# TETSUYA ABE - §â€ˆ³¢‡¤§‡¦...

date post

28-Feb-2021Category

## Documents

view

0download

0

Embed Size (px)

### Transcript of TETSUYA ABE - §â€ˆ³¢‡¤§‡¦...

CONSTRUCTIONS OF SMOOTHLY SLICE KNOTS

TETSUYA ABE

1. Abstract

A slice-ribbon conjecture is a long standing conjecture. In this note, we explain constructions of smoothly slice knots which might be non-ribbon and discuss related topics.

The idea of constructions is the following: Let HD be a non-trivial handle decomposition of the standard 4-ball B4 and h2 a 2-handle of HD. A slice knot K is obtained as the belt-sphere of h2. The cocore disk of h2 is a slice disk for K. There is no apparent reason for K to be ribbon. Typical examples are explained in Section 11.

If HD has, at least, two 2-handles, then we can construct more complicated slice knots. Let h21 and h

2 2 be 2-handles of HD. Let Ki (i = 1, 2) be the

belt-sphere of h2i . Then Ki is a slice knot. Furthermore any band sum of K1 and K2, denoted by K1♯bK2, is a slice knot

1. There is no apparent reason for K1♯bK2 to be ribbon. Gompf, Scharlemann and Thompson [GST] gave such a slice knot, which will be explained in Section 10.

1This is because the link consists of K1 and K2 bounds disjoint smooth disks in B 4.

1

2 TETSUYA ABE

The problem is how to obtain a “good” handle decomposition of B4. If we consider a complicated handle decomposition of B4, then we will obtain complicated slice knots. However it may be too difficult to check whether these slice knots are ribbon or not. One of the purposes of this note is to give simple and enough complicated handle decompositions of B4 explicitly.

2. Notations and organization

Throughout this note, we only consider the smooth category unless other- wise stated. The symbol A ≈ B means that A and B are diffeomorphic. We prefer to use the term“handle calculus, handle diagram” rather than “Kirby calculus, Kirby diagram”. We sometimes identify a given handle diagram with the corresponding handle decomposition or the 4-manifold itself represented by the handle decomposition.

Section 3–Section 12 are based on the talks given by the author in Mini-work shop on knot concordance, Sep. 17-20, 2013 at Tokyo Institute of Technology. In Section 13, we explain how to obtain ribbon presentations via handle cal- culus. The rest of this note, we give a brief overview of related topics.

3. 2-handles

We recall some definitions for the reader who is not familiar with handle theory.

A (4-dimensional) 2-handle h2 is a copy ofD2×D2, attached to the boundary of a 4-manifold X along ∂D2 ×D2 by an embedding ϕ : ∂D2 ×D2 → ∂X.

We call ∂D2 × D2 the attaching region of h2, ∂D2 × {0} the attaching sphere of h2, D2 × {0} the core of h2, {0} ×D2 the cocore of h2, {0} × ∂D2 the belt-sphere of h2. The belt-sphere of h2 is a knot in ∂(X ∪ϕ h2) which bounds a smooth disk(=cocore) in X ∪ϕ h2. A schematic picture may help us understanding, see the left picture in Figure 1.

Figure 1. A 2-handle and related terminologies.

CONSTRUCTIONS OF SMOOTHLY SLICE KNOTS 3

Note that the belt sphere of h2 is isotopic to the meridian of the attaching sphere of h2, see the right side picture in Figure 1. This fact is used to draw a slice knot in a handle diagram in Section 10.

4. The slice-ribbon conjecture

Let S3 be the 3-sphere and B4 the standard 4-ball such that ∂B4 = S3. A knot K in S3 is called smoothly slice if it bounds a properly embedded smooth disk D2 in B4. We call D2 a slice disk for K, see Figure 2.

Figure 2. A schematic picture of a slice knot K and a slice disk D2.

The class of slice knots are conjectured to be that of knots which are defined 3-dimensionally. A knot K in S3 is called ribbon if it bounds an immersed disk D in S3 with only ribbon singularities.

Lemma 4.1. A ribbon knot is slice.

Proof. Let K be a ribbon knot. By definition, it bounds an immersed disk D2

in S3 with only ribbon singularities. By pushing intD2 toward the interior of B4, we obtain a slice disk for K.

Here we give a more precise proof. Let N be the color neighborhood of B4. Then N is diffeomorphic to S3× [0, 1] and S3×{0} ≈ S3 = ∂B4. If we deform D2 as Figure 4, then we obtain a slice disk for K. □

The slice-ribbon conjecture� � All slice knots are ribbon.� �

Figure 3. A ribbon singularity and an example of a ribbon knot.

4 TETSUYA ABE

S 3 × {0} ≈ S3

S 3 × {

1

2 }

S 3 × {1}

Figure 4. A slice disk obtained from an immersed disk.

This conjecture is due to Fox [F2]. I can not guess whether this conjecture is true or not.

Positive direction. There are some results which suggest that the slice-ribbon conjecture is true.

