Vous êtes sur la page 1sur 122

Topological quantum field theory:

symmetries and defects


Nils Carqueville
Uni Wien & Erwin Schr
odinger Institute

geometric
Langlands

(quantum)
computing
beauty
beauty
(?)

cobordism
hypothesis

(higher)
categories

topological
quantum
field
theory

quantum
field theory

nature
knot
invariants

conformal
field theory

geometric
Langlands

(quantum)
computing
beauty
beauty
(?)

cobordism
hypothesis

(higher)
categories

topological
quantum
field
theory

quantum
field theory

nature
knot
invariants

conformal
field theory

geometric
Langlands

(quantum)
computing
beauty
beauty
(?)

cobordism
hypothesis

(higher)
categories

topological
quantum
field
theory

quantum
field theory

nature
knot
invariants

conformal
field theory

geometric
Langlands

(quantum)
computing
beauty
beauty
(?)

cobordism
hypothesis

(higher)
categories

topological
quantum
field
theory

quantum
field theory

nature
knot
invariants

conformal
field theory

geometric
Langlands

(quantum)
computing
beauty
beauty
(?)

cobordism
hypothesis

(higher)
categories

topological
quantum
field
theory

quantum
field theory

nature
knot
invariants

conformal
field theory

geometric
Langlands

(quantum)
computing
beauty
beauty
(?)

cobordism
hypothesis

(higher)
categories

topological
quantum
field
theory

quantum
field theory

nature
knot
invariants

conformal
field theory

geometric
Langlands

(quantum)
computing
beauty
beauty
(?)

cobordism
hypothesis

(higher)
categories

topological
quantum
field
theory

quantum
field theory

nature
knot
invariants

conformal
field theory

geometric
Langlands

(quantum)
computing
beauty
beauty
(?)

cobordism
hypothesis

(higher)
categories

topological
quantum
field
theory

quantum
field theory

nature
knot
invariants

conformal
field theory

geometric
Langlands

(quantum)
computing
beauty
beauty
(?)

cobordism
hypothesis

(higher)
categories

topological
quantum
field
theory

quantum
field theory

nature
knot
invariants

conformal
field theory

geometric
Langlands

(quantum)
computing
beauty
beauty
(?)

cobordism
hypothesis

(higher)
categories

topological
quantum
field
theory

quantum
field theory

nature
knot
invariants

conformal
field theory

geometric
Langlands

(quantum)
computing
beauty
beauty
(?)

cobordism
hypothesis

(higher)
categories

topological
quantum
field
theory

quantum
field theory

nature
knot
invariants

conformal
field theory

geometric
Langlands

(quantum)
computing
beauty
beauty
(?)

cobordism
hypothesis

(higher)
categories

topological
quantum
field
theory

quantum
field theory

nature
knot
invariants

conformal
field theory

Basic features of quantum physics


physical states described by vector space V

Basic features of quantum physics


physical states described by vector space V
observables (energy, position,. . . ) described by linear operators on V

Basic features of quantum physics


physical states described by vector space V
observables (energy, position,. . . ) described by linear operators on V
time evolution of V described by linear map Ut

Basic features of quantum physics


physical states described by vector space V
observables (energy, position,. . . ) described by linear operators on V
time evolution of V described by linear map Ut ,
in quantum mechanics (= 1-dimensional quantum field theory):
i

=H
t

(t) = Ut (0)

Ut = eiHt

Basic features of quantum physics


physical states described by vector space V
observables (energy, position,. . . ) described by linear operators on V
time evolution of V described by linear map Ut ,
in quantum mechanics (= 1-dimensional quantum field theory):
i

=H
t

(t) = Ut (0)

Ut+t0 = Ut Ut0

Ut = eiHt

Basic features of quantum physics


physical states described by vector space V
observables (energy, position,. . . ) described by linear operators on V
time evolution of V described by linear map Ut ,
in quantum mechanics (= 1-dimensional quantum field theory):
i

=H
t

(t) = Ut (0)

Ut+t0 = Ut Ut0
similar for any quantum field theory

Ut = eiHt

Basic features of quantum physics


physical states described by vector space V
observables (energy, position,. . . ) described by linear operators on V
time evolution of V described by linear map Ut ,
in quantum mechanics (= 1-dimensional quantum field theory):
i

=H
t

(t) = Ut (0)

Ut+t0 = Ut Ut0
similar for any quantum field theory
Idea: axiomatise key properties of path integral

Ut = eiHt

Categories
A category C consists of
a set of objects , , , . . .,
sets Hom(, ) of morphisms for all objects , ,
composition maps Hom(, ) Hom(, ) Hom(, )
which are associative and unital.

Abstracts

Categories
A category C consists of Equivariant completion
a set of objects , , , .Nils
. ., Carqueville
(joint work with Ingo Runkel)
sets Hom(, ) of morphisms for all objects , ,
composition maps Hom(, ) Hom(, ) Hom(, )
which are associative and unital.
Y0
X

Recall the basic setting of a bicategory B, which one may think of as a monoidal
category with labels. We denote objects of B by a, b, . . ., generic 1-morphisms by
X, Y, . . ., units by Ia , and 2-morphisms by , , . . .. Composition of 1-morphisms
is written as , and we have two (Poincare dual) graphical depictions at our
disposal:
Y
A

*
X

=
b

A
a

Abstracts

Categories
A category C consists of Equivariant completion
a set of objects , , , .Nils
. ., Carqueville
(joint work with Ingo Runkel)
sets Hom(, ) of morphisms for all objects , ,
composition maps Hom(, ) Hom(, ) Hom(, )
which are associative and unital.
Y0
X

A functor C C 0 maps objects to objects and morphisms to morphisms,


Recall the basic setting of a bicategory B, which one may think of as a monoidal
preserving
units and compositions.
category with labels. We denote objects of B by a, b, . . ., generic 1-morphisms by
X, Y, . . ., units by Ia , and 2-morphisms by , , . . .. Composition of 1-morphisms
is written as , and we have two (Poincare dual) graphical depictions at our
disposal:
Y
A

*
X

=
b

A
a

Abstracts

Categories
A category C consists of Equivariant completion
a set of objects , , , .Nils
. ., Carqueville
(joint work with Ingo Runkel)
sets Hom(, ) of morphisms for all objects , ,
composition maps Hom(, ) Hom(, ) Hom(, )
which are associative and unital.
Y0
X

A functor C C 0 maps objects to objects and morphisms to morphisms,


Recall the basic setting of a bicategory B, which one may think of as a monoidal
preserving
units and compositions.
category with labels. We denote objects of B by a, b, . . ., generic 1-morphisms by
Examples.
X, Y, . . ., units by Ia , and 2-morphisms by , , . . .. Composition of 1-morphisms
, and we have two (Poincare dual) graphical depictions at our
C is=written
Vectkas: objects
= k-vector spaces, morphisms = linear maps
disposal:

Y
A

*
X

=
b

A
a

Abstracts

Categories
A category C consists of Equivariant completion
a set of objects , , , .Nils
. ., Carqueville
(joint work with Ingo Runkel)
sets Hom(, ) of morphisms for all objects , ,
composition maps Hom(, ) Hom(, ) Hom(, )
which are associative and unital.
Y0
X

A functor C C 0 maps objects to objects and morphisms to morphisms,


Recall the basic setting of a bicategory B, which one may think of as a monoidal
preserving
units and compositions.
category with labels. We denote objects of B by a, b, . . ., generic 1-morphisms by
Examples.
X, Y, . . ., units by Ia , and 2-morphisms by , , . . .. Composition of 1-morphisms
, and we have two (Poincare dual) graphical depictions at our
C is=written
Vect as: objects
= k-vector spaces, morphisms = linear maps
disposal: k
C = Mfld: (smooth) manifolds and smooth maps
Y

*
X

=
b

Abstracts

Categories
A category C consists of Equivariant completion
a set of objects , , , .Nils
. ., Carqueville
(joint work with Ingo Runkel)
sets Hom(, ) of morphisms for all objects , ,
composition maps Hom(, ) Hom(, ) Hom(, )
which are associative and unital.
Y0
A

