SLICE KNOTS AND KNOT CONCORDANCE

ARUNIMA RAY

Abstract. These notes were prepared to accompany a sequence of three lectures at the
conference Winterbraids XI in Dijon, held in December 2021. In them, we provide an introduction
to slice knots and the equivalence relation of concordance. We explain some connections between
slice knots and exotic smooth structures on R4 . We also introduce filtrations of the knot
concordance groups and satellite operations.
arXiv:2311.12168v1 [math.GT] 20 Nov 2023

Overview
Slice knots were first defined in 1958 by Fox and Milnor and have since become a flourishing
field of study. While they were originally considered in the context of resolving singularities of
surfaces in 4-manifolds, numerous other connections to questions in 3- and 4-manifold topology
have been discovered. A notable highlight: every knot which is topologically slice but not
smoothly slice gives rise to an exotic smooth structure on R4 ; more on this in Section 2.
The goal of these lecture notes is to provide an overview of the basic notions in the study
of slice knots and knot concordance, leading up to a small selection of recent developments.
Due to the limitations of time we will barely scratch the surface of this active and vibrant area.
Nevertheless we will attempt to provide some references and pointers to other resources.

Exercises. There are exercises throughout the lecture notes, which are also compiled in a
list at the end. The problems are separated into three levels. Green △ exercises are usually
straightforward and should be attempted if you are seeing this material for the first time.
Prerequisites are courses in introductory geometric and algebraic topology. Orange □ exercises
are for readers who are already comfortable with some of the terminology; they may require
nontrivial input from outside these lectures, which we have tried to indicate as hints. Finally,
red ⃝ exercises are challenge problems. Open problems will be marked as such, and do not
intersect with the exercises.

Relationship to lectures. Some details and exercises in these notes were not mentioned in
the accompanying lectures. The order of topics has also been slightly modified. Many of the
exercises are new.

Conventions. Homeomorphism of manifolds is denoted by the symbol ≈, while diffeomorphism

is denoted by ∼=. All manifolds are assumed to be oriented. All knots are assumed to be tame and
oriented, i.e. (images of) smooth embeddings S 1 ,→ S 3 , where both the domain and codomain
are oriented by hypothesis. The set of nonnegative integers is denoted by N.

Acknowledgements. I am grateful to the organisers of Winterbraids XI in Dijon for an excellent

conference, especially considering the Covid-19 pandemic, and to the attendees for their lively
participation and many questions. Many thanks are due as well to the anonymous referee for
their helpful comments.

1. Slice and ribbon knots

Recall that a knot K : S 1 ,→ S 3 is trivial or the unknot if and only if it bounds an embedded
disc in S 3 . We will often conflate a knot and its image, and it will be important that both S 1
and S 3 are oriented. See Figure 1 for some examples of knots. Here, by definition, a knot is
trivial if it can be deformed via an ambient isotopy to the round unit circle S 1 ⊆ R3 ⊆ S 3 . For
more on classical knot theory, see e.g. [Fox62, Rol90, Liv93].
1
 SLICE KNOTS AND KNOT CONCORDANCE 2

Figure 1. The two knots on the left are trivial. The three knots on the right
are not trivial, and are respectively called the right-handed trefoil, the left-handed
trefoil, and the figure eight knot.

Slice knots are, in some sense, generalised trivial knots – we still ask for the knot to bound an
embedded disc, but it has the freedom of an additional (half-)dimension.
Definition 1.1. A knot K ⊆ S 3 = ∂B 4 is smoothly slice if it bounds a smoothly embedded disc,
called a smooth slice disc, in B 4 .1
We begin by discussing a number of elementary ways to construct new slice knots from a given
slice knot. Let K : S 1 ,→ S 3 be a knot. The knot obtained by reversing the orientation of S 1 is
called the reverse of K, denoted by rK. The knot obtained by reversing the orientation of S 3 is
called the mirror image of K, denoted by K. For example, the left-handed trefoil is the mirror
image of the right-handed trefoil. Both trefoils are isotopic to their reverses. The figure eight is
isotopic to both its reverse and its mirror image.
Proposition 1.2 (Exercise △ 3). Let K ⊆ S 3 be a knot. Show that K is smoothly slice if and
only if rK is smoothly slice if and only if K is smoothly slice.
Given knots K, J ⊆ S 3 , we can form their connected sum by taking the (oriented!) connected
sum of pairs (S 3 , K)#(S 3 , J); see Figure 2. Given diagrams of knots K and J we can form a
diagram of the connected sum by connecting the two diagrams by a trivial band, as shown in
the figure, taking care to match the original orientations. This is a special case of a band sum
defined later in this section.

Figure 2. The connected sum of the figure eight and the right-handed trefoil
from Figure 1.

Proposition 1.3 (Exercise △ 4). If the knots K, J ⊆ S 3 are smoothly slice, then so is K#J.
By ambient Morse theory, we can isotope any smooth slice disc so that the level sets with
respect to the radial function on B 4 are links, i.e. (the image of) a smooth embedding of a disjoint
union of circles in S 3 , except at finitely many radii where the level sets have singularities. Since
we are considering a disc in a 4-manifold, there are only three possible types of singularities:
minima, saddles, and maxima. See Figure 3 for a picture and [Mil65; Mil63; Mat02; Nic11;
GS99, Sections 4.2 and 6.2] for more on Morse theory.
Smooth slice discs, once they have been perturbed to respect the radial Morse function as
above, can be described in terms of movies. More specifically, we consider the radial level sets
as the radius increases from 0 to 1, with the latter corresponding to ∂B 4 = S 3 containing a
smoothly slice knot. At each minimum, an unknot is born, split from every other component, if
1In 1958, Fox and Milnor called these null-equivalent knots, in unpublished work. Fox adopted the term slice
knot, proposed by Edward Moise, in [Fox62].
 SLICE KNOTS AND KNOT CONCORDANCE 3

y, z y, z

x x

w w

Figure 3. The axes schematically indicate B 4 , where the coordinates x, y, z

describe R3 ⊆ S 3 and the coordinate w ≥ 0 indicates depth, so that S 3 occurs at
w = 0. In both cases we have a knot in R3 ⊆ S 3 = ∂B 4 . On the left we have a
trivial knot, as detected by the embedded disc in S 3 shaded grey. On the right
we have a smoothly slice knot, with a smooth slice disc lying in B 4 . The disc
is in a particularly nice position with respect to the radial function on B 4 . At
most w-slices, the disc appears as a link, i.e. an embedding of a disjoint union of
circles. At finitely many values of w, we see singularities. There are three types
of singularities: minima (red), saddles (orange), and maxima (blue). See Figure 4
for a more precise diagram.

any. At every saddle singularity, two portions of a link either merge together, or split apart. At
a maximum, an unknotted component, split from everything else, disappears. See Figure 4 for
an example.
We especially like smooth slice discs with only two types of singularities, as in the following
definition.
Definition 1.4. A knot K ⊆ S 3 = ∂B 4 is ribbon if it bounds a smoothly embedded disc in B 4
with only minima and saddles. Such a disc is called a ribbon disc.
Given a link L ⊆ S 3 , a band B is a copy of [0, 1] × [0, 1] embedded in S 3 , with {0, 1} × [0, 1]
lying on L (matching the orientations), and otherwise disjoint from L. The arc [0, 1] × {1/2} is
called the core of B. A band sum of L along the band B is the result of removing {0, 1} × [0, 1]
from L and gluing in [0, 1] × {0, 1}; see Figure 5. When the components of {0, 1} × [0, 1] lie on
distinct components of L the result is also called a fusion of L along B. Note that the core
[0, 1] × {1/2} of a band can be an interesting arc in S 3 ∖ L, and in particular can link with the
components of L. A band sum on L along B can be reversed by performing a dual band move,
and corresponds to removing [0, 1] × {0, 1} from the band sum and gluing {0, 1} × [0, 1] back in.
One notices that the result is isotopic to the original link L.
From the movie perspective, reading the ribbon disc from the bottom up, we see that any
ribbon knot is a fusion of some trivial link. (Why is it a fusion, rather than an arbitrary
band sum?) In other words, it can be obtained from an n-component trivial link for some
n (corresponding to minima) and a collection of n − 1 bands (corresponding to saddles), by
performing band sums, each of which joins two distinct components of the original link, and no
two of which join the same pair of components. Alternatively, we could run the movie backwards.
From this perspective, a ribbon knot is a knot where one can add n − 1 bands for some n
(corresponding to the dual band moves described above), so that the result is an n-component
unlink.
The following shows that ribbon knots can be described purely 3-dimensionally. This gives a
useful method to detect if a given knot is ribbon.
Proposition 1.5 (Exercise □ 20). A knot K ⊆ S 3 is ribbon if and only if it bounds a disc in S 3
with only ribbon singularities, i.e. singularities of the form shown in Figure 6.2
Proposition 1.6 (Exercise △ 6). For any knot K ⊆ S 3 , the knot K#rK is ribbon.
2Such singular discs in S 3 are also sometimes called ribbon discs.
 SLICE KNOTS AND KNOT CONCORDANCE 4

Figure 4. Describing a ribbon disc as a movie. At the very bottom, we have two
minima. At a slightly larger radius in B 4 , the minima give rise to (split) unknots,
i.e. an unlink. The subsequent cross sections, until the orange saddle singularity,
describe an isotopy of this unlink. At the saddle singularity, two portions of the
previous unlink touch each other, and at a slightly higher radius we see a ribbon
knot.

From Definition 1.4, one sees directly that any ribbon knot is smoothly slice. (For the equivalent
definition given in Proposition 1.5, one could push a small region of the ribbon disc in S 3 into
B 4 to produce an embedded disc, as indicated in Figure 6.) The converse is an important open
problem.
Conjecture 1.7 (Open, slice-ribbon conjecture, Fox [Fox61, Problem 25]). Every smoothly slice
knot is ribbon.
This conjecture has been established for certain families of knots, e.g. 2-bridge knots [Lis07]
and most 3-strand pretzel knots [GJ11, Lec15], but there is no general approach to attacking it at
the moment. Other work on the slice-ribbon conjecture using a similar strategy includes [Lec12,
Bry17,Lon14,AKPR21]. Some potential counterexamples have been proposed, e.g. [GST10,AT16].
A very promising candidate had been the (2, 1)-cable of the figure eight, which was known to be
nonribbon for twenty years [Miy94] but whose sliceness status was unknown. It was recently
shown to not be smoothly slice [DKM+ 22] (see also [ACM+ 23]). Notably it is still open whether
it is topologically slice (Definition 1.10).
 SLICE KNOTS AND KNOT CONCORDANCE 5

L
B

(a) (b)

Figure 5. (a) A link L (black) and a band B (dark green). The core of B is
shown in light green. (b) The result of the band sum on L along B is shown in
black. The arc guiding the dual band move reversing the previous band move is
shown in blue.

Figure 6. Left: A ribbon singularity (shown in blue) of a disc (shown in grey) in

S 3 . Right: The smaller disc region encircled in red on the horizontal sheet could
be pushed radially into B 4 to resolve the singularity.

The subtlety of the conjecture is seen by the following proposition. Specifically, to show that
a smoothly slice knot is ribbon it is not a good strategy to prove that a given smooth slice disc
is ribbon.
Proposition 1.8 (Exercise □ 22). There exist smooth slice discs that are not ambiently isotopic
(relative to the boundary) to any ribbon disc.

1.1. Topologically slice knots. So far we have worked strictly in the smooth setting. We can
however loosen this restriction slightly. We will need the following definition.
Definition 1.9. An embedding φ : (F, ∂F ) ,→ (M, ∂M ), i.e. a continuous map which is a
homeomorphism onto its image, of a surface F in a 4-manifold M is said to be locally flat if for
all x ∈ F there is a neighbourhood U of φ(x) such that (U, U ∩ φ(F )) is homeomorphic to either
(R4 , R2 ), in the case that x ∈ Int F , or to (R4+ , R2+ ), in the case that x ∈ ∂F .
The above gives rise to another notion of sliceness for knots.
Definition 1.10. A knot K ⊆ S 3 = ∂B 4 is topologically slice if it bounds a locally flat embedded
disc, called a topological slice disc, in B 4 .
It is straightforward to see that any smoothly slice knot is topologically slice. Via a deep
result of Quinn [Qui82, Theorem 2.5.1; FQ90, Theorem 9.3], a topological slice disc admits a
tubular neighbourhood, i.e. we could equivalently define a knot K to be topologically slice if
it bounds an embedded disc ∆ in B 4 , admitting a neighbourhood homeomorphic to ∆ × D2 ,
intersecting S 3 in a tubular neighbourhood of K.
Remark 1.11. We motivated slice knots by saying that they generalise trivial knots. Certainly
the trivial knot is topologically slice. We should however be careful not to be too general. Given
any knot K ⊆ S 3 , the coned disc cone(K) ⊆ cone(S 3 ) = B 4 is an embedded disc bounded by K
in B 4 . See Figure 7. As shown in the following proposition, this disc is not in general locally flat.
 SLICE KNOTS AND KNOT CONCORDANCE 6

Figure 7. Let K ⊆ S 3 be a knot. Thinking of B 4 as the cone on S 3 , we get an

embedded disc cone(K) ⊆ B 4 .

Proposition 1.12 (Exercise △ 7). Let K ⊆ S 3 be a knot. The coned disc cone(K) ⊆ cone(S 3 ) =
B 4 is locally flat if and only if K is the trivial knot.
The following result of Freedman and Quinn gives a powerful method to construct topologically
slice knots. For the definition of the Alexander polynomial, see e.g. [Rol90, Chapter 8]. (See [Fel16]
for a generalisation to the topological slice genus.)
Theorem 1.13 ([FQ90, Theorem 11.7B; GT04]). Let K ⊆ S 3 be a knot. If the Alexander
.
polynomial ∆K (t) = 1, then K is topologically slice.
As a consequence of the above theorem, the (untwisted) Whitehead double of any knot
(Figure 8) is topologically slice, since it has Alexander polynomial one. On the other hand, many
such knots are not smoothly slice.

