You are on page 1of 67



Abstract. With a 4-ended tangle T , we associate a Heegaard Floer invariant CFT∂ (T ), the
peculiar module of T . We classify peculiar modules in terms of immersed curves on the 4-
punctured sphere, based on an algorithm due to Hanselman, Rasmussen and Watson [HRW16].
We deduce that the peculiar module of any tangle T can be computed combinatorially using
nice diagrams and illustrate this result by computing the peculiar modules of the 2-stranded
pretzel tangles T2n,−(2m+1) for n, m > 0.
arXiv:1712.05050v1 [math.GT] 13 Dec 2017

Based on Zarev’s bordered sutured Heegaard Floer theory [Zar11], we prove a glueing for-
mula for peculiar modules which recovers link Floer homology HFL. [ We then study some
applications: Firstly, we show that peculiar modules detect rational tangles. Secondly, we give
short proofs of various skein exact triangles. Finally, we observe certain symmetries in the
peculiar modules of the pretzel tangles T2n,−(2m+1) and deduce that mutation of those tangles
preserves δ-graded, and for some orientations even bigraded link Floer homology.

Table of contents

Introduction 1
1. Preliminaries: Algebraic structures from dg categories 5
2. The invariant CFT∂ 14
3. Pairing 4-ended tangles 25
4. Curved complexes for marked surfaces 33
5. Applications 54
References 66


0.1. Peculiar modules. Let L be a link in the 3-sphere S 3 . Consider an embedded closed
3-ball B 3 ⊂ S 3 whose boundary intersects L transversely. Then, modulo a parametrization of
the boundary ∂B 3 , the embedding L ∩ B 3 ,→ B 3 is essentially what we call a tangle. In [Zib16],
I introduced a set of Alexander polynomials ∇sT and a Heegaard Floer theory HFT(T
[ ) for such
tangles T . They should be regarded as generalisations of the classical multivariate Alexander
polynomial [Ale28] and Ozsváth and Szabó’s and J. Rasmussen’s knot and link Floer homology
[OS03a, Ras03, OS05], respectively. Indeed, both tangle invariants have similar properties to
their corresponding link invariants. In particular, the graded Euler characteristic of HFT(T
[ )
recovers the polynomial invariants ∇T . Moreover, the polynomials ∇T satisfy a simple glueing
s s

theorem which allows one to prove results about the classical multivariate Alexander polynomial
of links, such as invariance under Conway mutation [Zib16, corollary 3.6]. Unfortunately, we
do not have a similar glueing theorem for the categorified invariants HFT(T
[ ).



The main objective of this paper is to resolve this problem in the case of 4-ended tangles,
ie tangles with four ends on ∂B 3 , see figure 1. For this purpose, we upgrade the tangle Floer
homology HFT(T
[ ) of 4-ended tangles T to an invariant which we call the peculiar module
of T and denote by CFT∂ (T ). This is done by adding some more structure maps. In fact,
we construct an even more general invariant CFT− (T ) := CFT− (T, M ), a generalised peculiar
module, for tangles T in homology 3-balls M with spherical boundary. As algebraic objects, both
generalized and ordinary peculiar modules are curved type D structures over certain algebras,
the generalized and ordinary peculiar algebras A− n and A , respectively. The algebra A is a
∂ ∂

quotient of A− n , obtained by setting certain variables equal
to 0. Similarly, CFT∂ (T ) can be recovered from CFT− (T ),
just as the hat version CFL
[ of link Floer homology can be
recovered from its −-version CFL− .
Both generalised and ordinary peculiar modules are Hee-
gaard Floer type invariants and, as such, rely on some choice
of Heegaard diagram. So the first goal is to prove that the in-
variants are independent of this choice. However, this follows
essentially from multi-pointed Heegaard Floer theory.

Theorem 0.1 (2.17). Given a 4-ended tangle T in a homology Figure 1. A diagram of a 4-
3-ball with spherical boundary, the bigraded chain homotopy ended tangle; in this case, the
(2, −3)-pretzel tangle T2,−3
types of CFT− (T ) and CFT∂ (T ) are invariants of T .

We do not have a glueing theorem for the generalised peculiar modules CFT− (T ), except
that one can recover CFL− of certain closures of the tangle T from CFT− (T ) (see remark 2.28).
Nonetheless, we do have a glueing formula for peculiar modules CFT∂ (T ). Its proof relies on
Zarev’s glueing theorem for his bordered sutured invariants [Zar11] and an identification of some
structure maps of certain bordered sutured invariants for tangles and peculiar modules. The
precise statement of the glueing theorem uses the -tensor product between type A and type D
structures familiar from bordered Heegaard Floer homology; for details, see definition 1.19.
Theorem 0.2 (3.6). Let T1 and T2 be two 4-ended tangles
and L the link obtained by glueing them together as illustrated
in figure 2. Then the link Floer homology HFL(L)
[ can be com- T1 T2
puted from CFT (T1 ) and CFT (T2 ). More precisely, there
∂ ∂

exists a bounded, strictly unital type AA structure P such that

[ ⊗ V i = CFT∂ (T1 )  P  CFT∂ (T2 )
Figure 2. A link obtained
where V is some 2-dimensional vector space and i ∈ {0, 1}. from two 4-ended tangles

0.2. Classification of peculiar modules. In section 4, we classify peculiar modules in terms
of immersed curves on the 4-punctured sphere.

Definition 0.3 (4.23). A loop on the 4-punctured sphere S = S 2 r 4D2 is a pair (γ, X), where
γ is an immersion of an oriented circle into S representing a non-trivial primitive element of
π1 (S) and X ∈ GLn (F2 ) for some integer n. For such loops (γ, X), we call X the local system
of the loop. A collection of loops is a set of loops {(γi , Xi )}i∈I such that the immersed curves
γi are pairwise non-homotopic.


Theorem 0.4 (4.31, 4.35). With every peculiar module (C, ∂), we can associate a collection of
loops L(C, ∂) = {(γi , Xi )}i∈I such that if (C 0 , ∂ 0 ) is another peculiar module with L(C 0 , ∂ 0 ) =
{(γj0 , Xj0 )}j∈J , (C, ∂) and (C 0 , ∂ 0 ) are homotopic iff there is a bijection ι : I → J such that γi is
homotopic to γι(i) 0 and Ai is similar to A0ι(i) for all i ∈ I.

Definition 0.5 (4.34). In particular, with any 4-ended
tangle T in a homology 3-ball M with spherical boundary, we
can associate a collection of loops, denoted by LT := LT,M ,
which is a tangle invariant up to homotopy of the underlying
curves and similarity of the local systems.

Question 0.6. The number of curves in LT is obviously a
tangle invariant. What is its geometric meaning?

The proof of theorem 0.4 is based on an algorithm due to
Hanselman, Rasmussen and Watson [HRW16] which they use
Figure 3. The three loops
to classify bordered Heegaard Floer invariants for 3-manifolds of L
T2,−3 (with the unique
with torus boundary. There are striking similarities between 1-dimensional local systems)
their bordered Heegaard Floer invariants and peculiar mod- on the 4-punctured sphere for
ules, and it would be interesting to see if there exists a closer the (2, −3)-pretzel tangle T2,−3
from figure 1.
connection between them apart from their formal properties.
Another important ingredient for the classification of peculiar modules is a classification of
morphisms between them in terms of the intersection theory of the corresponding collections of
loops. As a corollary of this second classification, we obtain the following two reformulations of
the glueing theorem.

Corollary 0.7 (4.44). With the same notation as in theorem 0.2,

[ ⊗ V i = HF(LT1 , Lm(T2 ) ) = H∗ (Mor(CFT∂ (T1 ), CFT∂ (m(T2 )))),

where m(T2 ) denotes the mirror of T2 and HF denotes Lagrangian intersection Floer homology.

We actually prove theorem 0.4 for all categories of curved complexes over arbitrary marked
surfaces, see definition 4.7. As a consequence, we can show that by passing to certain quotients
of the peculiar algebra A∂ , we do not lose information. This rather abstract observation has
the following very practical consequence.

Theorem 0.8 (4.38). Peculiar modules for 4-ended tangles can be computed combinatorially.

We illustrate this result in section 5.3 by computing CFT∂ (T ) for an infinite family of 2-
stranded pretzel tangles. The following definition is inspired by [HW15, definition 3.2].

Definition 0.9. A tangle T is called loop-type if all local systems in LT are similar to per-
mutation matrices.

All tangles for which I have so far computed the invariants are loop-type. This provokes
the following question, whose corresponding counterpart for 3-manifolds with torus boundary
is open as well [HRW16, section 2.4].

Question 0.10. Are all tangles loop-type?

For peculiar modules.12 triangle relating the (appropriately stabilised) link Floer ho- mologies of links that differ in these three tangles. the description of pe- culiar modules of rational tangles is particularly simple. it gives rise to an exact theorem 0. n . see figure 4.4.3. Baldwin and Levine conjectured the following. see figure 5 for an illustration.13 (Conway mutation). Given a link L in S 3 . CFT∂ (T−n ) CFT∂ (Tn ) n . Let Tn be the positive n- twist tangle.5]. otherwise we do not change any orientation. If the tangles are Figure 4. If this means that we need to reverse the orientation of the two open components of R. As another application. We say Figure 5. Conjecture 0. The basic tangles oriented and coloured consistently. T−n the negative n-twist tangle and T0 the trivial   tangle. Theorem 0. A 4-ended tangle T in the 3-ball is rational iff LT is a single embedded loop with the unique 1-dimensional local system. we choose an orientation of L0 that agrees with the one for L outside of R. This is an easy consequence of unlink detection of link Floer homology and should be regarded as its corresponding analogue for invariants of 4-ended tangles..4 CLAUDIUS BODO ZIBROWIUS 0. not invariant under mutation. In fact. However. this is true for any invari- ant of 4-ended tangles for which there exists a glueing theorem recovering link Floer homology. Then there is an exact triangle            . Furthermore.11 (5. see theorem 5. any tangle invariant like CFT∂ also detects orientable tangle genera.5). Peculiar modules detect rational tangles. defined in a suitable sense. However. this will be discussed in a forthcoming. and hence so is rational tangle detection for our invariants: Theorem 0. we reprove a result originally due to Manolescu [Mnl06].. see also 5. let L0 be the link obtained by cutting out a tangle diagram R −→ R R with four ends from a diagram of L and glueing it back in after a half-rotation.  . If L is oriented. in [BL11.12 (5.             CFT∂ (T0 ) ⊗ V (a) Tn (b) T−n (c) T0 where V is a 2-dimensional vector space. known as the δ- grading. one obtains (bi)graded ver. Then HFL(L) [ and HFL(L [ 0 ) agree after collapsing the bigrading to a single Z-grading. More colloquially. δ-graded link Floer homology is mutation invariant. conjecture 1.6. Let L be a link and let L0 be obtained from L by Conway mutation. Likewise. . namely the existence of an unoriented skein exact sequence. Conway mutation L0 is a Conway mutant of L and we call R the mutating tangle in this mutation. Similarly. we obtain the following slight generalisation of Ozsváth and Szabó’s oriented skein exact sequence [OS03a]. Definition 0.2). in general. Ozsváth and Szabó showed in [OS03b] that knot and link Floer homology is.for the skein exact triangle from sions of this triangle. then we also reverse the orientation of all other components of R during the mutation.14. Applications. separate pa- per [WZ].

An infinite family of δ-graded link Floer homology. Robert Lipshitz. Andy Manion. The paper was completed during my time as a CIRGET postdoctoral fellow at the Université de Sherbrooke. The main goal of this section is to develop some tools for dealing with this problem. In fact. .5.  2n    tangles for n. My special thanks go to Liam Watson for his continuing interest in my work. namely type D and curved type D structures. where he confirms conjecture 0.14 for various families of mutant pairs (different from the one in theorem 0. m > 0 is more involved and relies on theorem 0.15 (5. OS16. the calculation of the invariants for general n. Preliminaries: Algebraic structures from dg categories In this paper. Acknowledgements. However. If we reverse the orientation of one of the two strands. namely the cancellation lemma (1. I consider myself very fortunate to have been his student.15). OS17] and their corresponding decategorifications in terms of the representation theory of Uq (gl(1|1)) [EPV15. we often work in categories of various algebraic structures. 1. The result again simply follows from an observation that the peculiar invariants for the mutating tangles have a certain symmetry. pretzel tangles for n. preserves bi.. we often want to simplify these structures by replacing them by homotopy equivalent ones. 0. the latter for making the structure maps “look nicer”. In all settings.   graded link Floer homology.       2m + 1   . It is interesting to compare the ideas described in this paper to the tangle Floer theories defined in [PV14.2]. essentially by doing Gaussian elimination as in [BB14. oriented as in figure 6. I also thank Jonathan Hanselman.12). Some parts of this paper were written during my stay at the Isaac Newton Institute during the programme Homology Theories in Low Dimensions (EPSRC grant number EP/K032208/1). mutation of these tangles preserves Figure 6.23) and the clean-up lemma (1. m > 0 This generalises an earlier result from my thesis [Zib17] for the (2. Vera Vértesi and Marcus Zibrowius for helpful conversations. for which I thank the then Head of Department Martin Hyland. The former can be used to reduce the number of generators of an algebraic structure. My PhD was funded by an EPSRC scholarship covering tuition fees and a DPMMS grant for maintenance. we prove a slightly stronger version of this conjecture for an infinite family of mutating tangles. but also type A structures and various bimodules. Mnn16]. I also want to mention some impressive work of Lambert-Cole [Lam16. Ina Petkova. Theorem 0. Similar work by other people. −(2m + 1))-pretzel . 0.4. lemma 3. . I would therefore like to take the opportunity to thank my PhD supervisor Jake Rasmussen for his generous support. . after identifying the Alexander gradings of the two open strands. using entirely different techniques. m > 0.8. the definition of our generalised peculiar modules CFT− is primarily inspired by Ozsváth and Szabó’s invariants from [OS17]. PECULIAR MODULES FOR 4-ENDED TANGLES 5 In this paper.19. essentially by changing the basis.25). This paper grew out of the final chapter of my PhD thesis [Zib17]. Lam17]. −3)-pretzel tangle. see remark 2.. I thank my PhD examiners Ivan Smith and András Juhász for many valuable comments on and corrections to my thesis. Mutation of (2n.

The unit in Com is the complex 0 → F2 → 0.4 and 1. B) i∈Z endowed with differentials ∂i : Mori (A. For simplicity. Note that the identity morphisms have degree zero and lie in the kernel of ∂. where m : Morj (B. B). Let Com be the category of Z-graded chain complexes over F2 and grading preserving chain maps between them. B) = Mori (A. Since the hom-sets in H0 (C) are just quotients of those in C0 . C(A. see for example [Rie14].4. Next. we denote this category by H0 (C). Definition 1. Let C be a dg category. B)). B) := MorV (1V . so we do not need to keep track of signs. We also study its functoriality properties and interpret the box tensor product between type A and type D structures in this framework. C) denotes composition in C. C) ⊗ Mori (A. Definition 1. see definitions 1. we now get the usual notions of chain homotopies between morphisms and objects. However.6 CLAUDIUS BODO ZIBROWIUS In the category of ordinary chain complexes. with the correct sign conventions. We will spend the first part of this section explaining a general construction which turns any differential graded category into another such category in which the lemmas hold in some generality sufficient for our purposes. which is associative and unital. Example 1.3. For more details on enriched categor- ies. By passing to the underlying ordinary category. A differential graded (dg) category C over F2 is an enriched category over Com. both tools will be familiar to the reader as easy exercises in linear algebra.2. we state and prove the cancellation and clean-up lemmas.1. Next. we only work over the field F2 = Z/2. [Rie14. B) → Mori+j (A. the underlying ordinary category C0 of C has the same objects as C and its hom-sets are defined by C0 (A. the hom-objects are Z-graded F2 -vector spaces. To spell this out more explicitly. ie vector space homomorphisms satisfying ∂i−1 ∂i = 0 and (1) ∂ ◦ m = m ◦ (∂ ⊗ id + id ⊗∂). 1. obtained from C by replacing the hom-objects by their homologies with respect to the differential ∂. we show that the various different algebraic structures mentioned above arise naturally from this general construction.1. B) → Mori−1 (A. and the morphisms in Com are grading preserving. definition 3. Finally. So it might not be too surprising that they also work in quite general settings. supported in homological degree 0. Hence. .5.5]. Given an enriched category C over some monoidal category V. the hom-sets of C0 are those elements in the kernel of ∂0 . we pick out the degree 0 morphisms in H∗ (C). Therefore. The general construction. consider the enriched category H∗ (C) over the category of graded vector spaces and grading preserving morphisms between them. most (if not all) statements should also hold over fields of arbitrary characteristic. M Mor(A.


The reason why we need to pass to the underlying category is that otherwise, two objects could
be (chain) isomorphic through grading shifting morphisms.

Definition 1.4. (cp. [Bar04, section 6] and [Sei06, section I.3k]) Given a dg category C, we
define another dg category Mat(C) as follows. Its objects are formal direct sums
Oi [ni ],

where I is some finite index set and Oi [ni ] denotes the object Oi ∈ ob(C) with a formal grading
shift by an integer ni . Morphisms are given by
Morn ( Oi [ni ], Oj [nj ]) := Morn+ni −nj (Oi , Oj ).
i∈I j∈J (i,j)∈I×J

Compositions and differentials in Mat(C) are induced by those in C.

Definition 1.5. (cp. [Sei06, section I.3l]) Given a differential graded category C, we define an
auxiliary category Cxpre (C), the category of pre-complexes, which is an enriched category
over the category of Z-graded vector spaces and grading preserving morphisms between them.
Its objects are pairs (O, dO ), where O ∈ ob(C) and dO ∈ Mor−1 (O, O). The hom-objects are
the same as in C,
Mor((O, dO ), (O0 , dO0 )) = Mor(O, O0 ),
viewed as Z-graded vector spaces. On these, we can define a map

D : Mori ((O, dO ), (O0 , dO0 )) → Mori−1 ((O, dO ), (O0 , dO0 ))

by setting
D(f ) := dO0 ◦ f + f ◦ dO + ∂(f ).
We would like D to be a differential in order to turn Cxpre (C) into a dg category. However,
this only works in general if we restrict ourselves to a full subcategory of Cxpre (C). It is easy to
check that D is always compatible with multiplication in the sense of (1). So D is a differential
D2 (f ) = (d2O0 + ∂(dO0 )) ◦ f + f ◦ (d2O + ∂(dO ))
vanishes. This is, of course, the case for the full subcategory Cx0 (C) of Cxpre (C) consisting of
those objects (O, dO ) for which

(∗) d2O + ∂(dO )

vanishes. However, in some situations, other conditions on (∗) also work. For example, if we
replace Com by the category of Z/2-graded chain complexes, we can restrict to those objects
(O, dO ) for which (∗) is equal to the identity. Also, if the hom-objects are bimodules over an
algebra A, we can ask (∗) to be equal to a. idO for a fixed central algebra element a (of degree
−2) which commutes with all morphisms f . In both cases, D will be a differential. We call any
such full subcategory a category of complexes, denoted by Cx∗ (C), where ∗ ∈ {0, 1, a} is the
value of (∗). By construction, Cx∗ (C) is a dg category.

Remark 1.6. As usual, we can associate a directed graph to a category, where objects cor-
respond to vertices and arrows to morphisms. In the same way, we can think of complexes in
Cx∗ (Mat(C)) as graphs.


The point of the construction above is that after choosing a basis, we can interpret the
categories of type D, type A, type AA and curved type D structures as instances of Cx∗ (Mat(C))
for suitable choices of relatively simple differential graded categories C. But let us start with
an even simpler example: ordinary chain complexes.

Note of warning. In the following examples, our definitions only coincide with the usual
ones after passing to the underlying ordinary categories, see example 1.3. The advantage of our
point of view is that the conditions we usually impose on morphism and chain homotopies for
various algebraic structures arise naturally by viewing those morphisms as elements of chain

Example 1.7 (ordinary chain complexes over F2 ). Let C be the category with a single object •
in grading 0, Mor(•, •) = F2 and vanishing differential. Then (the underlying ordinary category
of) Cx0 (Mat(C)) is Com.

Example 1.8 (type D modules over dg F2 -algebras). Let A be a differential graded algebra
over F2 . Let C be the category with a single object • and morphisms being elements in A.
Composition is multiplication in A and the differential ∂ is induced by the differential on A.
We define the category of type D modules by Cx0 (Mat(C)). (Again, note that we need to pass
to the underlying ordinary category to obtain the definitions in [Zar09] and [LOT08].)

Example 1.9 (type D modules over dg I-algebras). Let us assume that A is an algebra over
some ring I ⊆ A of idempotents and fix a basis {ij }j∈J of idempotents of I, where J is some
index set. Let CAD be the category with one object for each basis element of I, and for any

two such elements i1 and i2 , let Mor(i1 , i2 ) := i2 .A.i1 . Again, composition is multiplication
in A and the differential ∂ is induced by the differential on A. We call Cx0 (Mat(CA D )) the

category of (right) type D modules over dg I-algebras. To obtain the category of left type D
modules, we only need to change the morphism spaces to Mor(i1 , i2 ) := i1 .A.i2 with the obvious
multiplication; however, we usually work with right type D modules in this paper, since we
interpret algebra elements as functions and thus read them from right to left.

