Elastic ep scattering and higher radiative corrections
Part II
Krzysztof M. Graczyk
Institute of Theoretical Physics Neutrino Physics Division
University of Wrocław Poland
December 6, 2011
Motivation
Jeśli nie przestaniecie udowadniać tego, co już zrobili inni, nabierać pewności, komplikować rozwiązań – po prostu dla przyjemności – wtedy, pewnego dnia rozejrzycie się, i stwierdzicie, że tego jeszcze nikt nie zrobił! To jest sposób zostania uczonym!
R. P. Feynman.
PT and Rosenbluth
There is a systematic discrepancy between ratio µpGE/GM data extracted from PT and cross section measurements!
Figure:
Taken from C. Perdrisat, V. Punjabi and M. Vanderhaeghen, Prog. Part.Nucl. Phys. 59 (2007) 694.
PT and Rosenbluth
I The two-photon exchange (TPE) correction (Born-like) is responsible for that!
I The PT data is less affected by TPE correction than cross section
measurements! P. A. M. Guichon and M. Vanderhaeghen, Phys. Rev. Lett. 91 (2003) 142303, P. G. Blunden, W. Melnitchouk and J. A. Tjon, Phys. Rev. Lett.
91 (2003) 142304.
electron/positron scattering off proton
I At least two new experiments dedicated to the investigation of the TPE contribution!
σ(e+p → e+p)
σ(e−p → e−p) ≈ 1 −2∆C2γ
σ1γ
. (1)
I J. Arrington et al., Two-photon exchange and elastic scattering of electrons / positrons on the proton. (Proposal for an experiment at VEPP-3), arXiv:nucl-ex/0408020;
I Jefferson Lab experiment E04-116, Beyond the Born Approximation: A Precise Comparison of e+p and e−p Scattering in CLAS, W. K. Brooks, et al., spokespersons.
The Proton Radius
I The Proton Radius is extracted from CODATA, as it has been already explained,
p
hr2i = 0.8768 ± 0.0069 fm (P. J. Mohr, B. N. Taylor and D. B. Newell, Rev.Mod. Phys. 80, 633 (2008).) I Lamb shift in muonic atom,
p
hr2i = 0.84184 ± 0.00067 fm, R. Pohl, A.
Antognini, F. Nez et al., Nature 466, 213 (2010).
I The results are 5σ away of each other!
I The Lamb shift is a small difference in energy between two energy levels2S1/2 and2P1/2 of the hydrogen atom. According to Dirac, the2S1/2and2P1/2 orbitals should have the same energies. However, the interaction between the electron and the vacuum causes a tiny energy shift on2S1/2. (see e.g. K.
Pachucki, Phys. Rev. A60 (1999) 3593.)
Lexp = 206.2949 ± 0.0032 meV (2)
Lth = 209.9779(49) − 5.2262
p
hr2i + 0.00913
p
hr3i(2) (3)
where hr3i(2) is the third Zemach moment defined as:
hr3i(2)=
Z
d3rd3r0|r − r0|3ρ(r0)ρ(r) (4)
p
hr2iLAMB = 0.84184 ± 0.00067 fm (5)
p
hr2iCODATA = 0.8768 ± 0.0069 fm (6)
p
hr2idipole = 0.81 fm (7)
p
hr2Ei
NN = 0.85 fm (8)
p
hrM2iNN = 0.82 fm (9)
(10) NN from K.M. Graczyk, Phys. Rev. C 84, 034314 (2011).
Electron Scattering off Coulomb Potential
Electron Scattering off Coulomb Potential
Our attention is concentrated on the first and second order Born diagrams.
|M|2 ≈ |M(1)|2
| {z }
α2
+ 2Re M(1)∗M(2)
| {z }
α3
+ |M(2)|2
| {z }
α4
,(11)
d σcou.
d Ω = |p0| 16π2|p|·1
2
X
|M|2 ≈ d σ(1)cou.d Ω +d σ(2)cou.
d Ω +d σ(3)cou.
