Holomorphic vector bundles over compact complex surfaces (Q1912497)

In modern algebraic and analytic geometry, the notion of vector bundles is of primary rank, having the same status, and in some sense dual, to that of subvariety. When studying a single compact complex manifold \(X\) two natural questions appear from the very beginning: which smooth vector bundles admit analytic structures and how to parametrize all or parts of the set of analytic vector bundles on \(X\). Even on surfaces (i.e. \(\dim_C X = 2)\) the above two questions are hard and they remain partially open. The book under review reports on the latest achievements in solving these questions on compact analytic surfaces. A vector bundle \(E\) on a compact complex surface \(X\) has three complete topological invariants: the rank \(r(E) \in \mathbb{N}\) and the Chern classes \(c_1 (E) \in H^2 (X, \mathbb{Z})\), \(c_2 (E) \in H^4 (X, \mathbb{Z})\) \((= \mathbb{Z})\). If \(E\) admits an analytic structure already \(c_1 (E)\) is restricted to belong to the Neron-Severi subgroup \(NS(X)\) of \(H^2 (X, \mathbb{Z})\). More precisely, \(NS(X)\) is the subgroup of \(H^2 (X, \mathbb{Z})\) generated by \(c_1 (L)\), where \(L\) runs over all holomorphic line bundles on \(X\). A classical result of Schwarzenberger asserts that the converse is also true, on every algebraic surface \(X\) (that is \(E\) carries an analytic structure if and only if \(c_1 (E) \in NS(X))\) [see \textit{R. L. E. Schwarzenberger}, Proc. Lond. Math. Soc., III. Ser. 11, 601-622 (1961; Zbl 0212.26003)]. For non-algebraic surfaces the situation is much more complicated, and this is actually the subject of the book by Brînzănescu. The strategy proposed by the author is roughly the following: compute the Neron-Severi group \(NS(X)\), then find the domain of the Chern classes of analytic vector bundles on \(X\) (given in general by inequalities derived from Riemann-Roch type formulas), and finally produce by a constructive method enough many analytic vector bundles on \(X\), so that all possible topological types are filled up. So far, this method proved to be successful for filtrable analytic vector bundles and under some geometric restriction on \(X\) [see \textit{C. Bănică} and \textit{J. Le Potier}, J. Reine Angew. Math. 378, 1-31 (1987; Zbl 0624.32017); \textit{M. Toma}, Publ. Res. Inst. Math. Sci. 27, No. 4, 533-550 (1991; Zbl 0749.32021), the author and \textit{P. Flondor}, Rev. Roum. Math. Pures Appl. 31, 507-512 (1986; Zbl 0617.14009)]. The contents of the book is the following. The first two chapters are devoted to preliminaries: Chern classes, coherent sheaves, the Enriques-Kodaira classification of surfaces, the Picard group. Chapter 3 contains detailed computations of the Neron-Severi group for tori, some elliptic surfaces and for primary Kodaira surfaces. Here the techniques are mainly from classical (by now) algebraic geometry. Chapter 4 collects all the known existence theorems, mainly for filtrable holomorphic structures on vector bundles. A quick review of Serre's extension method is added here. The last chapter is a survey (only with sketches of proofs) of the recent progress made in constructing and understanding moduli spaces for some classes of analytic vector bundles on surfaces. Here the techniques are more sophisticated, by combining algebraic geometry with differential geometry, gauge theory and analysis. The author belongs to the school of the late distinguished geometer Constantin Bănică. Vasile Brînzănescu has constantly contributed to this field in the last decade and now his students are continuing his work. His expertize and original thought are transparent in every page of the text. The style is extremely clear, the presentation is sober and lucid. Beginners will find the first two chapters very valuable and agreeable to read as an introduction to non-algebraic compact complex surfaces. For the experts, the whole book will be a useful reference. In short, a definite contribution to modern analytic geometry.