Remark 1.10. A priori, the definition in the previous example depends on a choice of basis
for I. In all examples in this paper, there is a natural choice of such a basis, so this is not an
issue. However, we can replace CAD above by the enlarged category C D , where there is an object
for every element in I. Then CA D is a full subcategory of C D and it is not hard to see that
Mat(CA D ) and Mat(C D ) are equivalent. Now, the construction of the category of complexes is
functorial (in the category of dg categories), so after all, the definition above does not depend
on a basis for I.

Example 1.11 (curved type D modules over dg I-algebras). We start with the same category
D as in the previous example, but we fix a central element a ∈ Z(A), the curvature, and
CA c
define the category of curved (right) type D modules with curvature ac as Cxac (Mat(CA
D )). For

a more explicit, but less concise definition, see definition 2.13.

Remark 1.12. In the Heegaard Floer community, the term “curved” seems to be the accepted
attribute for algebraic structures for which some differential is non-vanishing; however, the first


written reference (that I am aware of) in which this terminology is used is of very recent date

Example 1.13 (type A structures over an A∞ -algebra over I). Let A be an A∞ -algebra over
a ring of idempotents I over F2 . As in example 1.9, fix a basis {ik }k∈I of idempotents of I,
where I is some index set. Let CA
A be the category with one object for each basis element in I,

just as for type D structures. However, a morphism in a hom-object Mor(i1 , i2 ) of CA
A is given

by a sequence of vector space homomorphisms

(fi : i2 .A⊗i .i1 → F2 )i≥0

where composition is defined by
(f ◦ g)i (ai ⊗ · · · ⊗ a1 ) := fk (ai ⊗ · · · ⊗ aj+1 ).gj (aj ⊗ · · · ⊗ a1 ).

The differential ∂ is given by
X i−k
(∂(f ))i (ai ⊗ · · · ⊗ a1 ) := fj (ai ⊗ · · · ⊗ µk (al+k ⊗ · · · ⊗ al+1 ) ⊗ · · · ⊗ a1 ).
j+k=i+1 l=0

We call Cx0 (Mat(CAA )) the category of (left) type A structures. We define the category of strictly

unital type A structures by restricting to those objects (O, dO ) such that

dO (·, 1) = idO

dO (·, ai ⊗ · · · ⊗ a1 ) = 0 if i > 1 and aj = 1 for some j = 1, . . . , i
and morphisms to those satisfying

f (·, ai ⊗ · · · ⊗ a1 ) = 0 if i > 0 and aj = 1 for some j = 1, . . . , i.

Right type A structures are defined in an analogous way; we only define the hom-objects of the
underlying dg category to be given by sequences of vector space homomorphisms

(fi : i1 .A⊗i .i2 → F2 )i≥1 ,

and adapt the multiplication maps accordingly. In this paper, we restrict ourselves to left type A

Remark 1.14. When we describe type A structures as directed graphs, it is useful to fix a
basis of the algebra A. Then, we label an arrow corresponding to a morphism f by the formal
sum of those tuples/tensor products of basis elements of the algebra A on which f is non-zero.
In this language, composition of two morphisms f and g can be described as the sum of all
concatenations of labels for f and g (modulo 2).
Now consider a morphism fa whose only label is a tuple of basic algebra elements a =
(ai , . . . , a1 ). Then the arrow of ∂(fa ) is labelled by the formal sum (modulo 2) of all labels
obtained from a by replacing a single entry ai by sequences (a0j , . . . , a01 ) such that

ai = µj (a0j ⊗ · · · ⊗ a01 ).

Example 1.15 (bimodules of various kinds). We can form the categories of type DD, type DA,
type AD and type AA bimodules as follows. We start with the dg category where objects

9 and 1. Note that via the inclusion J . Similarly. we may define functors for bimodules of various types. Likewise. Suppose. since adding a zero-object in the underlying dg category does not change the result after applying Cx0 (Mat(·)). However. Note that the category Cx0 (Mat(CBA. We sometimes include the algebras over which the modules are defined in our notation. Then.13. In the proofs of the glueing formula in section 3. let CBD be the category D corresponding to the J -algebra B. F̃πA (ι2 ) otherwise. Similarly. This functor is well-defined and respects the differential if A and B are A∞ -algebras over I and J . to define an induced functor.ι1 → ι2 . Let us also fix an F2 -basis {ιi |i ∈ J} of J and extend it to a basis {ιi |i ∈ I} of I.→ I. For example. Multiplication is defined A as the product on the two factors and the differential is defined as usual by the Leibniz rule. we need to slightly modify CBA by adding a zero object. There is also a dual construction for type A structures: consider the category CA A corres- ponding to the I-algebra A from example 1. Definition 1. following the usual convention to use subscripts for type A and superscripts for type D sides. respectively. both of which respect the differentials on both sides. Thus. ι2 ) of the form ι2 . Functoriality and pairing for type A and type D structures. we have a subring J of the ring of idempotents I. Let us first consider the construction for (curved) type D structures: Consider the category CA corresponding to the I-algebra A from example 1. it induces a functor FπA : Cx0 (Mat(CA A )) → Cx0 (Mat(CBA.A⊗i .0 )) that likewise respects the differentials on both sides. let CBA be the category corresponding to the J -algebra B. we need to change the underlying algebras and rings of idempotents of the algebraic structures involved.ι1 → F2 is sent to   f ◦ ι2 . we can regard A also as a dg J -algebra. and π is an A∞ -algebra homomorphism.ι1 : ι2 . ι2 )   if ι1 and ι2 are in J and to 0 ∈ Mor F̃πA (ι1 ). we can define a functor F̃πA : CA A → CBA.0 )) agrees with the category of type A structures Cx0 (Mat(CBA )).16. multi-modules can be defined.A⊗i . Similarly.10 CLAUDIUS BODO ZIBROWIUS correspond to idempotents as before.0 as follows: an object ιi is sent to ιi if i ∈ J and to the zero-object otherwise.B ⊗i .ι1 → F2 ∈ Mor(ι1 .13.9. but where the hom-objects are defined as the corresponding products of the hom-objects of CAA and C D in examples 1. A basic morphism f ∈ Mor(ι1 . but we will not need those in the current paper. Let us call this new category CBA. the notation A M B means that M is a type AD A-B-bimodule where A acts on the left and B on the right. Let π : B → A be a dg J -algebra homomorphism.→ I and the J -algebra homomorphism π induce a functor F̃πD : CBD → CA D and hence also a functor FπD : Cx∗ (Mat(CBD )) → Cx∗ (Mat(CA D )). 1. Let A be a dg I-algebra and B be a dg J -algebra. The inclusion J .2. .π ⊗i .0 .

Questions 1. Remark 1. see [LOT10a. . Definition 1. type A.4] in terms of the graphs associated with M and N . For details.16. w1 ) to (v2 .18. the second component of a label on the arrow from (v1 . w1 ) and (v2 .17. .13]. Suppose we have an oriented labelled graph representing a type D structure N over B (with respect to the basis chosen above). PECULIAR MODULES FOR 4-ENDED TANGLES 11 Remark 1. definition 7. pairing type D and type A structures is invariant under homotopy up to homotopy since it is functorial.21. which we can then extend it to a basis of A. . we can define a pairing between other types of bimodules. I. if we have an oriented labelled graph representing a type A structure M over A. This is usually done by requiring that at least one of M or N is bounded: for type D structures. section 2. w2 ) is equal to the second component of the corresponding label of the arrow from w1 to w2 in N . FπA (M ) has the effect of replacing each label a = (an . w2 )- component of ∂  (v1 . The result of the pairing operation described above is a well-defined object in the corresponding category.A.3. that in all examples that we are concerned with in this paper. for type A structures. proposition 2. ie a chain complex. a1 ) by the formal sum of all labels whose images under π ⊗n are equal to a. ∂  ) from [Zar09. Note that for the differential to be well-defined. We now reformulate the definition of the chain complex (M  N. or B = J . we need to make sure that this number is finite.20. w2 ). see [LOT10a. the (v2 .3. see [LOT10a.9]. Then the graph representing FπD (N ) is obtained by replacing all algebra elements by their images under π. this means that there are only finitely many labels. Let M be a (right) type D structure and N a (left) type A structure over the same dg algebra A over a ring I of idempotents. the pairing is defined in the same way. B and π : B → A be as in definition 1. non-injective homomorphisms J → I also induce functors FπD : Cx∗ (Mat(CBD )) → Cx∗ (Mat(CAD )) between the categories of (curved) type D structures. Given two such pairs (v1 . Observation 1. type D structure or a bimodule. definition 2. If we start with a type AD or type DD bimodule M and a type AA or type AD bimodule N . lemma 2. this means that there are no loops in its graph. all modulo 2. Similarly. Similarly. together with a fixed basis of I and A. Is there a way to see that the statements in the previous remark are ”obvious“? D and C A ? How do I interpret the pairing on the level of the categories CA A . This induces a basis on the image of π. All we need to change in its construction is to add a zero-object to CA D .19 (pairing type D and type A structures). Choose a basis of the kernel of π and extend it to B. .J and π is the inclusion. w1 ) is equal to the number of sequences s of labels of consecutive arrows along a path from v1 to v2 in M such that s agrees with a label on the arrow from w1 to w2 in N . either J = I and π is a quotient map.3.10]. Let J . except that we need to record the labels for the remaining type A or type D sides: the first component is equal to the product (ie algebra product or concatenation) of the first components of the arrows along the corresponding path from v1 to v2 in M . A. however. The generators of M  N are defined by pairs of vertices in M and N labelled by the same idempotents. as long as we pair type A sides with type D sides and left structures with right structures.4] and [LOT08. In the previous definition. Furthermore. Note.

. (Z. Proof. b c a d Lemma 1.24. ε2 ) e and suppose it has the form shown in figure 7. A. ζ + bgc) is indeed an object of Cx∗ (Mat(C)): (ζ + bgc)2 + ∂(ζ + bgc) = ζζ + ζbgc + bgcζ + bgcbgc + ∂(ζ) + ∂(b)gc + b∂(g)c + bg∂(c) = ζζ + (ζb + ∂(b))gc + bg(cζ + ∂(c)) + ∂(ζ) + b∂(g)c + bgcbgc = ζζ + (bε1 + df )gc + bg(ε2 c + f a) + ∂(ζ) + b∂(g)c + bgcbgc = ζζ + dc + ba + ∂(ζ) + ( b (D(g) (((+ c = ∗. ε2 ). . Next. w2 ) on the left-hand side is equal to the number of all sequences b = (bn . . we consider the two chain maps F : (Z.12 CLAUDIUS BODO ZIBROWIUS Theorem 1. ε2 ) and f is the identity map. we observe that gcbg = gD(f )g = D(gf g) + D(g)f g + gf D(g) = D(g). let us check that (Z.3. f ject of Cx∗ (Mat(C)) for some differential graded category C (Y1 . ε1 ) = (Y2 . Let M be a (right) type D structure over B and N a (left) type A structure over A. . On the level of generators this identity is clear. . . w1 ) to (v2 . . For bimodules. . fix a label a = (am . . Let J . a1 ) from a vertex w1 to w2 in N . We usually apply this lemma to the case where (Y1 . FπD (M ) and FπA (N ) only possess generators belonging to idempotents in J . for lemma 1. We now state and (Z. δ) be an ob.23 (Cancellation Lemma). . w1 ) to (v2 .16. This follows almost tautologically from the interpretation of the induced functors in terms of oriented labelled graphs in observation 1. Then we have an identification FπD (M )A  A N ∼ = M B  B FπA (N ). The label a in N corresponds to labels b0 = (b0n . The contribution of a to the differential from (v1 . . ζ + bgc) → (X. . ζ) prove the two central lemmas mentioned at the beginning of this section. δ) → (Z. Next. M . note that the functors only act on one component of the morphism spaces.18.  1. Let (X.23 Remark 1. ζ) ∈ ob(Cxpre (Mat(C))) and f is an isomorphism Figure 7. The same holds true for bimodules of various types. ζ +bgc).22 (Pairing adjunction). b01 ) from w1 to w2 in FπA (N ) such that π(b0 ) := (π(b0n ). ( (( (gcbg) ( For the last step. Those labels contribute 1 to the differential from (v1 . π(b01 )) = a. since by definition. . . . . where (Y1 . δ) with inverse g. First of all. so only those generators of N that belong to idempotents in J survive the pairing. The object (X. ε1 ) (Y2 . b1 ) of labels from v1 to v2 such that π(b) := (π(bn ). . . (Y2 . I. δ) and G : (X. . π(b1 )) = a. ζ + bgc) . So the contributions agree. w2 ) on the right-hand side iff they agree with some b and do not contribute otherwise. Cancellation and Cleaning-up. ε1 ). Proof. δ) is chain homotopic to (Z. Then (X. along with two vertices v1 and v2 in M such that v1 and w1 belong to the same idempotent in J and so do v2 and w2 . . B and π : B → A be as in definition 1.

GF = 1Z and conversely. ε2 ) f f e e (Y1 . (O. PECULIAR MODULES FOR 4-ENDED TANGLES 13 1 1 (Z. where H is the homotopy given by the dashed line in figure 8c. for the second identity. . ε1 ) (c) The composition of F and G and the homotopy H (dashed arrow).  Lemma 1. Indeed. Now. ζ) (Z. Figure 8. Let (O. dO + D(h)). ζ) (Z. dO )) for which h2 . ζ) (Z. dO ) is chain homotopic to (O. ζ + bgc) c c bg d d b a b a gc (Y2 . dO ). ζ) c c bg d d b a b a (Y2 . ε2 ) gc (Y2 . (O. ζ + bgc) (Z. hD(h) and D(h)h vanish. ε1 ) (a) The chain map F (b) The chain map G 1 1 (Z. ζ + bgc) (Z. ε2 ) f f e e g (Y1 .25 (Clean-up Lemma). One easily checks that indeed D(F ) = 0 and D(G) = 0. the only non-trivial terms we need to compute are gc(ζ + bgc) + ε1 gc + ∂(gc) + a = g(cζ + ∂(c)) + (ε1 g + ∂(g))c + a + gcbgc = g(f a + ε2 c) + (ε1 g + ∂(g))c + a + gcbgc = (D(g) + gcbg) c = 0 for the first identity and similarly (ζ + bgc)bg + bgε2 + ∂(bg) + d = (ζb + ∂(b))g + b(gε2 + ∂(g)) + d + bgcbg = (df + bε1 )g + b(gε2 + ∂(g)) + d + bgcbg = b (D(g) + gcbg) = 0. ε1 ) (Y1 . respectively. dO ) be an object in Cx∗ (C) for some differential graded category C. ε1 ) (Y1 . Then for any morphism h ∈ Mor0 ((O.23 defined in figure 8a and 8b. ε2 ) (Y2 . Maps for the proof of lemma 1. it is not hard to check that F G = 1Z + D(H).

for the closed components. labelled a. together with a labelling of the arcs S 1 r im(T ) by some index set. dO + D(h)) is an object in Cx∗ (Mat(C)):     (dO + D(h))2 + ∂(dO + D(h)) = d2O + ∂(dO ) + dO D(h) + D(h)dO + ∂(D(h)) + D(h)D(h) The first term on the right gives (∗) and the last term vanishes. We also label these tangle components by variables t1 and t2 for the open components and t01 . We call those variables the colours of T . ∂ . b. . S 1 ⊂ S 2 = ∂M . . We call a choice of a single arc a site of the tangle T . A Heegaard diagram HT for a tangle T with n closed components in a Z-homology 3-ball with spherical boundary M is a tuple (Σg . which can be seen by applying the differential D to hD(h) = 0. dO ). d}. etc. .2.1. The chain isomorphisms between the two objects are given by 1+h 1+h (O. which are disjoint from αc and whose endpoints lie on Z. 1 b b Definition 2. b I qI q 2 b 3 such that the endpoints of the two intervals lie on a fixed circle b S 1 on the boundary of M . a c ball M with spherical boundary is an embedding  a    T: S 1 . and • β is a set of (g + n + 1) pairwise disjoint circles β1 . t02 . s3 for c and s4 for d. βg+n+1 on Σg . −3)-pretzel tangle in S sometimes it is more conventient to write s1 for a. Let us start by re- d calling the basic definitions from [Zib16. but Figure 9. . where • Σg is an oriented surface of genus g with 2n + 4 boundary components. the second for a and b. s2 for b. The invariant CFT∂ 2. which are partitioned into (n + 2) pairs. The images of the two intervals are called the open components.14 CLAUDIUS BODO ZIBROWIUS Proof. since hD(h) and D(h)h vanish. dO + D(h)) (O. • αa is a set of 4 pairwise disjoint arcs on Σg . α = αc ∪ αa . Z. 3 The (2. We number the ends of the open components such that the first endpoint is the point where the arcs d and a meet. β). c. Indeed. which can be seen by expanding D(h) = dO h + hdO + ∂(h) and using the fact that a term (∗) commutes with any morphism. . adapted 4 to 4-ended tangles. We can easily check that (O. The middle term also vanishes. . dO ) (O. Definition 2. . b.→ M.  2. . section 4]. We consider tangles up to ambient isotopy which keeps track of the labelling of the arcs. An orientation on a tangle is an choice of orientation on the two intervals and the circles. . Their composition is equal to 1 + h2 = 1. . • αc is a set of (g + n) pairwise disjoint circles α1 . . we usually use {a. denoted by Z. A 4-ended tangle T in a homology 3.1. Heegaard diagrams for tangles. αg+n on Σg . the images of any circles are called the closed components of the tangle. these two morphisms lie in the kernel of D. c. d. . dO + D(h)) and (O.

As in [Zib16].5. as illus- trated in figure 10: We collapse the four boundary components of Σ which meet the α-arcs. The difference between peculiar and ordinary Heegaard diagrams for tangles. and • stabilisation.6] are equivalent to the following moves for peculiar Heegaard diagrams: • isotopies of the α. see [Zib16. . The only reason for introducing peculiar Heegaard diagrams is to avoid any bordered Heegaard Floer theory. A peculiar Heegaard diagram for T is obtained from a tangle Heegaard diagram for T by a local modification around the punctures. and • αa is equal to S 1 r ν(∂T ) ⊂ ∂M . we obtain a multi-pointed Heegaard diagram. ie in the direction of the β-curves. Thus. PECULIAR MODULES FOR 4-ENDED TANGLES 15 back qi i pi front (a) An original tangle end (b) A new tangle end Figure 10. Definition 2. Remark 2. zj and wj . thereby joining the four α-arcs to a single α-circle S 1 . Remark 2. For each closed component. note that the number of α-circles and β-circles in a peculiar Heegaard diagram is the same. and connect these two points by an arc which intersects S 1 exactly once and no other curve. so the proof of invariance of the algebraic structures we are about to define is a minor adaptation of the one for link Floer homology. • handleslides of β-curves over β-curves and handleslides of α-curves over α-curves other than S 1 . and each tangle component belongs to exactly one such pair.4. we also contract the corresponding boundary components to points zj and wj . definition 4. It is obvious that we can go from a peculiar Heegaard diagram back to an ordinary tangle Heegaard diagram. As in [Zib16]. our convention on the orientation of the Heegaard surface is that its normal vector (determined using the right-hand rule) points in the positive direction. The former are obtained from the latter by local changes near the boundary of the Heegaard surface. we need to restrict ourselves to admissible dia- grams. the Heegaard moves from [Zib16. we usually draw the Heegaard surfaces such that the normal vector points into the projection plane. qi . Then we add a marked point for each tangle end on either side of S 1 . In particular. whose underlying Heegaard surface is now closed and carries basepoints pi . pi on the front and qi on the back. Obviously. However. Let T be a 4-ended tangle. ie diagrams whose non-zero periodic domains avoiding all basepoints have both negative and positive multiplicities. We impose the following condition on the data above: The 3-manifold obtained by attaching 2-handles to Σg × [0. lemma 4.and β-curves away from the marked points and the connecting arcs.3.7]. 1] along αc × {0} and β × {1} is equal to the tangle complement M r ν(T ) such that under this identification. • each pair of circles in Z is a pair of meridional circles for the same tangle component.

where s is a site of the tangle T . 4} into pairs.4 i=1 When comparing this to [Zib16. i=1. Given a peculiar Heegaard diagram HT for a 4-ended tangle T . the outward pointing end second.6. definition 4. For any two generators x and y of T. one considers the space of domains from x to y. see also [OS05. y) denotes the domains avoiding all basepoints and # M(φ) R denotes the number of points in the quotient of the 1-dimensional moduli spaces M(φ) by a natural R-action. section 5. Then. 2.3. The tangle Floer homology HFT(T [ ) is defined to be the homology of CFT(T [ ). adapted to 4-ended tangles. 3. For each element φ ∈ π2 (x. there . Let us briefly recall the definition of the tangle Floer homology HFT(T [ ) from [Zib16. wj . We partition T into four sets Ts . o2 }} of {1.9. one can define a moduli space M(φ) of holomorphic curves in Σ × [0. An ordered matching is a matching in which the pairs are ordered. 3. one may count holomorphic discs in Symg+n+1 (Σg ). This comes from the notation used in [AAEKO] in the context of the wrapped Fukaya category of the n-punctured sphere. zj . such that each α.2]. we order each pair of points such that the inward pointing end comes first. Definition 2. the one labelled t2 . let x(φ) denote the multiplicity of φ at x. depending on which of the former four arcs of the special α-circle S 1 are occupied. let T := T(HT ) be the set of tuples of intersection points. for every component t of the tangle.2. denoted by π2 (x. Furthermore. {i2 .19]. 2.y) R 2 µ(φ)=1   where π20 (x. (Alternatively. however. this gives the same theory. y) between two generators x and y in T and a basepoint x = pi . The differential in CFT(T [ ) is defined as X X   M(φ) ∂x = # y.8. 4.7. Given a domain φ ∈ π2 (x. y). the second consists of the two endpoints of the second open component of T . qi . A 4-ended tangle T gives rise to a matching PT as follows: The first pair consists of the two endpoints of the open component with colour t1 . Definition 2. Then. 1] × R representing φ.) The dimension of this moduli space is given by the Maslov index µ(φ). Given an orientation of the two open components of T . note that we now use peculiar Heegaard dia- grams to compute the Maslov index µ. let Ii = F2 [ιi ]/(ι2i = ιi ) and I ∂ := I1 × I2 × I3 × I4 be the ring of idempotents. sec- tion 5]. also called generators. y).and each β-curve is occupied by exactly one point. by the main result of [Lip05]. Definition 2. For i = 1.ιi be the F2 -vector space freely generated by elements in Tsi . with one idempotent for each of the four sites of a 4-ended tangle.16 CLAUDIUS BODO ZIBROWIUS Remark 2. o1 }. A matching P is a partition {{i1 . we define three gradings on the generators of CFT(T [ ): The δ-grading δ is a relative 1 Z-grading and defined by 2 X n X 1 δ(y) − δ(x) = µ(φ) − 2 (pi (φ) + qi (φ)) − (zi (φ) + wi (φ)). a reference to the labels pi and qi we have chosen for the marked points. The attribute “peculiar” should be considered as a homophone of “p-q-lier”. y φ∈π 0 (x. we define a right I ∂ -module CFT(T [ ) by letting CFT(T [ ).

