Maxima by Example:

Ch. 12, Dirac Algebra and Quantum Electrodynamics

Edwin L. Woollett
April 15, 2011
Contents
12.1 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
12.2 High Energy Physics Notation Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
12.2.1 Contraction and Trace Theorems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
12.3 Introduction to the Dirac Package, Version 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
12.4 Overview and Examples of the Dirac Package Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
12.4.1 Some Basic Package Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
12.4.2 Symbolic Expansions of D(a,b) using Dexpand . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
12.4.3 Some Examples of the Symbolic Expansion of G(a,b,c,...) Using Gexpand . . . . . . . . . . . . 14
12.4.4 Symbolic and Explicit Matrix Trace Examples: tr, Gtr, mat trace, m tr . . . . . . . . . . . . . 16
12.4.5 Symbolic and Explicit Matrix Contraction Examples Using Con . . . . . . . . . . . . . . . . . . 18
12.4.6 Examples of noncov, comp def, VP, and econ . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
12.4.7 Contraction of a Product of Traces: Con, scon, mcon, econ . . . . . . . . . . . . . . . . . . . . 22
12.4.8 Contraction Choices: Timing Comparisons for Polarized Squared Amplitudes . . . . . . . . . . . 23
12.4.9 Dirac Spinors and Helicity Amplitudes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
12.4.10 Polarized Amplitudes for (electron,positron) (muon, antimuon) . . . . . . . . . . . . . . . . . 26
12.5 moller0.mac: Scattering of Identical Scalar (Spin 0) Charged Particles . . . . . . . . . . . . . . . . . . . 32
12.6 moller1.mac: High Energy Elastic Scattering of Two Electrons . . . . . . . . . . . . . . . . . . . . . . . 37
12.7 moller2.mac: Arbitrary Energy Moller Scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
12.8 bhabha1.mac: High Energy Limit of Bhabha Scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
12.9 bhabha2.mac: Arbitrary Energy Bhabha Scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
12.10 photon1.mac: Photon Transverse Polarization 3-Vector Sums . . . . . . . . . . . . . . . . . . . . . . . 53
12.11 Covariant Physical External Photon Polarization 4-Vectors - A Review . . . . . . . . . . . . . . . . . . 54
12.12 compton0.mac: Compton Scattering by a Spin 0 Structureless Charged Particle . . . . . . . . . . . . . . 55
12.13 compton1.mac: Lepton-photon scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
12.14 pair1.mac: Unpolarized Two Photon Pair Annihilation . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
12.15 pair2.mac: Polarized Two Photon Pair Annihilation Amplitudes . . . . . . . . . . . . . . . . . . . . . . 66
12.16 moller3.mac: Squared Polarized Amplitudes Using Symbolic or Explicit Matrix Trace Methods . . . . . 75
12.17 List of Dirac Package Files and Example Batch Files . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
12.18 Test Suite Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
12.19 Using simplifying-new.lisp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

This version uses Maxima 5.23.2 Check http://www.csulb.edu/

woollett/ for the latest version of these notes. Send comments

and suggestions to woollett@charter.net
1
Preface
COPYING AND DISTRIBUTION POLICY
This document is part of a series of notes titled
"Maxima by Example" and is made available
via the authors webpage http://www.csulb.edu/woollett/
to aid new users of the Maxima computer algebra system.
NON-PROFIT PRINTING AND DISTRIBUTION IS PERMITTED.
You may make copies of this document and distribute them
to others as long as you charge no more than the costs of printing.
These notes (with some modifications) will be published in book form
eventually via Lulu.com in an arrangement which will continue
to allow unlimited free download of the pdf files as well as the option
of ordering a low cost paperbound version of these notes.
Feedback from readers is the best way for this series of notes to become more helpful to new users of Maxima. All
comments and suggestions for improvements will be appreciated and carefully considered.
LOADING FILES
The defaults allow you to use the brief version load(fft) to load in the
Maxima file fft.lisp.
To load in your own file, such as qxxx.mac
using the brief version load(qxxx), you either need to place
qxxx.mac in one of the folders Maxima searches by default, or
else put a line like:
file_search_maxima : append(["c:/work2/###.{mac,mc}"],file_search_maxima )$
in your personal startup file maxima-init.mac (see later in this chapter
for more information about this).
Otherwise you need to provide a complete path in double quotes,
as in load("c:/work2/qxxx.mac"),
We always use the brief load version in our examples, which are generated
using the XMaxima graphics interface on a Windows XP computer, and copied
into a fancy verbatim environment in a latex file which uses the fancyvrb
and color packages.
Maxima.sourceforge.net. Maxima, a Computer Algebra System. Version 5.23.2
(2011). http://maxima.sourceforge.net/
The author would like to thank Maxima developer Barton Willis for his initial suggestion to use the Maxima
code le simplifying.lisp and his patient help in understanding how to make use of that code. The new
code has also been simplied (as compared with the rather long-winded code in version 1) by using a set of functions
shared by Barton Willis which are used in making decision branches. These functions are
mplusp, mtimesp, mnctimesp, mexptp, mncexptp.
(The nc additions refer to non-commutative). General expansions of multiple term arguments have been simplied
by using Barton Willis methods which combine lambda and xreduce.
2
3
12.1 References
The following works are useful sources. Some abbreviations (which will be used in the following sections) are dened
here.
Aitchison, I.J.R., Relativistic Quantum Mechanics, Barnes and Noble,1972
A/H: Aitchison,I.J.R, and Hey, A.J.G., Gauge Theories in Particle Physics, Adam Hilger, 1989
B/D: Bjorken, James D. and Drell, Sidney D, Relativistic Quantum Mechanics, McGraw Hill, 1964
BLP: Berestetskii, V. B, Lifshitz, E. M., and Pitaevskii, L. P., Quantum Electrodynamics, Course of Theoretical
Physics, Vol. 4, 2nd. Ed. Pergamon Press, 1982
Berestetskii, V. B, Lifshitz, E. M., and Pitaevskii, L. P., Relativistic Quantum Theory, Part 1, Course of Theoretical
Physics, Vol. 4, Pergamon Press, 1971
De Wit, B. and Smith, J., Field Theory in Particle Physics, North-Holland, 1986
G/R: Greiner, Walter, and Reinhardt, Joachim, Quantum Electrodynamics, fourth ed., Springer, 2009
Grifths, Introduction to Elementary Particles, Harper and Row, 1987
H/M: Halzen, Francis and Martin, Alan D., Quarks and Leptons, John Wiley, 1984
I/Z: Itzykson, Claude and Zuber, Jean-Bernard, Quantum Field Theory, McGraw-Hill, 1980
Jauch, J.M., and Rohrlich, F., The Theory of Photons and Electrons, Second Expanded Edition, Springer-Verlag,
1976
Kaku, Michio, Quantum Field Theory, Oxford Univ. Press, 1993
Maggiore, Michele, A Modern Introduction to Quantum Field Theory, Oxford Univ. Press, 2005
P/S: Peskin, Michael E. and Schroeder, Daniel V., An Introduction to Quantum Field Theory, Addison-Wesley,
1995
Quigg, Chris, Gauge Theories of the Strong, Weak, and Electromagnetic Interactions, Benjamin/Cummings, 1983
Renton, Peter, Electroweak Interactions, Cambridge Univ. Press, 1990
Schwinger, Julian, Particles, Sources, and Fields, Vol. 1, Addison-Wesley, 1970
Sterman, George, An Introduction to Quantum Field Theory, Cambridge Univ. Press, 1993
Weinberg, Steven, The Quantum Theory of Fields, Vol.I, Cambridge Univ. Press, 1995
12.2 High Energy Physics Notation Conventions
We use the same conventions as P/S (Peskin and Schroeder) and Maggiore,
0123
= +1, the electromagnetic units are
Heaviside-Lorentz, and natural units are used, so e
2
= 4.
The only exception is that we interpret the Feynman rules as an expression for the quantity i M, following the conven-
tions of Grifths, and Halzen/Martin.
The diagonal elements of the metric tensor are (1, 1, 1, 1), and the helicity (chiral) representation used for explicit
matrix methods with the Dirac gamma matrices is (in 2 2 block form)

0
=
_
0 1
1 0
_
,
j
=
_
0
j

j
0
_
,
5
=
_
1 0
0 1
_
, (12.1)
4
in which
j
for j = 1, 2, 3 are the 2 2 Pauli matrices

1
=
_
0 1
1 0
_
,
2
=
_
0 i
i 0
_
,
3
=
_
1 0
0 1
_
. (12.2)
The above form for
5
is equivalent to

5
= i
0

3
. (12.3)
The lepton Dirac spinor is the four element column matrix u(p, ) where = +1 picks out positive helicity h = +1/2
and = 1 picks out negative helicity h = 1/2. The corresponding barred particle spinor is a four element row
matrix
u(p, ) = u

0
(12.4)
and the particle Dirac spinor normalization is
u(p, ) u(p,

) = 2 m

, ,

= 1, (12.5)
and we also have

u(p, ) u(p, ) = p / + m, (12.6)

in which a 4 4 unit matrix is understood to be multiplying the scalar mass symbol m, and p / = p

with use of the

summation convention.
The antilepton Dirac spinor is the four element column matrix v(p, ) in which = +1 picks out the positive helicity
h = +1/2 antiparticle state and = 1 picks out negative helicity h = 1/2 antiparticle state, is dened (with P/S
phase choice) via
v(p, ) =
5
u(p, ) (12.7)
The corresponding barred antiparticle spinor is a four element row matrix
v(p, ) = v

0
(12.8)
and the normalization is
v(p, ) v(p,

) = 2 m

,

= 1. (12.9)
We also have

v(p, ) v(p, ) = p / m (12.10)

We dene the 4 4 the spin projection matrix
S(p, ) =
1
2
(1 +
5
s /
R
(p)) (12.11)
in which the 1 is interpreted as the 4 4 unit matrix and in which s /
R
(p) = s
R

, and the right handed spin 4-vector

s

R
generated by the 4-vector p has the components
s

r
(p) =
_
|p|
m
,
E
m
p
_
(12.12)
5
in which E =
_
m
2
+p
2
.
For nite (non-zero) mass m we have
u(p, ) u(p, ) = S(p, ) (p / + m) = (p / + m) S(p, ) (12.13)
and again for nite mass,
v(p, ) v(p, ) = S(p, ) (p / m) = (p / m) S(p, ) (12.14)
In the case of zero mass leptons
u(p, ) u(p, ) =
1
2
(1 +
5
) p / (12.15)
and for zero mass antiparticles
v(p, ) v(p, ) =
1
2
(1
5
) p /. (12.16)
Calculation of unpolarized cross sections requires summing the absolute value squared of a polarized amplitude over the
helicity quantum numbers of the particles involved, and the result can be written as the matrix trace of a 4 4 matrix. A
simple example follows. We use a compressed notation in which, for example, u
1
stands for u(p
1
,
1
) and u
2
stands for
u(p
2
,
2
). Let be an arbitrary 4 4 matrix.
One can show that
( u
2
u
1
)

= u
1

u
2
(12.17)
in which

=
0

0
. (12.18)
Then

1

2
| u
2
u
1
|
2
=

1

2
u
2
u
1
u
1

u
2
(12.19)
We use the sum over
1
to replace u
1
u
1
by (p /
1
+ m), and are left with

2
u
2

u
2
,

.
= (p /
1
+ m)

. (12.20)
Writing out the matrix indices explicitly, the sum becomes (using the summation convention for repeated matrix indices)

2
u
2 a

a b
u
2 b
=

a b

2
u
2 b
u
2 a
=

a b
(p /
2
+ m)
b a
= Trace {

(p /
2
+ m)} (12.21)
Note that Trace (AB) = Trace (B A). We then have

1

2
| u
2
u
1
|
2
= Trace {(p /
2
+ m) (p /
1
+ m)

}. (12.22)
12.2.1 Contraction and Trace Theorems
The basic gamma matrix algebra relations are

+ 2 g

I
4
(12.23)
for Lorentz indices and which can have values (0, 1, 2, 3) and

5
. (12.24)
6
Contraction Theorems

= 4 I
4
(12.25)

= 2

(12.26)

= 4 g

I
4
(12.27)
A general contraction pattern for the case that the pair of contracting gamma matrices surround an odd number (3, 5, ...)
of gamma matrices: we get the order reversed times (2).

= 2

(12.28)
A general contraction pattern for the case that the pair of contracting gamma matrices surround an even number (4, 6, ...)
of gamma matrices: we get two terms with the rst term bringing the last gamma to the front, and the second term being
the rst term in reversed order, all multiplied by (2).

= 2 (

) (12.29)
Trace Theorems
The trace of the product of an odd number of gamma matrices is zero. For the trace of a product of an even number of
gamma matrices (including zero) is based on:
Tr(I
4
) = 4 (12.30)
Tr(

) = 4 g

(12.31)
Tr(

) = 4 (g

g

+ g

g

) (12.32)
For the trace of a product of an even number of gamma matrices, we can use the general formula for the reduction of a
trace of a product of (N = even) matrices to a sum involving the traces of products of (N - 2) matrices.
Lets use a more efcient notation here:
tr(n1, n2, n3, ..., nN) Tr(
n
1

n
2

n
3
. . .
n
N
) (12.33)
The reduction formula is then (see, for example, Barger and Phillips, p. 550 or Kaku, p. 752):
tr(n1, n2, n3, ..., nNm1, nN) = g
n
1
n
2
tr(n3, n4, n5, ..., nN)
g
n
1
n
3
tr(n2, n4, n5, ..., nN)
+ g
n
1
n
4
tr(n2, n3, n5, ..., nN)
. . . + g
n
1
n
N
tr(n3, n4, n5, ..., nNm1)
For example, if N = 6 we have
tr(n1, n2, n3, n4, n5, n6) = g
n
1
n
2
tr(n3, n4, n5, n6)
g
n
1
n
3
tr(n2, n4, n5, n6) + g
n
1
n
4
tr(n2, n3, n5, n6)
g
n
1
n
5
tr(n2, n3, n4, n6) + g
n
1
n
6
tr(n2, n3, n4, n5)
7
Gamma 5 Traces
Since
5
= i
0

3
is proportional to the product of an even number of gamma matrices, the trace of the product
of gamma5 by an odd number of gamma matrices is zero. Hence Tr(
5

) = 0 and Tr(
5

) = 0
When
5
is multiplied by an even number of gamma matrices, we have
Tr(
5
) = 0 (12.34)
Tr(
5

) = 0 (12.35)
Tr(
5

) = 4 i

(12.36)
For the cases that
5
is multiplied by six or eight gamma matrices, we have used the Chisholm-Kahane gamma matrix
identity which replaces a product of three Dirac gamma matrices by a series of four terms. With our (Peskin/Schroeder)
conventions,
0123
= +1 and
5
= +i
0

3
, that identity is (using the summation convention for the repeated
index ):

= g

+ g

(12.37)
Here is the symbolic calculation of the trace of the product of G5 with six gamma matrices.
(%i81) tr6:tr(G5,n1,n2,n3,n4,n5,n6);
(%o81) -4
*
%i
*
Eps(n1,n2,n3,n4)
*
Gm(n5,n6)+4
*
%i
*
Eps(n1,n2,n3,n5)
*
Gm(n4,n6)
-4
*
%i
*
Eps(n1,n2,n3,n6)
*
Gm(n4,n5)
-4
*
%i
*
Gm(n1,n2)
*
Eps(n3,n4,n5,n6)
+4
*
%i
*
Gm(n1,n3)
*
Eps(n2,n4,n5,n6)
-4
*
%i
*
Eps(n1,n4,n5,n6)
*
Gm(n2,n3)
Version 1 of the Dirac package produces a result we will call tr6comp (copied from a separate work session using
version 1):
(%i82) tr6comp : 4
*
%i
*
Eps(n3,n2,n1,n4)
*
Gm(n5,n6)-4
*
%i
*
Eps(n3,n2,n1,n5)
*
Gm(n4,n6)
+4
*
%i
*
Eps(n3,n2,n1,n6)
*
Gm(n4,n5)
-4
*
%i
*
Gm(n1,n2)
*
Eps(n3,n4,n5,n6)
+4
*
%i
*
Gm(n1,n3)
*
Eps(n2,n4,n5,n6)
-4
*
%i
*
Eps(n1,n4,n5,n6)
*
Gm(n2,n3)
If we take the difference, we do not get zero, because we are using a different algorithm in version 2, and the symbol Eps
is not declared antisymmetric by the package.
(%i83) tr6 - tr6comp;
(%o83) -4
*
%i
*
Eps(n3,n2,n1,n4)
*
Gm(n5,n6)-4
*
%i
*
Eps(n1,n2,n3,n4)
*
Gm(n5,n6)
+4
*
%i
*
Eps(n3,n2,n1,n5)
*
Gm(n4,n6)
+4
*
%i
*
Eps(n1,n2,n3,n5)
*
Gm(n4,n6)
-4
*
%i
*
Eps(n3,n2,n1,n6)
*
Gm(n4,n5)
-4
*
%i
*
Eps(n1,n2,n3,n6)
*
Gm(n4,n5)
However, use of noncov introduces eps4, which is declared antisymmetic, and shows the results are equivalent when
the antisymmetry is taken into account.
(%i84) noncov(%);
(%o84) 0
8
Here we calculate a trace of gamma5 multiplied by eight gamma matrices:
(%i85) tr8:tr(G5,n1,n2,n3,n4,n5,n6,n7,n8);
(%o85) -4
*
%i
*
Eps(n1,n2,n3,n4)
*
Gm(n5,n6)
*
Gm(n7,n8)
+4
*
%i
*
Eps(n1,n2,n3,n5)
*
Gm(n4,n6)
*
Gm(n7,n8)
-4
*
%i
*
Eps(n1,n2,n3,n6)
*
Gm(n4,n5)
*
Gm(n7,n8)
-4
*
%i
*
Gm(n1,n2)
*
Eps(n3,n4,n5,n6)
*
Gm(n7,n8)
+4
*
%i
*
Gm(n1,n3)
*
Eps(n2,n4,n5,n6)
*
Gm(n7,n8)
-4
*
%i
*
Eps(n1,n4,n5,n6)
*
Gm(n2,n3)
*
Gm(n7,n8)
+4
*
%i
*
Eps(n1,n2,n3,n4)
*
Gm(n5,n7)
*
Gm(n6,n8)
-4
*
%i
*
Eps(n1,n2,n3,n5)
*
Gm(n4,n7)
*
Gm(n6,n8)
+4
*
%i
*
Eps(n1,n2,n3,n7)
*
Gm(n4,n5)
*
Gm(n6,n8)
+4
*
%i
*
Gm(n1,n2)
*
Eps(n3,n4,n5,n7)
*
Gm(n6,n8)
-4
*
%i
*
Gm(n1,n3)
*
Eps(n2,n4,n5,n7)
*
Gm(n6,n8)
+4
*
%i
*
Eps(n1,n4,n5,n7)
*
Gm(n2,n3)
*
Gm(n6,n8)
-4
*
%i
*
Eps(n1,n2,n3,n4)
*
Gm(n5,n8)
*
Gm(n6,n7)
+4
*
%i
*
Eps(n1,n2,n3,n5)
*
Gm(n4,n8)
*
Gm(n6,n7)
-4
*
%i
*
Eps(n1,n2,n3,n8)
*
Gm(n4,n5)
*
Gm(n6,n7)
-4
*
%i
*
Gm(n1,n2)
*
Eps(n3,n4,n5,n8)
*
Gm(n6,n7)
+4
*
%i
*
Gm(n1,n3)
*
Eps(n2,n4,n5,n8)
*
Gm(n6,n7)
-4
*
%i
*
Eps(n1,n4,n5,n8)
*
Gm(n2,n3)
*
Gm(n6,n7)
+4
*
%i
*
Eps(n1,n2,n3,n6)
*
Gm(n4,n7)
*
Gm(n5,n8)
-4
*
%i
*
Eps(n1,n2,n3,n7)
*
Gm(n4,n6)
*
Gm(n5,n8)
-4
*
%i
*
Eps(n1,n2,n3,n6)
*
Gm(n4,n8)
*
Gm(n5,n7)
+4
*
%i
*
Eps(n1,n2,n3,n8)
*
Gm(n4,n6)
*
Gm(n5,n7)
-4
*
%i
*
Gm(n1,n2)
*
Gm(n3,n4)
*
Eps(n5,n6,n7,n8)
+4
*
%i
*
Gm(n1,n3)
*
Gm(n2,n4)
*
Eps(n5,n6,n7,n8)
-4
*
%i
*
Gm(n1,n4)
*
Gm(n2,n3)
*
Eps(n5,n6,n7,n8)
+4
*
%i
*
Eps(n1,n2,n3,n7)
*
Gm(n4,n8)
*
Gm(n5,n6)
-4
*
%i
*
Eps(n1,n2,n3,n8)
*
Gm(n4,n7)
*
Gm(n5,n6)
+4
*
%i
*
Gm(n1,n2)
*
Gm(n3,n5)
*
Eps(n4,n6,n7,n8)
-4
*
%i
*
Gm(n1,n3)
*
Gm(n2,n5)
*
Eps(n4,n6,n7,n8)
+4
*
%i
*
Gm(n1,n5)
*
Gm(n2,n3)
*
Eps(n4,n6,n7,n8)
-4
*
%i
*
Gm(n1,n2)
*
Eps(n3,n6,n7,n8)
*
Gm(n4,n5)
+4
*
%i
*
Gm(n1,n3)
*
Eps(n2,n6,n7,n8)
*
Gm(n4,n5)
-4
*
%i
*
Eps(n1,n6,n7,n8)
*
Gm(n2,n3)
*
Gm(n4,n5)
(%i86) length(%);
(%o86) 33
The present package function simp_tr1 (which controls tr1 via the lisp code in simplifying-new.lisp) does
not return an evaluation for the case of an even number equal to ten or more gamma matrices multiplying G5.
(%i2) tr10:tr(G5,n1,n2,n3,n4,n5,n6,n7,n8,n9,n10);
"simp_tr1: present version of tr only returns nonzero gamma5 traces for"
" G5 times a product of 4,6,or 8 gammas"
(%o2) Gamma5trace(G5,n1,n2,n3,n4,n5,n6,n7,n8,n9,n10)
(%i93) Mindex(n11,n12);
(%o93) done
(%i94) indexL;
(%o94) [n1,n2,n3,n4,n5,n6,n7,n8,n9,n10,la,mu,nu,rh,si,ta,al,be,ga,de,ep,n11,
n12]
(%i95) tr11:tr(G5,n1,n2,n3,n4,n5,n6,n7,n8,n9,n10,n11);
(%o95) 0
More compact expressions for gamma 5 traces (compared to the classical methods used here) can be found in the paper:
A new method for calculation of traces of Dirac gamma-matrices
in Minkowski space, by Alexander L. Bondarev,
http://arxiv.org/abs/hep-ph/0504223
9
12.3 Introduction to the Dirac Package, Version 2
Three independent Dirac algebra methods are available in our package. One can gain condence in an unexpected result
by doing the same calculation in multiple ways. The explicit Dirac spinor methods, which use an explicit representation
of the gamma matrices, are bug free, fast, and the route to polarized amplitudes (rather that the square of polarized am-
plitudes). Moreover, summing the absolute square of the polarized amplitudes over all helicity values leads quickly and
reliably to the unpolarized cross section in terms of the frame dependent kinematic variables such as scattering angle.
However the version available with this package is not a covariant method and is always frame dependent. The explicit
Dirac matrix methods using the Maxima function mat_trace are inherently less bug prone and also are faster in exe-
cution than the purely symbolic methods. However, the purely symbolic contraction and trace methods available in this
package are capable of introducing the kinematic invariants s, t, and u into the calculation in a simple way, or to reducing
the types of four vector dot products to two, leading to useful covariant results. In the following examples, we usually
use the fastest and most convenient method for a given problem. An example of using the matrix trace and contraction
methods, and the symbolic trace and contraction methods to generate the square of polarized amplitudes is provided in
the section dealing with the batch le moller3.mac.
Each physics calculation is written in the form of a batch le with an appropriate name, such as moller1.mac. One
can start a problem calculation in two steps by rst loading the Dirac package with load("dirac2.mac") (which
kills all denitions and arrays and resets options to the default) and then, for example, launching the batch le using
batch ("moller1.mac").
The batch les use code in the the launch le dirac2.mac to print out the year-month-day of the batch le run, and
also to print the Maxima version number. This code is:
load("new-timedate.lisp")$
mydate : apply (sconcat, rest (charlist (timedate(+7)),-15))$
_binfo% : "Maxima 5.23.2"$
This code also is in the authors maxima-init le, to allow a print out the year-month-day at the start of a Maxima
session screen, using
disp (mydate)$
as a line in maxima-init.
However the launch le dirac2.mac starts with the lines
kill(all);
reset();
if not lfreeof (rules, eps4rule1) then apply (kill, [eps4] )$
/
*
remove all global arrays currently existing
*
/
if length (arrays) > 0 then apply (remarray, arrays)$
(which is probably overkill), but removes the mydate dependent code in maxima-init also, hence that code again
appears in dirac2.mac.
The lisp le new-timedat.lisp, available on the authors webpage, was written by Maxima developer Leo Butler,
and allows Windows users to overcome a bug to get a year-month-day stamp using Maximas timedate function.
In our example batch les is the line
print (" ver: ",_binfo%," date: ",mydate )$
which prints out the Maxima version number and year-month-day as:
ver: Maxima 5.23.2 date: 2011-02-28
10
12.4 Overview and Examples of the Dirac Package Tools
Loading the Dirac package is accomplished by the command load("dirac2.mac") (or just load(dirac2) if you
have set up your init le as recommended in Chapter 1). The dirac2.mac le itself then loads ve package les:
simplifying-new.lisp, dgtrace2.mac, dgcon2.mac, dgeval2.mac, and dgmatrix2.mac.
This package uses symbols for both purely symbolic entities and also for purely numerical and matrix entities. The most
used (in the twelve example batch les) symbolic route package functions are tr, nc_tr, Con, comp_def, and
noncov.
The le dgfunctions2.txt has an alphabetical listing of the Dirac package functions with the name the program le
in which the code for that function will be found. One should then go to that le and use a name search for the name of
the function. Near the top and/or the bottom of each function code will be notes on syntax, purpose, functions which call
it, and (for the most important functions) some explicit examples of use and output. Once the program is loaded, you can
experiment with calls to any of these functions, using the interactive mode.
Symbols for Lorentz indices need to be declared using Mindex. However, a set of default Lorentz index symbols is de-
ned in dirac2.mac with the command:
Mindex (n1,n2,n3,n4,n5,n6,n7,n8,n9,n10,la,mu,nu,rh,si,ta,al,be,ga,de,ep)$,
and the package list indexL will show you the current Lorentz index symbols recognised. There is a function unindex
which allows you to remove recognised Lorentz index symbols from the list indexL. The Lorentz indices such as n1 or
mu, nu,... take values {0,1,2,3}.
dirac2.mac also assigns two default symbols to be recognised as symbols for particle masses, using the command:
Mmass (m,M)$. The package list massL will show you the current list of recognised mass symbols. If you want to
also use the symbols m1 and m2 for masses, you use the command Mmass (m1,m2); (we show this example below).
The list massL is then updated to contain four recognised mass symbols.
The launch le dirac2.mac also denes a default set of scalars with the command:
Mscalar (c1,c2,c3,c4,c5,c6,c7,c8,c9,c10)$. You can see the current list of recognised scalars in the
package list scalarL. (There is a function unscalar which both removes symbols from the list scalarL and also
removes the scalar property).
An atom (like p1 or a) which is not a recognised Lorentz index symbol, not a recognised mass symbol, and not a recog-
nised declared scalar, is treated as a 4-momentum (or Feynman slashed momentum, depending on the context) symbol.
Purely symbolic entities include the symbolic metric tensor Gm (mu,nu) corresponding to g

(or g

depending
on the context) , the symbolic dot product of four vectors D(p1,p2) corresponding to p
1
p
2
, the symbolic four di-
mensional antisymmetric tensor Eps(mu,nu,rh,si) corresponding to

. (Eps however is not treated as
antisymmetric until conversion to the numerical array eps4[n1,n2,n3,n4] occurs).
A symbolic contravariant four vector component is described with the notation UI(p,mu), corresponding to p