A functor C C 0 maps objects to objects and morphisms to morphisms,


Recall the basic setting of a bicategory B, which one may think of as a monoidal
preserving
units and compositions.
category with labels. We denote objects of B by a, b, . . ., generic 1-morphisms by
Examples.
X, Y, . . ., units by Ia , and 2-morphisms by , , . . .. Composition of 1-morphisms
, and we have two (Poincare dual) graphical depictions at our
C is=written
Vect as: objects
= k-vector spaces, morphisms = linear maps
disposal: k
C = Mfld: (smooth) manifolds and smooth maps
Y
Y
A
C = Bordn : (n 1)-dimensional
compact manifolds
M, N, . . . and
Z

A
a
*
b
c
=
b
a
a
c t Nb /diffeomorphisms
n-dimensional manifolds with = M
X

Topological quantum field theory (`


a la Atiyah-Segal)
An n -dimensional closed TQFT is a symmetric monoidal functor




Z : Bordn , t, VectC , C , C

Atiyah 1988

Topological quantum field theory (`


a la Atiyah-Segal)
An n -dimensional closed TQFT is a symmetric monoidal functor




Z : Bordn , t, VectC , C , C
Why?

Atiyah 1988

Topological quantum field theory (`


a la Atiyah-Segal)
An n -dimensional closed TQFT is a symmetric monoidal functor




Z : Bordn , t, VectC , C , C
Why?

Z(M ) = space of physical states

Atiyah 1988

Topological quantum field theory (`


a la Atiyah-Segal)
An n -dimensional closed TQFT is a symmetric monoidal functor




Z : Bordn , t, VectC , C , C
Why?

Z(M ) = space of physical states



Z : M N = time evolution

Atiyah 1988

Topological quantum field theory (`


a la Atiyah-Segal)
An n -dimensional closed TQFT is a symmetric monoidal functor




Z : Bordn , t, VectC , C , C
Why?

Z(M ) = space of physical states



Z : M N = time evolution

functoriality Z( 0 ) = Z() Z(0 ) captures Ut+t0 = Ut Ut0

Atiyah 1988

Topological quantum field theory (`


a la Atiyah-Segal)
An n -dimensional closed TQFT is a symmetric monoidal functor




Z : Bordn , t, VectC , C , C
Why?

Z(M ) = space of physical states



Z : M N = time evolution

functoriality Z( 0 ) = Z() Z(0 ) captures Ut+t0 = Ut Ut0

Z(M t M 0 ) = Z(M ) C Z(M 0 ) captures many-particle systems

Atiyah 1988

and
only if p C"n , quantum
and f (p) = 1 iffield
and only
if p C(`
.
m
"a
Topological
theory

la Atiyah-Segal)

nor
contain any critical
"-dimensional
AnCnm
closedpoints.
TQFT is a symmetric monoidal functor
al points of f have pairwise
 distinct critical
 values.



"
n

Z : Bord , t, Vect , , C

C
C
ndard techniques, one shows thatnevery open-closed cobordism
M 2Cobext ["n, m]
"
l Morse function f : M . Since M is compact and since all non-degenerate
re Why?
isolated, the set of critical points of f is a finite set. If neither a
nor
l values Z(M
of f , )the
N := f 1states
([a, b]) forms an open-closed cobordism with
= pre-image
space of physical
b}) and 1 N = 1 M N. If [a, b] does not contain any critical value of f , then
Z : M N = time evolution
eomorphic to the cylinder f 1 ({a}) [0, 1].
0 ) captures U
0
Ut Utpoints
functoriality
Z(
0 ) =
Z(the
g proposition
classifies
in terms
of Z()
Morsedata
non-degenerate
t+t0 =critical
0
0
on open-closed
cobordisms.
Z(M t M ) = Z(M ) C Z(M ) captures many-particle systems

" be a connected open-closed cobordism and f : M


8. Let M 2Cobext ["n, m]
function
that f=has
precisely
one of
critical
point.
Then Mdecomposable
is equivalent tointo
one
Bord2such
: objects
disjoint
unions
circles,
morphisms
open-closed cobordisms:

A
A
compositions
Atiyah 1988

(3.25)
A

2d closed TQFT: algebraic description


Theorem.
2d closed TQFTs are commutative Frobenius algebras.

d, the set of critical points of f is a finite set. If neither a


nor
f , the pre-image N := f 1 ([a, b]) forms an open-closed cobordism with
N = 1 M N . If [a, b] does not contain any critical value of f , then
c to the cylinder f 1 ({a}) [0, 1].
ion Theorem.
classifies in terms of Morse data the non-degenerate critical points
osed2d
cobordisms.
closed TQFTs are commutative Frobenius algebras.

2d closed TQFT: algebraic description

" be a connected open-closed cobordism and f : M


2Cobext ["n, m]
uch that f has precisely one critical point. Then M is equivalent to one
Proof sketch: For TQFT Z : Bord2 VectC , set H
d cobordisms:

Z
A

: H H H
C

(3.25)

ons

(3.26)

open-closed cobordisms embedded in 3 and are drawn in such a way


e drawing plane is f . The source is at the top, and the target at the
18

= Z(S 1 ) VectC .

1. The
object !npoints
= (0), of
i.e.fthe
C!n set.
= S If
, forms
a commutative
d, the set
of critical
is circle
a finite
neither
a
nor Frobenius algebra object.
f , the pre-image N := f 1 ([a, b]) forms an open-closed cobordism with
N = 1 M N . If [a, b] does not contain any critical value of f , then

(3.33)
=
=
c to the cylinder f 1 ({a}) [0,=
1].
ion Theorem.
classifies in terms of Morse data the non-degenerate critical points
osed2d
cobordisms.
closed TQFTs are commutative Frobenius algebras.

2d closed TQFT: algebraic description

" be a connected
:M
2Cobext ["n, m]
=open-closed cobordism and f
=
=
uch that f has precisely one critical point. Then M is equivalent to one
Proof
sketch:
For
TQFT
Z
:
Bord

Vect
,
set
H
=
2
C
dTHE
cobordisms:
CATEGORY OF
3 OPEN-CLOSED
THE CATEGORY
COBORDISMS
OF OPEN-CLOSED COBORDISMS

(3.34)

Z(S 1 ) VectC .
21

(3.35)
=is associative,
=
commutative, unital:
Z
: H H H
(3.25)
1
1. The object !n = (0), i.e.
1. The
the circle
objectC!n
(0),
, forms
i.e. thea circle
commutative
C!n = S 1Frobenius
, forms a commutative
algebra object.
Frobenius algebra o
!
n = S

ons

=
=

(3.33)
= (3.36)

(3.26)

2. The object !n =
(1), i.e. the interval C!n
= I, forms
a symmetric
Frobenius
algebra object.

(3.34)=
=
=
=
=
=

open-closed cobordisms embedded in 3 and are drawn in such a way

=
= at the
=
e drawing plane is f . The source
is at the top, and the target

18

(3.37)
(3.35)
(3.38)
(3.36)

1. The
object
!npoints
= (0), identities:
i.e.fthe
C!n set.
= S If
, forms
a commutative
d,
thesatisfy
set
of critical
of
is circle
a finite
neither
a
nor Frobenius algebra object.
ely,
the
zig-zag
f , the pre-image N := f 1 ([a, b]) forms an open-closed cobordism with
N = 1 M N . If [a, b] does not contain any critical value of f , then

(3.33)
=
=
c to the cylinder f 1 ({a}) [0,=
1].

ion Theorem.
classifies in terms of Morse data the non-degenerate
critical
points
=
=
osed2d
cobordisms.
closed TQFTs are commutative Frobenius algebras.

