Universal torsion, L
2-invariants, polytopes and the Thurston norm
Wolfgang Lück Bonn Germany
email wolfgang.lueck@him.uni-bonn.de http://131.220.77.52/lueck/
Münster, December, 2015
Review of classical L
2-invariants
LetG→X →X be aG-covering of a connected finite CW-complexX.
The cellular chain complex ofX is a finitely generated free ZG-chain complex:
· · ·−c−−n−1→M
In
ZG−→cn M
in−1
ZG−c−−n−1→ · · · The associatedL2-chain complex
C∗(2)(X):=L2(G)⊗ZGC∗(X)
has Hilbert spaces with isometric linearG-action as chain modules and boundedG-equivariant operators as differentials
· · · c
(2)
−−−n−1→M
I
L2(G) c
(2)
−−→n M
i
L2(G) c
(2)
−−−n−1→ · · ·
Review of classical L
2-invariants
LetG→X →X be aG-covering of a connected finite CW-complexX.
The cellular chain complex ofX is a finitely generated free ZG-chain complex:
· · ·−c−−n−1→M
In
ZG−→cn M
in−1
ZG−c−−n−1→ · · · The associatedL2-chain complex
C∗(2)(X):=L2(G)⊗ZGC∗(X)
has Hilbert spaces with isometric linearG-action as chain modules and boundedG-equivariant operators as differentials
· · · c
(2)
−−−n−1→M
I
L2(G) c
(2)
−−→n M
i
L2(G) c
(2)
−−−n−1→ · · ·
Definition (L2-homology andL2-Betti numbers) Define then-thL2-homologyto be the Hilbert space
Hn(2)(X):=ker(cn(2))/im(c(2)n+1).
Define then-thL2-Betti number
b(2)n (X):=dimN(G) Hn(2)(X)
∈R≥0.
The original notion is due toAtiyahand was motivated by index theory. He defined for aG-coveringM→M of a closed
Riemannian manifold b(2)n (M) := lim
t→∞
Z
F
tr e−t·∆n(x,x)
dvolM.
IfGis finite, we have
bn(2)(X) = 1
|G|·bn(X).
IfG=Z, we have
b(2)n (X) =dimC[Z](0) C[Z](0)⊗C[Z]Hn(X;C)
∈Z.
The original notion is due toAtiyahand was motivated by index theory. He defined for aG-coveringM→M of a closed
Riemannian manifold b(2)n (M) := lim
t→∞
Z
F
tr e−t·∆n(x,x)
dvolM.
IfGis finite, we have
bn(2)(X) = 1
|G|·bn(X).
IfG=Z, we have
b(2)n (X) =dimC[Z](0) C[Z](0)⊗C[Z]Hn(X;C)
∈Z.
The original notion is due toAtiyahand was motivated by index theory. He defined for aG-coveringM→M of a closed
Riemannian manifold b(2)n (M) := lim
t→∞
Z
F
tr e−t·∆n(x,x)
dvolM.
IfGis finite, we have
bn(2)(X) = 1
|G|·bn(X).
IfG=Z, we have
b(2)n (X) =dimC[Z](0) C[Z](0)⊗C[Z]Hn(X;C)
∈Z.
In the sequel3-manifoldmeans a prime connected compact orientable 3-manifold with infinite fundamental group whose boundary is empty or a union of tori and which is notS1×D2or S1×S2.
Theorem (Lott-Lück)
For every3-manifold M all L2-Betti numbers bn(2)(M)e vanish.
We are interested in the case where allL2-Betti numbers vanish, since then a very powerful secondary invariant comes into play, the so calledL2-torsion.
In the sequel3-manifoldmeans a prime connected compact orientable 3-manifold with infinite fundamental group whose boundary is empty or a union of tori and which is notS1×D2or S1×S2.
Theorem (Lott-Lück)
For every3-manifold M all L2-Betti numbers bn(2)(M)e vanish.
We are interested in the case where allL2-Betti numbers vanish, since then a very powerful secondary invariant comes into play, the so calledL2-torsion.
In the sequel3-manifoldmeans a prime connected compact orientable 3-manifold with infinite fundamental group whose boundary is empty or a union of tori and which is notS1×D2or S1×S2.
Theorem (Lott-Lück)
For every3-manifold M all L2-Betti numbers bn(2)(M)e vanish.
We are interested in the case where allL2-Betti numbers vanish, since then a very powerful secondary invariant comes into play, the so calledL2-torsion.
L2-torsion can be defined analytical in terms of the spectrum of the Laplace operator, generalizinganalytic Ray-Singer torsion. It can also be defined in terms of the cellularZG-chain complex, generalizingReidemeister torsion.
The definition ofL2-torsion is based on the notion of the Fuglede-Kadison determinantwhich is a generalization of the classical determinant to the infinite-dimensional setting. It is defined for a boundedG-equivariant operatorf:L2(G)m→L2(G)n to be the non-negative real number
det(2)(f)=exp 1
2 · Z
ln(λ)dνf∗f
∈R>0
whereνf∗f is the spectral measure of the positive operatorf∗f. IfGis finite andm=n, then det(2)(f) =|det(f)|1/|G|.
L2-torsion can be defined analytical in terms of the spectrum of the Laplace operator, generalizinganalytic Ray-Singer torsion. It can also be defined in terms of the cellularZG-chain complex, generalizingReidemeister torsion.