, and
is converted to the array component p[mu] by noncov (see below). The covariant component p

is described by
LI(p,mu). With our metric convention, p
0
= p
0
.
We use G (A,B,C,...) for the symbolic representation of a product of a combination of Dirac gamma matrices and
slashed 4-momentum matrices and use G(1) to symbolically denote the unit 4 x 4 matrix.
The expression G(mu,p,nu,q), for example, represents the purely symbolic product of four matrices

p/

q/, in
which the Feynman slashed matrix p/ = p

(using the summation convention). Using explicit Maxima matrix nota-

tion, this matrix product is Gam[mu] . sL(p) . Gam[nu] . sL(q) in which Gam[mu] is one of the four Dirac
gamma matrices and sL(p) is the Feynman slash matrix p/ formed from the explicit gamma matrices and the explicit
11
4-momentum array p[mu].
A symbolic product which includes the gamma5 matrix
5
, represented symbolically by G5, is G(G5,mu,p1,nu,p2),
for example, and corresponds to: Gam[5] . Gam[mu] . sL(p1) . Gam[nu] . sL(p2). The arrays Gam[mu]
and Gam[5] are 4 x 4 matrices, as is the matrix function sL(p), and they are dened in dgmatrix2.mac, where
you can also nd the denitions of the Dirac particle spinor (4-component column vector) UU, and the Dirac anti-particle
spinor (also a 4-component column vector) VV, (both dened using Maximas matrix function) and having a concrete
form which is generated from the specied 4-momentum and helicity of the particle or anti-particle. Many examples of
explicit matrix methods (based on dgmatrix2.mac code) will be found in the twelve batch le problems worked out.
The le dirac2.mac includes some denitions of useful utility functions, and also denes the explicit numerical metric
tensor array gmet[mu,nu] and the explicit four dimensional completely antisymmetric tensor array
eps4[mu,nu,rh,si] components (and antisymmetry property). (The Lorentz indices mu, nu,... take values 0,1,2,3.)
dirac2.mac also declares that the symbolic four vector dot product D(p,q) is symmetric, as is the symbolic metric
tensor Gm(mu,nu), as an aid in automatic simplication of expressions.
The function noncov converts Eps(mu,nu,rh,si) to eps4[mu,nu,rh,si] (with the convention that
0123
= 1),
converts Gm(mu,nu) to gmet[mu,nu], converts UI(p,mu) to p[mu], converts LI(p,mu) to
sum (gmet[mu,nu]
*
p[nu],nu,0,3). For example LI (p,0) --> p[0], LI (p,1) --> -p[1]. The
function noncov also converts D(p,q) into sum (sum (gmet[mu,nu]
*
p[mu]
*
q[nu],nu,0,3),mu,0,3),
which reduces to the single sum sum (gmet[mu,mu]
*
p[mu]
*
q[mu],mu,0,3), which becomes the explicit sum
p[0]
*
q[0] - p[1]
*
q[1] - p[2]
*
q[2] - p[3]
*
q[3]. The function noncov does nothing to G(...).
The reference frame dependent function comp_def (p(Ep,px,py,pz)) creates an array with the name p whose
elements can be viewed using listarray (p) and whose elements can be accessed by using p[0] to get the value
Ep, using p[1] to get the value px, etc.
The symbolic trace of a product of gamma matrices is usually carried out using the tr function. Thus the syntax
tr(mu,nu,rh,si) carries out the trace of a product of four Dirac gamma matrices, and the same calculation can
be carried out using explicit Maxima matrices using the syntax:
mat_trace (Gam[mu] . Gam[nu] . Gam[rh] . Gam[si]);
(although one must supply denite values for the Lorentz index symbols mu, nu, rh, si to get a simple result
from the explicit matrix method ).
A simpler-to-use function for the trace of explicit matrix expressions is m_tr. For example, m_tr (mu,p,q,nu) uses
the exact same syntax as tr but internally translates this into
mat_trace (Gam[mu] . sL(p) . sL(q) . Gam[nu]).
To obtain useful expressions from the explicit matrix methods, one should rst use comp_def to dene the frame de-
pendent components of relevant 4-momenta, then use either the m_tr or mat_trace syntax for the trace, and then use
Con (or mcon) for contraction.
A symbolic trace involving a gamma5 matrix, such as tr (G5,p,nu,rh,si) generates expressions proportional
to (for example) LI(p,N1)
*
Eps (N1,nu,rh,si), in which N1 is an unsummed dummy index created by the
code. (The sum over that dummy index is done by the later use of noncov.) The global symbol Nlast will al-
ways tell you what the last dummy symbol created is. The symbolic tr code creates these dummy indices in the order
N1,N2,... and establishes indexp and dummyp attributes, so that, for example, indexp (N1) --> true and
dummyp (N1) --> true.
An expression or sum of terms involving G(a,b,c,..)s can be converted into an expression (or sum of such) in
which each G(...) is replace by tr (...) by using the symbolic function Gtr.
Thus Gtr (G(a,b,c,d)) --> tr (a,b,c,d).
12
The symbolic trace function tr automatically contracts on repeated Lorentz indices, such as in tr (mu,p,q,mu),
before using the actual trace algorithms.
Symbolic contraction (without the trace) on repeated indices in a symbolic product of gamma matrices can be done using
Con, which recognises a symbolic case and then calls scon. Examples in which scon (s for symbolic) is called by
Con are: Con (G (mu,p,q,mu)) (equivalent to scon(G(mu,p,q,mu)))
and
Con (G (mu,G5,mu)) (equivalent to scon(G(mu,G5,mu))).
Con (and scon) automatically contract on all repeated Lorentz indices unless the user supplies the desired contraction
indices.
For example, Con (G(mu,nu,p,q,nu,mu),mu,nu) (or the same with Con --> scon) will carry out symbolic
contraction on both mu and nu and produce the same answer as Con (G(mu,nu,p,q,nu,mu)).
But Con (G(mu,nu,p,q,nu,mu),nu) will only contract on nu.
When using explicit Dirac gamma matrices and Maxima matrix methods, you must always supply the desired contrac-
tion indices to Con (which will recognise the explicit matrix case and call mcon with the desired contraction indices).
For example,
Con (mat_trace (Gam[mu] . sL(p) . sL(q) . Gam[mu]), mu) or the same with Con --> mcon .
A frequent case met with in the twelve examples is the symbolic contraction of a product of symbolic traces each produced
by tr. The most efcient symbolic method is to rst convert each of the symbolic trace expressions into expressions de-
pending on eps4, gmet, and 4-momentum components like p[1] by using noncov before the contraction. This is
such a common approach, that the package function nc_tr is dened to immediately apply noncov to each trace result
returned from the function tr1 (controlled by simp_tr1 via simplifying-new.lisp) called by TR1, which is
called by tr.
One then needs the contraction of products of such expressions, which is provided by using either the contraction
function econ (e for explicit) or (usually quicker in operation) mcon. You could simply use Con (which should
recognise the post-noncov case and call econ with the supplied contraction indices). As in the explicit matrix case, you
must supply the desired contraction indices (in this third case) to accomplish the desired contraction.
In the following introductory examples, carried out in an interactive context, you will see examples of all of the above for
both traces and contractions.
13
12.4.1 Some Basic Package Symbols
In the initial examples, no values have been assigned to the 4-momenta array components referred to (this will eventu-
ally be done using the package function comp_def), so we will see things like p[1] for example. The last line of
dirac2.mac is display2d:false$, which allows more information per screen, especially when displaying matri-
ces.
(%i1) load(dirac2);
dirac2.mac
simplifying-new.lisp
dgtrace2.mac
dgcon2.mac
dgeval2.mac
dgmatrix2.mac
scalarL = [c1, c2, c3, c4, c5, c6, c7, c8, c9, c10]
indexL = [n1, n2, n3, n4, n5, n6, n7, n8, n9, n10, la, mu, nu, rh, si, ta, al,
be, ga, de, ep]
massL = [m, M]
Nlast = 0
reserved program capital letter name use:
Chi, Con, D, Eps, G, G(1), G5, G5p, Gam, Gm, Gtr, LI, Nlast, UI, P, S, Sig
UU, VP, VV, I2, Z2, CZ2, I4, Z4, CZ4, RZ4, N1,N2,...
reserved array names: gmet, eps4
(%o1) "c:/work5/dirac2.mac"
(%i2) noncov(Eps(mu,nu,rh,si));
(%o2) eps4[mu,nu,rh,si]
(%i3) noncov (Eps(mu,nu,al,be));
(%o3) eps4[al,be,mu,nu]
(%i4) Eps(0,1,2,3);
(%o4) Eps(0,1,2,3)
(%i5) eps4 [0,1,2,3];
(%o5) 1
(%i6) eps4 [1,0,2,3];
(%o6) -1
(%i7) noncov (Gm (mu,nu));
(%o7) gmet[mu,nu]
(%i8) Gm (0,0);
(%o8) Gm(0,0)
(%i9) gmet[0,0];
(%o9) 1
(%i10) gmet[1,1];
(%o10) -1
(%i11) Gm (be,al);
(%o11) Gm(al,be)
(%i12) noncov (D(p,q));
(%o12) -p[3]
*
q[3]-p[2]
*
q[2]-p[1]
*
q[1]+p[0]
*
q[0]
(%i13) D (q,p);
(%o13) D(p,q)
(%i14) D(b,a);
(%o14) D(a,b)
(%i15) noncov (UI(p,0));
(%o15) p[0]
(%i16) noncov (UI(p,1));
(%o16) p[1]
(%i17) noncov (LI (p,0));
(%o17) p[0]
(%i18) noncov (LI (p,1));
(%o18) -p[1]
(%i19) massL;
(%o19) [m,M]
(%i20) Mmass (m1,m2);
(%o20) done
14
(%i21) massL;
(%o21) [m,M,m1,m2]
(%i22) map (massp,%);
(%o22) [true,true,true,true]
(%i23) indexL;
(%o23) [n1,n2,n3,n4,n5,n6,n7,n8,n9,n10,la,mu,nu,rh,si,ta,al,be,ga,de,ep]
(%i24) map (indexp,%);
(%o24) [true,true,true,true,true,true,true,true,true,true,true,true,true,true,
true,true,true,true,true,true,true]
(%i25) scalarL;
(%o25) [c1,c2,c3,c4,c5,c6,c7,c8,c9,c10]
(%i26) map (scalarp,%);
(%o26) [true,true,true,true,true,true,true,true,true,true]
(%i27) noncov (G (mu,p1,nu,p2) );
(%o27) G(mu,p1,nu,p2)
12.4.2 Symbolic Expansions of D(a,b) using Dexpand
The package uses D(a,b) to represent the symbolic four-vector dot product of two four vectors a and b. The package
function Dexpand will expand arguments of D which consist of two or more terms, and will do a mass expansion when
declared mass symbols (such as the default m and M symbols) are found, and will also pull out scalars declared using
Mscalar (the default set of scalars are called c1, c2, ...,c10).
Here are some interactive examples which use the package function Dexpand.
(%i2) Dexpand(D(a,b));
(%o2) D(a,b)
(%i3) Dexpand(D(a,b+c));
(%o3) D(a,c)+D(a,b)
(%i4) Dexpand(D(a,b+m));
(%o4) D(a)
*
m+D(a,b)
(%i5) Dexpand(D(a+m,b+M));
(%o5) m
*
M+D(a)
*
M+D(b)
*
m+D(a,b)
(%i6) Dexpand(D(2
*
a+m,b/3+M));
(%o6) m
*
M+2
*
D(a)
*
M+D(b)
*
m/3+2
*
D(a,b)/3
(%i7) Dexpand(D(2
*
c1
*
a+m,b/3+M));
(%o7) m
*
M+2
*
c1
*
D(a)
*
M+D(b)
*
m/3+2
*
c1
*
D(a,b)/3
(%i8) Dexpand ( D(a,b+c) + D(e,f+g) );
(%o8) D(e,g)+D(e,f)+D(a,c)+D(a,b)
(%i9) Dexpand ( r
*
D(a,b+c)/D(e,f) );
(%o9) D(a,c)
*
r/D(e,f)+D(a,b)
*
r/D(e,f)
(%i10) Dexpand ( r
*
D(a,b+c)/(s
*
D(e,f+g)) );
(%o10) D(a,c)
*
r/(D(e,g)
*
s+D(e,f)
*
s)+D(a,b)
*
r/(D(e,g)
*
s+D(e,f)
*
s)
12.4.3 Some Examples of the Symbolic Expansion of G(a,b,c,...) Using Gexpand
The package uses G(a,b,c,d) to represent the product of four Dirac matrices, in which the symbols a,b,c,d repre-
sent Feynman slashed four-vectors unless they have been declared index symbols using Mindex. You can see the current
list of index symbols available by looking at the contents of the list indexL.
(%i11) indexL;
(%o11) [n1,n2,n3,n4,n5,n6,n7,n8,n9,n10,la,mu,nu,rh,si,ta,al,be,ga,de,ep]
With this (default) list of recognised Lorentz indices, the expression G(mu,a,nu,b) would represent the matrix product
Gam[mu] . sL(a) . Gam[nu] . sL(b).
The package function Gexpand will expand symbolic products of Dirac gamma matrices, expanding arguments with
multiple terms, pulling out numbers and declared scalars, and making mass expansions when nding a term which is a
15
recognised mass symbol.
An additional expansion occurs (using Gexpand) if a slot contains the symbol S(sv) or S(sv,Sp) which are used to
symbolically represent heliciy projection matrices (the simpler rst case for particles treated as massless, and the more
complicated two argument version for the more general case of massive particles).
The symbol sv is a standin of the helicity quantum number taken to have values sv = 1 or sv = -1. The symbol Sp
represents the positive helicity spin 4-vector implied by a particles 4-momentum vector. Many examples of the use of
helicity projection operator methods will be shown in the reaction examples. When Gexpand encounters such a helicity
projection operator symbol an expansion is made using the symbolic gamma5 matrix G5.
(%i12) Gexpand (G (S(sv)));
(%o12) sv
*
G(G5)/2+G(1)/2
(%i13) Gexpand (G (S(sv,Sp)));
(%o13) sv
*
G(G5,Sp)/2+G(1)/2
Note especially that the symbol G(1) represents symbolically the unit 4 x 4 matrix, whose trace is 4.
The package trace functions tr, Gtr, and nc_tr all use the function Gexpand as the rst step in calculating a trace of
a product of Dirac gamma matrices symbolically.
Here are some examples of the use of Gexpand.
(%i14) Gexpand (G());
(%o14) G()
(%i15) Gexpand (G(1));
(%o15) G(1)
(%i16) Gexpand (G(-1));
(%o16) -G(1)
(%i17) Gexpand (G(-a));
(%o17) -G(a)
(%i18) Gexpand (G(m));
(%o18) G(1)
*
m
(%i19) Gexpand (G(a + m));
(%o19) G(1)
*
m+G(a)
(%i20) Gtr (%);
(%o20) 4
*
m
(%i21) Gexpand (G(a,b+c));
(%o21) G(a,c)+G(a,b)
(%i22) Gtr (%);
(%o22) 4
*
D(a,c)+4
*
D(a,b)
(%i23) Gexpand (G (a+m,b+M));
(%o23) G(1)
*
m
*
M+G(a)
*
M+G(b)
*
m+G(a,b)
(%i24) Gtr (%);
(%o24) 4
*
m
*
M+4
*
D(a,b)
(%i25) Gexpand (G (S(1),a,S(1),b));
(%o25) G(G5,a,G5,b)/4+G(G5,a,b)/4+G(a,G5,b)/4+G(a,b)/4
(%i26) Gexpand (G (2
*
c1
*
a,3
*
c2
*
b - c3
*
c/7));
(%o26) 6
*
c1
*
c2
*
G(a,b)-2
*
c1
*
c3
*
G(a,c)/7
(%i27) Gtr(%);
(%o27) 24
*
c1
*
c2
*
D(a,b)-8
*
c1
*
c3
*
D(a,c)/7
16
12.4.4 Symbolic and Explicit Matrix Trace Examples: tr, Gtr, mat trace, m tr
Some Symbolic Trace Examples
We next display some examples of symbolic traces.
The symbol G5 symbolically represents
5
.
The symbol Gm(mu,nu) whose arguments are both recognised Lorentz index symbols, represents symbolically either
g

or g

, depending on the context.

