Non-Runge Fatou-Bieberbach Domains in Stein Manifolds with the Density Property

This paper presents methods for constructing two types of non-Runge Fatou-Bieberbach domains—proper open subsets biholomorphic to either $\mathbb{C}^n$ or the ambient Stein manifold $X$ itself—within Stein manifolds possessing the density property, and provides concrete examples where these constructions are applicable.

The Gist

Here is an explanation of the paper "Non-Runge Fatou-Bieberbach Domains in Stein Manifolds with the Density Property," translated into everyday language with creative analogies.

The Big Picture: The "Magic Shape-Shifting" Universe

Imagine you live in a universe called $\mathbb{C}^n$ (a complex, multi-dimensional space). In this universe, there are special rules about how shapes can stretch, shrink, and move without tearing.

Usually, if you have a shape (a "domain") that fits inside this universe, you can stretch it to fill the whole universe only if it was already a "perfect" shape to begin with. This is a rule called the Runge property. Think of it like a puzzle: if a piece fits perfectly into the hole, it's a "Runge" piece. If it's a weird, jagged piece that could fit but leaves gaps or overlaps in a way that breaks the rules, it's "non-Runge."

For a long time, mathematicians thought that in these complex universes, you could never find a "non-Runge" piece that still looked exactly like the whole universe when you squinted at it.

The Breakthrough:
This paper proves that such "weird" shapes do exist. The authors show how to construct a shape that:

1. Is a proper subset of the universe (it doesn't fill the whole thing).
2. Is mathematically identical to the whole universe (you can stretch it to look exactly like the original).
3. Crucially: It is "non-Runge," meaning it has a weird, jagged edge that prevents it from being smoothly approximated by the rest of the universe.

The Cast of Characters

To understand how they did it, we need to meet the tools they used:

1. The "Density Property" (The Super-Stretchy Fabric):
   Imagine the universe is made of a fabric that is incredibly stretchy and flexible. If you have a small patch of this fabric, you can stretch and twist it in almost any direction you want using "automorphisms" (magic transformations).

   • Analogy: Think of a piece of Silk. If you have a small piece of silk, you can pull it, fold it, and twist it to match almost any shape you want nearby. The "Density Property" just means the universe is made of this super-flexible silk.

2. Fatou-Bieberbach Domains (The "Hole" that is actually the "Whole"):
   These are the weird shapes we are looking for.

   • Analogy: Imagine a Swiss Cheese. Usually, a hole in cheese is just a hole. But imagine a hole that, if you zoomed in, looked exactly like the entire block of cheese. That's a Fatou-Bieberbach domain. It's a "hole" that is actually a perfect copy of the whole universe, just sitting inside the universe.

3. The "Runge" vs. "Non-Runge" Distinction:

   • Runge: A hole that is "nice." You can approximate its shape perfectly using the cheese around it.
   • Non-Runge: A hole that is "nasty." It has a shape so weird that the cheese around it can't approximate it well. It's like trying to fit a square peg into a round hole, but the peg is made of the same material as the hole.

The Two Methods: How They Built the "Nasty" Holes

The authors describe two different ways to create these weird holes.

Method 1: The "Vacuum Cleaner" (Basin of Attraction)

• The Concept: Imagine a whirlpool in a bathtub. If you drop a toy in, it spirals down to the drain. The area that flows into the drain is the "basin of attraction."
• The Trick: The authors found a way to create a whirlpool in their universe where the "drain" is a specific point, but the whirlpool doesn't cover the whole bathtub.
• The "Non-Runge" Twist: They placed a "barrier" (a hypersurface, like a thin sheet of glass) inside the universe. They made sure the whirlpool sucked everything in except that barrier.
  • Because the barrier is there, the whirlpool (the domain) has a weird edge.
  • Because the universe is "stretchy" (Density Property), they could pull the whirlpool tight so it looks exactly like the whole universe, even though it's missing that barrier.
  • Result: A perfect copy of the universe that is missing a piece of glass, making it "non-Runge."

Method 2: The "Push-Out" (The Escapist)

• The Concept: Imagine you have a room (the universe) and a wall (the barrier). You want to push the wall out of the room so the room becomes empty, but you want to keep the room looking like a room the whole time.
• The Trick: This is harder. You need to push the wall away without tearing the room.
• The "Non-Runge" Twist: They used a technique called "push-out." They slowly pushed the barrier (the wall) out of the way using the "stretchy fabric" of the universe.
  • They created a sequence of moves that pushed the wall further and further away.
  • In the limit, the wall is gone, and the remaining space is a perfect copy of the original universe.
  • However, because the wall was pushed out in a specific, jagged way, the resulting space has a "memory" of that jaggedness. It's a copy of the universe that is technically missing the wall, making it "non-Runge."