d Ω (12)
T-Matrix: 1st Born
hp0|iT(1)
fi |pi = hp0| −i
Z
d4xeψ(x )γµψ(x )Aµ(x ) | pi (13)
= −ieu(p0)γµu(p)
Z
d4xei (p0−p)xAµ(x ) (14)
= −ieu(p0)γµu(p)˜Aµ(p0− p) (15)
= −(2πi)δ(Ef − Ei)eu(p0)γµu(p)˜Aµ(p0− p) (16)
where we have assumed that Aµis time independent, and ˜Aµ(p0− p) is the Fourier transform:
A˜µ(p0− p) = (2π)δ(Ef− Ei)˜Aµ(p0− p), (17) and
A˜µ(p0− p) =
Z
d3re−i(p0−p)·rAµ(r). (18)
T-Matrix: 2st Born
hp0| iT(2)| pi = −1 2hp0|
Z
d4xd4ye2ψ(x )γµΨ(x )Aµ(x )ψ(y )γνΨ(y )Aν(y ) | pi (19)
= − ie2
Z
d4l (2π)4u(p0)γµ(ˆl + me)γνu(p) l2− m2e+ i
A˜µ(p0− l)˜Aν(l − p)
(20)
Coulomb Potential and M-matrix Point-like Coulomb potential
Aµpoin(~r) = gµ0 Ze
4π|r|. (21)
A˜0poin(p0− p) =
Z
d3re−i(p0−p)·r Ze 4π|r| = Ze
4πq2
Z
d3r 4 e−i(p0−p)·r
1|r|(22)
= Ze
4πq2
Z
d3re−i(p0−p)·r4
1|r|
= −Ze
q2. (23)
Hence
hp0|iT(1)|pi = (2πi)δ(Ef − Ei)Ze2
q2 u(p0)γ0u(p). (24) hp0|iT(2)|pi = −(2πi)δ(p00− p0)Z2e4
Z
d3l (2π)3u(p0)(γ0E + ~γ · l + me)u(p) (l2− p2+ i )(p0− l)2(l − p)2, (25) where we have integrated the energy component of dl0, namely
Z
dl02π(2π)δ(p00− l0)(2π)δ(l0− p0) = 2πδ(p00− p0).
and we substituted, l0= E as well as l2= l02− l2− m2e= p2− l2. Hence
γ0(ˆl + me)γ0= γ0(γ0l0− ~γ · l + me)γ0= γ0l0+ ~γ · l + me→ γ0E + ~γ · l + me (26)
Coulomb Potential and M-matrix
In order to compute the cross section one has to compute the M matrix,
hp0| iT | pi = M(2πi)δ(p00− p0). (27)
Similarly as in the case of the T matrix, the M matrix can be written in as the perturbative series:
M = M(1)+ M(2)+ ... (28)
I First order:
M(1)=Ze2
q2 u(p0)γ0u(p). (29)
I Second order:
M(2)= −Z2e4
Z
d3l (2π)3u(p0)(γ0E + ~γ · l + me)u(p)
(p2− l2+ i )(p0− l)2(l − p)2. (30)
Arbitrary Potential – Form Factor
Suppose that the Coulomb potential is not point-like but has its own spherical distribution.
Aµ(~r) = gµ0φ(r), φ(r) = Z 4π
Z
d3r0 ρ(r0)
|r − r0|, Z =
Z
d3r ρ(r) (31)
The Fourier transformation of the potential reads
A0(q) = Ze 4π
Z
d3rd3r0e−iq·r ρ(r0)
|r − r0|= Ze 4πq2
Z
d3rd3r0e−iq·r4r
ρ(r0)
|r − r0|
= Ze q2
Z
d3rd3r0e−iq·rρ(r0)δ(3)(r − r0)
= Ze q2
Z
d3re−iq·rρ(r) ≡ Ze q2 F (q)
|{z}
form factor
. (32)
F (q) =
Z
d3re−iq·rρ(r) ρ(r) =
Z
d3rei q·rF (q). (33)
Arbitrary Potential – Form Factor
d σ1γ
d Ω = Z2α2 4β2p2sin4 θ2
1 − β2sin2
θ2
, (34)
where
β2=|p2|
E2 = velectron2 . (35)
Notice that for β → 0 we have well known Mott scattering formula,
d σ(1)coul .