2d closed TQFT: algebraic description

(3

" be a connected
:M
2Cobext ["n, m]
(3.34)
=open-closed cobordism and f
=
=
uch that f has precisely one critical point. Then M is equivalent to one
1 ) Vect .
Proof
sketch:
For
TQFT
Z
:
Bord

Vect
,
set
H
=
Z(S
2 the left C
C
dTHE
cobordisms:
ows
directly
from
relations,
and right unit laws,
CATEGORY
OF
3 the
OPEN-CLOSED
THEFrobenius
CATEGORY
COBORDISMS
OF OPEN-CLOSED
COBORDISMS
21 and the

left


nit laws. From
Equations
(3.40) and (3.37), the pairing can be shown to be symme

(3.35)
=is associative,
=
commutative, unital:
: H H H
iant, Z
1
1 (3.25)

1. The object !n = (0), i.e.


1. The
the circle
objectC!n
(0),, forms
i.e. thea circle
commutative
C!n = S Frobenius
, forms a commutative
algebra object.
Frobenius algebra o
!
n = S

ons

C
=

=
=

(3

(3.33)
= (3.36)

same holds for the copairing. Similarly, we define the closed pairing and the clo
.
(3.26)
:
2. The object !n = (1), i.e. theinterval C= I, forms a symmetric Frobenius algebra object.

!
n

pairing h, i = :=
Z

: H H :=C

open-closed cobordisms embedded in 3 and are drawn in such a way

=
= at the
=
e drawing plane is f . The source
is at the top, and the target

18

22

(3.34)
=

(3

(3.37)

(3.35)
(3.38)
(3.36)

1. The
object
!npoints
= (0), identities:
i.e.fthe
C!n set.
= S If
, forms
a commutative
d,
thesatisfy
set
of critical
of
is circle
a finite
neither
a
nor Frobenius algebra object.
ely,
the
zig-zag
f , the pre-image N := f 1 ([a, b]) forms an open-closed cobordism with
N = 1 M N . If [a, b] does not contain any critical value of f , then

(3.33)
=
=
c to the cylinder f 1 ({a}) [0,=
1].

ion Theorem.
classifies in terms of Morse data the non-degenerate
critical
points
=
=
osed2d
cobordisms.
closed TQFTs are commutative Frobenius algebras.

2d closed TQFT: algebraic description

(3

" be a connected
:M
2Cobext ["n, m]
(3.34)
=open-closed cobordism and f
=
=
uch that f has precisely one critical point. Then M is equivalent to one
1 ) Vect .
Proof
sketch:
For
TQFT
Z
:
Bord

Vect
,
set
H
=
Z(S
2 the left C
C
dTHE
cobordisms:
ows
directly
from
relations,
and right unit laws,
CATEGORY
OF
3 the
OPEN-CLOSED
THEFrobenius
CATEGORY
COBORDISMS
OF OPEN-CLOSED
COBORDISMS
21 and the

left


nit laws. From
Equations
(3.40) and (3.37), the pairing can be shown to be symme

(3.35)
=is associative,
=
commutative, unital:
: H H H
iant, Z
1
1 (3.25)

1. The object !n = (0), i.e.


1. The
the circle
objectC!n
(0),, forms
i.e. thea circle
commutative
C!n = S Frobenius
, forms a commutative
algebra object.
Frobenius algebra o
!
n = S

ons

C
=

=
=

(3

(3.33)
= (3.36)

same holds for the copairing. Similarly, we define the closed pairing and the clo
.
(3.26)
:
3 THE
OPEN-CLOSED
COBORDISMS
23
2. The
objectCATEGORY
!n = (1), i.e. OF
the
interval C= I, forms
a symmetric Frobenius algebra object.

!
n

pairing h, i = :=
Z

open-closed cobordisms embedded in 3 and are drawn in such a way

These also satisfy the


zig-zag identities,
=
= at the
=
e drawing plane is f . The source
is at the top, and the target

18

(3.34)
=

: H H :=C is non-degenerate:

22=

=
and the closed pairing is symmetric and invariant,

(3.35)
(3.50)

(3

(3.37)

(3.38)
(3.36)

2d closed TQFT
2d closed TQFTs are commutative Frobenius algebras.

2d closed TQFT
2d closed TQFTs are commutative Frobenius algebras.
Examples.
G finite Abelian group, k field, group algebra:
H = k[G] ,

hg, hi = g,h1

2d closed TQFT
2d closed TQFTs are commutative Frobenius algebras.
Examples.
G finite Abelian group, k field, group algebra:
H = k[G] ,

hg, hi = g,h1

M complex manifold, Dolbeault cohomology :


Z
H = H(M ) ,
h, i =

2d closed TQFT
2d closed TQFTs are commutative Frobenius algebras.
Examples.
G finite Abelian group, k field, group algebra:
H = k[G] ,

hg, hi = g,h1

M complex manifold, Dolbeault cohomology :


Z
H = H(M ) ,
h, i =

W R = C[x1 , . . . , xn ] isolated singularity, Milnor algebra:




f g dx1 . . . dxn
H = R/(x1 W, . . . , xn W ) ,
hf, gi = Res
x1 W . . . xn W

2d open/closed TQFT
Next step: add more structure to bordism category.

2d open/closed TQFT
Next step: add more structure to bordism category.
Let B be a set. Call elements a, b, c, . . . B boundary conditions.

2d open/closed TQFT
Next step: add more structure to bordism category.
Let B be a set. Call elements a, b, c, . . . B boundary conditions.
o/c

Bord2 (B) is the category with:


objects are disjoint unions of circles and intervals Iab with endpoints
labelled by a, b B:
a

b
Iab

2.2d
f (p)
= 0 if and only ifTQFT
p C"n , and f (p) = 1 if and only if p Cm
".
open/closed
3. Neither C"n nor Cm
" contain any critical points.

Next step: add more structure to bordism category.

4. The critical points of f have pairwise distinct critical values.

Let B be a set. Call elements a, b, c, . . . B boundary conditions.

Using the standard techniques, one shows that every open-closed cobordism M 2C
dmits
a o/c
special Morse function f : M . Since M is compact and since all non
Bord
2 (B) is the category with:
ritical points are isolated, the set of critical points of f is a finite set. If neither
disjoint
of circles
and
intervals
Iab with
endpoints cob
areobjects
critical are
values
of f , unions
the pre-image
N :=
f 1
([a, b]) forms
an open-closed
1
labelled
by
a,
b

B:
({a, b}) and 1 N = 1 M N . If [a, b] does not contain any critical valu
0N = f
1
([a, b]) is diffeomorphic to the cylinder f 1 ({a}) [0, 1].
b
b non-degenerate
c
The following proposition classifies ain terms
of Morsec datac the
cr
Iab
hat can occur on open-closed cobordisms.

" be a connected open-closed cobordism and


Proposition 3.8. Let M 2Cobext ["n, m]
special morphisms
Morse function
such that fdecomposable
has precisely one
are bordisms,
intocritical point. Then M is equiv
f the following open-closed cobordisms:

A
A
A
r to one of the compositions

2d open/closed TQFT: algebraic description


Theorem.
o/c
A 2d open/closed TQFT Z : Bord2 (B) VectC

Lazaroiu 2000, Lauda/Pfeiffer 2005, Moore/Segal 2006

2d open/closed TQFT: algebraic description


Theorem.
o/c
A 2d open/closed TQFT Z : Bord2 (B) VectC is
a commutative Frobenius algebra H = Z(S 1 )

Lazaroiu 2000, Lauda/Pfeiffer 2005, Moore/Segal 2006

2d open/closed TQFT: algebraic description


Theorem.
o/c
A 2d open/closed TQFT Z : Bord2 (B) VectC is

a commutative Frobenius algebra H = Z(S 1 ),


a Calabi-Yau category O with objects B and HomO (a, b) = Z(Iab )

Lazaroiu 2000, Lauda/Pfeiffer 2005, Moore/Segal 2006

Lazaroiu 2000, Lauda/Pfeiffer 2005, Moore/Segal 2006

ws that every open-closed cobordism M 2Cobext ["n, m]