The symbol UI(p,mu) (upper index) symbolically represents p

.
The symbol LI(p,mu) (lower index) symbolically represents p

.
The symbol Eps (la,mu,nu,rh) represents symbolically

.
(%i2) tr(1);
(%o2) 4
(%i3) Gtr (G(1));
(%o3) 4
(%i4) tr(mu);
(%o4) 0
(%i5) tr(mu,nu);
(%o5) 4
*
Gm(mu,nu)
(%i6) Gtr (G (mu,nu));
(%o6) 4
*
Gm(mu,nu)
(%i7) tr(p,q);
(%o7) 4
*
D(p,q)
(%i8) tr(p,mu);
(%o8) 4
*
UI(p,mu)
(%i9) noncov(%);
(%o9) 4
*
p[mu]
(%i10) tr(mu,p);
(%o10) 4
*
UI(p,mu)
(%i11) tr(G5);
(%o11) 0
(%i12) tr(G5,mu);
(%o12) 0
(%i13) tr (G5,mu,nu);
(%o13) 0
(%i14) tr (G5,mu,nu,rh);
(%o14) 0
(%i15) tr (G5,al,be,la,mu);
(%o15) -4
*
%i
*
Eps(al,be,la,mu)
(%i16) noncov(%);
(%o16) -4
*
%i
*
eps4[al,be,la,mu]
(%i17) tr(la,mu,nu,rh);
(%o17) 4
*
Gm(la,mu)
*
Gm(nu,rh)-4
*
Gm(la,nu)
*
Gm(mu,rh)+4
*
Gm(la,rh)
*
Gm(mu,nu)
(%i18) tr (a,b,c,d);
(%o18) 4
*
D(a,b)
*
D(c,d)-4
*
D(a,c)
*
D(b,d)+4
*
D(a,d)
*
D(b,c)
(%i19) tr(G5,p,be,la,mu);
(%o19) -4
*
%i
*
LI(p,N1)
*
Eps(N1,be,la,mu)
(%i136) tr(G5,mu,nu,rh,al);
(%o136) -4
*
%i
*
Eps(mu,nu,rh,al)
(%i137) tr(mu,G5,nu,rh,al);
(%o137) 4
*
%i
*
Eps(mu,nu,rh,al)
(%i138) tr(mu,nu,G5,rh,al);
(%o138) -4
*
%i
*
Eps(mu,nu,rh,al)
(%i139) tr(mu,nu,rh,G5,al);
(%o139) 4
*
%i
*
Eps(mu,nu,rh,al)
(%i140) tr(mu,nu,rh,al,G5);
(%o140) -4
*
%i
*
Eps(mu,nu,rh,al)
17
Some Simple Explicit Matrix Trace Examples
Next are some very simple examples of the use of explicit matrix trace methods. I4 is an explicit unit matrix in 4
dimensions, and Z4 is an explicit zero matrix in 4 dimensions.
(%i21) I4;
(%o21) matrix([1,0,0,0],[0,1,0,0],[0,0,1,0],[0,0,0,1])
(%i22) mat_trace(%);
(%o22) 4
(%i23) Z4;
(%o23) matrix([0,0,0,0],[0,0,0,0],[0,0,0,0],[0,0,0,0])
(%i24) mat_trace (%);
(%o24) 0
(%i25) a(mu,nu):=mat_trace(Gam[mu] . Gam[nu])$
(%i26) b(mu,nu):=4
*
gmet[mu,nu]$
(%i27) c(mu,nu):=is(equal(a(mu,nu),b(mu,nu)))$
(%i28) c(0,0);
(%o28) true
(%i30) for mu : 0 thru 3 do
for nu : 0 thru 3 do
if not c(mu,nu) then print(mu,nu," false")$
In the last step, we have shown that a(mu,nu) = b(mu,nu) for all 16 mu,nu pairs, since there were no false
messages printed to the screen.
More Symbolic Trace Examples
Returning to the symbolic trace with tr, recognised mass symbols get special treatment.
(%i31) tr(m+p1,mu,m+p2,mu);
(%o31) 16
*
m2-8
*
D(p1,p2)
(%i32) tr(m,m);
(%o32) 4
*
m2
(%i33) tr(m,M);
(%o33) 4
*
m
*
M
The arguments of tr(m+p1,mu,m+p2,mu) are examined to look for mass terms, and if any are found those argu-
ments are expanded and the mass symbols are factored out of the tr function. An argument like p + m in tr is treated
like the matrix sumsL(p) + m
*
I4 and a parallel treatment using explicit matrix methods needs to have the mass sym-
bol multiplied by I4.
The last two simple examples can be reproduced using mat_trace, but we have to be careful to use parentheses around
the quantity m
*
I4, since the Maxima function mat_trace does not know anything about special mass symbols.
(%i35) mat_trace((m
*
I4) . (m
*
I4));
(%o35) 4
*
m2
Likewise, numbers and recognised scalars which multiply and/or divide momentum symbols are factored out of tr
arguments.
(%i36) scalarL;
(%o36) [c1,c2,c3,c4,c5,c6,c7,c8,c9,c10]
(%i37) tr(c1
*
p,q);
(%o37) 4
*
c1
*
D(p,q)
(%i38) tr(-c1
*
p,q/c2);
(%o38) -4
*
c1
*
D(p,q)/c2
(%i39) tr(-c1
*
%i
*
p,q/c2);
(%o39) -4
*
%i
*
c1
*
D(p,q)/c2
(%i40) tr(-7
*
c1
*
%i
*
p,q/c2/10);
(%o40) -14
*
%i
*
c1
*
D(p,q)/(5
*
c2)
18
Factoring out scalars, etc., occurs after multiple term arguments are rst expanded.
(%i41) tr(a,b);
(%o41) 4
*
D(a,b)
(%i42) tr(a,b+c);
(%o42) 4
*
D(a,c)+4
*
D(a,b)
(%i43) tr(a,c1
*
b+c2
*
c+c3
*
d);
(%o43) 4
*
c3
*
D(a,d)+4
*
c2
*
D(a,c)+4
*
c1
*
D(a,b)
12.4.5 Symbolic and Explicit Matrix Contraction Examples Using Con
Examples of Symbolic Contraction of Repeated Lorentz Indices
We next show some examples of contractions on repeated Lorentz indices. Recall that G(1) symbolically represents the
unit 4 x 4 matrix and G5 represents the matrix
5
.
(%i4) Con (G (mu,p,q,mu));
(%o4) 4
*
G(1)
*
D(p,q)
(%i5) Gtr (%);
(%o5) 16
*
D(p,q)
(%i6) tr (mu,p,q,mu);
(%o6) 16
*
D(p,q)
(%i7) Con (G (mu,G5,mu));
(%o7) -4
*
G(G5)
(%i8) Con (G(n1,p,n2,n1),n1);
(%o8) 4
*
G(1)
*
UI(p,n2)
(%i9) Con (G(n1,p,n2,n1));
(%o9) 4
*
G(1)
*
UI(p,n2)
(%i10) Con (G(n1,n2,n2,n1),n1);
(%o10) 4
*
G(1)
*
Gm(n2,n2)
(%i11) Con (%);
(%o11) 16
*
G(1)
(%i12) Gtr (%);
(%o12) 64
(%i13) Con (UI(p,mu)
*
UI(q,mu));
(%o13) D(p,q)
(%i14) Con (UI(p,mu)
*
Gm(mu,nu));
(%o14) UI(p,nu)
(%i15) Con (G(a,mu,b,c,mu,a));
(%o15) 4
*
G(1)
*
D(a,a)
*
D(b,c)
(%i16) Con (G(n1,p+m,q+M,n1),n1);
(%o16) -2
*
G(p)
*
M+4
*
G(1)
*
m
*
M-2
*
m
*
G(q)+4
*
G(1)
*
D(p,q)
(%i17) Gtr (%);
(%o17) 16
*
m
*
M+16
*
D(p,q)
(%i18) tr(n1,p+m,q+M,n1);
(%o18) 16
*
m
*
M+16
*
D(p,q)
(%i19) Con (G(n1,2
*
a,b/3,n1),n1);
(%o19) 8
*
G(1)
*
D(a,b)/3
(%i20) Con (G(n5,n4,n3,n2,n1,n5,n1,n4,n3,n2));
(%o20) 256
*
G(1)
(%i21) Gtr (%);
(%o21) 1024
(%i22) tr (n5,n4,n3,n2,n1,n5,n1,n4,n3,n2);
(%o22) 1024
(%i25) Con (G (G5,mu,nu,nu,mu));
(%o25) 16
*
G(G5)
(%i2) scon (Gm (n1,n2)
*
Eps (n1,n3,n4,n5));
(%o2) Eps(n2,n3,n4,n5)
The last example shows a contraction between an index which appears in a Gm and a multiplying Eps. scon calls
get_rind to nd repeated indices and then simp_scon1 to automatically contract on repeated indices, which it will
19
do for a product of Gm(..) and Eps(..). looks for repeated indices.
If we try the same contraction with Con we get
(%i13) Con (Gm (n1,n2)
*
Eps (n1,n3,n4,n5));
(%o13) gmet[n1,n2]
*
eps4[n1,n3,n4,n5]
which occurs because Con calls Eps_facp, (which returns true), thus detecting the presence of an Eps factor in the
expression, and as a consequence, calls noncov immediately, followed by a call to econ. noncov converts Gm( ) to
gmet[ ] and Eps( ) to eps4[ ]. econ will not automatically contract; you must supply the contraction index.
(%i14) Con (Gm (n1,n2)
*
Eps (n1,n3,n4,n5),n1);
(%o14) eps4[n2,n3,n4,n5]
Note that the present version of scon will not simplify a product of two Eps factors.
(%i17) scon(Eps(n1,n2,n3,n4)
*
Eps(n1,n5,n6,n7));
(%o17) Eps(n1,n2,n3,n4)
*
Eps(n1,n5,n6,n7)
(%i18) scon(Eps(n1,n2,n3,n4)
*
Eps(n1,n5,n6,n7),n1);
(%o18) Eps(n1,n2,n3,n4)
*
Eps(n1,n5,n6,n7)
However, you can use Con (supplying the contraction index as well) which again will detect the Eps factors, then will
call noncov and econ to do the summatation over the contraction index.
(%i16) Con(Eps(n1,n2,n3,n4)
*
Eps(n1,n5,n6,n7),n1);
(%o16) -eps4[3,n2,n3,n4]
*
eps4[3,n5,n6,n7]-eps4[2,n2,n3,n4]
*
eps4[2,n5,n6,n7]
-eps4[1,n2,n3,n4]
*
eps4[1,n5,n6,n7]
+eps4[0,n2,n3,n4]
*
eps4[0,n5,n6,n7]
Alternatively, you can use noncov and then econ:
(%i19) noncov (Eps(n1,n2,n3,n4)
*
Eps(n1,n5,n6,n7));
(%o19) eps4[n1,n2,n3,n4]
*
eps4[n1,n5,n6,n7]
(%i20) econ (%,n1);
(%o20) -eps4[3,n2,n3,n4]
*
eps4[3,n5,n6,n7]-eps4[2,n2,n3,n4]
*
eps4[2,n5,n6,n7]
-eps4[1,n2,n3,n4]
*
eps4[1,n5,n6,n7]
+eps4[0,n2,n3,n4]
*
eps4[0,n5,n6,n7]
The present version of the Dirac package does not simplify the implied contraction presented by
Con (Eps (n1,n2,n3,n4)
*
Eps (n1,n5,n6,n7),n1). Using the corresponding expression with Eps(...)
replaced by eps4[...] and then using econ (or Con) will return a contracted expression in terms of eps4s.
(%i34) Con (Eps (n1,n2,n3,n4)
*
Eps (n1,n5,n6,n7));
(%o34) Eps(n1,n2,n3,n4)
*
Eps(n1,n5,n6,n7)
(%i37) Con (Eps (n1,n2,n3,n4)
*
Eps (n1,n5,n6,n7),n1);
(%o37) Eps(n1,n2,n3,n4)
*
Eps(n1,n5,n6,n7)
(%i35) noncov(%);
(%o35) eps4[n1,n2,n3,n4]
*
eps4[n1,n5,n6,n7]
(%i36) econ(%,n1);
(%o36) -eps4[3,n2,n3,n4]
*
eps4[3,n5,n6,n7]-eps4[2,n2,n3,n4]
*
eps4[2,n5,n6,n7]
-eps4[1,n2,n3,n4]
*
eps4[1,n5,n6,n7]
+eps4[0,n2,n3,n4]
*
eps4[0,n5,n6,n7]
(%i38) Con (eps4[n1,n2,n3,n4]
*
eps4[n1,n5,n6,n7],n1);
(%o38) -eps4[3,n2,n3,n4]
*
eps4[3,n5,n6,n7]-eps4[2,n2,n3,n4]
*
eps4[2,n5,n6,n7]
-eps4[1,n2,n3,n4]
*
eps4[1,n5,n6,n7]
+eps4[0,n2,n3,n4]
*
eps4[0,n5,n6,n7]
20
Examples of Explicit Matrix Contractions on Repeated Lorentz Indices Using Con
(%i26) Con (Gam[n1] . Gam[n1],n1);
(%o26) matrix([4,0,0,0],[0,4,0,0],[0,0,4,0],[0,0,0,4])
(%i27) is (equal (%,4
*
I4));
(%o27) true
(%i28) Con (G(n1,n1));
(%o28) 4
*
G(1)
(%i29) Con (Gam[mu] . Gam[5] . Gam[mu],mu);
(%o29) matrix([4,0,0,0],[0,4,0,0],[0,0,-4,0],[0,0,0,-4])
(%i30) is (equal (%,-4
*
Gam[5]));
(%o30) true
(%i32) Con (G(mu,G5,mu));
(%o32) -4
*
G(G5)
(%i23) Con(Gam[5] . Gam[mu] . Gam[nu] . Gam[nu] . Gam[mu],mu,nu);
(%o23) matrix([-16,0,0,0],[0,-16,0,0],[0,0,16,0],[0,0,0,16])
(%i24) is (equal (%,16
*
Gam[5]));
(%o24) true
12.4.6 Examples of noncov, comp def, VP, and econ
(%i2) noncov(D(p,q));
(%o2) -p[3]
*
q[3]-p[2]
*
q[2]-p[1]
*
q[1]+p[0]
*
q[0]
(%i3) VP (p,q);
(%o3) -p[3]
*
q[3]-p[2]
*
q[2]-p[1]
*
q[1]+p[0]
*
q[0]
(%i4) noncov(Gm(mu,nu));
(%o4) gmet[mu,nu]
(%i5) gmet[0,0];
(%o5) 1
(%i6) gmet[1,1];
(%o6) -1
(%i7) noncov(UI(p,mu));
(%o7) p[mu]
(%i8) noncov(LI(p,mu));
(%o8) p[3]
*
gmet[3,mu]+p[2]
*
gmet[2,mu]+p[1]
*
gmet[1,mu]+p[0]
*
gmet[0,mu]
(%i9) noncov(LI(p,0));
(%o9) p[0]
(%i10) noncov(LI(p,1));
(%o10) -p[1]
(%i11) noncov(Eps(mu,nu,rh,la));
(%o11) -eps4[la,mu,nu,rh]
(%i12) eps4[0,1,2,3];
(%o12) 1
(%i13) Eps(0,1,2,3);
(%o13) Eps(0,1,2,3)
(%i14) eps4[1,0,2,3];
(%o14) -1
PRODUCTS OF eps4s WITH econ AND noncov: COMPARISONS
To get summation of dummy indices N1,N2,.. etc from noncov, the code for sum_eps1 automatically recognises
that these symbols represent dummy summation variables.
(%i11) map (dummyp,[N1,N2,N3,N4,N5]);
(%o11) [true,true,true,true,true]
The use of noncov will automatically sum over all dummy indices.
(%i12) noncov(Eps(N1,N2,N3,N4)
*
Eps(N1,N2,N3,N4));
(%o12) 24
21
However, the above result is NOT what is meant by contraction. True contraction of indices includes both lowering one
of each pair of indices and then summing, and this operation is correctly carried out by econ.
(%i13) econ(eps4[n1,n2,n3,n4]
*
eps4[n1,n2,n3,n4],n1,n2,n3,n4);
(%o13) -24
Here are a few more comparisons between noncov and econ in this eps4 product context:
(%i14) noncov(Eps(N1,N2,N3,0)
*
Eps(N1,N2,N3,0));
(%o14) 6
(%i15) econ(eps4[n1,n2,n3,0]
*
eps4[n1,n2,n3,0],n1,n2,n3);
(%o15) -6
(%i16) noncov(Eps(N1,N2,N3,0)
*
Eps(N1,N2,N3,1));
(%o16) 0
(%i17) econ(eps4[n1,n2,n3,0]
*
eps4[n1,n2,n3,1],n1,n2,n3);
(%o17) 0
(%i18) noncov(Eps(N1,N2,0,1)
*
Eps(N1,N2,0,1));
(%o18) 2
(%i19) econ(eps4[n1,n2,0,1]
*
eps4[n1,n2,0,1],n1,n2);
(%o19) 2
EXAMPLE OF comp_def WITH 2 --> 2 SCATTERING KINEMATICS
Assume high energy scattering in the z-x plane in the center of momentum frame. Then the masses can be ignored, and
the magnitude of the 3-momentum is the same as the relativistic energy (outside the range of the force eld felt by each
particle). Remember that we are using natural units in which momentum, energy and mass all have the same units. The
incident particles move along the positive and negative z axis and the particle initially moving along the positive z axis
is scattered into the direction in the z-x plane which makes the angle th radians with the z axis.
The momenta of the incoming particles are (p1, p2) and the momenta of the outgoing particles are (p3, p4). We
use comp_def to dene relativistic energy and cartesian components of 3-momentumwith the syntax p(E,px,py,pz).
The kinematics are described in terms of the symbols E, th.
(%i15) assume(E > 0,th >= 0,th <= %pi)$
(%i16) comp_def(p1(E,0,0,E),p2(E,0,0,-E),p3(E,E
*
sin(th),0,E
*
cos(th)),
p4(E,-E
*
sin(th),0,-E
*
cos(th)))$
(%i17) map (listarray, [p1,p2,p3,p4]);
(%o17) [[E,0,0,E],[E,0,0,-E],[E,sin(th)
*
E,0,cos(th)
*
E],
[E,-sin(th)
*
E,0,-cos(th)
*
E]]
(%i18) p1[0];
(%o18) E
(%i19) p3[1];
(%o19) sin(th)
*
E
(%i20) noncov(D(p1,p2));
(%o20) 2
*
E2
(%i21) noncov(D(p2+p1,p2+p1));
(%o21) 4
*
E2
(%i22) VP(p2+p1,p2+p1);
(%o22) 4
*
E2
(%i23) factor(VP(p1-p3,p1-p3));
(%o23) 2
*
(cos(th)-1)
*
E2
22
12.4.7 Contraction of a Product of Traces: Con, scon, mcon, econ
In the example of the high energy scattering of electrons in the center of momentum frame, worked out in the batch le
moller1.mac later, we need the contraction of the product of two traces, followed by a reduction to the kinematic
variables.
We will obtain a correct answer three different ways here.
METHOD1: We rst nd the symbolic traces, and then the symbolic contractions, followed by use of noncov. We call
this method1.
(%i1) load(dirac2);
[launch data edited out here]
(%o1) "c:/work5/dirac2.mac"
(%i2) assume(E > 0,th >= 0,th <= %pi)$
(%i3) comp_def(p1(E,0,0,E),p2(E,0,0,-E),p3(E,E
*
sin(th),0,E
*
cos(th)),
p4(E,-E
*
sin(th),0,-E
*
cos(th)))$
(%i4) tr1 : tr(p3,mu,p1,nu);
(%o4) 4
*
UI(p1,mu)
*
UI(p3,nu)+4
*
UI(p1,nu)
*
UI(p3,mu)-4
*
Gm(mu,nu)
*
D(p1,p3)
(%i5) tr2 : tr(p4,mu,p2,nu);
(%o5) 4
*
UI(p2,mu)
*
UI(p4,nu)+4
*
UI(p2,nu)
*
UI(p4,mu)-4
*
Gm(mu,nu)
*
D(p2,p4)
(%i6) method1a : scon (tr1
*
tr2);
(%o6) 32
*
D(p1,p2)
*
D(p3,p4)+32
*
D(p1,p4)
*
D(p2,p3)
(%i7) method1b : Con (tr1
*
tr2);
(%o7) 32
*
D(p1,p2)
*
D(p3,p4)+32
*
D(p1,p4)
*
D(p2,p3)
(%i8) method1c : Con (tr1
*
tr2,mu,nu);
(%o8) 32
*
D(p1,p2)
*
D(p3,p4)+32
*
D(p1,p4)
*
D(p2,p3)
(%i9) method1 : noncov (method1a);
(%o9) 32
*
cos(th)2
*
E4+64
*
cos(th)
*
E4+160
*
E4
METHOD2: We use nc_tr. This is a special function which is not just tr followed by noncov, but rather as TR1
generates the trace of each term of the initial argument expansion, the function noncov is immediately applied to the
trace result, and the output is accumulated in a sum which is eventually returned by nc_tr.
The contraction of uncontracted repeated Lorentz indices can then be effected using mcon, or econ (which Con should
call automatically by recognising the typical after-effects of noncov).
We will call this method2. We will see that method2 produces an apparently different trigonometric result, but use of
trigsimp, followed by expand, results nally in the same answer as found for method1.
(%i10) nctr1 : nc_tr(p3,mu,p1,nu);
(%o10) 4
*
gmet[mu,nu]
*
cos(th)
*
E2-4
*
gmet[mu,nu]
*
E2+4
*
p1[mu]
*
p3[nu]
+4
*
p3[mu]
*
p1[nu]
(%i11) nctr2 : nc_tr(p4,mu,p2,nu);
(%o11) 4
*
gmet[mu,nu]
*
cos(th)
*
E2-4
*
gmet[mu,nu]
*
E2+4
*
p2[mu]
*
p4[nu]
+4
*
p4[mu]
*
p2[nu]
(%i12) method2a : mcon (nctr1
*
nctr2,mu,nu);
(%o12) 64
*
sin(th)2
*
E4+96
*
cos(th)2
*
E4+64
*
cos(th)
*
E4+96
*
E4
(%i13) method2b : econ (nctr1
*
nctr2,mu,nu);
(%o13) 64
*
sin(th)2
*
E4+96
*
cos(th)2
*
E4+64
*
cos(th)
*
E4+96
*
E4
(%i14) method2c : Con (nctr1
*
nctr2,mu,nu);
(%o14) 64
*
sin(th)2
*
E4+96
*
cos(th)2
*
E4+64
*
cos(th)
*
E4+96
*
E4
(%i15) method2 : expand (trigsimp (method2a));
(%o15) 32
*
cos(th)2
*
E4+64
*
cos(th)
*
E4+160
*
E4
(%i16) method2 - method1;
(%o16) 0
We now obtain the same answer using the explicit matrix route, calling this method3. In this case, Con will call mcon
for the contraction, and again, we must supply the desired contraction indices. Using the more compact (but still explicit
matrix) syntax of m_tr, we call the result method3a:
23
(%i17) method3a : Con(m_tr(p3,mu,p1,nu)
*
m_tr(p4,mu,p2,nu),mu,nu);
(%o17) 64
*
sin(th)2
*
E4+96
*
cos(th)2
*
E4+64
*
cos(th)
*
E4+96
*
E4
(%i18) method3a : expand (trigsimp (%));
(%o18) 32
*
cos(th)2
*
E4+64
*
cos(th)
*
E4+160
*
E4
(%i19) method3a - method1;
(%o19) 0
Now, using the transparently explicit matrix route with mat_trace, we call the result method3b.
(%i20) method3b : Con(mat_trace(sL(p3) . Gam[mu] . sL(p1) . Gam[nu])
*
mat_trace(sL(p4) . Gam[mu] . sL(p2) . Gam[nu]),mu,nu);
(%o20) 64
*
sin(th)2
*
E4+96
*
cos(th)2
*
E4+64
*
cos(th)
*
E4+96
*
E4
(%i21) method3b : expand (trigsimp (%));
(%o21) 32
*
cos(th)2
*
E4+64
*
cos(th)
*
E4+160
*
E4
(%i22) method3b - method1;
(%o22) 0
As a passing remark, it is easy to verify two properties of Gam[5] using explicit matrices. First, Gam[5] has zero trace,
and then that the matrix product of Gam[5] with itself results in the unit 4 x 4 matrix:
(%i23) mat_trace(Gam[5]);
(%o23) 0
(%i24) is (equal (Gam[5] . Gam[5],I4));
(%o24) true
12.4.8 Contraction Choices: Timing Comparisons for Polarized Squared Amplitudes
In the examples above, as well as in the section dealing with moller3.mac, which looks for consistency in different
methods of calculation of the square of polarized amplitudes in Moller scattering of nite mass polarized leptons, we have
examples which compare matrix methods (m_tr) with symbolic methods (nc_tr).
Here we present an example which shows that the symbolic route will take about 3 - 7 times as long as the explicit
matrix route, depending on the contraction function chosen.
The quickest symbolic contraction route is to force the use of mcon as the contraction function (in the context of the
contraction of a product of nc_trs).
Forcing the use of econ almost doubles the time required (as compared with using mcon), whereas using Con (which
interposes some intermediate poking at the expression to be contracted) increases the length of time over the forced use
of econ (which Con eventually calls).
We have commented out most of moller3.mac and used just the top section to set up the invariants and kinematics.
The work starting with (%i27) is interactive work.
(%i3) load(dirac2);
(%o1) "c:/work5/dirac2.mac"
(%i2) batch("moller3.mac");
read and interpret file: #pc:/work5/moller3.mac
(%i3) " moller3.mac "
(%i4) "
*********************************************
"
(%i5) print(" ver: ",_binfo%," date: ",mydate)
ver: Maxima 5.23.2 date: 2011-04-09
(%i6) " Maxima by Example, Ch. 12 "
(%i7) " Dirac Algebra and Quantum Electrodynamics "
(%i8) " Edwin L. Woollett "
24
(%i9) " http://www.csulb.edu/woollett "
(%i10) " woollett@charter.net "
(%i11) "SQUARED POLARIZED AMPLITUDES VIA SYMBOLIC AND MATRIX TRACE METHODS"
(%i12) " FOR ARBITRARY ENERGY MOLLER SCATTERING "
(%i13) " e(-,p1,sv1) + e(-,p2,sv2) --> e(-,p3,sv3) + e(-,p4,sv4) "
(%i14) " --------------------------------------------------------"
(%i15) invar(D(p1,p1) = m2,D(p2,p2) = m2,D(p3,p3) = m2,D(p4,p4) = m2,
D(p1,Sp1) = 0,D(Sp1,Sp1) = -1,D(p2,Sp2) = 0,D(Sp2,Sp2) = -1,
D(p3,Sp3) = 0,D(Sp3,Sp3) = -1,D(p4,Sp4) = 0,D(Sp4,Sp4) = -1)
(%i16) comp_def(p1(E,0,0,p),Sp1(p/m,0,0,E/m),p2(E,0,0,-p),Sp2(p/m,0,0,(-E)/m),
p3(E,p
*
sin(th),0,p
*
cos(th)),
Sp3(p/m,E
*
sin(th)/m,0,E
*
cos(th)/m),
p4(E,-p
*
sin(th),0,-p
*
cos(th)),
Sp4(p/m,(-E
*
sin(th))/m,0,(-E
*
cos(th))/m))
(%i17) p_Em(expr):=expand(ratsubst(E2-m2,p2,expr))
(%i18) E_pm(expr):=expand(ratsubst(m2+p2,E2,expr))
(%i19) s_th:VP(p2+p1,p2+p1)
(%o19) 4
*
E2
(%i20) t_th:VP(p1-p3,p1-p3)
(%o20) 2
*
p2
*
cos(th)-2
*
p2
(%i21) u_th:VP(p1-p4,p1-p4)
(%o21) -2
*
p2
*
cos(th)-2
*
p2
(%i22) t_th2:to_ao2(t_th,th)
(%o22) -4
*
p2
*
sin(th/2)2
(%i23) u_th2:to_ao2(u_th,th)
(%o23) 4
*
p2
*
sin(th/2)2-4
*
p2
(%i24) t_thE:p_Em(t_th)
(%o24) 2
*
cos(th)
*
E2-2
*
E2-2
*
m2
*
cos(th)+2
*
m2
(%i25) u_thE:p_Em(u_th)
(%o25) -2
*
cos(th)
*
E2-2
*
E2+2
*
m2
*
cos(th)+2
*
m2
(%o26) "moller3.mac"
(%i27) /
*
case RR --> LL
*
/
[sv1,sv2,sv3,sv4]:[1,1,-1,-1]$
(%i28) /
*
explicit matrix method always faster
*
/
M1n_m : trigsimp (E_pm (mcon ( m_tr (S(sv3,Sp3),p3+m,mu,S(sv1,Sp1),p1+m,nu)
*
m_tr (S(sv4,Sp4),p4+m,mu,S(sv2,Sp2),p2+m,nu), mu,nu)));
(%o28) -4
*
m4
*
sin(th)2-8
*
m4
*
cos(th)+8
*
m4
(%i29) time(%);
(%o29) [1.11]
(%i30) /
*
forcing use of econ with symbolic method --> six times
as long as explicit matrix method
*
/
M1n_s : trigsimp (E_pm (econ ( nc_tr (S(sv3,Sp3),p3+m,mu,S(sv1,Sp1),p1+m,nu)
*
nc_tr (S(sv4,Sp4),p4+m,mu,S(sv2,Sp2),p2+m,nu), mu,nu)));
(%o30) -4
*
m4
*
sin(th)2-8
*
m4
*
cos(th)+8
*
m4
(%i31) time(%);
(%o31) [6.25]
(%i32) /
*
forcing use of mcon here is the fastest symbolic method
*
/
M1n_sm : trigsimp (E_pm (mcon ( nc_tr (S(sv3,Sp3),
p3+m,mu,S(sv1,Sp1),p1+m,nu)
*
nc_tr (S(sv4,Sp4),
p4+m,mu,S(sv2,Sp2),p2+m,nu), mu,nu)));
(%o32) -4
*
m4
*
sin(th)2-8
*
m4
*
cos(th)+8
*
m4
(%i33) time(%);
(%o33) [3.03]
(%i34) /
*
letting Con choose method here ends up with choosing econ,
and Cons overhead results in the slowest symbolic method
*
/
M1n_sc : trigsimp (E_pm (Con ( nc_tr (S(sv3,Sp3),
p3+m,mu,S(sv1,Sp1),p1+m,nu)
*
nc_tr (S(sv4,Sp4),
p4+m,mu,S(sv2,Sp2),p2+m,nu), mu,nu)));
(%o34) -4
*
m4
*
sin(th)2-8
*
m4
*
cos(th)+8
*
m4
(%i35) time(%);
(%o35) [7.22]
25
(%i36) M1n_s - M1n_m;
(%o36) 0
(%i37) M1n_sm - M1n_m;
(%o37) 0
(%i38) M1n_sc - M1n_m;
(%o38) 0
12.4.9 Dirac Spinors and Helicity Amplitudes
Finally, we present an example of an expression constructed from Dirac spinors and matrices for given helicity quantum
numbers.
We generate an explicit Dirac spinor, a four element column matrix, denoted symbolically by u(p, ) using the func-
tion UU(E,p,theta,phi,sv) which descibes a spin 1/2 particle with relativistic energy E, 3-momentum magni-
tude p, whose 3-vector momentum is described by polar angles (theta,phi), where 0 <= theta <= %pi and
0 <= phi < 2
*
%pi, and whose helicity quantum number sv is 1 for helicity +1/2 and sv is -1 for helicity -1/2.
We normally describe scattering in the z-x plane, and if p
1
makes an angle th <= %pi with the positive z axis and
we are describing positive helicity (sv = 1) (or righthanded, R) we would use the syntax
up1 : UU(E1,p1,th,0,1).
If a different particle is moving in the z-x plane in a direction opposite to p1 with 3-momentum p2 and energy E2, we
would use the syntax (taking sv2 = +1)
up2 : UU (E2,p2,%pi - th, %pi, 1). For a negative helicity particle, we would use sv = -1.
A barred particle spinor (a row matrix) is generated by using sbar which has the denition
sbar (_uu%) := hc (_uu%) . Gam[0]$
in which hc is the package dened hermitian conjugate function.
We would use the syntax up1b : sbar (UU(E1,p1,th,0,1)) for example.
To generate the antiparticle spinor, denoted symbolically by v(p, ), for an antiparticle of energy E and 3-momentum p
we use the function VV(E,p,theta,phi,sv).
We normally describe scattering in the z-x plane, and if p
3
makes an angle th <= %pi with the positive z axis and
we are describing positive helicity (sv = 1) (or righthanded, R) we would use the syntax
vp3 : VV(E3,p3,th,0,1). For a negative helicity antiparticle, we would use sv = -1.
If a different antiparticle is moving in the z-x plane in a direction opposite to p
3
with 3-momentum p4 and energy E4,
we would use the syntax (taking for example sv4 = +1)
vp4 : VV (E4,p4,%pi - th, %pi, 1).
A barred antiparticle spinor (a row matrix) is generated by using sbar with the syntax
vp3b : sbar (VV (E3,p3,th,0,1)) for example.
26
12.4.10 Polarized Amplitudes for (electron,positron) (muon, antimuon)
High Energy Limit Dirac Spinor Polarized Amplitudes
In the following example, we construct spinors needed to calculate the matrix element for the RL --> RL process
of the process (electron, positron ) --> (muon, antimuon),
e(-,p1,1) + e(+,p2,-1) --> mu(-,p3,1) + mu(+,p4,-1), in the high energy limit in which we neglect
the mass of the particles compared with their relativistic energy. (See Peskin/Schroeder, p.131 ff for details.)
The incident electron has attributes p1,sv1 = 1 and is represented by the spinor up1 (a column vector), the incident
positron has attributes p2, sv2 = -1 and is represented by the barred antiparticle spinor vp2b (a row vector). The
outgoing muon has attributes p3 and sv3 = 1 and is represented by the barred particle spinor up3b (a row vector), the
outgoing antimuon has attributes p4 and sv4 = -1 and is represented by the antiparticle spinor vp4 (a column vector).
Since in the high energy limit the energy and 3-momentum magnitude are the same (ignoring mass), we replace p by E,
and we are working in the center of momentum frame, so each particle has the same energy.
(%i1) load(dirac2);
(%o1) "c:/work5/dirac2.mac"
(%i2) comp_def (p1(E,0,0,E),
p2(E,0,0,-E),
p3(E,E
*
sin(th),0,E
*
cos(th)),
p4(E,-E
*
sin(th),0,-E
*
cos(th)))$
(%i3) listarray(p3);
(%o3) [E,sin(th)
*
E,0,cos(th)
*
E]
(%i4) up1 : UU(E,E,0,0,1);
(%o4) matrix([0],[0],[sqrt(2
*
E)],[0])
(%i5) vp2b : sbar (VV (E,E,%pi,0,-1));
(%o5) matrix([0,-sqrt(2
*
E),0,0])
(%i6) a12 : vp2b.Gam[mu].up1;
(%o6) matrix([0,-sqrt(2
*
E),0,0]) . Gam[mu] . matrix([0],[0],[sqrt(2
*
E)],[0])
(%i7) up3b : sbar (UU (E,E,th,0,1));
(%o7) matrix([cos(th/2)
*
sqrt(2
*
E),sin(th/2)
*
sqrt(2
*
E),0,0])
(%i8) vp4 : VV (E,E,%pi-th,%pi,-1);
(%o8) matrix([0],[0],[-cos((%pi-th)/2)
*
sqrt(2
*
E)],[sin((%pi-th)/2)
*
sqrt(2
*
E)])
(%i9) a34 : up3b . Gam[mu] . vp4;
(%o9) matrix([cos(th/2)
*
sqrt(2
*
E),sin(th/2)
*
sqrt(2
*
E),0,0])
. Gam[mu]
. matrix([0],[0],[-cos((%pi-th)/2)
*
sqrt(2
*
E)],
[sin((%pi-th)/2)
*
sqrt(2
*
E)])
Both a12 and a34 are (in general) complex numbers. Ignoring the factor -e2, the numerator Mn of the transition
amplitude is given by the contraction of the product a12
*
a34 on the index mu.
(%i10) Mn : mcon (a12
*
a34,mu);
(%o10) 8
*
cos(th/2)2
*
E2
The denominator of the transition amplitude is called s_th and is the 4-vector dot product of the sum of the incident
4-momenta:
(%i11) s_th : VP (p1 + p2,p1 + p2);
(%o11) 4
*
E2
The amplitude Amp_RL_RL (again, we are ignoring -e2) is then
(%i12) Amp_RL_RL : Mn/s_th;
(%o12) 2
*
cos(th/2)2
27
Note that this amplitude is real.
As an aside, in the high energy limit the spinor normalizations are zero. For example:
(%i13) sbar(up1) . up1;
(%o13) 0
If we redene up1 for the arbitrary energy case, in which mass cannot be neglected, then the normalization becomes
2
*
m, where m is the mass of the particle.
(%i14) up1 : UU(E,p,0,0,1);
(%o14) matrix([sqrt(E-p)],[0],[sqrt(E+p)],[0])
(%i15) sbar (up1) . up1;
(%o15) 2
*
sqrt(E-p)
*
sqrt(E+p)
(%i16) rootscontract (%);
(%o16) 2
*
sqrt(E2-p2)
(%i17) ratsubst (m, sqrt(E2-p2), %);
(%o17) 2
*
m
which goes to zero as m goes to zero.
Trace Methods and High Energy Limit Helicity Projection Matrices and Operators
We can check the square of the polarized amplitude calculated above (using Dirac spinors for given helicity assignments)
by using either the mat_trace method, the m_tr method, or the purely symbolic method. The matrix method using
m_tr syntax is designed to have arguments which look the same as the purely symbolic case.
We rst use the mat_trace method using explicit Dirac matrices. For the helicity dependent matrix element squared via
trace methods, helicity projection operators must be inserted into the traces. We use P for the explicit helicity projection
matrix, which takes two forms, depending on whether massive or (formally) massless particles are considered. In the fol-
lowing, P(sv) is the massless case helicity projection matrix which becomes (I4 + sv
*
Gam[5])/2, and \mvsv
takes values 1, -1.
(%i18) is (equal (P(1), (I4 + Gam[5])/2 ));
(%o18) true
(%i19) is (equal (P(-1), (I4 - Gam[5])/2 ));
(%o19) true
In the massless limit P(+1) picks out negative physical helicity case (sv = -1) if we are dealing with an antiparticle.
(%i20) Mn_sq:(a12:mat_trace(P(1) . sL(p2)
. Gam[mu] . P(1) . sL(p1) . Gam[nu]),
a34:mat_trace(P(1) . sL(p3)
. Gam[mu] . P(1) . sL(p4) . Gam[nu]),
mcon (a12
*
a34,mu,nu),factor(%%));
(%o20) 16
*
(cos(th)+1)2
*
E4
(%i21) time(%);
(%o21) [0.03]
To compare with the absolute value squared of the transition matrix element calculated above using Dirac spinors, we
convert the a function of th/2 using the Dirac package function to_ao2 (expr,a), where the symbol for the angle
must be correctly inserted.
(%i22) Mn_sq : to_ao2 (Mn_sq, th);
(%o22) 64
*
cos(th/2)4
*
E4
28
Dividing this numerator squared expression by s_th2, we get what should be the same as Amp_RL_RL2 (since
Amp_RL_RL is real):
(%i23) M_sq : Mn_sq/s_th2;
(%o23) 4
*
cos(th/2)4
(%i24) M_sq - Amp_RL_RL2;
(%o24) 0
which shows agreement.
We next use the m_tr method, which is translated into an explicit mat_trace expression by m_tr, including the
replacement of S by P. The virtue of the m_tr method is simply that the notation is formally the same as is used in the
symbolic tr case.
(%i25) Mn_sq:(a12 : m_tr (S(1),p2,mu,S(1),p1,nu),
a34 : m_tr (S(1),p3,mu,S(1),p4,nu),
mcon (a12
*
a34,mu,nu), factor(%%));
(%o25) 16
*
(cos(th)+1)2
*
E4
(%i26) time(%);
(%o26) [0.02]
(%i27) Mn_sq : to_ao2 (Mn_sq, th);
(%o27) 64
*
cos(th/2)4
*
E4
(%i28) M_sq : Mn_sq/s_th2;
(%o28) 4
*
cos(th/2)4
(%i29) M_sq-Amp_RL_RL2;
(%o29) 0
which again shows agreement.
For a purely symbolic check, the use of a nc_tr method is recommended, since the helicity projection operators S(1)
and S(-1) eventually cause the introdution of two Eps factors which the present version of the Dirac package does not
symbolically reduce (ie scon does not effect the contraction of a product of Eps containing contraction indices). For
example,
(%i30) scon (LI(p1,N4)
*
LI(p2,N3)
*
Eps(N3,mu,N4,nu)
*
LI(p3,N7)
*
LI(p4,N8)
*
Eps(N7,mu,N8,nu), mu,nu);
(%o30) LI(p1,N4)
*
LI(p2,N3)
*
LI(p3,N7)
*
LI(p4,N8)
*
Eps(N3,mu,N4,nu)
*
Eps(N7,mu,N8,nu)
Here is the recommended symbolic approach, in which the immediate use of nc_tr allows mcon to do the contraction.
The factor S(sv) --> (1 + sv
*
G5)/2, effectively, is the massless helicity projection operator for the symbolic
method.
(%i31) Mn_sq:(a12 : nc_tr (S(1),p2,mu,S(1),p1,nu),
a34 : nc_tr (S(1),p3,mu,S(1),p4,nu),
mcon (a12
*
a34,mu,nu),
factor(%%));
(%o31) -16
*
(sin(th)2-2
*
cos(th)-2)
*
E4
(%i32) time(%);
(%o32) [0.34]
(%i33) Mn_sq : to_ao2 (Mn_sq,th);
(%o33) 64
*
cos(th/2)4
*
E4
(%i34) M_sq : Mn_sq/s_th2;
(%o34) 4
*
cos(th/2)4
(%i35) M_sq - Amp_RL_RL2;
(%o35) 0
In terms of time, the explicit matrix approach (m_tr) to these polarized squared amplitudes is about ve times faster than
the nc_tr symbolic approach.
29
Just as an experiment, the non-recommended and slow calculation, in which we apply noncov to the product of two
expressions produced by tr, and then use mcon to contract, looks like this:
(%i36) Mn_sq : (a12 : tr(S(1),p2,mu,S(1),p1,nu),
a34 : tr(S(1),p3,mu,S(1),p4,nu),
noncov (a12
*
a34),
mcon (%%,mu,nu),factor(%%));
(%o36) -16
*
(sin(th)2-2
*
cos(th)-2)
*
E4
(%i37) time(%);
(%o37) [3.89]
which took about eleven times as long as the recommended nc_tr method. Slightly longer times arise with the explicit
use of econ or Con:
(%i38) Mn_sq : (a12 : tr(S(1),p2,mu,S(1),p1,nu),
a34 : tr(S(1),p3,mu,S(1),p4,nu),
noncov (a12
*
a34),
econ (%%,mu,nu),factor(%%));
(%o38) -16
*
(sin(th)2-2
*
cos(th)-2)
*
E4
(%i39) time(%);
(%o39) [3.98]
(%i40) Mn_sq : (a12 : tr(S(1),p2,mu,S(1),p1,nu),
a34 : tr(S(1),p3,mu,S(1),p4,nu),
noncov (a12
*
a34),
Con (%%,mu,nu),factor(%%));
(%o40) -16
*
(sin(th)2-2
*
cos(th)-2)
*
E4
(%i41) time(%);
(%o41) [4.01]
Arbitrary Energy Case (e e mu mu) of Polarized Amplitudes via Dirac Spinors
We continue to examine the same physical process (e(-) e(+) --> mu(-) mu(+)) but for arbitary center of mo-
mentum frame energy, in which case we must take into account the mass of the electron m and the mass of the muon M.
We use (as before) the symbols p3,p4 for the nal muon 4-momenta, but now the symbol k for the magnitude of a
nal muon 3-momentum, and the symbol p for the magnitude of an initial electron 3-momentum. The overall process
description is:
e(-,p1,sv1) + e(+,p2,sv2) --> mu(-,p3,sv3) + mu(+,p4,sv4).
The total relativistic energy of the initial electron and positron, 2
*
E, is shared equally (center of momentum frame) and
passed on to the the nal muon and antimuon by energy conservation, and again shared equally, so each particle has
relativistic energy E.
Thus E = sqrt(p2+m2) = Emu = sqrt(k2 + M2).
(%i42) load(dirac2);
(%o1) "c:/work5/dirac2.mac"
(%i2) comp_def(p1(E,0,0,p),
p2(E,0,0,-p),
p3(E,k
*
sin(th),0,k
*
cos(th)),
p4(E,-k
*
sin(th),0,-k
*
cos(th)))$
The denominator of the transition amplitude is called s_th (the square of the total CMS energy) and is the 4-vector dot
product of the sum of the incident 4-momenta:
(%i3) s_th : VP (p1 + p2,p1 + p2);
(%o3) 4
*
E2
30
As in the high energy case, the incident electron has attributes p1,sv1 = 1 and is represented by the spinor up1 (a
column vector), the incident positron has attributes p2, sv2 = -1 and is represented by the barred antiparticle spinor
vp2b (a row vector). The outgoing muon has attributes p3 and sv3 = 1 and is represented by the barred particle spinor
up3b (a row vector), the outgoing antimuon has attributes p4 and sv4 = -1 and is represented by the antiparticle
spinor vp4 (a column vector).
(%i4) up1 : UU(E,p,0,0,1);
(%o4) matrix([sqrt(E-p)],[0],[sqrt(E+p)],[0])
(%i5) vp2b : sbar(VV(E,p,%pi,0,-1));
(%o5) matrix([0,-sqrt(E+p),0,sqrt(E-p)])
(%i6) a12 : vp2b . Gam[mu] . up1$
(%i7) up3b : sbar(UU(E,k,th,0,1));
(%o7) matrix([cos(th/2)
*
sqrt(E+k),sin(th/2)
*
sqrt(E+k),cos(th/2)
*
sqrt(E-k),
sin(th/2)
*
sqrt(E-k)])
(%i8) vp4 : VV(E,k,%pi-th,%pi,-1);
(%o8) matrix([cos((%pi-th)/2)
*
sqrt(E-k)],[-sin((%pi-th)/2)
*
sqrt(E-k)],
[-cos((%pi-th)/2)
*
sqrt(E+k)],[sin((%pi-th)/2)
*
sqrt(E+k)])
(%i9) a34 : up3b . Gam[mu] . vp4$
Both a12 and a34 are (in general) complex numbers. Ignoring the factor -e2, the numerator Mn of the transition am-
plitude is given by the contraction of the product a12
*
a34 on the index mu, and, as before, the amplitude Amp_RL_RL
(again, we are ignoring -e2) is Mn/s_th.
(%i10) Mn : mcon (a12
*
a34,mu);
(%o10) 8
*
cos(th/2)2
*
E2
(%i11) Amp_RL_RL : Mn/s_th;
(%o11) 2
*
cos(th/2)2
which is the same as the high energy limit previously discussed for this process.
Trace Methods and Arbitrary Energy Case Helicity Projection Matrices and Operators
The symbol Sp1 is the positive helicity spin 4-vector dened by the incident electron 4-momentum vector p1. The
symbol Sp2 is the positive helicity spin 4-vector dened by the incident positron 4-momentum vector p2. The symbol
Sp3 is the positive helicity spin 4-vector dened by the nal muon 4-momentum vector p3. The symbol Sp4 is the
positive helicity spin 4-vector dened by the nal antimuon 4-momentum vector p4.
(%i12) tr(S(1,Sp1));
(%o12) 2
(%i13) comp_def(Sp1(p/m,0,0,E/m),
Sp2(p/m,0,0,(-E)/m),
Sp3(k/M,E
*
sin(th)/M,0,E
*
cos(th)/M),
Sp4(k/M,(-E
*
sin(th))/M,0,(-E
*
cos(th))/M))$
(%i14) p_Em(expr) := expand (ratsubst (E2-m2,p2,expr))$
(%i15) k_EM(expr) := expand (ratsubst (E2-M2,k2,expr))$
In the contraction result, we can simplify by replacing k2 and p2 in terms of the equivalent quantities depending on E
and either m or M.
First the mat_trace method. In the explicit matrix method, a matrix factor P(sv1,Sp1), turns into the matrix
(I4 + sv1
*
Gam[5].sL(Sp1))/2, in which Sp1 is the positive helicity spin 4-vector implied by the particles
4-momentum p1.
(%i16) is (equal (P(1,Sp1),(I4 + Gam[5].sL(Sp1))/2));
(%o16) true
(%i17) is (equal (P(-1,Sp1),(I4 - Gam[5].sL(Sp1))/2));
(%o17) true
31
The matrix P(sv1,Sp1) is the matrix version of the nite mass spin projection operator, and requires a frame dependent
denition (using comp_def) of the components of the positive helicity spin 4-vector corresponding to a particles 4-
momentum.
(%i18) a12 : mat_trace(Gam[mu] . P(1,Sp1) . (sL(p1) + m
*
I4)
. Gam[nu] . P(-1,Sp2) . (sL(p2) - m
*
I4))$
(%i19) a34 : mat_trace (Gam[mu] . P(-1,Sp4) . (sL(p4) - M
*
I4)
. Gam[nu] . P(1,Sp3) . (sL(p3) + M
*
I4))$
(%i20) Mn_sq : (mcon (a12
*
a34,mu,nu),
k_EM(%%),p_Em(%%),
expand(trigsimp(%%)),factor(%%));
(%o20) 16
*
(cos(th)+1)2
*
E4
(%i21) time(%);
(%o21) [0.58]
(%i22) Mn_sq : to_ao2 (Mn_sq, th);
(%o22) 64
*
cos(th/2)4
*
E4
(%i23) M_sq : Mn_sq/s_th2;
(%o23) 4
*
cos(th/2)4
The mat_trace method requires careful attention to parentheses, and it is simpler (and also faster) to use the equivalent
m_tr method, which uses the same operator notation as tr, but actually translates everything into a mat_trace
expression.
(%i24) a12 : m_tr (mu,S(1,Sp1),p1 + m,nu,S(-1,Sp2),p2 - m)$
(%i25) a34 : m_tr (mu,S(-1,Sp4),p4 - M,nu,S(1,Sp3),p3 + M)$
(%i26) Mn_sq : (mcon (a12
*
a34,mu,nu),
k_EM(%%),p_Em(%%),
expand(trigsimp(%%)),factor(%%));
(%o26) 16
*
(cos(th)+1)2
*
E4
(%i27) time(%);
(%o27) [0.59]
(%i28) Mn_sq : to_ao2 (Mn_sq, th);
(%o28) 64
*
cos(th/2)4
*
E4
(%i29) M_sq : Mn_sq/s_th2;
(%o29) 4
*
cos(th/2)4
Finally we use the symbolic nc_tr method, which is a specially designed function which immediately takes noncov
of the trace of each term of the general expansion of the starting expression.
The symbolic method uses the helicity projection operators such as S(sv1,Sp1), used for a particle with helicity
quantum number sv1 = +/- 1 and 4-momentum p1 and corresponding positive helicity spin vector Sp1 dened by
p1. In the symbolic method, a factor S(sv1,Sp1), turns into the symbolic factor
(1 + sv1
*
G5
*
Sp1)/2, in which Sp1 is the positive helicity spin 4-vector implied by the particles 4-momentump1.
(%i36) tr(S(1,Sp1));
(%o36) 2
(%i37) tr(S(-1,Sp1));
(%o37) 2
(%i38) tr(mu,S(1,Sp1));
(%o38) 0
(%i39) tr(mu,nu,rh,G5,Sp1);
(%o39) 4
*
%i
*
Eps(mu,nu,rh,N4)
*
LI(Sp1,N4)
(%i40) tr(mu,nu,rh,S(1,Sp1));
(%o40) 2
*
%i
*
Eps(mu,nu,rh,N5)
*
LI(Sp1,N5)
(%i41) tr(mu,nu,rh,S(-1,Sp1));
(%o41) -2
*
%i
*
Eps(mu,nu,rh,N6)
*
LI(Sp1,N6)
32
We rst populate the list invarR used by the symbolic method for the immediate evaluation of 4-vector dot products
such as D(p1,p1), which is done by tr.
(%i42) invarR;
(%o42) []
(%i43) invar(D(p1,p1) = m2,
D(p2,p2) = m2,
D(p3,p3) = M2,
D(p4,p4) = M2,
D(p1,Sp1) = 0,
D(Sp1,Sp1) = -1,
D(p2,Sp2) = 0,
D(Sp2,Sp2) = -1,
D(p3,Sp3) = 0,
D(Sp3,Sp3) = -1,
D(p4,Sp4) = 0,
D(Sp4,Sp4) = -1)$
(%i44) invarR;
(%o44) [D(Sp4,Sp4) = -1,D(p4,Sp4) = 0,D(Sp3,Sp3) = -1,D(p3,Sp3) = 0,
D(Sp2,Sp2) = -1,D(p2,Sp2) = 0,D(Sp1,Sp1) = -1,D(p1,Sp1) = 0,
D(p4,p4) = M2,D(p3,p3) = M2,D(p2,p2) = m2,D(p1,p1) = m2]
(%i45) Mn_sq:(nc_tr(mu,S(1,Sp1),m+p1,nu,S(-1,Sp2),p2-m),a12:p_Em(%%),
nc_tr(mu,S(-1,Sp4),p4-M,nu,S(1,Sp3),M+p3),a34:k_EM(%%),
mcon(a12
*
a34,mu,nu),k_EM(%%),p_Em(%%),trigsimp(%%),factor(%%));
(%o45) -16
*
(sin(th)2-2
*
cos(th)-2)
*
E4
(%i46) time(%);
(%o46) [1.43]
(%i47) Mn_sq:to_ao2(Mn_sq,th);
(%o47) 64
*
cos(th/2)4
*
E4
(%i48) M_sq:Mn_sq/s_th2;
(%o48) 4
*
cos(th/2)4
We now proceed to use a batch le method to examine a series of basic simple examples of quantum electrodynamics
using our tools.
12.5 moller0.mac: Scattering of Identical Scalar (Spin 0) Charged Particles
The batch le moller0.mac treats the scattering event

