Concordance and 1-loop clovers (Q5949336)

Two knots \(K_{0}\) and \(K_{1}\) in the three sphere \(S^{3}\) are called concordant if there is a smoothly and properly embedded annulus \(A\) in \(S^{3}\times[0,1]\) such that \((S^{3}\times[i],S^{3}\times[i]\cap{A})=(S^{3},K_{i})\) for \(i=0,1\). It is an interesting and challenging problem to realize concordance in terms of moves in \(S^{3}\), that is, to find a set of moves such that \(K_{1}\) can be obtained from \(K_{0}\) by using a finite sequence of the moves. Recently M.~Goussarov and K.~Habiro introduced independently the theory of clover (Habiro uses clasper, which is now recognized, but here I use `clover' once for all) to describe finite type invariants for knots and three-manifolds geometrically [\textit{S. Garoufalidis, M. Goussarov}, and \textit{M. Polyak}, Calculus of clovers and finite type invariants of 3-manifolds, Geom. Topol. 5, 75-108 (2001); \textit{K. Habiro}, ibid. 4, 1-83 (2000; Zbl 0941.57015)]. A clover is a uni-trivalent graph embedded in \(S^{3}\) (or more generally a three-manifold) containing a knot. One can associate a zero-framed link replacing each edge with a Hopf link so that each trivalent vertex forms a Borromean link. A link component corresponding to a univalent vertex is called a leaf. Doing surgery along this link, which can be regarded as a move in \(S^{3}\), one can get a new knot. Let \(\mathfrak{c}^{n}\) be the class of clovers whose first Betti numbers are at least \(n\), and each of whose leaves bounds a disk and intersects the knot exactly in one point. Two knots are said to be \(\mathfrak{c}^{n}\)-equivalent if one can be obtained from the other by a finite sequence of surgeries corresponding to clovers in \(\mathfrak{c}^{n}\). It is already known [\textit{A.~Kricker}, Covering spaces over claspered knots, math.GT/9901029; \textit{H.~Murakami} and \textit{T.~Ohtsuki}, Finite type invariants of knots via their Seifert matrices. Asian J. Math. 5, No. 2, 379-386 (2001)] that \(\mathfrak{c}^{2}\) implies \(S\)-equivalence, i.e. if two knots are \(\mathfrak{c}^{2}\)-equivalent, then their Seifert matrices are \(S\)-equivalent. Necessary and sufficient conditions in terms of other classes of clovers are also known [\textit{S.~Naik} and \textit{T.~Stanford}, Double \(\Delta\)-moves and \(S\)-equivalence, math.GT/9903069; \textit{J.~Conant} and \textit{P.~Teichner}, Grope cobordism of classical knots, math.GT/0012118]. The main result of the paper under review is that \(\mathfrak{c}^{1}\) implies concordance. Moreover, it is also proved that the converse is not true. In fact, there are double-slice knots which are not \(\mathfrak{c}^{1}\)-equivalent to the unknot. Here a knot is double-slice if it is obtained as the intersection of \(\mathbb{R}^{3}\) and the unknotted two-sphere in \(\mathbb{R}^{4}\) and so a double-slice knot is concordant to the unknot.