The definition ofL2-torsion is based on the notion of the Fuglede-Kadison determinantwhich is a generalization of the classical determinant to the infinite-dimensional setting. It is defined for a boundedG-equivariant operatorf:L2(G)m→L2(G)n to be the non-negative real number
det(2)(f)=exp 1
2 · Z
ln(λ)dνf∗f
∈R>0
whereνf∗f is the spectral measure of the positive operatorf∗f. IfGis finite andm=n, then det(2)(f) =|det(f)|1/|G|.
L2-torsion can be defined analytical in terms of the spectrum of the Laplace operator, generalizinganalytic Ray-Singer torsion. It can also be defined in terms of the cellularZG-chain complex, generalizingReidemeister torsion.
The definition ofL2-torsion is based on the notion of the Fuglede-Kadison determinantwhich is a generalization of the classical determinant to the infinite-dimensional setting. It is defined for a boundedG-equivariant operatorf:L2(G)m→L2(G)n to be the non-negative real number
det(2)(f)=exp 1
2 · Z
ln(λ)dνf∗f
∈R>0
whereνf∗f is the spectral measure of the positive operatorf∗f. IfGis finite andm=n, then det(2)(f) =|det(f)|1/|G|.
Definition (L2-torsion)
Suppose thatX isL2-acyclic, i.e., allL2-Betti numbersbn(2)(X)vanish.
Let∆(2)n :Cn(2)(X)→Cn(2)(X)be then-Laplace operatorgiven by c(2)n+1◦ cn(2)∗
+ cn−1(2) ∗
◦c(2)n . Define theL2-torsion
ρ(2)(X):= 1 2 ·X
n≥0
(−1)n·n·ln det(2)(∆(2)n )
∈R.
Theorem (Lück-Schick)
Let M be a3-manifold. Let M1, M2, . . . , Mmbe the hyperbolic pieces in its Jaco-Shalen decomposition.
Then
ρ(2)(M) :=e − 1 6π ·
m
X
i=1
vol(Mi).
Universal L
2-torsion
Definition (K1w(ZG))
LetK1w(ZG)be the abelian group given by:
generators
Iff:ZGm →ZGmis aZG-map such that the induced bounded G-equivariantL2(G)m →L2(G)m map is a weak isomorphism, i.e., the dimensions of its kernel and cokernel are trivial, then it
determines a generator[f]inK1w(ZG).
relations
f1 ∗ 0 f2
= [f1] + [f2];
[g◦f] = [f] + [g].
DefineWhw(G):=K1w(ZG)/{±g |g∈G}.
Universal L
2-torsion
Definition (K1w(ZG))
LetK1w(ZG)be the abelian group given by:
generators
Iff:ZGm →ZGmis aZG-map such that the induced bounded G-equivariantL2(G)m →L2(G)m map is a weak isomorphism, i.e., the dimensions of its kernel and cokernel are trivial, then it
determines a generator[f]inK1w(ZG).
relations
f1 ∗ 0 f2
= [f1] + [f2];
[g◦f] = [f] + [g].
DefineWhw(G):=K1w(ZG)/{±g |g∈G}.
Definition (Weak chain contraction)
Consider aZG-chain complexC∗. Aweak chain contraction(γ∗,u∗)for C∗ consists of aZG-chain mapu∗:C∗→C∗ and aZG-chain
homotopyγ∗:u∗'0∗such thatu∗(2):C∗(2)→C∗(2)is a weak
isomorphism for alln∈Zandγn◦un =un+1◦γnholds for alln∈Z.
Definition (UniversalL2-torsion)
LetC∗ be a finite based freeZG-chain complex such that C∗(2)is L2-acyclic. Define itsuniversalL2-torsion
ρ(2)u (C∗)∈Ke1w(ZG) by
ρ(2)u (C∗) = [(uc+γ)odd]−[uodd], where(γ∗,u∗)is any weak chain contraction ofC∗.
Definition (Weak chain contraction)
Consider aZG-chain complexC∗. Aweak chain contraction(γ∗,u∗)for C∗ consists of aZG-chain mapu∗:C∗→C∗ and aZG-chain
homotopyγ∗:u∗'0∗such thatu∗(2):C∗(2)→C∗(2)is a weak
isomorphism for alln∈Zandγn◦un =un+1◦γnholds for alln∈Z.
Definition (UniversalL2-torsion)
LetC∗ be a finite based freeZG-chain complex such that C∗(2)is L2-acyclic. Define itsuniversalL2-torsion
ρ(2)u (C∗)∈Ke1w(ZG) by
ρ(2)u (C∗) = [(uc+γ)odd]−[uodd], where(γ∗,u∗)is any weak chain contraction ofC∗.
AnadditiveL2-torsion invariant(A,a)consists of an abelian group Aand an assignment which associates to a finite based free ZG-chain complexC∗, for whichC∗(2)isL2-acyclic, an element a(C∗)∈Asuch that for any based exact short sequence of such ZG-chain complexes 0→C∗ →D∗→E∗ →0 we get
a(D∗) =a(C∗) +a(E∗), and we havea · · · →0→ZG−−→±id ZG→0→ · · ·
=0.
We call an additiveL2-torsion invariant(U,u)universalif for every additiveL2-torsion invariant(A,a)there is precisely one group homomorphismf:U →Asatisfyingf(u(C∗)) =a(C∗)for any such ZG-chain complex.
Then(K1w(ZG), ρ(2)u )is the universal additiveL2-torsion invariant.
