A geometric perspective on plus-one generated arrangements of lines

1. Problem Statement and Context

The paper addresses the classification and characterization of plus-one generated (POG) arrangements of lines in the complex projective plane $\mathbb{P}^2$. This research sits within the broader context of the Terao Conjecture, which posits that the freeness of a hyperplane arrangement depends only on its combinatorial intersection lattice.

While "free" arrangements are well-understood (their associated vector bundles split as a direct sum of line bundles), the authors focus on arrangements that are "next-to-free" (NT-free). These are arrangements obtained by either deleting a line from a free arrangement or adding a line to a free arrangement.

• The Gap: Previous work (specifically by Abe) established that NT-free arrangements are either free or POG. However, the converse was not fully characterized: Under what geometric conditions is a POG arrangement actually NT-free?
• The Goal: To provide a geometric characterization of POG arrangements that are NT-free, moving beyond combinatorial conditions to properties of the associated vector bundles and their splitting types.

2. Methodology

The authors employ a geometric approach centered on the vector bundle associated to the arrangement ($T_A$), rather than purely algebraic or combinatorial methods.

• Vector Bundle Framework: They utilize the kernel of the Jacobian map, $T_A$, which is a rank 2 reflexive sheaf (vector bundle) over $\mathbb{P}^2$.
• Splitting Types: A core tool is Grothendieck's splitting theorem. For any line $l \subset \mathbb{P}^2$, the restriction $T_A|_l$ splits into a sum of line bundles $\mathcal{O}_l(\alpha) \oplus \mathcal{O}_l(\beta)$. The pair $(\alpha, \beta)$ is the splitting type.
• Resolution Analysis: They analyze the specific resolution of POG bundles:
  $$0 \to \mathcal{O}_{\mathbb{P}^2}(-1-d) \to \mathcal{O}_{\mathbb{P}^2}(-d) \oplus \mathcal{O}_{\mathbb{P}^2}(-b) \oplus \mathcal{O}_{\mathbb{P}^2}(-a) \to T_A \to 0$$
  where $(a, b)$ are the exponents and $d$ is the level.
• Ziegler Restrictions: The paper connects the splitting type of $T_A$ on a line $l$ to the exponents of the Ziegler restriction of the arrangement onto $l$. This allows them to translate bundle properties into arrangement properties.
• Exact Sequences: The authors use short exact sequences relating $T_A$ and $T_{A \setminus \{l\}}$ (deletion) or $T_{A \cup \{l\}}$ (addition) to derive constraints on the number of intersection points ($h$) on a line.

3. Key Contributions and Results

A. Characterization of Splitting Types (Section 2)

The authors establish how the splitting type of $T_A$ changes depending on the intersection of a line $l$ with a specific finite scheme $Z$ (defined by the resolution of $T_A$).

• Theorem 2.1: For a POG arrangement with exponents $(a, b)$ and level $d$, the splitting type of $T_A$ on a line $l$ depends on $|l \cap Z|$:
  • If $l \cap Z = \emptyset$: Splitting type is $(-a, 1-b)$.
  • If $|l \cap Z| = 1$: Splitting type is $(1-a, -b)$.
  • If $l = l_A^0$ (the unique line containing $Z$): Splitting type is $(-d, d+1-a-b)$.
• Corollary 2.2: There exists a unique line $l_A^0$ where the splitting type is distinct from the generic case. This line is central to the geometric characterization.

B. Deletion Results (Section 2)

The paper provides simplified geometric proofs for how deleting a line affects the POG status:

• Proposition 2.6 & 2.7: If a line $l$ is deleted and the number of intersection points $h$ equals $a+1$ or $b+1$ respectively, the resulting arrangement remains POG (with modified exponents and level).
• Proposition 2.8: If $h = d+1$, the resulting arrangement becomes free.
• Counter-examples: The authors show that if $h \leq a$, the resulting arrangement is generally neither free nor POG, highlighting the rigidity of the POG structure.

C. Geometric Characterization of NT-Free Arrangements (Section 3)

This is the paper's primary contribution. The authors provide necessary and sufficient conditions for a POG arrangement to be NT-free, based on the position of the unique line $l_A^0$.

Theorem 3.5 (Main Result):
Let $A$ be a POG arrangement with exponents $(a, b)$ and level $d > b$. Let $l_A^0$ be the unique line with the special splitting type.

1. Case 1: $l_A^0 \notin A$ (The line is not in the arrangement)
   $A$ is NT-free (specifically NT-free minus) if and only if:

   • $A \cup \{l_A^0\}$ is free.
   • Equivalently, $d = |A| - |A \cap l_A^0|$.

2. Case 2: $l_A^0 \in A$ (The line is in the arrangement)
   $A$ is NT-free (specifically NT-free plus) if and only if:

   • $A \setminus \{l_A^0\}$ is free.
   • Equivalently, $d + 1 = |A \cap l_A^0|$ (the number of intersections on the special line matches the level plus one).

Lemma 3.4 proves that if $l_A^0 \notin A$, $A$ cannot be NT-free plus, and if $l_A^0 \in A$, $A$ cannot be NT-free minus. This dichotomy simplifies the classification.

4. Significance and Impact

• Geometric Insight: The paper shifts the perspective from combinatorial lattice properties to the geometry of vector bundles. It demonstrates that the "special line" $l_A^0$ (derived from the bundle resolution) dictates the freeness properties of the arrangement.
• Simplified Proofs: The authors provide new, succinct proofs for known results regarding nearly free and POG arrangements using vector bundle techniques, making these complex algebraic properties more accessible via geometric intuition.
• Resolution of Open Questions: Theorem 3.5 answers the question of when a POG arrangement is NT-free, providing a concrete geometric test (checking the intersection count on $l_A^0$) rather than relying solely on abstract combinatorial conditions.
• Connection to Ziegler Restrictions: The work reinforces the deep link between the splitting type of the logarithmic vector field bundle and the exponents of Ziegler restrictions, validating the use of bundle theory as a primary tool in arrangement theory.

5. Conclusion

Măcinic and Vallès successfully characterize the subset of plus-one generated arrangements that are "next-to-free." By identifying a unique geometric feature—the line $l_A^0$ associated with the vector bundle's resolution—they establish a direct correspondence between the combinatorial data of the arrangement (intersection counts) and the geometric structure of the associated bundle. This provides a powerful new criterion for determining the freeness of arrangements near the boundary of the free locus.