Searcharxiv⌕ Search

arXiv subjects

Stefan Schroeer

Publications and source records attributed to Stefan Schroeer.

At least 37 records · Page 2Linked to original sources
arXiv2005

Toroidal crossings and logarithmic structures

We generalize Friedman's notion of d-semistability, which is a necessary condition for spaces with normal crossings to admit smoothings with regular total space. Our generalization deals with spaces that locally look like the boundary divisor in Gorenstein toroidal embeddings. In this situation, we replace d-semistability by the existence of global log structures for a given gerbe of local log structures. This leads to cohomological descriptions for the obstructions, existence, and automorphisms of log structures. We also apply toroidal crossings to mirror symmetry, by giving a duality construction involving toroidal crossing varieties whose irreducible components are toric varieties. This duality reproduces a version of Batyrev's construction of mirror pairs for hypersurfaces in toric varieties, but it applies to a larger class, including degenerate abelian varieties.

math.AG↗
arXiv2005

Topological methods for complex-analytic Brauer groups

Using methods from algebraic topology and group cohomology, I pursue Grothendieck's question on equality of geometric and cohomological Brauer groups in the context of complex-analytic spaces. The main result is that equality holds under suitable assumptions on the fundamental group and the Pontrjagin dual of the second homotopy group. I apply this to Lie groups, Hopf manifolds, and complex-analytic surfaces.

math.AG↗
arXiv2003

Curves with only triple ramification

I show that the set of smooth curves of genus g admitting branched coverings X->P^1 with only triple ramification points has dimension at least max(2g-3,g). In characteristic two, such curves have tame rational functions and an analog of Belyi's Theorem applies to them.

math.AG↗
arXiv2003

Some Calabi-Yau threefolds with obstructed deformations over the Witt vectors

I construct some smooth Calabi-Yau threefolds in characteristic two and three that do not lift to characteristic zero. These threefolds are pencils of supersingular K3-surfaces. The construction depends on Moret-Bailly's pencil of abelian surfaces and Katsura's analysis of generalized Kummer surfaces. The threefold in characteristic two turns out to be nonrigid.

math.AG↗
arXiv2003

The Brauer group of analytic K3 surfaces

We show that for complex analytic K3 surfaces any torsion class in H^2(X,O_X^*) comes from an Azumaya algebra. In other words, the Brauer group equals the cohomological Brauer group. For algebraic surfaces, such results go back to Grothendieck. In our situation, we use twistor spaces to deform a given analytic K3 surface to suitable projective K3 surfaces, and then stable bundles and hyperholomorphy conditions to pass back and forth between the members of the twistor family.

math.AG↗
arXiv2003

The strong Franchetta Conjecture in arbitrary characteristics

Using Moriwaki's calculation of the Q-Picard group for the moduli space of curves, I prove the strong Franchetta Conjecture in all characteristics. That is, the canonical class generates the group of rational points on the Picard scheme for the generic curve of genus g>2. Similar results hold for generic pointed curves. Moreover, I show that Hilbert's Irreducibility Theorem implies that there are many other nonclosed points in the moduli space of curves with such properties.

math.AG↗
arXiv2002

The bigger Brauer group is really big

I show that each etale n-cohomology class on noetherian schemes comes from a Cech cocycle, provided that any n-tuple of points admits an affine open neighborhood. Together with results of Raeburn and Taylor on the bigger Brauer group, this implies that for schemes such that each pair of points admits an affine open neighborhood, any etale G_m-gerbe comes from a central separable algebra. Such algebras are nonunital generalizations of Azumaya algebras. I also prove that, on normal noetherian schemes, each Zariski G_m-gerbe comes from a central separable algebra.

math.AG↗
arXiv2002

Existence of vector bundles and global resolutions for singular surfaces

We prove two results about vector bundles on singular algebraic surfaces. First, on proper surfaces there are vector bundles of rank two with arbitrarily large second Chern number and fixed determinant. Second, on separated normal surfaces any coherent sheaf is the quotient of a vector bundle. As a consequence, for such surfaces the Quillen K-theory of vector bundles coincides with the Waldhausen K-theory of perfect complexes. Examples show that, on nonseparated schemes, usually many coherent sheaves are not quotients of vector bundles.

math.AG↗
arXiv2001

Hilbert's Theorem 90 and algebraic spaces

In modern form, Hilbert's Theorem 90 tells us that R^1f_*(G_m)=0, where f is the canonical map between the etale site and the Zariski site of a scheme X. I construct examples showing that the corresponding statement for algebraic spaces does not hold. The first example is a nonseparated smooth 1-dimensional bug-eyed cover in Kollar's sense. The second example is a nonnormal proper algebraic space obtained by identifying points on suitable nonprojective smooth proper schemes.

math.AG↗
arXiv2001

The T^1-lifting theorem in positive characteristic

Replacing symmetric powers by divided powers and working over Witt vectors instead of ground fields, I generalize Kawamatas T^1-lifting theorem to characteristic p>0. Combined with the work of Deligne-Illusie on degeneration of the Hodge-de Rham spectral sequences, this gives unobstructedness for certain Calabi-Yau varieties with free crystalline cohomology modules.

math.AG↗
arXiv2001

What is missing in canonical models for proper normal algebraic surfaces?

Smooth surfaces have finitely generated canonical rings and projective canonical models. For normal surfaces, however, the graded ring of multicanonical sections is possibly nonnoetherian, such that the corresponding homogeneous spectrum is noncompact. I construct a canonical compactification by adding finitely many non-Q-Gorenstein points at infinity, provided that each Weil divisor is numerically equivalent to a Q-Cartier divisor. Similar results hold for arbitrary Weil divisors instead of the canonical class.

math.AG↗
arXiv2001

Normal del Pezzo surfaces containing a nonrational singularity

Working over a perfect field, I classify normal del Pezzo surfaces with base number one that contain a nonrational singularity. They form a huge infinite hierarchy; contractions of ruled surfaces lie on top of it. Descending the hierarchy hinges on a generalized version of elementary transformations. As an application, I determine the structure of 2-dimensional anticanonical models for arbitrary normal surfaces.

math.AG↗
arXiv2001

A characterization of semiampleness and contractions of relative curves

A line bundle with a base-point-free multiple is called semiample. I give a cohomological characterization of semiample line bundles. The result is a common generalization of the Fujita-Zariski criterion for semiampleness and the Grothendieck-Serre characterization of ampleness. Applying the Fujita-Zariski Theorem, I characterize contractible curves in 1-dimensional families.

math.AG↗
arXiv2001

There are enough Azumaya algebras on surfaces

Using Maruyama's theory of elementary transformations, I show that the Brauer group surjects onto the cohomological Brauer group for separated geometrically normal algebraic surfaces. As an application, I infer the existence of nonfree vector bundles on proper normal algebraic surfaces.

math.AG↗
arXiv2000

Remarks on the existence of Cartier divisors

Given an invertible sheaf, does it come from a Cartier divisor? This might fail in presence of embedded components. I give some examples and characterize those invertible sheaves that allow a Cartier divisor.

math.AG↗