Dice Question Streamline Icon: https://streamlinehq.com

Embedding every 3-manifold in a 4-manifold with non-trivial gauge-theoretic invariants

Determine whether every closed oriented 3-manifold Y admits a smooth embedding as a separating hypersurface in some closed smooth 4-manifold with non-trivial Donaldson invariants; similarly, determine whether every closed oriented 3-manifold Y embeds (as a separating hypersurface) in some closed smooth 4-manifold with non-trivial Seiberg–Witten invariants, and whether such embeddings exist into 4-manifolds with non-trivial Bauer–Furuta invariants.

Information Square Streamline Icon: https://streamlinehq.com

Background

The authors’ method for constructing exotic families of embeddings requires starting with an embedding of the given 3-manifold Y into a 4-manifold V that has a non-trivial gauge-theoretic invariant (Donaldson, Seiberg–Witten, or Bauer–Furuta). They therefore pose the general existence problem of whether such embeddings are possible for all 3-manifolds.

They emphasize that the embedding should be separating to avoid trivial constructions like Y ↪ (S1 × Y) # X when X itself has non-trivial invariants. They also note known obstructions: certain 3-manifolds do not embed smoothly in any symplectic 4-manifold, which blocks an easy route via symplectic geometry to positive answers (Daemi–Lidman–Miller).

References

\begin{question} Does every $3$-manifold embed into some $4$-manifold with non-trivial Donaldson invariants? Non-trivial Seiberg-Witten invariants? Non-trivial Bauer-Furuta invariants? As the referee points out, we should require that the $3$-manifold separates to avoid a trivial answer such as $Y \hookrightarrow (S1 \times Y) X$ where $X$ has a non-trivial gauge-theoretic invariant. \end{question}

Exotic families of embeddings (2501.12673 - Auckly et al., 22 Jan 2025) in Section 6 (Proof of Theorem B)