"
. Since M is compact and since all non-degenerate
itical points of f is a finite set. If neither a
nor
e N := f 1 ([a, b]) forms an open-closed cobordism with
. If [a, b] does not contain any critical value of f , then
r f 1 ({a}) [0, 1].
terms of Morse data the non-degenerate critical points
.
] be a connected open-closed cobordism and f : M
recisely one critical point. Then M is equivalent to one

Theorem.
o/c
A 2d open/closed TQFT Z : Bord2 (B) VectC is
a commutative Frobenius algebra H = Z(S 1 ),
a Calabi-Yau category O with objects B and HomO (a, b) = Z(Iab )
 
: H EndO (a))
(plus compatibility constraints on Z
(3.25)

(3.26)

2d open/closed TQFT: algebraic description

(3.45)

2d open/closed TQFT: algebraic description


Theorem.
o/c
A 2d open/closed TQFT Z : Bord2 (B) VectC is

elationsa commutative Frobenius algebra H = Z(S 1 ),


a Calabi-Yau
category
O be
with
objects B
andtheHomO (a, b) = Z(Iab )
e additional
diffeomorphisms
that can
constructed
from

3.10.
To simplify
Proof
sketch:the diagrams, we define:

Z
:=

: HomO (a,:=
b) HomO (b, a) C (3.46)

which we sometimes call the open pairing and open copairing,


dentities:

(3.47)

robenius relations, the left and right unit laws, and the left and
ons (3.40) and (3.37), the pairing can be shown to be symmetric
Lazaroiu 2000, Lauda/Pfeiffer 2005, Moore/Segal 2006

(3.45)

2d open/closed TQFT: algebraic description

= (1), i.e. the interval C!n = I, forms a symmetric Frobenius algebra object.

(3

Theorem.

(3.37)
=
=
= o/c
A 2d open/closed
TQFT
Z i.e.
: Bord

Vect
C is
2. The object
!n = (1),
the interval
C!n =
I, forms
a symmetric
Frobenius algebra obje
2 (B)

elationsa commutative Frobenius algebra H = Z(S 1 ),

=a Calabi-Yau category
= objects B and
=
= O (a, b)
= = Z(Iab )
with
e additional
diffeomorphisms that =can
be
constructed
from(3.38)
theHom

(3

3.10.
To simplify
Proof
sketch:the diagrams, we define:

Z
:=

(3.39)
= (a,:=
= Hom
=O (b, a) C symmetric,
= non-degenerate:
=
: Hom
b)
O
(3.46)

which we sometimes call


the open pairing and open copairing,

(3.40)
=
=
=
dentities:
forms an algebra homomorphism:

3. The zipper

(3

(3

(3

(3.47)

=
(3.41)
forms an algebra homomorphism:

robenius relations, the left and right unit laws, and the left and
ons (3.40) and (3.37), the pairing can be shown to be symmetric

describes the knowledge about the centre, c.f. (2.12):


Lazaroiu 2000, Lauda/Pfeiffer 2005, Moore/Segal 2006

(3

(3.45)

2d open/closed TQFT: algebraic description

= (1), i.e. the interval C!n = I, forms a symmetric Frobenius algebra object.

(3

Theorem.

(3.37)
=
=
= o/c
A 2d open/closed
TQFT
Z i.e.
: Bord

Vect
C is
2. The object
!n = (1),
the interval
C!n =
I, forms
a symmetric
Frobenius algebra obje
2 (B)

elationsa commutative Frobenius algebra H = Z(S 1 ),

=a Calabi-Yau category
= objects B and
=
= O (a, b)
= = Z(Iab )
with
e additional
diffeomorphisms that =can
be
constructed
from(3.38)
theHom

(3

3.10.
To simplify
Proof
sketch:the diagrams, we define:

Z
:=

(3.39)
= (a,:=
= Hom
=O (b, a) C symmetric,
= non-degenerate:
=
: Hom
b)
O
(3.46)

which we sometimes call


the open pairing and open copairing,

(3.40)
=
=
=
dentities:

(3

Examples.
forms
an algebra homomorphism:


=

=
(3.47)
k=field, G finite
group: O = mod k[G]
= Rep(G)

3. The zipper

(3

(3

=
(3.41)
forms an algebra homomorphism:

robenius relations, the left and right unit laws, and the left and
ons (3.40) and (3.37), the pairing can be shown to be symmetric

describes the knowledge about the centre, c.f. (2.12):


Lazaroiu 2000, Lauda/Pfeiffer 2005, Moore/Segal 2006

(3

(3.45)

2d open/closed TQFT: algebraic description

= (1), i.e. the interval C!n = I, forms a symmetric Frobenius algebra object.

(3

Theorem.

(3.37)
=
=
= o/c
A 2d open/closed
TQFT
Z i.e.
: Bord

Vect
C is
2. The object
!n = (1),
the interval
C!n =
I, forms
a symmetric
Frobenius algebra obje
2 (B)

elationsa commutative Frobenius algebra H = Z(S 1 ),

=a Calabi-Yau category
= objects B and
=
= O (a, b)
= = Z(Iab )
with
e additional
diffeomorphisms that =can
be
constructed
from(3.38)
theHom

(3

3.10.
To simplify
Proof
sketch:the diagrams, we define:

Z
:=

(3.39)
= (a,:=
= Hom
=O (b, a) C symmetric,
= non-degenerate:
=
: Hom
b)
O
(3.46)

which we sometimes call


the open pairing and open copairing,

(3.40)
=
=
=
dentities:

(3

(3

Examples.
forms
an algebra homomorphism:

=
=Rep(G)
(3.47)
k=field, G finite
group: O = mod k[G] =
M complex
manifold:

=
=
(3.41)
an algebra
homomorphism:
3. The zipper
O = bounded
derived forms
category
of coherent
sheaves on M

(3

robenius relations, the left and right unit laws, and the left and
ons (3.40) and (3.37), the pairing can be shown to be symmetric

describes the knowledge about the centre, c.f. (2.12):


Lazaroiu 2000, Lauda/Pfeiffer 2005, Moore/Segal 2006

(3

(3.45)

2d open/closed TQFT: algebraic description

= (1), i.e. the interval C!n = I, forms a symmetric Frobenius algebra object.

(3

Theorem.

(3.37)
=
=
= o/c
A 2d open/closed
TQFT
Z i.e.
: Bord

Vect
C is
2. The object
!n = (1),
the interval
C!n =
I, forms
a symmetric
Frobenius algebra obje
2 (B)

elationsa commutative Frobenius algebra H = Z(S 1 ),

=a Calabi-Yau category
= objects B and
=
= O (a, b)
= = Z(Iab )
with
e additional
diffeomorphisms that =can
be
constructed
from(3.38)
theHom

(3

3.10.
To simplify
Proof
sketch:the diagrams, we define:

Z
:=

(3.39)
= (a,:=
= Hom
=O (b, a) C symmetric,
= non-degenerate:
=
: Hom
b)
O
(3.46)

which we sometimes call


the open pairing and open copairing,

(3.40)
=
=
=
dentities:

Examples.
forms
an algebra homomorphism:

=
=Rep(G)
(3.47)
k=field, G finite
group: O = mod k[G] =
M complex
manifold:

=
=
(3.41)
an algebra
homomorphism:
3. The zipper
O = bounded
derived forms
category
of coherent
sheaves on M
robenius relations, the left and right unit laws, and the left and
(3.37),
C[x1 , .the
. . ,pairing
xn ] isolated
ons (3.40)Wand
can be singularity:
shown to be symmetric
describes the knowledge about the centre, c.f. (2.12):

O = homotopy category of matrix


= factorisations
=
Lazaroiu 2000, Lauda/Pfeiffer 2005, Moore/Segal 2006

(3

(3

(3

(3

2d defect TQFT
Fix a set D2 of phases, and a set D1 of defect conditions.

