Family 060 — Finding a spherical shell inside every class VII surface

← All digests

Family 060 · Algebraic and complex geometry
Importance 5/5 · Provisional editorial rating
Sources consulted 2026-10-12

1. The question

A complex surface has two complex dimensions and four real dimensions. Some compact complex surfaces can be studied using projective equations and positive Kähler forms; others lie outside that framework. Among the latter, class VII surfaces have long formed a difficult part of the classification problem.

The Global Spherical Shell conjecture proposes a concrete geometric feature inside every minimal class VII surface with positive second Betti number: a holomorphic copy of a neighborhood of the standard three-sphere in $\mathbb C^2$, placed so that its complement remains connected.

Family 060 claims this conjecture. The proposed route first produces enough compact rational curves and then applies an existing criterion that converts those curves into a shell. The hard distinction is between a divisor on an infinite cover and a divisor whose individual curve components are all compact. The manuscript’s analytic construction is designed to bridge that gap.

2. Background and essential definitions

The invariants defining class VII

The Betti number $b_j(X)$ is the dimension of the $j$-th real cohomology group. Roughly, $b_1$ measures independent loop information visible to abelian invariants, while $b_2$ counts independent real two-dimensional cohomology classes.

The canonical bundle $K_X$ is the line bundle of holomorphic two-forms. The Kodaira dimension $\kappa(X)$ measures the growth of sections of its tensor powers; $\kappa(X)=-\infty$ means every positive power has no nonzero holomorphic section.

In the convention used here, class VII means

$$b_1(X)=1,\qquad \kappa(X)=-\infty.$$

These surfaces are non-Kähler: a compact Kähler manifold has even first Betti number. This is why tools relying on a global Kähler form cannot simply be applied to the surface itself.

A surface is minimal when it contains no smooth rational curve of self-intersection $-1$. Such a curve can be contracted by reversing a point blowup. Minimality removes this elementary source of extra complexity; it does not mean that the surface has no rational curves at all.

Shells and rational curves

The standard sphere is

$$S^3=\{(z_1,z_2)\in\mathbb C^2:|z_1|^2+|z_2|^2=1\}.$$

A global spherical shell consists of a holomorphic embedding of an open neighborhood of this sphere from $\mathbb C^2\setminus\{0\}$ into $X$, with connected complement. The embedding includes a whole collar around the sphere, so it is much stronger than a smooth embedded sphere. Connectedness of the complement is the global, nonseparating condition.

A rational curve here is a compact irreducible complex curve whose normalization is $\mathbb P^1$. It may have singularities. Normalization separates branches and resolves the intrinsic singularities of the curve without requiring that its embedding in the surface be smooth.

An effective divisor is a locally finite sum of curves with nonnegative integer multiplicities. On a compact surface it has finitely many components. On an infinite cover it can have infinitely many, and a component can itself be noncompact unless something rules that out.

3. Development history and previous partial results

Kodaira’s classification placed class VII among the non-Kähler surfaces. For minimal class VII surfaces with $b_2=0$, work of Bogomolov, Li–Yau–Zheng, and Teleman led to the Hopf and Inoue classification. The positive-second-Betti-number problem is a separate part of that history.

Kato introduced the shell construction in 1977. One modifies a ball by blowups and glues suitable boundary neighborhoods holomorphically. Dloussky developed the description by blowup sequences and contracting germs. Nakamura’s classification program formulated the expectation that every minimal positive-$b_2$ class VII surface has such a shell. Kato, Dloussky.

Two existing curve criteria are central to M1. Enoki treats the case with a nonzero effective divisor of square zero. In the remaining case, Dloussky–Oeljeklaus–Toma prove that the presence of $b_2(X)$ rational curves gives a shell. This reduces the classification question to producing sufficient curves, even when none are known initially. Enoki, Dloussky–Oeljeklaus–Toma.