| {z } | {z } | {z } Rpre n Rpre n Rpre n We define a 12 Z-grading on pre An . and then extend linearly to Apre n . by setting 1 δ(ιi ) := 0. and then extending linearly to all of Apre n . . let Rpre n be the free polynomial ring generated by the variables pi . s ) of sites. n.11. where i = 1. by 1 Atk (ιs ) := 0.1.qi . PECULIAR MODULES FOR 4-ENDED TANGLES 17 is a relative Z-grading At . n. An . restricts to the multiplication map 0 00 in Rpre n : ιs00 . . a relative Z-grading. like a+1 for the single-variate or a( 2 . 4 and j = 1. qi := ιi .1. . Similarly. o1 }. n. .ιi . At0k (Uk ) := 1 and At0k (Vk ) := −1. for all triples (s. .pi qi . δ(pi ) = δ(qi ) := 2 and δ(Uj ) = δ(Vj ) := 1. 3. the reduced Alexander grading Ar .ιs0 ⊗I ∂ ιs0 . Peculiar algebras. which is called the Alexander grading.ιs → ιs00 . {i2 .9. . Ui := 1. . we define Atk for k = 1. . we obtain a relative Z-grading. we define relative 2 Z-gradings Atk for k = 1.1 and Vj0 := 1. the homological grading h.ιs := Rn for pairs (s. . Uj0 := 1. For each n ≥ 0. o2 }}. and similarly Alexander gradings At0k for k = 1.ιi−1 . By taking the sum of all Alexander gradings. 2. We sometimes denote the Alexander grading on generators by a superscript list of integers (or half-integers. given 1 an ordered matching P = {{i1 . s . Finally. . we want to achieve the same symmetry present in the decategorified 3 1 invariants [Zib16]). n by At0k (ιs ) := 0. 2. Definition 2. Given a ordered matching P = {{i1 . 2. . qi and Uj and Vj . is defined as h = 12 Ar − δ. .2.Vj0 . . . if eg. called Alexander gradings. An .14].ιs . o2 }}.pi . where i = 1. . s ) of 0 sites and whose algebra multiplication is defined by the unique I ∂ -I ∂ -bimodule homomorphism Apre pre pre n ⊗I ∂ An → An which.Uj0 . For n ≥ 0. 4 and j = 1. called the δ-grading. we define A− n to be the subalgebra of An generated as an algebra by the following elements: pi := ιi−1 . .10. These gradings give rise to a reduced Alexander grading and a homological grading as in definition 2.− 2 ) for the multivariate Alexander grading. . o1 }. 3. Apre pre pre n . we define At0k by At0k (y) − At0k (x) := At0k (φ) := wk (φ) − zk (φ). Atk (pik ) = Atk (qik ) := 2 and Atk (pok ) = Atk (qok ) := − 21 . Let Apre n be the I ∂ -I ∂ -algebra whose underlying I ∂ -I ∂ -bimodule structure is given by ιs0 Apre n . 2 by Atk (y) − Atk (x) := Atk (φ) := pok (φ) + qok (φ) − pik (φ) − qik (φ). 2. For k = 1. see [Zib16. definition 4. {i2 . pre Definition 2.

which was already defined in [Zib17]. where again. where µ denotes composition in A. we set p = p1 + p2 + p3 + p4 ∈ A− n and q = q1 + q2 + q3 + q4 ∈ A− n. Let A∂ be the quotient of A− n by the re. j = 1. the resulting complex does ∂ not satisfy the relation ∂ 2 = 0. qj+1 qj . For most part of this paper. Furthermore. 3. . where r ∈ Rn is a monomial. However.3. p2 p3 lations Ui = 0. 4 modulo 4 with an offset of 1. . Thus any element in A− 0 0 n can be uniquely written as a sum of elements of the form pi pi+1 . Let I be a ring of idempotents and A a Z-graded algebra over I. . q1 q4 p1 p4 ι1 • • ι3 Definition 2. the various gradings on Apre restrict to well-defined gradings on this subalgebra. δN ) is an I-module homomorphism M → N ⊗I A. As abstract algebraic structures. n and where 1 = ι1 + ι2 + ι3 + ι4 is the identity in I ∂ . n. we sometimes write the elements pi pi+1 . A morphism between two curved type D structures (M. . we obtain a slightly modified ∂ 2 -relation which enables us to promote CFT [ to more sophisticated homological invariants.9. Let us recall this definition here in slightly more down-to-earth terms.18 CLAUDIUS BODO ZIBROWIUS where we take the indices i = 1. For two such morphisms f and g. This the standard basis on A− n as a vector space over F2 . . We claim that we also obtain a tangle invariant if we add those curves to our differential. Also fix a central element ac ∈ Z(A) of degree −2. we will only be concerned with ι4 • a certain quotient of A− n . .r and qi qi−1 . . namely certain curved type D modules. recording their multiplicities at the basepoints by elements of the algebras A− n or A . . we take the indices modulo 4 with an offset of 1. [ we only consider holomorphic curves which stay away from the basepoints in our peculiar Heegaard diagrams. For the differential in CFT. Note that A− n is a faithful Rn -module. δM ) and (N. pj−1 pj as pi(i+1)···(j−1)j and qi qi−1 . 2.13. pj−1 pj . Definition 2. Peculiar modules. where Rn is the free polynomial ring in the variables Ui . so we can write for example p4 = p1234 + p2341 + p3412 + p4123 . . . Figure 11. . . alternative definition of A∂ 2. see also example 4. 2.12. . We call A− n the generalised peculiar algebra. Uj and Vj for i = 1. we defined curved type D modules in example 1. For con- venience. A (right) curved type D structure over A is a Z-graded I-module M together with a (right) I-module homomorphism δ : M → M ⊗I A of degree −1 satisfying (1M ⊗ µ) ◦ (δ ⊗ 1A ) ◦ δ = 1M ⊗ ac . to simplify notation.8 and figure 23. . . 4 and j = 1. This algebra can be q2 q3 • interpreted as the path algebra of the quiver in figure 11 with ι2 relations pi qi = 0 = qi pi . Instead. 3. . Uj0 = 0 and Vj0 = 0. We call ac the curvature of M . their composition is defined as (g ◦ f ) = (1M ⊗ µ2 ) ◦ (g ⊗ 1A ) ◦ f. . qj+1 qj as qi(i−1)···(j+1)j . Naturally.r. The quiver for an We call A∂ the peculiar algebra.

y). let gpqModP.15. giving us the usual notions of chain homotopy and homotopy equivalence. idM1 . Definition 2. pi i · qi i · Uj j · Vj . see section 4. respectively. the category of ordinary chain complexes over F2 . we can regard d0 and d1 as maps d0 : M0 → M1 ⊗I A and d1 : M1 → M0 ⊗I A.n be the category of δ-graded (and Alexander graded) curved complexes over A− n with curvature (3) p4 + q 4 + Ui1 Uo1 + Ui2 Uo2 . Furthermore. This gives us an enriched category over Com. N ) with a differential D defined by D(f ) = δN ◦ f + f ◦ δM . o2 }}. y∈T φ∈π2 (x.n whose underlying right I ∂ -module agrees with CFT(T [ ). see definition 1. In general. o1 }. which is an easy corollary of the classification in terms of immersed curves on the 4-punctured sphere. M ) in gpqModPT .3. since we have chosen ac to be central. This is for the simple reason that the total number of generators can be odd. Then (M0 ⊕ M1 . If M0 and M1 denote the k-vector spaces generated by an A-basis of M0 and M1 . It is interesting to compare curved type D structures to matrix factorisations as studied by Khovanov-Rozansky [KR04]. We call the objects of this category generalised peculiar modules. {i2 . in general. Given an (ordered) matching P = {{i1 . However. a matrix factorisation of a potential w ∈ A consists of two free A-modules M0 and M1 with two maps d0 (2) M0 M1 d1 such that d1 d0 = w.2. such splittings exist. i=1. Then indeed D2 = 0. The underlying ordinary category is obtained by restricting the morphism spaces to degree 0 elements in the kernel of D.3. We call the objects of this category peculiar modules. Definition 2. admit a splitting of the form (2). the differential ∂ on CFT− (T ) is defined by X X   M(φ) (4) ∂x = # · y ⊗I ∂ a(φ). see for example [HRW16.2 and example 1. Given a 4-ended tangle T with n closed components in a Z-homology 3-ball M with spherical boundary and an (admissible) peculiar Heegaard diagram for T . 5]. For example curved complexes associated with manifolds with torus boundary do not. let use define a generalised peculiar module CFT− (T ) := CFT− (T.ιs(x) . a(φ) is the preimage of Y n Y p (φ) q (φ) z (φ) wj (φ) ιs(y) . idM0 and d0 d1 = w. Remark 2.y) R µ(φ)=1 where for φ ∈ π2 (x. 9 fig. we cannot go in the other direction.14. Given an algebra A over some field k. PECULIAR MODULES FOR 4-ENDED TANGLES 19 We endow the space of morphisms Mor(M. let pqMod be the category of δ-graded (and Alexander graded) curved complexes over A∂ with curvature p4 + q 4 . p. d0 + d1 ) defines a curved type D structure over the k-algebra A.16.4 j=1 . for the curved type D structure invariants of tangles. However.

Proof. the grading on a tensor product is given by the sum of the gradings of the tensor factors. lemma 4. CFT− (T ) is indeed a well-defined generalised peculiar module. The proof is essentially the same as in link Floer homology. one can promote both Alexander and δ-gradings to absolute gradings via symmetries and the spectral sequence to HF(S d 3 ) [OS05]. We call (CFT (T ). Since CFT− is finitely generated. which we call the peculiar ∂ ∂ module of T . For each x ∈ CFT− .22. a(φ) lies in the image of inclusion map A− n .→ An . Hence CFT∂ (T ) is a well- defined a peculiar module. This follows from the observation that ∂φ intersected with the α-circle S 1 is a path on S 1 connecting the two points of x and y on S 1 .n → pqMod induced by the quotient map A− n → A is denoted by CFT (T ). Note that by the previous lemma. We can now argue as in the proof of [OS01. ∂) the generalised peculiar module − of T .) Proof.18. Furthermore. too. its bigraded chain homotopy type is an invariant of the tangle T .20. the difference of the δ-gradings of x and y determines the δ-grading of a(φ). So let us also fix the multiplicities of φ at the basepoints. (As usual.21. see [OS05. y).20 CLAUDIUS BODO ZIBROWIUS pre under the inclusion map A−n . Thus. Computations suggest that their invariant for one-sided 4-ended tangles is closely related to CFT− . pre Lemma 2.  Lemma 2. [ Remark 2. it might also be interesting to set up their theory for an odd number of tangle strands and then compare CFT− to their invariants of (1. one could simply fix absolute gradings on a specific test tangle. 3)-tangles.19. φ ∈ π2 (x. Alternatively. I expect that something similar can be done for our tangle invariants.  . Its image under the functor gpqModPT . using admissibility of the underlying Heegaard diagram. ∂ increases the δ-grading by 1 and preserves the Alexander grading. So. lemma 4. On link Floer homology.17. The generalized algebra A− n and the generalised invariant CFT are inspired − by Ozsváth and Szabó’s algebra and tangle invariant from [OS17]. we are only working with the relative gradings on CFT and CFT− inherited from those on CFT ∂ [ which were defined in [Zib16].13]. in this paper. Remark 2. Conceptually.2].→ An . there are only finitely many choices for the coefficients a(φ). Both statements follows directly from the definitions of the gradings of generators and algebra elements.44) and the absolute gradings on HFL.  Lemma 2. Proof. Theorem 2. it is sufficient to show that the coefficient of each y ∈ CFT− is a finite sum. throughout this paper. and then define absolute gradings on all other tangles via the pairing with this test tangle (using theorem 4. whose bigraded chain homotopy type is also an invariant of the tangle T . [ However. the sum on the right-hand side of (4) is finite. say a trivial tangle. all gradings on CFL should be regarded as relative.

25 (rational tangles). we can follow [OS05. Conversely. Checking the ∂ 2 -identity is analogous to the link case. By tightening the β-curve. since we only − allow handleslides of α-curves over α-curves other than S 1 . we obtain invariance as a (curved) type D module over the free polynomial ring Rpre n . so the remaining curve has to be constant. they contribute at least 2 to the Maslov index. and a pseudo-holomorphic curve ψ of Maslov index µ(ψ) = 2 and with non-negative multiplicities. However. Thus. (In particular. we only count bigons in the differential of the peculiar invariant. Then its Maslov index µ(φ) is equal to twice the sum of the coefficients. If there are boundary degenerations. equipped with a set α of (g + r) attaching circles. The remaining two α- injective boundary degenerations contribute the first two terms of (3) and the remaining two β-injective boundary degenerations contribute the last two terms. So in this case. For closed tangle components.3] and count ends of moduli spaces of Maslov index 2 curves. z) with µ(φ) = 2 and a(φ) = a for some fixed a ∈ A− n. where φ varies over those curves in π2 (x.  Fact 2. the β-curve can be read off from CFT∂ (T ). Then CFT∂ (T ) can be read off from this single β-curve as follows: the vertices of the graph of CFT∂ (T ) correspond to intersection points of this β-curve with the α-arcs. The same holds for β-injective boundary degenerations.5] Given a surface Σ of genus g. there is an even number of ends. hence x = z. we might actually view the β-curve as the invariant associated with the tangle. [OS05.  Example 2.5]. then by fact 2. More precisely. proof of lemma 4. this arrow pair is labelled by powers of p if the component goes via the front component of the 4-punctured sphere minus the α-arcs and by powers of q otherwise.24. example 4. Its arrows come in pairs labelled by powers of p or q. We fix two generators x and z and consider the disjoint union of moduli spaces M(φ). write φ as a linear combination of connected components of Σ r α. All other ends appear in pairs again. so the z ⊗ a-component of ∂ 2 x vanishes. As shown in [Zib16. we obtain an arrow pair connecting the vertices corresponding to the ends of this component. ie that every bigon covers some pi or qi . The same holds for β-injective boundary degenerations. Fact 2. we get a boundary degeneration for each component of Σ r α and Σ r β. so their contributions cancel. we can assume that there are no honest differentials in CFT∂ (T ). Figure 12 shows the peculiar modules of some very simple 4- ended tangles. By fact 2. . Thus. then ψ is equal to one of the connected components of Σ r α and the number of pseudo-holomorphic boundary degenerations in the homology class of ψ is odd. where r ≥ 1. for each component of the β-curve minus the α-arcs.) If there are no boundary degenerations.23 above.23.24. PECULIAR MODULES FOR 4-ENDED TANGLES 21 In the following. this fixes the multiplicities of φ at the basepoints. we need two analytical facts from [OS05]. However.4] Given a homology class φ of an α-injective boundary degen- eration. [OS05. It remains to show that the peculiar module is an invariant of the tangle T .  Proof of theorem 2. CFT− (T ) is essentially the chain complex associated with a multi-pointed Heegaard diagram in Heegaard Floer theory. these boundary degenerations come in pairs which cancel each other. lemma 5. every rational tangle T has a tangle Heegaard diagram with just a single β-curve. the same proof also works if we work over An . theorem 5. and only those that do not occupy both pi and qj .16.

We obtain a complex in which every arrow is paired with another one going in the opposite direction and every generator is connected along the arrows to exactly two other generators – just as for rational tangles! A schematic picture of this complex is shown in figure 14a. the complex is homotopic to the invariant consisting of the solid arrows only. which works surprisingly well. those could point in both directions. 2 ) δ − 2 b( 2 . 2 ) δ 0 a( 2 . respectively to the right. such that we obtain a collection of loops. we will compute the peculiar modules for more general pretzel tangles. so we can only get arrows in one direction. Next. −3)-pretzel tangle).25 for curved type D modules). But it is easy to see that we can homotope those dotted arrows away (using the clean-up lemma 1.0) δ 0 d(0. the connecting domains in one direction either have negative multiplicities or occupy both pi s and qi s. In subsection 5. by 1.3.0) q2 q341 q123 q4 p2 p143 p321 p4 p21 +q34 1 1 1 p214 1 1 1 1 1 q412 1 1 δ 2 b(− 2 . where the generators in each of its fields have the same Alexander grading and where moving one field down.26 (the (2. The superscripts of the generators specify the Alexander grading. we compute everything from definition.23. their Heegaard diagrams and peculiar modules. But there are also other contributing domains. First. we can deduce that all solid arrows contribute. Figure 13 shows the computation of CFT∂ (T ) for the (2. We can then apply the cancellation lemma 1. increases the Alexander grading corresponding to the colour t1 . Those correspond to the labelled arrows in figure 13c. using some more advanced methods. −3) pretzel tangle from figure 9. respectively t2 . in each case. From this and the ∂ 2 -relation. which we develop in section 4 as a corollary of the general classification of peculiar modules. They are ordered according to their Al- exander grading on an infinite chessboard. so in any case. . we compute the generators of the complex.− 2 ) p432 q234 p43 +q12 δ 0 b(0.− 2 ) δ 2 d( 2 . However. Basic rational tangles. where these arrow pairs have been replaced by single unoriented edges.− 2 ) δ 0 c(− 2 . In principle. There are only eight remaining arrows (the dotted ones) and they can only appear in pairs. 2 ) δ 0 c(− 2 .22 CLAUDIUS BODO ZIBROWIUS d b d b d b a b b c a b b c b b b b b b 1 1 p1 1 1 1 1 1 q1 1 1 1 δ 0 a( 2 .− 2 ) δ − 2 d(− 2 . we compute bigons and squares. 2 ) p3 q3 Figure 12. Example 2. Grading constraints tell us that we can only get additional morph- isms between those generators which are connected by the other arrows. here.

26 23 . A computation of a peculiar invariant for a non-rational tangle. −3)-pretzel tangle (b) A Heegaard diagram for the tangle on the left p2143 (46) p4 p214 (4) p214 (46) 3 3 q23 (1345) 3 δ −2 x2 d0 δ − 2 x2 c1 δ −1 by1 δ − 2 x2 c2 δ −1 by2 δ − 2 x2 c3 δ −2 x2 b0 δ −1 by3 q41 (1236) q3 q2 (126) p3 (26) q341 (133455) p3 (2) 1(2) q2 q234 (15) q1 (3) q41 (35) q41 (3) q2 q2 q23 (15) q23 (1) 3 3 3 3 3 3 δ − 2 a1 y1 δ − 2 a1 y2 δ − 2 a2 y1 δ − 2 a1 y3 δ − 2 a2 y2 δ − 2 a2 y3 p14 (4) p1 p14 (46) p1 p2 (2) p143 (46) PECULIAR MODULES FOR 4-ENDED TANGLES p1 p32 (26) p32 (2) 1(3) p1 (345) p4 q4 (3) p432 (122466) q4 (35) 3 δ −1 dy1 δ −2 x d0 δ − 32 −1 δ − 32 p32 (2345) 1 x1 c1 δ dy2 p14 (1246) x1 c2 δ −1 dy3 δ − 2 x1 c3 δ −2 x1 b0 q123 (15) q123 (1) q3 q1234 (15) (c) The peculiar invariant for the pretzel tangle above Figure 13. q1 q4 d d′ b b d t2 p1 p4 t1 3 3 b b 4 x1 y1 4 b b a1 6 c1 b a c p2 y2 b b b a2 x2 y3 c2 p3 b b c3 2 2 q3 b b′ b b q2 1 t1 b t2 1 5 (a) A tangle diagram for the (2. see example 2.

p3 . they are just a convenient way to see certain symmetries. respectively. In CFL(L). arranged according to their Alexander grading and labelled by their δ-grading. a T c ure 15. The dotted edges correspond to pairs of arrows labelled by powers of q. The meaning of both of these representations as loops will be discussed in section 4. p2 . Let T be a 4-ended tangle and L the link obtained by closing T at the sites a and c as shown in fig. [ we only . For the moment. q1 . the loops have been transferred onto separate 4-punctured spheres in such a way that the vertices lie on the four arcs that connect the punctures and the unoriented edges lie on the front or back of the spheres depending on whether they correspond to arrow pairs labelled by powers of p or q. The subfigures (b)–(d) show the three loops from (a) separately on 4-punctured spheres. depending on whether L has two or one the closure of a tangle T closed component(s) more than T . respectively. Then CFL(L) [ = F1 (CFT∂ (T )) = ⊗Vi ∼ ∼ F2 (CFT∂ (T )) Figure 15. q1 . q4 . p1 . Generators correspond to vertices. and where F1 and F2 are the functors induced by the quotient maps f1 : A∂ → F2 . The final result of the computation from example 2. 7→ 0. q4 . q3 . we obtain a similar formula by a cyclic permutation of the indices. p1 .24 CLAUDIUS BODO ZIBROWIUS −2 − 23 −1 − 23 −1 − 23 b b b b b b − 32 − 32 − 23 − 32 − 23 − 32 b b b b b b b b b b b b − 32 −1 − 32 −1 − 32 −2 (a) A schematic picture of the result. 7→ 1. p2 . For the other two opposite sites.26 and figure 13. p3 . A link obtained as where i = 0 or 1. In figures 14b-d. we obtain a Heegaard diagram for L. If we delete those basepoints in a peculiar Heegaard diagram of T that correspond to the variables that we set equal to 1.27. q2 . b b b b b b b b b b b b b b b b b b (b) (c) (d) Figure 14. q3 . p4 7→ 0 f2 : A∂ → F2 . Proposition 2. q2 . the solid ones to pairs of arrows labelled by powers of p. p4 7→ 1. Proof.

