Fake quadrics from irreducible lattices acting on the product of upper half planes (Q464135)

A fake quadric is a minimal surface of general type with the invariants of a quadric, i.e.\ \(q=p_g=0\), \(c_1^2=8\), \(c_2=4\). One method of constructing them is to take a quotient of \({\mathbb H}\times {\mathbb H}\) by a suitable subgroup \(\Gamma\) of \(\mathrm{PSL}(2,{\mathbb R})\times\mathrm{PSL}(2,{\mathbb R})\). The aim here is to take some steps towards classifying the fake quadrics that arise in this way. It is not expected that all fake quadrics arise in this way: indeed there are examples that are quotients of products of curves, which are not of this type although they do have the same universal cover. The methods parallels the approach of \textit{G. Prasad} and \textit{S.-K. Yeung} [Invent. Math. 168, No. 2, 321--370 (2007; Zbl 1253.14034)] to classifying fake projective planes. The groups \(\Gamma\) is a cocompact lattice and \textit{G. A. Margulis} [Discrete subgroups of semisimple Lie groups. Berlin etc.: Springer-Verlag (1991; Zbl 0732.22008)] showed that all such lattices are arithmetic. Associated with \(\Gamma\) is an invariant \(\chi(\Gamma)\), equal to the Euler characteristic of the quotient if \(\Gamma\) is torsion-free, so one needs to classify such lattices with \(\chi(\Gamma)=1\). Such a lattice has a finite index overlattice \(\Delta\) with \(\chi(\Delta)\leq 1\), and one studies the arithmetic data that define \(\Delta\). These data include a totally real number field \(k\) and the first stage in this paper serves to bound the degree (and some other parts of the data). The bound given here is \(\deg k\leq 33\), or \(\deg k\leq 25\) under GRH, but improvements are possible although that aspect is not the focus here. Instead, the author examines in detail the case where \(k\) is a quadratic field, and bounds the discriminant. In the end, the paper lists 39 possible conjugacy classes of lattices with \(k\) a quadratic field: the biggest discriminants arising are \(13\), \(15\), \(17\), \(21\) and \(24\). In all but two cases the class is non-empty (the question being still open in those other two cases). It should be noted, however, that different lattices in the same conjugacy class may lead to non-isomorphic fake quadrics, and the present paper does not attempt a classification at that level. The results are thus for several reasons some way from giving a classification, but they are quite strong and the paper also suggests how to make further progress.