The Real-World Examples (The "Where")

The paper isn't just theory; they showed exactly where to find these shapes in real mathematical objects:

1. $SL_n(\mathbb{C})$ (The Matrix Universe):
   Think of a giant library of all possible $n \times n$ matrices (grids of numbers) where the determinant is 1.

   • They found a specific "forbidden zone" (matrices where the trace is zero).
   • They showed that if you remove this zone, the remaining space is stretchy enough to create these weird "non-Runge" holes.

2. The Koras-Russell Cubic (The Curved Surface):
   This is a specific, twisted 3D shape defined by a complex equation.

   • They proved that even on this weird, curved surface, you can find these "non-Runge" copies of the whole surface.

3. Flexible Varieties (The "Bendable" Shapes):
   They showed that if a shape is "flexible" (can bend easily), you can almost always find these weird holes in it.

Why Does This Matter?

In the past, mathematicians thought the "Runge" rule was a hard law of the universe. This paper breaks that law.

• The Analogy: It's like discovering that you can build a house that is a perfect scale model of a city, but the house has a door that leads to a room that doesn't exist in the city.
• The Impact: This changes how we understand the flexibility of complex shapes. It shows that even in highly structured, "perfect" mathematical worlds, there is room for chaos and "jagged" edges that defy simple approximation.

Summary in One Sentence

The authors used the extreme flexibility of certain complex universes to construct "perfect copies" of those universes that are missing specific parts, creating shapes that are mathematically identical to the whole but have "jagged" edges that break the usual rules of smooth approximation.

Technical Summary

Here is a detailed technical summary of the paper "Non-Runge Fatou-Bieberbach Domains in Stein Manifolds with the Density Property" by Gaofeng Huang, Frank Kutzschebauch, and Feng Rong.

1. Problem Statement

The paper addresses a fundamental question in several complex variables regarding the existence of non-Runge Fatou-Bieberbach (FB) domains within Stein manifolds possessing the density property.

• Fatou-Bieberbach Domains: These are proper open subsets $\Omega \subsetneq X$ of a complex manifold $X$ that are biholomorphic to either $\mathbb{C}^n$ (First Kind) or to $X$ itself (Second Kind).
• Runge Property: A domain $\Omega \subset X$ is Runge if every holomorphic function on $\Omega$ can be uniformly approximated on compact subsets by holomorphic functions on $X$.
• The Gap: While the existence of FB domains in $\mathbb{C}^n$ and Stein manifolds with the density property is well-established (via the Andersén-Lempert theory), classical constructions (basins of attraction and the push-out method) inherently produce Runge domains.
• The Core Question: Do there exist FB domains that are non-Runge in Stein manifolds with the density property? While Wold (2008) answered this affirmatively for $\mathbb{C}^n$, this paper seeks to generalize the construction to arbitrary Stein manifolds with the density property.

2. Methodology and Theoretical Framework

The authors develop a framework combining the Andersén-Lempert theory with new criteria for density properties relative to subvarieties.

A. The Weak Density Property

The authors introduce a novel concept, the Weak Density Property tangent to a subvariety $Y$.

• Definition: A Stein space $X$ has the weak density property tangent to a closed analytic subvariety $Y$ (containing singularities) if there exists a coherent sheaf $\mathcal{W}$ of $OX$-submodules of vector fields tangent to $Y$, such that:
  1. Global sections of $\mathcal{W}$ lie in the closure of the Lie algebra generated by complete vector fields tangent to $Y$.
  2. The stalk of $\mathcal{W}$ at any point $x \in X \setminus Y$ coincides with the tangent sheaf $T_x X$.
• Significance: This is a weaker condition than the "relative density property" (which requires vector fields to vanish on $Y$). It allows for vector fields tangent to $Y$ that do not vanish, providing greater flexibility in moving compact subsets within $X \setminus Y$ while preserving $Y$.

B. Approximation Theorems with Interpolation

The paper establishes a version of the Andersén-Lempert Theorem (Theorem 2.7) that incorporates interpolation.

• Result: If $X$ has the weak density property tangent to $Y$, any continuous isotopy of injective holomorphic maps on a Runge domain $\Omega \subset X$ (where the image avoids $Y$) can be approximated by holomorphic automorphisms of $X$ that preserve $Y$.
• Application: This allows the authors to move specific geometric configurations (like totally real disks) into FB domains without "touching" or destroying the separating hypersurface $Y$.