S 1 × D2

Wh+
K Wh+ (K)

Figure 8. Whitehead doubling. On the left we have a knot Wh+ in a solid

torus S 1 × D2 . In the middle we show the figure eight knot. Given an arbitrary
knot K, the (positive clasped, untwisted) Whitehead double of K, denoted by
Wh+ (K), is the image of Wh+ under a map identifying S 1 × D2 with a tubular
neighbourhood of K, so that the longitude S 1 × {∗} of S 1 × D2 is sent to the
Seifert longitude of K. On the right we show the (positive clasped, untwisted)
Whitehead double of the figure eight. The negative clasped Whitehead double
Wh− (K) is defined in the same way, using as input the knot Wh− in S 1 × D2
obtained from Wh+ by changing both crossings. Whitehead doubling is a special
case of the satellite construction on knots (Definition 5.1).

Theorem 1.14. There exist knots that are topologically slice but not smoothly slice, e.g. the
positive clasped untwisted Whitehead double of the right-handed trefoil knot.
The above was first shown by Casson and Akbulut, in independent unpublished work, using
Donaldson’s diagonalisation theorem [Don83, Don86, Don87]. The first published accounts are by
Gompf [Gom86], Cochran–Lickorish [CL86], and Cochran–Gompf [CG88].
 SLICE KNOTS AND KNOT CONCORDANCE 7

However, for many examples of Whitehead doubles, it is still open whether they are smoothly
slice. For the next question, let L denote the left-handed trefoil and let 41 denote the figure
eight.
Question 1.15 (Open). Is Wh+ (L) or Wh± (41 ) smoothly slice?
It is not difficult to see that if the knot K is smoothly slice, then so is Wh± (K); see Exercise △ 8.
The converse is an interesting open question, generalising Question 1.15.
Conjecture 1.16 (Open [Kir97, Problem 1.38]). Let K ⊆ S 3 be a knot. The knot W h± (K) is
smoothly slice if and only if K is smoothly slice.
As mentioned earlier, Donaldson’s diagonalisation theorem was used to show that the positive
clasped untwisted Whitehead double of the right-handed trefoil is not smoothly slice. This
is a relatively subtle sliceness obstruction, since e.g. it does not obstruct topological sliceness.
There are numerous other slicing obstructions, including many topological sliceness obstructions.
We will return to this topic in Section 4, where we will explain a method to arrange sliceness
obstructions in order of strength, in some sense. We defer that discussion to first describe some
connections between slice knots and 4-manifold topology.

1.2. Links. The notions in this section can be generalised to links, i.e. embeddings ⊔S 1 ,→ S 3 .
For example, we have the following definition.
Definition 1.17. A link L ⊆ S 3 = ∂B 4 is said to be (strongly) smoothly slice if the components
bound a collection of pairwise disjoint smooth slice discs in B 4 . If the discs are only locally flat,
the link is said to be (strongly) topologically slice.
There is a parallel notion of weak sliceness of links, in both the topological and smooth settings.
For this we do not require a collection of pairwise disjoint embedded discs, but rather just some
embedded planar surface (either smooth or locally flat).
Unfortunately we will not have time to pursue links thoroughly. We limit ourselves to saying
that the open questions mentioned in this section have analogues for links: it is open whether
the slice-ribbon conjecture is true for links and whether the Whitehead double of a link L is
smoothly slice if and only if L is smoothly slice. It is also open whether the Whitehead double of
every link is topologically slice, and an answer would have deep consequences for fundamental
open questions in 4-manifold topology (see e.g. [KOPR21]).

2. Exotic smooth structures on R4

By definition, a smooth structure on a topological manifold M is a maximal atlas of charts
on M such that the transition maps are smooth. A priori a given topological manifold can
admit finitely many (including none) or infinitely many smooth structures. The study of
existence and uniqueness questions for smooth structures comprises the field of smoothing
theory. Notably, the sphere S 7 has 28 distinct smooth structures up to orientation preserving
diffeomorphism [Mil56, KM63], and whether S 4 has more than one smooth structure is the
content of the (still open) smooth 4-dimensional Poincaré conjecture. Smoothing theory is well
developed in dimensions five and higher, and for example, using technology of Kirby–Siebenmann
from [KS77], it can be shown that closed manifolds of dimension five and higher have finitely many
(possibly zero) smooth structures. Smoothing closed 4-manifolds is more complicated. There
exist topological 4-manifolds which do not admit any smooth structures, such as Freedman’s E8-
manifold [Fre82], while there are closed 4-manifolds admitting infinitely many smooth structures,
such as the K3 surface [FS98].
The existence and uniqueness questions for smooth structures can also be asked for noncompact
manifolds. The following result shows that Euclidean spaces are well behaved assuming n ̸= 4.
̸ 4 there is a unique smooth structure on Rn , up to
Theorem 2.1 ([Moi77, Sta62]). For n =
diffeomorphism.
By contrast, there are uncountably many smooth structures on R4 [Tau87].
 SLICE KNOTS AND KNOT CONCORDANCE 8

Definition 2.2. Let R4std denote R4 with its standard smooth structure. A smooth 4-manifold
which is homeomorphic to R4std , but not necessarily diffeomorphic to R4std , is called an R4 -
homeomorph. An R4 -homeomorph which is not diffeomorphic to R4std is called an exotic R4 , and
such a smooth structure is called an exotic smooth structure on R4 .
There are two distinct types of R4 -homeomorphs, as per the following definition.
Definition 2.3. Let R be an R4 -homeomorph. The manifold R is said to be small if it admits
a smooth embedding R ,→ R4std . It is called large otherwise.3
Knots and links feature prominently in the known constructions of exotic R4 s. In this section
we describe two such constructions, one of large exotic R4 s, and one of small exotic R4 s.

2.1. Large exotic R4 s using 0-traces. For any knot K we have the 0-trace, which is by
definition the smooth 4-manifold obtained by adding a 0-framed 2-handle to B 4 along K ⊆ S 3 ,
and then smoothing corners. One of the two key ingredients in the forthcoming construction is
the following characterisation of slice knots.
Lemma 2.4 (Trace embedding lemma, Exercise □ 23). Let K ⊆ S 3 be a knot. The 0-trace
X0 (K) admits a smooth (resp. locally collared) embedding into R4std if and only if K is smoothly
(resp. topologically) slice.
We will also need the following powerful result.
Theorem 2.5 ([Qui82, Corollary 2.2.3; FQ90, Theorem 8.2, Section 8.7]). Let M be a connec-
ted, noncompact 4-manifold. If desired, fix a smooth structure on any collection of connected
components of ∂M . There is a smooth structure on M extending the given smooth structure on
(a subset of) ∂M .
We are now ready for the construction.
Construction 2.6 ([Gom85, Lemma 1.1; GS99, Exercise 9.4.23]). Given a knot K ⊆ S 3 which
is topologically slice but not smoothly slice there exists R, a large exotic R4 , with a smooth
embedding X0 (K) ,→ R.
Proof. Let K be a knot which is topologically slice but not smoothly slice. Then there is a
locally collared embedding φ : X0 (K) ,→ R4 by the trace embedding lemma. Then the image
φ(X0 (K)) inherits a smooth structure from X0 (K) since it is homeomorphic to the latter. Since
φ is locally collared, we see that R4 ∖ Int φ(X0 (K)) is a manifold, and we can check using the
Mayer–Vietoris sequence that it is connected. Since X0 (K) is compact and R4 is not, we also
know that R4∖ Int φ(X0 (K)) is noncompact. So by Theorem 2.5, we extend the smooth structure
on φ(∂X0 (K)) to the rest of R4 ∖ Int φ(X0 (K)). This produces a smooth structure on R4 , and
we denote the corresponding smooth 4-manifold by R.
Suppose that R ∼ = R4std . Then by construction we have a smooth embedding X0 (K) ,→ R ∼ =
4
Rstd , so K is smoothly slice by the trace embedding lemma, which is a contradiction. Therefore
R is an exotic R4 .
Indeed, since X0 (K) ⊆ R has no smooth embedding in R4std , we see that R is large.4 □

2.2. Small exotic R4 s using ribbon disc exteriors. The construction in this section will use
Casson handles. These are smooth noncompact 4-manifolds, constructed by Casson in [Cas86] as
approximations of (open) 2-handles. The boundary of a Casson handle C, called the attaching
region ∂C, is identified with the open solid torus S 1 × Int D2 , and from the Kirby diagrams of
Casson handles (see, e.g. [Kir89, Chapter XII; GS99, Chapter 6; Fre82, Section 2]), one observes
that every Casson handle C admits a smooth embedding (C, ∂C) ,→ (D2 × D2 , S 1 × Int D2 ),
extending the aforementioned identification on the boundary. In [Fre82], Freedman showed that
any Casson handle C is homeomorphic, relative to its attaching region, to an open 2-handle,
3There are two slightly different definitions of small/large R4 -homeomorphs in the literature. Ours coincides
with the definition of Scorpan’s book [Sco05]. In the book by Gompf–Stipsicz [GS99], a large R4 -homeomorph is
one which contains a compact subset that does not embed in R4std , and an R4 -homeomorph is called small if it is
not large. It is not known whether the two definitions are equivalent.
4This holds for either definition of ‘large’.
 SLICE KNOTS AND KNOT CONCORDANCE 9

i.e. (C, ∂C) ≈ (D2 × Int D2 , S 1 × Int D2 ), again extending the aforementioned identification on
the boundary.
Construction 2.7. Given a smoothly slice link L with n components, and a collection of Casson
handles {C1 , . . . Cn }, there exists a small R4 -homeomorph RL .
Proof. Let {∆i } be a collection of smooth slice discs for L. Consider the complement B 4∖ ν∆i ,
S

of open tubular neighbourhoods of the discs. If we glue in 2-handles along the meridians of the
components of L, we will get back B 4 . However, we could instead glue in the Casson handles
{C1 , .S
. . Cn } along those meridians. Let RL denote the result of gluing in Casson handles to
B 4 ∖ ν∆i along the meridians of L, and then removing all the remaining boundary. Since
Casson handles are homeomorphic to open 2-handles relative to the attaching region we see that
RL is an R4 -homeomorph. Since every Casson handle embeds in a standard 2-handle, respecting
the attaching region, we also see that RL admits a smooth embedding into R4std , and so by
definition RL is small. □
For certain choices of L and Casson handles {C1 , . . . , Cn }, it can be shown that RL is not
diffeomorphic to R4std . For examples of this, see [DMF92, BG96]. Roughly speaking, showing that
such an RL is exotic involves embedding it appropriately within a simply connected h-cobordism
which is known to not be a smooth product. The known examples use ribbon links and are
therefore called ribbon R4 s. All known small exotic R4 s are ribbon R4 s. The simplest known
exotic ribbon R4 is built using the complement of a standard ribbon disc for the 946 knot (also
called the (3, −3, 3) pretzel knot), and the Casson handle built using self-plumbed 2-handles
within a single, positive self-plumbing at each stage [BG96] – this explicit construction yields
a description of the corresponding exotic R4 as the interior of an infinite but rather simple
handlebody.

2.3. Universal exotic R4 . Finally, there is the Freedman–Taylor universal exotic R4 [FT86],
denoted by U, which has the remarkable property that every R4 -homeomorph R admits a smooth
embedding R ,→ U. The construction of U also involves solving slicing problems for knots
and links (via Kirby diagrams for Casson handles) and applying Quinn’s result (Theorem 2.5).
Unfortunately the construction is beyond the scope of the lectures and we invite the reader to
learn more about the construction in [FT86] on their own.

3. The knot concordance groups

So far we have considered sliceness as a generalisation of triviality for knots. Similarly, there
is a generalisation of isotopy, called concordance. Specifically, we have the following definition.
Definition 3.1. Knots K, J ⊆ S 3 are said to be smoothly (resp. topologically) concordant if
they cobound a smooth (resp. locally flat) embedded annulus in S 3 × [0, 1]; see Figure 9. More
specifically, we consider K ⊆ S 3 as lying in S 3 × {1} and J ⊆ S 3 as lying in S 3 × {0}, and assert
the existence of a concordance A = S 1 × [0, 1] ,→ S 3 × [0, 1], with S 1 × {0} mapping to J and
S 1 × {1} mapping to K, where the embedding of A is either smooth or locally flat, as needed.
Proposition 3.2 (Exercise △ 9). A knot K ⊆ S 3 is smoothly (resp. topologically) slice if and
only if it is smoothly (resp. topologically) concordant to the unknot.
Recall that by our conventions, all the manifolds in the definition of concordance are oriented.
Using the standard orientation of [0, 1], we notice that S 3 × {1} inherits the standard orientation
of S 3 while S 3 × {0} inherits the opposite orientation. Similarly a concordance (i.e. a copy of
S 1 × [0, 1]) induces opposite orientations on its boundary components. Under these orientations,
we see that if K and J are concordant knots, then the claimed annulus joins K ⊆ S 3 × {1} = S 3
to rJ ⊆ S 3 × {0} = −S 3 . The latter knot is by definition rJ = rJ.
When speaking about slice knots we often ignore these orientation conventions, but they are
key to proving the following fact.
Proposition 3.3 (Exercise □ 24). Two knots K, J ⊆ S 3 are smoothly (resp. topologically)
concordant if and only if K#rJ is smoothly (resp. topologically) slice.
 SLICE KNOTS AND KNOT CONCORDANCE 10

S 3 × [0, 1]

Figure 9. A schematic picture showing a concordance between two knots J ⊆

S 3 × {0} and K ⊆ S 3 × {1}.