2d defect TQFT
Fix a set D2 of phases, and a set D1 of defect conditions.
Bord2def(D2 , D1 ) is (roughly) the category with:
objects = disjoint unions of circles with marked points labelled by
X, Y, Z, . . . D1 , and line segments labelled by , , , . . . D2

2d defect TQFT
Fix a set D2 of phases, and a set D1 of defect conditions.
Bord2def(D2 , D1 ) is (roughly) the category with:
objects = disjoint unions of circles with marked points labelled by
X, Y, Z, . . . D1 , and line segments labelled by , , , . . . D2

morphisms = bordisms with embedded 1-dim. submanifolds S


labelled by D1

2d defect TQFT
Fix a set D2 of phases, and a set D1 of defect conditions.
Bord2def(D2 , D1 ) is (roughly) the category with:
objects = disjoint unions of circles with marked points labelled by
X, Y, Z, . . . D1 , and line segments labelled by , , , . . . D2

morphisms = bordisms with embedded 1-dim. submanifolds S


labelled by D1 ; components of \ S labelled by D2

2d defect TQFT
Fix a set D2 of phases, and a set D1 of defect conditions.
Bord2def(D2 , D1 ) is (roughly) the category with:
objects = disjoint unions of circles with marked points labelled by
X, Y, Z, . . . D1 , and line segments labelled by , , , . . . D2

morphisms = bordisms with embedded 1-dim. submanifolds S


labelled by D1 ; components of \ S labelled by D2

2d defect TQFT
Fix a set D2 of phases, and a set D1 of defect conditions.
Bord2def(D2 , D1 ) is (roughly) the category with:
objects = disjoint unions of circles with marked points labelled by
X, Y, Z, . . . D1 , and line segments labelled by , , , . . . D2

morphisms = bordisms with embedded 1-dim. submanifolds S


labelled by D1 ; components of \ S labelled by D2
O1+

O2+ Y

in

out

O1!

in

Figure 7: A world-sheet with defect lines from (O1 , O2 ) to (O1! ). The shaded regions of the
annuli indicate the subsets O1+ , O2+ and O1! . The maps in and out are defined in the
shaded regions and map the solid circle |z| = 1 to the boundary of the world-sheet.
which preserves the orientation, the boundaries, and the one-dimensional submanifolds
with their orientation.
(W.iv) out is a smooth injective isometry from the disjoint union O1! ! O2! ! ! On!
to W with the same properties as in (W.iii).
We refer to the boundary components of W in the image of in as in-going, and to those in

2d defect TQFT
Fix a set D2 of phases, and a set D1 of defect conditions.
Bord2def(D2 , D1 ) is (roughly) the category with:
objects = disjoint unions of circles with marked points labelled by
X, Y, Z, . . . D1 , and line segments labelled by , , , . . . D2

morphisms = bordisms with embedded 1-dim. submanifolds S


labelled by D1 ; components of \ S labelled by D2
O1+

O2+ Y

in

out

O1!

in

Figure 7: A world-sheet with defect lines from (O1 , O2 ) to (O1! ). The shaded regions of the
annuli indicate the subsets O1+ , O2+ and O1! . The maps in and out are defined in the
shaded regions and map the solid circle |z| = 1 to the boundary of the world-sheet.

Figure 1: When a sum over intermedia


lines, the intermediate states lie in a t

which preserves the orientation, the boundaries, and the one-dimensional submanifolds
with their orientation.
(W.iv) out is a smooth injective isometry from the disjoint union O1! ! O2! ! ! On!
to W with the same properties as in (W.iii).
We refer to the boundary components of W in the image of in as in-going, and to those in

r
|!

2d defect TQFT
Fix a set D2 of phases, and a set D1 of defect conditions.
Bord2def(D2 , D1 ) is (roughly) the category with:
objects = disjoint unions of circles with marked points labelled by
X, Y, Z, . . . D1 , and line segments labelled by , , , . . . D2

morphisms = bordisms with embedded 1-dim. submanifolds S


labelled by D1 ; components of \ S labelled by D2
O1+

O2+ Y

in

O1!

out

in
!
1

Figure 7: A world-sheet with defect lines from (O , O ) to (O ). The


shaded 1:
regions
of the a sum over intermedia
Figure
When
A 2d annuli
defect
TQFT
monoidal
indicate
the subsets is
O a
, Osymmetric
and O . The maps
and functor
are defined in the
+
1

+
2

!
1

in

out

lines, the intermediate states lie in a t

shaded regions and map the solid circle |z| = 1 to the boundary of the world-sheet.

Z : Borddef (D , D ) Vect

which preserves the orientation, the boundaries,


2and the1one-dimensional submanifolds
C
2
with their orientation.
(W.iv) out is a smooth injective isometry from the disjoint union O1! ! O2! ! ! On!
to W with the same properties as in (W.iii).
We refer to the boundary components of W in the image of in as in-going, and to those in

r
|!

2d defect TQFT
Fix a set D2 of phases, and a set D1 of defect conditions.
Bord2def(D2 , D1 ) is (roughly) the category with:
objects = disjoint unions of circles with marked points labelled by
X, Y, Z, . . . D1 , and line segments labelled by , , , . . . D2

morphisms = bordisms with embedded 1-dim. submanifolds S


labelled by D1 ; components of \ S labelled by D2
O1+

O2+ Y

in

O1!

out

in
!
1

Figure 7: A world-sheet with defect lines from (O , O ) to (O ). The


shaded 1:
regions
of the a sum over intermedia
Figure
When
A 2d annuli
defect
TQFT
monoidal
indicate
the subsets is
O a
, Osymmetric
and O . The maps
and functor
are defined in the
+
1

+
2

!
1

in

out

lines, the intermediate states lie in a t

shaded regions and map the solid circle |z| = 1 to the boundary of the world-sheet.

Z : Borddef (D , D ) Vect

which preserves the orientation, the boundaries,


2and the1one-dimensional submanifolds
C
2
with their orientation.
(W.iv) out is a smooth injective isometry from the disjoint union O1! ! O2! ! ! On!

What istoaW2d
TQFT?
withdefect
the same properties
as in (W.iii).

We refer to the boundary components of W in the image of in as in-going, and to those in

r
|!

Bicategories
A 2-category B consists of
objects , , . . .
categories B(, ) of 1-morphisms X, Y, . . . and 2-morphisms , , . . .
functors : B(, ) B(, ) B(, )
units I B(, ) and for X B(, )
I X = X = X I , (Z Y ) X = Z (Y X)

Bicategories
A bicategory B consists of
objects , , . . .
categories B(, ) of 1-morphisms X, Y, . . . and 2-morphisms , , . . .
functors : B(, ) B(, ) B(, )
units I B(, ) and for X B(, ) coherent isomorphisms
I X
=X
= X I , (Z Y ) X
= Z (Y X)

Abstracts
Bicategories

A bicategory B consists of
Equivariant
objects
, , . completion
..
Nils
Carqueville
categories B(, ) of 1-morphisms X, Y, . . . and 2-morphisms , , . . .
(joint work with Ingo Runkel)
functors : B(, ) B(, ) B(, )
units I B(, ) and for X B(, ) coherent isomorphisms
I X
=X
= X I , (Z Y ) X
= Z (Y X)
Y0
X

*
Y

basic setting of a bicategory B, which one may think of as a monoidal


h labels. We denote objects of B by a, b, . . ., generic 1-morphisms by
s by Ia , and 2-morphisms by , , . . .. Composition of 1-morphisms
, and we have two (Poincare dual) graphical depictions at our
Y

=
b

A
a

Abstracts
Bicategories

A bicategory B consists of
Equivariant
objects
, , . completion
..
Nils
Carqueville
categories B(, ) of 1-morphisms X, Y, . . . and 2-morphisms , , . . .
(joint work with Ingo Runkel)
functors : B(, ) B(, ) B(, )
units I B(, ) and for X B(, ) coherent isomorphisms
I X
=X
= X I , (Z Y ) X
= Z (Y X)
Y0
X

Poincare dual to string diagram

*
Y

basic setting of a bicategory B, which one may think of as a monoidal


h labels. We denote objects of B by a, b, . . ., generic 1-morphisms by
s by Ia , and 2-morphisms by , , . . .. Composition of 1-morphisms
, and we have two (Poincare dual) graphical depictions at our
Y