C. Construction Strategies

The paper employs two distinct construction methods corresponding to the two kinds of FB domains:

1. First Kind (Biholomorphic to $\mathbb{C}^n$):

   • Strategy: Generalization of Wold's method.
   • Mechanism: Construct a set $W$ (union of two disjoint totally real disks) that is holomorphically convex in $X \setminus H$ but whose polynomially convex hull intersects the hypersurface $H$.
   • Process: Use the density property of $X \setminus H$ to map $W$ into a standard FB domain $\Omega \subset X \setminus H$. The preimage of this domain under the automorphism is the desired non-Runge FB domain. The non-Runge nature arises because the hull of $W$ intersects $H$, preventing approximation by functions on the whole space $X$.

2. Second Kind (Biholomorphic to $X$):

   • Strategy: Generalization of the "push-out" method.
   • Mechanism: Requires a specific topological condition called Property (PO).
   • Property (PO): For any compact $K' \subset K \subset X$ with $H \cap K' = \emptyset$ and $H \cap K \neq \emptyset$, there exists an isotopy of automorphisms moving $H$ out of $K$ while keeping $K'$ fixed (approximately).
   • Process: Iteratively push the hypersurface $H$ away from an exhausting sequence of compact sets using automorphisms that preserve $H$. The limit of this process yields a domain $\Omega \subset X \setminus H$ biholomorphic to $X$. The non-Runge property is guaranteed because the complement of $\Omega$ contains a "copy" of $H$.

3. Key Contributions and Results

Main Theorems

• Theorem 3.1 (First Kind): Let $X$ be a Stein manifold and $H \subset X$ a complex hypersurface such that $X \setminus H$ has the density property. Then, there exists a Fatou-Bieberbach domain of the first kind in $X \setminus H$ that is Runge in $X \setminus H$ but non-Runge in $X$.
• Theorem 4.1 (Second Kind): Let $X$ be a Stein manifold and $H \subset X$ a hypersurface. If:
  1. $X \setminus H$ has the density property,
  2. $X$ has the weak density property tangent to $H$, and
  3. $(X, H)$ satisfies Property (PO),
     Then, there exists a Fatou-Bieberbach domain of the second kind in $X \setminus H$ that is Runge in $X \setminus H$ but non-Runge in $X$.

Specific Examples

The authors provide concrete examples where these theorems apply:

1. Special Linear Groups ($SL_n(\mathbb{C})$):
   • For $H = \{A \in SL_n(\mathbb{C}) : \text{tr}(A) = 0\}$, the complement $SL_n(\mathbb{C}) \setminus H$ has the density property.
   • Result: There exists a non-Runge FB domain of the first kind in $SL_n(\mathbb{C})$.
2. Koras-Russell Cubic Threefold ($KR$):
   • Defined by $x^2y + x + s^2 + t^3 = 0$ in $\mathbb{C}^4$.
   • For $H = \{x=0\} \cap KR$, the complement has the density property.
   • Result: There exists a non-Runge FB domain of the first kind in $KR$.
3. Flexible Affine Varieties ($X \times \mathbb{C}$):
   • If $X$ is a smooth flexible affine algebraic variety, then $X \times \mathbb{C}$ admits a non-Runge FB domain of the second kind contained in $X \times \mathbb{C}^*$.
   • This applies to products like the Calogero-Moser space.

4. Significance and Discussion

• Generalization: The paper successfully extends Wold's 2008 result from $\mathbb{C}^n$ to a broad class of Stein manifolds with the density property, covering both types of FB domains.
• Theoretical Innovation: The introduction of the Weak Density Property is a significant theoretical advancement. It bridges the gap between the strict "relative density property" (vanishing on $Y$) and the full density property, enabling the approximation of maps that preserve a subvariety without requiring them to vanish on it.
• Topological Obstructions: The authors highlight a deep tension in constructing Second Kind domains. While the density property of the complement is often easy to verify, Property (PO) is topologically restrictive.
  • Example: In $SL_2(\mathbb{C})$, moving the hypersurface $H=\{a=0\}$ away from the compact set $SU(2)$ is obstructed by homotopy groups ($\pi_3$). This suggests that finding Second Kind examples is significantly harder than First Kind examples due to competing topological and analytic constraints.
• Open Problems: The paper concludes with Problem 5.11, asking whether a non-Runge FB domain of the second kind exists in $SL_n(\mathbb{C}) \setminus \{ \text{tr}(A)=0 \}$. This remains an open challenge, illustrating the difficulty of satisfying Property (PO) in high-symmetry algebraic groups.

In summary, this paper provides a robust toolkit for constructing non-Runge Fatou-Bieberbach domains in complex geometry, establishing that the phenomenon of "escaping" the Runge property is not limited to $\mathbb{C}^n$ but is a feature of Stein manifolds with rich automorphism groups, provided specific geometric and topological conditions are met.