AnadditiveL2-torsion invariant(A,a)consists of an abelian group Aand an assignment which associates to a finite based free ZG-chain complexC∗, for whichC∗(2)isL2-acyclic, an element a(C∗)∈Asuch that for any based exact short sequence of such ZG-chain complexes 0→C∗ →D∗→E∗ →0 we get
a(D∗) =a(C∗) +a(E∗), and we havea · · · →0→ZG−−→±id ZG→0→ · · ·
=0.
We call an additiveL2-torsion invariant(U,u)universalif for every additiveL2-torsion invariant(A,a)there is precisely one group homomorphismf:U →Asatisfyingf(u(C∗)) =a(C∗)for any such ZG-chain complex.
Then(K1w(ZG), ρ(2)u )is the universal additiveL2-torsion invariant.
AnadditiveL2-torsion invariant(A,a)consists of an abelian group Aand an assignment which associates to a finite based free ZG-chain complexC∗, for whichC∗(2)isL2-acyclic, an element a(C∗)∈Asuch that for any based exact short sequence of such ZG-chain complexes 0→C∗ →D∗→E∗ →0 we get
a(D∗) =a(C∗) +a(E∗), and we havea · · · →0→ZG−−→±id ZG→0→ · · ·
=0.
We call an additiveL2-torsion invariant(U,u)universalif for every additiveL2-torsion invariant(A,a)there is precisely one group homomorphismf:U →Asatisfyingf(u(C∗)) =a(C∗)for any such ZG-chain complex.
Then(K1w(ZG), ρ(2)u )is the universal additiveL2-torsion invariant.
The universalL2-torsion is asimple homotopy invariant.
It satisfies usefulsum formulasandproduct formulas. There are also formulas for appropriatefibrationsandS1-actions.
IfGis finite, we rediscover essentially the classicalReidemeister torsion.
We haveρ(2(Sf1) = (z−1)in Whw(Z)∼=Q(z±1)×/{±zn|n∈Z}.
We haveρ(2(Tfn) =0 forn≥2.
The universalL2-torsion is asimple homotopy invariant.
It satisfies usefulsum formulasandproduct formulas. There are also formulas for appropriatefibrationsandS1-actions.
IfGis finite, we rediscover essentially the classicalReidemeister torsion.
We haveρ(2(Sf1) = (z−1)in Whw(Z)∼=Q(z±1)×/{±zn|n∈Z}.
We haveρ(2(Tfn) =0 forn≥2.
The universalL2-torsion is asimple homotopy invariant.
It satisfies usefulsum formulasandproduct formulas. There are also formulas for appropriatefibrationsandS1-actions.
IfGis finite, we rediscover essentially the classicalReidemeister torsion.
We haveρ(2(Sf1) = (z−1)in Whw(Z)∼=Q(z±1)×/{±zn|n∈Z}.
We haveρ(2(Tfn) =0 forn≥2.
Theorem (Jaco-Shalen-Johannson decomposition)
Let M be a compact connected orientable irreducible3-manifold with infinite fundamental group whose boundary is empty or toroidal. Let M1, M2, . . . , Mr be its pieces in the Jaco-Shalen-Johannson
decomposition. Let ji:π1(Mi)→π1(M)be the injection induced by the inclusion Mi →M.
Then each Mi and M are L2-acyclic and we have
ρ(2)u (M) =e
r
X
i=1
(ji)∗ ρ(2)u (Mfi) .
Many other invariants come from the universalL2-torsion by applying a homomorphismK1w(ZG)→Aof abelian groups.
For instance, the Fuglede-Kadison determinant defines a homomorphism
det(2): Whw(ZG)→R
which maps the universalL2-torsionρ(2)u (X)to the (classical) L2-torsionρ(2)(X).
The fundamental square and the Atiyah Conjecture
Thefundamental squareis given by the following inclusions of rings
ZG //
N(G)
D(G) //U(G)
U(G)is thealgebra of affiliated operators. Algebraically it is just theOre localizationofN(G)with respect to the multiplicatively closed subset of non-zero divisors.
D(G)is thedivision closureofZGinU(G), i.e., the smallest subring ofU(G)containingZGsuch that every element inD(G), which is a unit inU(G), is already a unit inD(G)itself.
The fundamental square and the Atiyah Conjecture
Thefundamental squareis given by the following inclusions of rings
ZG //
N(G)
D(G) //U(G)
U(G)is thealgebra of affiliated operators. Algebraically it is just theOre localizationofN(G)with respect to the multiplicatively closed subset of non-zero divisors.
D(G)is thedivision closureofZGinU(G), i.e., the smallest subring ofU(G)containingZGsuch that every element inD(G), which is a unit inU(G), is already a unit inD(G)itself.
The fundamental square and the Atiyah Conjecture
Thefundamental squareis given by the following inclusions of rings
ZG //
N(G)
D(G) //U(G)
U(G)is thealgebra of affiliated operators. Algebraically it is just theOre localizationofN(G)with respect to the multiplicatively closed subset of non-zero divisors.
D(G)is thedivision closureofZGinU(G), i.e., the smallest subring ofU(G)containingZGsuch that every element inD(G), which is a unit inU(G), is already a unit inD(G)itself.
IfGis finite, its is given by
ZG //
CG
id
QG //CG
IfG=Z, it is given by
Z[Z] //
L∞(S1)
Q[Z](0) //L(S1)
IfGis finite, its is given by
ZG //
CG
id
QG //CG
IfG=Z, it is given by
Z[Z] //
L∞(S1)
Q[Z](0) //L(S1)
IfGis elementary amenable torsionfree, thenD(G)can be identified with the Ore localization ofZGwith respect to the multiplicatively closed subset of non-zero elements.