Similar to isotopy, concordance (both smooth and topological) is an equivalence relation.

However, unlike isotopy (Exercise □ 25), the equivalence classes under concordance form a group
under the connected sum operation.
Proposition 3.4 (Exercise △ 10). Smooth (resp. topological) concordance is an equivalence
relation on knots.
Definition 3.5 (Concordance groups). The smooth concordance group, denoted by C, consists of
smooth concordance classes of knots, under the operation of connected sum.
Similarly, the topological concordance group, denoted by C top , consists of topological concord-
ance classes of knots, under the operation of connected sum.
We often conflate a knot K with its concordance class [K]. We asserted but did not prove
that concordance classes form a group. This is the combination of the following proposition and
Exercise △ 2.
Proposition 3.6 (Exercise △ 11). Let K ⊆ S 3 be a knot. In either C diff or C top , the inverse
of [K] is the class of −K := rK. The identity element in C diff (resp. C top ) is the class of the
unknot, equivalently the class of any smoothly (resp. topologically) slice knot.
By Exercise △ 2, the concordance groups are abelian, so it is tempting to speculate that they
cannot be particularly complicated. This is somewhat of a red herring. There is an algebraic
version of the knot concordance group, called the algebraic concordance group, denoted by AC.
This group has a couple of equivalent definitions, one in terms of Seifert matrices and another in
terms of Alexander modules and Blanchfield forms – see [Lev69a, Lev69b, Kea73, Tro73, Kea75a,
Kea75b, Sto77, Tro78] for further details. Given the definitions, it is not hard to see that there is
a surjection
Φ : C top ↠ AC,
given by taking a knot to the class of its Seifert matrix, or of its Alexander module and Blanchfield
form. It was shown by J. Levine [Lev69b] (see also Stoltzfus [Sto77]) that
AC ∼ = Z∞ ⊕ Z/2∞ ⊕ Z/4∞ ,
so C top is quite large.
Notice that AC has many elements of order two. So do C diff and C top by the following
proposition.
Proposition 3.7 (Exercise △ 12). A knot K ⊆ S 3 is said to be negative amphichiral if it is
isotopic to rK. Every negative amphichiral knot has order at most two in C diff and C top . As a
result, the figure eight knot has order two in C diff and C top .
Conjecture 3.8 (Open, Gordon [Kir97, Problem 1.94]). Let K ⊆ S 3 be a knot. If K#K is
slice, then K is smoothly concordant to a negative amphichiral knot.
By contrast, there are no known candidates for knots of order four (note that AC has many
such elements). We have the following open question.
 SLICE KNOTS AND KNOT CONCORDANCE 11

Question 3.9 (Open, see [Kir97, Problem 1.32]). Does the map Φ split? In particular, does
C top (or C diff ) have any elements of order four? Indeed, are there any knots with finite order
̸= 0, 2 in C diff or C top ?
The kernel of the map Φ has also been an object of much study, and is the focus of the next
section.
Remark 3.10. One may also consider high-dimensional knots, i.e. embeddings S n ,→ S n+2 for
arbitrary n. As in the classical case, one then has corresponding concordance and algebraic
concordance groups. Indeed, for even n, every embedding S n ,→ S n+2 is slice, in the smooth,
piecewise linear, or smooth categories [Ker65, Théorème III.6; Ker71, Theorem 1]. For odd
n ≥ 5, the topological and piecewise linear concordance groups of such embeddings is precisely
in bijection with the corresponding algebraic concordance group (for n = 3 there is a discrepancy
of Z/2 related to Rochlin’s theorem) [Lev69b, Ker71]. In the smooth category, the odd high-
dimensional concordance groups are also closely determined by the corresponding algebraic
concordance group, with a discrepancy of Z/k for some k – this discrepancy is related to the
existence of exotic smooth structures on high-dimensional spheres and the famous Kervaire
invariant problem (see [HHR16] for an account of the history of this problem). Roughly speaking,
the map from knot concordance classes to the algebraic concordance group is still injective, as in
the case for topological or piecewise linear concordance in high dimensions, but we no longer get
a surjection since the construction in the realisation step may produce an exotic sphere rather
than a standard one.
From the definitions of the concordance groups it should be clear that there is a surjection
Ψ : C diff ↠ C top . (3.1)
As mentioned in Section 1, there exist knots that are topologically slice but not smoothly slice.
These (or rather, their smooth concordance classes) lie in the kernel of Ψ.
Definition 3.11. The subset T ⊆ C diff consists of the smooth concordance classes of topologically
slice knots.
The structure of T is another key area of research (see e.g. [Gom86, CG88, End95, HK12,
HLR12, CHH13, Hom14, CH15, Hom15, HKL16, OSS17, Hom19, CK21, DHST21, KL22]). One
method of studying this structure is described in the following section.

4. Filtrations of the knot concordance groups

4.1. Overview. As mentioned in Remark 3.10, the high-dimensional codimension two con-
cordance groups of odd-dimensional knots are controlled almost entirely by the corresponding
algebraic analogues, and in almost all cases they are isomorphic. Therefore, it was an interesting
question for several years whether the knot concordance groups C top and the algebraic knot
concordance group AC are isomorphic also in the classical dimension, in other words, whether
the map Φ is a bijection. Since it is easily shown to be a surjection, the main question was
whether Φ is injective. The first nontrivial elements in the kernel of Φ were given by Casson
and Gordon in [CG78, CG86]. Then Cochran, Orr, and Teichner [COT03] defined the solvable
filtration of C diff , denoted by {Fi } 1 , satisfying
i∈ 2 N
\
Fi ⊆ · · · ⊆ Fn+1 ⊆ Fn.5 ⊆ Fn ⊆ · · · ⊆ F0 ⊆ C diff .
1
i∈ 2 N

Remark 4.1. Cochran–Orr–Teichner defined the filtration in the locally flat category, for the
topological concordance group C top . In light of future developments we give the definition
in the smooth category. For the relationship between the two notions see Proposition 4.7
and Corollary 4.8.
We will give the precise definition of the terms of this filtration presently. For now, we note
that it is highly nontrivial (see Theorem 4.9), and it subsumes several other invariants, e.g. a
knot K is in F0 if and only if Arf(K) = 0; it is in F0.5 if and only if it is algebraically slice; and
 SLICE KNOTS AND KNOT CONCORDANCE 12

if K is in F1.5 then all its Casson–Gordon sliceness obstructions vanish (see Exercise ⃝ 31). We
will also see in Theorem 4.12 that T ⊆ 1 Fi .
T
i∈ 2 N

Question 4.2 (Open). Is T = 1 Fi ?

T
i∈ 2 N

On the other hand, as we mentioned in Section 1, we know that T =

̸ {1}. By Theorem 4.12
the solvable filtration cannot effectively distinguish between smooth concordance classes of
topologically slice knots. For this purpose, Cochran–Harvey–Horn [CHH13] defined the bipolar
filtration of T , denoted by {Tn }n∈N , satisfying
\
Ti ⊆ · · · ⊆ Tn+1 ⊆ Tn ⊆ · · · ⊆ T0 ⊆ T .
i∈N

We will define the bipolar filtration presently, and note here that it is also highly nontrivial (see
Theorem 4.15).
i∈N Ti = {1}?
T
Question 4.3 (Open). Is

4.2. Definitions of the filtrations. Both the solvable filtration of Cochran–Orr–Teichner and
the bipolar filtration of Cochran–Harvey–Horn are motivated by the following characterisation of
topologically slice knots. For a knot K ⊆ S 3 , the 3-manifold obtained by performing 0-framed
Dehn surgery on S 3 along K is denoted by S03 (K).
Proposition 4.4. Let K ⊆ S 3 be a knot. Then K is topologically slice if and only if S03 (K) = ∂W ,
where W is a compact, connected, oriented 4-manifold such that
(i) inclusion induces an isomorphism Z ∼ = H1 (S03 (K); Z) → H1 (W ; Z);
(ii) H2 (W ; Z) = 0; and
(iii) π1 (W ) is normally generated by the meridian µK ⊆ S03 (K).

Proof. Assume that K is topologically slice, with a topological slice disc ∆ ⊆ B 4 . As mentioned
before, by work of Quinn [Qui82, Theorem 2.5.1; FQ90, Theorem 9.3] we know that ∆ admits an
open tubular neighbourhood ν∆ ≈ ∆ × D̊2 . Then let W = B 4 ∖ ν∆. The boundary of W is a
union of the knot exterior S 3 ∖ νK and ∆ × S 1 , a solid torus. This shows that ∂W is the result
of some Dehn surgery on S 3 along K. In order to see that we indeed get S03 (K) as the boundary,
note that the disc ∆ × pt ⊆ ∆ × S 1 is attached to a pushoff of K which has zero linking number
with K, since in particular it bounds a disc in B 4 disjoint from ∆ × {0} ⊆ ν∆.
Then (iii) follows from the Seifert–van Kampen theorem since we obtain B 4 (which is simply
connected) from W by gluing in a thickened 2-cell (i.e. the neighbourhood ν∆) to W along
µK ⊆ ∂W .
To see (ii), consider the Mayer–Vietoris sequence for B 4 as the union W ∪ ν∆ = B 4 . Notice
that W ∩ ν∆ ∼ = D2 × S 1 . Then we have
H2 (D2 × S 1 ; Z) → H2 (W ; Z) ⊕ H2 (ν∆; Z) → H2 (B 4 ; Z),
so we see that H2 (W ; Z) = 0. From the same sequence, we also get that H 2 (W ; Z) ∼
= H 3 (W ; Z) =
∼
0. Therefore, H1 (W, ∂W ; Z) = H2 (W, ∂W ; Z) = 0. Apply these to the long exact sequence for
the pair (W, ∂W ), where we know that ∂W ∼ = S03 (K):
H2 (W, S03 (K); Z) → H1 (S03 (K); Z) → H1 (W ; Z) → H1 (W, S03 (K); Z)
This yields (i) and completes half of the proof.
For the converse direction, given W , we will glue in a 2-handle along µK ⊆ ∂W , with a framing
we describe presently, so that the result B is homeomorphic to B 4 , within which we will locate a
slice disc for the knot K ⊆ ∂B. To do so, we will need to choose a framing of µK with the goal
of recovering S 3 as ∂B. We do this explicitly next. Experts can safely skip this level of detail,
and instead work directly in a Dehn surgery diagram. We provide an explicit argument for the
benefit of newcomers to the field.
This construction will use multiple copies of the 2-disc D2 , which we indicate with subscripts,
i.e. each Di2 below indicates a copy of D2 . Use the Seifert longitude of K to identify a tubular
 SLICE KNOTS AND KNOT CONCORDANCE 13

neighbourhood νK ⊆ S 3 with K × D02 , so the Seifert longitude is given by K × x for some

x ∈ ∂D02 . By definition, we know that
˚2 )) ∪ 2 1 (D2 × S 1 ),
S03 (K) = (S 3 ∖ (K × D 0 ∂D ×S 1 1

with the surgery solid torus × D12 S1

attached to K × ∂D02 so that ∂D12 × y for some y ∈ S 1
is identified with K × x and z × S for z ∈ ∂D12 is identified with the meridian µK . Attach a
1

2-handle D12 × D22 to W along the core of the surgery solid torus, i.e. along 0 × S 1 , using the
surgery solid torus as the requisite framing of a tubular neighbourhood of the core, and call the
result B. In other words we have:
B := W ∪D2 ×S 1 D12 × D22 ,
1

Note that the core 0 × ⊆ S1

× D12 S1
is isotopic in S03 (K) to z × S 1 , which we saw is identified
with µK , so we are indeed attaching a 2-handle to W along µK as claimed.
Then ∂B is the result of surgery on S03 (K) along µ(K). More precisely, we have
∂B = S 3 (K) ∖ (D2 × S 1 ) ∪ 1 1 2


