Fano-K3 Pairs in Algebraic Geometry

• Fano-K3 pairs are geometric constructions linking smooth Fano varieties and K3 surfaces via either subvariety embeddings or fully faithful Fourier–Mukai functors.
• Their construction employs moduli spaces, wall-crossing techniques, and birational transformations, with outcomes influenced by the arithmetic of the genus.
• These pairs illuminate connections in Hodge theory, derived categories, and moduli problems, impacting research in hyperkähler and Calabi–Yau geometry.

A Fano-K3 pair is a geometric construction relating smooth Fano varieties and K3 surfaces, typically in one of two configurations: a Fano variety containing or associated to a K3 surface in its geometry, or a categorical embedding of the derived category of a K3 surface into that of a Fano (or weak Fano) variety. These pairs play a central role in bridging birational geometry, Hodge theory, derived categories, moduli problems, and connections to hyperkähler and Calabi-Yau geometry.

1. Foundational Definitions and Genus Arithmetic

A smooth projective Fano variety $Y$ is called a Fano-K3 partner to a K3 surface $X$ when there exists a fully faithful embedding of triangulated categories $D^b(X)\hookrightarrow D^b(Y)$ given by a Fourier-Mukai kernel, or equivalently, when the geometry of $Y$ encodes a K3 surface $X$ as a subvariety, anticanonical divisor, discriminant double cover, or via lattice-theoretic/Hodge-theoretic correspondences. In the setting of $X$ a complex K3 surface with Picard number $\rho(X)=1$ and ample generator $\Lambda$ of degree $2g-2$, the arithmetic of the genus $g$ modulo $4$ dictates the outcomes:

• If $g\not\equiv 3 \pmod{4}$: $X$ is a Fano visitor, i.e., there exists a smooth Fano $Y$ together with $D^b(X)\hookrightarrow D^b(Y)$ via a fully faithful functor.
• If $g\equiv 3 \pmod{4}$: Only a smooth weak Fano $Y$ arises; $-K_Y$ is big and nef but not ample (Aravena, 2024).

This dichotomy is realized through wall-crossing and birational transformations in moduli spaces constructed from the Mukai vector $v=(0,h,1-g)$, Gieseker moduli spaces, and Bridgeland stability conditions.

2. Construction Methodology: Moduli, Flips, and Flops

The Fano-K3 construction proceeds from the moduli space $\mathcal{M}=M_\Lambda(v)$ of $\Lambda$-Gieseker-stable sheaves on $X$, with $v$ as above. $\mathcal{M}$ is a $2g$-dimensional irreducible holomorphic symplectic manifold with Lagrangian fibration $\pi:\mathcal{M}\to |\Lambda|\cong\mathbb{P}^g$. Two distinguished divisor classes $f$ and $\lambda$ in $\operatorname{NS}(\mathcal{M})$ are constructed via the Mukai isomorphism, and the movable cone admits a wall-and-chamber decomposition indexed by arithmetic slopes $\mu(c,d)$, determined by explicit divisibility and positivity conditions.

A sequence of Bridgeland wall-crossings in space of stability conditions $\sigma_\alpha$ yields flops (the “Mukai flops”) across walls $\alpha_j$, producing a finite chain of birational models $\mathcal{M}_i$. Restricting antisymplectic involutions induced by the dualizing functor $\Psi$ gives fixed loci $M_0\cong\mathbb{P}^g$ and $M_1\cong\operatorname{Bl}_X\mathbb{P}^g$, with subsequent flips governed by eigenbundle ranks computable via Riemann–Roch.

3. Semiorthogonal Embeddings and Fourier–Mukai Functors

Each small flip $M_i\to M_{i+1}$ induces a fully faithful embedding $D^b(M_i)\hookrightarrow D^b(M_{i+1})$ by the Bondal–Orlov theorem. Since $D^b(X)$ embeds into $D^b(M_1)$ (exceptional divisor), a cascade of embeddings $$D^b(X)\hookrightarrow D^b(M_1)\hookrightarrow\cdots\hookrightarrow D^b(M_j)$$ is constructed where $M_j$ is chosen for $-K_{M_j}$ to be ample (Fano) or big and nef (weak Fano).