Gauge theory supplied major partial progress. Teleman proved the shell result for $b_2=1$ and the existence of a cycle of curves for $b_2=2$. A cycle result is not identical to the general criterion requiring a specified number of rational curves. More recently, Kurnosov–Spicer proved the shell conclusion for $b_2=3$ under the additional assumption of two distinct singular holomorphic foliations. M1 does not assume those foliations or a bound on $b_2$. Teleman’s first case, cycles, Kurnosov–Spicer.

Dloussky–Teleman’s work on infinite bubbling showed how exceptional divisors can limit to infinite sums of compact curves on covers. More general extension arguments still allowed the possibility of open Riemann surfaces as components. M1 identifies excluding those open components as the key additional analytic task. Infinite bubbling.

4. What the new manuscripts claim

Every positive-$b_2$ minimal class VII surface has a shell

M1 claims that every connected minimal compact complex surface satisfying

$$b_1(X)=1,\qquad b_2(X)>0,\qquad\kappa(X)=-\infty$$

contains a global spherical shell in the precise holomorphic and connected-complement sense above. There is no assumption that the surface is projective, has a known curve configuration, or carries a particular foliation.

The theorem is stated for all positive second Betti numbers. It concerns minimal surfaces; general blowups are not included by silently dropping minimality from its hypothesis.

Deformation and oriented smooth topology

Combining the claimed shell with Kato’s earlier deformation theorem, M1 obtains a small one-parameter deformation whose central fiber is $X$ and every nonzero fiber is a primary Hopf surface blown up exactly $b=b_2(X)$ times. Centers may be infinitely near, meaning later blowups can occur on exceptional curves of earlier ones.

A primary Hopf surface is a quotient of $\mathbb C^2\setminus\{0\}$ by an infinite cyclic group generated by a holomorphic contraction. The source deduces

$$\pi_1(X)\simeq\mathbb Z,$$

and an orientation-preserving diffeomorphism

$$X\cong (S^1\times S^3)\mathbin{ \# } b\,\overline{\mathbb{CP}}^{,2}.$$

The symbol $\#$ denotes connected sum, and the bar means the opposite of the usual complex orientation on the projective plane. Each point blowup contributes one such summand. This is a smooth topological description, not a holomorphic identification of the original minimal surface with a blown-up Hopf surface.

A consequence for aspherical complex surfaces

A manifold is aspherical when its universal cover is contractible. M1 combines its theorem with Albanese–Di Cerbo–Lombardi’s earlier work to claim that every connected closed aspherical complex surface satisfies

$$\chi_{\mathrm{top}}(S)\ge\frac95|\sigma(S)|.$$

Here $\chi_{\mathrm{top}}$ is the alternating sum of Betti numbers, and the signature $\sigma$ is the number of positive minus negative directions of the degree-two intersection form. If the Euler characteristic is positive, the surface is of general type; otherwise both invariants are zero.

The earlier inequality had a possible exceptional class VII case. The claimed shell theorem removes that case, rather than replacing the rest of the earlier argument. Albanese–Di Cerbo–Lombardi.

5. Why the advance matters

Importance: 5/5. If correct, the result would resolve a central remaining classification problem for compact complex surfaces. It would put all minimal positive-$b_2$ class VII surfaces within the geometric shell framework, with strong consequences for their topology and deformation behavior.

The contribution would go beyond identifying a numerical invariant. A holomorphic shell is an actual geometric region inside the surface, and its existence ties the surface to explicit blowup-and-gluing constructions. The claimed production of compact rational curves also addresses the long-standing difficulty of turning negative-definite topological data into effective complex geometry.

The aspherical consequence illustrates the reach beyond classification: resolving the exceptional complex-surface case sharpens an inequality comparing basic invariants of real four-manifolds carrying complex structures.

6. The main idea

Produce sections on a periodic cover

Enoki’s theorem first removes the square-zero-divisor case. In the remaining case, the manuscript works on an infinite cyclic cover $Y\to X$, obtained from a surjection of $\pi_1(X)$ onto $\mathbb Z$. This does not assume in advance that the whole fundamental group is $\mathbb Z$.

