Fermion bags, topology and index theorems
Shailesh Chandrasekharan (Duke University)
Supported by: US Department of Energy, Nuclear Physics Division
Lattice 2016, Southampton UK
work done in collaboration with V. Ayyar
Fermion bags, topology and index theorems Shailesh Chandrasekharan - - PowerPoint PPT Presentation
Fermion bags, topology and index theorems Shailesh Chandrasekharan (Duke University) work done in collaboration with V. Ayyar Lattice 2016, Southampton UK Supported by: US Department of Energy, Nuclear Physics Division Summary Summary The
Shailesh Chandrasekharan (Duke University)
Supported by: US Department of Energy, Nuclear Physics Division
Lattice 2016, Southampton UK
work done in collaboration with V. Ayyar
The concepts of topology and index theorems that arise in the context of QCD, have analogies in simple fermion lattice field theories with staggered fermions, when formulated in the fermion bag approach
The concepts of topology and index theorems that arise in the context of QCD, have analogies in simple fermion lattice field theories with staggered fermions, when formulated in the fermion bag approach
This connection gives a more complete perspective
including a mechanism where fermions acquire a mass through four-fermion condensates instead of fermion bilinear condensates.
The concepts of topology and index theorems that arise in the context of QCD, have analogies in simple fermion lattice field theories with staggered fermions, when formulated in the fermion bag approach
This connection gives a more complete perspective
including a mechanism where fermions acquire a mass through four-fermion condensates instead of fermion bilinear condensates. Can non-Abelian gauge theories also demonstrate this alternate mechanism of fermion mass generation?
Partition function of a non-Abelian gauge theory (formal, continuum, finite volume)
Partition function of a non-Abelian gauge theory (formal, continuum, finite volume)
background gauge field integration weight of the background gauge field anti-Hermitian operator depends on the gauge field
D(A) = (γµ∂µ − iAµ)
Z = Z [dA] e−SG (A) Z [dψ dψ] e−ψ D(A) ψ
Partition function of staggered fermion lattice field theories in the fermion bag approach.
Partition function of staggered fermion lattice field theories in the fermion bag approach.
Z = X
B
e−S(B) Z [dψ dψ] e−ψ W (B) ψ
sum over fermion bag configurations weight of a fermion bag configuration anti-Hermitian “fermion bag” matrix depends fermion bag configuration
For every gauge field configuration we can define a topological charge:
Q
A
For every gauge field configuration we can define a topological charge:
Q
A
γ5 D(A) = −D(A) γ5
Dirac operators satisfy: non-zero modes come in pairs
D(A) |z±i = 0, γ5 |z±i = ± |z±i
zero modes are eigenstates of number of modes
n± =
|z±i
D(A) |λi = iλ |λi
D(A) γ5 |λi = iλ γ5 |λi
γ5
The index of the Dirac operator : D(A)
(n+ − n−)
For every gauge field configuration we can define a topological charge:
Q
A
γ5 D(A) = −D(A) γ5
Dirac operators satisfy: non-zero modes come in pairs
D(A) |z±i = 0, γ5 |z±i = ± |z±i
zero modes are eigenstates of number of modes
n± =
|z±i
D(A) |λi = iλ |λi
D(A) γ5 |λi = iλ γ5 |λi
γ5
The index of the Dirac operator : D(A)
(n+ − n−)
Index Theorem: Q = (n+ − n−)
For every gauge field configuration we can define a topological charge:
Q
A
γ5 D(A) = −D(A) γ5
Dirac operators satisfy: non-zero modes come in pairs
D(A) |z±i = 0, γ5 |z±i = ± |z±i
zero modes are eigenstates of number of modes
n± =
|z±i
D(A) |λi = iλ |λi
D(A) γ5 |λi = iλ γ5 |λi
γ5
The index of the Dirac operator : D(A)
(n+ − n−)
Index Theorem: Q = (n+ − n−) D(A) has at least |Q| zero modes
