From Anton Fetisov's work I think we can extract a simple example. In fact this example should be universal for the class of examples he constructs.
Consider the free symmetric monoidal category onRecall that a connected second-countable one object-manifold with dualsboundary is a closed interval, an open interval, a half-open interval, or a circle. We can represent
Consider the objects of this category as a pair $(A^+, A^-)$whose objects are pairs of finitecountable sets, with $(A^+,A^-)$ and where a morphism $(A^+, A^-) \to (B^+, B^-)$ beingis a bijection betweensecond-countable 1-manifold with boundary, whose boundary equals $A^+ \circ B^-$$A^+ \cup A^- \cup B^+ \cup B^-$, and such that each closed interval component has one boundary point in $A^- \circ B^+$, plus a finite number of loops. Morphisms are composed by gluing the diagrams$A^+ \cup B^-$ and following the pathsone in $A^- \cup B^+$ (up to form a new bijection, while also possibly creating loopsisomorphism).
Suppose we want to add (countably) infinite products by generalizing this to infinite sets. Then we run into trouble, because when we compose the paths, we could get something that goes around infinitely in a spiral rather than ending up at any exit point. But if we just defineComposition of a morphism $(A^+, A^-) \to (B^+, B^-)$ to be a bijection between$(A^+ , A^-) \to (B^+, B^-)$ with a subset ofmorphism $A^+ \circ B^-$ and a subset of$(B^+,B^-) \to (C^+,C^-)$ is obtained by gluing the manifolds along $A^- \circ B^+$, plus a set$B$.
A symmetric monoidal structure is given by disjoint union of loops, then we can composesets and morphisms as before, throwing out any infinite paths. The identity is a union of closed intervals connecting each element to itself.
This satisfiesIf I did this with finite sets and compact manifolds, so that the axioms of aonly components were closed intervals and circles, I would have constructed the free symmetric monoidal abelian category on one generator and its dual. Because I allow infinite sets, with union as the tensor product structureI must necessarily allow connected components that are not closed.
The element $(\mathbb N,0)$$(\mathbb N,\emptyset)$ is idempotent andas $(\mathbb N,\emptyset) \otimes (\mathbb N,\emptyset)= (\mathbb N \cup \mathbb N,\emptyset \cup \emptyset) = (\mathbb N , \emptyset)$.
It has dual to $(0,\mathbb N)$$(\emptyset, \mathbb N)$, butwith the unit and counit each given by a union of closed intervals connecting the two copies of $\mathbb N$, where the relations hold by an infinite disjoint union of the usual stringy proof.
However, it is not self-dual dual, as $(\mathbb N,0)$$(\mathbb N,\emptyset)$ and $(0, \mathbb N)$$(\emptyset,\mathbb N)$ are not isomorphic, as every map between them has all elements connected to half-open intervals (because $B^- \cup A^+$ is empty), so the composition also has all elements connected to half-open intervals and thus isn't the identity.