30. −Z is defined to be the same arc diagram. a an even number of points on Z and M a matching of points in a. Mα ) and Zβ = (Zβ . M ).1. Also. We will assume some famili- arity with [Zar09. In [OS02]. aβ .28. where • Y is a sutured manifold with sutures Γ.21. more precisely. we compute the Heegaard Floer homology of the link L obtained by glueing T1 to T2 . φβ ). is the bigraded chain homotopy type of CFT∂ (T ) determined by ∇sT ? 3.31. where Z is a (possibly empty) set of oriented line segments. Review of bordered sutured Heegaard Floer theory. Mβ ) are arc diagrams. One can obtain the minus version of link Floer homology from the generalised peculiar modules from the same idea as in the proof of the previous proposition.  Remark 2. except that the orientation of the line segments is reversed. 3.and β-arcs is a tuple (Y. An arc diagram Z is a triple (Z.29. . [ using the generalised clock theorem [Zib15. but this is only needed to fix the absolute grading). So let us start by recalling Zarev’s Heegaard Floer theory for bordered sutured manifolds. A bordered sutured manifold with α. Observation 2. CFT∂ (T ) decomposes into the direct sum over the Alexander gradings of closed tangle components. The proof generalises immediately to HFT. By definition. Pairing 4-ended tangles In this section. The proof is essentially a calculation of a bordered sutured type AA bimodule. Zα . theorem 1.13].1. the signature. • Zα = (Zα . a. aα . Zar10.2. Γ. Observation 2. Ozsváth and Szabó showed that the link Floer homology of alternating links is completely determined by the Alexander polynomial (and. The graph G(Z) of an arc diagram Z is the graph obtained from the line segments Z by adding an edge between matched points in a. Given an alternating tangle T . • φα is an embedding of G(Zα ) into the closure of R− such that φα (Zα ) ⊂ Γ. dividing ∂Y into two oriented open surfaces-with-boundary R− and R+ such that ∂R− is equal to Γ as embedded oriented 1-manifolds. Given an arc diagram Z. we prove a glueing formula for CFT∂ : Given the peculiar modules of two 4-ended tangles T1 and T2 . Question 2. Zar11] and only give a short review of the basic geometric objects involved. Y is an oriented manifold and Γ ⊂ ∂Y are embedded oriented simple closed curves. φα . Definition 3. the differential of a peculiar module preserves the Alexander grading by lemma 2. Zβ . so we need to set those algebra elements equal to 0. Thus. to be precise. We end this section with some simple observations about CFT∂ . the Alexander grading corresponding to a closed tangle component vanishes on A∂ . up to at most one stabilisation. Definition 3. PECULIAR MODULES FOR 4-ENDED TANGLES 25 count those holomorphic curves that stay away from the remaining basepoints.

The following theorem summarizes his theory for bimodules. bordered by two arc diagrams −Z1 and Z2 . such that the map (5) π0 (Γ r (φα (Zα ) ∪ φβ (Zβ ))) → π0 (∂Y r (im(φα ) ∪ im(φβ ))) is surjective. proposition 3. in particular examples 1. consider a Heegaard diagram of the underlying sutured manifold. Definition 3.9 and 1.26 CLAUDIUS BODO ZIBROWIUS • φβ is an embedding of G(Zβ ) into the closure of R+ such that φβ (Zβ ) ⊂ Γ and φα (Zα ) ∩ φβ (Zβ ) = ∅. and let Y1 ∪Z2 Y2 be the 3-manifold obtained by glueing Y1 and Y2 together along tubular neigh- bourhoods of the images of Z2 on ∂Y1 . Each arc diagram can either play the role of a type D or type A side of the bimodule. For a discussion of type A and D structures as algebraic objects. This gives us an embedding of G(Zα ) into the Heegaard surface such that its image does not intersect the α- curves. theorem 12. and −Z2 and Z3 . it is true for general multimodules. Zarev defines a moving strands algebra A(Z).6]. Zarev defines various bimodules over the strands algebras corres- ponding to its arc diagrams. one obtains multimodules. which is phrased in terms of the - tensor product (see definition 1. Given an arc diagram Z. We proceed similarly for G(Zβ ) in R+ and the β-curves. simply by sliding them off those 2-handles. a basepoint. Then there are homotopy . respectively. If we drop this condition. we allow the sutured surfaces of bordered sutured manifolds to be degenerate in the sense that surgery along all edges between matched points may contain closed components. respectively −Z2 on ∂Y2 . However.19).10] or [Zar11. A central result of Zarev’s work is a pairing theorem. Let Y be a bordered sutured manifold. see [Zar09. in which case. Remark 3. this is true for each connected component of an arc diagram.3.3. If we restrict to non-degenerate sutured surfaces. We view the images of the edges connecting points in a as α-arcs. ie the sutures. Then we can embed the graph G(Zα ) into R− in such a way that it misses the 2-handles corresponding to the α-curves. Zarev’s invariants fail to be well-defined in general.5 ([Zar09. bordered by −Z1 and Z2 . Theorem 3. To be more precise. in fact. This allows us to consider more general bordered sutured manifolds. well defined up to homotopy equivalence: A(Z1 ) A(Z1 ) BSAA(Y )A(Z2 ) BSDA(Y \ \ )A(Z ) 2 A(Z1 ) BSAD(Y \ )A(Z2 ) A(Z1 ) \ )A(Z2 ) BSDD(Y Let Y1 and Y2 be two such manifolds. which we usually draw in green. We put a marked point.2]).4. in every open component of the boundary minus the image of Z. homological linear independence is automatically satisfied. A Heegaard diagram of a bordered sutured manifold is obtained from a Heegaard diagram of the underlying sutured manifold by adding the graphs of the arc diagrams to it. Condition (5) is called homological linear independence.13. theorem 3. we refer the reader to section 1. Then there are bimodules. Note that unlike Zarev. and given a bordered sutured Heegaard diagram. The image of Z lies on the boundary of the Heegaard diagram.

6. The algebra elements in the first and second components are also coloured red and blue. PECULIAR MODULES FOR 4-ENDED TANGLES 27 equivalences A(Z1 ) \ )A(Z ) A(Z2 ) BSDA(Y A(Z1 ) BSAA(Y1 ∪Z2 Y2 )A(Z3 ) ∼ BSAA(Y \ = \ )A(Z ) . Let P be the strictly a T1 c a T2 c unital type AA structure over A∂ defined in figure 17. for any given ordered matching. Similarly. The link obtained ander gradings t and t−1 and identical δ-gradings and where by glueing two 4-ended tangles i = |T1 | + |T2 | − |L| − 2 ∈ {0. Let T1 and T2 be two 4-ended tangles and L the link obtained by glueing them together such that their sites d d match up. The identity action is implicit.6. and the other is type D for A(Z2 ). 1}. A first glueing formula.Figure 16. This is the same convention that we use in the proof of theorem 3. there is a unique way to define a relative Alexander grading on P which is preserved by the differentials after inverting the algebra Alexander grading. where one is type A for A(Z2 ). . 2 3 etc. Theorem 3.2. Any combination of bimodules for Y1 and Y2 can be used. together along matching sites (q3 |p1 ) δ 0 ba δ 0 cd (q2 |−) (q23 |p1 ) (q34 |p1 ) (q4 |−) (q3 |p21 ) (q3 |p14 ) (−|p2 ) (−|p4 ) − 12 (p1 |q234 )+(q234 |p1 ) 1 δ aa δ 2 dd (p21 |q34 )+(q34 |p21 ) (p14 |q23 )+(q23 |p14 ) 1 1 δ − 2 bb δ 2 cc (p214 |q3 )+(q3 |p214 ) (−|q2 ) (−|q4 ) (p1 |q23 ) (p1 |q34 ) (p2 |−) (p21 |q3 ) (p14 |q3 ) (p4 |−) δ 0 ab δ 0 dc (p1 |q3 ) Figure 17. The idempotent of a generator xy is ιx in the first component and ιy in the second. The δ-grading of each generator is specified by the exponent of δ. as illustrated in figure 16. respectively. where we identify P with the bordered sutured type AA structure P 0 illustrated in figure 19c. The type AA structure P for theorem 3. Then b b CFL(L) [ ⊗ V i = CFT∂ (T1 )  P  CFT∂ (T2 ) where V is the 2-dimensional vector space supported in Alex.6. 2 3 A(Z1 ) \ )A(Z2 ) A(Z ) BSAA(Y A(Z1 ) BSDA(Y1 ∪Z2 Y2 )A(Z3 ) ∼ BSDD(Y \ = \ )A(Z ) . 3.

A decomposition of the complement of the link from figure 16 .28 CLAUDIUS BODO ZIBROWIUS d D a T1 c A T2 C b B (a) The parametrization on ∂XT1 (b) The parametrization on ∂XT2 D F d f A a c C E b e B (c) The parametrization on the boundary of the thickened 4-punctured sphere X Figure 18.

q3 /P1 .δ } bb {p4 . A Heegaard diagram for the bordered sutured manifold X from figure 18c and some computations of generators and domains .p1 /Q3 } {p4 .q3 /P1 } {q4 } {q3 /P1 .δ } {P2 } {P4 .δ } ab dc {p1 /Q3 } (c) The domains that contribute to the type AA structure P 0 Figure 19. The idempotent of a generator xy is ιx ∈ Iα0 and ιy ∈ Iβ0 .β } {q2 .Q2 .p1 /Q3 .P2 .q4 .p1 /Q3 . The orientation on the surface is such that the normal vector (de- termined by the right-hand rule) points out of the projection plane.P4 .Q4 .P4 .p1 /Q3 . aa : cCEβδ ab : bcACδ cc : aAF βδ cd : adACβ aa : bcBCδ ba : acBCδ cc : adADβ dc : acADβ aa : ceCβδ bb : acACδ cc : af Aβδ dd : acACβ (b) Generators in idempotents Iα0 and Iβ0 for the Heegaard diagram above.β .β }+{q2 .β } {q4 .Q2 .Q4 }+{q4 .δ } cc {Q2 .P2 } {q3 /P1 .β } {p1 /Q3 .P4 .β .q3 /P1 .δ } aa {p1 /Q3 .Q4 } {p2 } {p2 .β } {Q4 } {p1 /Q3 .Q2 .P2 } {p4 .δ }+{q2 .β } dd {p2 . PECULIAR MODULES FOR 4-ENDED TANGLES 29 f d δ δ p4 F b D b P1 δ q4 c b d P4 b p4 b p1 D Q4 f P4 b C F b c a C A E q2 a b p2 E b Q3 Q2 b q2 b B b A P2 β q3 β β b e Q2 e b b b B (a) A Heegaard diagram for X.q3 /P1 .p1 /Q3 .δ }+{q3 /P1 . {q3 /P1 } ba cd {q2 .p2 .q3 /P1 .δ } {p4 .

ιβ (BSAA(X)). f } and {β. respectively (see definition 1. respectively. generators of its type D structure belong to the idempotents that occupy one of the four arcs a. So let us compute FιAA α .and β-arcs are labelled by a.5) tells us that (6) SFC(XL ) ∼ [ T1 )A  A. e. the two square regions with vertices {d.Iα0 of A. So the goal is to identify the three tensor-factors on the right-hand sides with each other. In other words. the right hand side of (6) is equal to (7) BSD(X [ T1 )A0  A0 . Note however. we obtain the complement XL of the link from figure 16. There are two domains that have two labels. Thus. respectively. There are four pairs of regions that have the same label. F .B BSAA(X) = BSD(X \ [ T2 )B  BSD(X The left-hand side agrees with the left-hand term of the identity from theorem 3. observe that we can choose a Heegaard diagram for XT1 where the two α-arcs that have ends on the same suture do not intersect any β-curve. The regions adjacent to a basepoint are shaded light green (). Zarev’s glueing theorem (theorem 3.ιβ is the functor induced by the inclusions ια and ιβ . δ. \ ie those generators that occupy e.ιβ (BSAA(X)) \  BSD(X [ T2 )B0 where FιAA α .ιβ (BSAA(X)) \ first. by the pairing adjunction (theorem 1.ιβ (BSAA(X)) \ are shown in figure 19b. Note that each closed component of T1 and T2 carries two oppositely oriented meridional sutures. In addition to those sutures on closed components of T1 and T2 .22). B. Similarly define Iβ . Thus. First of all. C. we may use these labels to describe all domains that contribute to FιAA α . the sutured manifold XL carries one meridional suture at each of the four places where the tangles have been glued together. The strategy of the proof is to glue the tangle complements XT1 and XT2 to opposite sides of a thickened 4-punctured sphere X = I ×(S 2 r4D2 ). c or d. \ The generators of FιAA α . Let Iα0 be the subring of idempotents occupying the α-arcs e and f and ια : A0 . E and F . Pi .A.→ B. b. respectively. b. c. Moreover.30 CLAUDIUS BODO ZIBROWIUS Proof of theorem 3.16). Furthermore. In the following. D. the labels of the type D structures for XT1 and XT2 are contained in A0 . r∈D+ r∈D− . XT2 and X with the structures of bordered sutured manifolds specified by the arc diagrams in figure 18. we equip XT1 . Finally. Thus. D. By glueing XT1 to the inside and XT2 to the outside of the thickened 4-punctured sphere X as shown in figure 18c. namely q3 /P1 and p1 /Q3 . e.ιβ (BSAA(X)). α.→ A the subalgebra Iα0 . b. ie along {0}×(S 2 r4D2 ) and {1} × (S 2 r 4D2 ). that the coefficients of any paired regions agree in any domains connecting generators in FιAA α . Let A be the bordered sutured algebra corresponding to the arc diagram on X consisting of α-arcs and Iα the corresponding ring of idempotents. f . The same is true for XT2 and its β-arcs.6. as in figure 18c. let us write domains D as formal differences D+ − D− of unordered sets of regions D+ and D− with D+ ∩ D− = ∅ such that X X D= r− r.6. all other regions in the Heegaard diagram have at least one boundary component and are labelled by Qi . Intersection points are labelled by black Greek and Roman letters. A Heegaard diagram for X is shown in figure 19a. pi and qi . For this. Iβ0 and ιβ : B 0 . B} are labelled by δ and β .B0 FιAA α . f and A. d. E. BSD(X [ T1 ) and BSD(X[ T2 ) lie in the images of the functors FιD α and FιDα induced by the inclusions ια and ιβ .

figure 19c shows all connecting domains with non-negative multiplicities. δ } : dc → cc. Q2 } and their negatives. each with a single boundary puncture: {q2 . D2 = {p2 . D3 = {p4 . or respecively X+ + P+ ⊇ X− + P− . D2 . {Q2 . D2 − D3 = {p2 . P2 . Q2 . P2 }. . P4 . Q4 }. so P+ ⊇ X− . In particular. P4 }. since neither X− nor X+ contain any region more than once. q4 } − {P4 . β } : aa → ab. δ } : cd → cc. First of all. We claim that these are in fact all domains that connect these eight generators and have non-negative multiplicities. X+ ⊇ P− . {Q4 } : dc → dd. so P+ ⊆ X− . q2 . D1 − D2 = {p1 /Q3 . Q2 } − {q2 . q2 . q2 } − {P2 . P4 . q4 . The following domains consist of two bigons. p2 . Q2 }. It is now elementary to check that indeed. {P4 . we can compute the connecting domains shown in figure 19c. we can always choose a domain X = (X+ − X− ) : x → y with multiplicities in {−1. q2 . {p1 /Q3 } : ab → dc. q3 /P1 . q3 /P1 }. +1}. 0. Q4 } The periodic domains which have multiplicities ≥ −1 or ≤ +1 are given by D1 . p1 /Q3 } − {q3 /P1 . D1 + D3 = {p4 . p1 /Q3 . we can argue as follows: Pick any two generators x and y and calculate a connecting domain between them. Q4 . Q4 } − {p4 . q4 } − {P2 . Then X+ + P+ ⊆ X− + P− . All those domains above contribute to the type AA structure. β } : aa → ba. here are some bigons with a single boundary puncture: {p2 } : bb → ab. {p4 . To show this. we only need to consider periodic domains which have multiplicities ≥ −1 or ≤ +1. {q4 } : cd → dd. D1 − D2 + D3 = {p4 . eg by following along the arrows in figure 19c. X+ ⊆ P− . p4 . Let us compute the group of periodic domains of our Heegaard diagram. The polygonal regions q3 /P1 and p1 /Q3 connect the following generators: {q3 /P1 } : ba → cd. Q4 } − {q4 . D3 . We observe that if x and y are among those eight generators from figure 19c. P2 }. Suppose this new connecting domain (X+ + P+ ) − (X− + P− ) has non-negative (respectively non-positive) coefficients only. q3 /P1 . It is easy to see that it is freely generated by the following three domains: D1 = {p2 . We can obtain any other connecting domain between x and y by adding a periodic domain P = P+ − P− to X. P4 . From these. Q2 } − {q4 . PECULIAR MODULES FOR 4-ENDED TANGLES 31 Let us calculate some connecting domains between the generators. {P2 } : bb → ba. p1 /Q3 . D2 + D3 = {p2 .

the expression (7) simplifies to (8) BSD(X [ T2 )B0 [ T1 )A0  A0 . Now identify I ∂ with Iα0 and Iβ0 such that an idempotent for site s corresponds to an idempotent which does not occupy the α. by definition. In fact. Pi . Let us connect these two generators to those from figure 19c. \ Thus. the δ-gradings on both sides of our glueing formula agree. 2. . Finally.and respectively β-arc labelled s in figure 18c. let  1 if j = i. aa. Then by a final application of the pairing adjunction. 4} and x ∈ {pi . definition 4. j ∈ {1. By the addit- ivity of the δ-grading under glueing.32 CLAUDIUS BODO ZIBROWIUS There are four generators that we have not considered yet. each domain from figure 19c is recorded in P 0 by the algebra elements in A0 and B 0 corresponding to the algebra elements in A∂0 and A∂0 obtained by the product of all red and blue labels of the domain. Qi }.B0 P 0  BSD(X Let A∂0 be the quotient algebra obtained from A∂ by setting p3 = 0 = q1 . namely aa. there are unique I ∂ -algebra epimorphisms πα : A0 → A∂0 and πβ : B 0 → A∂0 . 3.ιβ (BSAA(X)).22). Similarly. is homotopic to FιAA α . {δ } : cc → cc. Thus the two generator pairs can be cancelled. respectively. The second identification follows likewise. P 0 is equal to the image of the type AA structure P from figure 17 (viewed as a bimodule over A∂0 ) under the induced functor FπAA α . Q2 } : aa → aa. a “universal” Alexander grading: For fixed i.19]. the remaining complex is the same as before. qi . There are for example the two domains {P2 . Note that. cc and cc.A 0 β 0 Let π : A → ∂ A∂0 be the quotient map defining A∂0 . we can define a relative Z4 -grading Auni . see [Zib16. Hence.and β-curves while at the same time switching the orientation of the Heegaard surface leaves the resulting complex unchanged.7 below and the observation that the bordered sutured Heegaard diagram for XT1 can be regarded as a tangle Heegaard diagram for T1 . {P4 . by the pairing adjuction (theorem 1. the δ-grading on P is calculated as usual. observing that switching the roles of α. the expression (8) is equal to (9) Fπα (BSD(X [ T1 ))A0  ∂ ∂ P  Fπ (BSD(X [ T2 ))A0 ∂ ∂ A . it is sufficient to identify Fπα (BSD(X [ T1 )) and Fπ (BSD(X β [ T2 )) with Fπ (CFT∂ (T1 )) and Fπ (CFT∂ (T2 )). They are connected by the following two contributing domains {β } : aa → aa. after cancellation. Auni j (x) = 0 if j 6= i. Under this identification. We can now use the A∞ -relations to deduce that all domains from figure 19c contribute and thus define a type AA structure P 0 which. up to a minor modification at the tangle ends. as the notation for our domains suggests. respectively. Q4 } : cc → cc.πβ . Thus. The first identification is immediate from lemma 3. We can use the same arguments as above to verify that there are no domains with non-negative domains leaving aa or terminating at cc other than {β } and {δ }.

