Quantum Cohomology Central Charge

• Quantum cohomology central charge is a functional on derived categories of Fano manifolds, built using the Gamma class, a modified Chern character, and a canonical quantum differential solution.
• Its explicit formulation in projective spaces utilizes the J-function and asymptotic analysis to bridge quantum D-modules with semiorthogonal decompositions via Bridgeland stability.
• Recent research shows that this central charge underpins the connection between enumerative geometry and derived category invariants, providing insights into exceptional collections.

The quantum cohomology central charge is a functional associated to objects in the derived category of coherent sheaves on a Fano manifold, constructed from a canonical fundamental solution of the quantum differential equation and incorporating the Gamma class and a modified Chern character. It plays a central role in the interaction between quantum cohomology, the theory of Bridgeland stability conditions, and semiorthogonal decompositions (SODs), especially for projective spaces. Recent advances have shown how paths in the stability space driven by this central charge can be canonically linked to Beĭlinson’s exceptional collection SOD via asymptotic analysis, thereby elucidating deep connections between enumerative geometry and derived categories (Zuliani, 2024).

1. Construction of the Quantum Cohomology Central Charge

Let $F$ be a Fano manifold of complex dimension $n$, with a fixed point $\tau$ in the small quantum parameter space (for example $\tau \in H^2(F,\mathbb{C})$). The small quantum product $\star_{\tau}$ by a divisor class $H$ is used to deform the usual cup product structure on cohomology. The deformed (Dubrovin) connection

$$\nabla_{\tau, \alpha} = \partial_\alpha + \frac{1}{z} (\alpha \star_{\tau}), \quad \nabla_{\tau, z\partial_z} = z\partial_z - \frac{1}{z} (c_1(F) \star_{\tau}) + \mu$$

acts on $$H^*(F,\mathbb{C}) \times \mathbb{P}^1 \to \mathbb{P}^1$$, where $\mu$ is the grading operator.

There is a unique fundamental solution $$S(\tau, z) \in \operatorname{End}(H^*(F))[[z^{-1}]]$$ defined for $z \neq 0$ such that $S(\tau, z) \to \operatorname{Id}$ as $z \to \infty$ and satisfying

$$(\nabla_\tau)(S(\tau, z) \cdot z^{-\mu} \cdot z^\rho \, \alpha) = 0$$

for all $\alpha \in H^*(F)$ with $\rho = (c_1(F) \cup)$. Iritani defines the flat section isomorphism

$$\Phi^\tau(\alpha)(z) = (2\pi)^{-n/2} \, S(\tau, z) \, z^{-\mu} \, z^{\rho} (\alpha).$$

For any object $E \in D^b(F)$, Iritani’s quantum cohomology central charge is defined as:

$$Z^\tau(E;z) := (2\pi z)^{n/2} \int_F \Phi^\tau(\hat{\Gamma}_F \cdot \operatorname{Ch}(E))(z)$$

where $\hat{\Gamma}_F$ is the Gamma class $\hat{\Gamma}_F = \prod_i \Gamma(1 + \delta_i)$ for the Chern roots $\{\delta_i\}$ of $T_F$, and $$\operatorname{Ch}(E) = (2\pi i)^{\deg/2} \operatorname{ch}(E)$$ is the modified Chern character. The integral is taken against the Poincaré volume form. For $\tau = 0$, this specializes to Iritani’s 2009 central charge formula.

2. Explicit Formulation in Projective Space

When $F = \mathbb{P}^{N-1}$, let $H = c_1(\mathcal{O}(1)), n=N-1$. The quantum product at $\tau = t H$ is given by:

• $H \star_t H^i = H^{i+1}$ for $0 \leq i \leq N-2$
• $H \star_t H^{N-1} = e^t \cdot 1$

The quantum differential equation for $s(t, z) \in H^*(\mathbb{P}^{N-1})$ is:

$$z \frac{\partial}{\partial t} S(t, z) = (H \star_t) S(t, z), \quad S(0, z) = \operatorname{Id}$$

A closed-form fundamental solution is obtained via the $J$-function:

$$J(t, z) = e^{t H / z} \sum_{d=0}^{\infty} e^{dt} \frac{\prod_{k=0}^{dN-1} (H + k z)}{\prod_{k=1}^{d} (H + k z)^N}$$

The fundamental solution is then the matrix with columns:

$$S(t, z) = \left[ J(t, z),\, z \partial_t J(t, z),\, \dots,\, z^{N-1} \partial_t^{N-1} J(t, z) \right]$$

which approaches the identity as $z \to \infty$. In the monomial basis $\{1, H, \dots, H^{N-1}\}$, its expansion reads $$S(t, z) = \operatorname{Id} + \frac{t}{z}(H \star_0) + O(t^2)$$.