+
(p
1
) +
+
(p
2
)
+
(p
3
) +
+
(p
4
) (12.38)
With the denitions
s = (p
1
+ p
2
)
2
, t = (p
1
p
3
)
2
, u = (p
1
p
4
)
2
, (12.39)
The symbolic method is used to derive the unpolarized differential cross section, starting with the invariant amplitude
M = M
1
+ M
2
, in which
M
1
=
e
2
(p
1
+ p
3
) (p
2
+ p
4
)
t
, (12.40)
and
M
2
=
e
2
(p
1
+ p
4
) (p
2
+ p
3
)
u
. (12.41)
The Feynman rules for scalar electrodynamics are discussed in: G/R, QED, p. 434 ff, Renton, p.180 ff, I/Z, p. 282 ff,
B/D, RQM, p. 195, Aitchison, p. 51 ff, A/H, p. 158 ff, Kaku, p. 158 ff, Quigg, p. 49, Schwinger, p. 284 ff, H/M, Ch. 4.
33
The list invarR (rules for invariant dot products) is rst populated using the Dirac package function invar, which will
be used by tr to replace invariant dot products such as D(p1,p1), D(p1,p2), etc by expressions in terms of the mass
m of the scalar particle, and the Mandelstam variables s, t, u.
We then use comp_def to dene the components of the 4-vectors in the center of momentum (CM) frame, and evaluate
the Mandelstam variables in that frame.
Some other Dirac package dened functions used in moller0.mac are take_parts, ev_Ds, pullfac, sub_stu,
and fr_ao2.
The batch le moller0.mac is used after loading in the Dirac package.
(%i1) load(dirac2);
dirac2.mac
simplifying-new.lisp
dgtrace2.mac
dgcon2.mac
dgeval2.mac
dgmatrix2.mac
scalarL = [c1, c2, c3, c4, c5, c6, c7, c8, c9, c10]
indexL = [n1, n2, n3, n4, n5, n6, n7, n8, n9, n10, la, mu, nu, rh, si, ta, al,
be, ga, de, ep]
massL = [m, M]
Nlast = 0
reserved program capital letter name use:
Chi, Con, D, Eps, G, G(1), G5, G5p, Gam, Gm, Gtr, LI, Nlast, UI, P, S, Sig
UU, VP, VV, I2, Z2, CZ2, I4, Z4, CZ4, RZ4, N1,N2,...
reserved array names: gmet, eps4
invar_flag = true
stu_flag = false
(%o1) "c:/work5/dirac2.mac"
(%i2) batch ("moller0.mac");
read and interpret file: #pc:/work5/moller0.mac
(%i3) " ======================================================"
(%i4) " file moller0.mac "
(%i5) " Maxima by Example, Ch. 12 "
(%i6) " Dirac Algebra and Quantum Electrodynamics "
(%i7) " Edwin L Woollett, woollett@charter.net "
(%i8) " http://www.csulb.edu/woollett "
(%i9) print(" ver: ",_binfo%," date: ",mydate)
ver: Maxima 5.23.2 date: 2011-04-04
(%i10) " ======================================================"
(%i11) " ELASTIC SCATTERING OF SPIN 0 PARTICLES (SAME CHARGE)"
(%i12) " FOR EXAMPLE: "
(%i13) " PI(+,p1) PI(+,p2) ---> PI(+,p3) PI(+,p4) "
(%i14) " POPULATE THE LIST invarR OF 4-VEC DOT PRODUCT VALUES, "
(%i15) " Using p1 + p2 = p3 + p4, s = (p1+p2)2 = (p3+p4)2 ,"
(%i16) " t = (p1-p3)2 = (p2-p4)2,"
(%i17) " and u = (p1-p4)2 = (p2-p3)2 "
(%i18) " ---------------------------------------------"
(%i19) invar(D(p1,p1) = m2,D(p1,p2) = s/2-m2,D(p1,p3) = m2-t/2,
D(p1,p4) = m2-u/2,D(p2,p2) = m2,D(p2,p3) = m2-u/2,
D(p2,p4) = m2-t/2,D(p3,p3) = m2,D(p3,p4) = s/2-m2,
D(p4,p4) = m2)
(%i20) "------------------------------------------------------"
(%i21) " factor out -e2 from Mfi, leaving Mfi = M1 + M2 "
(%i22) M1:D(p3+p1,p4+p2)/D(p1-p3,p1-p3)
(%i23) M2:D(p4+p1,p3+p2)/D(p1-p4,p1-p4)
(%i24) M1:ev_Ds(M1)
(%o24) s/t-u/t
34
(%i25) M1:pullfac(M1,1/t)
(%o25) (s-u)/t
(%i26) M2:ev_Ds(M2)
(%o26) s/u-t/u
(%i27) M2:pullfac(M2,1/u)
(%o27) (s-t)/u
(%i28) Mfi:M2+M1
(%o28) (s-u)/t+(s-t)/u
(%i29) " we get the result Mfi = (s-u)/t + (s-t)/u "
(%i30) "CM FRAME EVALUATION "
(%i31) assume(p > 0,th >= 0,th <= %pi)
(%i32) comp_def(p1(E,0,0,p),p2(E,0,0,-p),p3(E,p
*
sin(th),0,p
*
cos(th)),
p4(E,-p
*
sin(th),0,-p
*
cos(th)))
(%i33) s_th:noncov(D(p2+p1,p2+p1))
(%o33) 4
*
E2
(%i34) t_th:factor(noncov(D(p1-p3,p1-p3)))
(%o34) 2
*
p2
*
(cos(th)-1)
(%i35) u_th:factor(noncov(D(p1-p4,p1-p4)))
(%o35) -2
*
p2
*
(cos(th)+1)
(%i36) Mfi:sub_stu(Mfi)
(%o36) -4
*
E2/(p2
*
sin(th)2)-4/sin(th)2+2
(%i37) " We want to combine the first two terms "
(%i38) " and replace p2 by E2 - m2 in the numerator"
(%i39) Mfi_12:ratsimp(take_parts(Mfi,1,2))
(%o39) -(4
*
E2+4
*
p2)/(p2
*
sin(th)2)
(%i40) Mfi_12n:num(Mfi_12)
(%o40) -4
*
E2-4
*
p2
(%i41) Mfi_12n:factor(expand(subst(p2 = E2-m2,Mfi_12n)))
(%o41) -4
*
(2
*
E2-m2)
(%i42) Mfi_12:Mfi_12n/denom(Mfi_12)
(%o42) -4
*
(2
*
E2-m2)/(p2
*
sin(th)2)
(%i43) " We now add in the third term and extract -2 "
(%i44) Mfi:pullfac(part(Mfi,3)+Mfi_12,-2)
(%o44) -2
*
(2
*
(2
*
E2-m2)/(p2
*
sin(th)2)-1)
(%i45) " having absorbed e4 into a factor A which multiplies"
(%i46) " |Mfi|2, and using e2 = 4
*
%pi
*
alpha,"
(%i47) " in general, A = alpha2
*
(pf/pi)/(4
*
Ecm2) "
(%i48) " but here, pf = pi, and Ecm = 2
*
E, so "
(%i49) A:alpha2/(16
*
E2)
(%i50) " We can now write down the differential scattering cross section:"
(%i51) dsigdo:A
*
Mfi2
(%i52) (display2d:true,display(dsigdo),display2d:false)
2 2
2 2 (2 E - m ) 2
alpha (------------- - 1)
2 2
p sin (th)
dsigdo = ---------------------------
2
4 E
(%i53) " which agrees with Itzykson and Zuber, p. 286."
(%i54) " The factor 1/sin(th)2 can be displayed in terms of th/2"
(%i55) " using: "
(%i56) fr_ao2(1/cos(th/2)2+1/sin(th/2)2,th)
(%o56) 4/sin(th)2
(%i57) " and, of course p2 = E2 - m2 "
(%o57) "moller0.mac"
35
Since this is the rst example of using a batch le to program the calculations, we show here the appearance of the batch
le moller0.mac.
To save space, we will only show the batch le run output for the remaining examples, even though the batch le output is
rather compressed (minimumwhite space) and symbols are rearranged in the way Maxima likes to organize printed output.
Of course you can edit your own batch le run output to include more white space which would make the results more
readable and easier to follow.
/
*
file moller0.mac
pi(+) pi(+) --> pi(+) pi(+)
in scalar electrodynamics, ie,
ignoring structure of pions
*
/
/
*
references:
Renton p.182
I/Z p. 286
Schwinger, pp. 285 - 289
*
/
" ======================================================"$
" file moller0.mac "$
" Maxima by Example, Ch. 12 "$
" Dirac Algebra and Quantum Electrodynamics "$
" Edwin L Woollett, woollett@charter.net "$
" http://www.csulb.edu/woollett "$
print (" ver: ",_binfo%," date: ",mydate )$
" ======================================================"$
" ELASTIC SCATTERING OF SPIN 0 PARTICLES (SAME CHARGE)"$
" FOR EXAMPLE: "$
" PI(+,p1) PI(+,p2) ---> PI(+,p3) PI(+,p4) "$
" POPULATE THE LIST invarR OF 4-VEC DOT PRODUCT VALUES, "$
" Using p1 + p2 = p3 + p4, s = (p1+p2)2 = (p3+p4)2 ,"$
" t = (p1-p3)2 = (p2-p4)2,"$
" and u = (p1-p4)2 = (p2-p3)2 "$
" ---------------------------------------------"$
invar (D(p1,p1) = m2,
D(p1,p2) = s/2 - m2,
D(p1,p3) = m2 - t/2,
D(p1,p4) = m2 - u/2,
D(p2,p2) = m2,
D(p2,p3) = m2 - u/2,
D(p2,p4) = m2 - t/2,
D(p3,p3) = m2,
D(p3,p4) = s/2 - m2,
D(p4,p4) = m2)$
"------------------------------------------------------"$
" factor out -e2 from Mfi, leaving Mfi = M1 + M2 "$
M1 : D(p1+p3,p2+p4)/D(p1-p3,p1-p3)$
M2 : D(p1+p4,p2+p3)/D(p1-p4,p1-p4)$
36
M1 : ev_Ds (M1);
M1 : pullfac (M1,1/t);
M2 : ev_Ds (M2);
M2 : pullfac (M2,1/u);
Mfi : M1 + M2;
" we get the result Mfi = (s-u)/t + (s-t)/u "$
"CM FRAME EVALUATION "$
assume ( p > 0, th >= 0, th <= %pi )$
comp_def ( p1( E,0,0,p),
p2( E,0,0,-p),
p3 (E,p
*
sin(th),0,p
*
cos(th)),
p4 (E,-p
*
sin(th),0,-p
*
cos(th)) )$
s_th : noncov (D(p1+p2,p1+p2));
t_th : factor (noncov (D(p1-p3,p1-p3)));
u_th : factor (noncov (D(p1-p4,p1-p4)));
Mfi : sub_stu(Mfi);
" We want to combine the first two terms "$
" and replace p2 by E2 - m2 in the numerator"$
Mfi_12 : ratsimp (take_parts (Mfi,1,2));
Mfi_12n : num (Mfi_12);
Mfi_12n : factor (expand (subst (p2 = E2 - m2, Mfi_12n)));
Mfi_12 : Mfi_12n/denom(Mfi_12);
" We now add in the third term and extract -2 "$
Mfi : pullfac (Mfi_12 + part(Mfi,3),-2);
" having absorbed e4 into a factor A which multiplies"$
" |Mfi|2, and using e2 = 4
*
%pi
*
alpha,"$
" in general, A = alpha2
*
(pf/pi)/(4
*
Ecm2) "$
" but here, pf = pi, and Ecm = 2
*
E, so "$
A : alpha2/(16
*
E2)$
37
" We can now write down the differential scattering cross section:"$
dsigdo : A
*
Mfi2$
(display2d:true, display (dsigdo), display2d:false)$
" which agrees with Itzykson and Zuber, p. 286."$
" The factor 1/sin(th)2 can be displayed in terms of th/2"$
" using: "$
fr_ao2(1/sin(th/2)2 + 1/cos(th/2)2, th );
" and, of course p2 = E2 - m2 "$
12.6 moller1.mac: High Energy Elastic Scattering of Two Electrons
The batch le moller1.mac treats the high energy limit of the scattering event
e

(p
1
,
1
) + e

(p
2
,
2
) e

(p
3
,
3
) + e

(p
4
,
4
) (12.42)
We consider the unpolarized ultra-relativistic limit in which the masses of the electrons can be ignored (|p| >> m).
Some references are: C. Moller, 1932, Schwinger, PSF I, p. 300 - 305, Greiner-Reinhardt, QED, 4th, Sec. 3.3, Grifth
Sec. 7.6, Jauch and Rohrlich, p. 252-257, Renton Sec. 4.3.1, B/D Sec. 7.9, BLP, Sec. 81.
(Note that BLPs symbol r
e
should be replaced by /mto compare with our results - they use (along with Greiner/Reinhardt)
Gaussian units for electromagnetic equations, whereas we use (along with Peskin/Schroeder and Bjorken/Drell) rational-
ized Heaviside-Lorentz electromagnetic units. e
2
HL
= 4 e
2
G
See Greiner/Reinhardt, Sec 4.2.).
The invariant amplitude M = M
1
M
2
, has two terms. Using the denitions
s = (p
1
+ p
2
)
2
, t = (p
1
p
3
)
2
, u = (p
1
p
4
)
2
, (12.43)
and the abbreviations
u
1
= u(p
1
,
1
), u
2
= u(p
2
,
2
), u
3
= u(p
3
,
3
), u
4
= u(p
4
,
4
) (12.44)
we have
M
1
=
e
2
u
3

