Grothendieck-Lefschetz theory, set-theoretic complete intersections and rational normal scrolls (Q615850)

The authors study set-theoretical complete intersection varieties, defined over fields of arbitrary characteristic. For a projective variety \(Y\), the authors define the invariant \(\mathrm{ara}(Y)\) (the arithmetic rank of \(Y\)) as the minimal cardinality of a set of polynomials \(S\) whose zero-locus coincides with \(Y\). Thus, \(Y\) is set-theoretical complete intersection exactly when \(\mathrm{ara}(Y)\) equals the codimension of \(Y\). Using an algebraic approach to the Grothendieck-Lefschetz theory, the authors determine several necessary conditions on the Picard group of a variety, to be set theoretical-complete intersection. Using these characterizations, the authors are able to produce many examples of surfaces, which are not set-theoretical complete intersection. The authors point then their attention to rational normal scrolls. Rational normal scrolls of dimension \(2\) are known to be set-theoretical complete intersection. For any dimension, the authors show that a rational normal scroll embedded in \(\mathbb P^N\) has \(\mathrm{ara}(Y)=N-2\). It follows that rational normal scrolls of dimension greater than two are not set-theoretical complete intersection.