For every fermion bag configuration we can define a topological charge:
Q
B
For every fermion bag configuration we can define a topological charge:
Q
B Fermion bag matrix satisfies: non-zero modes come in pairs zero modes are eigenstates of number of modes
n± =
|z±i
The index of the Dirac operator : (n+ − n−)
W (B) Ξ |λi = iλ Ξ |λi W (B) |λi = iλ |λi
W (B) |z±i = 0, Ξ |z±i = ± |z±i Ξ
Ξ W (B) = − W (B) Ξ
W (B)
For every fermion bag configuration we can define a topological charge:
Q
B Fermion bag matrix satisfies: non-zero modes come in pairs zero modes are eigenstates of number of modes
n± =
|z±i
The index of the Dirac operator : (n+ − n−)
W (B) Ξ |λi = iλ Ξ |λi W (B) |λi = iλ |λi
W (B) |z±i = 0, Ξ |z±i = ± |z±i Ξ
Ξ W (B) = − W (B) Ξ
W (B)
Index Theorem:
Q = (n+ − n−)
For every fermion bag configuration we can define a topological charge:
Q
B Fermion bag matrix satisfies: non-zero modes come in pairs zero modes are eigenstates of number of modes
n± =
|z±i
The index of the Dirac operator : (n+ − n−)
W (B) Ξ |λi = iλ Ξ |λi W (B) |λi = iλ |λi
W (B) |z±i = 0, Ξ |z±i = ± |z±i Ξ
Ξ W (B) = − W (B) Ξ
W (B)
W(B) has at least |Q| zero modes Index Theorem:
Q = (n+ − n−)
Consider free massive staggered fermions:
S = 1 2 X
x,α
ηx,α ⇣ ψxψx+α − ψx+αψx ⌘ + m X
x
ψxψx
X
x,y
ψxDxyψy
Consider free massive staggered fermions:
S = 1 2 X
x,α
ηx,α ⇣ ψxψx+α − ψx+αψx ⌘ + m X
x
ψxψx
X
x,y
ψxDxyψy
D is an anti-Hermitian matrix of the form
D = ✓ C −C T ◆
even
even
Consider free massive staggered fermions:
S = 1 2 X
x,α
ηx,α ⇣ ψxψx+α − ψx+αψx ⌘ + m X
x
ψxψx
X
x,y
ψxDxyψy
D is an anti-Hermitian matrix of the form
D = ✓ C −C T ◆
even
even
even
even
Ξ = ✓ 1 −1 ◆
Consider free massive staggered fermions:
S = 1 2 X
x,α
ηx,α ⇣ ψxψx+α − ψx+αψx ⌘ + m X
x
ψxψx
X
x,y
ψxDxyψy
D is an anti-Hermitian matrix of the form
D = ✓ C −C T ◆
even
even
even
even
Ξ = ✓ 1 −1 ◆
DΞ = −ΞD
Partition function:
Z = Z [dψ dψ] e−S
Partition function:
Z = Z [dψ dψ] e−S Z = Z [dψ dψ] e−ψDψ e−m P
x ψxψx
Femion Bag Approach:
= Z [dψ dψ] e−ψDψ Y
x
X
B
mk Z [dψ dψ] e−ψ W (B) ψ
Partition function:
Z = Z [dψ dψ] e−S Z = Z [dψ dψ] e−ψDψ e−m P
x ψxψx
Femion Bag Approach:
= Z [dψ dψ] e−ψDψ Y
x
X
B
mk Z [dψ dψ] e−ψ W (B) ψ
configuration B space-time defects free fermion bags k = total number of defects B = B1 + B2 + …
Partition function:
Z = Z [dψ dψ] e−S Z = Z [dψ dψ] e−ψDψ e−m P
x ψxψx
Femion Bag Approach:
= Z [dψ dψ] e−ψDψ Y
x
X
B
mk Z [dψ dψ] e−ψ W (B) ψ
Z = X
B
exp(−S(B)) Z [dψ dψ] e−ψ W (B) ψ
Defining
S(B) = −k log(m)
configuration B space-time defects free fermion bags k = total number of defects B = B1 + B2 + …
Partition function:
Z = Z [dψ dψ] e−S Z = Z [dψ dψ] e−ψDψ e−m P
x ψxψx
Femion Bag Approach:
= Z [dψ dψ] e−ψDψ Y
x
X
B
mk Z [dψ dψ] e−ψ W (B) ψ
Z = X
B
exp(−S(B)) Z [dψ dψ] e−ψ W (B) ψ
Defining
S(B) = −k log(m)
W (B) = ✓ C(B) −C(B)T ◆
even
even
anti-Hermitian configuration B space-time defects free fermion bags k = total number of defects B = B1 + B2 + …
Topology and index theorem for a fermion bag
Topology and index theorem for a fermion bag
Topological charge of a fermion bag
Q = neven − nodd
Topology and index theorem for a fermion bag
Topological charge of a fermion bag
Q = neven − nodd
Fermion bag Dirac operator
W (B) = ✓ C(B) −C(B)T ◆
even
even
Topology and index theorem for a fermion bag
Topological charge of a fermion bag
Q = neven − nodd
Fermion bag Dirac operator
W (B) = ✓ C(B) −C(B)T ◆
even
even
even
even
ΞB = ✓ 1 −1 ◆
define
Topology and index theorem for a fermion bag
Topological charge of a fermion bag
Q = neven − nodd
Fermion bag Dirac operator
W (B) = ✓ C(B) −C(B)T ◆
even
even
even
even
ΞB = ✓ 1 −1 ◆
define
Let n± be the number of zero modes of W (B)
that are also eigenvalues of ΞB with eigenvalues ±1.