M. the Alexander grading on both sides of the glueing formula agree. since arcs in A give rise to nontrivial elements in π1 (S. D contributes iff D0 does. ∂S r M ).→ (S. PECULIAR MODULES FOR 4-ENDED TANGLES 33 D ←→ D′ Figure 20. In the first case. Let F c (S. M. M ) is (the image of) an embedding of an oriented closed interval (I. A side s also lies on the boundary of some face in F (S. M ) where S is an oriented surface with boundary and without closed components and M is a (possibly empty) set of basepoints on ∂S. For each face f in F (S.1. given an ordered matching. Now. for a suitable choice of complex structure. A) for which f is closed. A). Given an arc system A on a marked surface (S. For each arc a ∈ A. M. Furthermore. Then. An arc a of a marked surface (S. M. we denote this face by f (s). A). we denote the arc corresponding to a side s by a(s). A). by its additivity under glueing. we call s1 (a) and s2 (a) the two sides of a. The set of all sides s with f (s) = f is denoted by S(f ). Let D and D0 be two domains which only differ in a small region of multiplicity 1 as shown in figure 20. Then. M. let nf be the number of components of f ∩∂S. M ). in the second case an open face. . We also fix a foliation Fa = I × I of N (a) such that s1 (a). let n = lcm(nj . nf ≥ 2 or f contains a basepoint in M . Suppose. A marked surface is a pair (S. M. choose a closed neighbourhood N (a) of a such that N (a) ∩ N (a0 ) = ∅ for all a0 ∈ A r {a}. S Furthermore. A domain with multiplicity 1 near a tangle end Since the three generating periodic domains have vanishing Alexander gradings. Curved complexes for marked surfaces Definition 4. Proof.7]. We call A (together with a choice of fixed N (a) and foliation Fa as above) an arc system if all faces f ∈ F (S. we can also define such a grading Auni on P which is preserved by the arrows in P. A) are topological discs containing at most one point in M . ∂S r M ) is non-trivial.7. we can specialize this universal Alexander grading to a relative Z2 -grading which corresponds to the Alexander grading induced by that ordered matching. A)). ∂S r M ) such that its class [a] ∈ π1 (S.  4.  Lemma 3. ∂I) . Again. s2 (a) and a are leaves of Fa . A) be the subset of those f ∈ F (S. proposition 2. we call the closures of the connected components of S r a∈A N (a) faces and denote the set of all faces by F (S. We now introduce some more notation. we call f a closed face. A is a set of pairwise disjoint arcs on a marked surface (S. j ∈ F (S. This follows from the same arguments as [Han13. Let s1 (a) and s2 (a) be the two components of ∂N (a) in S such that s1 (a) lies to the right of the oriented arc a and s2 (a) to its left. by definition. M. M ).

which correspond to the arrows in the quiver Q(S. A) below. p1 a5 p5 p4 p2 a1 p3 a2 a3 p7 p8 p6 a4 (b) The quiver Q(S. The oriented arcs in A are drawn in red. except that they require nf ≥ 3. M ).34 CLAUDIUS BODO ZIBROWIUS p1 s2 (a5 ) a5 s1 (a5 ) s2 (a1 ) p2 s2 (a2 ) p5 p4 s2 (a3 ) a1 a2 a3 p7 p8 s1 (a1 ) p3 s1 (a2 ) p6 m2 s1 (a3 ) s2 (a4 ) a4 s1 (a4 ) m1 (a) An arc system A on the marked surface (S. Example 4. A)) and remov- ing the interior of each face f as well as any component of Figure 22. M ) above Figure 21. m2 }). Given an arc system A on a marked surface p1 p2 (S. .4 Remark 4. The boundary components are labelled by the elementary algebra elements pi . M. M. M. . M.1 and 4. we have labelled the components of ∂S ∩ f which do not contain the two points m1 and m2 by variables pi .2. M ) with an arc system A. For each face f . bounded by solid arrows representing the two sides of ai . M.3. . A typical neigh- ∂S ∩ f containing a basepoint in M (which defines the edges bourhood of an arc a . These correspond to the basic algebra elements from the next definition. labelled by ai . The neighbourhood of each arc ai is shaded in light red. their arc systems allow faces of arbitrary genus. An illustration of the definitions 4. A). it simply serves as an example to illustrate the generality of the arguments in this section. Definition 4. in this case labelled by m1 and m2 . A) obtained by contract. {m1 . a ∂S ∂S ing the closed neighbourhood of each arc in A to a single q1 q2 point (which defines the vertices of Q(S. consider the graph Q(S. Also. In this particular example. Figure 21a shows an example of a marked surface (S. Our arc systems correspond to their admissible arc systems. M ) = (D2 r3D2 . which are the faces in F (S.4. 5. Our conventions are slightly different from those in [HKK14]: Their markings M correspond to the subsets ∂S r M . We usually draw the boundary of the surface in green and mark elements in M by dashes through the boundary. i = 1. Note that this particular marked surface will be irrelevant for our tangle invariants. A) for the arc system A on the marked surface (S. . the arcs cut the surface into three components.

we define pst = 0. b) := ιb . M. b ∈ A. A)/R∞ . A) as a deformation retract of the surface S with a single puncture in every closed face. Let X U= Uf . If γts contains a point in M . PECULIAR MODULES FOR 4-ENDED TANGLES 35 of Q(S. Definition 4. This defines a partial order on all f -pure algebra elements. An R≥0 -grading on A is called a δ-grading. ie if there exists some algebra element q such that p = qp0 or p = p0 q. A) into a quiver. where R = {p1 q1 = 0 = q2 p2 | arcs a ∈ A} and R∞ = {p1 p2 = 0 = q2 q1 | arcs a ∈ A}. Uf is homogeneous of degree 2 and the subalgebra of grading 0 is I. define Us to be the algebra element corresponding to the path that starts at the side s and traverses the boundary of f exactly once. A typical neighbourhood of the image of an arc a ∈ A is shown in figure 22. Definition 4. Let Uf be the sum of the algebra elements Us over all s ∈ S(f ). A)/R and A∞ := FQ(S. denote the idempotent corresponding to a by ιa and let I be the ring of all idempotents. pst = ιa(s) . We can consider A and A∞ as two categories: The underlying objects are given by the arcs in A. we can define a n2 Z-grading δ by setting δ(p) = n2f for any elementary algebra elements p with f (p) = f . Let A be an arc system on a marked surface (S. Then for any a. Definition 4. Given a side s of a closed face f . The standard basis of A as a vector space over F2 is given by non-zero elements of the form pt Ufk . M. MorA (a. M ). if for all f ∈ F c . γts is homotopic to a path of arrows in Q(S. Let A be an arc system on a marked surface (S. Such gradings always exist: For example. M. Given two sides s and t in S(f ). M. Every arrow in the quiver corresponds an algebra element which we call the elementary algebra elements.ιa and multiplication is given by algebra multiplication.5. f ∈F c (S. A)). For each arc a ∈ A. A) the following two path algebras with relations: A := FQ(S. M.A. Note that if s = t. consider a path γts on ∂f connecting s to t when oriented according to the boundary orientation of ∂f and missing some point on ∂S ∩ f . M ) and f ∈ F (S. M. We associate with the triple (S. Geometrically.ιa and MorA∞ (a. For fixed f .7. M ). pure. we think of Q(S. A) and thus to an element in the algebra A. where s and t are two sides of a face f and k is some non-negative integer. M. By choosing the induced boundary orientation on ∂f .6. A∞ . M ). A). we say p is longer than p0 if the former is divisible by the latter. Otherwise. s we call these elements of A f -pure. Let A be an arc system on a marked surface (S. as we can see. b) := ιb . Note that we follow the convention to read algebra elements from right to left. set Uf = 0. The f -pure elements which are not equal to the identity are called strictly f -pure. or generally for all faces. each arc a is the starting (ending) vertex of at most two arrows. M. the 1-cells inherit an orientation from (S.A) . For open faces f .M. Given two f -pure elements p and p0 for the same face f . We say p is strictly longer than p0 if we additionally require that q does not act as the identity. which we denote by pst . which turns Q(S.

we fix a δ-grading on A in the defin- p2 p3 ition of the category CC(S.8 pqMod.33 and corollary 4. This plays the role of the Z-grading in section 1. A) of curved complexes. 1 d 4 face into exactly two faces each of which has four bound. However. ∅) and let A be q1 ι4 q4 the arc system consisting of four arcs which divide the sur. theorem 32] to our setting. M ) with arc system A as follows: vertices corresponding to generators of an object are mapped to pairwise disjoint points on the arcs in A corresponding to the idempotents of the generators. which is the main goal of this subsection and an adaptation of [HRW16. there is also a homological grading h. Usually. A))) by CC(S. we will think about objects and morphisms in CC(S.26.38. M. of peculiar modules is a special case of CC(S. except that the 2 b 3 q ι2 q3 differential here increases δ-grading by 1. A) the cat- Figure 23. face and arc system which are because. an A∞ -structure is defined on A∞ . M. which then gives rise to the notion of twisted complexes using an A∞ -version of Cx0 (Mat(·)). A) traversing p more than once is set to zero. We work instead with the full algebra A. M.9. in this particular case. A). So actually. Example 4. For each face f .11. Remark 4. ι1 a c ι3 Remark 4. The grad- ing h is defined as a combination of the δ-grading and the Alexander grading. In particular. each arrow labelled by a strictly f -pure element a = pst Ufk is mapped to an oriented path in S connecting the images of the endpoints such that the induced path in Q(S. In what follows. Let (S. the path γ can be drawn such that it only intersects the face f and only the two sides s and t and that it only intersects A in the endpoints. The marked sur- egory of curved cocomplexes. In [HKK14]. see definition 2. A posterior.10. the graphs can be immersed into the marked surface (S. The main technical difference is that in [HRW16]. The latter is preserved by the differentials. and for see example 4. A) and call it the category of curved complexes associated with (S. as shown in figure 23.8. We have chosen this convention. M. Arrows labelled by idempotents are mapped into the neighbourhoods of the corresponding arcs in A. M. M. so for the present section it is irrelevant. 4. In fact. see corollary 4. . M. there is a special elementary algebra element p such that any f -pure algebra element in A that corresponds to a path in Q(S. It is worthwhile to compare our definition of curved complexes associated with marked surfaces to extendable type D structures studied in [HRW16]. A) in terms of labelled graphs as in section 1.36 CLAUDIUS BODO ZIBROWIUS We denote the dg category CxU (Mat(A(S.9. we will see that these two perspectives coincide and that it is actually sufficient to work with marked surfaces and arc systems in which each face has a basepoint. as seen in the previous example. M. A) = pqMod. which decreases along differentials by 1. p1 p4 ary components. the algebra A is truncated in the sense that for any face f . one should call CC(S. M ) = (S 2 r 4D2 . Remark 4.1. A). the category pqMod relevant for peculiar modules. A) corresponds to the algebra element a. Then A = A∂ and CC(S. we will make use of their algorithm below in theorem 4. with 2 respect to the δ-grading. M. Curved complexes as immersed curves with local systems. M.

B0 (f ) = B(f ). i) on the sides specified by the first components and ordered by the second components according to the orientation of the respective side. M. i)|s ∈ S(f ). then s 6= s0 . we usually represent the elements in B(f ) by pairwise different dots •(s. see figure 24a. Then. {πf }). M ). M. .M. na in B(f (s1 (a))) with the standard basis vectors of Va .14.12. A). A pre-curve L is a pair ({Pa }a∈A .M.. the map πf can be interpreted as a set of paths on the face f as follows: We draw an embedded path between each pair of points matched by the fixed point free involution πf |B0 (f ) .na . we draw an embedded path from each of the remaining points in B(f ) to m. • if πf ((s.A) ). the second condition in the previous definition says that no path returns to the same side. i) = 1 − δ(pst ) for (s. we define a map Π0 : Cpre ( . {πf }f ∈F (S.ei 1 }i=1. which we denote by s (a) {eai := ei 1 }i=1. M. We call the paths matching point in B(f ) to point in B(f ) ∪ (M ∩ f ) f -joins. i) ∈ B0 (f ) and πf (s. With this graphical interpretation.. Given a pre-curve L = ({Pa }a∈A . In this case. i)) = (s0 .. j). A)) as follows: Given a pre-curve ({Pa }. • for each face f ∈ F (S.. where B(f ) := {(s. M.. If f is a closed face. i0 ) for some (s.. M. This is the preferred basis on Va . or if we do not specify f . A) by Cpre (S. Pa ∈ GLna (F2 ) for some fixed na ∈ Z≥0 . a compatible δ-grading on such a pre-curve is an R- grading on the sets B(f ) satisfying the following properties: • (Pa )ij = 0 if δ(s1 (a). Given an arc system A on a marked surface (S. A) → ob(CC(S. However. i). j) − δ(s. We denote the set of all δ-graded pre-curves of (S. PECULIAR MODULES FOR 4-ENDED TANGLES 37 Definition 4. A). . i) ∈ B0 (f ). We then regard the invertible matrix Pa as a base change matrix which gives rise to a second basis s (a) s (a) {ei 2 := Pa−1 . there is a single point m ∈ M ∩ f . .A) ). the corresponding arcs might be the same.13. i) = (t. {πf }f ∈F (S. πf (B0 (f )) = B0 (f ) and πf |B0 (f ) : B0 (f ) → B0 (f ) is a fixed point free involution. i = 1. Given a δ-grading on the algebra A. there is a map πf : B(f ) → B(f ) ∪ (M ∩ f ). satisfying the following properties: • for B0 (f ) := πf−1 (B(f )). so there are no unmatched points left. M ).. . Let A be an arc system on a marked surface (S. 1 ≤ i ≤ na(s) }. where • for each arc a ∈ A. We identify the elements (s1 (a). we just talk about joins. Definition 4.. as illustrated in figure 24a. j) 6= δ(s2 (a). For an illustration. Remark 4. let the ιa -component of its image under Π0 be given by the vector space Va = Fna . Otherwise. i) and • δ(t.

. . M. A): Indeed. so the ∂ 2 -identity holds.. . Otherwise. Figure 24. i) ∈ M .. It is now straightforward to check that this defines a curved complex in CC( .. The first condition on the δ- grading for pre-curves ensures that the δ-gradings induced on each vector space Va = Fna by the bases of the two sides s1 (a) and s2 (a) agree. the second condition on the δ-grading for pre-curves ensures that the differential increases δ-grading by 1. If πf (s) (s. An example illustrating the definition above is shown in figure 24. which we identify with the elements (s2 (a). . Any 0 components involving non-trivial powers of Uf . na in B(f (s2 (a))). the pss0 -component of ∂(esi ) for s0 ranging over all elements in S(f (s)) vanishes. i = 1. i). Finally. An illustration of the map Π0 from definition 4. The δ-grading on the sets B(f ) induces a δ-grading on the bases {esi }i=1.15. vanish. M. i) under πf (s) . A).14 of Va . Example 4. the only non-zero compositions of arrows in this pre-complex correspond to the fixed point free involutions πf for closed faces f . A) from figure 21 ea25 p1 p2 p5 p4 p7 p4 p1 ea15 p4 p2 p5 p7 ea11 ea12 ea13 ea23 p7 p6 p3 ea14 (b) The curved complex corresponding to the pre-curve shown above. f ∈ F (S. there is exactly one non- vanishing such component. namely esi0 ⊗ pss0 . . i0 ) is the image of (s. Note that any δ-grading of the algebra needs to take the value 1 on p7 . The differential of the curved complex is now easy to describe in terms of these two bases of the underlying I-module: Let esi be a generator.38 CLAUDIUS BODO ZIBROWIUS   1 0 p1 Pa5 = 0 1 b b Pa5 b b p2 b b p5 p4 b b Pa1 Pa2 Pa3 p7 p8 b b b b p6 2 m p3 b Pa4 b   1 1 Pa3 = 1 m 1 0 (a) An example of a pre-curve for the triple (S. M. where (s0 ..

We have now split off a graph with two vertices and one or two arrows between them. We can now apply induction on the number of vertices in G to see that we can split G into pairs of . or do a corresponding base change. starting at etj .25. Indeed.X). Similarly. for every arc a ∈ A. Proof. ∂). note that there is at most one arrow going into esi . we can consider the subgraph G of (X. If q 6= ιt . Suppose there is another arrow in G. we apply this lemma to h1 . see figure 26. in both cases. s 6= t. so h21 = 0. The homotopy h1 h1 ∂h1 = 0 as well. Let us choose a δ-homogeneous basis {esi } of ιs . set na = dim(ιa . In fact. p = pst with s 6= t. we can achieve that there are no arrows labelled esi etj by prt Ufm going into etj . For closed faces f . any arrow labelled by an idempotent does not have any other labels. since X is δ-graded. t ∈ S(f ) and n ≥ 0. ∂). Note that this only affects arrows in the subgraph G of the complex (X. we see that the compositions with the differential on X have erk b δ-grading 1. so h1 satisfies the hypothesis of lemma 1. PECULIAR MODULES FOR 4-ENDED TANGLES 39 Proposition 4. For open faces. Let us call the source of this arrow esi and the target etj . iff the face f is closed. namely the one going to esi labelled by pts and that Figure 26. The homotopy h2 there is such an arrow iff f is closed.16. clearly n = 0 and since there are no identity labelled arrows.23). we can either perform a similar homotopy h2 . So instead. First of all. If t = r and q = ιt . Thus. which removes the arrow esi to erk . any such arrows have to be at pst b b least as long as as p. We can argue similarly that there is at most one arrow leaving etj . there is such an arrow. so we can reduce the number of generators by cancellation (lemma 1. Consider an arrow in G labelled by a shortest algebra element p = pst Ufn 6= 0 among all labels of arrows in G. Hence. pst b b there exists some algebra element q such that qpst = psr Ufm . we consider such a homotopy as a base change of {eti }. so a shortest arrow leaving esi has to be labelled by an element strictly shorter than Uf . Then by assumption. ∂) in CC is homotopic to one in the image of the map Π0 from the previous definition. namely one labelled by pts . Consider an arrow h1 labelled by q going from etj to erk . we may assume without loss of generality that the differential ∂ does not contain any identity components. Every (δ-graded) curved complex (X. ∂) consisting of the vertices corresponding to {esi } with s ∈ S(f ) and those arrows labelled by f -pure elements pst Ufn for s. labelled by psr Ufm esi etj leaving esi and going to a generator erk . q psr Ufm Finally. such a homotopy would also affect arrows in the rest of the complex (X. ∂ 2 (esi ) contains a term Uf ⊗ a. because of the erk b ∂ 2 -relation. the composition with 0 q would be a power of Uf and thus have an even δ-grading. Interpreting (X. otherwise. Then h1 is a morphism of δ-grading 0 and the only algebra elements q s m p r Uf of δ-grading 0 are the idempotents. Step 1. So Figure 25.X for all sides s. if there were an arrow from erk to etj labelled by prt Ufm for some integer m0 . ∂) as a graph. Fix a face f . Furthermore. so by the same argument as above. We now repeat this precedure until there are no arrows in G leaving esi . Then.

we might also see arrows labelled by 1 + Uf . we can now define the matrices Pa . Next.  i  . 1  Eij := (δi0 j 0 + δii0 δjj 0 )i0 j 0 =  . j ≤ n and i 6= j. While their proofs rely on the fact that extendable type D structures are defined over the truncated algebra A (see remark 4. let Pij ∈ GLn (F2 ) be the permutation matrix corresponding to the permutation (i.18.26 and definition 4.  Remark 4. This defines the map πf . .  j   . (Such a decomposition always exists.. Pa1 b b for some la ≥ 0. For each arc a ∈ A. These are simply the base change matrices corresponding to the bases for the two sides s1 (a) and s2 (a) for each arc a. we repeatedly apply the clean-up lemma. there are no singletons.10).19 below. Definition 4. {πf }) is a well-defined pre-curve.   1 Furthermore. once there are no arrows left. Proposition 4. M ) b b with an arc system A. in step 1 of the proof. . la . Given three positive integers i. Pak is either equal to some Eij or Pij . where for each k = 1. In some sense. connect the dots on the two sides s1 and s2 corresponding to points in B(f (s1 )) and B(f (s2 )) by pairwise disjoint arcs which are . definition 17] of a special form. to made precise in the proof of theorem 4. Note that if f is closed. (Note that these need not be the same as the ones chosen at the beginning of step 1. for some i = i(k) and j = j(k). After doing step 1 for all faces f . the image of this pre-curve under Π0 is homotopic to (X. A train-track in a elimination.16 above corresponds to propositions 22 and 24 in [HRW16].) Secondly. To verify that all hypotheses of the lemma are satisfied. by construction. we first reduce the complex. Thus we can label the k th segment by the matrix Pak . so the whole map πf is a fixed point free involution. Definition 4. . in the proof above. Finally.) It is now straightforward to check that ({Pa }. . .) Divide the neighbourhood N (a) of a along some neighbourhood of an arc leaves of Fa into la segments and order them along a path from s1 := s1 (a) to s2 := s2 (a). ∂). n satisfying i. If we dropped the δ-grading. Step 2. . j.17. using the fact that any arrow labelled by an idempotent is not labelled by another algebra element. A δ-grading is induced by the δ-grading on the basis elements.40 CLAUDIUS BODO ZIBROWIUS vertices and. . our pre-curves correspond to train-tracks from [HRW16. We now introduce a graphical notation for s2 · · · Pij · · · Eij · · · s1 the matrices Pa in a pre-curve on a marked surface (S..19. this is basically Gaussian Figure 27. j). choose a decom- position of Pa into elementary matrices i b b > b b j Pa := Pala . we define the elementary matrix Eij ∈ GLn (F2 ) by i j   1    . singletons. our proof relies in essential points on the δ-grading of our curved complexes over the full algebra A: Indeed. which do not have an inverse in A! (One might be able to rectify this by using power series. we use the δ-grading again.

so πf (s. A). . j).19 transverse to the foliation Fa . If we add the joins to this picture. since Pij = Eij Eji Eij . where L0 is the pre-curve obtained from L by one of the following three moves: (M1) Multiply Pa on the right/left by Pij . We call this a switch. we do not meet πf (s. we draw a pair of arrows in the k th segment from the j th arc to the ith arc. as shown in figure 28. It is well-defined up to homotopy through train-tracks and the following moves. in such a case. assume also that if we follow along a path along the boundary of f (s). a double-arrow connecting the other two strands needs to be added. (T3) Any two double-arrows can be moved past one another unless one strand is the start of one and the end of the other. {πf }) be a pre-curve on a marked surface (S. and precompose πf with the map that permutes (s. Some train-track moves corresponding to a change of matrix decomposition in definition 4. i) and (s. while leaving the other arcs unchanged. assume that (s. Then multiply Pa on the right/left by Eij . depending on whether s = s1 (a) or s = s2 (a). M. which are illustrated in figure 28: (T1) A switch can be expressed in terms of double-arrows. j) lie in M and they belong to the same side s0 of some arc a0 . s ∈ S(f ). We call this a double-arrow. We call it the train-track associated with a pre-curve. depending on whether s0 = s1 (a0 ) or s0 = s2 (a0 ). j) lies in M or they both belong to different sides of f . Without loss of generality. i) and (s. (M2) Suppose neither πf (s. starting at s. j). j) = (s0 . and multiply Pa0 on the right/left by Ei0 j 0 . j 0 ) for some integers i0 and j 0 . Then Π0 (L) is homotopic to Π0 (L0 ). (T2) Two adjacent identical double-arrows can be cancelled. Then multiply Pa on the right/left by Eij depending on whether s = s1 (a) or s = s2 (a). the result is an immersion of a graph with both oriented and unoriented edges. If L is δ-graded. since Eij Eij is the identity.20. j) have the same δ-grading. This corresponds to the identities     E 0 0E if i 6= j 0 and j 6= i0  i j ij Eij Ei0 j 0 = E 0 0E E 0 i j ij i j if i = j 0 and j 6= i0    E 0 0 E E 0 ij ij ij if i 6= j 0 and j = i0 . M ) with arc system A. Let f ∈ F (S. PECULIAR MODULES FOR 4-ENDED TANGLES 41 ←→ ←→ (T1) (T3) Figure 28. Let L = ({Pa }. (M3) Suppose at least one of πf (s. i) before πf (s. Lemma 4. j ≤ na . i) and πf (s. change the ith and j th arcs in the k th segment such that they cross once and remain transverse to the foliation. i0 ) and πf (s. a = a(s) and fix some distinct integers i and j with 1 ≤ i. If Pak = Eij for some i 6= j. i) = (s0 . depending on whether s = s1 (a) or s = s2 (a). i) nor πf (s. If Pak = Pij for some i 6= j. whose ends are tangential to the two strands and which intersect each leaf of Fa at most once. Both double-arrows and switches are illustrated in figure 27.