Let $T$ generate the deck group. A proper height function satisfies $h(Ty)=h(y)+1$, making the cover resemble an endless chain of repeated regions. Finite cyclic quotients and the negative-definite intersection form provide many line-bundle classes with compact support.

The source equips those bundles with holomorphic structures and solves a weighted Dolbeault problem. The Dolbeault operator measures failure to be holomorphic; Fourier–Laplace analysis in the repeated direction converts the problem into character-twisted problems on the compact base. An index calculation and vanishing estimates produce sections normalized to approach one at the positive end and zero at the negative end.

Their zero divisors supply the desired geometric representatives. At this point, however, their components might still run infinitely far down the cover.

Force every component to be compact

The manuscript controls divisor area in successive height slabs by an exponential bound. It then associates currents to holomorphic differentials on the normalization of a component. A current is a generalized differential form that can integrate test forms along a curve.

Weighted cohomology and a residue argument exclude certain nonzero currents. First this forces genus zero. If an open normalization omits at least two points of the sphere, a differential analogous to $dz/z$ gives the forbidden current. The only remaining open possibility is a complex plane.

The plane case uses translates of a component and bundle-valued differentials. A cylinder estimate puts all their currents into one fixed finite-dimensional cohomology space. Distinct translated curves would give linearly independent classes there, producing a contradiction. Uniformity of the estimate is vital: different bounds for each translate would not yield a common finite-dimensional obstruction.

Thus every component is compact and rational, although an entire divisor upstairs may still have infinitely many components.

Count enough curves downstairs

For a cyclic quotient of degree $n$, the construction provides at least $nb_2(X)-C_0$ independent compact classes, with $C_0$ fixed. If there were only $r<b_2(X)$ rational curves downstairs, their lifts would supply at most $nr$ translation orbits. For large $n$, that is too few to account for the independent classes.

A complementary curve-class bound gives exactly $b_2(X)$ rational curves in the remaining case. The earlier Dloussky–Oeljeklaus–Toma criterion then supplies the shell. The last step is established classification theory; the manuscript’s proposed new work is the section construction, compactness, and count.

7. Scope, limitations, and open questions

The theorem is specific to minimal compact complex surfaces of class VII with positive $b_2$. The $b_2=0$ classification is earlier work, and this is not an all-dimensional classification theorem for non-Kähler manifolds.

The deformation statement permits nonminimal nearby fibers even though the central surface is minimal. Minimality is not preserved automatically in every direction of deformation. Likewise, an oriented diffeomorphism fixes smooth topology without identifying holomorphic structures.

The proof’s infinite cyclic cover is an auxiliary cover, not initially asserted to be universal. The fundamental-group conclusion comes after the shell through Kato’s theorem. Keeping that order avoids presuming a main consequence inside the construction.

Detailed moduli, explicit contracting germs for individual surfaces, and finer analytic invariants remain meaningful questions even under the shell classification. The theorem supplies the common framework rather than a single holomorphic normal form for every surface.

8. Sources and further reading

M1’s Theorem 1.1 states the shell claim, and Corollaries 1.2–1.3 give the deformation and aspherical consequences. Sections 3–4 construct compact classes and sections, Sections 5–6 address the compactness of components, and Section 7 counts curves and invokes the existing shell criterion.

The primary papers linked above trace the development from Kato’s construction through Enoki, Dloussky–Oeljeklaus–Toma, Teleman’s gauge-theoretic advances, and the recent conditional $b_2=3$ result. Dloussky’s classification discussion gives further context for the curve and foliation approaches.

Source manuscripts

• M1. OpenAI, Global Spherical Shells on Minimal Surfaces of Class VII (September 24, 2026).

Sources consulted on 12 October 2026 at repository revision fd4aeeb. Download the manuscript citations.