Date post: | 13-Feb-2017 |
Category: |
Documents |
Upload: | truongthien |
View: | 237 times |
Download: | 1 times |
An invitation to Scattering Amplitudes
Sangmin LeeSeoul National University
29 January 2015Joint Winter Conference on Particle Physics, String and Cosmology
References
N. Arkani-Hamed, talks and lectures
F. Cachazo, talks and lectures
M. Spradlin, “An Introduction to Amplitudes,”Scattering Amplitudes in Hong Kong, Nov. 2014
J. Bourjaily, overview talk, Grassmannian Geometry of Scattering Amplitudes, Dec. 2014
H. Elvang, Y.-t. Huang, “Scattering Amplitudes in Gauge Theory and Gravity”Cambridge Univ. Press, Mar. 2015
Progress in scattering amplitudes of gauge theory and gravity
Simple results
Efficient algorithms
Deeper understanding
Old and new ideas are being unified!
Field theory vs S-matrix theory
a f = Âi
S f iai , SS† = 1
[Jacob Bourjaily, talk 2014]
The final result was summarized in 8 pages.
220 Feynman diagrams, thousands of terms.
=hiji4
h12ih23i · · · hn1i d4(P)
Simplification and generalization
+
++
+++
+
++
i �
� j
n-point MHV amplitude
[Parke-Taylor ’86][Giele-Berends ’88]
[Marcus Spradlin, talk 2014]
1. Simplifications do not happen by accident.
2. This is an experimental science.
“The Philosophy of Amplitudeology”
N=4 SYM
One-loop
6-gluon
amplitude
(schematically)
in modern notation.
= [12345]Li2(u)+ [12356]Li2(v)+ [13456]Li2(w)
[Zvi Bern, KITP colloquium] [Marcus Spradlin, talk 2014]
Color decomposition
Full amplitude (permutation symmetry)
“Color-ordered” amplitude(cyclic symmetry only)
Color ordered amplitude
Color ordered Feynman rules (Less diagrams, simpler Feynman rules)
Ma1,a2,...,ann (p1, p2, . . . , pn) = Â
s�Sn/Zn
Tr(Ts(1) · · · Ts(n))An(s(1), · · · , s(n))
Spinor helicity notation
h = +1 h = �1
[Berends, Kleiss, Troost, Wu, Xu... 80’s]
Massless momentum in bi-spinor notation: paa = lala
Lorentz invariants: hiji ⌘ eab(li)a(lj)b , [ij] ⌘ eab(li)a(lj)b
Helicity weights:
(l, l) ! (tl, t�1l) , e+ ! t�2e+ , e� ! t+2e�
=) An(tili, t�1i li, hi) =
n
’i=1
t�2hii
!An(li, li, hi)
3-gluon amplitude
p1 + p2 + p3 = 0
1-2-
3+
An(tili, t�1i li; hi) =
’i
t�2hii
!An(li, li; hi)
An(sli, sli; hi) = s�2(n�4)An(li, li; hi)
A3 =h12i3
h23ih31i .
A3 is the unique Lorentz invariant satisfying the following scaling properties:
Caution: momentum complexified
Twistor
Twistors “linearize” conformal symmetry
Twistor transformation
Paa = lala ! la∂
∂µa
Kaa =∂
∂la
∂
∂la! µa
∂
∂la
A(⇥, ⇥) ! A(⇥, µ) =Z
d2⇥ eiµ� ⇥� A(⇥, ⇥)
ZA ⌘ (⇥�, µ�)
MAB = ZA ∂
∂ZB � (trace) 2 SU(2, 2)
[Penrose 67, Witten 03]
Half-way Fourier transform
e
ip·x !Z
e
ix
aalalae
�iµalad
2l µ d(µa � x
aala)
BCFW recursion [Britto-Cachazo-Feng(-Witten) 04(05)]
Higher point amplitude from lower ones
Idea: on-shell deformation of momenta
lj ! lj � zll , ll ! ll + zlj
(p2j = 0, p2
l = 0, p1 + · · ·+ pn = 0 unaffected)
An = An(z = 0) =Z dz
2piAn(z)
z
(deformed contour, fall-off at infinity, residue theorem)
j lJ Lj l. . .
12 n
. ..
n-1
....
..
1PJ(0)2
.. .... .
.
. .. .
�h +h
pj(zJ)
ÂJ[L=All
pl(zJ)
BCFW and gauge symmetry
No need for 4-point vertex!