d Ω ≈ Z2α2
4β2p2sin4 θ2
. (36)Notice that if instead of the point-like potential the one given by (31) is discussed the cross section in the first order Born approximation reads
d σ(1)
d Ω = α2
4β2p2sin4 θ2
1 − β2sin2
θ2
F2(q). (37)
I Point-like static potential:
d σ
d Ω= α2E
4E3sin4 θ2
. (38)I Spatial-charge distribution but still static:
d σ
d Ω= α2E 4E3sin4 θ2
F2(q2). (39)
I Proton with spin 1/2:
d σ
d Ω= α2E0 4E3sin4 θ2 ·
cos2θ 2+ Q2
2M2sin2θ 2
, (40)
(recoil correction,spin 1/2 correction)
I Proton with spin 1/2, and magnetic anomalous moment
d σ d ΩLAB
= α2E0 4E3sin4 θ2 ·
cos2θ 2
F12+ Q2 4M2F22
+ Q2 2M2sin2θ
2(F1+F2)2
(41) (γµ,σµν).
Notice that for Q2→ 0, sin2 θ2→ 0, cos2 θ2 → 1 the static potential F (q2) form factor has the same meaning as F1(Q2), and GE because GE= F1− τ F2.
Photon Mass
M(2) = −Z2e4
Z
d3l (2π)3u(p0)(γ0E + ~γ · l + me)u(p)
(p2− l2+ i )(p0− l)2(l − p)2 (42)
= Z2e4u(p0) (γ0E I1+ ~γ · I2) u(p) (43)
I Long-range character of the Coulomb forces → infrared singularities I One has to extract the divergent term from the amplitude
I R. H. Dalitz Proc. R. Soc. Lond. A206, 509 (1951).
1 q2 → 1
q2+ µ2. (44)
It corresponds to the screened Coulomb interaction
Aµ(r ) = gµ0Zeexp(−µ|r|) 4π|r| = −
Z
d4q (2π)3δ(q0)
q2− µ2eiq·r (45) The problem is seen already in QM, if the potential V (r ) does not converge faster then 1/r then the partial wave solution can not be obtained. For potential 1/r distorted wave functions are obtained!
Properties of the integrals I1and I2
Notice that I1and I2are symmetric under exchange p ↔ p,
I1=
Z
d3l (2π)31
(l2− p2− i)(p0− l)2(l − p)2, I2=
Z
d3l (2π)3l
(l2− p2− i)(p0− l)2(l − p)2 (46) hence I2∼ p + p0. Notice that u(p0, s0)~γu(p, s) = χ†s0(p + p0+ i q × τ )χs.
Feynman’s Identity 1
(a + λ)(b + λ) = −∂
∂λ
Z
1 0d α 1
αa + (1 − α)b + λ= −
Z
1 0d α ∂
∂(αa)
1 αa + (1 − α)b + λ
(47) Then
I1 = −
Z
1 0d α ∂
∂µ2
Z
d3l (2π)31
(l2− p2− i)((l − P)2+ M02) (48)
= −
Z
1 0d α ∂
∂M02
Z
d3l (2π)31
(l2− p2− i)((l − P)2+ M02) (49)
Ik2 =
Z
1 0d α
∂2∂Pk − Pk ∂
∂M02
Z
d3l (2π)31
(l2− p2− i)((l − P)2+ M02)(50) where
Properties of the integrals I1and I2
I1 = −
Z
1 0d α ∂
∂M02I (52)
Ik2 =
Z
1 0d α
∂2∂Pk − Pk ∂
∂M02
I (53)
I =
Z
d3l (2π)31
(l2− p2− i)((l − P)2+ M02) (54) Notice that
I ∼
Z
d3l l4|M|2
|M|2≈ |M(1)|2+2 Re M(1)∗M(2)
| {z }
∗
+ |M(2)|2, (55)
X
spin
∗ = Z3e6
|q|2
X
