The Kobayashi-Hitchin correspondence for nef and big classes

This paper establishes a complete proof of the Kobayashi-Hitchin correspondence for nef and big classes by introducing the concepts of adapted closed positive $(1,1)$-currents and $T$-adapted Hermitian-Yang-Mills metrics, thereby proving that a holomorphic vector bundle is slope polystable if and only if it admits such a metric, a result that extends to singular settings and yields new insights into projective flatness and the Bogomolov-Gieseker inequality.

The Gist

Imagine you are an architect trying to build a perfect, stable skyscraper (a holomorphic vector bundle) on a piece of land that isn't perfectly flat. Some parts of the land are lush and green (the ample locus), but other parts are swampy, rocky, or even have sinkholes (the singularities or the non-Kähler locus).

For decades, mathematicians have known how to build these skyscrapers if the land is perfectly flat and smooth. This is the famous Kobayashi-Hitchin correspondence: it says that a building is structurally stable (mathematically "polystable") if and only if you can find a specific type of "perfect tension" in its steel beams (a Hermitian-Yang-Mills metric) that balances all the forces perfectly.

The Problem:
What happens when the land is rough? What if the "ground" (the geometric class $\alpha$) is "nef and big"—meaning it's mostly good, but has some nasty, undefined singularities where the ground is broken? Previous methods required the ground to be smooth or the holes to be very specific (like cone-shaped). If the ground was messy and the holes were weird, the old blueprints didn't work.

The Solution (This Paper):
Satoshi Jinnouchi has written a new blueprint that works even on this messy, broken ground. Here is how the paper works, broken down into simple concepts:

1. The "Adapted" Tool (The Flexible Ruler)

In the past, mathematicians tried to measure the land with a rigid ruler. If the ground was bumpy, the ruler broke.
Jinnouchi introduces a new tool called an "Adapted Current."

• The Metaphor: Imagine a flexible, stretchy measuring tape that can mold itself to the shape of the swampy ground. It doesn't need to be perfectly smooth; it just needs to know where the "good" ground is and how to handle the "bad" spots without snapping.
• This tool allows the mathematician to define what "smoothness" means even when the underlying geometry is broken.

2. The "Adapted" Metric (The Custom Suit)

Once you have your flexible ruler, you need to dress your skyscraper.

• The Metaphor: Usually, you try to put a standard, stiff suit on a person. If the person is twisted or has a weird shape, the suit tears.
• Jinnouchi defines a "T-Adapted Hermitian-Yang-Mills Metric." This is a custom-tailored suit made of "smart fabric." It stretches and shrinks exactly where the ground is bad (near the singularities) so that the tension (the forces holding the building together) remains balanced everywhere else.
• The paper proves that if your building is "stable" (it won't collapse under its own weight), you can always find this custom suit. Conversely, if you can find this suit, the building is guaranteed to be stable.

3. The "Weak" Sub-bundles (The Invisible Support Beams)

To prove this, the author had to invent a way to look at the building's internal structure through the fog of the broken ground.

• The Metaphor: Imagine trying to see the steel beams inside a building through a thick, smoky window. You can't see them clearly, but you can see their shadows.
• Jinnouchi uses "Weak Holomorphic Sub-bundles." These are like the "shadows" of the building's internal supports. Even though the ground is broken, these shadows are strong enough to prove that the building is stable. He uses a famous mathematical trick (the Chern-Weil formula) to calculate the weight of these shadows, proving that the building is balanced.

4. The "Jordan-Hölder" Filter (The Sorting Hat)

What if the building isn't perfectly stable, but it's "okay" (semistable)?

• The Metaphor: Imagine a messy pile of Lego bricks. Some are red, some are blue. They aren't a single perfect tower, but they can be sorted into separate, perfect towers.
• The paper shows that any "okay" building can be broken down into its purest, stable components (a Jordan-Hölder filtration). Once you separate them, each piece gets its own perfect custom suit. This proves that even messy structures have an underlying order.

5. The Big Payoff: Why Does This Matter?

This isn't just about abstract math; it solves real problems in physics and geometry:

• Singular Spaces: It allows mathematicians to study shapes that have "cracks" or "kinks" (like singular Kähler-Einstein metrics), which appear in string theory and the study of the universe's shape.
• The Bogomolov-Gieseker Inequality: This is a rule about how much "curvature" a building can have. The paper proves that if a building hits the maximum limit of this rule, it must be "projectively flat"—meaning it's essentially a perfect, flat sheet wrapped around a sphere. This is a new discovery even for perfect, smooth buildings!

Summary

Think of this paper as a universal construction manual.