Build up higher point amplitude from lower ones
An = ÂJ,±
A±J (zJ)
1PJ(0)2 A⌥
L (zJ)
j lJ Lj l. . .
12 n
. ..
n-1
....
..
1PJ(0)2
.. .... .
.
. .. .
�h +h
pj(zJ)
ÂJ[L=All
pl(zJ)
Grassmannian
Momentum conservation in spinor-helicity
BCFW deformation is a special case of
li ! Sijl
j , li ! lj(S�1)ji , S 2 GL(n, C)
[Britto, Cachazo, Feng, Witten 04-05]
[Arkani-Hamed,Cachazo,Cheung,Kaplan 09]
P-conservation at each vertex: Grassmannian
Cmi 2 Gr(k, n)
l (2-plane)
l (2-plane)C (k-plane)
n
Âi=1
(pi)aa =n
Âi=1
lialia =
�l1a · · · lna
�0
B@l1
a...
lna
1
CA
Grassmannian Integral [Arkani-Hamed,Cachazo,Cheung,Kaplan 09]
C = k
(i + k � 1)(i)
Mi = em1···mk Cm1(i)Cm2(i+1) · · ·Cmk(i+k�1)
(C · W)m = CmiWi
n
GL(k)⇥ GL(n)
“Contour integral over Grassmannian”
An,k(W) =Z dk⇥nC
vol [GL(k)]d4k|4k(C · W)
M1
M2
· · · Mn�1
Mn
W = (µ, ⇥; �)
Dual Super-conformal
Dual super-conformal symmetry
yi+1 � yi = pi
y2
y3
yn+1 ⌘ y1
yn
p1
p2
First noticed in perturbation theory
Physical explanation from AdS/CFT, T-duality & Wilson loops.[Alday-Maldacena 07; Berkovits-Maldacena 08, Beisert-Ricci-Tseytlin-Wolf 08]
Original + dual super-conformal >> Yangian [Drummond-Henn-Plefka 09]
... mutually non-local and non-linear (spin-chain analogy)
[Drummond-Henn-Korchemsky-Sokatchev 07]
Yangian and Integrability
Heisenberg spin chain
Original + dual super-conformal Yangian of SL(4|4)
H = �J Âi
~Si · ~Si+1
integrable = infinite number of conserved charges
Yangian of SU(2)
(Near) integrability of string theory in AdS5
On-shell diagrams and permutations[Arkani-Hamed,Bourjaily, Cachazo,Goncharov,Postnikov,Trnka 12]
All tree amplitudes & loop integrands from 3-point amplitudes via “BCFW bridging”
Permutation, positive Grassmannian
On-shell diagrams
TREE 4-point MHV =
??? =
Amplituhedron
Amplitude as “volume”of some polytope.
Hidden symmetries become manifest.
Unitarity and locality “emerge” from positivity of the polytope.
[Arkani-Hamed,Trnka 13]
Scattering equation [Cachazo,He,Yuan 13]
Âb 6=a
pa · pbza � zb
= 0 (za 2 CP1)
Valid for massless spin 0, 1, 2 particles in all dimensions
Generalizes Witten’s twistor string theory
Same equation appears in the high energy string scattering
[Gross,Mende 87]
KK/BCJ/KLT
Kleiss-Kuijf identities
Bern-Carrasco-Johansson identities
#(independent amplitudes) : (n � 1)! ! (n � 2)!
#(independent amplitudes) : (n � 2)! ! (n � 3)!
[Bern,Carrasco,Johansson 08]
f abc = Tr(TaTbTc)� Tr(TbTaTc) , f abe fecd + f bce fe
ad + f cae febd = 0
An(pi, hi, ai) = Âs2Sn/Zn
Tr(Tas(1) · · · Tas(n) )An(s(1h1), · · · , s(nhn))
A4 =nscs
s+
ntctt
+nucu
ucs + ct + cu = 0
ns + nt + nu = 0
color (Jacobi)
kinematics An(gauge) = Â
I
cInIDI
=) Mn(gravity) = ÂI
nI nIDI
!
BCJ doubling
[Kleiss-Kuijf 89]
Loop integration, dLog, …
Polylogarithms
Symbols
Cluster coordinates
Lik+1(z) =Z z
0Lik(t)
dtt
Beyond D=4, N=4 SYM
Stringy corrections
Other dimensions (e.g. D=3, N=6 ABJM)
Non-SUSY theories
Massive particles
…
Outlook
We will see a lot more within our lifetime!