u
1
u
4

u
2
t
, (12.45)
and
M
2
=
e
2
u
4

u
1
u
3

u
2
u
. (12.46)
The unpolarized differential cross section in the CM frame is rst found using symbolic methods, rst in terms of
s, t, u and then in terms of the scattering angle th.
The polarized amplitudes corresponding to denite electron helicity assignments are then calculated using explicit Dirac
spinors and matrices, and the sum of the squares of these amplitudes reproduces the unpolarized symbolic calculation.
Finally one polarized squared amplitude is calculated using the symbolic methods. We will surpress the information
output which dirac2.mac puts on the screen after loading in the package les.
38
(%i58) load(dirac2);
(%o1) "c:/work5/dirac2.mac"
(%i2) batch ("moller1.mac");
read and interpret file: #pc:/work5/moller1.mac
(%i3) " ======================================================"
(%i4) " file moller1.mac "
(%i5) " Maxima by Example, Ch. 12 "
(%i6) " Dirac Algebra and Quantum Electrodynamics "
(%i7) " Edwin L Woollett, woollett@charter.net "
(%i8) " http://www.csulb.edu/woollett "
(%i9) print(" ver: ",_binfo%," date: ",mydate)
ver: Maxima 5.23.2 date: 2011-04-04
(%i10) " ======================================================"
(%i11) " MOLLER SCATTERING "
(%i12) " HIGH ENERGY LIMIT, CENTER OF MOMENTUM FRAME, NEGLECT MASSES "
(%i13) " e(-,p1,sv1) + e(-,p2,sv2) --> e(-,p3,sv3) + e(-,p4,sv4) "
(%i14) " m = electron mass is set to zero."
(%i15) " ------------------------------------------ "
(%i16) "NON-POLARIZED DIFFERENTIAL CROSS SECTION: SYMBOLIC METHODS"
(%i17) " POPULATE THE LIST invarR OF 4-VEC DOT PRODUCT VALUES, "
(%i18) " Using p1 + p2 = p3 + p4, s = (p1+p2)2 = (p3+p4)2 ,"
(%i19) " t = (p1-p3)2 = (p2-p4)2,"
(%i20) " and u = (p1-p4)2 = (p2-p3)2 "
(%i21) " CASE HIGH ENERGY (HE) LIMIT E >> m "
(%i22) " ---------------------------------------------"
(%i23) invar(D(p1,p1) = 0,D(p1,p2) = s/2,D(p1,p3) = (-t)/2,D(p1,p4) = (-u)/2,
D(p2,p2) = 0,D(p2,p3) = (-u)/2,D(p2,p4) = (-t)/2,D(p3,p3) = 0,
D(p3,p4) = s/2,D(p4,p4) = 0)
(%i24) "------------------------------------------------------"
(%i25) " factor out -e2 from Mfi, leaving Mfi = M1 - M2 "
(%i26) " With a sum over all helicities implied,"
(%i27) " |Mfi|2 = M1n/t2 + M2n/u2 -M12n/(t
*
u) - M21n/(t
*
u) "
(%i28) " M1n = t2
*
M1
*
conj(M1), M2n = u2
*
M2
*
conj(M2) "
(%i29) " M12n = (t
*
u)
*
M1
*
conj(M2), M21n = (t
*
u)
*
M2
*
conj(M1), and: "
(%i30) M1n:factor(Con(tr(p3,mu,p1,nu)
*
tr(p4,mu,p2,nu),mu,nu))
(%o30) 8
*
(u2+s2)
(%i31) M2n:factor(Con(tr(p4,mu,p1,nu)
*
tr(p3,mu,p2,nu),mu,nu))
(%o31) 8
*
(t2+s2)
(%i32) " NOTE AUTOMATIC PRETRACE CONTRACTION OF REPEATED "
(%i33) " LORENTZ INDICES WITHIN A SINGLE TRACE OCCURS USING tr."
(%i34) M12n:factor(tr(p3,mu,p1,nu,p4,mu,p2,nu))
(%o34) -8
*
s2
(%i35) M21n:factor(tr(p4,mu,p1,nu,p3,mu,p2,nu))
(%o35) -8
*
s2
(%i36) MfiSQ:pullfac((-M21n)/(t
*
u)+(-M12n)/(t
*
u)+M2n/u2+M1n/t2,8)
(%o36) 8
*
((u2+s2)/t2+2
*
s2/(t
*
u)+(t2+s2)/u2)
(%i37) " We have absorbed e4 into A, with e2 = 4
*
%pi
*
alpha "
(%i38) " Averaging over initial spins means we need to divide A by 4"
(%i39) " to get the unpolarized differential cross section (CM, HE)"
(%i40) A:alpha2/(4
*
s)
(%i41) dsigdo_unpol_CM_HE:A
*
MfiSQ/4
(%o41) alpha2
*
((u2+s2)/t2+2
*
s2/(t
*
u)+(t2+s2)/u2)/(2
*
s)
(%i42) (display2d:true,display(dsigdo_unpol_CM_HE),display2d:false)
2 2 2 2 2
2 u + s 2 s t + s
alpha (------- + ---- + -------)
2 t u 2
t u
dsigdo_unpol_CM_HE = ---------------------------------
2 s
39
(%i43) " which agrees with Renton function of s,t, and u "
(%i44) " on page 159, Eq.(4.54) "
(%i45) " CONVERSION TO EXPLICIT FUNCTION OF SCATTERING ANGLE "
(%i46) assume(p > 0,th >= 0,th <= %pi)
(%i47) comp_def(p1(E,0,0,E),p2(E,0,0,-E),p3(E,E
*
sin(th),0,E
*
cos(th)),
p4(E,-E
*
sin(th),0,-E
*
cos(th)))
(%i48) s_th:VP(p2+p1,p2+p1)
(%o48) 4
*
E2
(%i49) t_th:factor(VP(p1-p3,p1-p3))
(%o49) 2
*
(cos(th)-1)
*
E2
(%i50) u_th:factor(VP(p1-p4,p1-p4))
(%o50) -2
*
(cos(th)+1)
*
E2
(%i51) " To get an expression we can compare with Renton and G/R, "
(%i52) " work with one term at a time "
(%i53) " sub_stu replaces s by s_th, t by t_th "
(%i54) " and u by u_th "
(%i55) M1SQ:sub_stu(M1n/t2)
(%o55) 8
*
cos(th)2/(cos(th)2-2
*
cos(th)+1)+16
*
cos(th)/(cos(th)2-2
*
cos(th)+1)
+40/(cos(th)2-2
*
cos(th)+1)
(%i56) " convert to (th/2) form "
(%i57) M1SQ:factor(ratsimp(to_ao2(M1SQ,th)))
(%o57) 8
*
(cos(th/2)4+1)/sin(th/2)4
(%i58) " ---------------- "
(%i59) M2SQ:sub_stu(M2n/u2)
(%o59) 8
*
cos(th)2/(cos(th)2+2
*
cos(th)+1)-16
*
cos(th)/(cos(th)2+2
*
cos(th)+1)
+40/(cos(th)2+2
*
cos(th)+1)
(%i60) M2SQ:ratsimp(to_ao2(M2SQ,th))
(%o60) (8
*
cos(th/2)4-16
*
cos(th/2)2+16)/cos(th/2)4
(%i61) M2SQ:factor(ratsubst(1-sin(th/2)2,cos(th/2)2,num(M2SQ)))/denom(M2SQ)
(%o61) 8
*
(sin(th/2)4+1)/cos(th/2)4
(%i62) " ---------------------- "
(%i63) M12INT:-2
*
sub_stu(M12n/(t
*
u))
(%o63) -64/(cos(th)2-1)
(%i64) M12INT:factor(M12INT)
(%o64) -64/((cos(th)-1)
*
(cos(th)+1))
(%i65) M12INT:subst([cos(th)-1 = -2
*
sin(th/2)2,1+cos(th) = 2
*
cos(th/2)2],
M12INT)
(%o65) 16/(cos(th/2)2
*
sin(th/2)2)
(%i66) " pull out a factor of 8 from each term. "
(%i67) MfiSQ:pullfac(M12INT+M2SQ+M1SQ,8)
(%o67) 8
*
((sin(th/2)4+1)/cos(th/2)4+2/(cos(th/2)2
*
sin(th/2)2)
+(cos(th/2)4+1)/sin(th/2)4)
(%i68) dsigdo_unpol_CM_HE:A
*
MfiSQ/4
(%o68) alpha2
*
((sin(th/2)4+1)/cos(th/2)4+2/(cos(th/2)2
*
sin(th/2)2)
+(cos(th/2)4+1)/sin(th/2)4)
/(2
*
s)
(%i69) (display2d:true,display(dsigdo_unpol_CM_HE),display2d:false)
4 th 4 th
sin (--) + 1 cos (--) + 1
2 2 2 2
alpha (------------ + ----------------- + ------------)
4 th 2 th 2 th 4 th
cos (--) cos (--) sin (--) sin (--)
2 2 2 2
dsigdo_unpol_CM_HE = --------------------------------------------------------
2 s
(%i70) " which agrees with Renton, page 159, Eq.(4.55) "
(%i71) " as well as G/R, page 138, Eq. (3-139) "
(%i72) "-------------------------------------------------"
40
(%i73) " Convert to simple function of th displayed by G/R "
(%i74) MfiSQ:fr_ao2(MfiSQ,th)
(%o74) 16
*
cos(th)4/sin(th)4+96
*
cos(th)2/sin(th)4+144/sin(th)4
(%i75) MfiSQ:factor(ratsimp(MfiSQ))
(%o75) 16
*
(cos(th)2+3)2/sin(th)4
(%i76) dsigdo_unpol_CM_HE:A
*
MfiSQ/4
(%o76) alpha2
*
(cos(th)2+3)2/(s
*
sin(th)4)
(%i77) dsigdo_unpol_CM_HE:subst(s = s_th,dsigdo_unpol_CM_HE)
(%o77) alpha2
*
(cos(th)2+3)2/(4
*
sin(th)4
*
E2)
(%i78) (display2d:true,display(dsigdo_unpol_CM_HE),display2d:false)
2 2 2
alpha (cos (th) + 3)
dsigdo_unpol_CM_HE = ----------------------
4 2
4 sin (th) E
(%i79) " which agrees with G/R, page 139, top. "
The next section of moller1.mac uses explicit Dirac spinors and matrices to calculate polarized amplitudes.
(%i80) " --------------------------------------------------"
(%i81) " HE POLARIZED AMPLITUDES USING EXPLICIT DIRAC SPINORS "
(%i82) " ------------------------------------------------- "
(%i83) " case HE RR --> RR, ie, (+1,+1) --> (+1,+1) polarized amplitude "
(%i84) " Define the needed Dirac spinors and barred spinors."
(%i85) (up1:UU(E,E,0,0,1),up3b:sbar(UU(E,E,th,0,1)),up2:UU(E,E,%pi,0,1),
up4b:sbar(UU(E,E,%pi-th,%pi,1)))
(%i86) " For example, following Halzen/Martin p. 120 ff :"
(%i87) " For helicity quantum numbers RR -> RR, the amplitude Mfi"
(%i88) " (pulling out -e2) has the form Mfi = M1 - M2 "
(%i89) " where M1 = Mt/t and M2 = Mu/u, "
(%i90) " where t = (p1-p3)2 and u = (p1-p4)2, and "
(%i91) Mt:(a13:up3b . Gam[mu] . up1,a24:up4b . Gam[mu] . up2,mcon(a13
*
a24,mu),
trigsimp(%%))
(%o91) -8
*
E2
(%i92) Mu:(a14:up4b . Gam[mu] . up1,a23:up3b . Gam[mu] . up2,mcon(a14
*
a23,mu),
trigsimp(%%))
(%o92) 8
*
E2
(%i93) Mfi:Mt/t-Mu/u
(%o93) -8
*
E2/u-8
*
E2/t
(%i94) " Now replace s, t, and u by explicit functions of th "
(%i95) Mfi:sub_stu(Mfi)
(%o95) -8/(cos(th)2-1)
(%i96) Mfi_RR_RR:ts(Mfi,th)
(%o96) 8/sin(th)2
(%i97) " Now automate production of HE polarized amplitudes "
(%i98) " as functions of t and u "
(%i99) " Our definition of the amplitude does not include a factor of -e2."
(%i100) he_me(sp1v,sp2v,sp3v,sp4v):=block([up1,up2,up3b,up4b,_mu%,Mt,Mu,temp],
up1:UU(E,E,0,0,sp1v),up3b:sbar(UU(E,E,th,0,sp3v)),
up2:UU(E,E,%pi,0,sp2v),up4b:sbar(UU(E,E,%pi-th,%pi,sp4v)),
a13:up3b . Gam[_mu%] . up1,a24:up4b . Gam[_mu%] . up2,
Mt:(Con(a13
*
a24,_mu%),expand(trigsimp(%%)))/t,
a14:up4b . Gam[_mu%] . up1,a23:up3b . Gam[_mu%] . up2,
Mu:(Con(a14
*
a23,_mu%),expand(trigsimp(%%)))/u,temp:Mt-Mu,
if temp # 0 then temp:pullfac(temp,-8
*
E2),temp)
(%i101) " test he_me for the case already worked out above "
(%i102) he_me(1,1,1,1)
(%o102) -8
*
(1/u+1/t)
*
E2
(%i103) sub_stu(%)
(%o103) -8/(cos(th)2-1)
41
(%i104) ts(%,th)
(%o104) 8/sin(th)2
(%i105) " which agrees with our previous work."
(%i106) " ---------------------------------------- "
(%i107) " AN ALTERNATIVE PATH TO THE UNPOLARIZED CROSS SECTION "
(%i108) " IS TO SUM THE ABSOLUTE VALUE SQUARED OF EACH "
(%i109) " OF THE POLARIZED AMPLITUDES, WHICH WE NOW DO."
(%i110) " The function Avsq(expr) is defined in dgmatrix.mac "
(%i111) " and computes the absolute value squared."
(%i112) " (We could also use here abs(expr)2 )"
(%i113) " Each nonzero amplitude is proportional to 8
*
E2
*
e2"
(%i114) " here since we have not replaced t or u in the denominators"
(%i115) " by a function of th."
(%i116) " We have already pulled out -e2 from our amplitude def"
(%i117) " Pull out also the factor 8
*
E2 temporarily here."
(%i118) block([sL,s1,s2,s3,s4,temp],sL:[1,-1],mssq:0,print(" "),
print(" svp1 svp2 svp3 svp4 amplitude "),print(" "),
for s1 in sL do
for s2 in sL do
for s3 in sL do
for s4 in sL do
(temp:expand(he_me(s1,s2,s3,s4)/(8
*
E2)),
mssq:Avsq(temp)+mssq,print(" "),
print(s1,s2,s3,s4," ",temp)),
mssq:expand(fr_ao2(mssq,th)))
svp1 svp2 svp3 svp4 amplitude
1 1 1 1 -1/u-1/t
1 1 1 -1 0
1 1 -1 1 0
1 1 -1 -1 0
1 -1 1 1 0
1 -1 1 -1 cos(th/2)2/t
1 -1 -1 1 sin(th/2)2/u
1 -1 -1 -1 0
-1 1 1 1 0
-1 1 1 -1 sin(th/2)2/u
-1 1 -1 1 cos(th/2)2/t
-1 1 -1 -1 0
-1 -1 1 1 0
-1 -1 1 -1 0
-1 -1 -1 1 0
-1 -1 -1 -1 -1/u-1/t
(%i119) " -------------------------------------- "
42
(%i120) mssq
(%o120) 4/(t
*
u)+cos(th)2/(2
*
u2)-cos(th)/u2+5/(2
*
u2)+cos(th)2/(2
*
t2)
+cos(th)/t2+5/(2
*
t2)
(%i121) mssq:sub_stu(mssq)
(%o121) cos(th)4/(4
*
sin(th)4
*
E4)+3
*
cos(th)2/(2
*
sin(th)4
*
E4)
+9/(4
*
sin(th)4
*
E4)
(%i122) mssq:factor(ratsimp(mssq))
(%o122) (cos(th)2+3)2/(4
*
sin(th)4
*
E4)
(%i123) " restore (8
*
E2)2 = 64
*
E4 to mssq "
(%i124) mssq:64
*
E4
*
mssq
(%o124) 16
*
(cos(th)2+3)2/sin(th)4
(%i125) dsigdo_unpol_CM_HE:A
*
mssq/4
(%o125) alpha2
*
(cos(th)2+3)2/(s
*
sin(th)4)
(%i126) dsigdo_unpol_CM_HE:subst(s = s_th,%)
(%o126) alpha2
*
(cos(th)2+3)2/(4
*
sin(th)4
*
E2)
(%i127) (display2d:true,display(dsigdo_unpol_CM_HE),display2d:false)
2 2 2
alpha (cos (th) + 3)
dsigdo_unpol_CM_HE = ----------------------
4 2
4 sin (th) E
(%i128) " which agrees with the symbolic calculation."
The last section of moller1.mac uses symbolic methods to nd the square of polarized amplitudes. An argument
S(1) or S(-1) inside tr is effectively interpreted as
1
2
(1 +
5
), where = 1 is the argument of S.
(%i129) " ================================ "
(%i130) " SYMBOLIC POLARIZED AMPLITUDE SQUARED "
(%i131) " -----------------------------------------"
(%i132) " Instead of summing over helicities, insert helicity "
(%i133) " projection operators appropriate to the zero mass limit "
(%i134) " CASE RR --> RR "
(%i135) M1n:trigsimp(Con(tr(S(1),p3,mu,S(1),p1,nu)
*
tr(S(1),p4,mu,S(1),p2,nu),
mu,nu))
(%o135) 64
*
E4+(8
*
t-8
*
t
*
cos(th))
*
E2+4
*
t2
(%i136) M2n:trigsimp(Con(tr(S(1),p4,mu,S(1),p1,nu)
*
tr(S(1),p3,mu,S(1),p2,nu),
mu,nu))
(%o136) 64
*
E4+(8
*
cos(th)+8)
*
u
*
E2+4
*
u2
(%i137) " NOTE AUTOMATIC PRETRACE CONTRACTION OF REPEATED "
(%i138) " LORENTZ INDICES WITHIN A SINGLE TRACE OCCURS USING tr."
(%i139) M12n:tr(S(1),p3,mu,S(1),p1,nu,S(1),p4,mu,S(1),p2,nu)
(%o139) -4
*
s2
(%i140) M21n:tr(S(1),p4,mu,S(1),p1,nu,S(1),p3,mu,S(1),p2,nu)
(%o140) -4
*
s2
(%i141) MfiSQ:(-M21n)/(t
*
u)+(-M12n)/(t
*
u)+M2n/u2+M1n/t2
(%o141) (64
*
E4+(8
*
cos(th)+8)
*
u
*
E2+4
*
u2)/u2
+(64
*
E4+(8
*
t-8
*
t
*
cos(th))
*
E2+4
*
t2)/t2+8
*
s2/(t
*
u)
(%i142) MfiSQ:sub_stu(MfiSQ)
(%o142) 64/sin(th)4
(%i143) " which is the square of the Dirac spinor amplitude found above "
(%o144) "moller1.mac"
43
12.7 moller2.mac: Arbitrary Energy Moller Scattering
The batch le moller2.mac works out the case of arbitrary energy particles in the CM frame. See the references and
kinematic notation in the previous section dealing with moller1.mac.
The rst section of moller2.mac works out the unpolarized differential cross section for arbitrary energy, rst in an
arbitrary frame in terms of s, t, u, and then in the CM frame, using symbolic methods.
(%i145) load(dirac2);
(%o1) "c:/work5/dirac2.mac"
(%i2) batch ("moller2.mac");
read and interpret file: #pc:/work5/moller2.mac
(%i3) " ======================================================"
(%i4) " file moller2.mac "
(%i5) " Maxima by Example, Ch. 12 "
(%i6) " Dirac Algebra and Quantum Electrodynamics "
(%i7) " Edwin L Woollett, woollett@charter.net "
(%i8) " http://www.csulb.edu/woollett "
(%i9) print(" ver: ",_binfo%," date: ",mydate)
ver: Maxima 5.23.2 date: 2011-04-04
(%i10) " MOLLER SCATTERING "
(%i11) " ARBITRARY ENERGY, CENTER OF MOMENTUM FRAME "
(%i12) " e(-,p1,sv1) + e(-,p2,sv2) --> e(-,p3,sv3) + e(-,p4,sv4) "
(%i13) " ------------------------------------------ "
(%i14) " SYMBOLIC TRACES FOR UNPOLARIZED DIFFERENTIAL CROSS SECTION "
(%i15) " Supply s, t, and u expressions for dot products "
(%i16) " ---------------------------------------------- "
(%i17) invar(D(p1,p1) = m2,D(p1,p2) = s/2-m2,D(p1,p3) = m2-t/2,
D(p1,p4) = m2-u/2,D(p2,p2) = m2,D(p2,p3) = m2-u/2,
D(p2,p4) = m2-t/2,D(p3,p3) = m2,D(p3,p4) = s/2-m2,
D(p4,p4) = m2)
(%i18) M1n:Con(tr(m+p3,mu,m+p1,nu)
*
tr(m+p4,mu,m+p2,nu),mu,nu)
(%o18) 8
*
u2-32
*
m2
*
u+32
*
m2
*
t+8
*
s2-32
*
m2
*
s+64
*
m4
(%i19) M2n:Con(tr(m+p4,mu,m+p1,nu)
*
tr(m+p3,mu,m+p2,nu),mu,nu)
(%o19) 32
*
m2
*
u+8
*
t2-32
*
m2
*
t+8
*
s2-32
*
m2
*
s+64
*
m4
(%i20) " NOTE AUTOMATIC PRETRACE CONTRACTION OF REPEATED "
(%i21) " LORENTZ INDICES WITHIN A SINGLE TRACE OCCURS USING tr."
(%i22) M12n:tr(m+p3,mu,m+p1,nu,m+p4,mu,m+p2,nu)
(%o22) -16
*
m2
*
u-16
*
m2
*
t-8
*
s2+48
*
m2
*
s-32
*
m4
(%i23) M21n:tr(m+p4,mu,m+p1,nu,m+p3,mu,m+p2,nu)
(%o23) -16
*
m2
*
u-16
*
m2
*
t-8
*
s2+48
*
m2
*
s-32
*
m4
(%i24) " expressed as a function of s, t, and u "
(%i25) MfiSQ:(-M21n)/(t
*
u)+(-M12n)/(t
*
u)+M2n/u2+M1n/t2
(%o25) (8
*
u2-32
*
m2
*
u+32
*
m2
*
t+8
*
s2-32
*
m2
*
s+64
*
m4)/t2
+(32
*
m2
*
u+8
*
t2-32
*
m2
*
t+8
*
s2-32
*
m2
*
s+64
*
m4)/u2
+2
*
(16
*
m2
*
u+16
*
m2
*
t+8
*
s2-48
*
m2
*
s+32
*
m4)/(t
*
u)
(%i26) " REPLACE s, t, and u WITH EXPLICIT FUNCTIONS OF th "
(%i27) " ------------------------------------------------- "
(%i28) assume(E > 0,p > 0,th >= 0,th <= %pi)
(%i29) comp_def(p1(E,0,0,p),p2(E,0,0,-p),p3(E,p
*
sin(th),0,p
*
cos(th)),
p4(E,-p
*
sin(th),0,-p
*
cos(th)))
(%i30) s_th:VP(p2+p1,p2+p1)
(%o30) 4
*
E2
(%i31) t_th:factor(VP(p1-p3,p1-p3))
(%o31) 2
*
p2
*
(cos(th)-1)
(%i32) u_th:factor(VP(p1-p4,p1-p4))
(%o32) -2
*
p2
*
(cos(th)+1)
(%i33) " ------------------------------------------------- "
(%i34) MfiSQ_th:factor(trigsimp(sub_stu(MfiSQ)))
(%o34) 16
*
(8
*
E4+2
*
m2
*
cos(th)2
*
E2-10
*
m2
*
E2+p4
*
cos(th)4+6
*
p4
*
cos(th)2
+10
*
m2
*
p2
*
cos(th)2+m4
*
cos(th)2+p4-2
*
m2
*
p2+3
*
m4)
/(p4
*
sin(th)4)
44
(%i35) " show this equals BLPs expression in Eq. (81.10)"
(%i36) MfiSQ_cmp:16
*
(p2+E2)2
*
((p2/(p2+E2))2
*
(4/sin(th)2+1)
-3/sin(th)2+4/sin(th)4)
/p4
(%i37) trigsimp(expand(MfiSQ_th-MfiSQ_cmp))
(%o37) ((48
*
sin(th)2+64)
*
E4+((96
*
p2-32
*
m2)
*
sin(th)2-128
*
p2-128
*
m2)
*
E2
+(-144
*
p4-160
*
m2
*
p2-16
*
m4)
*
sin(th)2+64
*
p4
+128
*
m2
*
p2+64
*
m4)
/(p4
*
sin(th)4)
(%i38) ratsubst(p2+m2,E2,%)
(%o38) 0
(%i39) " which confirms equality."
(%i40) A:alpha2/(4
*
s)
(%i41) A:sub_stu(A)
(%o41) alpha2/(16
*
E2)
(%i42) " Averaging over initial spins means we need to divide A by 4"
(%i43) " to get the unpolarized differential cross section "
(%i44) dsigdo_unpol_CM:A
*
MfiSQ_cmp/4
(%o44) alpha2
*
(E2+p2)2
*
(p4
*
(4/sin(th)2+1)/(E2+p2)2-3/sin(th)2+4/sin(th)4)
/(4
*
p4
*
E2)
(%i45) (display2d:true,display(dsigdo_unpol_CM),display2d:false)
4 4
p (-------- + 1)
2
2 2 2 2 sin (th) 3 4
alpha (E + p ) (----------------- - -------- + --------)
2 2 2 2 4
(E + p ) sin (th) sin (th)
dsigdo_unpol_CM = -----------------------------------------------------------
4 2
4 p E
(%i46) " which agrees with BLP, p. 323, Eq (81.10). "
(%i46) " ========================================================="
The next section of moller2.mac uses explicit Dirac spinors and matrices to calculate polarized amplitudes.
(%i47) " ========================================================="
(%i48) " POLARIZED DIRAC SPINOR AMPLITUDES "
(%i49) " ---------------------------------------------------------"
(%i50) E_pm(expr):=expand(ratsubst(m2+p2,E2,expr))
(%i51) p_Em(expr):=expand(ratsubst(E2-m2,p2,expr))
(%i52) Ep_m(expr):=expand(ratsubst(m,sqrt(E-p)
*
sqrt(p+E),expr))
(%i53) Ep_Mm(expr):=(expand(ratsubst(M2/4-m2,p2,expr)),
expand(ratsubst(M/2,E,%%)))
(%i54) " -----------------------------------------------"
(%i55) " convert t_th and u_th to th/2 forms "
(%i56) t_th2:to_ao2(t_th,th)
(%o56) -4
*
p2
*
sin(th/2)2
(%i57) u_th2:to_ao2(u_th,th)
(%o57) -4
*
p2
*
cos(th/2)2
(%i58) " dirac spinor amplitude given global s1,s2,s3,s4 "
(%i59) dA():=(
(up1:UU(E,p,0,0,s1),up3b:sbar(UU(E,p,th,0,s3)),up2:UU(E,p,%pi,0,s2),
up4b:sbar(UU(E,p,%pi-th,%pi,s4))),
Mt:(a13:up3b . Gam[mu] . up1,a24:up4b . Gam[mu] . up2,
mcon(a13
*
a24,mu)),Mt:Ep_m(Mt),M1:Mt/t_th2,
Mu:(a14:up4b . Gam[mu] . up1,a23:up3b . Gam[mu] . up2,
mcon(a14
*
a23,mu)),Mu:Ep_m(Mu),M2:Mu/u_th2,Mfi:M1-M2)
(%i60) " Print a table of polarized amplitudes."
45
(%i61) " Accumulate sum of squares mssq."
(%i62) block([sL,sv1,sv2,sv3,sv4,temp],sL:[1,-1],mssq:0,print(" "),
print(" svp1 svp2 svp3 svp4 amplitude "),print(" "),
for sv1 in sL do
for sv2 in sL do
for sv3 in sL do
for sv4 in sL do
([s1,s2,s3,s4]:[sv1,sv2,sv3,sv4],temp:dA(),
temp:E_pm(temp),mssq:Avsq(temp)+mssq,
print(" "),print(s1,s2,s3,s4," ",temp)),
mssq:E_pm(mssq),mssq:expand(fr_ao2(mssq,th)))
svp1 svp2 svp3 svp4 amplitude
1 1 1 1 m2
*
sin(th/2)2/(p2
*
cos(th/2)2)+2
*
sin(th/2)2/cos(th/2)2
+m2
*
cos(th/2)2/(p2
*
sin(th/2)2)
+2
*
cos(th/2)2/sin(th/2)2+4
1 1 1 -1 m
*
sin(th/2)
*
E/(p2
*
cos(th/2))-m
*
cos(th/2)
*
E/(p2
*
sin(th/2))
1 1 -1 1 m
*
sin(th/2)
*
E/(p2
*
cos(th/2))-m
*
cos(th/2)
*
E/(p2
*
sin(th/2))
1 1 -1 -1 2
*
m2/p2
1 -1 1 1 m
*
cos(th/2)
*
E/(p2
*
sin(th/2))-m
*
sin(th/2)
*
E/(p2
*
cos(th/2))
1 -1 1 -1 m2
*
cos(th/2)2/(p2
*
sin(th/2)2)+2
*
cos(th/2)2/sin(th/2)2-m2/p2
1 -1 -1 1 m2
*
sin(th/2)2/(p2
*
cos(th/2)2)+2
*
sin(th/2)2/cos(th/2)2-m2/p2
1 -1 -1 -1 m
*
sin(th/2)
*
E/(p2
*
cos(th/2))-m
*
cos(th/2)
*
E/(p2
*
sin(th/2))
-1 1 1 1 m
*
cos(th/2)
*
E/(p2
*
sin(th/2))-m
*
sin(th/2)
*
E/(p2
*
cos(th/2))
-1 1 1 -1 m2
*
sin(th/2)2/(p2
*
cos(th/2)2)+2
*
sin(th/2)2/cos(th/2)2-m2/p2
-1 1 -1 1 m2
*
cos(th/2)2/(p2
*
sin(th/2)2)+2
*
cos(th/2)2/sin(th/2)2-m2/p2
-1 1 -1 -1 m
*
sin(th/2)
*
E/(p2
*
cos(th/2))-m
*
cos(th/2)
*
E/(p2
*
sin(th/2))
-1 -1 1 1 2
*
m2/p2
-1 -1 1 -1 m
*
cos(th/2)
*
E/(p2
*
sin(th/2))-m
*
sin(th/2)
*
E/(p2
*
cos(th/2))
-1 -1 -1 1 m
*
cos(th/2)
*
E/(p2
*
sin(th/2))-m
*
sin(th/2)
*
E/(p2
*
cos(th/2))
-1 -1 -1 -1 m2
*
sin(th/2)2/(p2
*
cos(th/2)2)+2
*
sin(th/2)2/cos(th/2)2
+m2
*
cos(th/2)2/(p2
*
sin(th/2)2)
+2
*
cos(th/2)2/sin(th/2)2+4
(%i63) mssq
(%o63) -192
*
m2/(p2
*
sin(th)2)-48
*
m4/(p4
*
sin(th)2)-128/sin(th)2
+256
*
m2/(p2
*
sin(th)4)+64
*
m4/(p4
*
sin(th)4)
+256/sin(th)4+16
(%i64) " SHOW THIS IS THE SAME AS MfiSQ_th computed above with traces "
(%i65) MfiSQ_p:E_pm(MfiSQ_th)
(%o65) 16
*
cos(th)4/sin(th)4+192
*
m2
*
cos(th)2/(p2
*
sin(th)4)
+48
*
m4
*
cos(th)2/(p4
*
sin(th)4)
+96
*
cos(th)2/sin(th)4+64
*
m2/(p2
*
sin(th)4)
+16
*
m4/(p4
*
sin(th)4)+144/sin(th)4
(%i66) trigsimp(mssq-MfiSQ_p)
(%o66) 0
46
(%i67) " which shows equality."
(%i68) " ========================================================="
(%o68) "moller2.mac"
12.8 bhabha1.mac: High Energy Limit of Bhabha Scattering
This batch le treats the high energy limit of the scattering event
e

(p
1
,
1
) + e
+
(p
2
,
2
) e

(p
3
,
3
) + e
+
(p
4
,
4
). (12.47)
This limit is that in which the masses of the particles can be ignored (|p| >> m). In practice this means treating the
leptons as massless particles.
Some references are: H.J. Bhabha, 1936, BLP QED, Sec. 81, B/D RQM, Sec. 7.9, G/R QED, Sec. 3.4, Renton EI, Sec.
4.3, and Schwinger, PSF I, p. 306 - 309.
The invariant amplitude M = M
1
M
2
has two terms. Using the denitions
s = (p
1
+ p
2
)
2
, t = (p
1
p
3
)
2
, u = (p
1
p
4
)
2
, (12.48)
and the abbreviations
u
1
= u(p
1
,
1
), v
4
= v(p
4
,
4
), u
3
= u(p
3
,
3
), v
2
= v(p
2
,
2
) (12.49)
we have
M
1
=
e
2
u
3