• Old Manual: "Only build on flat, smooth land."
• New Manual (Jinnouchi): "You can build on broken, rocky, swampy land too. Just use our 'Adapted' tools to measure the ground and 'Adapted' suits to balance the forces. If the building is stable, the suit exists. If the suit exists, the building is stable."

This breakthrough means mathematicians can now explore the most rugged, complex, and "broken" corners of geometry with confidence, knowing their tools will hold up.

Technical Summary

Here is a detailed technical summary of the paper "The Kobayashi-Hitchin correspondence for nef and big classes" by Satoshi Jinnouchi.

1. Problem Statement

The classical Kobayashi-Hitchin correspondence (also known as the Donaldson-Uhlenbeck-Yau theorem) establishes an equivalence between the algebraic notion of slope polystability for a holomorphic vector bundle $E$ and the analytic existence of a Hermitian-Yang-Mills (HYM) metric on $E$. This correspondence was originally proven for smooth Kähler metrics on compact Kähler manifolds.

However, significant gaps remain in the theory when dealing with:

1. Nef and Big Classes: The correspondence had not been fully established for cohomology classes $\alpha$ that are nef and big but not necessarily Kähler (i.e., not strictly positive everywhere).
2. Singular Settings: Extending the correspondence to singular varieties (e.g., compact normal Kähler varieties with log terminal singularities) and reflexive sheaves.
3. Singular Currents: Existing approaches often require the background metric to be smooth or have specific types of singularities (e.g., conical or orbifold). There was a lack of a framework for adapted currents that are not strictly positive and possess singularities that cannot be explicitly described.

This paper aims to provide a complete proof of the Kobayashi-Hitchin correspondence for nef and big classes on compact Kähler manifolds and extend it to singular settings, introducing new analytic tools to handle the singularities of the underlying currents.

2. Methodology and Key Definitions

The author introduces a novel framework based on adapted currents and adapted HYM metrics, inspired by the work of Chen-Chiu-Hallgren-Székelyhidi-Tó-Tong (CCHTST25) and Uhlenbeck-Yau.

A. Adapted Closed Positive (1,1)-Currents

Let $X$ be a compact Kähler manifold and $\alpha$ a nef and big class. A closed positive $(1,1)$-current $T \in \alpha$ is called adapted if:

1. Regularity on Ample Locus: $T$ is smooth Kähler on the ample locus $\text{Amp}(\alpha) = X \setminus \text{EnK}(\alpha)$, where $\text{EnK}(\alpha)$ is the non-Kähler locus.
2. Lower Bound: After a sequence of blow-ups $\pi: Y \to X$ resolving the singularities of the non-Kähler locus into a simple normal crossing (snc) divisor $D$, the pullback $\pi^*T$ satisfies a lower bound: $\pi^*T \ge |s_D|^{2m} h_D \omega_Y$ for some $m \in \mathbb{Z}_{\ge 0}$, where $s_D$ is the defining section of $D$.
3. Approximation: There exists a sequence of smooth Kähler metrics $\omega_i$ in a class approximating $\pi^*\alpha$ that converges to $\pi^*T$ locally smoothly on $Y \setminus D$, satisfying uniform $L^p$ bounds on the density ratio $(\omega_i^n / \omega_Y^n)$.

B. T-Adapted Hermitian-Yang-Mills Metrics

A smooth Hermitian metric $h$ on a vector bundle $E$ (defined on the ample locus) is a $T$-adapted HYM metric if:

1. HYM Equation: $\sqrt{-1}\Lambda_T F_h = \lambda \text{Id}$ on $\text{Amp}(\alpha)$.
2. Finite Energy: $\int_{\text{Amp}(\alpha)} |F_h|^2_{h,T} T^n < \infty$.
3. Adapted Conditions: The metric satisfies specific growth and integrability conditions near the singular locus $D$, controlled by the defining section $s_D$ and the current $T$. Specifically, the metric and its derivatives must be bounded relative to powers of $|s_D|$.

C. T-Weak Holomorphic Subbundles

The paper generalizes the concept of weak holomorphic subbundles (introduced by Uhlenbeck-Yau) to the setting where the base metric is a closed positive current $T$. This allows for the definition of destabilizing subsheaves even in the presence of singular currents, which is crucial for the stability analysis.

3. Key Contributions and Results

Main Theorem (The Kobayashi-Hitchin Correspondence)

Theorem 1.4: Let $X$ be a compact Kähler manifold, $\alpha$ a nef and big class, and $T$ an adapted current in $\alpha$. A holomorphic vector bundle $E$ is $\alpha^{n-1}$-slope polystable if and only if $E$ admits a $T$-adapted HYM metric. Furthermore, such a metric is unique up to constant scaling.