0 1 2 (S × D )
S ×∂D2 2
 
˚2 )
∪ 2 1 (D2 × S 1 ) ∖ (D2 × S 1 ) ∪ 1

3∖ 1 2
= S (K × D 0 ∂D ×S 1 1 S ×∂D2 (S × D2 )
1 2

= S 3 ∖ (K × D02 ) ∪S 1 ×∂D2 (S 1 × D22 ),

2

where we can check that w × ∂D22 for w ∈ S 1 is attached to the meridian µK . In other words,
we have performed the ∞-framed Dehn surgery on S 3 along K, producing ∂B = S 3 . In this copy
of S 3 , we still have the knot K, represented as ∂D12 × y, so in particular the disc D12 ⊆ B is a
slice disc for K.
Moreover, by our hypotheses, we also see that B is homotopy equivalent to B 4 . Then by the
topological 4-dimensional Poincaré conjecture [Fre82, Theorem 1.6], we know that B ≈ B 4 . This
completes the proof. □
We now give the definition of the solvable filtration. Recall that for a group G, the ith derived
subgroup, denoted G(i) , is inductively defined by setting G(0) = G and G(i+1) := [G(i) , G(i) ] for
each i ≥ 0.
Definition 4.5 ([COT03, Definition 1.2]). Let K ⊆ S 3 be a knot and n ∈ N. We say that K
is (smoothly) n-solvable if S03 (K) = ∂W where W is a smooth, compact, connected, oriented
4-manifold such that
(i) the inclusion induces an isomorphism Z = ∼ H1 (S 3 (K); Z) → H1 (W ; Z); and
0
(ii) H2 (W ; Z) has a basis consisting of smoothly embedded, closed, connected, oriented
surfaces {Li , Di }ki=1 , for some k, such that
(a) each Li and Di has trivial normal bundle, and
(b) the surfaces {Li , Di }i are pairwise disjoint, except that for each i, the surface Li
intersects Di transversely once with positive sign; and
(c) for each i, we have π1 (Li ) ⊆ π1 (W )(n) and π1 (Di ) ⊆ π1 (W )(n) , with respect to the
inclusion induced maps.
The set of n-solvable knots is denoted by Fn and the manifold W is called an n-solution.
We say that K is (smoothly) n.5-solvable if, in addition, for each i, we have π1 (Di ) ⊆
π1 (W )(n+1) with respect to the inclusion induced maps. The set of n.5-solvable knots is denoted
by Fn.5 and the manifold W is called an n.5-solution.
We leave it to the reader to show that, for each n ∈ 12 N, the set Fn is a subgroup of C diff
(Exercise △ 13). Note that condition (i) of Definition 4.5 and of Proposition 4.4 are the same.
Condition (ii) in Definition 4.5 can be seen as a generalisation of condition (ii) in Proposition 4.4
– roughly speaking, if we assume that each surface Li is a sphere, then one could do surgery on W
along {Li }, i.e. remove a tubular neighbourhood Li × D2 and glue in S 1 × D3 , for each i, and the
resulting 4-manifold with boundary S03 (K) would have trivial second homology. The condition
that π1 (Li ) ⊆ π1 (W )(n) then measures how far away the homology class of Li is from being
represented by an immersed sphere: in the case that π1 (Li ) is mapped to the trivial subgroup
 SLICE KNOTS AND KNOT CONCORDANCE 14

of π1 (W ), one could perform ambient surgery along half a symplectic basis of curves on Li to
transform it to an immersed sphere.
Remark 4.6. We could also add in Condition (iii) from Proposition 4.4 to Definition 4.5, without
any loss to known results about the solvable filtration. A priori (but only conjecturally) that
would lead to a different filtration.
In Definition 4.5, if the manifold W is not required to be smooth and the surfaces {Li , Di }i are
only required to be locally flat embedded, we say that K is topologically n-solvable or n.5-solvable.
top
Denote the set of such knots by Fntop and Fn.5 respectively. A priori the filtrations {Fntop }n∈ 1 N
2
and {Fn }n∈ 1 N appear to be distinct, but indeed, they coincide, as mentioned in [CHL09, p. 1423;
2
CHH13, p. 2105]. Since an explicit proof has not appeared in the literature before, we provide
one next.
Proposition 4.7. Let K ⊆ S 3 be a knot and n ∈ 12 N. Then K is smoothly n-solvable if and
only if it is topologically n-solvable.

Proof. By forgetting smoothness, we see that if K ∈ Fn then K ∈ Fntop .

Assume that K ∈ Fntop and let W denote a topological n-solution. First we note that the Kirby–
Siebenmann invariant ks(W ) = 0. To see this, observe that W is spin, since H1 (W ; Z) ∼ = Z has
no 2-torsion and the intersection form is even. The Kirby–Siebenmann invariant of a compact,
oriented, spin topological 4-manifold can be computed using the signature and the Rochlin
invariant of the boundary with respect to the induced spin structure [FQ90, Proposition 10.2B].
To avoid the Rochlin invariant and having to specify spin structures, we instead glue the 0-trace
of K, denoted by XK , to W , along the common boundary S03 (K); call the result Y . Since XK is
smooth, we know that ks(Y ) = ks(W ). Using the Mayer–Vietoris sequence,
∼
=
H1 (S03 (K); Z) H1 (XK ; Z) ⊕ H1 (W ; Z) H1 (Y ; Z) 0,

where we used that H1 (XK ; Z) = 0, we see that H1 (Y ; Z) = 0. Using the Mayer–Vietoris

sequence again, noting that H3 (Y ; Z) ∼ = H 1 (Y ; Z) = 0, and that the inclusion induced map
H2 (S0 (K); Z) → H2 (XK ; Z) is an isomorphism while H2 (S03 (K); Z) → H2 (W ; Z) is trivial, we can
3

further conclude that H2 (Y ; Z) = ∼ H2 (W ; Z) and indeed that the intersection forms on Y and on
W coincide. So we have that Y is spin, since it has even intersection form and trivial first homology.
By the definition of an n-solution, the signature σ(Y ) = σ(W ) = 0. For this we can either use that
the intersection form of Y agrees with that of W , or use Novikov additivity. Putting everything
together, we see that ks(W ) = ks(Y ) = 0 by [FQ90, Proposition 10.2B]. Therefore, the connected
sum of W with sufficiently many copies of S 2 × S 2 is smoothable [FQ90, Theorem 8.6] (see
also [FNOP19a, Theorem 8.6; PR21, Section 21.4.5]). Choose r ≥ 0 so that W ′ = W #rS 2 × S 2 is
smoothable. Choose a smooth structure on W ′ . We will show that W ′ is a (smooth) n-solution.
As required, the manifold W ′ is smooth, compact, connected and oriented. We also know
that inclusion induces an isomorphism H1 (W ; Z) ∼ = H1 (W ′ ; Z). So it remains only to check the
k
conditions on the second homology. Let {Li , Di }i=1 denote a basis for H2 (W ; Z) provided by the
definition of an n-solution. By requiring that the connected sum operation is performed away
from {Li , Di }i , we may assume that each Li and Di also lies in W ′ . Let {Ai , Bi }ri=1 denote the
standard basis for H2 (rS 2 × S 2 ; Z) given by the S 2 -factors, oriented so that each Ai intersects
the corresponding Bi with positive sign. Of course each Ai and Bi has trivial normal bundle.
We also know that π1 (W ′ ) ∼ = π1 (W ). Therefore, the set {Li , Di }ki=1 ∪ {Ai , Bi }ri=1 is a basis for
′
H2 (W ; Z) satisfying all the conditions of Definition 4.5, except that the elements of {Li , Di }i
may not be smoothly embedded in W ′ .
In general, it may not be possible to change {Li , Di }i by a smooth ambient isotopy so that the
result consists of smooth embeddings. However, by [FQ90, Theorem 8.1A] (see also [FQ90, p. 115]),
we may modify each Li and Di by ambient isotopies and finger moves until the resulting immersions
are smooth. The new intersections are paired by framed, embedded Whitney discs. Tube each
Li (or Di ) to itself, along one of each pair of Whitney arcs, as shown in Figure 10; call the
resulting surfaces {L′i , Di′ }i . Then by construction each L′i and Di′ is smoothly embedded in W ′
as needed. We also note that the intersections are as before, since the modification has occurred
 SLICE KNOTS AND KNOT CONCORDANCE 15

Li
µ

Li L′i
λ

Figure 10. Left: The result of performing a finger move on Li . The Whitney
disc is shaded in grey. Right: The surface L′i is produced after tubing along a
Whitney arc. The blue curves show the new generators of π1 (L′i ), denoted by µ
and λ, compared to π1 (Li ).

along arcs which can be assumed to have interiors disjoint from {Li , Di } and one another. Of
course, the genera of the surfaces have increased so we have to ensure that the new elements of
the fundamental group still map to π1 (W )(n) or π1 (W )(n+1) , as appropriate. We consider the
case of a surface L′i – the case of a Di′ is similar. As shown in Figure 10, for each finger move,
there are two new generators of the π1 (L′i ) compared to π1 (Li ), given by the curves µ and λ.
However, both µ and λ are null-homotopic in W , the latter due to the existence of the Whitney
disc and the former due to a meridional disc for L′i . Therefore, the images of π1 (Li ) and π1 (L′i )
in π1 (W ) under the inclusion induced maps coincide. This completes the proof that W ′ is a
smooth n-solution. □

Consider the natural map Ψ : Fn → Fntop , where for each knot K the smooth concordance
class of K is sent to the topological concordance class, i.e. induced by the map Ψ in (3.1). Then
Proposition 4.7 has the following straightforward corollary.
Corollary 4.8. For each n ∈ 1 N, the map Ψ induces an isomorphism Fn /Fn+0.5 ∼
2 = F top /F
n
top
.
n+0.5
We now state the results mentioned in Section 4.1, as well as a few more open questions.
Theorem 4.9 ([COT03, COT04, CT07, CHL09, CHL11a]). For each n ∈ N, there is a subgroup
of Fn /Fn.5 isomorphic to Z∞ ⊕ (Z/2)∞ .
For other related work, see also [Jia81, Liv99, DPR21]. By contrast, little is known about the
‘other half’ of the filtration.
Question 4.10 (Open). For some n ∈ N, is the group Fn.5 /Fn+1 nontrivial?
For example, it was shown in [DMOP19] that genus one knots which are 0.5-solvable (equival-
ently, algebraically slice) are also 1-solvable. We note that there is an analogue of the solvable
filtration for m-component (string) links, denoted by {Fnm } 1 , where for large enough m it is
n∈ 2 N
known that Fn.5 m /F m is nontrivial [Ott14].
n+1
Examples of knots lying in Fn for large n are constructed using satellite operations. We will
describe this further in Section 5. For now, we note that the examples from [COT03, COT04,
CHL09, CHL11a] all have genus one. The examples from [DPR21] are cables of genus one knots,
and have large Seifert genus and smooth slice genus, but may well have topological slice genus
one. In general it is difficult to bound the topological slice genus of knots lying deep in the
solvable filtration.
Question 4.11 (Open, Cha [Cha08, Remark 5.6]). For arbitrary n > 2 and g > 1, does there
exist a knot in Fn with topological slice genus at least g?
The n = 0 and 1 cases can be shown using Levine–Tristram and Casson–Gordon signatures
respectively, while the n = 2 case was shown by Cha–Miller–Powell in [CMP21].
 SLICE KNOTS AND KNOT CONCORDANCE 16

Membership in the various levels of the solvable filtration can be obstructed using von
Neumann ρ-invariants, defined by Cheeger and Gromov in [CG85]. For a detailed discussion, see
[CHL11b, Section 5; CT07, Section 2; COT03, Section 2].
Theorem 4.12. T Let K ⊆ S 3 be a topologically slice knot. Then K lies in Fn for every n. In
other words, T ⊆ 1 Fn .
n∈ 2 N

Proof. Since K is topologically slice, by Proposition 4.4 we can see that K ∈ Fntop for every n by
definition. Then by Proposition 4.7, we also know that K ∈ Fn for every n. □
The above theorem implies that the solvable filtration is not useful for studying the structure
of T . However, there is another filtration which can be used for this purpose.
Definition 4.13 ([CHH13, Definition 5.1]). Let K ⊆ S 3 be a knot and n ∈ N. We say that K is
n-positive if S03 (K) = ∂W where W is a smooth, compact, connected, oriented 4-manifold such
that
(i) the inclusion induces an isomorphism Z ∼ = H1 (S03 (K); Z) → H1 (W ; Z);
(ii) H2 (W ; Z) has a basis consisting of smoothly embedded, disjoint, closed, connected,
oriented surfaces {Si }ki=1 , for some k, such that
(a) each Si has normal bundle with Euler number +1; and
(b) for each i, we have π1 (Si ) ⊆ π1 (W )(n) with respect to the inclusion induced maps;
and
(iii) π1 (W ) is normally generated by the meridian µK ⊆ S03 (K).
The set of n-positive knots is denoted by Pn and the manifold W is called an n-positon.
We say that K is n-negative if each surface Si instead has normal bundle with Euler number
−1. The set of n-negative knots is denoted by Nn and the manifold W is called an n-negaton.
Let Bn := Pn ∩ Nn . Knots in Bn are said to be n-bipolar.
We leave it to the reader to show that, for each n ∈ N, the sets Pn and Nn are submonoids
of C diff , and Bn is a subgroup of C diff (Exercise △ 13). As for the solvable filtration, note that
condition (i) of Definition 4.13 and of Proposition 4.4 are the same. Indeed, now we further have
the same condition (iii). Condition (ii) in Definition 4.13 is again a generalisation of condition
(ii) in Proposition 4.4 – roughly speaking, if we assume that each surface Si is a sphere, then
one could perform a blow down operation on W along {Si }, i.e. remove a neighbourhood of Si
diffeomorphic to the D2 -bundle over S 2 with Euler number ±1, and then glue in B 4 , for each i,
and the resulting 4-manifold with boundary S03 (K) would have trivial second homology. The
condition that π1 (Si ) ⊆ π1 (W )(n) as before measures how far away the homology class of Si is
from being represented by an immersed sphere: in the case that π1 (Si ) is mapped to the trivial
subgroup of π1 (W ), one could perform ambient surgery along half a symplectic basis of curves
on Si to transform it to an immersed sphere.
Definition 4.14. For each n ∈ N define Tn := T ∩ Bn . The corresponding filtration {Tn }n∈N is
called the bipolar filtration of T .
As desired, the bipolar filtration is highly nontrivial on T , as seen in the theorem below.
Theorem 4.15 ([CK21], see also [CHH13, CH15]). For each n, there is a subgroup of Tn /Tn+1
isomorphic to Z∞ .
As before, the examples of knots lying in Tn for large n used in the theorem above are
constructed via a generalisation of the satellite operation, which we will describe in Section 5. As
in the case of the solvable filtration. membership in the various levels of the positive, negative, and
bipolar filtrations of C diff can be obstructed using the von Neumann ρ-invariants. However, these
are not sufficient to prove nontriviality of the bipolar filtration of T . The proof of Theorem 4.15
combines von Neumann ρ-invariants with the Heegaard–Floer d-invariant. For more details,
see [CHH13, CH15, CK21].
There are several open questions about the bipolar filtration, such as the following.
Question 4.16 (Open). Is there a (short) list of invariants characterising knots lying in B0 ?
Note that there are explicit characterisations of knots in F0 and F0.5 (Exercise ⃝ 31).
 SLICE KNOTS AND KNOT CONCORDANCE 17

Question 4.17 (Open). For n ∈ N, is there a subgroup of Tn /Tn+1 isomorphic to (Z/2)∞ ? (See
Theorems 4.9 and 4.15.)
Question 4.18 (Open). For arbitrary n ≥ 0 and g > 1, does there exist a knot in Tn with
smooth slice genus at least g? (See Question 4.11.)
Unlike Question 4.11, the above appears to be open even in the case n = 0.