Definition 4. In terms of train-tracks from definition 4. All three parts of the lemma can be t shown in the same way. depending on whether the joins end on sides or points in M . pushing the double-arrow along parallel joins. where the equivalence ∼ on the left hand side is given by the moves (M1) to (M3) of the previous lemma and ∼ on the right hand side denotes homotopy equivalence. Remark 4. we can Figure 30. only the local pictures of L0 are shown.22. namely by doing a base change. we see that we can reformulate (M2) as follows: if there is only one double-arrow in the picture for (M2) in figure 29. thus.20. For (M3). we can interpret the three moves from the lemma above graphically. For the other moves. we do two pst′ b such base changes. For (M1). M. one for s and one for s0 . M. as illustrated in figure 29.21. we do one such base change. L is obtained by removing all double-arrows. However. there are four different cases to consider. The homotopy for homotope such arrows away using the homotopy h from the the proof of invariance of (M2) same figure. the result after such a pts b s h = ptt′ base change obviously agrees with Π0 (L). namely the ones shown in figure 30. the b t′ result then differs in at most two arrows from the one for Π0 (L). which b in our graphical notation shifts the outermost segment of N (a) adjacent to s into the face f . proof or lemma 4. (M1) shows a local picture of the pre- curve L on the left and of L0 on the right.20. the map Π0 induces a well-defined map Π0 : Cpre (S.19. A))/ ∼. Graphical calculus for lemma 4. For L00 . then we may remove that double-arrow and replace it by the other double-arrow.42 CLAUDIUS BODO ZIBROWIUS Pij j b b j b j j′ b ←→ b b b b i i i i′ (M1) (M2) m∈M b b b j b j b j b j m∈M b b b b i i i i b b m∈M (M3a) (M3b) (M3c) (M3d) Figure 29. Using (T2). For (M2).  . As a consequence of the previous lemma. and again Π0 (L0 ) agrees with Π0 (L). A)/ ∼→ ob(CC(S.

A)/ ∼ as follows: Given a single curve (γ.23. A). M ) and X = id ∈ GLn (F2 ) for some positive integer n satisfying the following properties: • γ restricted to each component of the preimage of each face is an embedding and • γ restricted to each component of the preimage of the neighbourhood N (a) of each arc a is an embedding. X) are given by the intersection of these curves with f . A) → Cpre (S. We consider loops up to homotopy of the underlying immersed curves through curves. Definition 4. M.25. M. Suppose x and y are joined by a component of γ r A. for each face f ∈ F (S. in case (2) a path. A). we call (γ. M ) be a marked surface with an arc system A. M. say. ∂I) into (S. and set it equal to X −1 otherwise. we define a map Π00 : C(S. M. M. or (2) γ is an immersion of an interval (I. respectively. Let (S. A δ-grading on a curve (γ. A) is a pair (γ. where either (1) γ is an immersion of an oriented circle into S. On all other arcs. intersecting each leaf of Fa exactly once. Given an arc system A on (S. We denote the set of all collections of δ-graded curves up to equivalence by C(S. X). Then such a component is mapped to a path in Q(S. A curve on (S. PECULIAR MODULES FOR 4-ENDED TANGLES 43 Definition 4. Then we ask that 1 − δ(pst ) = δ(y) − δ(x). M ). the joins of Π00 (γ. X). We also define the map Π : C(S. defining a non-trivial element of π1 (S. We say that a curve (γ. we consider the local systems of loops up to matrix similarity. Then pick an intersection point x of an arc a with γ. We define this block to be equal to X if γ goes through x from the right of a to its left. Then. The definition of C(S.24. M. A collection of (δ-graded) curves is a finite set of (δ-graded) curves such that all underly- ing curves are pairwise non-homotopic as unoriented curves. we choose the identity matrix. M ). Note that the local system of paths only records some positive integer n. Let the matrix Pa be the diagonal block matrix with blocks of dimension dim X such that all blocks are equal to the identity matrix except the one corresponding to the intersection point x. Note that the definition of this composition is independent of the choice of a and x up to homotopy in CC(S. X) a loop. we extend Π00 to collections of curves by taking unions on both sides. Remark 4. Furthermore. A))/ ∼ as the composition of Π00 and Π0 . Finally. M. A) → ob(CC(S. In case (1). from x to y. X) is an R-grading on the set of intersection points of the underlying curve with arcs in A satisfying the following property: Let x and y be two intersection points of δ-grading δ(x) and δ(y). M. A). choose a small immersed tubular neighbourhood of γ and replace γ by dim X parallel copies thereof in this neighbourhood. A) corresponding to a pure algebra element of the form pst . representing a non-trivial primitive element of π1 (S) and X ∈ GLn (F2 ) for some n. A) is essentially independent of A except for the δ-grading. M. which can be seen by repeatedly . M. X) is supported on the immersed curve γ and call γ the underlying curve and X its local system.

L0 ) to be        (10) V0 (n · n0 )⊗ t (γ.42. {πf }) for which the maps πf are not necessarily involutions. Train-tracks associated with pre-curves ({Pa }. (Minimality can be achieved through a sequence of homotopies of the underlying curves which strictly reduce the number of intersection points by removing any immersed bigon in the sense of definition 4. Classification of morphisms between curved complexes. The same algorithm also works without any changes for pre-curves ({Pa }. {πf }) in our setting for which the maps πf are all involutions correspond to certain types of train-tracks in the language of [HRW16].2. that if γ and γ 0 are parallel. transversely and away from the neighbourhoods N (a) of the arcs a ∈ A. let δ(γ. X) and L0 = (γ 0 . Proof. .26. γ 0 ) denote the vector space spanned by intersection points between γ and γ 0 . γ 0 ). Once all arrows only connect parallel immersed curves. Any peculiar module is homotopic to a peculiar module which is in the image of Π. (M3c) and (M3d) from figure 29 can be regarded as special cases of (M3a). ie loops.  4. Each intersection point can be δ-graded as follows: follow the first curve from the intersection point to the first point x that lies on A. let Vδ (m) be an m-dimensional vector space in δ-grading δ ∈ R.20 namely (M3b). section 3. Homotope γ and γ 0 such that they intersect minimally. Let L = (γ. we set HF(L.7] for simplifying train-tracks. but the δ-grading does not depend on this choice. If γ and γ 0 are parallel. γ 0 ) be the unique real number one needs to add to the δ-grading of each intersection point of γ with arcs in A such that γ and γ 0 agree as δ-graded curves. M. The output of the algorithm is train-tracks whose double-arrows only connect parallel immersed curves. proposition 24] on train-tracks correspond exactly to our moves (T1) to (T3) and (M1) to (M3) on our train-tracks.) Let t (γ. we see that conjugation of the local systems X of a loop (γ.27. If γ and γ 0 are parallel. A) with dim X =: n and dim X 0 =: n0 . Both x and y are δ-graded. γ 0 ) ⊕ (V0 (1) ⊕ V−1 (1)) ⊗ Vδ(γ. Definition 4. Similarly. Then the moves from [HRW16. their orientations agree. Note s(x) that there are two choices for x. we define HF(L. Define the δ-grading of this intersection point to be δ(y) − δ(x) + δ(ps(y) ). Otherwise. Let M 0 be the set of basepoints on ∂S obtained from M by small pushoffs in the direction of the boundary orientation on ∂S and require that if γ 0 is a path.44 CLAUDIUS BODO ZIBROWIUS applying (M2). So we can apply the algorithm explained in [HRW16. For any non-negative integer m. L0 ) equal to (11) V0 (n · n0 )⊗ t (γ. we can remove all arrows on paths by applying moves (M2) followed by (M3d). since the additional moves for such pre-curves from lemma 4. Theorem 4. Then go back along that path to the intersection point. namely those in which all arrows come in pairs and only sit in a neighbourhood of the arcs. since the geometric object agree.γ 0 ) dim ker (X −1 )t ⊗ X 0 − id . even though they represent algebraic objects defined over slightly different algebras. turn left onto the second curve and follow it to the first point y that lies on A. X 0 ) be two δ-graded curves on a marked surface with arc system (S. its ends lie on points in M 0 . Assume. X) does not change the homotopy type of the image under Π.

the blue arrows only contribute to H∗ (Mor(Π(L). If we replace the joins in each picture by parallel copies thereof. the vector space of such morphisms is generated by one element.4. They fall into four different classes. A) and L0 = (γ 0 . We can now piece those local pictures together. blue: Finally. These configurations contribute exactly one generator to H∗ (Mor(Π(L). However. Then H∗ (Mor(Π(L). all morphisms in the ker- nel of the restricted differential lie in the image of morphisms labelled by strictly f -pure algebra elements.28. Thus. these morphisms are illustrated in figure 31. as they can be homotoped away along identity arrows. These configurations do not contribute any morphisms to H∗ (Mor(Π(L).29. . The notation HF(L. In this case. Again. Remark 4. it does not lie in the image of D. Π(L0 ))) ∼ = HF(L. f. k and n. the morphisms also lie in the kernel of the unrestricted D. the vector space of such morphisms is generated by one element. Moreover. h. such a morphism does not necessarily lie in the kernel of D. t. black: For configurations shown in figures 31b. red: For configurations shown in figures 31c. Π(L0 ))) iff the red arrows on one side of an arc are matched up by the local system on that arc with red arrows on the other side of the arc. the red arrows contribute iff the two segments of γ and γ 0 intersect once. These morphisms also lie in the kernel of the unrestricted differential D and not in the image thereof. as long as they intersect transversely and minimally. but possibly in the image thereof. L0 ). i and m. and t and t0 are either sides of f or points in M ∩ f or M 0 ∩ f . L0 ) is independent of the representatives of the underlying immersed curves of L and L0 . In subsection 4. we will get rid of the minimality condition.5) to the face f and consider the morphisms in the kernel of this restricted differential modulo the images of morphisms labelled by strictly f -pure algebra elements. and this intersection is counted with the multiplicity of the product of the dimensions of the local systems. Thus. the vector space of such morphisms is also generated by one element. Π(L0 ))) iff the blue dotted arrows match up. this corresponds to a a single intersection point between γ and γ 0 . if γ and γ 0 are not parallel loops. s0 and t0 . l and p.4. Then. indicated by the colours of the arrows: empty: For configurations shown in figures 31a. L0 ) will be justified in subsection 4. respectively. namely the morphism represented by the black arrow(s). we simply need to multiply the number of morphisms by the multiplicities of the joins. As a consequence of the theorem above. Let us restrict the differential D defined on the morphism spaces (see definition 1. j and o. Let us discuss these pictures in some more detail. we obtain that HF(L. Π(L0 ))). Let f be a face. Proof. for configurations shown in figures 31d. for the various configurations of s. Morphisms of red type only constitute a morphism in H∗ (Mor(Π(L). Let L = (γ. the difference to the previous case is that at least one of the arrows of the morphism is labelled by an idempotent. g. PECULIAR MODULES FOR 4-ENDED TANGLES 45 Theorem 4. s0 and t and t0 (if they do not correspond points in M ∩ f or M 0 ∩ f ). Similarly. consider all possible morphisms represented by arrows labelled by f -pure algebra elements between the generators corresponding to s. Π(L0 ))). B) be two curves. Consider a single f -join of Π00 (L) between s and t and a single f -join of Π00 (L0 ) between s0 and t0 where s and s0 are sides of f . e.

The red B arrows only contribute iff XA = BX. there are potentially some L′ additional morphisms that do not correspond to intersection b b points. Generators of morphism spaces. If γ and γ 0 are parallel loops. The last row shows those configurations with two ends of paths. The first row shows all configurations without an end of a path. there might be a point of M in the face. Two pairs of blue ar- Figure 32. Additional morph- rows are homotopic iff their difference lies in the subgroup isms for parallel loops H := {Y A + BY |Y ∈ G} of G. (However. some of those arrows are actually zero. the curves carry local systems and b b L the arrows labelled by X represent unions of arrows corres. A ponding to some matrix X in Matm×n (F2 ) =: G. respectively. here. In figure 32. however. we have redrawn this picture. namely those corresponding to local pictures of the X X X X form illustrated in figure 31e. We can also regard H as a .) The second and third rows show those configurations with either an end of a path on M or M 0 . in which case.46 CLAUDIUS BODO ZIBROWIUS s′ b t′ b tb t′ b L′ L L′ L′ L′ s′ b b t′ s b s′ b t′ b b t t b b s L s′ b s b L s b b t b b L′ b L b L s s′ t′ t (a) (b) (c) (d) (e) s′ b t′ b M t′ b L′ L L′ L′ s b s′ b t′ b M s b M L s′ b s b L b b L′ b L s s′ t′ M (f) (g) (h) (i) s′ b M′ t′ b M′ L′ L L′ L′ s b s′ b M′ b t s b b t L s′ b s b L b b L′ L b s s′ M′ t′ (j) (k) (l) (m) s b t′ b L′ L′ L′ M′ b t′ M′ M′ M M t L L M b b b L s′ t (n) (o) (p) Figure 31.

We can regard any X ∈ Matm×n (F2 ) as a linear transformation V → W between two vector spaces V and W of dimensions n and m. two P P i (A ⊗ Xvi −1 )t v ⊗ BXvi i vi i ∈ ∈ ∈ (A−1 )t ⊗B A−1 V ⊗W V ⊗W Avj ∈ V V vj vi∗ (Avi )∗ ∼ ∼ F2 P ∈ i (A V∗⊗W V∗⊗W −1 )∗ v ∗ ⊗ BXvi i δij ∈ ∈ = P P ∗ i vi ⊗ Xvi = X BXA−1 = i (Avi ) ∗ ⊗ BXvi (b) (a) Figure 33. Commutative diagrams for the final identification in the proof of lemma 4. The left hand side of (12) is therefore equal to the dimension of the kernel of the linear automorphism of V ∗ ⊗ W that sends X to BXA−1 . we can express BXA−1 P as i (Avi )∗ ⊗ BXvi . so the diagram in figure 33a indeed commutes. so it its equal to the number of contributing red arrows.30. respectively. Then X ∈ V ∗ ⊗ W can be P written as i vi∗ ⊗ Xvi . ∅) with an arc system A. Corollary 4. the number of homotopy classes of blue arrows is equal to # H G = # ker α. Then. a separate argument would be needed in this case. where V ∗ denotes the dual vector space of V . Using the basis {Avi }. Dualizing a linear transformation corresponds to transposing the corresponding matrix.32. However. The arguments that we use in the proof of theorem 4. We are now ready to prove theorem 0. Given two matrices A ∈ GLn (F2 ) and B ∈ GLm (F2 ) for some positive integers n and m. A0j )}j∈J be two collections of loops. Now we conclude with the following lemma. For this. So it suffices to show that this linear transformation can be identified with the map (A−1 )t ⊗ B.31 rely on different growth properties of the dimensions of the two summands in (10) from definition 4.  Lemma 4.33. then (12) dim{X ∈ Matm×n (F2 )|XA = BX} = dim(ker((A−1 )t ⊗ B + id)) Proof. Theorem 4. Consider a marked surface (S.30 . As the diagram in figure 33b shows. Remark 4. Ai )}i∈I and L0 = {(γj0 . as shown in figure 33a. Then Π0 (L) is homotopic to Π0 (L0 ) iff there is a bijection ι : I → J such that γi is homotopic 0 to γι(i) and Ai is similar to A0ι(i) . ∅. Let M be a set of points on ∂S. the second term simply vanishes for paths and thus.27 under pairing with particular test curves. Classification of curved complexes.31. such that every face f ∈ (S. fix some basis {vi } of V and let {vi∗ } be the dual basis in V ∗ . PECULIAR MODULES FOR 4-ENDED TANGLES 47 quotient of G via the map α : Y 7→ Y A + BY . A) contains at most one point in M . and thus as an element in V ∗ ⊗ W .3. We expect the same theorem to hold for general curves ie loops and paths.  4.4 as a special case of the following general classification. Let L = {(γi . Thus. (Avi )∗ = (A−1 )∗ vi∗ .

35. A) → CC(S. j ∈ {1. By theorem 4. 4} as well as one of zj and wj for each closed component of T .  Before we prove theorem 4.48 CLAUDIUS BODO ZIBROWIUS objects in CC(S.8. Thus the images of two homotopic objects are homotopic. in conjunction with example 4.  . We can use a nice peculiar Heegaard diagram for a tangle to compute the peculiar module CFT∂ (T ). A)/{pm = 0|m ∈ M } = A(S. ∅. The induced functor is a functor of dg categories. ∅. The complex CFT∂ (T ) corresponding to those domains is exactly the image of CFT∂ (T ) under the functor induced by the quotient map A∂ → A∂ /(pi = 0 = qj ). Theorem 4. if all regions except those containing these basepoints are bigons or squares.  Corollary 4.M denote the collection of loops associated with CFT∂ (T ) by theorem 4.34. ∅. these represent homotopic objects in both CC(S. Every 4-ended tangle T has a nice peculiar Heegaard diagram with respect to any choice of basepoints. Theorem 4. So by corollary 4. A peculiar Heegaard diagram for T is nice with respect to this choice of special basepoints. Given a tangle T . The images of such objects are represented by the same loops with local systems. LT is a well-defined tangle invariant up to homotopy of the underlying im- mersed curves and similarity of the local systems. A) and CC(S.31. Proof. This can be done algorithmically. the calculation of all domains that miss those basepoints is purely combinatorial. we can recover the homotopy type of CFT∂ (T ) from CFT∂ (T ). Since all regions away from the special basepoints are bigons or squares. let us pause for a moment to consider what the previous two results tell us about our tangle invariant CFT∂ . A) iff the curves are the same and the local systems are equivalent. where pm is the algebra element corresponding to the boundary component in which m ∈ M lies. M. A) → A(S. as it is induced by the quotient map A(S.33.26. Peculiar modules for 4-ended tangles can be computed combinatorially. So we may assume that the two objects are direct sums of loops with local systems. Given a 4-ended tangle T in a homology 3-ball M with spherical boundary.33. M. Definition 4. A) are homotopic iff their images under the induced functor CC(S. let LT := LT. by finding a curve with local system representing CFT∂ (T ) as described at the beginning of this section.  Definition 4. Proof of corollary 4. A) are homotopic. Proof.38. 2.36. 3. ∅. M.37. where they describe an algorithm for niceifying any pointed Heegaard diagram with one basepoint in each component of the Heegaard surface minus the α-circles. A).31. This is an application of Sarkar and Wang’s main result in [SW06]. pick two basepoints pi and qj for some i. ∅.