• Stability Definition: The paper uses $\alpha^{n-1}$-slope stability, defined via the intersection number $\int_X c_1(F) \wedge \alpha^{n-1}$, which is well-defined for nef and big classes.
• Uniqueness: The proof establishes that if two $T$-adapted HYM metrics exist, their Chern connections coincide, leading to uniqueness.

Applications to Singular Settings

By utilizing resolution of singularities, the main theorem implies the correspondence for reflexive sheaves on compact normal Kähler varieties with log terminal singularities (Corollary 1.6).

• Corollary 1.8: Specifically, if $X$ admits a singular Kähler-Einstein metric $\omega$ (which is an adapted current), then a reflexive sheaf $E$ is $\{\omega\}^{n-1}$-slope polystable if and only if it admits an $\omega$-adapted HYM metric. This is a new result for singular varieties.

Jordan-Hölder Filtration

Theorem 1.7: For a torsion-free sheaf $E$ that is $\alpha^{n-1}$-slope semistable (but not necessarily stable), there exists a Jordan-Hölder filtration. The associated graded sheaf admits a unique $T$-adapted HYM metric. This provides a canonical structure for semistable sheaves in the nef and big setting.

A Priori Estimates and Projective Flatness

• Corollary 1.9: The paper derives precise $L^\infty$ and $L^2$ estimates for the $T$-adapted HYM metric near the non-Kähler locus. It shows that the Lelong-type number of the metric's potential vanishes relative to the singularity of the current.
• Theorem 1.12 (Bogomolov-Gieseker Equality): If an $\alpha^{n-1}$-slope polystable bundle $E$ satisfies the equality case of the Bogomolov-Gieseker inequality ($(2rc_2(E) - rc_1(E)^2) \cdot \alpha^{n-2} = 0$), then $E$ is projectively flat on the ample locus of $\alpha$. This result is new even for projective manifolds when $\alpha$ is the first Chern class of a nef and big line bundle.

4. Proof Strategy

The proof follows the continuity method and compactness arguments, adapted to the singular setting:

1. Approximation: The author approximates the nef and big class $\alpha$ and the current $T$ by a sequence of smooth Kähler classes $\alpha_i = \alpha + \frac{1}{i}\omega_0$ and smooth metrics $\omega_i$.
2. Existence of Approximate Solutions: By the classical Uhlenbeck-Yau theorem, for each smooth $\omega_i$, there exists an $\omega_i$-HYM metric $h_i = h_0 \Psi_i^2$.
3. Uniform Estimates: The core difficulty is obtaining uniform estimates for $\Psi_i$ and $\Psi_i^{-1}$ as $i \to \infty$.
   • The author uses a mean-value type inequality (Guo-Phong-Sturm) to control the $C^0$ norm of $\Psi_i$ weighted by the defining section $s_D$.
   • A reductio ad absurdum argument is employed: If the $L^1$ norm of the logarithm of the metric diverges, one can construct a non-trivial limit endomorphism $u_\infty$ that defines a $T$-weak holomorphic subbundle.
   • Using a generalized Chern-Weil formula for these subbundles, the author shows that such a subbundle would destabilize the bundle $E$, contradicting the assumption of $\alpha^{n-1}$-slope stability.
4. Convergence: The uniform bounds allow the extraction of a subsequence converging to a limit $\Psi_\infty$, which defines the $T$-adapted HYM metric.
5. Uniqueness: The uniqueness proof involves showing that the difference between two such metrics satisfies a differential equation that forces them to be proportional, utilizing the stability of the bundle and the specific growth conditions of the adapted metrics.

5. Significance

• Generalization: This work significantly extends the Kobayashi-Hitchin correspondence beyond smooth Kähler metrics to the broader and more complex setting of nef and big classes.
• Singular Geometry: It provides a robust analytic framework for studying vector bundles on singular varieties (specifically klt varieties) without requiring explicit descriptions of the singularities (e.g., no need to assume conical or orbifold structures).
• New Tools: The introduction of adapted currents and $T$-weak holomorphic subbundles offers new tools for complex differential geometry, particularly for handling non-strictly positive currents.
• Applications: The results have immediate implications for the study of the equality cases of the Bogomolov-Gieseker inequality and the Miyaoka-Yau inequality in singular settings, linking algebraic stability to the existence of canonical metrics in a singular context.

In summary, Jinnouchi's paper resolves a major open problem in complex geometry by establishing a complete correspondence between algebraic stability and analytic metrics for nef and big classes, paving the way for further developments in the study of canonical metrics on singular spaces.