mathrm{SU}(2,1)\) are known to fall into exactly 9 commensurability classes. In 2016, almost 30 years after the Deligne-Mostow examples, the author in collaboration with \textit{J. R. Parker} and \textit{J. Paupert} [Invent. Math. 203, No. 3, 681--771 (2016; Zbl 1337.22007)] produced five new commensurability classes of non-arithmetic lattices in \(\mathrm{SU}(2,1)\). In that work, they followed the basic spirit of Mostow's original approach. They started with carefully chosen complex reflection groups and showed them to be lattices by constructing fundamental domains for them using Poincaré's polyhedron theorem. The non-arithmeticity is shown using Vinberg's arithmeticity criterion involving the trace fields. The aim of the present paper is to give a new proof of the construction of the author and Parker and Paupert mentioned above. This new proof avoids discussion of fundamental domains altogether. This is achieved by a new realization that the lattices in question are actually associated to the automorphism group of the Klein quartic that attains the Hurwitz bound \(168\). In order to realize this identification, the author uses a description due to Naruki of the fundamental group of the complement of the Klein configuration of mirrors in \(\mathbb{P}^2\). The author's proof of the main result is obtained by checking that equality holds in the logarithmic Bogomolov-Miyaoka-Yau inequality. In particular, a substantial part of the paper is towards checking that the hypotheses of a result of Kobayashi, Nakamura and Sakai holds good. The introduction of the paper very lucidly explains the basic approach as well as the complications involved.