Topology and index theorem for a fermion bag
Topological charge of a fermion bag
Q = neven − nodd
Index theorem:
Q = n+ − n−
Fermion bag Dirac operator
W (B) = ✓ C(B) −C(B)T ◆
even
even
even
even
ΞB = ✓ 1 −1 ◆
define
Let n± be the number of zero modes of W (B)
that are also eigenvalues of ΞB with eigenvalues ±1.
Topology and index theorem for a fermion bag
Topological charge of a fermion bag
Q = neven − nodd
Index theorem:
Q = n+ − n−
Fermion bag Dirac operator
W (B) = ✓ C(B) −C(B)T ◆
even
even
has at least zero modes.
|Q|
W (B)
even
even
ΞB = ✓ 1 −1 ◆
define
Let n± be the number of zero modes of W (B)
that are also eigenvalues of ΞB with eigenvalues ±1.
Partition function gets contribution only from Q = 0 sector.
Z = Z [dAQ=0] e−SG (AQ=0) Det(D(AQ=0))
Partition function gets contribution only from Q = 0 sector.
Z = Z [dAQ=0] e−SG (AQ=0) Det(D(AQ=0))
Chiral Condensate gets contribution from the sector with topological charge one.
hψψi = 1 Z Z [dA dψ dψ] e−S n 1 V Z d4x ψ(x)ψ(x)
1 ZV Z [dAQ=1] e−SG (AQ=1) ⇣ Y
λ
λ(AQ=1) ⌘
Partition function gets contribution only from Q = 0 sector.
Z = Z [dAQ=0] e−SG (AQ=0) Det(D(AQ=0))
Chiral Condensate gets contribution from the sector with topological charge one.
hψψi = 1 Z Z [dA dψ dψ] e−S n 1 V Z d4x ψ(x)ψ(x)
1 ZV Z [dAQ=1] e−SG (AQ=1) ⇣ Y
λ
λ(AQ=1) ⌘
Anomalous chiral symmetry breaking!
S = 1 2 X
x,α
ηx,α ⇣ ψxψx+α − ψx+αψx ⌘ + m X
x
ψxψx Partition function gets contribution from Q = 0 sectors.
Z = X
BQ=0
ek log(m) Det(W (BQ=0))
Chiral condensate get contribution from Q=1 sectors!
hψψi = 1 ZV X
BQ=1
ek log(m) ⇣ Y
λ
λ(BQ=1) ⌘
S = 1 2 X
x,α
ηx,α ⇣ ψxψx+α − ψx+αψx ⌘ + m X
x
ψxψx Partition function gets contribution from Q = 0 sectors.
Z = X
BQ=0
ek log(m) Det(W (BQ=0))
Chiral condensate get contribution from Q=1 sectors!
hψψi = 1 ZV X
BQ=1
ek log(m) ⇣ Y
λ
λ(BQ=1) ⌘
fermion bag with Q = 1
S = 1 2 X
x,α
ηx,α ⇣ ψxψx+α − ψx+αψx ⌘ + m X
x
ψxψx Partition function gets contribution from Q = 0 sectors.
Z = X
BQ=0
ek log(m) Det(W (BQ=0))
Chiral condensate get contribution from Q=1 sectors!
hψψi = 1 ZV X
BQ=1
ek log(m) ⇣ Y
λ
λ(BQ=1) ⌘
fermion bag with Q = 1 Anomalous chiral symmetry breaking ∼ Explicit chiral symmetry breaking
Chiral condensate vanishes due to additional chiral symmetries! hψψi = 0
Chiral susceptibility is non-zero and gets contribution from the sectors, Q = 0 and Q=1.
