Weak Del Pezzo Surfaces
- Weak del Pezzo surfaces are smooth projective surfaces defined by a nef and big anticanonical divisor, classified by degree values from 1 to 9.
- They feature structured Picard groups and ADE-type root systems, with (-1) and (-2)-curves determining the singularities on their anticanonical models.
- Their derived categories support full exceptional collections and 2-tilting bundles, linking geometric constructions to non-commutative crepant resolutions.
A weak del Pezzo surface is a smooth projective surface over an algebraically closed field such that its anticanonical divisor is nef and big, i.e., for all irreducible curves , , and . If is ample, is called a (strong) del Pezzo surface. The classification, geometry, and homological structures of weak del Pezzo surfaces link classical Lie theory, birational geometry, and modern non-commutative algebra. Their subtlety lies in the possible presence of -curves, mild surface singularities of Du Val type on their anticanonical models, and the richness of their derived categories.
1. Anticanonical Model and Classification
Given as above, the degree satisfies 0. Up to isomorphism, every weak del Pezzo surface is either:
- 1 (degree 8), the Hirzebruch surface 2 (degree 8), or
- the blow-up of 3 in 4 points in "almost general position" (no three collinear, etc., for the strong case, with relaxations in the weak case allowing special configurations giving 5-curves) (Tomonaga, 30 Oct 2025).
The anticanonical linear system 6 is base point free; it gives rise to the anticanonical map 7, whose image 8 is the normal anticanonical model. If 9 is not ample, then 0 acquires rational double point (ADE) singularities, each coming from contracting an effective 1-curve on 2 (Blache et al., 2023).
The negative curves on 3 are only 4- and 5-curves. The set of 6-classes (those 7 with 8) always forms a root system of ADE-type, its configuration governed by the combinatorics of 9 and its subsystems, depending on the degree and singularity content (Elagin et al., 2017).
2. Picard Group Structure and Root Subsystems
The Picard group 0 for a weak del Pezzo surface blown up in 1 points has basis 2 (class of a line, exceptional divisors), with intersection numbers 3, 4, 5, 6 (7), and
8
with 9 (Perepechko, 2023, Tomonaga, 30 Oct 2025).
The root system of 0-classes 1 has ADE-type depending on degree: | 2 | 3 type | |---|-------------| | 7 | 4 | | 6 | 5 | | 5 | 6 | | 4 | 7 | | 3 | 8 | | 2 | 9 | | 1 | 0 |
The configuration of effective roots encodes the singularities on 1; the subsystems, through their Dynkin diagrams, describe precisely which ADE singularities are present (Lubbes, 2013, Lubbes, 2018, Elagin et al., 2017).
3. Construction by Blow-Up and Special Loci
Any weak del Pezzo surface of degree 2 is the blow-up of 3 at 4 points 5 in almost general position. Special alignments among these points (three collinear, six on a conic, points infinitely near, etc.) produce effective 6-curves, whose classes are explicitly described (e.g., 7 for three collinear). Each configuration maps directly to a root subsystem, and hence to the Du Val singularities on the anticanonical image 8.
A notably efficient encoding of such singularity types is via C1 labels—combinatorial structures labeling root subsystem types—which classify isomorphism classes of weak del Pezzo surfaces up to Weyl group action (Lubbes, 2013). These enable algorithmic construction of weak del Pezzo surfaces with prescribed singularity configuration.
4. Homological and Non-Commutative Structures
For any weak del Pezzo surface 9, the derived category 0 admits full exceptional collections of line bundles, and the existence of a 1-tilting bundle (a vector bundle whose endomorphism algebra has 2) is equivalent to being weak del Pezzo. Concretely:
- If 3 is 4-tilting, 5 is 6-representation-infinite, so its 7-Calabi–Yau completion (Ginzburg DG algebra) yields a non-commutative crepant resolution (NCCR) of the corresponding Du Val cone 8 of 9 (Tomonaga, 30 Oct 2025).