spin
u(p0)γ0u(p)
∗I1E u(p0)γ0u(p) + I2· u(p0)~γu(p)
= Z3e6
|q|2
"
I1E Tr(ˆpγ0pˆ0γ0) +
3
X
k=1
Ik2Tr(ˆpγ0pˆ0γk)
#
= Z3e6
|q|2
4I1E3(1 + β2cos θ) + 4E I2· (p + p0)(56)
= 4Z3e6E3cos2 θ2 q2
(1 + β2cos θ)cos2 θ2 I1+I2· (p + p0) E2cos2 θ2
(57)
|M|2: useful expressions
Trˆpγ0pˆ0γ0 = 4
2EiEf − p · p0
(58)
= 4(E2+ p · p0) = 4E2(1 + β2cos θ) (59) Trˆpγ0pˆ0γk = 4
pkE + p0kE
(60)
3
X
k=1
IkTrˆpγ0pˆ0γk = 4E I2· (p + p0) (61)
(62)
(p + p0)2E = 2|p|(p2+ p · p0) = 4|p|3 β cos2θ
2 (63)
p2+ 2p · p0 = |p|2(4 cos2θ
2− 1) (64)
I
I(p2, P, M02) =
Z
d3l (2π)31
(l2− p2− i)((l − P)2+ M02(µ2)) (65)
=
Z
1−1
dt
Z
∞ 0dl (2π)2
l2 (l2− p2− i)
1
l2− 2tPl + P2+ M02 (66)
= 1
2
Z
1−1
dt
Z
∞0
dl (2π)2
l2 (l2− p2− i)
1
l2− 2tPl + P2+ M02
+1 2
Z
1−1
dt
Z
∞ 0dl (2π)2
l2 (l2− p2− i)
1
l2+ 2tPl + P2+ M20(67)
t → −t,
Z
1−1
dt →
Z
1−1
dt (68)
(69)
I
I = 1
2
Z
1−1
dt
Z
∞0
dl (2π)2
l2 (l2− p2− i)
1
l2− 2tPl + P2+ M02
+1 2
Z
1−1
dt
Z
0−∞
dl (2π)2
l2 (l2− p2− i)
1
l2− 2tPl + P2+ M02,(70)
l → −l ,
Z
∞0
dl →
Z
0−∞
dl (71)
I = 1
2
Z
1−1
dt
Z
∞−∞
dl (2π)2
l2 (l2− p2− i)
1
l2− 2tPl + P2+ M02 (72)
l2− p2− i = (l − p − i )(l + p + i ) (73)
= 1
2
Z
1−1
dt
Z
∞−∞
dl (2π)2
l2
(l − p − i )(l + p + i )(l − l+)(l − l−)(74), where
l± = Pt ± i
p
(1 − t2)P2+ M02 (75)
M02 = p2+ µ2− P2= µ2+ 4α(1 − α)q2 (76)
I
Z
∞−∞
(...)
| {z }
+
Z
R
(...)
| {z }
= 2πi
X
Res(top semicircle) (77)
I
I = ip
8π
Z
1−1
dt 1
p2− 2tPp + P2+ M02+ 1 8π
Z
1−1
dt l+2
(l+2− p2)
p
(1 − t2)P2+ M02
(78)
= i
16πP
ln p2+ 2Pp + P2+ M02− ln p2− 2Pp + P2+ M02
+1 8π
Z
1−1
dt l+2
(l+2− p2)
p
(1 − t2)P2+ M02
, (79)In the second integral we do the change of the variables, t → l+, indeed
dl+= − iPl+dt
p
(1 − t2)P2+ M20(80)
where 1 8π
Z
1−1
dt l+2
(l+2− p2)
p
(1 − t2)P2+ M02
=i 8πP
Z
dl+
l+
(l+2− p2) = i
16πPln(l+2−p2) (81)
I
I = i
16πP
ln p2+ 2Pp + P2+ M02− ln p2− 2Pp + P2+ M02
+ i 16πP
ln((P + iM0)2− p2) − ln((P − iM0)2− p2), (82)
= i
8πPln
p + P + iM0
p − P + iM0
(83)
I1
I1= −
Z
1 0d α ∂ 2M0∂M0
I (84)
Hence
I1 = 1 8π
Z
1 0d α 2PM0
h
1p + P + iM0
− 1
p − P + iM0
i
= − 1
8π
Z
1 0d α 1 M0
1
(p + iM0)2− P2 (85)
= − 1
8π
Z
1 0d α 1 M0
1
−µ2+ i 2pM0
(86)
= 1
8πµ2
Z
1 0d α M0
− 1
8πµ2
Z
1 0d α 2pi
−µ2+ i 2pM0
(87)
I1
M02 = q2α(1 − α) + µ2=
q2 4 + µ21 − q2
q2 4 + µ2
α −1 2
2(88)
α0 = 2|q|
p
q2+ 4µ2α −1 2
(89)
Now,
Z
d α M0= 1
|q|
Z
d α0p
1 − α02= 1
|q|arcsin(α0)= 1
|q|arcsin 2|q|
p
q2+ 4µ2α −1 2
!