5. Satellite operations on knots

In this final section, we discuss satellite operations on knots. Recall that Whitehead doubling
is a special case of the satellite operation, and satellite operations can be used to construct
examples of knots deep in the solvable and bipolar filtrations.
Definition 5.1. A pattern P is a knot in the solid torus S 1 × D2 . Given a knot K, called the
companion, the satellite knot P (K) is obtained by tying the solid torus into the knot K, in an
appropriately untwisted manner, as shown in Figure 8. More precisely, we map S 1 × D2 to a
tubular neighbourhood νK of K, so that S 1 × x for x ∈ ∂D2 is mapped to the Seifert longitude
of K. The image of P under this map is the satellite knot P (K) by definition. Note that the
knot P (K) inherits an orientation from P and K.
The algebraic intersection number of P with a generic meridional disc of S 1 × D2 is called the
winding number of P . For example, the winding number of the Whitehead doubling patterns
Wh± is zero.
The satellite operation is compatible with concordance, and yields well defined functions on
the concordance groups, called satellite operators (see Exercise △ 15).
We next give an alternative view of the satellite construction. In this case, we begin with a
knot R ⊆ S 3 and a curve η ⊆ S 3 ∖ R so that η is unknotted when considered in S 3 ⊇ S 3 ∖ R.
Then note that the complement in S 3 of a tubular neighbourhood νη of η can be canonically
identified with the solid torus, so R ⊆ S 3 ∖ νη is a pattern. So, given any arbitrary knot K, we
could construct the satellite knot with respect to this pattern. In this case, we denote the result
by Rη (K), and call it the result of infection on R by K along η. This procedure is depicted
in Figure 11. There is a generalisation of the infection procedure where as input we can take
not just knots but arbitrary string links. For more on this procedure, see [CO94, Section 1;
Coc04, p. 385; CFT09, Section 2.2].

R K Rη (K)

Figure 11. On the left, we have a ribbon knot R with an unknot η ⊆ S 3 ∖ R.

Given a knot K, we show on the right the result of infection on R by K along η.

As previously indicated, the infection procedure can be used to construct knots in the various
filtrations from the previous section. The key tool is the following proposition.
Proposition 5.2 ([COT04, Proposition 3.1; CHL09, Theorem 7.1; CHL11b, Proposition 2.7;
CHH13, Proposition 3.3]). Let R ⊆ S 3 be a knot and η ⊆ S 3 ∖ R be a curve which is unknotted
when considered in S 3 ⊇ S 3 ∖ R. Let K ⊆ S 3 be an arbitrary knot, and let n, i ≥ 0 be arbitrary.
(1) if R, K ∈ Fn and η ∈ π1 (S 3 ∖ R)(i) , then Rη (K) ∈ Fn+i .
(2) if R, K ∈ Pn and η ∈ π1 (S 3 ∖ R)(i) , then Rη (K) ∈ Pn+i .
(3) if R, K ∈ Nn and η ∈ π1 (S 3 ∖ R)(i) , then Rη (K) ∈ Nn+i .
(4) if R, K ∈ Bn and η ∈ π1 (S 3 ∖ R)(i) , then Rη (K) ∈ Bn+i .
 SLICE KNOTS AND KNOT CONCORDANCE 18

(5) if R, K ∈ Tn and η ∈ π1 (S 3 ∖ R)(i) , then Rη (K) ∈ Tn+i .

We leave the proof for the ambitious reader (Exercise □ 27). Armed with this proposition,
we are now able to construct knots deep in any of the solvable, positive, negative, or bipolar
filtrations, by the following strategy. Let us consider the solvable filtration for concreteness and
attempt to build a knot in Fn for some given n ∈ N. Start with any knot K with Arf(K) = 0,
e.g. the connected sum of two copies of the right-handed trefoil knot. Recall that we know that
K ∈ TF0 by Exercise ⃝ 31. Then choose a ribbon knot R. By Theorem 4.12, we know that
R∈ 3∖ R such that ℓk(R, η) = 0. By definition of the linking
1 Fi . Find an unknotted η ⊆ S
i∈ 2 N
number we know that η ∈ π1 (S 3∖ R)(1) . So then Rη (K) ∈ F1 by Proposition 5.2. Then we iterate.
In other words, to find a knot in Fn , we need the n-fold iterated satellite Rη (Rη (· · · (K) · · · )).
Note that these iterated satellites have bounded Seifert genus (Exercise □ 28) and we only needed
the k = 1 case of the proposition.
A priori it might seem that satellites are quite special. However, by [CFT09, Proposition 1.7],
any algebraically slice knot is concordant to a knot of the form Rη (J), where J is a string link
with pairwise linking numbers zero, the knot R is ribbon, and η is now an unlink, where each
component has linking number zero with R. Therefore, the construction explained above is
rather general.
We finish this section by surveying a few results and open questions about the effect of satellite
operators on the knot concordance groups. We already saw questions about the Whitehead
doubling satellite operator in Section 1. E.g. Conjecture 1.16 can be rephrased as asking whether
the Whitehead doubling satellite operator is injective.
Question 5.3 (Open). Is there a pattern P with winding number ̸= ±1 such that P (K) is
slice if and only if K is slice, in either category? More generally, such that P (K) and P (J) are
concordant if and only if K and J are concordant?
Examples of winding number ±1 satellite operators which are injective on C top were given
in [CDR14]. From another perspective, one could ask when satellite operators are surjective on the
concordance groups. For winding numbers ̸= ±1 it is not difficult to find nonsurjective satellite
operators using algebraic invariants [DR16, Proposition 3.1]. The first example of a nonsurjective
satellite operator on C diff with winding number ±1 was given by A. Levine in [Lev16]. Other
work has considered the invertibility of satellite operators, such as [DR16, MP18].
The existence of injective nonsurjective satellite operators is evidence towards the following
conjecture. Further evidence was provided in [CHL11b] by showing that a large class of winding
number zero satellite operators, called robust doubling operators, are injective on large subgroups
of C diff .
Conjecture 5.4 (Open, Cochran–Harvey–Leidy [CHL11b]). The knot concordance groups have
the structure of a fractal.
We say a set has a fractal structure if it admits self-similarities at arbitrarily small scales,
following [BGN03, Definition 3.1]. The results mentioned above show that infinite classes of
winding number ±1 satellite operators, as well as the robust doubling operators, are candidate
self-similarities for C top and C diff . In order to understand the notion of “scale” one needs a suitable
metric on the knot concordance groups. This has been studied in [CH18, CHP17, CHPR23]. One
might also consider the behaviour of satellite operators under iteration, such as in [Ray15, Che20].
Since knots under concordance form a group and satellite operators are maps on the concordance
groups, it is tempting to assume that they are homomorphisms. However, this is far from the
case. In [Gom86], Gompf showed that the Whitehead doubling operator is not a homomorphism
on C diff , despite being the zero map on C top . Modern tools from Heegaard–Floer homology can
be used to obstruct satellite operators on C diff from being homomorphisms [Lev16, Hed09]. In
the topological category, obstructions can be obtained from Casson–Gordon invariants [Mil23].
Most generally, we have the following conjecture. (For a recent, partial resolution, see [JKM23].)
Conjecture 5.5 (Open, Hedden [Con16, Wor16]). Let P ⊆ S 1 × D2 be a pattern. If the induced
satellite operator P : C diff → C diff is a homomorphism, then it agrees with either the zero map,
the identity map, or the reversal map (taking each knot to the concordance class of its reverse).
 SLICE KNOTS AND KNOT CONCORDANCE 19

Finally, there is substantial interest in understanding whether satellite operators preserve

linear independence. We highlight two conjectures of Hedden and Pinzón-Caicedo.
Conjecture 5.6 (Open, Hedden–Pinzón-Caicedo [HPC21, Conjecture 2]). The image of every
nonconstant satellite operator on C diff generates an infinite rank subgroup of C diff .
The winding number ̸= 0 case is not difficult to prove using algebraic invariants. [HPC21]
provides a criterion guaranteeing that winding number zero satellite operators have infinite rank
images, generalising [HK12, PC17].
Conjecture 5.7 (Open, Hedden–Pinzón-Caicedo [HPC21, Conjecture 3]). For any nonconstant
winding number zero operator P , there exists a knot K for which the set {P (nK) | n ∈ Z} has
infinite rank in C diff .
The above was partially confirmed in [DHMS22] using tools from Heegaard–Floer homology.

Epilogue
There are many aspects of the field of slice knots and knot concordance that we were unable to
cover in these few lectures. So we end these notes with a brief, but surely incomplete, overview
of those topics.
First, readers interested in more information on classical knot concordance should look at the
excellent survey by Livingston [Liv05]. In particular, one will find there a much more detailed
description of the algebraic concordance group and the Casson–Gordon invariants. Many recent
developments in knot concordance, especially in the smooth setting, use Heegaard–Floer homology.
Readers interested in those techniques should look at the excellent survey of Hom [Hom17].
Obstructions to sliceness can also be obtained from Khovanov homology [Ras10].
Quite a bit of recent research has considered knots in 3-manifolds other than S 3 and/or sliceness
in 4-manifolds other than B 4 . A notable highlight is the proposed strategy of Freedman–Gompf–
Morrison–Walker [FGMW10] to attack the 4-dimensional Poincaré conjecture: find a knot K and
a homotopy 4-ball B with ∂B = S 3 such that K is smoothly slice in B but not in B 4 . This would
imply that B is not diffeomorphic to B 4 , disproving the smooth 4-dimensional Poincaré conjecture.
It is worth noting that the following weaker question is also open: does there exist a knot K ⊆ S 3
which is slice in an integer homology 4-ball but not in B 4 ? However there do exist rationally
slice knots, such as the figure eight knot, i.e. knots that are slice in a rational homology ball, but
not necessarily slice in B 4 . For more on this topic see e.g. [Kaw09, Cha07, HKPS22, HKP23].
More recent work of Manolescu and Piccirillo [MP21] explains how one can use sliceness of
knots and links to address questions about exotic smooth structures on closed 4-manifolds other
than S 4 . Other work concerning slicing knots in general 4-manifolds includes [Nor69, Yas91,
Yas92, CT14, CN20, Rao20, MMP20, Pic20, KR21, KMRS21, MMRS22, MMSW23, HR23]. When
studying concordance in more general 3-manifolds, one could also quotient out by the action of
connected sum by knots in S 3 . This leads to the notion of almost concordance, previously called
piecewise linear I-equivalence [Rol85; Hil12, Section 1.5; Cel18; FNOP19b; Yil18; NOPP19].
In light of Proposition 4.4 it is natural to ask to what extent the homeomorphism or homology
cobordism class of the 0-surgery determines the concordance class of a knot. This is roughly
the Akbulut–Kirby conjecture, For more on this, see e.g. [Kir97, Problem 1.19; KL99; CFHH13;
Yas15; MP18].
We introduced ribbon discs in Section 1. There is a relative version called ribbon concordance
in the smooth setting and homotopy ribbon concordance in the topological setting (see also
Exercise □ 21). These are not necessarily symmetric relations. A flurry of recent work in the
smooth setting [Zem19, LZ19, MZ21, DLVVW22, Kan22] culminated in Agol showing that ribbon
concordance is a partial order on knots [Ago22], confirming a conjecture of Gordon [Gor81]. It is
still open whether homotopy ribbon concordance is a partial order as well [FP20, FKL+ 22].
We are also interested in investigating sliceness and concordance within standard families of
knots. For example, we saw in Section 1 that the slice-ribbon conjecture holds for certain families.
The families of algebraic knots [Rud76, Lit84, HKL12, CKP23], 2-bridge knots [CG86, Lis07, Mil18,
FM16], pretzel knots [GJ11, Lec15, Lec12, Bry17, Lon14, Mil17a, Mil17b, KST20, KLS22], and
(strongly) quasipositive knots [Rud93, Hay19, BF19] have received particular attention.
 SLICE KNOTS AND KNOT CONCORDANCE 20

One may also study specific types of slice discs, e.g. those that have certain symmetries. There
has been a renewed interest recently in equivariant sliceness and concordance [Nai97, CK99,
DN06, BI22, DMS22, BC22, Mal22, MP23, DP23, DF23, DiP23].
Finally so far we have only discussed the existence of slice discs. One could equally well study
the uniqueness question. In other words, can one quantify the number of slice discs for a given
slice knot? Recent work in this area includes [CP21, SS22, JZ20, MP19, Akb22, HS21, LS22].
 SLICE KNOTS AND KNOT CONCORDANCE 21

Exercises
Introductory problems.
Exercise △ 1. Let K ⊆ S 3 be a knot, with tubular neighbourhood νK.
(1) Show that S 3 ∖ νK is a homology circle, i.e. H∗ (S 3 ∖ νK; Z) ∼
= H∗ (S 1 ; Z).
3
(2) Show that π1 (S ∖ νK) is normally generated by an arbitrary meridian of K, i.e. it is
generated by the set of conjugates of the meridian.
Do the above properties generalise to higher-dimensional knots S n ,→ S n+2 ? How about knots
with arbitrary codimension?
Exercise △ 2. Show that the connected sum operation is commutative and associative. In
other words, given knots J, K, L ⊆ S 3 , show that J#K is isotopic to K#J and J#(K#L) is
isotopic to (J#K)#L.
Exercise △ 3. Prove Proposition 1.2: Let K ⊆ S 3 be a knot. Show that K is smoothly slice if
and only if rK is smoothly slice if and only if K is smoothly slice.
Note that the analogous statement also holds for topological sliceness.
Exercise △ 4. Prove Proposition 1.3: If the knots K, J ⊆ S 3 are smoothly slice, then so is
K#J.
As before, the analogous statement also holds for topological sliceness.

