Modern Approaches to Scattering Amplitudes

Transcription

1 Ph.D. Workshop, 10 October 2017

2 Outline Scattering Amplitudes 1 Scattering Amplitudes Introduction The Standard Method 2 On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations 3 Amplitudes as Volumes Amplitudes from Positive Geometry

3 Scattering Of Particles Introduction The Standard Method Cross Sections and Rates dσ dω = (2π)4 m i m f p f p i A i,f 2 A i,f = α in β out α, β = (p µ, h, c)

4 Scattering Of Particles Introduction The Standard Method Cross Sections and Rates dσ dω = (2π)4 m i m f p f p i A i,f 2 A i,f = α in β out α, β = (p µ, h, c)

5 Scattering Of Particles Introduction The Standard Method Cross Sections and Rates dσ dω = (2π)4 m i m f p f p i A i,f 2 A i,f = α in β out α, β = (p µ, h, c)

6 Scattering Of Particles Introduction The Standard Method Cross Sections and Rates dσ dω = (2π)4 m i m f p f p i A i,f 2 A i,f = α in β out α, β = (p µ, h, c)

7 Scattering Of Particles Introduction The Standard Method Cross Sections and Rates dσ dω = (2π)4 m i m f p f p i A i,f 2 A i,f = α in β out α, β = (p µ, h, c)

8 Feynman Diagrams Scattering Amplitudes Introduction The Standard Method Lagrangian L = d 4 xl(φ, µ φ), e.g. L = d 4 x µ φ µ φ + g 3 φ 3 + g 4 φ g 3 g 4

9 Cumbersome computation! Introduction The Standard Method Feynman Expansion A n = Γ Graphs Value(Γ) IR and UV infinities: need to renormalize Factorial growth of the number of diagrams n!

10 Cumbersome computation! Introduction The Standard Method Feynman Expansion A n = Γ Graphs Value(Γ) IR and UV infinities: need to renormalize Factorial growth of the number of diagrams n!

11 Cumbersome computation! Introduction The Standard Method Feynman Expansion A n = Γ Graphs Value(Γ) IR and UV infinities: need to renormalize Factorial growth of the number of diagrams n!

12 Introduction The Standard Method Unnecessarily Cumbersome Computation? A[g 1 g 2 g + 3 g g + n ] = ij (p i + p j ) n1

13 Particles and Fields Scattering Amplitudes On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations Particles: Experimentally accesible quantum states, well defined Mass, Charges and Free Propagation Fields: Tools used to construct Lorentz Invariant dynamics: φ(x) = d 3 p exp(ip µ x µ )u( p )â( p ) Û(Λ)(φ(x)) = S(Λ)(φ(Λ 1 x)), Λ SO(1, 3) L = d 4 x L( µ φ, φ) is invariant. Problem: No massless vector field (e.g. A µ in QED) Û(Λ)(A µ (x)) = Λ µ ν A ν (Λ 1 x) + µ χ Solution: Coupling with a conserved current J µ L L + J µ µ χ = L µ J µ χ = L

14 Particles and Fields Scattering Amplitudes On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations Particles: Experimentally accesible quantum states, well defined Mass, Charges and Free Propagation Fields: Tools used to construct Lorentz Invariant dynamics: φ(x) = d 3 p exp(ip µ x µ )u( p )â( p ) Û(Λ)(φ(x)) = S(Λ)(φ(Λ 1 x)), Λ SO(1, 3) L = d 4 x L( µ φ, φ) is invariant. Problem: No massless vector field (e.g. A µ in QED) Û(Λ)(A µ (x)) = Λ µ ν A ν (Λ 1 x) + µ χ Solution: Coupling with a conserved current J µ L L + J µ µ χ = L µ J µ χ = L

15 Particles and Fields Scattering Amplitudes On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations Particles: Experimentally accesible quantum states, well defined Mass, Charges and Free Propagation Fields: Tools used to construct Lorentz Invariant dynamics: φ(x) = d 3 p exp(ip µ x µ )u( p )â( p ) Û(Λ)(φ(x)) = S(Λ)(φ(Λ 1 x)), Λ SO(1, 3) L = d 4 x L( µ φ, φ) is invariant. Problem: No massless vector field (e.g. A µ in QED) Û(Λ)(A µ (x)) = Λ µ ν A ν (Λ 1 x) + µ χ Solution: Coupling with a conserved current J µ L L + J µ µ χ = L µ J µ χ = L

16 Gauge redundancy Scattering Amplitudes On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations Particles know nothing about gauge freedom Scattering Amplitudes are gauge invariant (Ward Identities p µ A µ = 0) Fields transform with the gauge Feynman diagrams are not gauge invariant Feynman Expansion A n = Γ Graphs Value(Γ) Complicate pattern of cancellations among gauge dependent terms!