χ = 1 ZV Z [dA dψ dψ] e−S n Z d4xψ(x)ψ(x)
d4yψ(y)ψ(y)
1 ZV Z [dAQ=0] e−SG (AQ=0) Det(D(AQ=0)) n X
λ
1 λ2
Chiral condensate vanishes due to additional chiral symmetries! hψψi = 0
Chiral susceptibility is non-zero and gets contribution from the sectors, Q = 0 and Q=1.
χ = 1 ZV Z [dA dψ dψ] e−S n Z d4xψ(x)ψ(x)
d4yψ(y)ψ(y)
1 ZV Z [dAQ=0] e−SG (AQ=0) Det(D(AQ=0)) n X
λ
1 λ2
Spontaneous chiral symmetry breaking requires χ ∼ Σ2V Chiral condensate vanishes due to additional chiral symmetries! hψψi = 0
Chiral susceptibility is non-zero and gets contribution from the sectors, Q = 0 and Q=1.
χ = 1 ZV Z [dA dψ dψ] e−S n Z d4xψ(x)ψ(x)
d4yψ(y)ψ(y)
1 ZV Z [dAQ=0] e−SG (AQ=0) Det(D(AQ=0)) n X
λ
1 λ2
Spontaneous chiral symmetry breaking requires χ ∼ Σ2V Chiral condensate vanishes due to additional chiral symmetries! hψψi = 0 This implies that the smallest non-zero eigenvalues in the Q=0 must scale as λ ∼
1 ΣV
Action (Nf = 2 model)
S = 1 2 X
x,α,i=1,2
ηx,α ⇣ ψx,iψx+α,i − ψx+α,iψx,i ⌘ − U X
x
ψx,1ψx,1ψx,2ψx,2
Ayyar, SC, PRD91 (2015) 6, 065035, PRD93 (2016) 8, 081701, arXiv:1606.06312
Z = Z [dψ] e− 1
2 ψT Mψ Y
x
X
B
ek log(U) ⇣ Det(W (B)) ⌘2
Partition function Action (Nf = 2 model)
S = 1 2 X
x,α,i=1,2
ηx,α ⇣ ψx,iψx+α,i − ψx+α,iψx,i ⌘ − U X
x
ψx,1ψx,1ψx,2ψx,2
Ayyar, SC, PRD91 (2015) 6, 065035, PRD93 (2016) 8, 081701, arXiv:1606.06312
Z = Z [dψ] e− 1
2 ψT Mψ Y
x
X
B
ek log(U) ⇣ Det(W (B)) ⌘2
Partition function
ψxψx
Action (Nf = 2 model)
S = 1 2 X
x,α,i=1,2
ηx,α ⇣ ψx,iψx+α,i − ψx+α,iψx,i ⌘ − U X
x
ψx,1ψx,1ψx,2ψx,2
Ayyar, SC, PRD91 (2015) 6, 065035, PRD93 (2016) 8, 081701, arXiv:1606.06312
Z = Z [dψ] e− 1
2 ψT Mψ Y
x
X
B
ek log(U) ⇣ Det(W (B)) ⌘2
Partition function Chiral condensate vanishes
hψψi = 0
ψxψx
Action (Nf = 2 model)
S = 1 2 X
x,α,i=1,2
ηx,α ⇣ ψx,iψx+α,i − ψx+α,iψx,i ⌘ − U X
x
ψx,1ψx,1ψx,2ψx,2
Ayyar, SC, PRD91 (2015) 6, 065035, PRD93 (2016) 8, 081701, arXiv:1606.06312
hψx,aψx,b ψy,aψy,bi = Z [dψdψ] e−S ψx,aψx,b ψy,aψy,b
Fermion bilinear correlation function If B = B1 + B2 + … then, it vanishes unless x and y are in the same bag!
hψx,aψx,b ψy,aψy,bi = Z [dψdψ] e−S ψx,aψx,b ψy,aψy,b
Fermion bilinear correlation function If B = B1 + B2 + … then, it vanishes unless x and y are in the same bag! Non-zero!
hψx,aψx,b ψy,aψy,bi = Z [dψdψ] e−S ψx,aψx,b ψy,aψy,b
Fermion bilinear correlation function If B = B1 + B2 + … then, it vanishes unless x and y are in the same bag! Non-zero! Zero!
hψx,aψx,b ψy,aψy,bi = Z [dψdψ] e−S ψx,aψx,b ψy,aψy,b
Fermion bilinear correlation function If B = B1 + B2 + … then, it vanishes unless x and y are in the same bag! At sufficiently large coupling U all two point correlation functions will exponentially decay! Non-zero! Zero!
As in QCD Chiral Susceptibility can get contributions from Q=0,1 sectors.
