Projective superspace varieties, superspace quadrics and non-splitting (Q2176907)

The paper studies the splitting problem of projective superspace varieties. An affine superspace subvariety is proved to be isomorphic to its split model, i.e, it is split. However, in projective case, it fails in general. One tool for this study this problem, is the super manifold non-splitting theorem. This theorem says a complex supermanifold with a non-vanishing obstruction in degree \(2\), is non-split. If a ringed space \(\mathfrak{X}=(X,\mathcal{O}_\mathfrak{X})\) is split with \(\mathcal{O}_\mathfrak{X}=\bigwedge T^*_{X,-}\), we say \(S(X,T^*_{X,-})\) is the split model associated to \(\mathfrak{X}=(X,\mathcal{O}_{\mathfrak{X}})\). One can associate an obstruction sheaf \(\mathcal{Q}_{T^*_{X,-}}\) to each model \((X,T^*_{X,-})\). If the degree of \(\mathcal{O}_{\mathfrak{X}}\) is even \(2k\),concerning \(\mathbb{Z}\)-grading, we have an isomorphism \(\mathcal{Q}_{T^*_{X,-}}^{(2k)}\cong T_X\otimes \bigwedge^{2k}T^*_{X,-}\), where \(T_X\) is the tangent space to \(X\). So in this paper, the main study is on the even degree cases. Define the projective superspace as follows. Let \(\overset{\rightarrow}{b}=(b_1,\cdots,b_n)\) be an \(n\)-tuple of integers. Define the action of \(\mathbb{C}^{\times}\) on \(\mathbb{C}^{m+1|n}\) with the coordinate \((x^\mu,\theta_a)\) to be \[(x^\mu,\theta_a)\stackrel{\lambda}{\mapsto}(\lambda x^\mu,\lambda^{b_a}\theta_a).\] The quotient space of \(\mathbb{C}^{m+1|n}-{(0|0)}\) under this action, which is denoted by \(\mathbb{P}_{\mathbb{C}}^{m|n}(1|\overset{\rightarrow}{b})\), is a super (analytic) space. Theorem 1. For \(n\)-tuples of integers \(\overset{\rightarrow}{b}=(b_1,\cdots,b_n)\), e have the following isomorphism. \[\mathbb{P}_{\mathbb{C}}^{m|n}(1|\overset{\rightarrow}{b})\cong S(\mathbb{P}_{\mathbb{C}}^m; \oplus_j \mathcal{O}_{\mathbb{P}_{\mathbb{C}}^m}(-b_j)).\] If \(F=(f^\alpha)\) be a finite collection of homogeneous polynomials with even degree, such that each \(f^\alpha\) is irreducible (means \(f^\alpha(x|0)\) is irreducible), then \(V(F)=\{(x|\theta)| f^\alpha(x|\theta)=0,\, \forall \alpha\}\) is a projective suerspace subvariety of \(\mathbb{P}_{\mathbb{C}}^{m|n}(1|\overset{\rightarrow}{b})\). If \(V(F)\) is irreducible, it is called homogeneously non-reduced if there exist an \(\alpha\) and a \(k\) such that \(\dfrac{\partial f^\alpha}{\partial\theta_k}\neq 0\). Lemma 2. If a sub variety of \(\mathbb{P}_{\mathbb{C}}^{m|n}(1|\overset{\rightarrow}{b})\) be homogeneously non-reduced then it is non-split. Then the attention is turn out to the quadratic surfaces in projective superspaces. Theorem 3. Let \(Q\) be a quadratic non-singular homogeneously non-reduced hypersurface in a projective superspace. Then \(Q\) is non-split. Using the same Segre embedding, the results are generalised to the product spaces of projective superspaces. Theorem 4. Let \(\overset{\rightarrow}{b}\) and \(\overset{\rightarrow}{b'}\) be two \(n\)-tuples of positive integers, defined as above. The following embedding is defined \[\mathbb{P}_{\mathbb{C}}^{m|n}(1|\overset{\rightarrow}{b})\times \mathbb{P}_{\mathbb{C}}^{m'|n'}(1|\overset{\rightarrow}{b'})\hookrightarrow \mathbb{P}_{\mathbb{C}}^{m''|n''}(1|\overset{\rightarrow}{b''}),\] for some positive \(m''\), \(n''\), and \(\overset{\rightarrow}{b''}\). Corollary 5. Let \(Q\) be a quadratic non-singular homogeneously non-reduced hypersurface in the product of two projective superspaces. Then \(Q\) is non-split.