   .31. Theorem 4. Proof. A diagonal block matrix of the form   Cf 1  . The kernel of the matrix (13) has the same dimension as the kernel of ((A−1 )t ⊗ Cf + id). 0 .   1 an−1 Note that Cf is invertible iff a0 6= 0. Two matrices are similar. we may assume without loss of generality that there exists a bijection ι : I → J such that γi = γι(i) . so we are done. Proof. So by allowing 0-dimensional local systems. a1  Cf :=   . Note that this does not change HF(L. .39.   .  1 . For every j ∈ J such that there is no i ∈ I with γj = γi . Let n = deg f . Fix such a bijection ι.  C fr is in Frobenius normal form if fi+1 |fi for all i = 1.  Lemma 4. Ai ) is a loop such that dim Ai 6= dim Aι(i) .  Proof of theorem 4. By performing row and column operations.   . Suppose there exists some i ∈ I such that (γi . PECULIAR MODULES FOR 4-ENDED TANGLES 49 We now turn to the proof of theorem 4. . we can bring (A−1 )t ⊗ Cf + id) into block diagonal form with the first block of dimension (n − 1)m equal to the identity matrix and the second block of dimension m equal to the expression (13) id +an−1 (A−1 )t · · · + a0 ((A−1 )t )n .41.. Every matrix is similar to a matrix in Frobenius normal form. . and A ∈ GLm (F2 ) for some integer m.31.39. . Now multiply (13) by (At )n to obtain f (At ). proposition 36].. L00 ) nor HF(L0 . where f1 . Given a polynomial f ∈ F2 [x]. dim(ker((A−1 )t ⊗ Cf + id)) = dim(ker(f (A))). . add a “formal” curve to L which is supported on γj and has a 0-dimensional local system. This is standard linear algebra. Transposing a matrix does not change the dimension of its kernel. where Cf is the companion matrix of f from definition 4. Given a polynomial n−1 X f = xn + ai xi ∈ F2 [x]. . .. L00 ) for any curve L00 with local system. Let p be the minimal polynomial of the matrix Ai and p0 be the minimal polynomial of the . r − 1.  ∈ GLn (F2 ). Do the same for L0 . where the ai are the coefficients of f as in definition 4. .40.39. . i=0 define the companion matrix Cf of f to be the matrix   0 a0    . This follows from the same arguments as [HRW16.. if they have the same Frobenius normal form. fr ∈ F2 [x]. Definition 4.

Cfm ) has dimension   X m−1 X r X (14) #γi ∩ γi0 · dim Ai0 · N + 2  dim ker(fm (Cfj )) + dim ker(fm (Cfj )) i0 ∈I j=1 j=m and   X m−1 X r X (15) #γi ∩ γi0 · dim A0i0 · N + 2  dim ker(fm (Cfj0 )) + dim ker(fm (Cfj0 )) . . . j=m j=m Hence. we obtain r r 0 X X dim Cfj = dim Cfj0 . However. in particular dim Ai = dim Aι(i) . Cfm ). So we have now established that if L and L0 represent homotopic curved complexes. r − 1 and l0 = 1. Then Ai and Aι(i) are similar. Now. fr0 0 such that fl+1 |fl and fl00 +1 |fl00 for all l = 1. 0 By lemma 4. if r = r0 and fj = fj0 for all j = 1. Hence. . The first sums coincide by the results that we have already established. . . . let fN (x) := (xN −deg p−deg p + 1) · p(x) · p0 (x). r r 0 X X dim Cfj = dim A0i = dim Aι(i) = dim Cfj0 . j=1 j=1 By minimality of m. .50 CLAUDIUS BODO ZIBROWIUS matrix A0ι(i) . since fj |fm for j > m. . . So it remains to show that then also Ai and Aι(i) are similar for all i ∈ I. γi is homotopic to γι(i) 0 and dim Ai = dim Aι(i) . the dimensions of the morphism spaces from the “test curve” (γi . m hence fm (Cfm 0 ) 6= 0. . CfN ) to L and L0 are equal to X #γi ∩ γi0 · dim Ai0 · N + 2 dim Ai i0 ∈I and X #γi ∩ γj0 · dim A0j · N + 2 dim A0ι(i) . The summands in the third sum of (14) are equal to dim Cfj = deg fj .41. we see that they coincide iff their coefficients coincide. fr and f10 . . the third sum in (15) is at most as large as the third sum in (14). For this. let us assume that both the two matrices are in Frobenius normal form defined by polynomials f1 . Assume without loss of generality that f 0 6 |f . . m Contradiction. . . fm 0 6 |f . we can find a bijection ι : I → J such that for all i ∈ I. The second sums agree by minimality of m. . Then the space of morphisms from L and L0 to (γi . i0 ∈I j=1 j=m respectively. r0 − 1. so dim ker(fm (Cfm 0 )) < dim Cf 0 . Then for N > deg p + deg p0 . . j∈J respectively. . By considering these two terms as linear functions in N . . . r. where fm = 0 if m > r and fm 0 = 0 if m > r 0 . .  . Suppose this is not the case. Then there exists some minimal m such that fm 6= fm 0 . Let N = deg fm . Then m m consider the “test curve” given by (γi .

there would be no identity morphism for such L. M ) be a marked surface with arc ϑ′ system A. An immersed annulus is an orientation- preserving immersion α : S 1 × [0. A). • the restriction of ι to the non-negative real part of ∂D2 is a path from x = ι(−i) to y = ι(+i) on ϑ such that the orientation agrees with orientation of any double-arrows in ϑ. in terms of the collections of curves L and L0 representing those objects. L) vanishes for some objects L. M. • the restriction of ι to the non-positive real part of ∂D2 is a path from y to x on ϑ0 such that the orientation agrees with orientation of any double-arrows in ϑ0 . A) for a marked surface (S. so in particular. note that [HRW16] use slightly different orientation con- ventions: because of the way they express their glueing theorem. y∈ϑ∩ϑ0 . In theorem 4. ϑ tracks up to homotopy and the moves (M2).4. ∂) → (S. We call an element ι ∈ M(x. L0 ). For example. M ) with arc system A. in the following. Let us assume that all intersection points between ϑ and ϑ0 are transverse. ϑ0 ) defined by X d(x) = #M(x. The glueing formula revisited. M. M. X 0 ) on (S. Definition 4. y) be the set of orientation-preserving maps ι : (D2 . The following discussion is an adaptation of the arguments in [Abo06] to our setting. We denoted this space by HF(L. we gave a formula for the homology of the morphism space between two objects in CC(S. If there are no immersed annuli. we may. y. in their setting. we treat both curves the same. A bigon x → y faces of (S. However. Since the train-track associated with a curve with x y local system of (S. namely the Lagrangian intersection Floer homology of (the train-tracks associated with) L and L0 . they find it more convenient to interpret the two collections of curves L and L0 differently. whose orientation agrees with the orientation of the double-arrows in ϑ. Let CF(ϑ. This does not happen with our conventions. consider these train. Given two such intersection points x. y) y. which restricts to S 1 × {0}. respectively ϑ0 .42. because it is also equal to the homology of a finite-dimensional chain complex associated directly with L and L0 . to an immersed circle on ϑ. Let (S. A) only has double-arrows between par- b b allel curves. • x and y are convex corners of the image of ι. y) an immersed bigon. A). let M(x. X) and L0 = (γ 0 . respectively S 1 × {1}. which follows more standard conventions in Lagrangian intersection Floer theory. We may then also allow double-arrows to lie in the Figure 34. let d be the endomorphism on CF(ϑ. respectively ϑ0 . M. PECULIAR MODULES FOR 4-ENDED TANGLES 51 4. 1] → S. using the language of [HRW16]. (T1). ϑ ∪ ϑ0 ) satisfying the following properties: • ι is an immersion away from ±i ∈ ∂D2 ⊂ C. namely one in terms of a type D structure and the other in terms of a type A structure. (T2) and (T3) only. So let ϑ and ϑ0 be train-tracks associated with two curves with local systems L = (γ. HF(L. ϑ0 ) be the F2 vector space generated by the intersection points between ϑ and ϑ0 .28.

ϑ0 ).52 CLAUDIUS BODO ZIBROWIUS ϑ′ ϑ′ ϑ ϑ ϑ ϑ ←→ ←→ ϑ′ ϑ′ (a) A finger move (b) A bridge move Figure 35. L0 ). The first creates creates/kills a bigon region. ϑ0 ). ϑ∪ϑ0 ) as follows: Given a bigon ι. d) remains invariant Theorem 4. The first corresponds to a cancellation and the other to a base change in (CF(ϑ. ϑ0 ). If there are no immersed annuli. An illustration of the identification of bigons between parallel curves and the matrix (X −1 )t ⊗ X 0 . d) does not change under the moves on each train-track that take place locally. define a map ϕ : M(x. the second moves a single double-arrow past a strand of the other train-track. The last part follows from homotoping the train-tracks such that they intersect minimally without any immersed annuli. The proof that d is a differential is the same as in [Abo06]. ϑ0 ). Up to conventions. y) → H2 (S. d) does not change under finger and bridge moves. one can easily check that the chain complex (CF(ϑ. the homotopy type of (CF(ϑ. the endomorphism d is a well-defined differ- ential. this is the same as [HRW16.27. Since train-tracks are not allowed to back-track. the homology of this space is given by HF(L. Since ϕ is injective. L0 ) from definition 4. away from the other train-track. contradiction! So d is a well-defined homomorphism. Furthermore. It remains to check that the homotopy type of (CF(ϑ. d) is an invariant of L and L0 . We leave the details to the reader. d). This map is well-defined. ϑ0 ). injective and independent of the basepoints. see figure 35. For any two intersection points x and y of the two train-tracks ϑ and ϑ0 . define the coefficient of each component of Sr(ϑ∪ϑ0 ) in ϕ(ι) to be the number of points in the preimage of some basepoint in r. figure 43]. Finally. Proof. having an infinite number of bigons between two fixed intersection points x and y would imply that some bigon can be restricted to a square by removing neighbourhoods of ±i such that the images of the new boundary map to the same path. along which we can glue the square to obtain an immersed annulus.43. Furthermore. the two paths from x to y obtained by restricting a bigon ι to the boundary of the disc intersect the same arcs in A in the same order. Consider two parallel components and assume without loss of generality that their orientations agree with those in figure 36 and that all double-arrows on     i i ϑ j X′ j  b y  x b    i′ i′ ϑ′ j ′ X j′   Figure 36. Two moves for which the homotopy type of (CF(ϑ. so we may write CF(L. .

so (X −1 )ij · Xj0 0 i0 counts the number of bigons from x to y within this local picture. CFT∂ (m(T2 )))). (X −1 )ij counts the number of paths from x to y on ϑ and Xj0 0 i0 counts the number of paths from y to x on ϑ0 . respectively.  q1 q4 dd d d 1 4 a q1 q4 dc c q2 q3 aa p1 p4 cd a ab p2 p3 c cc b b 2 3 bb p3 p4 p2 p1 ba q2 q3 Figure 37. i0 ) be two intersection points on either sides of the local systems in this picture.  Theorem 4. modulo 2. The first equality is then seen by identifying the intersection points between those curves and the connecting bigons with the generators and differentials of CFT∂ (T1 )  P  CFT∂ (T2 ) from theorem 3. The second equality is an application of theorem 4. we can use any representatives of the curves in LT1 and Lm(T2 ) . Then. Lm(T2 ) ) = H∗ (Mor(CFT∂ (T1 ).28. The boundary of the picture is identified to a point. j 0 ) and y = (i. Geometric interpretation of the type AA structure P for pairing in the wrapped Fukaya category of the 4-punctured sphere. j 0 ) to (i. where m(T2 ) denotes the mirror image of T2 . i0 ) outside this area if i0 = i and j 0 = j. .43. For example.44. as long as there are no immersed annuli. CFL(L) [ ⊗ V i = HF(LT1 . Proof. Let x = (j. By theorem 4. we may homotope the curves in LT1 and Lm(T2 ) into neighbourhoods of the blue and red curves in figure 37. Finally. there is a bigon from (j. PECULIAR MODULES FOR 4-ENDED TANGLES 53 these components are contained in the boxes labelled by the corresponding local systems X and X 0 .6. The blue curves denote a 1-skeleton and the red ones a Hamiltonian translate thereof with reversed roles of pi and qi . which has the same rank as its transpose (X −1 )t ⊗ X 0 − id.6. Thus X −1 ⊗ (X 0 )t − id is the differential. The orientation is chosen such that the normal vector (determined by the right-hand rule) points into the projection plane. With the same notation as in theorem 3.

suppose LT is a single loop which corresponds to an embedded loop on the 4-punctured sphere. So D separates the two strands in T1 or T2 . Rational tangle detection. see example 2. we obtain the 2-component unlink q .       alisation of Ozsváth and Szabó’s exact triangle [OS03a] which    . Let S be the 4-punctured sphere along which we glue T1 and T2 . Then either T1 or T2 is a rational tangle. so by theorem 4. Skein exact sequences. there are exactly two punctures in each disc. Proof. By observation 5. so LT agrees with LT 0 for some rational tangle T 0 . Then also Lm(T ) agrees with Lm(T 0 ) . The Alexander grading on CFT∂ (T. categorifies the oriented skein relation for the Alexander poly.18. However. Let T1 and T2 be two 4-ended tangles without closed components that glue together to the 2-component unlink. We now proceed by induction on the number of circles in S ∩ U . see re- mark 2. n  .1. it separates two punctures from the other two. Let L be the link obtained by pairing T with its mirror m(T ) as in figure 16.44 HFL(L) [ = HF(LT . Theorem 5. Thus either T1 or T2 is rational.25. Conversely. D ∪ D0 bounds a 3-ball.12 . each loop in LT lies in the kernel of π1 (∂M r ∂T ) → π1 (M r ν(T )) → H1 (M r ν(T )).1. for theorem 0. Lm(T ) ) = HF(LT 0 . If γ bounds a disc D0 in S.   nomial. The only-if direction is simply a calculation. If we glue T 0 to m(T 0 ). Observation 5.n  .   . They must obviously be unknotted. since the connected sum of two knots is the unknot iff both knots are unknots.1. Lm(T 0 ) ) = HFL( [ q ). Proof. A 4-ended tangle T in the 3-ball is rational iff LT is a single embedded loop with the unique 1-dimensional local system.3.2. It divides the sphere into two disc components. where the first map is induced by the inclusion and the second is the Abelianization map. The following lemma finishes the proof.54 CLAUDIUS BODO ZIBROWIUS 5. If γ does not bound a disc in S. We start with a slight gener. this intersection is non-empty. so the graded version of the following theorem is (a) Tn (b) T−n (c) T0 not quite as strong as Ozsváth and Szabó’s result in the case Figure 38. So we can always find a curve γ that bounds a disc D in U which does not contain any other curves in U ∩ S. we remind the reader that all gradings on         link Floer homology should be regarded as relative. First of all. which we can use as a homotopy for U to remove γ (along with any other components of S ∩ U in D0 ). since the loop is not nullhomotopic. Let U be the sphere that separates the two unknot components and assume that S and U intersect transversely in a disjoint union of circles. so we are done by the induction hypothesis. each of which has at least one puncture.  . M ) implies that for a tangle T . Applications 5.. We now apply the fact that link Floer homology detects unlinks [OS06a].2.    5. The basic tangles n = 1. so L is the 2-component unlink.  Lemma 5. since U is separating.

given three links Ln . They are both shown in figure 40. Moreover. for n even. respectively.5. Then there is an exact triangle shown in figure 39a. It is straightforward to compute CFT∂ (Tn ) and CFT∂ (T−n ) from genus 0 Heegaard diagrams. which agree outside a closed 3-ball and in this closed 3-ball agree with the 4-ended tangles Tn . depending on whether the two strands in Ti belong to different or the same components in Li . where for i ∈ {n. L−n and L0 in S 3 . labelled alternatingly by pi s and qi s. Furthermore. Let Tn be the positive n-twist tangle. The skein exact triangles from theorem 5. T−n the negative n-twist tangle and T0 the trivial tangle. For even n. The morphism ϕn : CFT∂ (Tn ) → CFT∂ (T−n ) . Proof. where t is the colour of the two open strands. li is either 0 or 1. ϕn preserves the (single- variate) Alexander grading and changes δ. 0}.4 Theorem 5. the other two maps preserve the all three gradings. see figure 38. Similar results hold for other orientations.4 (n-twist skein exact triangle). the two components are connected by similar sequences along the dotted lines. the former on the left. respectively. The horizontal arrows in figure 40 1 1 1 p3 δ − 2 c1−n p321 δ 2 c1−n p21 +q34 δ 0 b−n δ 0 d−n p214 p4 q341 q2 q1 q234 1 1 1 δ − 2 a1−n δ 2 a1−n 1 1 1 δ − 2 cn−1 δ 2 cn−1 q4 q123 q412 q3 p43 +q12 p432 δ 0 dn p2 δ 0 bn 1 p1 1 1 p143 δ − 2 an−1 δ 2 an−1 Figure 40. respectively. PECULIAR MODULES FOR 4-ENDED TANGLES 55 CFT∂ (Tn ) ϕn CFT∂ (T−n ) HFL(L [ n ) ⊗ V ln HFL(L [ −n ) ⊗ V l−n CFT∂ (T0 ) ⊗ V [ 0 ) ⊗ V l0 +1 HFL(L (a) (b) Figure 39. If n is odd. −n. Remark 5. connected by pairs of morphisms.and homological gradings by +1 and −1. the dashed lines denote a sequence of alternating generators in sites a and c. T−n and T0 . let V be a 2- dimensional vector space supported in degrees δ 0 tn and δ 0 t−n . the latter on the right. then the above triangle together with the glueing theorem induces an exact triangle shown in figure 39b. also multivariate Alexander gradings are preserved.

where for i ∈ {0. Moreover. X}. they agree with the mapping cones of maps from the negative crossing to the trivial tangle. Now apply the same arguments as in the proof of theorem 5.  Proposition 5.25.4. we get two copies of p43 +q12 CFT∂ (T0 ) = δ 0 b0 δ 0 d0 . Note that Manolescu uses slightly different conventions from ours.6 (resolution skein exact triangle). Proof. The map ϕ is given by (the horizontal arrows in) the following diagram on the left: q412 q3 b c b c q3 p2 p2 p321 p21 +q34 p43 +q12 p32 +q41 p14 +q23 = ∼ p4 p143 p4 q1 q1 d a d a q234 Using lemma 1. . we see that it is homotopic to the diagram on the right. which agree outside a closed 3-ball and in this closed 3-ball agree with the 4-ended tangles T0 = . Like Manolescu’s triangle. 1.  Next. Theorem 5. There is an exact triangle shown in figure 41a. T1 = and TX = . By cancelling all identity components of the mapping cone of ϕn .7. depending on whether the two strands in Ti belong to different or the same components in Li . which gives rise to the exact triangle of the required form. Since these loops are invariant under taking the mirror.6 describe ϕn . theorem 1]. we can write CFT∂ (T0 ) ⊗ V as a cone of CFT∂ (Tn ) and CFT∂ (T−n ). There are two morphisms 1     δ − 2 t±1 CFT∂ → CFT∂ whose mapping cones are homotopic to the peculiar modules represented by the “figure-8” loops shown in figure 42. respectively. ours does not preserve any gradings. we give a new proof of a theorem by Manolescu [Mnl06.56 CLAUDIUS BODO ZIBROWIUS ϕ CFT∂ ( ) CFT∂ ( ) HFL(L [ 0 ) ⊗ V l0 HFL(L [ 1 ) ⊗ V l1 CFT∂ ( ) HFL(L [ X ) ⊗ V lX (a) (b) Figure 41. so the two triangles only look the same after reversing the direction of the three arrows. The skein exact triangles from theorem 5. respectively. li is either 0 or 1. p21 +q34 Thus. given three links L0 . together with the glueing theorem induces the exact triangle from figure 41b. which is CFT∂ ( ). then the above triangle. L1 and LX in S 3 .

reversing all arrows.7 Proof. b2−n b4−n p1 1 1 ··· p2 bn 1 dn p2 p2 q412 q412 a1−n a3−n p143 p143 p143 q4 p14 p14 ··· an−1 q412 q3 q3 q3 q2 q341 q2 q341 c1−n c3−n q341 p43 +q12 q2 p3 q41 p3 q41 ··· p3 cn−1 p21 +q34 p214 p214 p214 b−n b2−n ··· bn−2 bn Figure 44. but in the opposite direction. up to b an additional tensor factor. A complex of peculiar modules homotopic to CFT∂ (Tn ) . as a complex in the objects and . swapping pi and qi and reversing the Alexander grading leaves both of them invariant. Also note that the proposition above gives rise to an exact triangle similar to the one in [OS07].  Remark 5. we obtain the two peculiar modules shown in figure 42. The two maps between the peculiar invariants of the two tangles are shown in figure 43. with both strands oriented upwards. After cancelling the identity arrows in both mapping cones. Doing this to the mapping cone gives us maps between the mirrors of the two tangles. Two figure-8 loops (a) and (c) and their peculiar modules (b) and (d) 1 1 1 δ − 2 d+1 1 1 δ 2 d+1 δ − 2 d−1 δ 2 d+1 p432 p4 p432 q1 p1 δ 0 a0 p21 +q34 p1 δ 0 a0 q4 p4 q123 p43 +q12 q123 p43 +q12 q341 q341 p21 +q34 δ 0 c0 q2 δ 0 c0 p2 q2 p3 p3 q3 p214 p214 1 1 − 12 1 δ − 2 b−1 b−1 δ b+1 δ b−1 2 1 δ 2 (a) (b) Figure 43. as the number of generators agree for the second loop.8. Moreover. Also. The two maps from proposition 5. we can write the n-twist tangle Tn from figure 38a. It is interesting to compare the “figure-8” curve to the local Heegaard diagram for a singular crossing in [OSS07]. cancelling the b identity components in the complex from figure 44 gives us a loop representing Tn . PECULIAR MODULES FOR 4-ENDED TANGLES 57 p2 p432 1 1 δ 0 a0 δ − 2 b+1 δ 0 a0 δ 2 d+1 p143 b b p1 b b q2 q341 q3 q412 q234 q1 q123 q4 p3 p321 1 1 δ 0 c0 δc0 b b b b δ 2 b−1 δ − 2 d−1 p214 p4 (a) (b) (c) (d) Figure 42. Indeed.