(90)
Z
1 0d α M0
= 2
|q|arcsin |q|
p
q2+ 4µ2!
= 2
|q|arcsin
1
q
1 +4µq22
→π
|q| (91)
where
√ 1 ≈ 1 +x
, arcsin(1 + x ) ≈ π (1 + x )
I1
Z
1 0d α 1
−µ2+ i 2pM0
= 1
2p
Z
1 0d α 1
−µ2p2 + iM0
(92)
= 1
2p
Z
10
d α 1
−µ2
2p + i
p
µ2+ α(1 − α)q2
(93)
= 1
p|q|
Z
120
d α0 1
−2p|q|µ2 + i
q
1 4+µ2q2 − α02 (94)
α0 = α −1
2 (95)
From Mathematica
Z
dx−B + i√
A2− x2 = −i arctan
x√A2− x2
+ (96)
−iB tanh−1
√ Bx
−A2−B2
√
A2−x2
+ B arctan
√ x−A2−B2
√−A2− B2
∗ =
Z
120
d α0 1
−2p|q|µ2 + i
q
14+µq22− α02
(97)
= 1
µ2
−i arctan
|q|2µ
−
i2p|q|µ2
q
1 4+µ2q2 + µ4
4p2q2
arctan
−i 2
q
1 4+µ2q2 + µ4
4p2q2
−−i tanh−1 −i µ 4p
1 +4µ2 q2 + µ4
p2q2
−12!#)
(98)
= −i π 2µ2− 1
p|q|tanh−1
1 +4µ2 q2 + µ4
p2q2
−12!
=−i π 2µ2− 1
p|q|lnp sinθ2
µ (99)
tan(−iz) =1
i tanh(z), tanh(iz) = i tan(z)
1 +4µ2 q2 + µ4
p2q2
−12≈ 1 −1 2
4µ2 q2 + µ4p2q2
≈ 1 −2µ2
q2 (100)
4µ2 µ4 −12!
1
2µ2
1 q2 2p sinθ
I1
I1 = 1 8π
π|q|µ2− 2i
|q|
−i π 2µ2 − 1
p|q|ln2p sinθ2 µ
(101)
= i
4πq2pln2p sinθ2
µ (102)
= i
16π sin2 θ2p3ln2p sinθ2
µ (103)
It is divergent when µ → 0, but does not contribute to the spin averaged interference term 2Re(M(1)M(2)∗)
I2
p · I2 =
Z
d3l (2π)3p · l
(l2− p2− i)((p0− l)2+ µ2)((l − p)2+ µ2) (104) p0· I2 =
Z
d3l (2π)3p0· l
(l2− p2− i)((p0− l)2+ µ2)((l − p)2+ µ2). (105) Notice that
1
(l − p)2+ µ2 − 1
(l2− p2− i) = 2l · p − 2p2− µ2
(l2− p2− i)((l − p)2+ µ2) (106) 1
(l − p0)2+ µ2 − 1
(l2− p2− i) = 2l · p0− 2p2− µ2
(l2− p2− i)((l − p0)2+ µ2) (107) hence
l · p
(l2− p2− i)((l − p)2+ µ2) = 1 2
h
1(l − p)2+ µ2− 1
(l2− p2− i)+ 2p2+ µ2 (l2− p2− i)((l − p)2+ µ2)
i
l · p0
2− p2− i)((l − p02 2 = 1
h
102 2− 1
2− p2− i)+ 2p2+ µ2 2− p2− i)((l − p02 2
i
I2
p · I2 = 1
2
Z
d3l (2π)31 ((p0− l)2+ µ2)
1(l − p)2+ µ2)
− 1
(l2− p2− i)
+
p2+µ2 2
I1
= 1
2(A − Ba) +
p2+µ2 2
I1 (108)
p0· I2 = 1 2
Z
d3l (2π)31 ((p − l)2+ µ2)
1(l − p0)2+ µ2− 1 (l2− p2− i)
+
p2+µ2 2
I1
= 1
2(A − Bb) +
p2+µ2 2
I1 (109)
where
A =
Z
d3l (2π)31
((p0− l)2+ µ2)(l − p)2+ µ2) (110)
Ba =
Z
d3l (2π)31
((p0− l)2+ µ2)(l2− p2− i) (111)
Bb =
Z
d3l (2π)31
((p − l)2+ µ2)(l2− p2− i) (112)
It is easy to show that Ba= Bb≡ B.