A universal family $\mathcal{E}\in D^b(X\times\mathcal{M})$ yields functors

$$\Phi_\mathcal{E}(-) = R\,p_{\mathcal{M},*}( \mathcal{E} \otimes p_X^*(-) )$$

and composition with the Bondal–Orlov functors gives a single kernel $\Phi_K: D^b(X)\hookrightarrow D^b(Y)$ for the sought Fano $Y$.

4. Geometry: Anticanonical Class, Ample Cone, and Invariants

On each $M_i$, the Picard group is generated by $H$ (pullback of $\mathcal{O}_{\mathbb{P}^g}(1)$) and the exceptional divisor $E$. One writes $\mathcal{O}_i(m,n) := (m+n)H - nE$. Restriction formulas yield

$$f|_{M_i} = \mathcal{O}_i(0,-1),\quad \lambda|_{M_i} = \mathcal{O}_i(2,\lfloor g/2 \rfloor - 1),$$

and the anticanonical divisor

$-K_{M_i} = \mathcal{O}_i(4, g-3).$

$M_j$ is Fano precisely if $4$ and $g-3$ lie in the positive interior cone defined by wall rays. The arithmetic criterion—examined via inequalities in the $\mu$-chamber structure—determines Fano/weak Fano types according to $g\bmod 4$.

Geometric invariants of $Y$ (dimension $g$) include Picard number ($2$ for $1\le i\le\nu$), Fano index (divides $\gcd(4,g-3)$), intersection numbers $H^g$, $H^{g-1}\cdot E$, $E^g$, all computable from the blow-up and flip geometry.

5. Birational and Derived Symmetries: The Special Involution

For $g=4k+3$, a “derived involution” $\Phi:D^b(X)\to D^b(X)$ is constructed via the rank-2 Mukai partner $Y\cong M(2,-h,2k+1)$, interchanging $M_i\cong M_{\nu+1-i}$ and manifesting the failure of ampleness in $-K_{M_j}$. The final blow-down $M_\nu\to M_{\nu+1}\cong\mathbb{P}^g$ centralizes the K3 surface $X$ in the exceptional locus, and only a weak Fano structure is obtained.

6. Extensions and Classification: Higher Moduli Spaces and Connection to Other Fano-K3 Geometries

A variant of the construction (twisted by Belmans–Fu–Raedschelders) shows various moduli spaces $M_H(v_i)$—notably Hilbert schemes $\operatorname{Hilb}^d X$ for $d\leq(g+1)/4$—are also Fano or weak Fano visitors, admitting $D^b(M_H(v_i))\hookrightarrow D^b(Y)$. Other geometric models for Fano-K3 pairs arise in:

• Prime Fano varieties of genus 10: moduli spaces of vector bundles on genus 10 K3 surfaces arising as double covers of planes branched over sextics (Kapustka et al., 2010).
• Trisecant flop descriptions of special cubic fourfolds and their associated minimal K3 surfaces (Russo et al., 2019).
• K-stability and modular compactifications for quartic K3 surfaces viewed as anti-canonical divisors in $\mathbb{P}^3$ (Ascher et al., 2021).
• Conic bundle structures and even nodal discriminants relating Fano fourfolds and minimal nodal surfaces generalizing Kummer surfaces (Bernardara et al., 2024).

7. Categorical, Hodge-Theoretic, and Enumerative Consequences

Many Fano-K3 pairs originate from deep categorical and Hodge-theoretic properties:

• Kuznetsov’s component $\mathcal{K}u(X)$ of a Fano threefold $X$ constitutes a noncommutative K3 (Enriques category), with CY$_2$ covers identified as derived categories of K3 surfaces (Bayer et al., 2022).
• Fano-K3 pairs underlie rationality questions, period mappings, and moduli problems, often connected via derived equivalences and Hodge isometries, influencing the geometry of irreducible holomorphic symplectic manifolds and modularity in enumerative invariants (Doran et al., 2024).

Key geometric, categorical, and enumerative invariants are thus organized by genus arithmetic, wall-and-chamber decompositions in moduli spaces, derived embedding criteria, and birational transformations such as Mukai flops and trisecant flops. Fano-K3 pairs remain central in ongoing developments in algebraic geometry, moduli theory, and the categorification of geometric equivalence (Aravena, 2024).