17 Gauge redundancy Scattering Amplitudes On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations Particles know nothing about gauge freedom Scattering Amplitudes are gauge invariant (Ward Identities p µ A µ = 0) Fields transform with the gauge Feynman diagrams are not gauge invariant Feynman Expansion A n = Γ Graphs Value(Γ) Complicate pattern of cancellations among gauge dependent terms!

18 QED Compton Scattering On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations A 4 ( f + f γ + γ ) = 2e 2 24 [q 4 ( 1] 1 3] 3 ) q 3 [31] + 13 [13] q 3 3 [q 4 4] + 2e 2 2q 3 [3 ( 1] 1 4] 4 ) 4 [q 4 1] 14 [14] q 3 3 [q 4 4] = 2e

19 On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations Amplitudes as building blocks for Amplitudes Gauge Invariant Expansion? A n = gauge invariants GI Amplitudes are natural invariant objects Feynman diagrams suggest a way: Look at poles and residues of Amplitudes

20 On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations Poles and Residues of Tree Amplitudes I

21 On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations Poles and Residues of Tree Amplitudes I

22 On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations Poles and Residues of Tree Amplitudes I

23 On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations Poles and Residues of Tree Amplitudes I

24 On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations Poles and Residues of Tree Amplitudes I

25 On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations Poles and Residues of Tree Amplitudes II To sum up: A n (p) has a pole if P 2 I = 0, P I = i I p i. The residue is given by amplitudes with fewer legs: Res A n (p) = A nl A nr, n L + n R = n + 2. A nl A nr P 2 I Tempting Conjecture: A n = I Not so easy: in general one gets spurious poles A more refined strategy: Britto-Cachazo-Feng-Witten recursion relations.

26 On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations Poles and Residues of Tree Amplitudes II To sum up: A n (p) has a pole if P 2 I = 0, P I = i I p i. The residue is given by amplitudes with fewer legs: Res A n (p) = A nl A nr, n L + n R = n + 2. A nl A nr P 2 I Tempting Conjecture: A n = I Not so easy: in general one gets spurious poles A more refined strategy: Britto-Cachazo-Feng-Witten recursion relations.

27 On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations Poles and Residues of Tree Amplitudes II To sum up: A n (p) has a pole if P 2 I = 0, P I = i I p i. The residue is given by amplitudes with fewer legs: Res A n (p) = A nl A nr, n L + n R = n + 2. A nl A nr P 2 I Tempting Conjecture: A n = I Not so easy: in general one gets spurious poles A more refined strategy: Britto-Cachazo-Feng-Witten recursion relations.

28 On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations BCFW Recursion Relations in a Nutshell Goal Exploit knowledge of singularity structure in a constructive way Deform momenta p i (z) = p i + zr i, z C, in such a way that 1 p i (z) are on-shell, i.e. p i (z) 2 = 0 2 Satisfy momentum conservation i p i(z) = 0 3 Mandelstam invariants are linear in z: P 2 I (z) = P2 I + zd I Poles of An(z) z = 0 z I = P2 I D I z : Res Res A n (0) = A n A n z I = An L (z I )A nr (z I ) PI 2 Use Cauchy theorem in CP 1 : A n = A nl (z I )A nr (z I ) P 2 I

29 BCFW in Action Scattering Amplitudes On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations Parke-Taylor formula = n1 Amplitudes beyond Maximally Helicity Violating order

30 BCFW in Action Scattering Amplitudes On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations Parke-Taylor formula = n1 Amplitudes beyond Maximally Helicity Violating order

31 Beyond Tree Level Scattering Amplitudes On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations BCFW for the integrand Graphs d 4 l d 4 l Graphs Unitarity Cuts A n,1 = i d i(box) + i c i(triangle) + sum i b i (bubble) + rational The Symbol, Cluster Algebras QCD loop computations (Dixon) Unexpected UV finiteness of N = 8 SUGRA (Bern, Carrasco, Johansson)