The general case uses the combinatorial algorithm Figure 45. preserves bigraded link Floer homology.  Corollary 5. mutation in general only preserves δ-graded link Floer homology. The peculiar modules of (2n. figure 10] in the context of instanton tangle Floer homology in the pillowcase. Proof of corollary. For example. Then.11 corresponds to a recolouring of the vertices: • ↔ • and • ↔ •. Finally.38. Peculiar modules of (2n. Then. if we reverse the orientation of one strand. a relabelling of the sites as in theorem 5. For n = m = 1. t1 t2   Theorem 5..11.26 directly from the definition of peculiar modules.4 to locally modify tangles until we obtain a complex of peculiar modules of trivial and 1-crossing tangles. Let T be a tangle in the closed 3-ball B 3 and T 0 the tangle obtained by relabelling the sites such that a and c. all generators on the diagonals from bottom-left to top-right have the same Alexander grading.   Remark 5.     2m + 1   pretzel tangles for n. Herald. .10. from theorem 5. Furthermore.58 CLAUDIUS BODO ZIBROWIUS Similarly.9. but the generators in the same Alexander gradings now sit on the diagonals that go from top-left to the bottom-right. 5. −(2m + 1))-pretzel tangles for n. . orient T 0 such that the orientation at the first tangle ends of T and T 0 (and hence any others) agree. one (only) needs to get rid of these extra factors. Kirk associate with a trivial tangle in [HHK13. if CFT∂ (T ) and CFT∂ (T 0 ) are (graded) chain-homotopic. The invariants of the (2n. −(2m+1))-pretzel tangles simply have the desired symmetry. If T is oriented. −(2m + 1)).9 Theorem 5. This follows directly from the definition of mutation and the glueing theorem. up to a large number of tensor factors from glueing. we can obtain a complex for T−n by applying the mirror operation. It is also interesting to compare our “figure-8” curve to the curve that Hedden. it is also easy to find such complexes for other orientations of T−n and Tn .. and b and d are interchanged. oriented as in figure 45. after identifying the Alexander gradings of the two open strands.12.  . Note that these complexes look very much like the ones we get in Bar-Natan’s Khovanov homology of tangles [Bar04]. are . or the two objects and . mutation of these tangles preserve (graded) link Floer homology. oriented as in figure 45. Proof. this calculation was already t1 t2 done in example 2.6 together with the skein exact sequence from theorem 5.  2n    represented by the curves shown in figure 46. We assume that every tangle can be written as a complex in the two objects and b . we can iteratively use the type AA glueing structure from theorem 3. This can be seen as follows: In terms of the loops on our infinite chessboard. In fact. we do not need to rotate the curves. in the case of the (2. Mutation about (2n. An orientation reversal of both tangle strands corresponds to a rotation of the chessboard by π. m > 0. this is indeed possible. After identifying the Alexander gradings of the two tangle strands.3. depending on the orientation. −(2m+1))-pretzel tangles. by either changing the orientation on all strands or leaving them all the same.. The pretzel tangle for computing peculiar modules from corollary 4. If we reverse the orientation of one of the two strands. m > 0. −3)-pretzel tangle.

What we need to decide is how they are connected.9. −(2m + 1))-pretzel tangle. Thus. We use the same conventions as in example 2. ie the red and green vertices. for most cases. there is only one way to connect them such that the result is a peculiar module. The peculiar module of the (2n. the vertices of the graphs in figure 46 are fixed. PECULIAR MODULES FOR 4-ENDED TANGLES 59 t−n n−2m−2 1 t2 t−n n+2m 1 t2 b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b tn1 t−n−2m 2 tn1 t−n+2m+2 2 (a) n ≤ m + 1 t−n n−2m−2 1 t2 t−n n+2m 1 t2 b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b t1n−2m−2 t−n 2 b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b t−n+2m+2 1 tn2 b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b tn1 t−n−2m 2 tn1 t−n+2m+2 2 (b) n > m + 1 Figure 46. shown in fig- ure 45. The generators of the peculiar module can already determined from the decategorified invariants and from the observation of two obvious differentials that can be can- celled as in example 2. The only . Proof of theorem 5. The diagonals connecting pairs of red and green generators of the same Alexander gradings should be continued in such a way that they do not intersect each other. see also figures 13c and 14a.26. All generators for site a and c. are in the same δ-grading. Because of the restrictions given by the gradings.26.

Figure 47.60 CLAUDIUS BODO ZIBROWIUS ′ d d d q1 d q4 | | | | p4 a1 c1 | | a2 x1 c1 | | | x2 c1 | | y1 c2 | | y1 c2 | | y1 c2 | | y2 | y2 | y2 | c2m | c2m | y 2m c2m | c2m+1 | | y2m c2m+1 | | y 2m y 2m+1 c2m+1 | | a2n−1 | | a2n x2n−1 y2m+1 y 2m+1 | | | x2n p2 p3 q3 | | b′ | | | | b b b (a) A niceified Heegaard diagram for a (2n. −(2m + 1))-pretzel tangle with n. k ≤ 2m + 1. The first step of the calculation of the peculiar modules of (2n. The generators of the second and third row can be cancelled. −(2m + 1))- pretzel tangles . where 1 ≤ i ≤ 2n and 1 ≤ j. m > 0 ai yj b0 yj xi b xi cj d0 yj xi d byj y j b cj yk y k cj dyj yj d byj y j b cj yk y k cj dyj yj d (b) Generators of the Heegaard diagram above.

the Heegaard diagram remains invariant under the following operation: in the names of the generators. This symmetry corresponds to mutation about the horizontal axis. exchange underlining and overlining. The result is shown in figure 47a. We start with a Heegaard diagram obtained by glueing two Heegaard diagrams for the rational tangles with 2n twist. exchange the letters d and b. and q3 and p4 . respectively −(2m + 1) twist together. and thus can be cancelled along the identity arrows connecting those generators of the same site and with the same indices. as well as p3 and q4 . Note that the nice Heegaard diagram has the same symmetry as the tangles. −(2m+1))-pretzel tangle in non-extremal t1 -Alexander grading question is how the diagonal strings of red and green generators connect the generators on the top left to the generators on the bottom right of each of the subfigures of figure 46. replace i by 2n + 1 − i. exchange p1 and q2 . Finally.38. We will use this symmetry in the following to simplify some parts of the computation. respectively. in the algebra. Some differentials for the computation of the (2n. For this. The generators of the second and third row are the ones that were created during the first and second handleslide. we are going to apply the algorithm from corollary 4. setting p1 = 0 and q2 . and thus the complex inherits this symmetry. . We can niceify the diagram by doing two handleslides of the β-curve for the 2n-twist rational tangle across the other β-curve. and j and k by 2m + 2 − j and 2m + 2 − k. The generators of this diagram are shown in figure 47b. which leaves the tangle invariant up to exchanging the two sites b and d. respectively. More precisely. PECULIAR MODULES FOR 4-ENDED TANGLES 61 xi+1 cj−1 xi+1 d xi+1 c2m q14 q14 q1     ay ai yj   ay  i 1  i 2m+1  p23 p23 p2 xi−1 cj+1 xi−1 c2 xi−1 b (a) 1 < i < 2n and 1 < j < 2m + 1 (b) 1 < i < 2n and j = 1 (c) 1 < i < 2n and j = 2m + 1 ai+1 yj−1 ai+1 y2m p23 p23   p4   q3 xi cj xi d xc xc xi b    i 1  i 2m+1  q14 q14 ai−1 yj+1 ai−1 y2 (d) 1 < i < 2n and 1 < j < 2m + 1 (e) 1 < i < 2n and j = 1 (f) 1 < i < 2n and j = 2m + 1 ai+1 y2m+1   p4 xi d   xi c1 p 2 q3   q1 xi c2m+1 xb i  ai−1 y1 (g) 1 ≤ i < 2n (h) 1 < i ≤ 2n Figure 48.

.62 CLAUDIUS BODO ZIBROWIUS x2 c2m x3 c2m−1 y 2m+1 c2m−1 q14 q14 q14   a y a y  1 2m+1   2 2m  p2 p23    q4 1 p3 q3 x1 c2m d0 y2m+1 y1b by1 x c x1 b    1 2m+1  (a) x2 c2j x3 c2j−1 y 2m+1 c2j−1 q14 q14 q14     q4 a1 y2j+1 a2 y2j dy2j+3     p23 p23 q4  1 1  q4 0 x1 c2j d0 y2j+1 y c c2j+2 y1 x c d y2j+2  1 2j+2   1 2j+1  (b) 1 ≤ j < m x3 c2j−2 y 2m+1 c2j−2 q14 q14   a1 y2j a2 y2j−1   p23 q4  q4 0 x1 c2j−1 d0 y 2j x1 c2j d y2j+1   (c) 1 ≤ j ≤ m x2 d q1  a y  1 1 p23   y 1 c2   1  p   0 4 dy c y  1 2 1  1 1  p4  q4 x  1 d  x c  1 1 d0 y2 (d) Figure 49. Some differentials for the computation of the (2n. −(2m+1))-pretzel tangle at generators in maximal t1 -Alexander grading before cancellation. the dashed arrows iff n > 1. The dotted arrows appear iff n = 1.

Secondly. generators in the same shaded regions share the same Alexander bigrading. (There are two observations that one might find useful when determining the contributing differentials: firstly. all other contributions come from squares. we compute all differentials that start/end at some selected generators. Since the Heegaard diagram is nice. Some differentials for the computation of the (2n. the only bigons in the diagram contribute arrows labelled by the elementary algebra elements q3 and p4 . which are enclosed in those figures by boxes.) p4 p3 x2n d x2n c1 b0 y1 x2n c2 q14 q143 q14 a2n−1 y2 a2n y1 p23 p23 x2n−2 c3 x2n−1 c2 (a) p3 p3 b0 y2j x2n c2j+1 b0 y2j+1 x2n c2j+2 q14 q143 q14 a2n−1 y2j+2 a2n y2j+1 p23 p23 x2n−2 c2j+3 x2n−1 c2j+2 (b) 1 ≤ j < m p3 p3 b0 y2j−1 x2n c2j b0 y2j x2n c2j+1 q14 q143 a2n−1 y2j+1 a2n y2j p23 x2n−2 c2j+2 (c) 1 ≤ j ≤ m p3 b0 y2m x2n c2m+1 q14 a2n y2m+1 p2 x2n−1 b (d) Figure 50. Note that in all figures. the only contributing domains are bigons and squares. so the computation is purely combinatorial and rather straightforward. The dashed arrows appear iff n > 1. . the differentials starting/ending at all marked generators in those figures are included. We therefore ask the reader to check for themselves that indeed. −(2m+1))-pretzel tangle at generators in minimal t1 -Alexander grading after cancellation. which necessarily have boundary on both β-curves. PECULIAR MODULES FOR 4-ENDED TANGLES 63 In figures 48 and 49.

there is no arrow x1 c2j → a1 y2j . where “e” stands for “even” and “o” for “odd”. In figure 50c. p3 : x1 c2j+1 → d0 y2j+2 . The only possible arrow labelled by a power of p leaving d0 y2m+1 can go to a1 y2m+1 . Basic homotopies for the final step of the proof of theorem 5. the pictures in figure 48 do not change.9 . −(2m + 1))- pretzel tangles. the ∂ 2 -relation at x1 c2j+1 in the peculiar module tells us that this does not happen.64 CLAUDIUS BODO ZIBROWIUS e b b o b b e b b o b b e b b o b b e b b oo b oe b b eo oo b b ee oe b b eo oo b b ee oe b b eo oo b b ee b eo o e oo o b b b b b b b b b b b b b b b b b b b b b b b b b b b b eo oo b e o oo b b b b b b b b b b b eb eo ob b b b b b b b b b b eo oo b b b b b b b b b b b eo b ob b b b b b b e oo b b b b b b b b b b b b b b b b b b b eo oo b b b b b b b e o oo b b b b b eb eo ob b b b b b b b b b b eo oo b b b b b b b b b b b eb eo b ob b b b b b b b b b b b b b b e oo b oe b b eo oo b b ee oe b b eo oo b b ee oe b b eo oo b b ee b eo b b b b b b b b b b b b b b e o e o e o o Figure 51. there has to be a contribution p234 : d0 y2j+1 → a1 y2j+1 . we consider the effect of cancelling generators. Next. Some of the vertices are labelled according to the parity of the indices of the generators they correspond to. first those corresponding to undoing the handleslides and then any remaining identity arrows. Similarly. h=q p2 q3 b b b b b b b b = = p3 q2 b b b b b b b b h=p (a) (b) q p2 b b h=p b b b b b b q = p2 = q2 p h=q b b b b b b q2 b b p (c) (d) Figure 52. The last step of the calculation of the peculiar modules of (2n. Again. It might also contribute another arrow. Because of the ∂ 2 -relation in the peculiar module. In figure 49a. this arrow has to be there. stemming from the arrow q4 : y 1 c2j+2 → dy2j+3 . cancellation only contributes an arrow p23 : x1 c2m+1 → a1 y2m+1 . Cancellation contributes an arrow p23 : x1 c2j+1 → a1 y2j+1 . Obviously. however. we can argue for figure 49b.

we cancel two arrows.  . namely x1 d → d0 y1 and y 1 c2 → c2 y1 . respectively. in figure 49d. we argue for those generators in minimal Alexander grading corresponding to the colour t1 . To see that this is indeed the result of putting all pieces together. we have indicated the parity of the generator indices for some generators. This only contributes one arrow. namely p23 : x1 c1 → a1 y1 . we suggest starting at the ends of figure 49d and 50d which sit at the bottom left and top right corners of figure 51. Similarly. PECULIAR MODULES FOR 4-ENDED TANGLES 65 but again. and then connecting these subcomplexes. The corresponding subcomplexes after cancellation are shown in figure 50. we may apply the symmetry of the Heegaard diagram. We can now do some basic homotopies shown in figure 52 to obtain the desired result. the cancellation must contribute an arrow p234 : d0 y2j → a1 y2j . see the caption of the figure. We have now obtained a reduced complex. Alternatively. To this. Finally. we add some arrows labelled by basic algebra elements that lie in the kernel of A∂ → A∂ /(p1 = 0 = q2 ) to turn the complex into the peculiar module shown in figure 51. To see how the diagonal strings of red and green generators connect the generators on the top left to the generators on the bottom right of figure 51.

Szabó. Lett. Floer homology and singular knots. Watson. arXiv: 0401268v2 [Lam16] P. Rasmussen. Manion. The pillowcase and traceless representations of knot groups II: a Lagrangian-Floer theory in the pillowcase. 159 (2004). arXiv: 1611. Kontsevich. Heegaard Floer homology and alternating knots. L. Matrix factorizations and link homology.06559v1 [OS17] . Efimov. 18 (2014). Thurston.02011 [Lam17] . Math. 8 (2008). I. 231 (2012).8611v2 [Han13] J.0687v5 [LOT10a] . no. Math. P. Math. Vértesi: Quantum gl(1|1) and tangle Floer homology. Topol. Geom. D. Bordered knot algebras with matchings. Bordered Floer homology for manifolds with torus boundary via immersed curves. Szabó. 2 (2009). Hanselman.4322) A2311 [Ale28] J. A cube of resolutions for knot Floer homology. 26 (2013). A. 30 (1928). P.3852v1) [OS16] . Abouzaid. Adv. Amer.5199v2) [EPV15] A.00597v1 [OSS07] P. Ann. Alexander. On the Fukaya Categories of Higher Genus Surfaces. On the decategorification of Ozsváth and Szabó’s bordered theory for knot Floer homology. Watson. 14 (2007). Ozsváth. Kirk. 4. P.6696 [HRW16] J. C. arXiv: 1707.2661v3) . Manolescu. Flat surfaces and stability structures. arxiv: 1508. arXiv: 1608. Geom. D. 186 (2004). arXiv: 1501. and a bigraded knot invariant. arXiv: 1409. genus bounds. 615–692 (arXiv: 0512286v2) [OS06a] . 9 (2005). A combinatorial spanning tree model for knot Floer homology. A. J. Topol. Topol. J. A. Math. Geom. Knot Floer homology. Khovanov. Amer.08001 [Mnl06] C. Soc. Bordered Heegaard Floer homology and graph manifolds. Twisting. Khovanov’s homology for tangles and cobordisms. Herald. A calculus for bordered Floer homology. Baldwin. 225–254 (arXiv: 0209149v3) [OS03a] . 7 (2003). mutation and knot Floer homology. S. Holomorphic discs and topological invariants of closed 3-manifolds. An unoriented skein exact triangle for knot Floer homology. bordered algebras. 1443–1499 (arXiv: 0410495v2) [BB14] D. L. arXiv: 0606598 [AAEKO] M. A.00028v1 [KR04] M. 1051-1083 (arXiv: 1103. Link Floer homology and the Thurston norm. Z. no. Bordered Heegaard Floer homology: Invariance and pairing.03483v1 [HKK14] F. arXiv: 1604. Lipshitz. Bimodules in bordered Heegaard Floer homology. Hanselman. Katzarkov.0598v4) [Mnn16] A. M. arXiv: 1701. Z. 2 (2009). Abouzaid. arXiv: 1305. Homological mirror symmetry for punctured spheres. Math. 1. A. 58–116 (arXiv: 0209056v4) [OS03b] . 1027–1158 (arXiv: 0101206v4) [OS02] . Kauffman states. Math. L. Holomorphic disks and knot invariants. 275–306 [Bar04] D. Bar-Natan. (arXiv: 1003. Lipshitz. J. I Petkova.0164v1) [HHK15] . Orlov. Ozsváth. 865–910 (arXiv: 0705. Geom. Haiden. Algebr.1695 [BL11] J. H. 19 (2015) 525–724. Res. Burgos-Soto. and mutation. D. Hedden. arXiv: 0303225v2 [OS05] . L.00880 [Lip05] R. Topol. arXiv: 1310. Geom.03466v2 [HW15] J. Soc. Ellis. Rozansky. Holomorphic disks and link invariants. Adv. Khovanov homology for alternating tangles. I. Geometry & Topology 10 (2006) 955–1096. On Conway mutation and link homology. Ozsváth. arXiv: 0601618v3 [OS07] . arXiv: 1510. Topol. arXiv: 0810. (arXiv: 0502404v2) [LOT08] R. Topol. Lambert-Cole. Trans. 211–287 (arXiv: 1301. A cylindrical reformulation of Heegaard Floer homology. arXiv: 1603. V. Auroux.66 CLAUDIUS BODO ZIBROWIUS References [Abo06] M. Stipsicz. Katzarkov. The pillowcase and perturbations of traceless representations of knot groups. Topological invariants of knots and links. Levine. L.05445v1 [HHK13] M. Hanselman. 1886–1939 (arXiv: 1105. J. W. 380–404 (arXiv: 0705. Topol. 839–852 (arXiv: 0609531v3) [OS01] P. Bar-Natan.

arXiv: 1410. Soc. Essay for the Smith-Knight & Rayleigh-Knight Prize Competition 2015.04915v1) [Zib16] . A. Orientable genera of 4-ended tangles. arXiv: 1601. 171 (2010). Sarkar. Seidel. V. arXiv: 1610. of Math. Link cobordisms and functoriality in link Floer homology. C. PhD thesis (2011). Zibrowius. Categorical homotopy theory.05207v1 [Zib15] C. University of Cambridge. 2008 [WZ] L. On a polynomial Alexander invariant for tangles and its categorification. arXiv: 0908. Riehl. PECULIAR MODULES FOR 4-ENDED TANGLES 67 [PV14] I. Zürich lectures in advanced mathematics. no. Watson.3496v1 [Zar11] . 2017 (arXiv: 1610.07494) . Cambridge University Press (2014) [SW06] S. Harvard (arXiv: 0306378v1) [Rie14] E. Wang. Zarev. J. Floer homology and knot complements.. B.1106v2 [Zar10] . Combinatorial tangle Floer homology. Columbia University (doi: 10. European Math. in preparation [Zar09] R. Petkova. Joining and gluing sutured Floer homology. Zemke. arXiv: 1010. Zibrowius. Bordered Sutured Floer Homology. An algorithm for computing some Heegaard Floer homologies. PhD thesis (2003). Fukaya Categories and Picard-Lefschetz Theory.7916/D83R10V4) [Zem16] I. Ann.2161v2 [Ras03] J. 2. Cambridge (arXiv: 1601. 1213–1236 (arXiv: 0607777v4) [Sei06] P. Bordered Floer homology for sutured manifolds. On a Heegaard Floer theory for tangles. Heegaard diagrams for tangles.04915 [Zib17] . Rasmussen. Vértesi.