Wildness of codimension 1 submanifolds of euclidean space

For $n\geq 5$, such a diffeomorphism exists iff the complement of $S$ has an unbounded connected component. Indeed, the existence of such a diffeomorphism implies there is an unbounded component of the complement of $S$. Conversely, suppose that the complement of $S$ has an unbounded component. Choose a proper arc $\gamma:[0,\infty)\to\mathbb R^n$ in the complement of $S$ (i.e. going to infinity). Since $n\geq 5$, the sweepout of a generic proper homotopy from $\gamma$ to a standard coordinate ray is properly embedded, hence is an isotopy (this might still work for $n=4$, but it's definitely a bit more delicate, and I haven't thought about it carefully). Using this isotopy on $S$, we reduce to the case where $S$ is disjoint from a standard coordinate ray, which is easy.

A sufficient (but not necessary) condition for the complement of $S$ to have an unbounded component is that $S$ have finitely many components (this implies the complement of $S$ also has finitely many components, hence one must be unbounded).

This may be overkill but you could argue as follows for the special case of $M = \mathbb{R}^n$. A theorem of Kirby (On the set of non-locally flat points of a submanifold of codimension one. Annals, 88, (1968), 281-290) says that an embedding of an $n$-manifold, $n\geq 3$ can't fail to be locally flat at just one point. So take the one-point compactification $M^+$ of $M$, which is a sphere embedded in $S^{n+1} =$ one-point compactification of $\mathbb{R}^{n+1}$. By Kirby's theorem, it is locally flat.

Using local flatness at $p$, choose two balls on either side of $p$ joined by an arc (say a fiber of the normal bundle) that meets $M^+$ transversally in one point. The region between the two balls is a product $S^n \times (0,1)$, and then the region you want is obtained by removing the arc. (To get the picture you're asking about, you'd probably want to remove the inner balls of nested pairs of balls...)

A caveat is that this may be a circular argument from your point of view; Kirby uses a certain amount of machinery that was in the air in the mid-sixties. But I suspect that this machinery doesn't include the result you're trying to prove, since the assumptions on dimension don't match the ones for engulfing or h-cobordism techniques.

As you probably know, Siebenmann's thesis can be extended to dimension 5 (subject to a fundamenal group condition) by using the 4-dimensional work of Freedman. Unfortunately, one loses smoothness in the final conclusion, so your hope for a diffeomorphism would fall short.

An alternative to using Siebenmann's thesis might be to apply his Open Collar Theorem from the later paper "On detecting open collars". The beauty of that approach is that it is valid in the smooth category down to and including dimension 5.