32 Beyond Tree Level Scattering Amplitudes On-Shell vs Off-Shell On-Shell building blocks BCFW Recursion Relations BCFW for the integrand Graphs d 4 l d 4 l Graphs Unitarity Cuts A n,1 = i d i(box) + i c i(triangle) + sum i b i (bubble) + rational The Symbol, Cluster Algebras QCD loop computations (Dixon) Unexpected UV finiteness of N = 8 SUGRA (Bern, Carrasco, Johansson)

33 Two Puzzles from BCFW Amplitudes as Volumes Amplitudes from Positive Geometry Spurious Poles: Individual terms have spurious poles that cancels in the sum. Non trivial relations: Different deformation schemes relations among Gauge invariant objects. The simplest example: 6-point NMHV amplitude in N = 4 Super Yang-Mills A 6 [1, 2, 3, 4, 5, 6] = [3, 1, 6, 5, 4] + [3, 2, 1, 6, 5] + [3, 2, 1, 5, 4] = [2, 3, 4, 6, 1] + [2, 3, 4, 5, 6] + [2, 4, 5, 6, 1]

34 Amplitudes as Volumes Amplitudes from Positive Geometry Kinematics in (Bosonized) Twistor Space Kinematical Space Space-Time Bosonized Twistor Space Variables p i R 1,3 Z i CP 4, 0 = (0, 0, 0, 0, 1) Physical Poles ( j a=i p a) 2 = 0 0, i, i + 1, j, j + 1 = 0 [i, j, k, l, m] := 1 4! d 4 ψ i,j,k,l,m 4 0,i,j,k,l 0,j,k,l,m 0,k,l,m,i 0,l,m,i,j 0,m,i,j,k A 6 [1, 2, 3, 4, 5, 6] = [2, 3, 4, 6, 1] + [2, 3, 4, 5, 6] + [2, 4, 5, 6, 1] [2, 3, 4, 6, 1] 0, 4, 6, 1, 2, 0, 2, 3, 4, 6 The Integrand is the volume of a 4-simplex in CP 4, with co-dimension 1 boundaries Z i W = 0! (Hodges) The Amplitude is the volume of a polytope.

35 Amplitudes as Volumes Amplitudes from Positive Geometry Kinematics in (Bosonized) Twistor Space Kinematical Space Space-Time Bosonized Twistor Space Variables p i R 1,3 Z i CP 4, 0 = (0, 0, 0, 0, 1) Physical Poles ( j a=i p a) 2 = 0 0, i, i + 1, j, j + 1 = 0 [i, j, k, l, m] := 1 4! d 4 ψ i,j,k,l,m 4 0,i,j,k,l 0,j,k,l,m 0,k,l,m,i 0,l,m,i,j 0,m,i,j,k A 6 [1, 2, 3, 4, 5, 6] = [2, 3, 4, 6, 1] + [2, 3, 4, 5, 6] + [2, 4, 5, 6, 1] [2, 3, 4, 6, 1] 0, 4, 6, 1, 2, 0, 2, 3, 4, 6 The Integrand is the volume of a 4-simplex in CP 4, with co-dimension 1 boundaries Z i W = 0! (Hodges) The Amplitude is the volume of a polytope.

36 Amplitudes as Volumes Amplitudes from Positive Geometry Kinematics in (Bosonized) Twistor Space Kinematical Space Space-Time Bosonized Twistor Space Variables p i R 1,3 Z i CP 4, 0 = (0, 0, 0, 0, 1) Physical Poles ( j a=i p a) 2 = 0 0, i, i + 1, j, j + 1 = 0 [i, j, k, l, m] := 1 4! d 4 ψ i,j,k,l,m 4 0,i,j,k,l 0,j,k,l,m 0,k,l,m,i 0,l,m,i,j 0,m,i,j,k A 6 [1, 2, 3, 4, 5, 6] = [2, 3, 4, 6, 1] + [2, 3, 4, 5, 6] + [2, 4, 5, 6, 1] [2, 3, 4, 6, 1] 0, 4, 6, 1, 2, 0, 2, 3, 4, 6 The Integrand is the volume of a 4-simplex in CP 4, with co-dimension 1 boundaries Z i W = 0! (Hodges) The Amplitude is the volume of a polytope.