=
b

A
a

Abstracts
Bicategories

A bicategory B consists of
Equivariant
objects
, , . completion
..
Nils
Carqueville
categories B(, ) of 1-morphisms X, Y, . . . and 2-morphisms , , . . .
(joint work with Ingo Runkel)
functors : B(, ) B(, ) B(, )
units I B(, ) and for X B(, ) coherent isomorphisms
I X
=X
= X I , (Z Y ) X
= Z (Y X)
Y0
X

Poincare dual to string diagram

*
Y

basic setting of a bicategory B, which one may think of as a monoidal


h labels. We denote objects of B by a, b, . . ., generic 1-morphisms by
s by Ia , and 2-morphisms by , , . . .. Composition of 1-morphisms

, and we have two (Poincare dual) graphical


depictions
at our
Y

=
b

A
a

Abstracts
Bicategories

A bicategory B consists of
Equivariant
objects
, , . completion
..
Nils
Carqueville
categories B(, ) of 1-morphisms X, Y, . . . and 2-morphisms , , . . .
(joint work with Ingo Runkel)
functors : B(, ) B(, ) B(, )
units I B(, ) and for X B(, ) coherent isomorphisms
I X
=X
= X I , (Z Y ) X
= Z (Y X)
Y0
X

Poincare dual to string diagram

*
Y

basic setting of a bicategory B, which one may think of as a monoidal


h labels. We denote objects of B by a, b, . . ., generic 1-morphisms by
s by Ia , and 2-morphisms by , , . . .. Composition of 1-morphisms

, and we have two (Poincare dual) graphical


depictions
at our
Y

=
b

A
a

Abstracts
Bicategories

A bicategory B consists of
Equivariant
objects
, , . completion
..
Nils
Carqueville
categories B(, ) of 1-morphisms X, Y, . . . and 2-morphisms , , . . .
(joint work with Ingo Runkel)
functors : B(, ) B(, ) B(, )
units I B(, ) and for X B(, ) coherent isomorphisms
I X
=X
= X I , (Z Y ) X
= Z (Y X)
Y0
X

Poincare dual to string diagram

*
Y

basic setting of a bicategory B, which one may think of as a monoidal


h labels. We denote objects of B by a, b, . . ., generic 1-morphisms by
s by Ia , and 2-morphisms by , , . . .. Composition of 1-morphisms

, and we have two


(Poincare dual) graphical
depictions
at our
Y

=
b

A
a

Abstracts
Bicategories

A bicategory B consists of
Equivariant
objects
, , . completion
..
Nils
Carqueville
categories B(, ) of 1-morphisms X, Y, . . . and 2-morphisms , , . . .
(joint work with Ingo Runkel)
functors : B(, ) B(, ) B(, )
units I B(, ) and for X B(, ) coherent isomorphisms
I X
=X
= X I , (Z Y ) X
= Z (Y X)
Y0
X

Poincare dual to string diagram

*
Y

Y0

basic setting of a bicategory B, which one may think of as a monoidal


h labels. We denote objects of B by a, b, . . ., generic 1-morphisms by
s by Ia , and 2-morphisms by , , . . .. Composition of 1-morphisms
e dual) graphical

, and we have two


(Poincar
depictions
at our
Y

=
b

A
a

Abstracts
Bicategories

A bicategory B consists of
Equivariant
objects
, , . completion
..
Nils
Carqueville
categories B(, ) of 1-morphisms X, Y, . . . and 2-morphisms , , . . .
(joint work with Ingo Runkel)
functors : B(, ) B(, ) B(, )
units I B(, ) and for X B(, ) coherent isomorphisms
I X
=X
= X I , (Z Y ) X
= Z (Y X)
Y0
X

Poincare dual to string diagram

*
Y

Y0

basic setting of a bicategory B, which one may think of as a monoidal


h labels. We denote objects of B by a, b, . . ., generic 1-morphisms by
s by Ia , and 2-morphisms by , , . . .. Composition of 1-morphisms
e dual) graphical

, and we have two


(Poincar
depictions
at our
Y

=
b

A
a

Abstracts
Bicategories

A bicategory B consists of
Equivariant
objects
, , . completion
..
Nils
Carqueville
categories B(, ) of 1-morphisms X, Y, . . . and 2-morphisms , , . . .
(joint work with Ingo Runkel)
functors : B(, ) B(, ) B(, )
units I B(, ) and for X B(, ) coherent isomorphisms
I X
=X
= X I , (Z Y ) X
= Z (Y X)
Y0
X

Poincare dual to string diagram

*
Y

Y0

basic setting of a bicategory B, which one may think of as a monoidal


h labels. We denote objects of B by a, b, . . ., generic 1-morphisms by
s by Ia , and 2-morphisms by , , . . .. Composition of 1-morphisms
e dual) graphical

, and we have two


(Poincar
depictions
at our
Y

=
b

A
a

Adjunction
X B(, ) has a right adjoint if there is X B(, )

Adjunction
X B(, ) has a right adjoint if there is X B(, ) and
X

= evX : X X I

= coevX : I X X

Adjunction
X B(, ) has a right adjoint if there is X B(, ) and
X

= evX : X X I

= coevX : I X X

such that the Zorro moves hold:

Zorro

Adjunction
X B(, ) has a right adjoint if there is X B(, ) and
X

= evX : X X I

= coevX : I X X

such that the Zorro moves hold:


=

Adjunction
X B(, ) has a right adjoint if there is X B(, ) and
X

= evX : X X I

= coevX : I X X

such that the Zorro moves hold:


=

If X is also a left adjoint then the quantum dimensions are

diml (X) =

Adjunction
X B(, ) has a right adjoint if there is X B(, ) and
X

= evX : X X I

= coevX : I X X

such that the Zorro moves hold:


=

If X is also a left adjoint then the quantum dimensions are

End(I )

diml (X) =

Adjunction
X B(, ) has a right adjoint if there is X B(, ) and
X

= evX : X X I

= coevX : I X X

such that the Zorro moves hold:


=

If X is also a left adjoint then the quantum dimensions are

End(I )

diml (X) =

End(I )

dimr (X) =
X

Adjunction
X B(, ) has a right adjoint if there is X B(, ) and
X

= evX : X X I

= coevX : I X X

such that the Zorro moves hold:


=

If X is also a left adjoint then the quantum dimensions are

End(I )

diml (X) =

End(I )

dimr (X) =

B is pivotal if X
= X coherently for all X.

Defect bicategories
Theorem.
A TQFT Z : Borddef
2 (D2 , D1 ) VectC gives a pivotal bicategory BZ .

Davydov/Kong/Runkel 2011

Defect bicategories
Theorem.
A TQFT Z : Borddef
2 (D2 , D1 ) VectC gives a pivotal bicategory BZ .
Proof sketch:
objects = elements of D2 = theories

Davydov/Kong/Runkel 2011

Defect bicategories
Theorem.
A TQFT Z : Borddef
2 (D2 , D1 ) VectC gives a pivotal bicategory BZ .

Proof sketch:
objects = elements of D2 = theories
1-morphisms X : = (lists of) elements of D1 = defect lines:

...
Xn

Davydov/Kong/Runkel 2011

2
X3

1
X2

X1

Defect bicategories
Theorem.
A TQFT Z : Borddef
2 (D2 , D1 ) VectC gives a pivotal bicategory BZ .

Proof sketch:
objects = elements of D2 = theories
1-morphisms X : = (lists of) elements of D1 = defect lines:

...
Xn

2
X3

1
X2

...

X1

2-morphisms = junction fields: Hom(X, Y ) = Z

Y2
Y1
X1
X2

Xn
...

Davydov/Kong/Runkel 2011

Ym

Defect bicategories
Theorem.
A TQFT Z : Borddef
2 (D2 , D1 ) VectC gives a pivotal bicategory BZ .