In general the Ore localization does not exist and in these cases D(G)is the right replacement.
Conjecture (Atiyah Conjecture for torsionfree groups)
Let G be a torsionfree group. It satisfies theAtiyah ConjectureifD(G) is a skew-field.
Fix a natural numberd ≥5. Then a finitely generated torsionfree groupGsatisfies the Atiyah Conjecture if and only if for any G-coveringM→M of a closed Riemannian manifold of dimension d we haveb(2)n (M)∈Zfor everyn≥0.
The Atiyah Conjecture implies for a torsionfree groupGthat the rational group ring has no non-trivial zero-divisors.
Notice that the Farrell-Jones Conjecture implies for a torsionfree groupGthat the group ring over any field of characteristic zero has no non-trivial idempotents.
Conjecture (Atiyah Conjecture for torsionfree groups)
Let G be a torsionfree group. It satisfies theAtiyah ConjectureifD(G) is a skew-field.
Fix a natural numberd ≥5. Then a finitely generated torsionfree groupGsatisfies the Atiyah Conjecture if and only if for any G-coveringM→M of a closed Riemannian manifold of dimension d we haveb(2)n (M)∈Zfor everyn≥0.
The Atiyah Conjecture implies for a torsionfree groupGthat the rational group ring has no non-trivial zero-divisors.
Notice that the Farrell-Jones Conjecture implies for a torsionfree groupGthat the group ring over any field of characteristic zero has no non-trivial idempotents.
Conjecture (Atiyah Conjecture for torsionfree groups)
Let G be a torsionfree group. It satisfies theAtiyah ConjectureifD(G) is a skew-field.
Fix a natural numberd ≥5. Then a finitely generated torsionfree groupGsatisfies the Atiyah Conjecture if and only if for any G-coveringM→M of a closed Riemannian manifold of dimension d we haveb(2)n (M)∈Zfor everyn≥0.
The Atiyah Conjecture implies for a torsionfree groupGthat the rational group ring has no non-trivial zero-divisors.
Notice that the Farrell-Jones Conjecture implies for a torsionfree groupGthat the group ring over any field of characteristic zero has no non-trivial idempotents.
Theorem (Linnell, Schick)
1 LetC be the smallest class of groups which contains all free groups, is closed under extensions with elementary amenable groups as quotients and directed unions.
Then every torsionfree group G which belongs toCsatisfies the Atiyah Conjecture, actually even overC.
2 If G is residually torsionfree elementary amenable, then it satisfies the Atiyah Conjecture.
This theorem and results byWaldhausenshow for the
fundamental groupπof a 3-manifold (with the exception of some graph manifolds) that it satisfies the Atiyah Conjecture and that Wh(π)vanishes.
Theorem (Linnell, Schick)
1 LetC be the smallest class of groups which contains all free groups, is closed under extensions with elementary amenable groups as quotients and directed unions.
Then every torsionfree group G which belongs toCsatisfies the Atiyah Conjecture, actually even overC.
2 If G is residually torsionfree elementary amenable, then it satisfies the Atiyah Conjecture.
This theorem and results byWaldhausenshow for the
fundamental groupπof a 3-manifold (with the exception of some graph manifolds) that it satisfies the Atiyah Conjecture and that Wh(π)vanishes.
Identifying K
1w( Z G) and K
1(D(G))
Theorem (Linnell-Lück)
If G belongs toC, then the natural map
K1w(ZG)−∼=→K1(D(G)) is an isomorphism.
Its proof is based on identifyingD(G)as an appropriate Cohn localization ofZGand the investigating localization sequences in algebraicK-theory.
There is aDieudonné determinantwhich induces an isomorphism detD:K1(D(G))−→ D(G)∼= ×/[D(G)×,D(G)×].
Identifying K
1w( Z G) and K
1(D(G))
Theorem (Linnell-Lück)
If G belongs toC, then the natural map
K1w(ZG)−∼=→K1(D(G)) is an isomorphism.
Its proof is based on identifyingD(G)as an appropriate Cohn localization ofZGand the investigating localization sequences in algebraicK-theory.
There is aDieudonné determinantwhich induces an isomorphism detD:K1(D(G))−→ D(G)∼= ×/[D(G)×,D(G)×].
Identifying K
1w( Z G) and K
1(D(G))
Theorem (Linnell-Lück)
If G belongs toC, then the natural map
K1w(ZG)−∼=→K1(D(G)) is an isomorphism.
Its proof is based on identifyingD(G)as an appropriate Cohn localization ofZGand the investigating localization sequences in algebraicK-theory.
There is aDieudonné determinantwhich induces an isomorphism detD:K1(D(G))−→ D(G)∼= ×/[D(G)×,D(G)×].
Identifying K
1w( Z G) and K
1(D(G))
Theorem (Linnell-Lück)
If G belongs toC, then the natural map
K1w(ZG)−∼=→K1(D(G)) is an isomorphism.
Its proof is based on identifyingD(G)as an appropriate Cohn localization ofZGand the investigating localization sequences in algebraicK-theory.
There is aDieudonné determinantwhich induces an isomorphism detD:K1(D(G))−→ D(G)∼= ×/[D(G)×,D(G)×].
In particular we get forG=Z
K1w(Z[Z])∼=Q(z±1)×
It turns out that in the caseG=Zthe universal torsion is the same as theAlexander polynomialof an infinite cyclic covering, as it occurs for instance in knot theory.
Twisting L
2-invariants
Consider aCW-complexX withπ=π1(M). Fix an element φ∈H1(X;Z) =hom(π;Z).
