Product-quotient surfaces: new invariants and algorithms (Q260120): Difference between revisions

A \textit{product-quotient surface} \(S\) is the minimal resolution of the singularities of \(X:=(C_1\times C_2)/G\), where \(C_1\), \(C_2\) are two Riemann surfaces of genus at least two and \(G\) is a finite group acting on \(C_1\times C_2\) diagonally. \(X\) is called the \textit{quotient model} of the product-quotient surface. In the last years several authors have been studying product-quotient surfaces, quite some literature is nowadays available and many surfaces of general type have been constructed in this way (especially in the boundary case \(\chi=1\), i.e. \(p_g=q\)). It is interesting to classify all product-quotient surfaces of general type with \(p_g=0\). In the case when \(S\) is minimal, this has been done in a previous paper by the authors [Math. Comput. 81, No. 280, 2389--2418 (2012; Zbl 1306.14017)]. The paper under review aims to classify all non-minimal product-quotient surfaces of general type with geometric genus zero. In order to achieve such classification, the authors suggest a new approach to the systematic, computer-aided construction, introducing a new invariant: the integer \(\gamma\), which depends only on the singularities of the quotient model \( X=(C_1\times C_2)/G\). The key observation is that for all minimal product-quotient surfaces, \(H^{1,1}(S)\) is generated by algebraic curves coming from the construction: the classes of the two fibers and the Hirzebruch-Jung strings arising from the minimal resolution of singularities of \(X\); whereas for the single already known non-minimal example ([loc. cit.]) this is not the case. The authors study then the subspace of \(H^{1,1}(S)\) generated by the divisors mentioned above, and prove that its codimension is even and equal to \(2(p_g(S)+\gamma)\). Using this new insight, the authors developed and implemented an algorithm in the computer algebra software MAGMA which constructs all regular product-quotient surfaces with given values of \(\gamma\) and \(p_g\). Running the program for \(\gamma=1,2,3 \), they construct a substantial number of new regular product-quotient surfaces with \(p_g=0\), raising the number of known families of product-quotient surfaces of general type with \(p_g=0\) to 75. The authors conjecture that this is the complete list of product-quotient surfaces of general type with \(p_g=q=0\). They also formulate other two interesting conjectures: (i) a product-quotient surface \(S\) is minimal if and only if \(p_g(S)+\gamma=0\); (ii) there exists an explicit function \(\Gamma(p_g,q)\) such that, for the quotient model \(X\) of every product-quotient surfaces \(S\) of general type it holds \(\gamma\leq \Gamma(p_g(S),q(S))\). The former conjecture is proved for \(p_g(S)=0\), and the latter under some additional hypotheses. Finally the authors introduce a duality among product-quotient surfaces and prove that the dual surface of a surface of geometric genus zero has maximal Picard number, thus providing several new examples of surfaces with maximal Picard number.