3. Integral and Pairing Formulation

For any $E \in D^b(\mathbb{P}^{N-1})$, the pairing presentation of the central charge is:

$$Z^t(E;z) = (2\pi z)^{-n/2} \langle S(t, -z) z^{-\mu} z^\rho(\Gamma_P \operatorname{Ch}(E)), 1 \rangle_P$$

or equivalently,

$$Z^t(E;z) = (2\pi z)^{n/2} \int_{\mathbb{P}^{N-1}} S(t, z) z^{-\mu} z^\rho(\Gamma_P \operatorname{Ch}(E))$$

where $\Gamma_P$ denotes the Gamma class for $\mathbb{P}^{N-1}$. At $t=0$, this reduces to the classical formula:

$$Z^0(E;z) \simeq (2\pi i)^{-n/2}\int_{\mathbb{P}^{N-1}} e^{H/z} \Gamma_P \operatorname{Ch}(E)$$

with the symbol $\simeq$ indicating dependence on the branch of $\log z$.

4. Bridgeland Stability, Quasi-Convergent Paths, and SODs

A Bridgeland stability condition $\sigma=(Z,\mathcal{P})$ on $D^b(\mathbb{P}^{N-1})$ is specified by a group homomorphism $Z: K(D) \to \mathbb{C}$ and a slicing $\mathcal{P}$. The notion of quasi-convergent paths $\sigma_r = (Z_r, \mathcal{P}_r)$ in $\operatorname{Stab}(D)$, introduced by Halpern-Leistner–Jiang–Robotis, is characterized by:

• Each object in $D^b$ admits a limit-semistable filtration with factors whose phases are separated by a positive gap as $r \to 0$;
• Any two such limit-semistable objects have well-defined relative log-masses asymptotically.

Any quasi-convergent path yields a canonical SOD:

$$D^b(\mathbb{P}^{N-1}) = \langle \mathcal{A}_1, \dots, \mathcal{A}_m \rangle$$

where $\mathcal{A}_i$ are generated by the limit-semistable factors grouped by phase.

In this framework, one takes $Z_r(E) := Z^t(E;z=r)$. As $r \to 0$, under Gamma Conjecture II, for exceptional objects $E_j$ from the Gamma basis,

$$\ln Z^t(E_j; r) = \frac{n}{2} \ln(2\pi r) - u_j(t)/r + o(1/r)$$

with $\{u_j(t)\}$ being the eigenvalues of $c_1(\mathbb{P}^{N-1}) \star_t$. Generic $t$ ensures these eigenvalues lie on distinct rays in $\mathbb{C}$.

Theorem 4.2 (Zuliani, 2024) establishes that, assuming distinct imaginary parts of $-u_j(t)$, one can select phases $$\phi_j(r) \simeq ( \operatorname{Im}(-u_j(t))/\pi r ) +$$ constant such that the phase differences traverse the value $1$ exactly once as $r$ varies. By Macrì's exceptional-collection construction, there exists an algebraic path $\sigma_r$ for small $r$ where each $E_j$ is $\sigma_r$-stable, and no other objects remain stable as $r \to 0$; this induces the full Beĭlinson SOD:

$$D^b(\mathbb{P}^{N-1}) = \langle E_0, E_1, \dots, E_{N-1} \rangle$$

The quantum cohomology central charges $Z^t$ are the explicit homomorphisms steering the path from the geometric chamber (skyscraper sheaves stable) to the algebraic chamber (exceptional collection stable), with the Beĭlinson SOD as the limit boundary point in the compactified stability space.

5. Foundational Results and Their Implications

The intricate structure of quantum cohomology central charges is underpinned by a suite of key results:

• Proposition 2.3.1 (Galkin–Golyshev–Iritani): Guarantees existence and uniqueness of the canonical fundamental solution $S(\tau,z)$ of the Dubrovin connection.
• Asymptotic expansion (Equation (6), Proposition 4.1): For objects in a Gamma basis,

$$\ln Z^\tau(E_j; r) \sim \frac{n}{2} \ln(2\pi r) - u_j(\tau)/r$$

as $r \to 0$.

• Macrì’s Theorem 3.1: For any full strong exceptional collection $\{E_j\}$ with mass/phase data $({m_j}, {\phi_j})$ and $\phi_{j+1} - \phi_j > 1$, there is a unique Bridgeland stability condition with $E_j$ stable in prescribed phases.
• Proposition 2.20 (Halpern-Leistner–Jiang–Robotis): Any quasi-convergent path induces a SOD ordered by asymptotic phases.

Together, these constructions situate the quantum cohomology central charge $Z^\tau(-;r)$ as the organizing principle for stability conditions on $D^b(\mathbb{P}^{N-1})$, directly encoding the information of the small quantum product and realizing explicit semiorthogonal decompositions as geometric limits in the stability manifold.

6. Context and Connections

The quantum cohomology central charge represents the convergence of techniques in enumerative geometry, representation of quantum invariants, derived categories, and stability theory. Its central role in connecting flat quantum D-module solutions to algebraic decompositions formalizes the relationship between enumerative invariants and homological algebraic structures. In particular, the explicit use of the fundamental solution of the quantum differential equation and the Gamma class points to the deep geometric content encoded in these central charges.

A plausible implication is that these methods may generalize to broader classes of Fano varieties beyond projective spaces, provided suitable foundational solutions exist and Gamma conjectures hold. The construction aligns with ongoing research on the interaction between Frobenius manifolds, stability conditions, and derived category invariants.