Fort∈(0,∞), letφ∗Ct be the 1-dimensionalπ-representation given by
w·λ:=tφ(w)·λ forw ∈π, λ∈C.
One cantwisttheL2-chain complex ofX with this representation, or, equivalently, apply the following ring homomorphism to the cellularZG-chain complex before passing to the Hilbert space completion
CG→CG, X
g∈G
λg·g 7→ X
g∈G
λ·tφ(g)·g.
Notice that for irrationalt the relevant chain complexes do not have coefficients inQGanymore and theDeterminant Conjecture does not apply. Moreover, the Fuglede-Kadison determinant is in
Twisting L
2-invariants
Consider aCW-complexX withπ=π1(M). Fix an element φ∈H1(X;Z) =hom(π;Z).
Fort∈(0,∞), letφ∗Ct be the 1-dimensionalπ-representation given by
w·λ:=tφ(w)·λ forw ∈π, λ∈C.
One cantwisttheL2-chain complex ofX with this representation, or, equivalently, apply the following ring homomorphism to the cellularZG-chain complex before passing to the Hilbert space completion
CG→CG, X
g∈G
λg·g 7→ X
g∈G
λ·tφ(g)·g.
Notice that for irrationalt the relevant chain complexes do not have coefficients inQGanymore and theDeterminant Conjecture does not apply. Moreover, the Fuglede-Kadison determinant is in
Twisting L
2-invariants
Consider aCW-complexX withπ=π1(M). Fix an element φ∈H1(X;Z) =hom(π;Z).
Fort∈(0,∞), letφ∗Ct be the 1-dimensionalπ-representation given by
w·λ:=tφ(w)·λ forw ∈π, λ∈C.
One cantwisttheL2-chain complex ofX with this representation, or, equivalently, apply the following ring homomorphism to the cellularZG-chain complex before passing to the Hilbert space completion
CG→CG, X
g∈G
λg·g 7→ X
g∈G
λ·tφ(g)·g.
Notice that for irrationalt the relevant chain complexes do not have coefficients inQGanymore and theDeterminant Conjecture does not apply. Moreover, the Fuglede-Kadison determinant is in
Twisting L
2-invariants
Consider aCW-complexX withπ=π1(M). Fix an element φ∈H1(X;Z) =hom(π;Z).
Fort∈(0,∞), letφ∗Ct be the 1-dimensionalπ-representation given by
w·λ:=tφ(w)·λ forw ∈π, λ∈C.
One cantwisttheL2-chain complex ofX with this representation, or, equivalently, apply the following ring homomorphism to the cellularZG-chain complex before passing to the Hilbert space completion
CG→CG, X
g∈G
λg·g 7→ X
g∈G
λ·tφ(g)·g.
Notice that for irrationalt the relevant chain complexes do not have coefficients inQGanymore and theDeterminant Conjecture does not apply. Moreover, the Fuglede-Kadison determinant is in
Twisting L
2-invariants
Consider aCW-complexX withπ=π1(M). Fix an element φ∈H1(X;Z) =hom(π;Z).
Fort∈(0,∞), letφ∗Ct be the 1-dimensionalπ-representation given by
w·λ:=tφ(w)·λ forw ∈π, λ∈C.
One cantwisttheL2-chain complex ofX with this representation, or, equivalently, apply the following ring homomorphism to the cellularZG-chain complex before passing to the Hilbert space completion
CG→CG, X
g∈G
λg·g 7→ X
g∈G
λ·tφ(g)·g.
Notice that for irrationalt the relevant chain complexes do not have coefficients inQGanymore and theDeterminant Conjecture does not apply. Moreover, the Fuglede-Kadison determinant is in
Defineφ-twistedL2-torsion function
ρ(Xe;φ): (0,∞)→R by sendingt to theCt-twistedL2-torsion.
Its value att =1 is just theL2-torsion.
On the analytic side this corresponds for closed Riemannian manifoldMto twisting with the flat line bundleMe ×πCt →M. It is obvious that some work is necessary to show that this is a well-defined invariant since theπ-action onCt isnotisometric.
Defineφ-twistedL2-torsion function
ρ(Xe;φ): (0,∞)→R by sendingt to theCt-twistedL2-torsion.
Its value att =1 is just theL2-torsion.
On the analytic side this corresponds for closed Riemannian manifoldMto twisting with the flat line bundleMe ×πCt →M. It is obvious that some work is necessary to show that this is a well-defined invariant since theπ-action onCt isnotisometric.
Defineφ-twistedL2-torsion function
ρ(Xe;φ): (0,∞)→R by sendingt to theCt-twistedL2-torsion.
Its value att =1 is just theL2-torsion.
On the analytic side this corresponds for closed Riemannian manifoldMto twisting with the flat line bundleMe ×πCt →M. It is obvious that some work is necessary to show that this is a well-defined invariant since theπ-action onCt isnotisometric.
Theorem (Lück)
Suppose thatX is Le 2-acyclic.
1 The L2torsion functionρ(2):=ρ(2)(Xe;φ) : (0,∞)→Ris well-defined.
2 The limitslim supt→∞ ρln(t(2)(t)) andlim inft→0ρln(t(2)(t)) exist and we can define thedegree ofφ
deg(X;φ)∈R to be their difference.