u
1
v
2

v
4
t
, (12.50)
and
M
2
=
e
2
v
2

u
1
u
3

v
4
s
. (12.51)
The batch le bhabha1.mac uses the Dirac package functions pullfac, VP, to_ao2, Avsq, and fr_ao2.
The rst section of bhabha1.mac derives the unpolarized differential cross section, rst in an arbitrary frame in terms
of s, t, u, and then in the CM frame in terms of the scattering angle th, using symbolic methods.
(%i69) load(dirac2);
(%o1) "c:/work5/dirac2.mac"
(%i2) batch ("bhabha1.mac");
read and interpret file: #pc:/work5/bhabha1.mac
(%i3) " ======================================================"
(%i4) " file bhabha1.mac "
(%i5) " Maxima by Example, Ch. 12 "
(%i6) " Dirac Algebra and Quantum Electrodynamics "
(%i7) " Edwin L Woollett, woollett@charter.net "
(%i8) " http://www.csulb.edu/woollett "
(%i9) print(" ver: ",_binfo%," date: ",mydate)
ver: Maxima 5.23.2 date: 2011-04-04
(%i10) " BHABHA SCATTERING "
(%i11) " HIGH ENERGY LIMIT, CENTER OF MOMENTUM FRAME, NEGLECT MASSES "
(%i12) " e(-,p1,s1) + e(+,p2,s2) --> e(-,p3,s3) + e(+,p4,s4) "
(%i13) " m = electron and positron mass is set to zero."
(%i14) " ------------------------------------------ "
47
(%i15) "NON-POLARIZED DIFFERENTIAL CROSS SECTION: SYMBOLIC METHODS"
(%i16) " POPULATE THE LIST invarR OF 4-VEC DOT PRODUCT VALUES, "
(%i17) " Using p1 + p2 = p3 + p4, s = (p1+p2)2 = (p3+p4)2 ,"
(%i18) " t = (p1-p3)2 = (p2-p4)2,"
(%i19) " and u = (p1-p4)2 = (p2-p3)2 "
(%i20) " CASE HIGH ENERGY (HE) LIMIT E >> m "
(%i21) " ---------------------------------------------"
(%i22) invar(D(p1,p1) = 0,D(p1,p2) = s/2,D(p1,p3) = (-t)/2,D(p1,p4) = (-u)/2,
D(p2,p2) = 0,D(p2,p3) = (-u)/2,D(p2,p4) = (-t)/2,D(p3,p3) = 0,
D(p3,p4) = s/2,D(p4,p4) = 0)
(%i23) "------------------------------------------------------"
(%i24) " With a sum over all helicities implied,"
(%i25) " |Mfi|2 = M1n/t2 + M2n/s2 -M12n/(t
*
s) - M21n/(t
*
s) "
(%i26) " M1n = t2
*
M1
*
conj(M1), M2n = s2
*
M2
*
conj(M2) "
(%i27) " M12n = (t
*
s)
*
M1
*
conj(M2), M21n = (t
*
s)
*
M2
*
conj(M1), and: "
(%i28) M1n:factor(Con(tr(p3,mu,p1,nu)
*
tr(p2,mu,p4,nu),mu,nu))
(%o28) 8
*
(u2+s2)
(%i29) M2n:factor(Con(tr(p2,mu,p1,nu)
*
tr(p3,mu,p4,nu),mu,nu))
(%o29) 8
*
(u2+t2)
(%i30) " NOTE AUTOMATIC PRETRACE CONTRACTION OF REPEATED "
(%i31) " LORENTZ INDICES WITHIN A SINGLE TRACE OCCURS USING tr."
(%i32) M12n:factor(tr(p3,mu,p1,nu,p2,mu,p4,nu))
(%o32) -8
*
u2
(%i33) M21n:factor(tr(p2,mu,p1,nu,p3,mu,p4,nu))
(%o33) -8
*
u2
(%i34) MfiSQ:pullfac((-M21n)/(t
*
s)+(-M12n)/(t
*
s)+M2n/s2+M1n/t2,8)
(%o34) 8
*
((u2+t2)/s2+(u2+s2)/t2+2
*
u2/(s
*
t))
(%i35) " We have absorbed e4 into A, with e2 = 4
*
%pi
*
alpha "
(%i36) " Averaging over initial spins means we need to divide A by 4"
(%i37) " to get the unpolarized differential cross section (CM, HE)"
(%i38) A:alpha2/(4
*
s)
(%i39) dsigdo_unpol_CM_HE:A
*
MfiSQ/4
(%o39) alpha2
*
((u2+t2)/s2+(u2+s2)/t2+2
*
u2/(s
*
t))/(2
*
s)
(%i40) (display2d:true,display(dsigdo_unpol_CM_HE),display2d:false)
2 2 2 2 2
2 u + t u + s 2 u
alpha (------- + ------- + ----)
2 2 s t
s t
dsigdo_unpol_CM_HE = ---------------------------------
2 s
(%i41) " which agrees with Rentons function of s,t, and u "
(%i42) " on page 159, Eq.(4.54) "
(%i43) " CONVERSION TO EXPLICIT FUNCTION OF SCATTERING ANGLE "
(%i44) assume(E > 0,th >= 0,th <= %pi)
(%i45) comp_def(p1(E,0,0,E),p2(E,0,0,-E),p3(E,E
*
sin(th),0,E
*
cos(th)),
p4(E,-E
*
sin(th),0,-E
*
cos(th)))
(%i46) s_th:VP(p2+p1,p2+p1)
(%o46) 4
*
E2
(%i47) t_th:factor(VP(p1-p3,p1-p3))
(%o47) 2
*
(cos(th)-1)
*
E2
(%i48) u_th:factor(VP(p1-p4,p1-p4))
(%o48) -2
*
(cos(th)+1)
*
E2
(%i49) " CONVERT FROM s, t, u TO th form "
(%i50) "--------------------------------------------------"
(%i51) MfiSQ_th:(sub_stu(MfiSQ),factor(%%))
(%o51) 4
*
(cos(th)2+3)2/(cos(th)-1)2
(%i52) A_th:subst(s = s_th,A)
(%o52) alpha2/(16
*
E2)
(%i53) dsigdo_unpol_th:A_th
*
MfiSQ_th/4
(%o53) alpha2
*
(cos(th)2+3)2/(16
*
(cos(th)-1)2
*
E2)
48
(%i54) (display2d:true,display(dsigdo_unpol_th),display2d:false)
2 2 2
alpha (cos (th) + 3)
dsigdo_unpol_th = ----------------------
2 2
16 (cos(th) - 1) E
(%i55) "------------------------------------------------------------"
(%i56) " which agrees with Renton, p. 160 "
The next section of bhabha1.mac uses explicit Dirac spinors and matrices to compute the polarized amplitudes.
(%i57) " HIGH ENERGY POLARIZED AMPLITUDES USING EXPLICIT DIRAC SPINORS "
(%i58) " ------------------------------------------------- "
(%i59) t_th2:to_ao2(t_th,th)
(%o59) -4
*
sin(th/2)2
*
E2
(%i60) " dirac spinor amplitude given global s1,s2,s3,s4 "
(%i61) dA():=(
(up1:UU(E,E,0,0,s1),up3b:sbar(UU(E,E,th,0,s3)),
vp2b:sbar(VV(E,E,%pi,0,s2)),vp4:VV(E,E,%pi-th,%pi,s4)),
Mt:(a13:up3b . Gam[_mu%] . up1,a42:vp2b . Gam[_mu%] . vp4,
mcon(a13
*
a42,_mu%),expand(trigsimp(%%))),M1:Mt/t_th2,
Ms:(a12:vp2b . Gam[_mu%] . up1,a43:up3b . Gam[_mu%] . vp4,
mcon(a12
*
a43,_mu%),expand(trigsimp(%%))),M2:Ms/s_th,M1-M2)
(%i62) " example: RR --> RR "
(%i63) [s1,s2,s3,s4]:[1,1,1,1]
(%i64) dA()
(%o64) 2/sin(th/2)2
(%i65) " Make table of polarized amplitudes. "
(%i66) " Accumulate sum mssq of square of amplitudes."
(%i67) block([sL,sv1,sv2,sv3,sv4,temp],sL:[1,-1],mssq:0,print(" "),
print(" s1 s2 s3 s4 amplitude "),print(" "),
for sv1 in sL do
for sv2 in sL do
for sv3 in sL do
for sv4 in sL do
([s1,s2,s3,s4]:[sv1,sv2,sv3,sv4],temp:dA(),
mssq:Avsq(temp)+mssq,print(" "),
print(s1,s2,s3,s4," ",temp)),
mssq:expand(fr_ao2(mssq,th)))
s1 s2 s3 s4 amplitude
1 1 1 1 2/sin(th/2)2
1 1 1 -1 0
1 1 -1 1 0
1 1 -1 -1 0
1 -1 1 1 0
1 -1 1 -1 2
*
cos(th/2)2/sin(th/2)2-2
*
cos(th/2)2
1 -1 -1 1 -2
*
sin(th/2)2
1 -1 -1 -1 0
-1 1 1 1 0
49
-1 1 1 -1 -2
*
sin(th/2)2
-1 1 -1 1 2
*
cos(th/2)2/sin(th/2)2-2
*
cos(th/2)2
-1 1 -1 -1 0
-1 -1 1 1 0
-1 -1 1 -1 0
-1 -1 -1 1 0
-1 -1 -1 -1 2/sin(th/2)2
(%i68) " Sum of squares of polarized amplitudes:"
(%i69) mssq
(%o69) 4
*
cos(th)4/(cos(th)2-2
*
cos(th)+1)
+24
*
cos(th)2/(cos(th)2-2
*
cos(th)+1)+36/(cos(th)2-2
*
cos(th)+1)
(%i70) " COMPARE WITH SYMBOLIC RESULT MfiSQ_th CALCULATED ABOVE"
(%i71) trigsimp(mssq-MfiSQ_th)
(%o71) 0
(%i72) " WHICH SHOWS THEY ARE EQUIVALENT."
(%o72) "bhabha1.mac"
12.9 bhabha2.mac: Arbitrary Energy Bhabha Scattering
The batch le bhabha2.mac works out bhabha scattering for arbitrary particle energy in the CM frame. See the references
and kinematic notation in the previous section dealing with bhabha1.mac. The batch le bhabha1.mac uses the
Dirac package functions pullfac, VP, sub_stu, to_ao2, Avsq, and fr_ao2.
In addition, some kinematic substitution functions are dened in this batch le
The function E_pm(expr) makes the replacement E2 --> m2 + p2,
The function p_Em (expr) makes the replacement p2 --> E2 - m2\mv,
The function Ep_m (expr) makes the replacement sqrt (E-p)
*
sqrt (E+p) --> m,
The function Ep_Mm (expr) makes the two replacements: 1. E --> M/2, 2. p2 --> M2/4 - m2,
where M is the total center of momentum frame energy.
The rst section of bhabha2.mac calculates the differential scattering cross section in the CM frame using symbolic
methods.
(%i73) load(dirac2);
(%o1) "c:/work5/dirac2.mac"
(%i2) batch ("bhabha2.mac");
read and interpret file: #pc:/work5/bhabha2.mac
(%i3) " ======================================================"
(%i4) " file bhabha2.mac "
(%i5) " Maxima by Example, Ch. 12 "
(%i6) " Dirac Algebra and Quantum Electrodynamics "
(%i7) " Edwin L Woollett, woollett@charter.net "
(%i8) " http://www.csulb.edu/woollett "
(%i9) print(" ver: ",_binfo%," date: ",mydate)
ver: Maxima 5.23.2 date: 2011-04-04
(%i10) " BHABHA SCATTERING "
(%i11) " Finite mass, Arbitrary Energy, CENTER OF MOMENTUM FRAME "
(%i12) " e(-,p1,s1) + e(+,p2,s2) --> e(-,p3,s3) + e(+,p4,s4) "
50
(%i13) " ------------------------------------------ "
(%i14) " SYMBOLIC TRACES FOR UNPOLARIZED DIFFERENTIAL CROSS SECTION "
(%i15) " Supply s, t, and u expressions for dot products "
(%i16) " ---------------------------------------------- "
(%i17) invar(D(p1,p1) = m2,D(p1,p2) = s/2-m2,D(p1,p3) = m2-t/2,
D(p1,p4) = m2-u/2,D(p2,p2) = m2,D(p2,p3) = m2-u/2,
D(p2,p4) = m2-t/2,D(p3,p3) = m2,D(p3,p4) = s/2-m2,
D(p4,p4) = m2)
(%i18) " With a sum over all helicities implied,"
(%i19) " |Mfi|2 = M1n/t2 + M2n/s2 -M12n/(t
*
s) - M21n/(t
*
s) "
(%i20) " M1n = t2
*
M1
*
conj(M1), M2n = s2
*
M2
*
conj(M2) "
(%i21) " M12n = (t
*
s)
*
M1
*
conj(M2), M21n = (t
*
s)
*
M2
*
conj(M1), and: "
(%i22) M1n:Con(tr(m+p3,mu,m+p1,nu)
*
tr(p2-m,mu,p4-m,nu),mu,nu)
(%o22) 8
*
u2-32
*
m2
*
u+32
*
m2
*
t+8
*
s2-32
*
m2
*
s+64
*
m4
(%i23) M2n:Con(tr(p2-m,mu,m+p1,nu)
*
tr(m+p3,mu,p4-m,nu),mu,nu)
(%o23) 8
*
u2-32
*
m2
*
u+8
*
t2-32
*
m2
*
t+32
*
m2
*
s+64
*
m4
(%i24) " NOTE AUTOMATIC PRETRACE CONTRACTION OF REPEATED "
(%i25) " LORENTZ INDICES WITHIN A SINGLE TRACE OCCURS USING tr."
(%i26) M12n:tr(m+p3,mu,m+p1,nu,p2-m,mu,p4-m,nu)
(%o26) -8
*
u2+48
*
m2
*
u-16
*
m2
*
t-16
*
m2
*
s-32
*
m4
(%i27) M21n:tr(p2-m,mu,m+p1,nu,m+p3,mu,p4-m,nu)
(%o27) -8
*
u2+48
*
m2
*
u-16
*
m2
*
t-16
*
m2
*
s-32
*
m4
(%i28) " sum over spins of |Mfi|2 in terms of s,t,u "
(%i29) MfiSQ:(-M21n)/(t
*
s)+(-M12n)/(t
*
s)+M2n/s2+M1n/t2
(%o29) (8
*
u2-32
*
m2
*
u+8
*
t2-32
*
m2
*
t+32
*
m2
*
s+64
*
m4)/s2
+(8
*
u2-32
*
m2
*
u+32
*
m2
*
t+8
*
s2-32
*
m2
*
s+64
*
m4)/t2
+2
*
(8
*
u2-48
*
m2
*
u+16
*
m2
*
t+16
*
m2
*
s+32
*
m4)/(s
*
t)
(%i30) " replace s, t, u in terms of scatt. angle th "
(%i31) assume(E > 0,p > 0,th >= 0,th <= %pi)
(%i32) comp_def(p1(E,0,0,p),p2(E,0,0,-p),p3(E,p
*
sin(th),0,p
*
cos(th)),
p4(E,-p
*
sin(th),0,-p
*
cos(th)))
(%i33) E_pm(expr):=expand(ratsubst(m2+p2,E2,expr))
(%i34) s_th:VP(p2+p1,p2+p1)
(%o34) 4
*
E2
(%i35) t_th:factor(VP(p1-p3,p1-p3))
(%o35) 2
*
p2
*
(cos(th)-1)
(%i36) u_th:factor(VP(p1-p4,p1-p4))
(%o36) -2
*
p2
*
(cos(th)+1)
(%i37) " ------------------------------------------------- "
(%i38) MfiSQ_th:sub_stu(MfiSQ)
(%i39) " replace E2 by m2 + p2 "
(%i40) MfiSQ_p:E_pm(MfiSQ_th)
(%i41) MfiSQ_p:trigsimp(MfiSQ_p)
(%o41) (4
*
p8
*
sin(th)4+(-8
*
m2
*
p6
*
cos(th)-32
*
p8-72
*
m2
*
p6-52
*
m4
*
p4)
*
sin(th)2+(32
*
m2
*
p6+80
*
m4
*
p4+56
*
m6
*
p2)
*
cos(th)
+64
*
p8+160
*
m2
*
p6+128
*
m4
*
p4+40
*
m6
*
p2+16
*
m8)
/((p8+2
*
m2
*
p6+m4
*
p4)
*
cos(th)2
+(-2
*
p8-4
*
m2
*
p6-2
*
m4
*
p4)
*
cos(th)+p8+2
*
m2
*
p6+m4
*
p4)
(%i42) " ------------------------------------------------"
(%i43) " compare to Greiner, Reinhardt expression p. 148 "
(%i44) GR1:(2
*
p4
*
(cos(th/2)4+1)+4
*
p2
*
m2
*
cos(th/2)2+m4)/(p4
*
sin(th/2)4)
(%o44) (2
*
p4
*
(cos(th/2)4+1)+4
*
m2
*
p2
*
cos(th/2)2+m4)/(p4
*
sin(th/2)4)
(%i45) GR2:(p4
*
(cos(th)2+1)+4
*
p2
*
m2+3
*
m4)/E4
(%o45) (p4
*
(cos(th)2+1)+4
*
m2
*
p2+3
*
m4)/E4
(%i46) GR3:(-(4
*
p4
*
cos(th/2)4+8
*
p2
*
m2
*
cos(th/2)2+3
*
m4))
/(E2
*
p2
*
sin(th/2)2)
(%o46) (-4
*
p4
*
cos(th/2)4-8
*
m2
*
p2
*
cos(th/2)2-3
*
m4)/(p2
*
sin(th/2)2
*
E2)
(%i47) " replace th/2 and E2 for comparison "
(%i48) GR1_r:fr_ao2(GR1,th)
(%i49) GR2_r:E_pm(GR2)
(%i50) GR3_r:E_pm(fr_ao2(GR3,th))
(%i51) GR:trigsimp(expand(GR3_r+GR2_r+GR1_r))
51
(%i52) trigsimp(GR-MfiSQ_p/4)
(%o52) 0
(%i53) " which shows the equivalence we need "
(%i54) " dividing MfiSQ by 4 comes from averaging over"
(%i55) " the initial helicities "
(%i56) " We have absorbed e4 into A, with e2 = 4
*
%pi
*
alpha "
(%i57) " to get the unpolarized differential cross section (CM)"
(%i58) A:alpha2/(4
*
s_th)
(%i59) GR_expr:GR3+GR2+GR1
(%i60) dsigdo_unpol_CM:A
*
GR_expr
(%i61) (display2d:true,display(dsigdo_unpol_CM),display2d:false)
4 4 th 2 2 2 th 4
- 4 p cos (--) - 8 m p cos (--) - 3 m
2 2 2
dsigdo_unpol_CM = (alpha (-----------------------------------------
2 2 th 2
p sin (--) E
2
4 2 2 2 4
p (cos (th) + 1) + 4 m p + 3 m
+ ----------------------------------
4
E
4 4 th 2 2 2 th 4
2 p (cos (--) + 1) + 4 m p cos (--) + m
2 2 2
+ -------------------------------------------))/(16 E )
4 4 th
p sin (--)
2
(%i62) " which agrees with G/R, Exer. 3.8, eqn. 21, p. 148 "
The next section of bhabha2.mac uses explicit Dirac spinors and matrices to compute polarized amplitudes.
(%i63) " =========================================================== "
(%i64) " POLARIZED DIRAC SPINOR AMPLITUDES "
(%i65) " ---------------------------------------------------------"
(%i66) p_Em(expr):=expand(ratsubst(E2-m2,p2,expr))
(%i67) Ep_m(expr):=expand(ratsubst(m,sqrt(E-p)
*
sqrt(p+E),expr))
(%i68) Ep_Mm(expr):=(expand(ratsubst(M2/4-m2,p2,expr)),
expand(ratsubst(M/2,E,%%)))
(%i69) " -----------------------------------------------"
(%i70) " convert t_th to th/2 form "
(%i71) t_th2:to_ao2(t_th,th)
(%o71) -4
*
p2
*
sin(th/2)2
(%i72) " dirac spinor amplitude given global s1,s2,s3,s4 "
(%i73) dA():=(
(up1:UU(E,p,0,0,s1),up3b:sbar(UU(E,p,th,0,s3)),
vp2b:sbar(VV(E,p,%pi,0,s2)),vp4:VV(E,p,%pi-th,%pi,s4)),
Mt:(a13:up3b . Gam[mu] . up1,a42:vp2b . Gam[mu] . vp4,
mcon(a13
*
a42,mu),Ep_m(%%),E_pm(%%),trigsimp(%%)),M1:Mt/t_th2,
Ms:(a12:vp2b . Gam[mu] . up1,a43:up3b . Gam[mu] . vp4,
mcon(a12
*
a43,mu),Ep_m(%%),E_pm(%%),trigsimp(%%)),M2:Ms/s_th,
M2-M1)
(%i74) " example: RR --> RR "
(%i75) [s1,s2,s3,s4]:[1,1,1,1]
(%o75) [1,1,1,1]
(%i76) dA()
(%o76) (4
*
m2-8
*
m2
*
sin(th/2)2)/(4
*
E2)
+(-4
*
m2
*
cos(th/2)2-8
*
p2)/(4
*
p2
*
sin(th/2)2)
52
(%i77) " =========================================================== "
(%i78) " table of polarized amplitudes generates global mssq "
(%i79) block([sL,sv1,sv2,sv3,sv4,temp],sL:[1,-1],mssq:0,print(" "),
print(" s1 s2 s3 s4 amplitude "),print(" "),
for sv1 in sL do
for sv2 in sL do
for sv3 in sL do
for sv4 in sL do
([s1,s2,s3,s4]:[sv1,sv2,sv3,sv4],temp:dA(),
print(" "),print(s1,s2,s3,s4," ",temp),
temp:E_pm(temp),mssq:Avsq(temp)+mssq),
mssq:E_pm(mssq),mssq:expand(fr_ao2(mssq,th)),mssq:trigsimp(mssq))
s1 s2 s3 s4 amplitude
1 1 1 1 (4
*
m2-8
*
m2
*
sin(th/2)2)/(4
*
E2)+(-4
*
m2
*
cos(th/2)2-8
*
p2)
/(4
*
p2
*
sin(th/2)2)
1 1 1 -1 2
*
m
*
cos(th/2)
*
sin(th/2)/E-m
*
cos(th/2)
*
E/(p2
*
sin(th/2))
1 1 -1 1 m
*
cos(th/2)
*
E/(p2
*
sin(th/2))-2
*
m
*
cos(th/2)
*
sin(th/2)/E
1 1 -1 -1 (4
*
m2-8
*
m2
*
sin(th/2)2)/(4
*
E2)+m2/p2
1 -1 1 1 m
*
cos(th/2)
*
E/(p2
*
sin(th/2))-2
*
m
*
cos(th/2)
*
sin(th/2)/E
1 -1 1 -1 (8
*
p2+8
*
m2)
*
cos(th/2)2/(4
*
E2)+(-8
*
p2-4
*
m2)
*
cos(th/2)2
/(4
*
p2
*
sin(th/2)2)
1 -1 -1 1 (8
*
p2+8
*
m2)
*
sin(th/2)2/(4
*
E2)-m2/p2
1 -1 -1 -1 m
*
cos(th/2)
*
E/(p2
*
sin(th/2))-2
*
m
*
cos(th/2)
*
sin(th/2)/E
-1 1 1 1 2
*
m
*
cos(th/2)
*
sin(th/2)/E-m
*
cos(th/2)
*
E/(p2
*
sin(th/2))
-1 1 1 -1 (8
*
p2+8
*
m2)
*
sin(th/2)2/(4
*
E2)-m2/p2
-1 1 -1 1 (8
*
p2+8
*
m2)
*
cos(th/2)2/(4
*
E2)+(-8
*
p2-4
*
m2)
*
cos(th/2)2
/(4
*
p2
*
sin(th/2)2)
-1 1 -1 -1 2
*
m
*
cos(th/2)
*
sin(th/2)/E-m
*
cos(th/2)
*
E/(p2
*
sin(th/2))
-1 -1 1 1 (4
*
m2-8
*
m2
*
sin(th/2)2)/(4
*
E2)+m2/p2
-1 -1 1 -1 2
*
m
*
cos(th/2)
*
sin(th/2)/E-m
*
cos(th/2)
*
E/(p2
*
sin(th/2))
-1 -1 -1 1 m
*
cos(th/2)
*
E/(p2
*
sin(th/2))-2
*
m
*
cos(th/2)
*
sin(th/2)/E
-1 -1 -1 -1 (4
*
m2-8
*
m2
*
sin(th/2)2)/(4
*
E2)+(-4
*
m2
*
cos(th/2)2-8
*
p2)
/(4
*
p2
*
sin(th/2)2)
(%i80) "------------------------------------------"
(%i81) " sum of squares of polarized amplitudes:"
(%i82) mssq
(%o82) (4
*
p8
*
sin(th)4+(-8
*
m2
*
p6
*
cos(th)-32
*
p8-72
*
m2
*
p6-52
*
m4
*
p4)
*
sin(th)2+(32
*
m2
*
p6+80
*
m4
*
p4+56
*
m6
*
p2)
*
cos(th)
+64
*
p8+160
*
m2
*
p6+128
*
m4
*
p4+40
*
m6
*
p2+16
*
m8)
/((p8+2
*
m2
*
p6+m4
*
p4)
*
cos(th)2
+(-2
*
p8-4
*
m2
*
p6-2
*
m4
*
p4)
*
cos(th)+p8+2
*
m2
*
p6+m4
*
p4)
(%i83) " SHOW THIS IS THE SAME AS MfiSQ_th computed above with traces "
(%i84) MfiSQ_p:E_pm(MfiSQ_th)
53
(%i85) trigsimp(mssq-MfiSQ_p)
(%o85) 0
(%i86) " which shows equality."
(%i87) " =========================================================== "
(%o88) "bhabha2.mac"
12.10 photon1.mac: Photon Transverse Polarization 3-Vector Sums
In this section we use Maxima to work out some well known properties of photon polarization 3-vectors corresponding to
physical (external) photons, using the batch le photon1.mac. Some references to this subject are: G/R QED p. 187,
Weinberg I p. 352, 360, 368, Renton p. 166, and Gross p. 315.
Let e
ks
be a real photon polarization 3-vector which is transverse to the initial photon 3-momentumk, with s = 1 being
parallel to the scattering plane, and s = 2 being perpendicular to the scattering plane.
Let e
k

r
be a real photon polarization 3-vector which is transverse to the nal photon 3-momentum k

with r = 1 being
parallel to the scattering plane, and r = 2 being perpendicular to the scattering plane. Then we have the relations
2

s=1
(e
ks
e
k

r
)
2
= 1
_

k e
k

r
_
2
(12.52)
2

s,r=1
(e
ks
e
k

r
)
2
= 1 +cos
2
(12.53)
Here we are using real polarization 3-vectors to describe states of linear polarization.
(%i89) load(dirac2);
(%o1) "c:/work5/dirac2.mac"
(%i2) batch ("photon1.mac");
read and interpret file: #pc:/work5/photon1.mac
(%i3) display2d:false
(%i4) " ======================================================"
(%i5) " file photon1.mac "
(%i6) " Maxima by Example, Ch. 12 "
(%i7) " Dirac Algebra and Quantum Electrodynamics "
(%i8) " Edwin L Woollett, woollett@charter.net "
(%i9) " http://www.csulb.edu/woollett "
(%i10) print(" ver: ",_binfo%," date: ",mydate)
ver: Maxima 5.23.2 date: 2011-04-04
(%i11) "-----------------------------------------------------"
(%i12) " PROPERTIES AND SUMS OF TRANSVERSE POLARIZATION VECTORS "
(%i13) " ----------------------------------------------------- "
(%i14) " define 3-vectors as lists "
(%i15) " scattering plane is z-x plane "
(%i16) k_vec:[0,0,k]
(%i17) kp_vec:[kp
*
sin(th),0,kp
*
cos(th)]
(%i18) " e[1] X e[2] = k_vec/k "
(%i19) e[1]:[1,0,0]
(%i20) e[1] . e[1]
(%o20) 1
(%i21) e[1] . k_vec
(%o21) 0
(%i22) e[2]:[0,1,0]
54
(%i23) e[2] . e[2]
(%o23) 1
(%i24) e[2] . e[1]
(%o24) 0
(%i25) e[2] . k_vec
(%o25) 0
(%i26) " ep[1] X ep[2] = kp_vec/kp "
(%i27) " parallel to the scattering plane "
(%i28) ep[1]:[cos(th),0,-sin(th)]
(%i29) trigsimp(ep[1] . ep[1])
(%o29) 1
(%i30) ep[1] . kp_vec
(%o30) 0
(%i31) " perpendicular to the scattering plane "
(%i32) ep[2]:[0,1,0]
(%i33) ep[2] . ep[2]
(%o33) 1
(%i34) ep[2] . ep[1]
(%o34) 0
(%i35) ep[2] . kp_vec
(%o35) 0
(%i36) "============================================="
(%i37) " photon transverse polarization sum over both leads to "
(%i38) " (1 + cos(th)2 ) "
(%i39) sum(sum((e[s] . ep[r])2,s,1,2),r,1,2)
(%o39) cos(th)2+1
(%i40) "============================================="
(%i41) " Sum over only e[1] and e[2] values leads to "
(%i42) " sum ( (e[s] . ep[r])2,s,1,2) = 1 - (ku . ep[r])2 "
(%i43) " where ku is unit vector along k_vec "
(%i44) ku:k_vec/k
(%o44) [0,0,1]
(%i45) " first for r = 1 "
(%i46) sum((e[s] . ep[1])2,s,1,2)
(%o46) cos(th)2
(%i47) trigsimp(1-(ku . ep[1])2)
(%o47) cos(th)2
(%i48) " next for r = 2 "
(%i49) sum((e[s] . ep[2])2,s,1,2)
(%o49) 1
(%i50) 1-(ku . ep[2])2
(%o50) 1
(%i51) "============================================="
(%o51) "photon1.mac"
12.11 Covariant Physical External Photon Polarization 4-Vectors - A Review
We will use the following relation for covariant physical photon polarization 4-vectors which are 4-orthogonal to the
photon 4-momentum vector. This formalism (Sterman, p. 220, De Wit/Smith, p.141, Schwinger, p. 73 and pp. 291 - 294,
Jauch/Rohrlich, p. 440) uses gauge freedom within the Lorentz gauge.
The two physical polarization states of real external photons corresond to = 1, 2, and for , = 0, 1, 2, 3,
P

(k) =

=1,2
e

k
e

k
= g

+
(k

+ k

)
k

k
, (12.54)
where there exists a reference frame in which if k

= (k
0
, k) then

k

= (k
0
, k).
55
Let n

be a unit timelike 4-vector which is 4-orthogonal to e

k
. Then dene

= k

+ 2 n

(k n). (12.55)
In an arbitrary frame, k k = 0,

k

k = 0, k e
k
= 0,

k e
k
= 0, n e
k
= 0, n n = +1, k

k = 2 (k n)
2
and
e

k
e
k
=

, ,

= 1, 2 (12.56)
An alternative but less compact form of the physical photon polarization tensor P

(k) is
P

(k) =

=1,2
e

k
e

k
= g

(k n)
2
+
(k

+ k

)
k n
(12.57)
We can always identify the unit timelike 4-vector n

with either the total 4-momentum vector (thus working in the CM

frame) or the 4-momentum of a particular particle (thus working in the rest frame of that particle).
Thus if in a chosen frame, {p

} = {m, 0, 0, 0}, we identify n

= p

/m, so that {n

} = {1, 0, 0, 0}, then in that

frame

k
0
= k
0
and

k
j
= k
j
.
Then n e
k
= 0 implies that e
0
k
= 0, and then k e
k
= 0 reduces to k e
k
= 0 (for = 1, 2), and the two
polarization 3-vectors e
k
must be chosen perpendicular to k and to each other: e

k 1
e
k 2
= 0.
12.12 compton0.mac: Compton Scattering by a Spin 0 Structureless Charged Particle
In this section we use the batch le compton0.mac to work out details of Compton scattering of scalar charged par-
ticles. Some references to this subject can be found in G/R QED, p 435, B/D RQM, p. 193, Schwinger PSF I, p290,
Aitchison RQM, p87, De Wit/Smith, p 147, I/Z p. 286.
Three diagrams contribute to Compton scattering of a scalar charged particle which we can for convenience think of as

+
(p
1
+ (k
1
,
1
)
+
(p
2
+ (k
2
,
2
) (12.58)
Let
s = (k
1
+ p
1
)
2
, t = (k
1
k
2
)
2
, u = (k
1
p
2
)
2
. (12.59)
Four momentum conservation means p
1
+ k
1
= p
2
+ k
2
. If we ignore a factor e
2
and use real photon polarization
4-vectors, the invariant amplitude is
M = e

k
2

2
_
2 g

(2 p
2
+ k
2
)

(2 p
1
+ k
1
)

s m
2

(2 p
1
k
2
)

(2 p
2
k
1
)

u m
2
_
e

k
1

1
(12.60)
Use of k
1
e
k
1

1
= 0, k
2
e
k
2

2
= 0, s m
2
= 2 p
1
k
1
, and u m
2
= 2 p
2
k
1
leads to
M = 2 e

k
2

2
T

k
1

1
(12.61)
where
T

= g

p
2
p
1
k
1
p
1
+
p
2
p
1
k
1
p
2
(12.62)
We form the absolute square of M and sum over
2
rst,

2
|M|
2
= 4 T

k
1

1
P

(k
2
) T

k
1

1
= 4 T

k
1

1
e

k
1

1
(12.63)
56
Using the compact form of P

(k
2
), only the rst term contributes at this stage. The second term contribution is
proportional to
T

k
1

1
k

2
T

k
1

1
. Using k
1
p
1
= k
2
p
2
and k
1
p
2
= k
2
p
1
,
k

2
T

= (k
2
+ p
2
p
1
)

= k
1
(12.64)
But then k
1
e

k
1

1
= 0. Likewise the third term contribution is zero because k

2
T

= k
1
, and k
1
e

k
1

1
= 0.
We now sum over
1
.

1
,
2
|M|
2
= 4 T

P

(k
1
) T

= 4 (T

2) , (12.65)
in which all three terms of P

(k
1
) contribute.
For example, the second term is proportional to
_
T

1

k

1
T

_
/
_
k
1


k
1
_
. The numerator is equal to the de-
nominator, since k

1
T

= (k
1
+ p
1
p
2
)

= k
2
, and k
2
T

= k
1
, as we saw above.
In the compton0.mac code, the expression T

is represented by t_munu, written in terms of the constructs D, UI,

and Gm which the Dirac package recognises. The contraction T

becomes Con (t_munu2).