Proof sketch:
objects = elements of D2 = theories
1-morphisms X : = (lists of) elements of D1 = defect lines:
...

Xn

X3

X2

2-morphisms = junction fields: Hom(X, Y ) = Z

Davydov/Kong/Runkel 2011

Y2
Y1

Ym

X1
X2

Xn
...

Fact. Correlators of Z are string diagrams in BZ !


Y0

...

X1

Generalised orbifolds: main idea


Main result.
Let B be pivotal, X B(, ) have invertible dimr (X), set A = X X.

Carqueville/Runkel 2012

Generalised orbifolds: main idea


Main result.
Let B be pivotal, X B(, ) have invertible dimr (X), set A = X X.
Everything about theory can be recovered from defect A.

Carqueville/Runkel 2012

Generalised orbifolds: main idea


Main result.
Let B be pivotal, X B(, ) have invertible dimr (X), set A = X X.
Everything about theory can be recovered from defect A.
Think of A as generalised symmetry.

Carqueville/Runkel 2012

Generalised orbifolds: main idea


Main result.
Let B be pivotal, X B(, ) have invertible dimr (X), set A = X X.
Everything about theory can be recovered from defect A.
Think of A as generalised symmetry.
Introducing X-bubbles in -correlator is scaling by dimr (X). Blowing up
all X-bubbles produces -correlator with A-defect network.
!

Carqueville/Runkel 2012

Generalised orbifolds: main idea


Main result.
Let B be pivotal, X B(, ) have invertible dimr (X), set A = X X.
Everything about theory can be recovered from defect A.
Think of A as generalised symmetry.
Introducing X-bubbles in -correlator is scaling by dimr (X). Blowing up
all X-bubbles produces -correlator with A-defect network.
!

Carqueville/Runkel 2012

Generalised orbifolds: main idea


Main result.
Let B be pivotal, X B(, ) have invertible dimr (X), set A = X X.
Everything about theory can be recovered from defect A.
Think of A as generalised symmetry.
Introducing X-bubbles in -correlator is scaling by dimr (X). Blowing up
all X-bubbles produces -correlator with A-defect network.
!

Carqueville/Runkel 2012

=Z

Generalised orbifolds: main idea


Main result.
Let B be pivotal, X B(, ) have invertible dimr (X), set A = X X.
Everything about theory can be recovered from defect A.
Think of A as generalised symmetry.
Introducing X-bubbles in -correlator is scaling by dimr (X). Blowing up
all X-bubbles produces -correlator with A-defect network.
!

=Z

=Z

Carqueville/Runkel 2012

Generalised orbifolds
Let B be pivotal. Its equivariant completion Beq

Carqueville/Runkel 2012, Fr
ohlich/Fuchs/Runkel/Schweigert 2009, (Fjelstad/Fr
ohlich/)Fuchs/Runkel/Schweigert 200x

Generalised orbifolds
Let B be pivotal. Its equivariant completion Beq has:
objects: pairs (, A) with B and A B(, ) separable
Frobenius algebra:
: AA A

: I A

Carqueville/Runkel 2012, Fr
ohlich/Fuchs/Runkel/Schweigert 2009, (Fjelstad/Fr
ohlich/)Fuchs/Runkel/Schweigert 200x

Generalised orbifolds
Let B be pivotal. Its equivariant completion Beq has:
objects: pairs (, A) with B and A B(, ) separable
Frobenius algebra:
: AA A

: I A

= =

Carqueville/Runkel 2012, Fr
ohlich/Fuchs/Runkel/Schweigert 2009, (Fjelstad/Fr
ohlich/)Fuchs/Runkel/Schweigert 200x

Generalised orbifolds
Let B be pivotal. Its equivariant completion Beq has:
objects: pairs (, A) with B and A B(, ) separable
Frobenius algebra:
: AA A
=

: I A
=

= =

Carqueville/Runkel 2012, Fr
ohlich/Fuchs/Runkel/Schweigert 2009, (Fjelstad/Fr
ohlich/)Fuchs/Runkel/Schweigert 200x

Generalised orbifolds
Let B be pivotal. Its equivariant completion Beq has:
objects: pairs (, A) with B and A B(, ) separable
Frobenius algebra:
: AA A
=

: I A
=

= =

(encoding Pachner moves)

Carqueville/Runkel 2012, Fr
ohlich/Fuchs/Runkel/Schweigert 2009, (Fjelstad/Fr
ohlich/)Fuchs/Runkel/Schweigert 200x

Generalised orbifolds
Let B be pivotal. Its equivariant completion Beq has:
objects: pairs (, A) with B and A B(, ) separable
Frobenius algebra:
: AA A
=

: I A
=

= =

(encoding Pachner moves)

1-morphisms: X B(, ) that are bimodules


X
X

Carqueville/Runkel 2012, Fr
ohlich/Fuchs/Runkel/Schweigert 2009, (Fjelstad/Fr
ohlich/)Fuchs/Runkel/Schweigert 200x

Generalised orbifolds
Let B be pivotal. Its equivariant completion Beq has:
objects: pairs (, A) with B and A B(, ) separable
Frobenius algebra:
: AA A
=

: I A
=

= =

(encoding Pachner moves)

1-morphisms: X B(, ) that are bimodules


X

Carqueville/Runkel 2012, Fr
ohlich/Fuchs/Runkel/Schweigert 2009, (Fjelstad/Fr
ohlich/)Fuchs/Runkel/Schweigert 200x

Generalised orbifolds
Let B be pivotal. Its equivariant completion Beq has:
objects: pairs (, A) with B and A B(, ) separable
Frobenius algebra:
: AA A
=

: I A
=

= =

(encoding Pachner moves)

1-morphisms: X B(, ) that are bimodules


X

Carqueville/Runkel 2012, Fr
ohlich/Fuchs/Runkel/Schweigert 2009, (Fjelstad/Fr
ohlich/)Fuchs/Runkel/Schweigert 200x

Generalised orbifolds
Let B be pivotal. Its equivariant completion Beq has:
objects: pairs (, A) with B and A B(, ) separable
Frobenius algebra:
: AA A
=

: I A
=

= =

(encoding Pachner moves)

1-morphisms: X B(, ) that are bimodules


X

horizontal composition: tensor product over algebra


Carqueville/Runkel 2012, Fr
ohlich/Fuchs/Runkel/Schweigert 2009, (Fjelstad/Fr
ohlich/)Fuchs/Runkel/Schweigert 200x

Generalised orbifolds
Let B be pivotal. Its equivariant completion Beq has:
objects: pairs (, A) with B and A B(, ) separable
Frobenius algebra:
: AA A
=

: I A
=

= =

(encoding Pachner moves)

1-morphisms: X B(, ) that are bimodules


X

horizontal composition: tensor product over algebra


2-morphisms: Hom(X, Y ) that are bimodule maps

Carqueville/Runkel 2012, Fr
ohlich/Fuchs/Runkel/Schweigert 2009, (Fjelstad/Fr
ohlich/)Fuchs/Runkel/Schweigert 200x

Constructing separable Frobenius algebras


Lemma.


Let G be finite group and let Dg B(, ) gG define group action
() Dg : B(, ) B(, )

Dg Dg 0
= Dgg0

Constructing separable Frobenius algebras


Lemma.


Let G be finite group and let Dg B(, ) gG define group action
() Dg : B(, ) B(, )

Then we have a separable Frobenius algebra


M
AG =
Dg
gG

Dg Dg 0
= Dgg0

Constructing separable Frobenius algebras


Lemma.


Let G be finite group and let Dg B(, ) gG define group action
() Dg : B(, ) B(, )

Dg Dg 0
= Dgg0

Then we have a separable Frobenius algebra


M

and
mod(AG ) Beq (, AG ), (, I )
AG =
Dg
= B(, )G
gG

Constructing separable Frobenius algebras


Lemma.


Let G be finite group and let Dg B(, ) gG define group action
() Dg : B(, ) B(, )

Dg Dg 0
= Dgg0

Then we have a separable Frobenius algebra


M

and
mod(AG ) Beq (, AG ), (, I )
AG =
Dg
= B(, )G
gG

Theorem.
Assume X B(, ) has invertible dimr (X).