(a) (b) (c)

Figure 12. Some ribbon knots

Exercise △ 5. Show that the knots in Figure 12 are ribbon.

Exercise △ 6. Prove Proposition 1.6: For any knot K ⊆ S 3 , the knot K#rK is ribbon.
Exercise △ 7. Prove Proposition 1.12: Let K ⊆ S 3 be a knot. The coned disc cone(K) ⊆
cone(S 3 ) = B 4 is locally flat if and only if K is the trivial knot.
Hint: Use the fact (without proof) from classical knot theory that π1 (S 3 ∖ K) ∼= Z if and only
if K is the trivial knot.
Exercise △ 8. Let K ⊆ S 3 be a knot. Prove that if K is smoothly or topologically slice, then
the same holds for the untwisted Whitehead doubles Wh± (K).
Exercise △ 9. Prove Proposition 3.2: A knot K ⊆ S 3 is smoothly (resp. topologically) slice if
and only if it is smoothly (resp. topologically) concordant to the unknot.
Exercise △ 10. Prove Proposition 3.4: Smooth (resp. topological) concordance is an equivalence
relation on knots.
Exercise △ 11. Prove Proposition 3.6: In either C diff or C top , the inverse of [K] is the class of
−K := rK. The identity element in C diff (resp. C top ) is the class of the unknot, equivalently the
class of any smoothly (resp. topologically) slice knot.
Exercise △ 12. Prove Proposition 3.7: A knot K ⊆ S 3 is said to be negative amphichiral if it
is isotopic to rK. Every negative amphichiral knot is order two in C diff and C top . As a result,
the figure eight knot is order two in C diff and C top .
Exercise △ 13.
(1) For each n ∈ 12 N, show that the set Fn is a subgroup of C diff .
(2) For each n ∈ N, show that the sets Pn and Nn are submonoids of C diff , and Bn is a
subgroup of C diff .
 SLICE KNOTS AND KNOT CONCORDANCE 22

Exercise △ 14. Let K ⊆ S 3 be a smoothly slice knot. Show that K lies in Tn for all n.
Exercise △ 15. Fix a pattern P ⊆ S 1 × D2 . Show that the satellite operation is well defined
on concordance on either category, i.e. there are well defined functions P : C diff → C diff and
P : C top → C top .
Exercise △ 16. Let P ⊆ S 1 × D2 be a pattern so that the induced satellite operator P : C diff →
C diff (resp. C top → C top ) is a homomorphism. Show that P (U ) is smoothly (resp. topologically)
slice, where U denotes the unknot.

Moderate problems.
∼ Z if and only if K is the
Exercise □ 17. Prove the hint from Exercise △ 7, that π1 (S 3 ∖ K) =
trivial knot.
Hint: Use Dehn’s lemma.
Exercise □ 18. Learn enough classical knot theory to show that the right-handed trefoil,
left-handed trefoil, and the figure eight from Figure 1 are nontrivial and distinct knots.
Hint: This will likely involve learning about some classical knot invariants, such as 3-
colourability, the knot group, the Seifert genus, the signature, the Alexander polynomial, . . . .
Exercise □ 19. Fix n ≥ 4. Prove that every smooth S 1 ,→ S n bounds a smoothly embedded
disc in S n .
Hint: The case n = 4 is the most challenging. Think about the types of singularities that arise
in the generic case and try to get rid of them.
Exercise □ 20. Prove Proposition 1.5: A knot K ⊆ S 3 is ribbon if and only if it bounds a disc
in S 3 with only ribbon singularities, i.e. singularities of the form shown in Figure 6.
Exercise □ 21. A knot K ⊆ S 3 is homotopy ribbon if there exists a topologically locally flat
disc D ⊆ B 4 bounded by K such that the inclusion induced map π1 (S 3 \ νK) → π1 (B 4 \ νD) is
surjective. Prove that ribbon implies homotopy ribbon.
Hint: Consider handle decompositions for ribbon disc complements.
Exercise □ 22. Prove Proposition 1.8: There exist smooth slice discs that are not ambiently
isotopic (relative to the boundary) to any ribbon disc.
Hint: Begin with the standard smooth slice disc for the unknot. Use the fact that there exist
2-knots in S 4 with nonabelian fundamental group of the complement (the intrepid reader could
try to prove the latter claim). Use Exercise □ 21. For a further challenge, construct examples of
such discs bounded by nontrivial knots.
Exercise □ 23. Prove the trace embedding lemma (Lemma 2.4): Let K ⊆ S 3 be a knot. The
0-trace X0 (K) admits a smooth (resp. locally collared) embedding into R4 if and only if K is
smoothly (resp. topologically) slice.
Wonder whether a similar argument would apply to the n-traces of the knots, denoted by
Xn (K), obtained by attaching an n-framed 2-handle to B 3 along a knot K ⊆ S 3 = ∂B 4 , for an
arbitrary n ∈ Z.
Hint: It is easier to prove the version of the trace embedding lemma in S 4 rather than R4 .
Avoid using the smooth 4-dimensional Schoenflies conjecture (which is still open!) in the S 4
version by noting that the closure of the complement of a smoothly embedded 4-ball in S 4 is
itself diffeomorphic to a 4-ball, by Palais’s disc theorem.
Exercise □ 24. Prove Proposition 3.3: Two knots K, J ⊆ S 3 are smoothly (resp. topologically)
concordant if and only if K#rJ is smoothly (resp. topologically) slice.
Hint: It is possible to isotope a concordance so that it contains a straight arc, i.e. one of the
form pt × [0, 1] ⊆ S 3 × [0, 1] (why?). Remove an open tubular neighbourhood of this arc. For
the other direction, build a concordance between K#rJ#J and J, and use Exercises 6 and 9.
Exercise □ 25. Show that isotopy classes of knots do not form a group under connected sum.
They do however form a monoid.
Exercise □ 26. Reprove Proposition 4.4 using the trace embedding lemma (Lemma 2.4).
 SLICE KNOTS AND KNOT CONCORDANCE 23

Exercise □ 27. Prove Proposition 5.2: Let R ⊆ S 3 be a knot and η ⊆ S 3 ∖ R be a curve which
is unknotted when considered in S 3 ⊇ S 3 ∖ R. Let K ⊆ S 3 be an arbitrary knot, and let n, i ≥ 0
be arbitrary
(1) if R, K ∈ Fn and η ∈ π1 (S 3 ∖ R)(i) , then Rη (K) ∈ Fn+i .
(2) if R, K ∈ Pn and η ∈ π1 (S 3 ∖ R)(i) , then Rη (K) ∈ Pn+i .
(3) if R, K ∈ Nn and η ∈ π1 (S 3 ∖ R)(i) , then Rη (K) ∈ Nn+i .
(4) if R, K ∈ Bn and η ∈ π1 (S 3 ∖ R)(i) , then Rη (K) ∈ Bn+i .
(5) if R, K ∈ Tn and η ∈ π1 (S 3 ∖ R)(i) , then Rη (K) ∈ Tn+i .
Exercise □ 28. Let R ⊆ S 3 be an arbitrary knot and η ⊆ S 3 ∖ R be a curve which is unknotted
when considered in S 3 ⊇ S 3 ∖ R. Assume that ℓk(R, η) = 0. Show that the Seifert genus of the
knots produced by infection on R along η is bounded. In other words, there exists g, such that
for any knot K, the Seifert genus of Rη (K) is at most g. Come up with a candidate g.
Challenge problems.
Exercise ⃝ 29. See Exercise □ 17. Are there any restrictions on the fundamental group of a
slice/ribbon knot? In other words, is every knot group the knot group of a slice/ribbon knot?
Hint: Consider the Alexander module.
Exercise ⃝ 30. Prove that every smooth 2-knot is smoothly slice, i.e. every smooth S 2 ,→ S 4
bounds a smoothly embedded B 3 in B 5 . Prove that every topological 2-knot is topologically
slice. This was originally proven by Kervaire [Ker65, Théorème III.6; Ker71, Theorem 1].
Hint: Use the fact that every 3-manifold is obtained by even-framed surgery on some link in
3
S .
Exercise ⃝ 31. Let K ⊆ S 3 be a knot.
(1) Show that K ∈ F0 if and only if Arf(K) = 0.
(2) Show that K ∈ F0.5 if and only if K is algebraically slice.
(3) Show that if K ∈ F1.5 then all its Casson–Gordon sliceness obstructions vanish.
 SLICE KNOTS AND KNOT CONCORDANCE 24

References
+
[ACM 23] P. Aceto, N. A. Castro, M. Miller, J. Park, and A. Stipsicz, Slice obstructions from genus bounds in
definite 4-manifolds, arXiv e-prints (Mar. 2023), arXiv:2303.10587, available at 2303.10587.
[Ago22] I. Agol, Ribbon concordance of knots is a partial ordering, Comm. Amer. Math. Soc. 2 (2022),
374–379.
[Akb22] S. Akbulut, Corks and exotic ribbons in B 4 , Eur. J. Math. 8 (2022), S494–S498.
[AKPR21] P. Aceto, M. H. Kim, J. Park, and A. Ray, Pretzel links, mutation, and the slice-ribbon conjecture,
Math. Res. Lett. 28 (2021), no. 4, 945–966.
[AT16] T. Abe and K. Tagami, Fibered knots with the same 0-surgery and the slice-ribbon conjecture, Math.
Res. Lett. 23 (2016), no. 2, 303–323.
[BC22] K. Boyle and W. Chen, Equivariant topological slice disks and negative amphichiral knots, arXiv
e-prints (July 2022), arXiv:2207.12593, available at 2207.12593.
[BF19] M. Borodzik and P. Feller, Up to topological concordance, links are strongly quasipositive, J. Math.
Pures Appl. (9) 132 (2019), 273–279.
[BG96] Ž. Bižaca and R. E. Gompf, Elliptic surfaces and some simple exotic R4 ’s, J. Differential Geom. 43
(1996), no. 3, 458–504.
[BGN03] L. Bartholdi, R. Grigorchuk, and V. Nekrashevych, From fractal groups to fractal sets, Fractals in
Graz 2001, 2003, pp. 25–118.
[BI22] K. Boyle and A. Issa, Equivariant 4-genera of strongly invertible and periodic knots, J. Topol. 15
(2022), no. 3, 1635–1674.
[BKK+ 21] S. Behrens, B. Kalmár, M. H. Kim, M. Powell, and A. Ray (eds.), The disc embedding theorem,
Oxford University Press, Oxford, 2021.
[Bry17] K. Bryant, Slice implies mutant ribbon for odd 5-stranded pretzel knots, Algebr. Geom. Topol. 17
(2017), no. 6, 3621–3664.
[Cas86] A. J. Casson, Three lectures on new-infinite constructions in 4-dimensional manifolds, À la recherche
de la topologie perdue, 1986, pp. 201–244. With an appendix by L. Siebenmann.
[CDR14] T. D. Cochran, C. W. Davis, and A. Ray, Injectivity of satellite operators in knot concordance, J.
Topol. 7 (2014), no. 4, 948–964.
[Cel18] D. Celoria, On concordances in 3-manifolds, J. Topol. 11 (2018), no. 1, 180–200.
[CFHH13] T. D. Cochran, B. D. Franklin, M. Hedden, and P. D. Horn, Knot concordance and homology cobordism,
Proc. Amer. Math. Soc. 141 (2013), no. 6, 2193–2208.
[CFT09] T. Cochran, S. Friedl, and P. Teichner, New constructions of slice links, Comment. Math. Helv. 84
(2009), no. 3, 617–638.
[CG78] A. J. Casson and C. McA. Gordon, On slice knots in dimension three., Algebraic and geometric
topology (Proc. Sympos. Pure Math., Stanford Univ., Stanford, Calif., 1976), Part 2, 1978, pp. 39–53,.
[CG85] J. Cheeger and M. Gromov, Bounds on the von Neumann dimension of L2 -cohomology and the
Gauss-Bonnet theorem for open manifolds, J. Differential Geom. 21 (1985), no. 1, 1–34.
[CG86] A. J. Casson and C. McA. Gordon, Cobordism of classical knots, À la recherche de la topologie
perdue, 1986, pp. 181–199. With an appendix by P. M. Gilmer.
[CG88] T. D. Cochran and R. E. Gompf, Applications of Donaldson’s theorems to classical knot concordance,
homology 3-spheres and property P , Topology 27 (1988), no. 4, 495–512.
[CH15] T. D. Cochran and P. D. Horn, Structure in the bipolar filtration of topologically slice knots, Algebr.
Geom. Topol. 15 (2015), no. 1, 415–428.
[CH18] T. Cochran and S. Harvey, The geometry of the knot concordance space, Algebr. Geom. Topol. 18
(2018), no. 5, 2509–2540.
[Cha07] J. C. Cha, The structure of the rational concordance group of knots, Mem. Amer. Math. Soc. 189
(2007), no. 885, x+95.
[Cha08] J. C. Cha, Topological minimal genus and L2 -signatures, Algebr. Geom. Topol. 8 (2008), no. 2,
885–909.
[Che20] W. Chen, An infinite-rank summand from iterated Mazur pattern satellite knots, arXiv e-prints (Oct.
2020), arXiv:2010.11277, available at 2010.11277.
[CHH13] T. D. Cochran, S. Harvey, and P. Horn, Filtering smooth concordance classes of topologically slice
knots, Geom. Topol. 17 (2013), no. 4, 2103–2162.
[CHL09] T. D. Cochran, S. Harvey, and C. Leidy, Knot concordance and higher-order Blanchfield duality,
Geom. Topol. 13 (2009), no. 3, 1419–1482.
[CHL11a] T. D. Cochran, S. Harvey, and C. Leidy, 2-torsion in the n-solvable filtration of the knot concordance
group, Proc. Lond. Math. Soc. (3) 102 (2011), no. 2, 257–290.
[CHL11b] T. D. Cochran, S. Harvey, and C. Leidy, Primary decomposition and the fractal nature of knot
concordance, Math. Ann. 351 (2011), no. 2, 443–508.
[CHP17] T. D. Cochran, S. Harvey, and M. Powell, Grope metrics on the knot concordance set, J. Topol. 10
(2017), no. 3, 669–699.
[CHPR23] T. D. Cochran, S. L. Harvey, M. Powell, and A. Ray, Tower metrics on the smooth concordance set
of knots, 2023+. In preparation.
 SLICE KNOTS AND KNOT CONCORDANCE 25