In 2007, Lisca [Li] proved that the slice-ribbon conjecture is true for two- bridge knots by a gauge theoretic method. Note that this result is based on the work of [CG2]. Further development was done by Greene and Jabuka [GJ] and Lecuona [Le1], [Le2].

Open problem. Is the slice-ribbon conjecture true for three-bridge knots ?

Negative direction. There are some results which suggest that the slice-ribbon conjecture is not true. Indeed, there exist slice knots which may not be ribbon. Such slice knots are obtained a byproduct of the study of the smooth Poincaré conjecture in dimension four. See also Section 14.

5. The smooth Poincaré conjecture in dimension four

1980’s, Freedman classified simply-connected closed 4-manifolds. As a corol- lary, Freedman solve the topological Poincaré conjecture in dimension four.

Theorem 5.1 (The topological Poincaré conjecture in dimension four). Let X be a topological 4-manifold. If X is homotopic to S4, then X is homeomorphic to S4.

CONSTRUCTIONS OF SMOOTHLY SLICE KNOTS 5

On the other hand, the smooth Poincaré conjecture in dimension four (SPC4) is not yet solved. The statement is the following.

The smooth Poincaré conjecture in dimension four (SPC4)� � If a smooth 4-manifold X is homeomorphic to S4, then X is

diffeomorphic to S4.� � This conjecture is one of the biggest unsolved problems in low dimensional

topology. Note that Donaldson proved that there exists an exotic R4 = S4 \ {pt}, which suggests that there might exist an exotic S4. Furthermore it is well known that there exist uncountably many exotic R4’s due to Taubes.

6. Cappell-Shaneson (homotopy) 4-spheres

For the matrix

An =

0 1 00 1 1 1 0 n+ 1

∈ SL(3,Z), Cappell and Shaneson2 associated a homotopy 4-sphere Σn. The homotopy 4-sphere Σn is obtained as follows: First, consider the mapping torus of the punctured 3-torus T 30 with the diffeomorphism induced by An. Then by gluing a S2 ×D2 to it with the non-trivial diffeomorphism of S2 × S1, we obtain Σn.

In 1991, Gompf [G1] proved that

Σ0 ≈ S4.

For a long time, many people thought that Σ1 might be an exotic S 4. In 2010,

Akbulut [A1] proved that

Σn ≈ S4.

7. Another story: an idea to disprove SPC4

This is a story before Akbulut’s work [A1] and this section can be skipped.

Let W be a homotopy 4-ball with ∂W ≈ S3. If there exists a knot K in S3 such that

• K bounds a smooth disk in W , and • K does not bound a smooth disk in B4,

2They constructed more homotopy 4-spheres, see [G2]. In this note, we only consider Σn.

6 TETSUYA ABE

then W is an exotic 4-ball. In particular, the 4-manifold W ∪idB4 is an exotic 4-sphere 3. In [FGMW], Freedman, Gompf, Morrison and Walker constructed a candidate of such a knot.

The construction is the following: Consider Σ1 and its handle decomposition

h0 ∪ h11 ∪ h12 ∪ h21 ∪ h22 ∪ h4

given by Gompf in [G1], where h0 is a 0-handle, h 1 i (i = 1, 2) is a 1-handle, h

2 j

(j = 1, 2) is a 2-handle, h4 is a 4-handle. Note that

h0 ∪ h11 ∪ h12 ∪ h21 ∪ h22 is a homotopy 4-ball whose boundary is diffeomorphic to S3 and we denote it by W1. Let Ki (i = 1, 2) be the belt-sphere of h

2 i . Then Ki bounds a smooth

disk in W1. Freedman, Gompf, Morrison and Walker consider a band sum of K1 and K2, denoted by K1♯bK2, where b is a certain band in S

3, see [FGMW]. The knot K1♯bK2 has the following properties:

• K1♯bK2 bounds a smooth disk in W1, and • K1♯bK2 may not bound a smooth disk in B4.

The problem is how to prove that K1♯bK2 does not bound a smooth disk in B4. In this century, two strong obstructions for sliceness are introduced. One of them is the τ -invariant which is derived from the knot Floer homology. The other is the s-invariant which is derived from the Khovanov homology. The properties of these invariants are the following.

Theorem 7.1. Let K be a smoothly slice knot. Then

τ(K) = s(K) = 0.

Furthermore, if K bounds a smooth disk in a homotopy 4-ball, then

τ(K) = 0.

The point is that, in 2010, it was not known whether s(K) is zero or not for a knot K which bounds a smooth disk in a homotopy 4-ball. Therefore there was a hope that s(K1♯bK2) ̸= 0 which implies that the smooth Poincaré conjecture in dimension four is false.

Freedman, Gompf, Morrison andWalker calculated the s-invariant ofK1♯bK2 using a supercomputer, turning out to be s(K1♯bK2) = 0. Soon after their work, as described before, Akbulut [A1] proved that W1 (and Σ1) is standard.

3A proof of this statement is the following. Suppose that W ∪id B4 is standard, that is, which is diffeomorphic to the standard 4-sphere B4 ∪id B4. The embedding of B4 into a connected 4-manifold is unique up

Recommended

*View more*