3 There is aφ-twisted Fuglede-Kadison determinant det(2)tw,φ:K1w(ZG)→map((0,∞),R) which sendsρ(2)u (Xe)toρ(2)(Xe;φ).
Definition (Thurston norm)
LetMbe a 3-manifold andφ∈H1(M;Z)be a class. Define its Thurston norm
xM(φ)=min{χ−(F)|F embedded surface inM dual toφ}
where
χ−(F) = X
C∈π0(M)
max{−χ(C),0}.
Thurstonshowed that this definition extends to the real vector spaceH1(M;R)and defines aseminormon it.
IfF →M −→p S1is a fiber bundle with connected closed surface F 6∼=S2andφ=π1(p), then
xM(φ) =−χ(F).
Definition (Thurston norm)
LetMbe a 3-manifold andφ∈H1(M;Z)be a class. Define its Thurston norm
xM(φ)=min{χ−(F)|F embedded surface inM dual toφ}
where
χ−(F) = X
C∈π0(M)
max{−χ(C),0}.
Thurstonshowed that this definition extends to the real vector spaceH1(M;R)and defines aseminormon it.
IfF →M −→p S1is a fiber bundle with connected closed surface F 6∼=S2andφ=π1(p), then
xM(φ) =−χ(F).
Theorem (Friedl-Lück,Liu)
Let M be a3-manifold. Then for everyφ∈H1(M;Z)we get the equality deg(M;φ) =xM(φ).
Polytopes
Consider a finitely generated abelian free abelian groupA. Let AR:=R⊗ZAbe the real vector space containingAas a spanning lattice;
ApolytopeP ⊆ARis a convex bounded subset which is the convex hull of a finite subsetS;
It is calledintegral, ifSis contained inA;
TheMinkowski sumof two polytopesP andQis defined by P+Q={p+q|p∈P,q ∈Q};
It iscancellative, i.e., it satisfiesP0+Q=P1+Q =⇒ P0=P1;
Polytopes
Consider a finitely generated abelian free abelian groupA. Let AR:=R⊗ZAbe the real vector space containingAas a spanning lattice;
ApolytopeP ⊆ARis a convex bounded subset which is the convex hull of a finite subsetS;
It is calledintegral, ifSis contained inA;
TheMinkowski sumof two polytopesP andQis defined by P+Q={p+q|p∈P,q ∈Q};
It iscancellative, i.e., it satisfiesP0+Q=P1+Q =⇒ P0=P1;
Polytopes
Consider a finitely generated abelian free abelian groupA. Let AR:=R⊗ZAbe the real vector space containingAas a spanning lattice;
ApolytopeP ⊆ARis a convex bounded subset which is the convex hull of a finite subsetS;
It is calledintegral, ifSis contained inA;
TheMinkowski sumof two polytopesP andQis defined by P+Q={p+q|p∈P,q ∈Q};
It iscancellative, i.e., it satisfiesP0+Q=P1+Q =⇒ P0=P1;
Polytopes
Consider a finitely generated abelian free abelian groupA. Let AR:=R⊗ZAbe the real vector space containingAas a spanning lattice;
ApolytopeP ⊆ARis a convex bounded subset which is the convex hull of a finite subsetS;
It is calledintegral, ifSis contained inA;
TheMinkowski sumof two polytopesP andQis defined by P+Q={p+q|p∈P,q ∈Q};
It iscancellative, i.e., it satisfiesP0+Q=P1+Q =⇒ P0=P1;
Polytopes
Consider a finitely generated abelian free abelian groupA. Let AR:=R⊗ZAbe the real vector space containingAas a spanning lattice;
ApolytopeP ⊆ARis a convex bounded subset which is the convex hull of a finite subsetS;
It is calledintegral, ifSis contained inA;
TheMinkowski sumof two polytopesP andQis defined by P+Q={p+q|p∈P,q ∈Q};
It iscancellative, i.e., it satisfiesP0+Q=P1+Q =⇒ P0=P1;
Polytopes
Consider a finitely generated abelian free abelian groupA. Let AR:=R⊗ZAbe the real vector space containingAas a spanning lattice;
ApolytopeP ⊆ARis a convex bounded subset which is the convex hull of a finite subsetS;
It is calledintegral, ifSis contained inA;
TheMinkowski sumof two polytopesP andQis defined by P+Q={p+q|p∈P,q ∈Q};
It iscancellative, i.e., it satisfiesP0+Q=P1+Q =⇒ P0=P1;
TheNewton polytope
N(p)⊆Rn of a polynomial
p(t1,t2, . . . ,tn) = X
i1,...,in
ai1,i2,...,in·t1i1t2i2· · ·tnin
innvariablest1,t2, . . . ,tnis defined to be the convex hull of the elements{(i1,i2, . . .in)∈Zn |ai1,i2,...,in 6=0};
One has
N(p·q) =N(p) +N(q).
TheNewton polytope
N(p)⊆Rn of a polynomial
p(t1,t2, . . . ,tn) = X
i1,...,in
ai1,i2,...,in·t1i1t2i2· · ·tnin
innvariablest1,t2, . . . ,tnis defined to be the convex hull of the elements{(i1,i2, . . .in)∈Zn |ai1,i2,...,in 6=0};
One has
N(p·q) =N(p) +N(q).
Definition (Polytope group)
LetPZ(A)be the Grothendieck group of the abelian monoid of integral polytopes inAR.
Denote byPZ,Wh(A)the quotient ofPZ(A)by the canonical homomorphismA→ PZ(A)sendingato the class of the polytope {a}.
InPZ,Wh(A)we consider polytopes up to translation with an element inA.