[CK21] J. C. Cha and M. H. Kim, The bipolar filtration of topologically slice knots, Adv. Math. 388 (2021),
Paper No. 107868, 32.
[CK99] J. C. Cha and K. H. Ko, On equivariant slice knots, Proc. Amer. Math. Soc. 127 (1999), no. 7,
2175–2182.
[CKP23] A. Conway, M. H. Kim, and W. Politarczyk, Nonslice linear combinations of iterated torus knots,
Algebr. Geom. Topol. 23 (2023), no. 2, 765–802.
[CL86] T. D. Cochran and W. B. R. Lickorish, Unknotting information from 4-manifolds, Trans. Amer.
Math. Soc. 297 (1986), no. 1, 125–142.
[CMP21] J. C. Cha, A. N. Miller, and M. Powell, Two-solvable and two-bipolar knots with large four-genera,
Math. Res. Lett. 28 (2021), no. 2, 331–382.
[CN20] A. Conway and M. Nagel, Stably slice disks of links, J. Topol. 13 (2020), no. 3, 1261–1301.
[CO94] T. D. Cochran and K. E. Orr, Homology boundary links and Blanchfield forms: concordance classific-
ation and new tangle-theoretic constructions, Topology 33 (1994), no. 3, 397–427.
[Coc04] T. D. Cochran, Noncommutative knot theory, Algebr. Geom. Topol. 4 (2004), 347–398.
[Con16] Conference organisers, Problem list, Conference on 4-manifolds and knot concordance, Max Planck
Institute for Mathematics, 2016.
[COT03] T. D. Cochran, K. E. Orr, and P. Teichner, Knot concordance, Whitney towers and L2 -signatures,
Ann. of Math. (2) 157 (2003), no. 2, 433–519.
[COT04] T. D. Cochran, K. E. Orr, and P. Teichner, Structure in the classical knot concordance group,
Comment. Math. Helv. 79 (2004), no. 1, 105–123.
[CP21] A. Conway and M. Powell, Characterisation of homotopy ribbon discs, Adv. Math. 391 (2021), Paper
No. 107960, 29.
[CT07] T. D. Cochran and P. Teichner, Knot concordance and von Neumann ρ-invariants, Duke Math. J.
137 (2007), no. 2, 337–379.
[CT14] T. D. Cochran and E. Tweedy, Positive links, Algebr. Geom. Topol. 14 (2014), no. 4, 2259–2298.
[DF23] A. Di Prisa and G. Framba, A new invariant of equivariant concordance and results on 2-bridge
knots, arXiv e-prints (Mar. 2023), arXiv:2303.08794, available at 2303.08794.
[DHMS22] I. Dai, M. Hedden, A. Mallick, and M. Stoffregen, Rank-expanding satellites, Whitehead doubles, and
Heegaard Floer homology, arXiv e-prints (Sep. 2022), arXiv:2209.07512, available at 2209.07512.
[DHST21] I. Dai, J. Hom, M. Stoffregen, and L. Truong, More concordance homomorphisms from knot Floer
homology, Geom. Topol. 25 (2021), no. 1, 275–338.
[DiP23] A. Di Prisa, Equivariant algebraic concordance of strongly invertible knots, arXiv e-prints (Mar. 2023),
arXiv:2303.11895, available at 2303.11895.
[DKM+ 22] I. Dai, S. Kang, A. Mallick, J. Park, and M. Stoffregen, The (2, 1)-cable of the figure-eight knot is not
smoothly slice, arXiv e-prints (July 2022), arXiv:2207.14187, available at 2207.14187.
[DLVVW22] A. Daemi, T. Lidman, D. S. Vela-Vick, and C.-M. M. Wong, Ribbon homology cobordisms, Adv. Math.
408 (2022), Paper No. 108580, 68.
[DMF92] S. De Michelis and M. H. Freedman, Uncountably many exotic R4 ’s in standard 4-space, J. Differential
Geom. 35 (1992), no. 1, 219–254.
[DMOP19] C. W. Davis, T. Martin, C. Otto, and J. Park, Every genus one algebraically slice knot is 1-solvable,
Trans. Amer. Math. Soc. 372 (2019), no. 5, 3063–3082.
[DMS22] I. Dai, A. Mallick, and M. Stoffregen, Equivariant knots and knot Floer homology, arXiv e-prints
(Jan. 2022), arXiv:2201.01875, available at 2201.01875.
[DN06] J. F. Davis and S. Naik, Alexander polynomials of equivariant slice and ribbon knots in S 3 , Trans.
Amer. Math. Soc. 358 (2006), no. 7, 2949–2964.
[Don83] S. K. Donaldson, An application of gauge theory to four-dimensional topology, J. Differential Geom.
18 (1983), no. 2, 279–315.
[Don86] S. K. Donaldson, Connections, cohomology and the intersection forms of 4-manifolds, J. Differential
Geom. 24 (1986), no. 3, 275–341.
[Don87] S. K. Donaldson, The orientation of Yang-Mills moduli spaces and 4-manifold topology, J. Differential
Geom. 26 (1987), no. 3, 397–428.
[DP23] A. Di Prisa, The equivariant concordance group is not abelian, Bull. Lond. Math. Soc. 55 (2023),
no. 1, 502–507.
[DPR21] C. W. Davis, J. Park, and A. Ray, Linear independence of cables in the knot concordance group,
Trans. Amer. Math. Soc. 374 (2021), no. 6, 4449–4479.
[DR16] C. W. Davis and A. Ray, Satellite operators as group actions on knot concordance, Algebr. Geom.
Topol. 16 (2016), no. 2, 945–969.
[End95] H. Endo, Linear independence of topologically slice knots in the smooth cobordism group, Topology
Appl. 63 (1995), no. 3, 257–262.
[Fel16] P. Feller, The degree of the Alexander polynomial is an upper bound for the topological slice genus,
Geom. Topol. 20 (2016), no. 3, 1763–1771.
[FGMW10] M. Freedman, R. Gompf, S. Morrison, and K. Walker, Man and machine thinking about the smooth
4-dimensional Poincaré conjecture, Quantum Topol. 1 (2010), no. 2, 171–208.
 SLICE KNOTS AND KNOT CONCORDANCE 26

[FKL+ 22] S. Friedl, T. Kitayama, L. Lewark, M. Nagel, and M. Powell, Homotopy ribbon concordance, Blanchfield
pairings, and twisted Alexander polynomials, Canad. J. Math. 74 (2022), no. 4, 1137–1176.
[FM16] P. Feller and D. McCoy, On 2-bridge knots with differing smooth and topological slice genera, Proc.
Amer. Math. Soc. 144 (2016), no. 12, 5435–5442.
[FNOP19a] S. Friedl, M. Nagel, P. Orson, and M. Powell, A survey of the foundations of four-manifold theory in
the topological category, arXiv e-prints (Oct. 2019), arXiv:1910.07372, available at 1910.07372.
[FNOP19b] S. Friedl, M. Nagel, P. Orson, and M. Powell, Satellites and concordance of knots in 3-manifolds,
Trans. Amer. Math. Soc. 371 (2019), no. 4, 2279–2306.
[Fox61] R. H. Fox, Some problems in knot theory, Topology of 3-manifolds and related topics (Proc. The
Univ. of Georgia Institute, 1961), 1961, pp. 168–176.
[Fox62] R. H. Fox, A quick trip through knot theory, Topology of 3-manifolds and related topics (Proc. The
Univ. of Georgia Institute, 1961), 1962, pp. 120–167.
[FP20] S. Friedl and M. Powell, Homotopy ribbon concordance and Alexander polynomials, Arch. Math.
(Basel) 115 (2020), no. 6, 717–725.
[FQ90] M. H. Freedman and F. Quinn, Topology of 4-manifolds, Princeton Mathematical Series, vol. 39,
Princeton University Press, Princeton, NJ, 1990.
[Fre82] M. H. Freedman, The topology of four-dimensional manifolds, J. Differential Geometry 17 (1982),
no. 3, 357–453.
[FS98] R. Fintushel and R. J. Stern, Knots, links, and 4-manifolds, Invent. Math. 134 (1998), no. 2, 363–400.
[FT86] M. H. Freedman and L. R. Taylor, A universal smoothing of four-space, J. Differential Geom. 24
(1986), no. 1, 69–78.
[GJ11] J. Greene and S. Jabuka, The slice-ribbon conjecture for 3-stranded pretzel knots, Amer. J. Math.
133 (2011), no. 3, 555–580.
[Gom85] R. E. Gompf, An infinite set of exotic R4 ’s, J. Differential Geom. 21 (1985), no. 2, 283–300.
[Gom86] R. E. Gompf, Smooth concordance of topologically slice knots, Topology 25 (1986), no. 3, 353–373.
[Gor81] C. McA. Gordon, Ribbon concordance of knots in the 3-sphere, Math. Ann. 257 (1981), no. 2, 157–170.
[GS99] R. E. Gompf and A. I. Stipsicz, 4-manifolds and Kirby calculus, Graduate Studies in Mathematics,
vol. 20, American Mathematical Society, Providence, RI, 1999.
[GST10] R. E. Gompf, M. Scharlemann, and A. Thompson, Fibered knots and potential counterexamples to
the property 2R and slice-ribbon conjectures, Geom. Topol. 14 (2010), no. 4, 2305–2347.
[GT04] S. Garoufalidis and P. Teichner, On knots with trivial Alexander polynomial, J. Differential Geom. 67
(2004), no. 1, 167–193.
[Hay19] K. Hayden, Cross-sections of unknotted ribbon disks and algebraic curves, Compos. Math. 155 (2019),
no. 2, 413–423.
[Hed09] M. Hedden, On knot Floer homology and cabling. II, Int. Math. Res. Not. IMRN 12 (2009), 2248–2274.
[HHR16] M. A. Hill, M. J. Hopkins, and D. C. Ravenel, On the nonexistence of elements of Kervaire invariant
one, Ann. of Math. (2) 184 (2016), no. 1, 1–262.
[Hil12] J. Hillman, Algebraic invariants of links, Second, Series on Knots and Everything, vol. 52, World
Scientific Publishing Co. Pte. Ltd., Hackensack, NJ, 2012.
[HK12] M. Hedden and P. Kirk, Instantons, concordance, and Whitehead doubling, J. Differential Geom. 91
(2012), no. 2, 281–319.
[HKL12] M. Hedden, P. Kirk, and C. Livingston, Non-slice linear combinations of algebraic knots, J. Eur.
Math. Soc. (JEMS) 14 (2012), no. 4, 1181–1208.
[HKL16] M. Hedden, S.-G. Kim, and C. Livingston, Topologically slice knots of smooth concordance order two,
J. Differential Geom. 102 (2016), no. 3, 353–393.
[HKP23] J. Hom, S. Kang, and J. Park, Topologically and rationally slice knots, arXiv e-prints (Apr. 2023),
arXiv:2304.06265, available at 2304.06265.
[HKPS22] J. Hom, S. Kang, J. Park, and M. Stoffregen, Linear independence of rationally slice knots, Geom.
Topol. 26 (2022), no. 7, 3143–3172.
[HLR12] M. Hedden, C. Livingston, and D. Ruberman, Topologically slice knots with nontrivial Alexander
polynomial, Adv. Math. 231 (2012), no. 2, 913–939.
[Hom14] J. Hom, The knot Floer complex and the smooth concordance group, Comment. Math. Helv. 89
(2014), no. 3, 537–570.
[Hom15] J. Hom, An infinite-rank summand of topologically slice knots, Geom. Topol. 19 (2015), no. 2,
1063–1110.
[Hom17] J. Hom, A survey on Heegaard Floer homology and concordance, J. Knot Theory Ramifications 26
(2017), no. 2, 1740015, 24.
[Hom19] J. Hom, Correction to the article An infinite-rank summand of topologically slice knots, Geom. Topol.
23 (2019), no. 5, 2699–2700.
[HPC21] M. Hedden and J. Pinzón-Caicedo, Satellites of infinite rank in the smooth concordance group, Invent.
Math. 225 (2021), no. 1, 131–157.
[HR23] M. Hedden and K. Raoux, Knot Floer homology and relative adjunction inequalities, Selecta Math.
(N.S.) 29 (2023), no. 1, Paper No. 7, 48.
 SLICE KNOTS AND KNOT CONCORDANCE 27