A
We use the Feynman’s trick,
1 AB=
Z
1 0dx [Ax + B(1 − x )]2
A =
Z
d3l (2π)31
((p0− l)2+ µ2)(l − p)2+ µ2)=
Z
1 0dx
Z
d3l (2π)31
l2− 2l · (x p + (1 − x )p0) + µ22=
Z
1 0dx
Z
d3l (2π)31
l2− 2l · (x p + (1 − x )p0) + p2+ µ22, l0= l − (x p + (1 − x )p0) (113)
=
Z
1 0dx
Z
d3l0 (2π)31
l0 2+ M022, M2
0= µ2+ q2x (1 − x ) (114)
=
Z
1 0dx
Z
∞−∞
dl 4π2
l2
l2+ M202 =1 8π
Z
1 0dx M0 = 1
8π
Z
1 0dx
p
µ2+ q2x (1 − x )(115)
= 1
8|q| (116)
I2
Analogically w obtain
B = i 8πp
ln
2pµ
− iπ 2
= I(p2, p0, µ2) (117)
hence
(p + p0) · I2 = A − B + 2p2I1= 1 8|q|
− i 8πp
ln
2pµ
− iπ 2
+ ip
4πq2ln2p sinθ2
µ (118)
= 1 − sinθ2 8|q|
− i
16πp
2 ln
2pµ
− 1
sin2 θ2 ln2p sinθ2 µ
| {z }
divergent
(119)
σ(2)
d σcou(2)
d Ω = 1
16π2
X
spin
Re M(1)∗M(2)
(120)
= 1
16π2 Z3e6
|q|2
4ReI1E3(1 + β2cos θ) + 4E ReI2· (p + p0)(121)
= 2Z3πα3E
|q|3
1 − sinθ 2
(122)
= Z3πα3 q2β
1 − sinθ 2
(123)
Rcoul .(2) =σcoul .(2) σ1γ
= πZαβ 1 − sinθ2
1 − β2sin2 θ2
sinθ
2 (124)
0 0.002 0.004 0.006 0.008 0.01 0.012
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 R(2) cou.
sin(θ/2) rel
non-rel
0 0.002 0.004 0.006 0.008 0.01 0.012
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 R(2) cou.
ε
rel, Q2=0.01 non-rel, Q2=0.01
Form-Factors: Time-like-Region
Interesting property:
ImF1,2= p3t
√
tΓ∗1,2(t) (125)
ptpion momentum in the crossed (t−)channel, Γ1,2– P-amplitudes for the ππ − NN.
Form-Factors: Time-like-Region
F (q2) = 1 2πi
I
C dt F (t)
t − q2 = 1 2πi
Z
4m2π∞
dt F (t − i ) t − q2− i+
Z
∞ 4m2πdt F (t + i ) t − q2+ i
!
(126)
= 1
π
Z
∞ 4m2πdtImF (t + i )
t − q2 , F (s∗) = F∗(s) (127)
= F (0) +q2 π
Z
∞ 4m2πdtImF (t + i )
t(t − q2) (128)
hr2i = 6 π
Z
∞ 4m2πdtImF1p(t) t2
| {z }
≈0.65 fm2
+ F2p(0) 4M2
|{z}
3κp 2M2≈0.12 fm2
(129)