The batch le compton0.mac uses the Dirac package functions VP and ev_Ds.
(%i3) load(dirac2);
(%o1) "c:/work5/dirac2.mac"
(%i2) batch ("compton0.mac");
read and interpret file: #pc:/work5/compton0.mac
(%i3) " ======================================================"
(%i4) " file compton0.mac "
(%i5) " Maxima by Example, Ch. 12 "
(%i6) " Dirac Algebra and Quantum Electrodynamics "
(%i7) " Edwin L Woollett, woollett@charter.net "
(%i8) " http://www.csulb.edu/woollett "
(%i9) print(" ver: ",_binfo%," date: ",mydate)
ver: Maxima 5.23.2 date: 2011-04-04
(%i10) " COMPTON SCATTERING OF STRUCTURELESS SPIN 0 PARTICLES "
(%i11) " which we will call (pi,gamma) scattering "
(%i12) " gamma(k1,s1) + pi(p1) --> gamma(k2,s2) + pi(p2) "
(%i13) " POPULATE THE LIST invarR OF 4-VEC DOT PRODUCT VALUES. "
(%i14) " Express 4-vector dot products in terms of s, t, u, and m2 "
(%i15) " Using s = (p1+k1)2 = (p2+k2)2, t = (k2-k1)2 = "
(%i16) " (p2-p1)2, u = (k2-p1)2 = (p2-k1)2, s + t + u = 2
*
m2 "
(%i17) invar(D(p1,p1) = m2,D(p1,k1) = (s-m2)/2,D(p1,k2) = (-(u-m2))/2,
D(p1,p2) = (u+s)/2,D(p2,p2) = m2,D(p2,k1) = (-(u-m2))/2,
D(p2,k2) = (s-m2)/2,D(k1,k1) = 0,D(k1,k2) = (u+s)/2-m2,
D(k2,k2) = 0)
(%i18) " 1. UNPOLARIZED RESULT IN ARBITRARY FRAME (s,t,u form) "
(%i19) " amplitude is M_fi "
(%i20) " M_fi = -2
*
e2
*
( D(e2_cc,e1) - D(e2_cc,p2)
*
D(e1,p1)/D(k1,p1) "
(%i21) " + D(e2_cc,p1)
*
D(e1,p2)/D(k1,p2) ) "
(%i22) " The parenthesis expression can be written as the "
(%i23) " contraction (assume here real pol. vectors ) "
(%i24) " sum (sum (e2[mu]
*
T[mu,nu]
*
e1[mu],mu,0,3),nu,0,3) "
(%i25) " where T[mu,nu] is the expression t_munu : "
57
(%i26) " -------------------------------------------------"
(%i27) t_munu:UI(p1,mu)
*
UI(p2,nu)/D(k1,p2)-UI(p1,nu)
*
UI(p2,mu)/D(k1,p1)
+Gm(mu,nu)
(%i28) "========================================================="
(%i29) " Note Schwingers expression (3-12.93) used here, "
(%i30) " which he derives on page 73, and further discusses p.294"
(%i31) " The Lorentz gauge photon polarization 4-vector always has"
(%i32) " the properties: 1). D(k,e(k,s)) = 0, "
(%i33) " 2). D(n,e(k,s)) = 0, 3). D(kbar,e(k,s)) = 0 "
(%i34) " 4). D(e(k,s)_cc,e(k,r)) = -kroneker-delta(s,r) "
(%i35) " Within the Lorentz gauge one still has gauge freedom "
(%i36) " since one can choose the unit timelike vector nmu to"
(%i37) " have any components, with kbarmu = -kmu + 2
*
nmu
*
D(k,n). "
(%i38) " In an frame in which n = (1,0,0,0) (which --> D(n,n) = 1)"
(%i39) " the spatial components of kbar are the opposite of those "
(%i40) " of k. One can identify n with any convenient 4-vector in "
(%i41) " the problem "
(%i42) "==========================================================="
(%i43) " Evaluate Mssq = abs. value squared of M_fi."
(%i44) " pull out factor e4 and sum over polarizations of both photons "
(%i45) " The result is : "
(%i46) Mssq:ratsimp(4
*
(Con(t_munu2)-2))
(%o46) ((8
*
s2+8
*
m4)
*
u2-32
*
m6
*
u+8
*
m4
*
s2-32
*
m6
*
s+40
*
m8)
/((s2-2
*
m2
*
s+m4)
*
u2+(-2
*
m2
*
s2+4
*
m4
*
s-2
*
m6)
*
u+m4
*
s2-2
*
m6
*
s
+m8)
(%i47) " LAB FRAME EVALUATION "
(%i48) " initial pion rest frame evaluation in terms of"
(%i49) " scattering angle of final photon relative to initial"
(%i50) " photon direction (z - x plane) "
(%i51) " kinematics with initial pion p1 at rest. "
(%i52) " initial photon k1 moving along positive z axis. "
(%i53) assume(m > 0,k > 0,kp > 0,th >= 0,th <= %pi)
(%i54) comp_def(p1(m,0,0,0),k1(k,0,0,k),k2(kp,kp
*
sin(th),0,kp
*
cos(th)),
p2(-kp+k+m,-kp
*
sin(th),0,k-kp
*
cos(th)))
(%i55) " ---------------------------------------------------------"
(%i56) " conservation of relativistic energy gives "
(%i57) " m + k = E_p2 + kp "
(%i58) " and conservation of 3 or 4-momentum implies "
(%i59) " kp = m
*
k/(m + k
*
(1-cos(th))) "
(%i60) " ---------------------------------------"
(%i61) " find replacement rule for kp using conservation of 4-mom"
(%i62) kp_rule:solve(VP(-k2+k1+p1,-k2+k1+p1) = ev_invar(D(p2,p2)),kp)
(%o62) [kp = -k
*
m/(k
*
cos(th)-m-k)]
(%i63) s_th:VP(k1+p1,k1+p1)
(%o63) m2+2
*
k
*
m
(%i64) t_th:(VP(k1-k2,k1-k2),ev(%%,kp_rule))
(%o64) 2
*
k2
*
m/(k
*
cos(th)-m-k)-2
*
k2
*
m
*
cos(th)/(k
*
cos(th)-m-k)
(%i65) u_th:(VP(k2-p1,k2-p1),ev(%%,kp_rule))
(%o65) 2
*
k
*
m2/(k
*
cos(th)-m-k)+m2
(%i66) " replace s by s_th, etc "
(%i67) Mssq_lab:factor(sub_stu(Mssq))
(%o67) 4
*
(cos(th)2+1)
(%i68) " having pulled out e4 from Mssq "
(%i69) A:alpha2
*
(kp/k)2/(4
*
m2)
(%o69) alpha2
*
kp2/(4
*
k2
*
m2)
(%i70) " -------------------------------------- "
(%i71) " To get unpolarized differential cross section, "
(%i72) " divide sum over spins Mssq_lab by 2 from average "
(%i73) " over polarization of the initial photon. "
(%i74) dsigdo_lab_unpol:A
*
Mssq_lab/2
58
(%i75) (display2d:true,display(dsigdo_lab_unpol),display2d:false)
2 2 2
alpha kp (cos (th) + 1)
dsigdo_lab_unpol = -------------------------
2 2
2 k m
(%i76) " which agrees with G/R QED p. 437 Eq.(14) "
(%i77) " since r_02 = alpha2/m2 "
(%i78) " ======================================="
(%i79) "---------------------------------------------------"
(%i80) " 2. LINEAR PHOTON POLARIZATION CASE "
(%i81) " ------ lab frame, Coulomb gauge --------------"
(%i82) " Leave real photon polarization 4-vectors e1,e2 in amplitude. "
(%i83) " replace p1 dot products using lab frame coulomb gauge."
(%i84) " replace p2 dot products with e1,e2 "
(%i85) " using 4-mom conservation relation. "
(%i86) " The D(p1,e1) and D(p1,e2) replacements depend on our "
(%i87) " lab frame Coulomb gauge choices. "
(%i88) " Recall that D has property symmetric, so "
(%i89) " D(a,b) = D(b,a). "
(%i90) " choose real polarization vectors => linear pol."
(%i91) invar(D(e1,e1) = -1,D(k1,e1) = 0,D(e2,e2) = -1,D(k2,e2) = 0,
D(p1,e1) = 0,D(p1,e2) = 0,D(e1,p2) = -D(e1,k2),
D(e2,p2) = D(e2,k1))
(%i92) Ampl:D(e2,c)
*
D(e1,d)/(u-m2)+D(e2,a)
*
D(e1,b)/(s-m2)-2
*
D(e1,e2)
(%o92) D(c,e2)
*
D(d,e1)/(u-m2)+D(a,e2)
*
D(b,e1)/(s-m2)-2
*
D(e1,e2)
(%i93) " the amplitude is real so just need Ampl2 for cross section "
(%i94) Ampsq:expand(Ampl2)
(%o94) D(c,e2)2
*
D(d,e1)2/(u2-2
*
m2
*
u+m4)
+2
*
D(a,e2)
*
D(b,e1)
*
D(c,e2)
*
D(d,e1)/(s
*
u-m2
*
u-m2
*
s+m4)
-4
*
D(c,e2)
*
D(d,e1)
*
D(e1,e2)/(u-m2)
+D(a,e2)2
*
D(b,e1)2/(s2-2
*
m2
*
s+m4)
-4
*
D(a,e2)
*
D(b,e1)
*
D(e1,e2)/(s-m2)+4
*
D(e1,e2)2
(%i95) " replace a,b,c,d "
(%i96) Ampsq1:subst([a = k2+2
*
p2,b = k1+2
*
p1,c = 2
*
p1-k2,d = 2
*
p2-k1],Ampsq)
(%o96) D(e1,2
*
p2-k1)2
*
D(e2,2
*
p1-k2)2/(u2-2
*
m2
*
u+m4)
+2
*
D(e1,2
*
p1+k1)
*
D(e1,2
*
p2-k1)
*
D(e2,2
*
p1-k2)
*
D(e2,2
*
p2+k2)
/(s
*
u-m2
*
u-m2
*
s+m4)-4
*
D(e1,e2)
*
D(e1,2
*
p2-k1)
*
D(e2,2
*
p1-k2)/(u-m2)
+D(e1,2
*
p1+k1)2
*
D(e2,2
*
p2+k2)2/(s2-2
*
m2
*
s+m4)
-4
*
D(e1,e2)
*
D(e1,2
*
p1+k1)
*
D(e2,2
*
p2+k2)/(s-m2)+4
*
D(e1,e2)2
(%i97) " expand dot products and use invariants list invarR "
(%i98) Ampsq2:ev_Ds(Ampsq1)
(%o98) 4
*
D(e1,e2)2
(%i99) " we see that there is no further reference to s and u in this result"
(%i100) " multiply by A to get polarized cross section "
(%i101) dsigdo_lab_pol:A
*
Ampsq2
(%i102) (display2d:true,display(dsigdo_lab_pol),display2d:false)
2 2 2
alpha D (e1, e2) kp
dsigdo_lab_pol = ---------------------
2 2
k m
(%i103) " which agrees with G/R QED p. 437, Eq.(12) "
(%i104) " since r_02 = alpha2/m2 "
(%i105) " In Coulomb gauge lab frame the four vector dot product square "
(%i106) " D(e2,e1)2 reduces to dot(e2_vec,e1_vec)2 "
(%i107) " We recover the unpolarized result above if we sum over both "
(%i108) " e1 and e2 values and divide by 2 since we are averaging over "
(%i109) " the e1 values (see above section on transverse pol. vector sums)"
59
(%i110) "----------------------------------------------------------------"
(%i111) " If we only average over the polarization of the incident photon"
(%i112) " and leave the polarization of the final photon arbitrary, we "
(%i113) " replace (1+cos(th)2) by ( 1 - dot(ku,e2_vec)2) which gives "
(%i114) " 1 for e2_vec chosen perpendicular to the scattering plane, and "
(%i115) " gives sin(th)2 for e2_vec chosen parallel to the scattering "
(%i116) " plane, leading to the scattered photons polarized preferentially"
(%i117) " in a direction perpendicular to the scattering plane."
(%o117) "compton0.mac"
12.13 compton1.mac: Lepton-photon scattering
The scattering event is
e

(p
1
,
1
) + (k
1
,
1
) e

(p
2
,
2
) + (k
2
,
2
). (12.66)
The invariant amplitude is M = M
1
+ M
2
, where, ignoring the charge squared factor e
2
,
M
1
=
u
2

1
u
1
u m
2
,
1
= e /
1
(p /
1
k/
2
+ m) e /

2
), u = (p
1
k
2
)
2
(12.67)
and
M
2
=
u
2

2
u
1
s m
2
,
2
= e /

2
(p /
1
+ k/
1
+ m) e /
1
, s = (p
1
+ k
1
)
2
(12.68)
The rst section of compton1.mac calculates the unpolarized differential cross section, proportional to

1
,
2
,
1
,
2
|M
1
+ M
2
|
2
. (12.69)
One the the four terms is

|M
1
|
2
. If we rst sum on
1
, the result is proportional to

1
u
2
e /
1

a
u
1
u
1

b
e /
1
u
2
=
_

1
e

k
1

1
e

k
1

1
_
u
2

a
u
1
u
1

b

u
2
(12.70)
A convenient, simplifying feature of the lepton photon interaction is that only the rst (metric tensor) term of P

(k
1
)
need be retained, since the other terms produce zero contribution (as is discussed in quantum eld theory texts). Thus for
lepton photon interactions we effectively have
P

(k
1
) g

(12.71)
Thus the sum on
1
above reduces to
u
2

a
u
1
u
1

b

u
2
. (12.72)
After summing over
2
,
1
,
2
as well, and again ignoring the factor e
4
,

1
,
2
,
1
,
2
|M
1
|
2
=
Trace {(p /
2
+ m)

(p /
1
k/
2
+ m)

(p /
1
+ m)

(p /
1
k/
2
+ m)

}
(u m
2
)
2
(12.73)
The compton1.mac code uses the symbol m1sq for the trace in the numerator, and uses the symbol Mssq_unpol
for

|M
1
+ M
2
|
2
(see also the run line (%i32)). The Dirac package functions VP, D_sub, take_parts,ts, and
pullfac are used in this code.
60
(%i1) load(dirac2);
(%o1) "c:/work5/dirac2.mac"
(%i2) batch("compton1.mac");
file_search1: compton1.mac not found in file_search_maxima.
-- an error. To debug this try: debugmode(true);
(%i3) batch("compton1.mac");
read and interpret file: #pc:/work5/compton1.mac
(%i4) " ======================================================"
(%i5) " file compton1.mac "
(%i6) " Maxima by Example, Ch. 12 "
(%i7) " Dirac Algebra and Quantum Electrodynamics "
(%i8) " Edwin L Woollett, woollett@charter.net "
(%i9) " http://www.csulb.edu/woollett "
(%i10) print(" ver: ",_binfo%," date: ",mydate)
ver: Maxima 5.23.2 date: 2011-04-04
(%i11) " --------------------------------------------"
(%i12) " UNPOLARIZED LEPTON - PHOTON SCATTERING "
(%i13) " for example : "
(%i14) " ---------------------------------------------"
(%i15) " e(-,p1,m) + gamma(k1) --> e(-,p2,m) + gamma(k2) "
(%i16) " ----------------------------------------------"
(%i17) " Arbitrary Frame Covariant Symbolic Methods "
(%i18) " in an arb frame, D(k1,p1) = D(k2,p2) "
(%i19) " and D(k1,p2) = D(k2,p1) "
(%i20) " In an arbitrary frame we can express everything "
(%i21) " in terms of m2, D(k1,p1) and D(k2,p1) "
(%i22) " with a view to then evaluating in the rest frame"
(%i23) " of the initial electron: p1 = (m,0,0,0) (lab frame) "
(%i24) " ---------------------------------------------"
(%i25) invar(D(p1,p1) = m2,D(p2,p2) = m2,D(k1,k1) = 0,D(k2,k2) = 0,
D(p1,p2) = -D(k2,p1)+D(k1,p1)+m2,D(k1,p2) = D(k2,p1),
D(k2,p2) = D(k1,p1),D(k1,k2) = D(k1,p1)-D(k2,p1))
(%i26) m1sq:tr(m+p2,mu,m-k2+p1,nu,m+p1,nu,m-k2+p1,mu)
(%o26) 32
*
m4-32
*
D(k2,p1)
*
m2+32
*
D(k1,p1)
*
D(k2,p1)
(%i27) m2sq:tr(m+p2,mu,m+k1+p1,nu,m+p1,nu,m+k1+p1,mu)
(%o27) 32
*
m4+32
*
D(k1,p1)
*
m2+32
*
D(k1,p1)
*
D(k2,p1)
(%i28) m1m2sq:tr(m+p2,mu,m-k2+p1,nu,m+p1,mu,m+k1+p1,nu)
(%o28) 32
*
m4-16
*
D(k2,p1)
*
m2+16
*
D(k1,p1)
*
m2
(%i29) m2m1sq:tr(m+p2,mu,m+k1+p1,nu,m+p1,mu,m-k2+p1,nu)
(%o29) 32
*
m4-16
*
D(k2,p1)
*
m2+16
*
D(k1,p1)
*
m2
(%i30) m1m2sq-m2m1sq
(%o30) 0
(%i31) " The interference terms are equal, so "
(%i32) " -----------------------------------------------"
(%i33) " Form the expression :"
(%i34) " Mssq_unpol = m1sq/(u-m2)2 + m2sq/(s-m2)2 "
(%i35) " + 2
*
m1m2sq/((u-m2)
*
(s-m2)), "
(%i36) " with u-m2 = -2
*
D(k2,p1), s-m2 = 2
*
D(k1,p1), "
(%i37) " using u = (p1 - k2)2, s = (p1 + k1)2 "
(%i38) " ( we use sloppy notation p2 for D(p,p) here )"
(%i39) " -----------------------------------------------"
(%i40) Mssq_unpol:expand((-2
*
m1m2sq)/(4
*
D(k2,p1)
*
D(k1,p1))
+m2sq/(4
*
D(k1,p1)2)+m1sq/(4
*
D(k2,p1)2))
(%o40) -16
*
m4/(D(k1,p1)
*
D(k2,p1))+8
*
m4/D(k2,p1)2+8
*
m4/D(k1,p1)2
-16
*
m2/D(k2,p1)+16
*
m2/D(k1,p1)
+8
*
D(k2,p1)/D(k1,p1)+8
*
D(k1,p1)/D(k2,p1)
(%i41) " pull factor of 8 out of definition "
(%i42) Mssq_unpol8:expand(Mssq_unpol/8)
(%o42) -2
*
m4/(D(k1,p1)
*
D(k2,p1))+m4/D(k2,p1)2+m4/D(k1,p1)2-2
*
m2/D(k2,p1)
+2
*
m2/D(k1,p1)+D(k2,p1)/D(k1,p1)
+D(k1,p1)/D(k2,p1)
61
(%i43) "---------------------------------------------------------"
(%i44) " LAB FRAME EVALUATION UNPOLARIZED "
(%i45) " rest frame of initial electron "
(%i46) " --------------------------------------------------------"
(%i47) assume(m > 0,k > 0,kp > 0,th >= 0,th <= %pi)
(%i48) comp_def(p1(m,0,0,0),k1(k,0,0,k),k2(kp,kp
*
sin(th),0,kp
*
cos(th)),
p2(-kp+k+m,-kp
*
sin(th),0,k-kp
*
cos(th)))
(%i49) " use conservation of 4-momentum to find kp "
(%i50) kp_rule:solve(VP(-k2+k1+p1,-k2+k1+p1) = ev_invar(D(p2,p2)),kp)
(%o50) [kp = -k
*
m/(k
*
cos(th)-m-k)]
(%i51) " To compare results, only replace kp in terms 1-5 "
(%i52) " leave kp alone in terms 6-7 "
(%i53) " D_sub (e,[D1,D2,..] uses noncov on each D(p1,p2) in list"
(%i54) M1t5:take_parts(Mssq_unpol8,1,5)
(%o54) -2
*
m4/(D(k1,p1)
*
D(k2,p1))+m4/D(k2,p1)2+m4/D(k1,p1)2-2
*
m2/D(k2,p1)
+2
*
m2/D(k1,p1)
(%i55) M6t7:take_parts(Mssq_unpol8,6,7)
(%o55) D(k2,p1)/D(k1,p1)+D(k1,p1)/D(k2,p1)
(%i56) M6t7:D_sub(M6t7,[D(k1,p1),D(k2,p1)])
(%o56) kp/k+k/kp
(%i57) M1t5:D_sub(M1t5,[D(k1,p1),D(k2,p1)])
(%o57) -2
*
m2/(k
*
kp)+m2/kp2+m2/k2-2
*
m/kp+2
*
m/k
(%i58) M1t5:trigsimp(ev(M1t5,kp_rule))
(%o58) cos(th)2-1
(%i59) " ts is our alternative trigsimp function "
(%i60) M1t5:ts(M1t5,th)
(%o60) -sin(th)2
(%i61) " restore overall factor of 8 "
(%i62) Mssq_unpol:8
*
(M6t7+M1t5)
(%o62) 8
*
(-sin(th)2+kp/k+k/kp)
(%i63) A:alpha2
*
(kp/k)2/(4
*
m2)
(%o63) alpha2
*
kp2/(4
*
k2
*
m2)
(%i64) " -------------------------------------- "
(%i65) " To get unpolarized differential cross section, "
(%i66) " divide sum over spins Mssq_unpol by 4 from average "
(%i67) " over spin of initial electron and initial photon. "
(%i68) dsigdo_lab_unpol:A
*
Mssq_unpol/4
(%i69) (display2d:true,display(dsigdo_lab_unpol),display2d:false)
2 2 2 kp k
alpha kp (- sin (th) + -- + --)
k kp
dsigdo_lab_unpol = ---------------------------------
2 2
2 k m
(%i70) " which agrees with BLP (86.9) "
To calculate the differential cross section for given photon (linear) polarizations (but summing over the inital and nal
elecron helicities), we retain the photon polarization vector factors e /
k
1

1
and e /
k
2

2
inside the trace created by summing
over the electron helicities.
We already have the basic dening inner products e
k
1

1
e
k
1

1
= 1, k
1
e
k
1

1
= 0, e
k
2

2
e
k
2

2
= 1, and
k
2
e
k
2

2
= 0.
Since we want a result appropriate to the rest frame of the initial electron p
1
, we identify n

= p

1
/m so that the compo-
nents of n

are (1, 0, 0, 0).

Then n e
k
1

1
= 0 implies that e
0
k
1

1
= 0 and n e
k
2

2
= 0 implies that e
0
k
2

2
= 0.
62
Then e
k
1

1
= (0, e
k
1

1
) and e
k
2

2
= (0, e
k
2

2
).
Since the components of p

1
are (m, 0, 0, 0), we then have p
1
e
k
1

1
= 0 and p
1
e
k
2

2
= 0. And using 4-momentum
conservation, the former equation implies 0 = e
k
1

1
(k
2
+ p
2
k
1
), and hence e
k
1

1
p
2
= e
k
1

1
k
2
.
Finally we replace e
k
2

2
p
2
= e
k
2

2
k
1
, using again 4-momentum conservation and the previous relations.
In our code, e1 stands for e
k
1

1
and e2 stands for e
k
2

2
.
(%i71) "======================================================"
(%i72) "----UNPOLARIZED ELECTRONS, POLARIZED PHOTONS ------------"
(%i73) "------- rest frame of initial electron p1 ------------- "
(%i74) " Leave photon polarization 4-vectors e1,e2 in trace. "
(%i75) " replace p1 dot products using p1 rest frame choice."
(%i76) " replace p2 dot products with e1,e2 "
(%i77) " using 4-mom conservation relation. "
(%i78) " The D(p1,e1) and D(p1,e2) replacements depend on our "
(%i79) " p1 rest frame choice."
(%i80) " Recall that D has property symmetric, so "
(%i81) " D(a,b) = D(b,a). "
(%i82) invar(D(e1,e1) = -1,D(k1,e1) = 0,D(e2,e2) = -1,D(k2,e2) = 0,
D(p1,e1) = 0,D(p1,e2) = 0,D(e1,p2) = -D(e1,k2),
D(e2,p2) = D(e2,k1))
(%i83) M1sq:tr(m+p2,e1,m-k2+p1,e2,m+p1,e2,m-k2+p1,e1)
(%o83) 8
*
D(k1,p1)
*
D(k2,p1)-16
*
D(e1,k2)2
*
D(k2,p1)
(%i84) M2sq:tr(m+p2,e2,m+k1+p1,e1,m+p1,e1,m+k1+p1,e2)
(%o84) 8
*
D(k1,p1)
*
D(k2,p1)+16
*
D(e2,k1)2
*
D(k1,p1)
(%i85) M1m2sq:tr(m+p2,e1,m-k2+p1,e2,m+p1,e1,m+k1+p1,e2)
(%o85) -16
*
D(e1,e2)2
*
D(k1,p1)
*
D(k2,p1)+8
*
D(k1,p1)
*
D(k2,p1)
+8
*
D(e2,k1)2
*
D(k2,p1)
-8
*
D(e1,k2)2
*
D(k1,p1)
(%i86) M2m1sq:tr(m+p2,e2,m+k1+p1,e1,m+p1,e2,m-k2+p1,e1)
(%o86) -16
*
D(e1,e2)2
*
D(k1,p1)
*
D(k2,p1)+8
*
D(k1,p1)
*
D(k2,p1)
+8
*
D(e2,k1)2
*
D(k2,p1)
-8
*
D(e1,k2)2
*
D(k1,p1)
(%i87) M2m1sq-M1m2sq
(%o87) 0
(%i88) " The interference terms are equal, so "
(%i89) Mssq_pol:expand((-2
*
M1m2sq)/(4
*
D(k2,p1)
*
D(k1,p1))
+M2sq/(4
*
D(k1,p1)2)+M1sq/(4
*
D(k2,p1)2))
(%o89) 2
*
D(k2,p1)/D(k1,p1)+2
*
D(k1,p1)/D(k2,p1)+8
*
D(e1,e2)2-4
(%i90) " avoid use of noncov(D(e1,e2)) which would expand D(e1,e2)"
(%i91) " in terms of undefined components, by instead using ev"
(%i92) " or D_sub for selected replacements. "
(%i93) Mssq_pol:D_sub(Mssq_pol,[D(k1,p1),D(k2,p1)])
(%o93) 2
*
kp/k+2
*
k/kp+8
*
D(e1,e2)2-4
(%i94) Mssq_pol:pullfac(Mssq_pol,2)
(%o94) 2
*
(kp/k+k/kp+4
*
D(e1,e2)2-2)
(%i95) "To get electron unpolarized, photon polarized differential cross section,"
(%i96) " divide Mssq_pol by 2 from average over helicity of initial electron. "
(%i97) dsigdo_lab_pol:A
*
Mssq_pol/2
(%i98) (display2d:true,display(dsigdo_lab_pol),display2d:false)
2 2 kp k 2
alpha kp (-- + -- + 4 D (e1, e2) - 2)
k kp
dsigdo_lab_pol = ---------------------------------------
2 2
4 k m
(%i99) " which agrees with the Klein-Nishina form quoted by Weinberg (8.7.39)"
63
(%i100) "============================================================"
(%i101) " Now assume unpolarized initial photon. Average the "
(%i102) " polarized photon dsigdo(r,s) over the two possible values "
(%i103) " of r which specifies the transverse polarization "
(%i104) " vector of incident photon e1_vec. With nmu = (1,0,0,0)"
(%i105) " D(e1,e2)2 --> { dot3 (e1_vec,e2_vec) }2 "
(%i106) " and sum (D(e1(r),e2(s))2,r,1,2) --> 1 - {dot3(ku,e2_vec(s)}2"
(%i107) " with ku = unit 3-vector along the z axis, dot3(a,b) = 3-vec dot prod. "
(%i108) " and dsigdo(s) = sum(dsigdo(r,s),r,1,2)/2 "
(%i109) " The overall result is recovered with: "
(%i110) dsigdo_s:subst(4
*
D(e1,e2)2 = 2-2
*
dot3(ku,e2_vec)2,dsigdo_lab_pol)
(%i111) (display2d:true,display(dsigdo_s),display2d:false)
2 2 2 kp k
alpha kp (- 2 dot3 (ku, e2_vec) + -- + --)
k kp
dsigdo_s = --------------------------------------------
2 2
4 k m
(%i112) "---------------------------------------------------------"
(%i113) " which agrees with Weinberg (8.7.40) "
(%i114) " Summing the above over the two possible values of "
(%i115) " final photon polarization s then leads to our "
(%i116) " unpolarized result above "
(%o116) "compton1.mac"
12.14 pair1.mac: Unpolarized Two Photon Pair Annihilation
The pair annihilation to two photons event is
e

(p
1
,
1
) + e
+
(p
2
,
2
) (k
1
,
1
) + (k
2
,
2
). (12.74)
With the denitions
s = (p
1
+ p
2
)
2
, t = (p
1
k
1
)
2
, u = (p
1
k
2
)
2
, (12.75)
and the abbreviations
u
1
= u(p
1
,
1
), v
2
= v(p
2
,
2
), e