[HS21] K. Hayden and I. Sundberg, Khovanov homology and exotic surfaces in the 4-ball, arXiv e-prints
(Aug. 2021), arXiv:2108.04810, available at 2108.04810.
[Jia81] B. J. Jiang, A simple proof that the concordance group of algebraically slice knots is infinitely generated,
Proc. Amer. Math. Soc. 83 (1981), no. 1, 189–192.
[JKM23] R. Johanningsmeier, H. Kim, and A. N. Miller, A partial resolution of Hedden’s conjecture on satellite
homomorphisms, arXiv e-prints (Aug. 2023), arXiv:2308.06890, available at 2308.06890.
[JZ20] A. Juhász and I. Zemke, Distinguishing slice disks using knot Floer homology, Selecta Math. (N.S.)
26 (2020), no. 1, Paper No. 5, 18.
[Kan22] S. Kang, Link homology theories and ribbon concordances, Quantum Topol. 13 (2022), no. 1, 183–205.
[Kaw09] A. Kawauchi, Rational-slice knots via strongly negative-amphicheiral knots, Commun. Math. Res. 25
(2009), no. 2, 177–192.
[Kea73] C. Kearton, Classification of simple knots by Blanchfield duality, Bull. Amer. Math. Soc. 79 (1973),
952–955.
[Kea75a] C. Kearton, Blanchfield duality and simple knots, Trans. Amer. Math. Soc. 202 (1975), 141–160.
[Kea75b] C. Kearton, Cobordism of knots and Blanchfield duality, J. London Math. Soc. (2) 10 (1975), no. 4,
406–408.
[Ker65] M. A. Kervaire, Les nœuds de dimensions supérieures, Bull. Soc. Math. France 93 (1965), 225–271.
[Ker71] M. A. Kervaire, Knot cobordism in codimension two., Manifolds–Amsterdam 1970 (Proc. Nuffic
Summer School), 1971, pp. 83–105.
[Kir89] R. C. Kirby, The topology of 4-manifolds, Lecture Notes in Mathematics, vol. 1374, Springer-Verlag,
Berlin, 1989.
[Kir97] R. Kirby, Problems in low-dimensional topology, Geometric topology (Athens, GA, 1993), 1997,
pp. 35–473.
[KL22] T. Kim and C. Livingston, Knot reversal acts non-trivially on the concordance group of topologically
slice knots, Selecta Math. (N.S.) 28 (2022), no. 2, Paper No. 38, 17.
[KL99] P. Kirk and C. Livingston, Twisted knot polynomials: inversion, mutation and concordance, Topology
38 (1999), no. 3, 663–671.
[KLS22] M. H. Kim, C. Lee, and M. Song, Non-slice 3-stranded pretzel knots, J. Knot Theory Ramifications
31 (2022), no. 3, Paper No. 2250018, 10.
[KM63] M. A. Kervaire and J. W. Milnor, Groups of homotopy spheres. I, Ann. of Math. (2) 77 (1963),
504–537.
[KMRS21] A. Kjuchukova, A. N. Miller, A. Ray, and S. Sakallı, Slicing knots in definite 4-manifolds, arXiv
e-prints (Dec. 2021), arXiv:2112.14596, available at 2112.14596.
[KOPR21] M. H. Kim, P. Orson, J. Park, and A. Ray, Open problems, The disc embedding theorem, 2021,
pp. 353–382.
[KR21] M. R. Klug and B. M. Ruppik, Deep and shallow slice knots in 4-manifolds, Proc. Amer. Math. Soc.
Ser. B 8 (2021), 204–218.
[KS77] R. C. Kirby and L. C. Siebenmann, Foundational essays on topological manifolds, smoothings, and
triangulations., Princeton University Press, Princeton, N.J.; University of Tokyo Press, Tokyo„ 1977.
With notes by John Milnor and Michael Atiyah.
[KST20] K. Kishimoto, T. Shibuya, and T. Tsukamoto, Sliceness of alternating pretzel knots and links,
Topology Appl. 282 (2020), 107317, 6.
[Lec12] A. G. Lecuona, On the slice-ribbon conjecture for Montesinos knots, Trans. Amer. Math. Soc. 364
(2012), no. 1, 233–285.
[Lec15] A. G. Lecuona, On the slice-ribbon conjecture for pretzel knots, Algebr. Geom. Topol. 15 (2015),
no. 4, 2133–2173.
[Lev16] A. S. Levine, Nonsurjective satellite operators and piecewise-linear concordance, Forum Math. Sigma
4 (2016), Paper No. e34, 47.
[Lev69a] J. Levine, Invariants of knot cobordism, Invent. Math. 8 (1969), 98–110; addendum, ibid. 8 (1969),
355.
[Lev69b] J. Levine, Knot cobordism groups in codimension two, Comment. Math. Helv. 44 (1969), 229–244.
[Lis07] P. Lisca, Lens spaces, rational balls and the ribbon conjecture, Geom. Topol. 11 (2007), 429–472.
[Lit84] R. A. Litherland, Cobordism of satellite knots, Four-manifold theory (Durham, N.H., 1982), 1984,
pp. 327–362.
[Liv05] C. Livingston, A survey of classical knot concordance, Handbook of knot theory, 2005, pp. 319–347.
[Liv93] C. Livingston, Knot theory, Carus Mathematical Monographs, vol. 24, Mathematical Association of
America, Washington, DC, 1993.
[Liv99] C. Livingston, Order 2 algebraically slice knots, Proceedings of the Kirbyfest (Berkeley, CA, 1998),
1999, pp. 335–342.
[Lon14] L. Long, Slice Ribbon Conjecture, Pretzel Knots and Mutation, Ph.D. Thesis, 2014.
[LS22] R. Lipshitz and S. Sarkar, Khovanov homology of strongly invertible knots and their quotients, arXiv
e-prints (Mar. 2022), arXiv:2203.13895, available at 2203.13895.
 SLICE KNOTS AND KNOT CONCORDANCE 28

[LZ19] A. S. Levine and I. Zemke, Khovanov homology and ribbon concordances, Bull. Lond. Math. Soc. 51
(2019), no. 6, 1099–1103.
[Mal22] A. Mallick, Knot Floer homology and surgery on equivariant knots, arXiv e-prints (Jan. 2022),
arXiv:2201.07299, available at 2201.07299.
[Mat02] Y. Matsumoto, An introduction to Morse theory, Translations of Mathematical Monographs, vol. 208,
American Mathematical Society, Providence, RI, 2002. Translated from the 1997 Japanese original
by Kiki Hudson and Masahico Saito, Iwanami Series in Modern Mathematics.
[Mil17a] A. N. Miller, Distinguishing mutant pretzel knots in concordance, J. Knot Theory Ramifications 26
(2017), no. 7, 1750041, 24.
[Mil17b] A. N. Miller, The topological sliceness of 3-strand pretzel knots, Algebr. Geom. Topol. 17 (2017),
no. 5, 3057–3079.
[Mil18] A. N. Miller, A note on the topological sliceness of some 2-bridge knots, Math. Proc. Cambridge
Philos. Soc. 164 (2018), no. 1, 185–191.
[Mil23] A. N. Miller, Homomorphism obstructions for satellite maps, Trans. Amer. Math. Soc. Ser. B 10
(2023), 220–240.
[Mil56] J. Milnor, On manifolds homeomorphic to the 7-sphere, Ann. of Math. (2) 64 (1956), 399–405.
[Mil63] J. Milnor, Morse theory., Princeton University Press, Princeton, N.J., 1963. Based on lecture notes
by M. Spivak and R. Wells.
[Mil65] J. Milnor, Lectures on the h-cobordism theorem., Princeton University Press, Princeton, N.J., 1965.
Notes by L. Siebenmann and J. Sondow.
[Miy94] K. Miyazaki, Nonsimple, ribbon fibered knots, Trans. Amer. Math. Soc. 341 (1994), no. 1, 1–44.
[MMP20] C. Manolescu, M. Marengon, and L. Piccirillo, Relative genus bounds in indefinite four-manifolds,
arXiv e-prints (Dec. 2020), arXiv:2012.12270, available at 2012.12270.
[MMRS22] M. Marengon, A. N. Miller, A. Ray, and A. I. Stipsicz, A note on surfaces in CP2 and CP2 #CP2 ,
arXiv e-prints (Oct. 2022), arXiv:2210.12486, available at 2210.12486.
[MMSW23] C. Manolescu, M. Marengon, S. Sarkar, and M. Willis, A generalization of Rasmussen’s invariant,
with applications to surfaces in some four-manifolds, Duke Math. J. 172 (2023), no. 2, 231–311.
[Moi77] E. E. Moise, Geometric topology in dimensions 2 and 3, Graduate Texts in Mathematics, Vol. 47,
Springer-Verlag, New York-Heidelberg, 1977.
[MP18] A. N. Miller and L. Piccirillo, Knot traces and concordance, J. Topol. 11 (2018), no. 1, 201–220.
[MP19] A. N. Miller and M. Powell, Stabilization distance between surfaces, Enseign. Math. 65 (2019), no. 3-4,
397–440.
[MP21] C. Manolescu and L. Piccirillo, From zero surgeries to candidates for exotic definite four-manifolds,
arXiv e-prints (Feb. 2021), arXiv:2102.04391, available at 2102.04391.
[MP23] A. N. Miller and M. Powell, Strongly invertible knots, equivariant slice genera, and an equivariant
algebraic concordance group, J. Lond. Math. Soc. (2) 107 (2023), no. 6, 2025–2053.
[MZ21] M. Miller and I. Zemke, Knot Floer homology and strongly homotopy-ribbon concordances, Math. Res.
Lett. 28 (2021), no. 3, 849–861.
[Nai97] S. Naik, Equivariant concordance of knots in S 3 , KNOTS ’96 (Tokyo), 1997, pp. 81–89.
[Nic11] L. Nicolaescu, An invitation to Morse theory, Second, Universitext, Springer, New York, 2011.
[NOPP19] M. Nagel, P. Orson, J. Park, and M. Powell, Smooth and topological almost concordance, Int. Math.
Res. Not. IMRN 23 (2019), 7324–7355.
[Nor69] R. A. Norman, Dehn’s lemma for certain 4-manifolds, Invent. Math. 7 (1969), 143–147.
[OSS17] P. S. Ozsváth, A. I. Stipsicz, and Z. Szabó, Concordance homomorphisms from knot Floer homology,
Adv. Math. 315 (2017), 366–426.
[Ott14] C. Otto, The (n)-solvable filtration of link concordance and Milnor’s invariants, Algebr. Geom. Topol.
14 (2014), no. 5, 2627–2654.
[PC17] J. Pinzón-Caicedo, Independence of satellites of torus knots in the smooth concordance group, Geom.
Topol. 21 (2017), no. 6, 3191–3211.
[Pic20] J. Pichelmeyer, Genera of knots in the complex projective plane, J. Knot Theory Ramifications 29
(2020), no. 12, 2050081, 32.
[PR21] M. Powell and A. Ray, The development of topological 4-manifold theory, The disc embedding theorem,
2021, pp. 295–330.
[Qui82] F. Quinn, Ends of maps. III. Dimensions 4 and 5, J. Differential Geometry 17 (1982), no. 3, 503–521.
[Rao20] K. Raoux, τ -invariants for knots in rational homology spheres, Algebr. Geom. Topol. 20 (2020),
no. 4, 1601–1640.
[Ras10] J. Rasmussen, Khovanov homology and the slice genus, Invent. Math. 182 (2010), no. 2, 419–447.
[Ray15] A. Ray, Satellite operators with distinct iterates in smooth concordance, Proc. Amer. Math. Soc. 143
(2015), no. 11, 5005–5020.
[Rol85] D. Rolfsen, Piecewise-linear I-equivalence of links, Low-dimensional topology (Chelwood Gate, 1982),
1985, pp. 161–178.
[Rol90] D. Rolfsen, Knots and links, Mathematics Lecture Series, vol. 7, Publish or Perish, Inc., Houston,
TX, 1990. Corrected reprint of the 1976 original.
 SLICE KNOTS AND KNOT CONCORDANCE 29

[Rud76] L. Rudolph, How independent are the knot-cobordism classes of links of plane curve singularities,
Notices Amer. Math. Soc 23 (1976), 410.
[Rud93] L. Rudolph, Quasipositivity as an obstruction to sliceness, Bull. Amer. Math. Soc. (N.S.) 29 (1993),
no. 1, 51–59.
[Sco05] A. Scorpan, The wild world of 4-manifolds, American Mathematical Society, Providence, RI, 2005.
[SS22] I. Sundberg and J. Swann, Relative Khovanov-Jacobsson classes, Algebr. Geom. Topol. 22 (2022),
no. 8, 3983–4008.
[Sta62] J. Stallings, The piecewise-linear structure of Euclidean space, Proc. Cambridge Philos. Soc. 58
(1962), 481–488.
[Sto77] N. W. Stoltzfus, Unraveling the integral knot concordance group, Mem. Amer. Math. Soc. 12 (1977),
no. 192, iv+91.
[Tau87] C. H. Taubes, Gauge theory on asymptotically periodic 4-manifolds, J. Differential Geom. 25 (1987),
no. 3, 363–430.
[Tro73] H. F. Trotter, On S-equivalence of Seifert matrices, Invent. Math. 20 (1973), 173–207.
[Tro78] H. F. Trotter, Knot module and Seifert matrices., Knot theory (Proc. Sem., Plans-sur-Bex, 1977),
1978, pp. 291–299,.
[Wor16] Workshop Organisers, Final report, Synchronizing smooth and topological 4-manifolds, 2016.
[Yas15] K. Yasui, Corks, exotic 4-manifolds and knot concordance, arXiv e-prints (May 2015),
arXiv:1505.02551, available at 1505.02551.
[Yas91] A. Yasuhara, (2, 15)-torus knot is not slice in CP2 , Proc. Japan Acad. Ser. A Math. Sci. 67 (1991),
no. 10, 353–355.
[Yas92] A. Yasuhara, On slice knots in the complex projective plane, Rev. Mat. Univ. Complut. Madrid 5
(1992), no. 2-3, 255–276.
[Yil18] E. Z. Yildiz, A note on knot concordance, Algebr. Geom. Topol. 18 (2018), no. 5, 3119–3128.
[Zem19] I. Zemke, Knot Floer homology obstructs ribbon concordance, Ann. of Math. (2) 190 (2019), no. 3,
931–947.

Max-Planck-Institut für Mathematik, Vivatsgasse 7, 53111 Bonn, Germany

Email address: [email protected]
URL: http://people.mpim-bonn.mpg.de/aruray/