- The classical Beilinson and extended Dynkin quiver algebras appear as endomorphism algebras in degrees 0 and lower. The tilting bundle construction proceeds via mutation of full exceptional collections followed by universal extensions to kill higher Exts.
This connects the geometry of weak del Pezzo surfaces to the representation theory of finite-dimensional algebras and non-commutative algebraic geometry, especially via the parallel between geometric blown-up models and NCCR theory.
5. Real, Arithmetic, and Moduli Aspects
The classification of weak del Pezzo surfaces over the real numbers or finite fields naturally extends to their moduli, arithmetic, and automorphism content:
- Real Forms: Real weak del Pezzo surfaces are classified by pairs of Dynkin diagrams encoding the invariants under the real involution and the arrangement of real effective root subsystems. Explicit formulas describe the associated Néron–Severi lattices, root system splittings, and cone polyhedral structure (Lubbes, 2018).
- Arithmetic Types: Over finite fields, the Frobenius action on the Picard group, decomposed into conjugacy classes in the Weyl group (stabilizers of root subsystems), defines the "arithmetic type" of the surface. The explicit realization of all types (at least for 1) has been achieved, often via blow-up constructions and pencils of quadrics, with point-count and zeta function formulas depending only on this type (Blache et al., 2023).
- Moduli Stratification: The moduli of weak del Pezzo surfaces stratify according to root subsystem type and arithmetic invariants, with further refinement given by real structure, configuration of 2- and 3-curves, and automorphism group actions.
6. Applications: Codes, Universal Torsors, and Affine Cones
Weak del Pezzo surfaces appear in several applied and explicit contexts:
- Algebraic Geometric Codes: Codes are constructed by evaluating sections of the anticanonical sheaf at rational points of 4, with minimum distance depending on the structure of 5-curves and genus-1 linear systems. The arithmetic of blow-ups and contractions controls the rational point count algorithmically (Blache et al., 2023).
- Universal Torsors and Cox Rings: For an explicit weak del Pezzo pair 6, universal torsor and Cox ring descriptions provide a platform for studying integral points, heights, and forms of Manin's conjecture. For example, for quartic surfaces with 7 singularities, integral point counting corresponds to lattice points in certain polytopes defined via the universal torsor parameterization (Derenthal et al., 2021).
- Automorphism and Flexibility: Affine cones over weak del Pezzo surfaces polarized by very ample divisors may be generically flexible—i.e., have infinitely transitive automorphism group actions—when there exist sufficient cylinder structures covering the ample cone. This property is amenable to combinatorial and algorithmic verification (Perepechko, 2023).
- Fibrations and Cylinders: The existence of 8-cylinders in fibrations of weak del Pezzo surfaces is highly constrained, with only degree 9 minimal weak del Pezzo surfaces (such as 0 or certain 1-type) admitting such cylinders (Sawahara, 2019).
7. Singularities, Log Versions, and Volume Bounds
A weak (log) del Pezzo surface may be equipped with a boundary divisor 2 such that 3 is big and nef. The singularities of the anticanonical model 4 are rational double (ADE) points; the classification of possible log pairs, volume bounds 5, and their relation to the Picard number of minimal resolutions have been completely worked out. For 6-lc pairs, optimal upper bounds on the anticanonical volume are achieved only in the 7 or special Hirzebruch surface cases (Jiang, 2013). These bounds are sensitive to the number and arrangement of negative curves and directly inform classification and boundedness results in Fano geometry.
This synthesis demonstrates that weak del Pezzo surfaces form a central class linking birational surface theory, Lie-theoretic combinatorics, derived category methods, and applied algebraic geometry across arithmetic, geometric, and homological dimensions (Tomonaga, 30 Oct 2025, Lubbes, 2013, Elagin et al., 2017, Blache et al., 2023, Perepechko, 2023, Lubbes, 2018, Sawahara, 2019, Derenthal et al., 2021, Jiang, 2013, Blache et al., 2023).