Given a homomorphism of finitely generated abelian groups f:A→A0, we obtain a homomorphisms of abelian groups
PZ(f) :PZ(A)→ PZ(A0), [P]7→[idR⊗Zf(P)];
and analogously forP (A).
Definition (Polytope group)
LetPZ(A)be the Grothendieck group of the abelian monoid of integral polytopes inAR.
Denote byPZ,Wh(A)the quotient ofPZ(A)by the canonical homomorphismA→ PZ(A)sendingato the class of the polytope {a}.
InPZ,Wh(A)we consider polytopes up to translation with an element inA.
Given a homomorphism of finitely generated abelian groups f:A→A0, we obtain a homomorphisms of abelian groups
PZ(f) :PZ(A)→ PZ(A0), [P]7→[idR⊗Zf(P)];
and analogously forP (A).
Definition (Polytope group)
LetPZ(A)be the Grothendieck group of the abelian monoid of integral polytopes inAR.
Denote byPZ,Wh(A)the quotient ofPZ(A)by the canonical homomorphismA→ PZ(A)sendingato the class of the polytope {a}.
InPZ,Wh(A)we consider polytopes up to translation with an element inA.
Given a homomorphism of finitely generated abelian groups f:A→A0, we obtain a homomorphisms of abelian groups
PZ(f) :PZ(A)→ PZ(A0), [P]7→[idR⊗Zf(P)];
and analogously forP (A).
Definition (Polytope group)
LetPZ(A)be the Grothendieck group of the abelian monoid of integral polytopes inAR.
Denote byPZ,Wh(A)the quotient ofPZ(A)by the canonical homomorphismA→ PZ(A)sendingato the class of the polytope {a}.
InPZ,Wh(A)we consider polytopes up to translation with an element inA.
Given a homomorphism of finitely generated abelian groups f:A→A0, we obtain a homomorphisms of abelian groups
PZ(f) :PZ(A)→ PZ(A0), [P]7→[idR⊗Zf(P)];
and analogously forP (A).
Example (A=Z)
An integral polytope inZRis just an interval[m,n]form,n∈Z satisfyingm≤n.
The Minkowski sum becomes
[m1,n1] + [m2,n2] = [m1+m2,n1+n2].
One obtains isomorphisms of abelian groups
PZ(Z) −→∼= Z2 [[m,n]]7→(n−m,m).
PZ,Wh(Z) −→∼= Z, [[m,n]]7→n−m.
We obtain an injection PZ(A)→ Y
φ∈homZ(A,Z)
PZ(Z), x 7→ φ(x)
φ.
It implies thatPZ(A)is torsionfree and not divisible.
ConjecturallyPZ(Zn)is always a free abelian group.
We obtain a well-defined homomorphism of abelian groups Q[Zn](0)×
→ PZ(Zn), p
q 7→[N(p)]−[N(q)].
We want to generalize it to the so called polytope homomorphism.
We obtain an injection PZ(A)→ Y
φ∈homZ(A,Z)
PZ(Z), x 7→ φ(x)
φ.
It implies thatPZ(A)is torsionfree and not divisible.
ConjecturallyPZ(Zn)is always a free abelian group.
We obtain a well-defined homomorphism of abelian groups Q[Zn](0)×
→ PZ(Zn), p
q 7→[N(p)]−[N(q)].
We want to generalize it to the so called polytope homomorphism.
We obtain an injection PZ(A)→ Y
φ∈homZ(A,Z)
PZ(Z), x 7→ φ(x)
φ.
It implies thatPZ(A)is torsionfree and not divisible.
ConjecturallyPZ(Zn)is always a free abelian group.
We obtain a well-defined homomorphism of abelian groups Q[Zn](0)×
→ PZ(Zn), p
q 7→[N(p)]−[N(q)].
We want to generalize it to the so called polytope homomorphism.
We obtain an injection PZ(A)→ Y
φ∈homZ(A,Z)
PZ(Z), x 7→ φ(x)
φ.
It implies thatPZ(A)is torsionfree and not divisible.
ConjecturallyPZ(Zn)is always a free abelian group.
We obtain a well-defined homomorphism of abelian groups Q[Zn](0)×
→ PZ(Zn), p
q 7→[N(p)]−[N(q)].
We want to generalize it to the so called polytope homomorphism.
Polytope homomorphism
Consider the projection
pr:G→H1(G)f :=H1(G)/tors(H1(G)).
LetK be its kernel.
After a choice of a set-theoretic section of pr we get isomorphisms ZK ∗H1(G)f −∼=→ ZG;
S−1 D(K)∗H1(G)f ∼=
−→ D(G),
where here and in the sequelS−1denotes Ore localization with respect to the multiplicative closed set of non-trivial elements.
Polytope homomorphism
Consider the projection
pr:G→H1(G)f :=H1(G)/tors(H1(G)).
LetK be its kernel.
After a choice of a set-theoretic section of pr we get isomorphisms ZK ∗H1(G)f −∼=→ ZG;
S−1 D(K)∗H1(G)f ∼=
−→ D(G),
where here and in the sequelS−1denotes Ore localization with respect to the multiplicative closed set of non-trivial elements.
Polytope homomorphism
Consider the projection
pr:G→H1(G)f :=H1(G)/tors(H1(G)).
LetK be its kernel.
After a choice of a set-theoretic section of pr we get isomorphisms ZK ∗H1(G)f −∼=→ ZG;
S−1 D(K)∗H1(G)f ∼=
−→ D(G),
where here and in the sequelS−1denotes Ore localization with respect to the multiplicative closed set of non-trivial elements.