As in QCD Chiral Susceptibility can get contributions from Q=0,1 sectors. As in QCD spontaneous chiral symmetry breaking requires
χ ∼ Σ2V
As in QCD Chiral Susceptibility can get contributions from Q=0,1 sectors. As in QCD spontaneous chiral symmetry breaking requires
χ ∼ Σ2V
This implies that the smallest non-zero eigenvalues in the Q=0 must sector scale as λ ∼ 1 ΣV
As in QCD Chiral Susceptibility can get contributions from Q=0,1 sectors. A necessary condition is that fermion bags span the entire volume for SSB! As in QCD spontaneous chiral symmetry breaking requires
χ ∼ Σ2V
This implies that the smallest non-zero eigenvalues in the Q=0 must sector scale as λ ∼ 1 ΣV
As in QCD Chiral Susceptibility can get contributions from Q=0,1 sectors. A necessary condition is that fermion bags span the entire volume for SSB! As in QCD spontaneous chiral symmetry breaking requires
χ ∼ Σ2V
This implies that the smallest non-zero eigenvalues in the Q=0 must sector scale as λ ∼ 1 ΣV But at large U fermion bags are small. Smallest eigenvalues λ 6⇠ 1 ΣV
As in QCD Chiral Susceptibility can get contributions from Q=0,1 sectors. A necessary condition is that fermion bags span the entire volume for SSB! As in QCD spontaneous chiral symmetry breaking requires
χ ∼ Σ2V
This implies that the smallest non-zero eigenvalues in the Q=0 must sector scale as λ ∼ 1 ΣV But at large U fermion bags are small. Smallest eigenvalues λ 6⇠ 1 ΣV At large U, unlike QCD no SSB, but fermions are still massive!
In 3D there is evidence for a single exotic transition
U = ∞ U = 0 Scenario B Uc Symmetric Massive Symmetric Massless
PRD91 (2015) 6, 065035, PRD 93 (2016), 081701
In 4D we find evidence for a narrow intermediate spontaneously broken phase
Symmetric Massive Massless Fermions U = 0 U = ∞ Broken Massive
arXiv:1606.06312
In 3D there is evidence for a single exotic transition
U = ∞ U = 0 Scenario B Uc Symmetric Massive Symmetric Massless
PRD91 (2015) 6, 065035, PRD 93 (2016), 081701
In 4D we find evidence for a narrow intermediate spontaneously broken phase
Symmetric Massive Massless Fermions U = 0 U = ∞ Broken Massive
arXiv:1606.06312
In 3D there is evidence for a single exotic transition
U = ∞ U = 0 Scenario B Uc Symmetric Massive Symmetric Massless
PRD91 (2015) 6, 065035, PRD 93 (2016), 081701
In 2D we find evidence for a single asymptotically free (massive) phase
Symmetric Massive U = 0 U = ∞
work in progress
Fermion bilinear condensates form due to anomalous or explicit chiral symmetry breaking through topological space-time defects. Examples: Nf=1 QCD, Nf =1 massive free staggered fermions
Bilinear Condensate Mechanism: A
Fermion bilinear condensates form due to anomalous or explicit chiral symmetry breaking through topological space-time defects. Examples: Nf=1 QCD, Nf =1 massive free staggered fermions
Bilinear Condensate Mechanism: A
Fermion bilinear condensates form due to spontaneous breaking of symmetries that prevent their formation. Nf = 2 QCD is an example of this phenomena.
Bilinear Condensate Mechanism: B
Fermion bilinear condensates form due to anomalous or explicit chiral symmetry breaking through topological space-time defects. Examples: Nf=1 QCD, Nf =1 massive free staggered fermions
Bilinear Condensate Mechanism: A
Fermion bilinear condensates form due to spontaneous breaking of symmetries that prevent their formation. Nf = 2 QCD is an example of this phenomena.
Bilinear Condensate Mechanism: B Four-fermion Condensate Mechanism: C
Four-fermion condensate can form while preserving symmetries that ensure fermion bilinear condensates vanish. But, fermions can still become massive. Nf = 2 interacting fermion model is one example.
The concepts of topology and index theorems that arise in the context of QCD, have analogies in simple fermion lattice field theories with staggered fermions, when formulated in the fermion bag approach
This connection gives a more complete perspective
including a mechanism where fermions acquire a mass through four-fermion condensates instead of fermion bilinear condensates. Can non-Abelian gauge theories also demonstrate this alternate mechanism of fermion mass generation?