A = X X is separable Frobenius algebra in B(, ).

Carqueville/Runkel 2012

Constructing separable Frobenius algebras


Lemma.


Let G be finite group and let Dg B(, ) gG define group action
() Dg : B(, ) B(, )

Dg Dg 0
= Dgg0

Then we have a separable Frobenius algebra


M

and
mod(AG ) Beq (, AG ), (, I )
AG =
Dg
= B(, )G
gG

Theorem.
Assume X B(, ) has invertible dimr (X).

A = X X is separable Frobenius algebra in B(, ).

X : (, A)  (, I ) : X are mutually inverse in Beq .

Carqueville/Runkel 2012

Constructing separable Frobenius algebras


Lemma.


Let G be finite group and let Dg B(, ) gG define group action
() Dg : B(, ) B(, )

Dg Dg 0
= Dgg0

Then we have a separable Frobenius algebra


M

and
mod(AG ) Beq (, AG ), (, I )
AG =
Dg
= B(, )G
gG

Theorem.
Assume X B(, ) has invertible dimr (X).

A = X X is separable Frobenius algebra in B(, ).

X : (, A)  (, I ) : X are mutually inverse in Beq .


We say that , are orbifold equivalent, .
Carqueville/Runkel 2012

Applications
Theorem.
There is a pivotal bicategory L (of Landau-Ginzburg models)

Carqueville/Murfet 2012

Applications
Theorem.
There is a pivotal bicategory L (of Landau-Ginzburg models) with
objects = isolated singularities W C[x1 , . . . , xn ] C[x]

Carqueville/Murfet 2012

Applications
Theorem.
There is a pivotal bicategory L (of Landau-Ginzburg models) with
objects = isolated singularities W C[x1 , . . . , xn ] C[x]
1-morphisms X : W (x) V (z) are matrix factorisations of
V W , i. e. X Mat(C[x, z]) odd with X 2 = (V W ) id

Carqueville/Murfet 2012

Applications
Theorem.
There is a pivotal bicategory L (of Landau-Ginzburg models) with
objects = isolated singularities W C[x1 , . . . , xn ] C[x]
1-morphisms X : W (x) V (z) are matrix factorisations of
V W , i. e. X Mat(C[x, z]) odd with X 2 = (V W ) id, and
"
 Q
 #
Q
str
i xi X
j zj X dx
dimr (X) = Res
x1 W . . . xn W

Carqueville/Murfet 2012

Applications
Theorem.
There is a pivotal bicategory L (of Landau-Ginzburg models) with
objects = isolated singularities W C[x1 , . . . , xn ] C[x]
1-morphisms X : W (x) V (z) are matrix factorisations of
V W , i. e. X Mat(C[x, z]) odd with X 2 = (V W ) id, and
"
 Q
 #
Q
str
i xi X
j zj X dx
dimr (X) = Res
x1 W . . . xn W
Theorem. L(0, W )
= Db (CQ) for simple singularity W and Dynkin
quiver Q of same ADE type.

Carqueville/Murfet 2012, Kajiura/Saito/Takahashi 2005

Applications
Theorem.
There is a pivotal bicategory L (of Landau-Ginzburg models) with
objects = isolated singularities W C[x1 , . . . , xn ] C[x]
1-morphisms X : W (x) V (z) are matrix factorisations of
V W , i. e. X Mat(C[x, z]) odd with X 2 = (V W ) id, and
"
 Q
 #
Q
str
i xi X
j zj X dx
dimr (X) = Res
x1 W . . . xn W
Theorem. L(0, W )
= Db (CQ) for simple singularity W and Dynkin
quiver Q of same ADE type.
152

150

154

I Singularity Theory

I Singularity Theory

I Singularity Theory

A1

A2

A3

A4D4

D5

D6

ED67

E7

E8

Fig. 2.10. Real pictures of two-dimensional Ak -singularities Fig. 2.11. Real pictures of two-dimensional Dk -singularities
Fig. 2.12. Real pictures of two-dimensional Ek -singularities

Remark 2.50.2.
During the following classification
we shall2.53.
makeLet
several
Theorem 2.48. Let f m2 C{x} and k 1, then the following
are equivTheorem
f times
m3 C{x, y}. The following are equivalent:
use of the so-called Tschirnhaus transformation: let(3)
A be a ring and
alent:
has a unique linear factor (of multiplicity 3) and (f ) 8,
(a) f
r 3
d
d1
(3)
(3)
(a) crk(f ) 1 and (f ) = k,
+ ... +
0f A[x]
f = d x + d1 x
x and if f = x3 then f
/ %x, y 2 &3 = %x3 , x2 y 2 , xy 4 , y 6 &.
(b)
r
r
(b) f xk+1
+ x22 + . . . + x2n , that is, f is of type Ak ,
(c) f g with g {x3 + y 4 , x3 + xy 3 , x3 + y 5 }, that is, f is of type E6 , E7
1
c k+1
a
polynomial
of
degree
d
with
coefficients
in
A.
Assume
that
the
quotient
2
2
(c) f x1 + x2 + . . . + xn .
E8 .
cx yields a polynomial
:= d1 /(dd ) exists in A. Then, substituting(d)
x by
g with g {x3 + y 4 , x3 + xy 3 , x3 + y 5 }.
f
Moreover, f is of type A2012,
) = 0. It is of type Ak offordegree
some2005
2 iff
1 iff crk(f
dk with
no term of degree d 1. In other words, the isomorphism
Carqueville/Murfet
Kajiura/Saito/Takahashi

Applications
Theorem.
There is a pivotal bicategory L (of Landau-Ginzburg models) with
objects = isolated singularities W C[x1 , . . . , xn ] C[x]
1-morphisms X : W (x) V (z) are matrix factorisations of
V W , i. e. X Mat(C[x, z]) odd with X 2 = (V W ) id, and
"
 Q
 #
Q
str
i xi X
j zj X dx
dimr (X) = Res
x1 W . . . xn W
Theorem. L(0, W )
= Db (CQ) for simple singularity W and Dynkin
quiver Q of same ADE type.
Theorem.
x`1 + x1 x22 z12` + z22
x31

x31 + x42 z112 + z22


+ x1 x2

x31

x52

z118
z130

z22
z22

D`+1 A2`1

E6 A11




Db (CQD`+1 )
= Db (CQA2`1 )
Db (CQE )
= Db (CQA )
6

11

E7 A17 D (CQE7 )
= Db (CQA17 )

E8 A29 Db (CQE8 )
= Db (CQA29 )
b

Carqueville/Murfet 2012, Kajiura/Saito/Takahashi 2005, Carqueville/Runkel 2012, Carqueville/Ros Camacho/Runkel 2013

Group
Flavio Montiel Montoya (Phd since October 2014):
Landau-Ginzburg models and matrix factorisations
categorified link invariants (`a la Khovanov-Rozansky)

Group
Flavio Montiel Montoya (Phd since October 2014):
Landau-Ginzburg models and matrix factorisations
categorified link invariants (`a la Khovanov-Rozansky)

Gregor Schaumann (postdoc starting October 2015):


fusion categories and tricategories
3d defect TQFT

Group
Flavio Montiel Montoya (Phd since October 2014):
Landau-Ginzburg models and matrix factorisations
categorified link invariants (`a la Khovanov-Rozansky)

Gregor Schaumann (postdoc starting October 2015):


fusion categories and tricategories
3d defect TQFT
??? (12 Phds starting October 2015):
Calabi-Yau completions?
2d/4d correspondences?
3d TQFT?

Theorem. (Orbifold equivalence )


Assume X B(, ) has invertible dimr (X). Then:

A = X X is separable Frobenius algebra in B(, ).


X : (, A)  (, I ) : X are mutually inverse in Beq .


B ,
= Beq (, I ), (, A) for all B.
!

=Z

=Z

Vous aimerez peut-être aussi