On the fundamental groups of perforated surfaces

Abstract: A perforated surface is the complement $\mathringΣ:=Σ\setminus A$ of a countable dense subset $A$ in a connected paracompact surface $Σ$. It is known that the topological type of $Σ\setminus A$ is independent of the choice of $A$. Any perforated surface is one-dimensional, connected, locally path connected, and is not semi-locally simply connected at any of its points. In this paper we obtain a classification theorem for perforated surfaces, using the classification theorem for surfaces. We show that any connected covering of a perforated surface $\mathring Σ$ arises from a covering of a surface $Σ'$ such that $\mathringΣ\cong \mathringΣ'$. We show that the fundamental group of perforated surfaces are large. We also show that the fundamental groups of $\mathring Σ$, the Sierpiński curve and the Menger curve are not Hopfian.

• The paper establishes that perforated surfaces, defined as a surface minus a dense subset, have a canonical homeomorphism type with 2^(aleph_0) distinct isomorphism classes.
• It develops covering space theory by showing any connected cover is itself perforated and that deck transformation groups can model any countable quotient.
• The paper demonstrates that the fundamental groups are large, locally free, and non-Hopfian, exhibiting intricate embedding structures and myriad non-isomorphic normal subgroups.

Fundamental Groups of Perforated Surfaces

Introduction

The paper "On the fundamental groups of perforated surfaces" (2604.13544) investigates the topology and algebraic properties of perforated surfaces, specifically their fundamental groups and covering space theory. Perforated surfaces are defined as $\mathring{\Sigma} = \Sigma \setminus A$, where $\Sigma$ is a connected paracompact surface and $A$ is a countable dense subset. The study leverages the established classification theorems for surfaces and one-dimensional wild spaces, with perforated surfaces serving as prototypical examples of spaces with intricate local and global topological behavior.

Classification of Perforated Surfaces

Perforated surfaces are shown to be locally path-connected, one-dimensional, and wild (i.e., non-semi-locally simply connected everywhere). Notably, the homeomorphism type of $\mathring{\Sigma}$ does not depend on the dense subset $A$, thus endowing these spaces with canonical structure. The paper establishes a refined classification theorem for perforated surfaces, extending the classical Kerékjártó–Richards theorem via analysis of ends and their orientability types—planar, non-planar, and non-orientable ends.

The key result is that any perforated surface is homeomorphic to the complement of a countable dense subset in a unique surface (up to homeomorphism) subject to constraints on its end spectrum. In particular, the cardinality of the distinct homeomorphism types of perforated surfaces (with fixed genus and orientability class) matches the classification of closed subsets of the Cantor set, yielding $2^{\aleph_0}$ many isomorphism classes.

Covering Spaces and Deck Transformations

The non-semi-locally simply connected nature of $\mathring{\Sigma}$ precludes universal covers, yet the paper demonstrates that these spaces admit a rich theory of connected coverings. It is proven that any connected covering space of a perforated surface is either of finite or countably infinite degree and is itself a perforated surface. The deck transformation groups of regular coverings are shown to realize any countable group as a quotient of the fundamental group by an appropriate normal subgroup.

Moreover, there exist $2^{\aleph_0}$ many inequivalent regular covering projections, highlighting a complex normal subgroup structure within $\pi_1(\mathring{\Sigma})$. The paper provides explicit methods for constructing covers and analyzes the relationship between open covers and normal subgroup generation.

Fundamental Group Structure and Embeddings

A principal finding is that the fundamental groups of perforated surfaces are large and complex. $\pi_1(\mathring{\Sigma})$ is locally free, resides inside an inverse limit of free groups, and is residually finite. For any surface $\Sigma$0, it is shown that the countable free $\Sigma$1-product $\Sigma$2, where $\Sigma$3 is the Sierpiński curve, and free products indexed by continuum, as $\Sigma$4, embed in $\Sigma$5.

There are numerous embeddings and retractions between fundamental groups of classical one-dimensional spaces (e.g., Hawaiian earring, Menger curve), establishing that these groups can be viewed as retracts of $\Sigma$6. The space of normal subgroups is highly nontrivial—most are uncountable, and there is a surfeit of non-isomorphic groups realized as $\Sigma$7 for various surfaces, including continuously many.

Non-Hopficity

A strong algebraic result is proven: $\Sigma$8 is not Hopfian for any surface $\Sigma$9, indicating the existence of non-injective surjective endomorphisms. This extends to other wild spaces such as the Sierpiński curve, Menger curve, and Sierpiński gasket. The proofs are constructive, employing explicit self-maps and retractions to demonstrate the failure of the Hopf property, even though these groups are locally free (which precludes certain pathologies but does not guarantee Hopficity in the infinite rank case).

Implications and Future Directions

The results have significant implications for both geometric group theory and manifold topology. The identification of large, non-Hopfian, residually finite, locally free groups arising from wild one-dimensional spaces expands the known landscape of fundamental groups associated with pathological topological spaces. Moreover, the embedding structure and surjectivity properties suggest deep connections between the topology of ends and group-theoretic properties.

Practically, these findings impact the study of aspherical spaces, wild embedding problems, and the classification of manifolds via their end spectra. The inability to realize simply connected covers motivates further exploration of alternative notions of covering and higher categorical structures. The embedding results and non-Hopfian property hint at new invariants for wild spaces and may inform future endeavors in the classification of noncompact surfaces, their automorphism groups, and the interplay with descriptive set theory.

Future avenues include extensions to higher dimensions, the homological and cohomological analysis of wild spaces, and applications to exotic phenomena in geometric topology and infinite group theory.

Conclusion

This paper provides a comprehensive analysis of the fundamental groups of perforated surfaces, delivering classification, embedding, and covering results that elucidate their large-scale algebraic and topological complexity. With continuously many non-isomorphic groups realized and explicit identification of the non-Hopfian property, the work lays foundational groundwork for future research in wild topology, group embeddings, and classification theory.