37 Amplitudes as Volumes Amplitudes from Positive Geometry Kinematics in (Bosonized) Twistor Space Kinematical Space Space-Time Bosonized Twistor Space Variables p i R 1,3 Z i CP 4, 0 = (0, 0, 0, 0, 1) Physical Poles ( j a=i p a) 2 = 0 0, i, i + 1, j, j + 1 = 0 [i, j, k, l, m] := 1 4! d 4 ψ i,j,k,l,m 4 0,i,j,k,l 0,j,k,l,m 0,k,l,m,i 0,l,m,i,j 0,m,i,j,k A 6 [1, 2, 3, 4, 5, 6] = [2, 3, 4, 6, 1] + [2, 3, 4, 5, 6] + [2, 4, 5, 6, 1] [2, 3, 4, 6, 1] 0, 4, 6, 1, 2, 0, 2, 3, 4, 6 The Integrand is the volume of a 4-simplex in CP 4, with co-dimension 1 boundaries Z i W = 0! (Hodges) The Amplitude is the volume of a polytope.

38 Triangles and Polytopes in the Plane Amplitudes as Volumes Amplitudes from Positive Geometry Z i CP 2, i = 1, 2, 3 1, 2, 3 = {W Z 1,2,3 W = 0} Area(T ) = 1 2 1,2,3 0,1,2 0,2,3 0,3,1 =: [1, 2, 3] [4, 1, 3] [2, 1, 3] = Area(P) = [2, 4, 3] [1, 2, 4]

39 Triangles and Polytopes in the Plane Amplitudes as Volumes Amplitudes from Positive Geometry Z i CP 2, i = 1, 2, 3 1, 2, 3 = {W Z 1,2,3 W = 0} Area(T ) = 1 2 1,2,3 0,1,2 0,2,3 0,3,1 =: [1, 2, 3] [4, 1, 3] [2, 1, 3] = Area(P) = [2, 4, 3] [1, 2, 4]

40 Triangles and Polytopes in the Plane Amplitudes as Volumes Amplitudes from Positive Geometry Z i CP 2, i = 1, 2, 3 1, 2, 3 = {W Z 1,2,3 W = 0} Area(T ) = 1 2 1,2,3 0,1,2 0,2,3 0,3,1 =: [1, 2, 3] [4, 1, 3] [2, 1, 3] = Area(P) = [2, 4, 3] [1, 2, 4]

41 Triangles and Polytopes in the Plane Amplitudes as Volumes Amplitudes from Positive Geometry Z i CP 2, i = 1, 2, 3 1, 2, 3 = {W Z 1,2,3 W = 0} Area(T ) = 1 2 1,2,3 0,1,2 0,2,3 0,3,1 =: [1, 2, 3] [4, 1, 3] [2, 1, 3] = Area(P) = [2, 4, 3] [1, 2, 4]

42 Triangles and Polytopes in the Plane Amplitudes as Volumes Amplitudes from Positive Geometry Z i CP 2, i = 1, 2, 3 1, 2, 3 = {W Z 1,2,3 W = 0} Area(T ) = 1 2 1,2,3 0,1,2 0,2,3 0,3,1 =: [1, 2, 3] [4, 1, 3] [2, 1, 3] = Area(P) = [2, 4, 3] [1, 2, 4]

43 Triangles and Polytopes in the Plane Amplitudes as Volumes Amplitudes from Positive Geometry Z i CP 2, i = 1, 2, 3 1, 2, 3 = {W Z 1,2,3 W = 0} Area(T ) = 1 2 1,2,3 0,1,2 0,2,3 0,3,1 =: [1, 2, 3] [4, 1, 3] [2, 1, 3] = Area(P) = [2, 4, 3] [1, 2, 4]

44 6-Points NMHV as a Volume Amplitudes as Volumes Amplitudes from Positive Geometry Back to higher dimension A NMHV 6 = Volume(P) P = {(1, 2, 3, 4); (1, 2, 3, 6); (1, 3, 4, 6); (1, 4, 5, 2); (1, 4, 5, 6); (1, 2, 5, 6)} P is triangulated by 3 4-simplices, non-local vertices cancels. Different BCFW representations give different triangulations of P Beyond NMHV amplitudes?

45 The Scattering Amplitudes Amplitudes as Volumes Amplitudes from Positive Geometry Gr + (4, 4 + k) A = C Z, C Gr + (4, n), Z Gr + (n, 4 + k)

46 Summary Scattering Amplitudes Amplitudes as Volumes Amplitudes from Positive Geometry Scattering Amplitudes are among the most important quantities in theoretical physics. Efficient way to compute them have immediate applications to experimental physics In a virtuous circle, these methods shred new lights upon hidden structures (and hopefully principles) in QFT

47 Thanks for your attention! Amplitudes as Volumes Amplitudes from Positive Geometry