Givenx =P
h∈H1(G)f uh·h∈ D(K)∗H1(G)f, define itssupport supp(x):={h∈H1(G)f |h∈H1(G)f),uh 6=0}.
The convex hull of supp(x)defines apolytope P(x)⊆R⊗ZH1(G)f =H1(M;R).
We haveP(x ·y) =P(x) +P(y)forx,y ∈(D(K)∗H1(G)f. Hence we can define a homomorphism of abelian groups
P0:
S−1 D(K)∗H1(G)f×
→ PZ(H1(G)f),
by sendingx ·y−1to[P(x)]−[P(y)].
Givenx =P
h∈H1(G)f uh·h∈ D(K)∗H1(G)f, define itssupport supp(x):={h∈H1(G)f |h∈H1(G)f),uh 6=0}.
The convex hull of supp(x)defines apolytope P(x)⊆R⊗ZH1(G)f =H1(M;R).
We haveP(x ·y) =P(x) +P(y)forx,y ∈(D(K)∗H1(G)f. Hence we can define a homomorphism of abelian groups
P0:
S−1 D(K)∗H1(G)f×
→ PZ(H1(G)f),
by sendingx ·y−1to[P(x)]−[P(y)].
Givenx =P
h∈H1(G)f uh·h∈ D(K)∗H1(G)f, define itssupport supp(x):={h∈H1(G)f |h∈H1(G)f),uh 6=0}.
The convex hull of supp(x)defines apolytope P(x)⊆R⊗ZH1(G)f =H1(M;R).
We haveP(x ·y) =P(x) +P(y)forx,y ∈(D(K)∗H1(G)f. Hence we can define a homomorphism of abelian groups
P0:
S−1 D(K)∗H1(G)f×
→ PZ(H1(G)f),
by sendingx ·y−1to[P(x)]−[P(y)].
Givenx =P
h∈H1(G)f uh·h∈ D(K)∗H1(G)f, define itssupport supp(x):={h∈H1(G)f |h∈H1(G)f),uh 6=0}.
The convex hull of supp(x)defines apolytope P(x)⊆R⊗ZH1(G)f =H1(M;R).
We haveP(x ·y) =P(x) +P(y)forx,y ∈(D(K)∗H1(G)f. Hence we can define a homomorphism of abelian groups
P0:
S−1 D(K)∗H1(G)f×
→ PZ(H1(G)f),
by sendingx ·y−1to[P(x)]−[P(y)].
The composite
K1w(ZG)−→∼= K1(D(G))−→ D(G)∼= × ∼−=→
S−1 D(K)∗H1(G)f× P0
−→ PZ(H1(G)f) factories to thepolytope homomorphism
P: Whw(G)→ PZ,Wh(H1(G)f).
Definition (Dual Thurston polytope)
LetMbe a 3-manifold. Define thedual Thurston polytopeto be subset ofH1(M;R)
T(M) :={v ∈H1(M;R)|φ(v)≤xM(φ)for allφ∈H1(M;R)}.
Thurstonhas shown that the dual Thurston polytope is always an integral polytope.
The Thurston seminormxM obviously determines the dual Thurston polytope.
The converse is also true, namely, we have xM(φ) := 1
2·sup{φ(x0)−φ(x1)|x0,x1∈T(M)}.
Definition (Dual Thurston polytope)
LetMbe a 3-manifold. Define thedual Thurston polytopeto be subset ofH1(M;R)
T(M) :={v ∈H1(M;R)|φ(v)≤xM(φ)for allφ∈H1(M;R)}.
Thurstonhas shown that the dual Thurston polytope is always an integral polytope.
The Thurston seminormxM obviously determines the dual Thurston polytope.
The converse is also true, namely, we have xM(φ) := 1
2·sup{φ(x0)−φ(x1)|x0,x1∈T(M)}.
Definition (Dual Thurston polytope)
LetMbe a 3-manifold. Define thedual Thurston polytopeto be subset ofH1(M;R)
T(M) :={v ∈H1(M;R)|φ(v)≤xM(φ)for allφ∈H1(M;R)}.
Thurstonhas shown that the dual Thurston polytope is always an integral polytope.
The Thurston seminormxM obviously determines the dual Thurston polytope.
The converse is also true, namely, we have xM(φ) := 1
2·sup{φ(x0)−φ(x1)|x0,x1∈T(M)}.
Definition (Dual Thurston polytope)
LetMbe a 3-manifold. Define thedual Thurston polytopeto be subset ofH1(M;R)
T(M) :={v ∈H1(M;R)|φ(v)≤xM(φ)for allφ∈H1(M;R)}.
Thurstonhas shown that the dual Thurston polytope is always an integral polytope.
The Thurston seminormxM obviously determines the dual Thurston polytope.
The converse is also true, namely, we have xM(φ) := 1
2·sup{φ(x0)−φ(x1)|x0,x1∈T(M)}.
Theorem (Friedl-Lück)
Let M be a3-manifold. Then the image of the universal L2-torsion ρ(2)u (M)e under the polytope homomorphism
P: Whw(π1(M))→ PZ,Wh(H1(π1(M))f) is represented by the dual of the Thurston polytope.
Higher order Alexander polynomials
Higher order Alexander polynomialswere introduced for a coveringG→M→M of a 3-manifold byHarveyandCochran, provided thatGoccurs in the rational derived series ofπ1(M).
At least thedegreeof these polynomials is a well-defined invariant ofM andG.
We can extend this notion of degree also to the universal covering ofM and can prove the conjecture that the degree coincides with the Thurston norm.