1
= e

k
1
,
1
, e

2
= e

k
2
,
2
, (12.76)
M = M
1
+ M
2
, with
M
1
=
e
2
v
2
e /

2
(p /
1
k/
1
+ m) e /

1
u
1
t m
2
(12.77)
and
M
2
=
e
2
v
2
e /

1
(p /
1
k/
2
+ m) e /

2
u
1
u m
2
. (12.78)
Some references for this calculation are Grifth p. 241, G/R QED: (p. 192 rest frame of electron, p. 198 CM frame), P/S
p. 168, BLP p. 370, Schwinger p. 314, Kaku p. 170, I/Z p. 230, and B/D p. 132.
The batch le pair1.mac works out the unpolarized differential cross section rst in an arbitrary frame, and then in
the CM frame. To compare our result with G/R in the CM frame, we express the result in terms of (E,v), where
p = E
*
v, E2 = m2 + p2, with v (the dimensionless velocity of either of the incident leptons relative to the
CM frame) having values 0 v < 1.
The Dirac package functions D_sub and ts are used.
64
(%i117) load(dirac2);
(%o1) "c:/work5/dirac2.mac"
(%i2) batch ("pair1.mac");
read and interpret file: #pc:/work5/pair1.mac
(%i3) " ======================================================"
(%i4) " file pair1.mac "
(%i5) " Maxima by Example, Ch. 12 "
(%i6) " Dirac Algebra and Quantum Electrodynamics "
(%i7) " Edwin L Woollett, woollett@charter.net "
(%i8) " http://www.csulb.edu/woollett "
(%i9) print(" ver: ",_binfo%," date: ",mydate)
ver: Maxima 5.23.2 date: 2011-04-04
(%i10) " --------------------------------------------"
(%i11) " UNPOLARIZED PAIR ANNIHILATION "
(%i12) " ---------------------------------------------"
(%i13) " e(-,p1) + e(+,p2) --> gamma(k1) + gamma(k2) "
(%i14) " ----------------------------------------------"
(%i15) " Arbitrary Frame Covariant Symbolic Methods "
(%i16) " ---------------------------------------------"
(%i17) " POPULATE THE LIST invarR OF 4-VEC DOT PRODUCT VALUES. "
(%i18) " Express 4-vector dot products in terms of t, u, and m2 "
(%i19) " Using s = (p1+p2)2 = (k1+k2)2, t = (p1-k1)2 = "
(%i20) " (k2-p2)2, u = (p1-k2)2 = (k1-p2)2, s + t + u = 2
*
m2 "
(%i21) " ---------------------------------------------------"
(%i22) invar(D(p1,p1) = m2,D(p1,k1) = (m2-t)/2,D(p1,k2) = (m2-u)/2,
D(p1,p2) = (-(u+t))/2,D(p2,p2) = m2,D(p2,k1) = (m2-u)/2,
D(p2,k2) = (m2-t)/2,D(k1,k1) = 0,D(k1,k2) = m2-(u+t)/2,
D(k2,k2) = 0)
(%i23) "------------------------------------------------------"
(%i24) " sum |M|2 over spins of electrons and polarizations "
(%i25) " of the photons, which leads to the expression: "
(%i26) " -------------------------------------------------"
(%i27) " m1sq/(t-m2)2 + m2sq/(u-m2)2 + "
(%i28) " (m1m2sq+m2m1sq)/((t-m2)
*
(u-m2))"
(%i29) " -------------------------------------------------"
(%i30) " -- where (ignoring e4) --- "
(%i31) m1sq:tr(p2-m,mu,m-k1+p1,nu,m+p1,nu,m-k1+p1,mu)
(%o31) 8
*
t
*
u-8
*
m2
*
u-24
*
m2
*
t-8
*
m4
(%i32) m2sq:tr(p2-m,mu,m-k2+p1,nu,m+p1,nu,m-k2+p1,mu)
(%o32) 8
*
t
*
u-24
*
m2
*
u-8
*
m2
*
t-8
*
m4
(%i33) m1m2sq:tr(p2-m,mu,m-k1+p1,nu,m+p1,mu,m-k2+p1,nu)
(%o33) -8
*
m2
*
u-8
*
m2
*
t-16
*
m4
(%i34) m2m1sq:tr(p2-m,mu,m-k2+p1,nu,m+p1,mu,m-k1+p1,nu)
(%o34) -8
*
m2
*
u-8
*
m2
*
t-16
*
m4
(%i35) m1m2sq-m2m1sq
(%o35) 0
(%i36) Mssq:ratsimp(2
*
m1m2sq/((t-m2)
*
(u-m2))+m2sq/(u-m2)2+m1sq/(t-m2)2)
(%o36) ((8
*
t-8
*
m2)
*
u3+(24
*
m4-56
*
m2
*
t)
*
u2+(8
*
t3-56
*
m2
*
t2+112
*
m4
*
t)
*
u
-8
*
m2
*
t3+24
*
m4
*
t2-48
*
m8)
/((t2-2
*
m2
*
t+m4)
*
u2+(-2
*
m2
*
t2+4
*
m4
*
t-2
*
m6)
*
u+m4
*
t2-2
*
m6
*
t
+m8)
(%i37) "----- replace t and u in terms of dot products in arb frame ------ "
(%i38) Mssq_dot:expand(subst([t = m2-2
*
D(k1,p1),u = m2-2
*
D(k2,p1)],Mssq))
(%o38) -16
*
m4/(D(k1,p1)
*
D(k2,p1))-8
*
m4/D(k2,p1)2-8
*
m4/D(k1,p1)2
+16
*
m2/D(k2,p1)+16
*
m2/D(k1,p1)
+8
*
D(k2,p1)/D(k1,p1)+8
*
D(k1,p1)/D(k2,p1)
(%i39) "=================================================="
65
(%i40) " specialize to center of momentum frame "
(%i41) "---------------------------------------------------"
(%i42) " kinematics in center of momentum frame "
(%i43) " p_vec_mag : E
*
v, v = velocity rel to CM frame "
(%i44) " E2 = m2 + p2 = m2/(1 - v2) "
(%i45) " initial electron 3-mom in dir. of pos z axis"
(%i46) " e(-,p1) e(+,p2) --> gamma(k1) gamma(k2) "
(%i47) " CM energy = 2
*
E shared equally by incident"
(%i48) " fermions and outgoing photons "
(%i49) assume(m > 0,E > 0,th >= 0,th <= %pi)
(%i50) comp_def(p1(E,0,0,v
*
E),p2(E,0,0,-v
*
E),k1(E,E
*
sin(th),0,E
*
cos(th)),
k2(E,-E
*
sin(th),0,-E
*
cos(th)))
(%i51) " pf/pi = E/p = E/(vE) = 1/v "
(%i52) " dsigdo_CM = A
*
|M|2 for given pol, where "
(%i53) " having absorbed e4 into A, e2 = 4
*
%pi
*
alpha"
(%i54) A:alpha2/(16
*
v
*
E2)
(%o54) alpha2/(16
*
v
*
E2)
(%i55) " convert to a function of angle th between "
(%i56) " p1_vec and k1_vec "
(%i57) Mssq_th:D_sub(Mssq_dot,[D(k1,p1),D(k2,p1)])
(%o57) -16
*
m4/((E2-cos(th)
*
v
*
E2)
*
(cos(th)
*
v
*
E2+E2))
+8
*
(E2-cos(th)
*
v
*
E2)/(cos(th)
*
v
*
E2+E2)+16
*
m2/(cos(th)
*
v
*
E2+E2)
-8
*
m4/(cos(th)
*
v
*
E2+E2)2+16
*
m2/(E2-cos(th)
*
v
*
E2)
-8
*
m4/(E2-cos(th)
*
v
*
E2)2+8
*
(cos(th)
*
v
*
E2+E2)/(E2-cos(th)
*
v
*
E2)
(%i58) Mssq_th:trigsimp(subst(E2 = m2/(1-v2),Mssq_th))
(%o58) -((16
*
sin(th)4+16)
*
v4-32
*
sin(th)2
*
v2-16)
/(cos(th)4
*
v4-2
*
cos(th)2
*
v2+1)
(%i59) " simplify the denominator - override default factor "
(%i60) mt1d:denom(Mssq_th)
(%o60) cos(th)4
*
v4-2
*
cos(th)2
*
v2+1
(%i61) mt1d:ratsubst(x,v2
*
cos(th)2,mt1d)
(%o61) x2-2
*
x+1
(%i62) mt1d:factor(mt1d)
(%o62) (x-1)2
(%i63) mt1d:subst(x = v2
*
cos(th)2,mt1d)
(%o63) (cos(th)2
*
v2-1)2
(%i64) " simplify the numerator "
(%i65) mt1n:num(Mssq_th)
(%o65) -(16
*
sin(th)4+16)
*
v4+32
*
sin(th)2
*
v2+16
(%i66) " pull out factor of 16 from numerator "
(%i67) mt1n16:factor(mt1n)/16
(%o67) -sin(th)4
*
v4-v4+2
*
sin(th)2
*
v2+1
(%i68) " change sin(th) to cos(th) in numerator "
(%i69) mt1n16:expand(ts(mt1n16,th))
(%o69) -cos(th)4
*
v4+2
*
cos(th)2
*
v4-2
*
v4-2
*
cos(th)2
*
v2+2
*
v2+1
(%i70) " simplified num/denom expression: "
(%i71) Mssq_unpol:16
*
mt1n16/mt1d
(%o71) 16
*
(-cos(th)4
*
v4+2
*
cos(th)2
*
v4-2
*
v4-2
*
cos(th)2
*
v2+2
*
v2+1)
/(cos(th)2
*
v2-1)2
(%i72) " divide by 4 from averaging over initial spin "
(%i73) " and polarization "
(%i74) dsigdo_CM:A
*
Mssq_unpol/4
(%i75) (display2d:true,display(dsigdo_CM),display2d:false)
dsigdo_CM =
2 4 4 2 4 4 2 2 2
alpha (- cos (th) v + 2 cos (th) v - 2 v - 2 cos (th) v + 2 v + 1)
------------------------------------------------------------------------
2 2 2 2
4 v (cos (th) v - 1) E
66
(%i76) " which agrees with Greiner and Reinhardt, p.201, eqn (17)"
(%i77) "----------------------------------------------------------"
(%o77) "pair1.mac"
12.15 pair2.mac: Polarized Two Photon Pair Annihilation Amplitudes
The pair annihilation to two photon event is
e

(p
1
,
1
) + e
+
(p
2
,
2
) (k
1
,
1
) + (k
2
,
2
). (12.79)
With the denitions
s = (p
1
+ p
2
)
2
, t = (p
1
k
1
)
2
, u = (p
1
k
2
)
2
. (12.80)
If we use explicit Dirac spinors and matrices and photon polarization 3-vectors to calculate polarized amplitudes in which
both the leptons and photons have denite helicities in the CM frame, we start with the polarized amplitude already writ-
ten down in the previous section dealing with pair1.mac.
With the abbreviations
u
1
= u(p
1
,
1
), v
2
= v(p
2
,
2
), e

1
= e

k
1
,
1
, e

2
= e

k
2
,
2
, (12.81)
M = M
1
+ M
2
, with
M
1
=
e
2
v
2
e /

2
(p /
1
k/
1
+ m) e /

1
u
1
t m
2
(12.82)
and
M
2
=
e
2
v
2
e /

1
(p /
1
k/
2
+ m) e /

2
u
1
u m
2
. (12.83)
Here we use explicit Dirac spinor methods in the CM frame, letting each lepton have the energy E = M/2, with M the
total CM frame energy, and then each lepton has 3-momentum magnitude p = E
*
v = M
*
v/2. We also have the di-
mensionless lepton CM frame speed v = sqrt( 1 - 4
*
m2/M2 ). Each outgoing photon has energy k = M/2
and each photon has 3-momentum magnitude k.
We use the photon polarization conventions of F. Gross. (See Franz Gross, Rel. Q. Mech. and Fld. Theory, pages 54,
218, and Schwinger, PSF I, Sec. 3-13-98.)
The outgoing photons are chosen to be travelling along the +z and z axes. We let e( z, ) be the photon polarization
3-vector describing a photon travelling in the +z direction, with = 1 being a positive helicity state and = 1 being
a negative helicity state. Then
e( z, 1) =
1

2
( x + i y), (12.84)
and
e( z, 1) =
1

2
( x i y). (12.85)
We let e( z, ) be the photon polarization 3-vector describing a photon travelling in the z direction, with = 1 being
a positive helicity state and = 1 being a negative helicity state. Then
e( z, 1) =
1

2
( x i y) = R
y
() e( z, 1), (12.86)
67
and
e( z, 1) =
1

2
( x + i y) = R
y
() e( z, 1). (12.87)
For our calculation, since the photons are in the nal state, we need complex conjugates of these. The rst example
discussed is RR --> RR, which means two right-handed electrons (actually positive helicity) turn into two positive
helicity photons. For this rst example, we dene CM frame components with the line:
comp_def(p1(M/2,M
*
v
*
sin(th)/2,0,M
*
v
*
cos(th)/2),k1(M/2,0,0,M/2),
k2(M/2,0,0,(-M)/2),e1_cc(0,(-1)/sqrt(2),%i/sqrt(2),0),
e2_cc(0,1/sqrt(2),%i/sqrt(2),0))
which shows that the photon with 4-momentumk1 is travelling along the positive z axis, and the polarization state of this
photon is described by the 4-vector e1_cc with the components (0,(-1)/sqrt(2),%i/sqrt(2),0) in which the
three spatial components agree with e( z, 1)

.
Likewise, the photon with 4-momentum k2 is travelling along the negative z axis, and the symbol e2_cc is the 4-
vector describing its positive helicity state, with components (0,1/sqrt(2),%i/sqrt(2),0) whose spatial com-
ponents agree with e( z, 1)

.
The cases considered are RR --> RR, RR --> LL, RR --> RL, RR --> LR, RL --> RL, RL --> LR, and
RL --> RR (this last case has amplitude equal to zero).
Since we want to compare our results with Schwingers results, some involved manipulations are necessary in each
case to arrive at a comparable form.
(%i78) load(dirac2);
(%o1) "c:/work5/dirac2.mac"
(%i2) batch ("pair2.mac");
read and interpret file: #pc:/work5/pair2.mac
(%i3) " ======================================================"
(%i4) " file pair2.mac "
(%i5) " Maxima by Example, Ch. 12 "
(%i6) " Dirac Algebra and Quantum Electrodynamics "
(%i7) " Edwin L Woollett, woollett@charter.net "
(%i8) " http://www.csulb.edu/woollett "
(%i9) print(" ver: ",_binfo%," date: ",mydate)
ver: Maxima 5.23.2 date: 2011-04-04
(%i10) " --------------------------------------------"
(%i11) " POLARIZED PAIR ANNIHILATION in CM Frame "
(%i12) " ---------------------------------------------"
(%i13) " e(-,p1) + e(+,p2) --> gamma(k1) + gamma(k2) "
(%i14) " ----------------------------------------------"
(%i15) " AMPLITUDES USING EXPLICIT DIRAC SPINORS "
(%i16) " ----------------------------------------------"
(%i17) " Define the needed Dirac spinors and barred spinors. "
(%i18) " For example, RR --> RR amplitude without the factor e2, and"
(%i19) " either t-m2 or u-m2 in the denominator . The photon 3-momenta"
(%i20) " magnitudes are equal to M/2, where M = 2
*
E = CM frame energy."
(%i21) " M >= 2
*
m, and low energy corresponds to M at its lower limit,"
(%i22) " for which v is close to value 0, and high energy corresponds"
(%i23) " to M >> m, for which v approaches its maximum value 1 "
(%i24) " The lepton 3-momentum magnitudes are each equal to E
*
v = M
*
v/2 "
(%i25) " where v is the dimensionless velocity with "
(%i26) " v2 = p2/E2 = (E2-m2)/E2 = 1- 4
*
m2/M2 "
(%i27) " All particle energies are equal to M/2."
(%i28) " ----------------------------------------------------"
68
(%i29) " CASE: e(-,p1,+1) e(+,p2,+1) --> gamma(k1,+1) gamma(k2,+1) "
(%i30) " ----------------------------------------------------"
(%i31) " Both the electron and positron have helicity +1. "
(%i32) " Each of the final gammas have helicity +1. "
(%i33) "For a given set of helicity quantum numbers, the amplitude is the"
(%i34) " sum of two terms: Mtn/(t-m2) + Mun/(u-m2) , where "
(%i35) " Mtd = t-m2 = -2
*
D(k1,p1), Mud = u-m2 = -2
*
D(k2,p1). "
(%i36) " --------------------------------------------------"
(%i37) " To simplify the photon helicity 3-vectors, let k1_3vec be "
(%i38) " along the positive z axis, and k2_3vec be along the "
(%i39) " minus z axis, p1_3vec in the z-x plane at an angle of "
(%i40) " th radians to k1_3vec."
(%i41) (up1:UU(M/2,M
*
v/2,th,0,1),vp2b:sbar(VV(M/2,M
*
v/2,%pi-th,%pi,1)))
(%i42) " ------------------------------------------------------"
(%i43) " We need to define the components of the four vectors "
(%i44) " p1, k1, k2, e1_cc, and e2_cc relative to the CM frame"
(%i45) " For the photon polarization 4-vectors, we use the phase"
(%i46) " conventions of Franz Gross, Rel. Q. Mech. and Fld. Theory"
(%i47) " pages 54, 218, and 230. e1_cc means complex conjugate of e1. "
(%i48) " Without the cc, e1 means here e (z_hat,+1) "
(%i49) " and e2 means here e (-z_hat,+1) "
(%i50) comp_def(p1(M/2,M
*
v
*
sin(th)/2,0,M
*
v
*
cos(th)/2),k1(M/2,0,0,M/2),
k2(M/2,0,0,(-M)/2),e1_cc(0,(-1)/sqrt(2),%i/sqrt(2),0),
e2_cc(0,1/sqrt(2),%i/sqrt(2),0))
(%i51) " denominators "
(%i52) Mtd:-2
*
pullfac(noncov(D(k1,p1)),M2/4)
(%o52) -(1-cos(th)
*
v)
*
M2/2
(%i53) Mud:-2
*
pullfac(noncov(D(k2,p1)),M2/4)
(%o53) -(cos(th)
*
v+1)
*
M2/2
(%i54) g1:sL(e2_cc) . (m
*
I4-sL(k1)+sL(p1)) . sL(e1_cc)
(%o54) matrix([2
*
m,0,-2
*
(M/2-cos(th)
*
v
*
M/2),0],[0,0,0,0],
[-2
*
(cos(th)
*
v
*
M/2-M/2),0,2
*
m,0],[0,0,0,0])
(%i55) g2:sL(e1_cc) . (m
*
I4-sL(k2)+sL(p1)) . sL(e2_cc)
(%o55) matrix([0,0,0,0],[0,2
*
m,0,-2
*
(cos(th)
*
v
*
M/2+M/2)],[0,0,0,0],
[0,-2
*
(-cos(th)
*
v
*
M/2-M/2),0,2
*
m])
(%i56) " M1 = Mtn/Mtd "
(%i57) Mtn:vp2b . g1 . up1
(%o57) sin((%pi-th)/2)
*
sqrt((v+1)
*
M/2)
*
(2
*
m
*
cos(th/2)
*
sqrt((v+1)
*
M/2)
-2
*
cos(th/2)
*
sqrt(-(v-1)
*
M/2)
*
(cos(th)
*
v
*
M/2-M/2))
-sin((%pi-th)/2)
*
sqrt(-(v-1)
*
M/2)
*
(2
*
m
*
cos(th/2)
*
sqrt(-(v-1)
*
M/2)
-2
*
cos(th/2)
*
sqrt((v+1)
*
M/2)
*
(M/2-cos(th)
*
v
*
M/2))
(%i58) expand(ratsubst(cos(th/2),sin((%pi-th)/2),%))
(%o58) -cos(th/2)2
*
cos(th)
*
sqrt(1-v)
*
v
*
sqrt(v+1)
*
M2
+cos(th/2)2
*
sqrt(1-v)
*
sqrt(v+1)
*
M2+2
*
m
*
cos(th/2)2
*
v
*
M
(%i59) expand(ratsubst(2
*
m/M,sqrt(1-v2),rootscontract(%)))
(%o59) -2
*
m
*
cos(th/2)2
*
cos(th)
*
v
*
M+2
*
m
*
cos(th/2)2
*
v
*
M+2
*
m
*
cos(th/2)2
*
M
(%i60) pullfac(%,2
*
m
*
M
*
cos(th/2)2)
(%o60) 2
*
m
*
cos(th/2)2
*
(-cos(th)
*
v+v+1)
*
M
(%i61) expand(ratsubst((cos(th)+1)/2,cos(th/2)2,%))
(%o61) -m
*
cos(th)2
*
v
*
M+m
*
v
*
M+m
*
cos(th)
*
M+m
*
M
(%i62) Mtn:pullfac(%,m
*
M)
(%o62) m
*
(-cos(th)2
*
v+v+cos(th)+1)
*
M
(%i63) M1:Mtn/Mtd
(%o63) -2
*
m
*
(-cos(th)2
*
v+v+cos(th)+1)/((1-cos(th)
*
v)
*
M)
(%i64) " M2 = Mun/Mud "
69
(%i65) Mun:vp2b . g2 . up1
(%o65) cos((%pi-th)/2)
*
sqrt((v+1)
*
M/2)
*
(2
*
m
*
sin(th/2)
*
sqrt((v+1)
*
M/2)
-2
*
sin(th/2)
*
sqrt(-(v-1)
*
M/2)
*
(-cos(th)
*
v
*
M/2-M/2))
-cos((%pi-th)/2)
*
sqrt(-(v-1)
*
M/2)
*
(2
*
m
*
sin(th/2)
*
sqrt(-(v-1)
*
M/2)
-2
*
sin(th/2)
*
sqrt((v+1)
*
M/2)
*
(cos(th)
*
v
*
M/2+M/2))
(%i66) expand(ratsubst(sin(th/2),cos((%pi-th)/2),%))
(%o66) sin(th/2)2
*
cos(th)
*
sqrt(1-v)
*
v
*
sqrt(v+1)
*
M2
+sin(th/2)2
*
sqrt(1-v)
*
sqrt(v+1)
*
M2+2
*
m
*
sin(th/2)2
*
v
*
M
(%i67) expand(ratsubst(2
*
m/M,sqrt(1-v2),rootscontract(%)))
(%o67) 2
*
m
*
sin(th/2)2
*
cos(th)
*
v
*
M+2
*
m
*
sin(th/2)2
*
v
*
M+2
*
m
*
sin(th/2)2
*
M
(%i68) pullfac(%,2
*
m
*
M
*
sin(th/2)2)
(%o68) 2
*
m
*
sin(th/2)2
*
(cos(th)
*
v+v+1)
*
M
(%i69) Mun:expand(ratsubst((1-cos(th))/2,sin(th/2)2,%))
(%o69) -m
*
cos(th)2
*
v
*
M+m
*
v
*
M-m
*
cos(th)
*
M+m
*
M
(%i70) Mun:pullfac(%,m
*
M)
(%o70) m
*
(-cos(th)2
*
v+v-cos(th)+1)
*
M
(%i71) M2:Mun/Mud
(%o71) -2
*
m
*
(-cos(th)2
*
v+v-cos(th)+1)/((cos(th)
*
v+1)
*
M)
(%i72) " Add amplitudes and simplify trig functions "
(%i73) Mfi:trigsimp(M2+M1)
(%o73) (4
*
m
*
v+4
*
m)/((cos(th)2
*
v2-1)
*
M)
(%i74) " factor numerator and pull out minus sign in denom "
(%i75) Mfi_n:factor(num(Mfi))
(%o75) 4
*
m
*
(v+1)
(%i76) Mfi_d:pullfac(expand(denom(Mfi)),-M)
(%o76) -(1-cos(th)2
*
v2)
*
M
(%i77) Mfi:Mfi_n/Mfi_d
(%o77) -4
*
m
*
(v+1)/((1-cos(th)2
*
v2)
*
M)
(%i78) (display2d:true,
disp("CASE: e(-,p1,+1) e(+,p2,+1) --> gamma(k1,+1) gamma(k2,+1) "),
display(Mfi),display2d:false)
CASE: e(-,p1,+1) e(+,p2,+1) --> gamma(k1,+1) gamma(k2,+1)
4 m (v + 1)
Mfi = - -------------------
2 2
(1 - cos (th) v ) M
(%i79) " which agrees with Schwinger, PSF I, (3-13-98)"
(%i80) "----------------------------------------------"
(%i81) "CASE: e(-,p1,+1) e(+,p2,+1) --> gamma(k1,-1) gamma(k2,-1) "
(%i82) " -------------------------------------------------"
(%i83) comp_def(e1_cc(0,1/sqrt(2),%i/sqrt(2),0),
e2_cc(0,(-1)/sqrt(2),%i/sqrt(2),0))
(%i84) listarray(e1_cc)
(%o84) [0,1/sqrt(2),%i/sqrt(2),0]
(%i85) listarray(e2_cc)
(%o85) [0,-1/sqrt(2),%i/sqrt(2),0]
(%i86) " the spinors and denominators are the same "
(%i87) " recompute the g1 and g2 matrices "
(%i88) g1:sL(e2_cc) . (m
*
I4-sL(k1)+sL(p1)) . sL(e1_cc)
(%o88) matrix([0,0,0,0],[0,2
*
m,0,-2
*
(cos(th)
*
v
*
M/2-M/2)],[0,0,0,0],
[0,-2
*
(M/2-cos(th)
*
v
*
M/2),0,2
*
m])
(%i89) g2:sL(e1_cc) . (m
*
I4-sL(k2)+sL(p1)) . sL(e2_cc)
(%o89) matrix([2
*
m,0,-2
*
(-cos(th)
*
v
*
M/2-M/2),0],[0,0,0,0],
[-2
*
(cos(th)
*
v
*
M/2+M/2),0,2
*
m,0],[0,0,0,0])
(%i90) " ------------------------------------"
(%i91) " M1 = Mtn/Mtd "
70
(%i92) Mtn:vp2b . g1 . up1
(%o92) cos((%pi-th)/2)
*
sqrt((v+1)
*
M/2)
*
(2
*
m
*
sin(th/2)
*
sqrt((v+1)
*
M/2)
-2
*
sin(th/2)
*
sqrt(-(v-1)
*
M/2)
*
(M/2-cos(th)
*
v
*
M/2))
-cos((%pi-th)/2)
*
sqrt(-(v-1)
*
M/2)
*
(2
*
m
*
sin(th/2)
*
sqrt(-(v-1)
*
M/2)
-2
*
sin(th/2)
*
sqrt((v+1)
*
M/2)
*
(cos(th)
*
v
*
M/2-M/2))
(%i93) expand(ratsubst(sin(th/2),cos((%pi-th)/2),%))
(%o93) sin(th/2)2
*
cos(th)
*
sqrt(1-v)
*
v
*
sqrt(v+1)
*
M2
-sin(th/2)2
*
sqrt(1-v)
*
sqrt(v+1)
*
M2+2
*
m
*
sin(th/2)2
*
v
*
M
(%i94) expand(ratsubst(2
*
m/M,sqrt(1-v2),rootscontract(%)))
(%o94) 2
*
m
*
sin(th/2)2
*
cos(th)
*
v
*
M+2
*
m
*
sin(th/2)2
*
v
*
M-2
*
m
*
sin(th/2)2
*
M
(%i95) pullfac(%,2
*
m
*
M
*
sin(th/2)2)
(%o95) 2
*
m
*
sin(th/2)2
*
(cos(th)
*
v+v-1)
*
M
(%i96) expand(ratsubst((1-cos(th))/2,sin(th/2)2,%))
(%o96) -m
*
cos(th)2
*
v
*
M+m
*
v
*
M+m
*
cos(th)
*
M-m
*
M
(%i97) Mtn:pullfac(%,m
*
M)
(%o97) m
*
(-cos(th)2
*
v+v+cos(th)-1)
*
M
(%i98) M1:Mtn/Mtd
(%o98) -2
*
m
*
(-cos(th)2
*
v+v+cos(th)-1)/((1-cos(th)
*
v)
*
M)
(%i99) "------------------------------"
(%i100) " M2 = Mun/Mud "
(%i101) Mun:vp2b . g2 . up1
(%o101) sin((%pi-th)/2)
*
sqrt((v+1)
*
M/2)
*
(2
*
m
*
cos(th/2)
*
sqrt((v+1)
*
M/2)
-2
*
cos(th/2)
*
sqrt(-(v-1)
*
M/2)
*
(cos(th)
*
v
*
M/2+M/2))
-sin((%pi-th)/2)
*
sqrt(-(v-1)
*
M/2)
*
(2
*
m
*
cos(th/2)
*
sqrt(-(v-1)
*
M/2)
-2
*
cos(th/2)
*
sqrt((v+1)
*
M/2)
*
(-cos(th)
*
v
*
M/2-M/2))
(%i102) expand(ratsubst(cos(th/2),sin((%pi-th)/2),%))
(%o102) -cos(th/2)2
*
cos(th)
*
sqrt(1-v)
*
v
*
sqrt(v+1)
*
M2
-cos(th/2)2
*
sqrt(1-v)
*
sqrt(v+1)
*
M2+2
*
m
*
cos(th/2)2
*
v
*
M
(%i103) expand(ratsubst(2
*
m/M,sqrt(1-v2),rootscontract(%)))
(%o103) -2
*
m
*
cos(th/2)2
*
cos(th)
*
v
*
M+2
*
m
*
cos(th/2)2
*
v
*
M-2
*
m
*
cos(th/2)2
*
M
(%i104) pullfac(%,2
*
m
*
M
*
cos(th/2)2)
(%o104) 2
*
m
*
cos(th/2)2
*
(-cos(th)
*
v+v-1)
*
M
(%i105) expand(ratsubst((cos(th)+1)/2,cos(th/2)2,%))
(%o105) -m
*
cos(th)2
*
v
*
M+m
*
v
*
M-m
*
cos(th)
*
M-m
*
M
(%i106) Mun:pullfac(%,m
*
M)
(%o106) m
*
(-cos(th)2
*
v+v-cos(th)-1)
*
M
(%i107) M2:Mun/Mud
(%o107) -2
*
m
*
(-cos(th)2
*
v+v-cos(th)-1)/((cos(th)
*
v+1)
*
M)
(%i108) " Add amplitudes and simplify trig functions "
(%i109) Mfi:trigsimp(M2+M1)
(%o109) (4
*
m
*
v-4
*
m)/((cos(th)2
*
v2-1)
*
M)
(%i110) " factor numerator and pull out minus sign in denom "
(%i111) Mfi_n:pullfac(num(Mfi),-4
*
m)
(%o111) -4
*
m
*
(1-v)
(%i112) Mfi_d:pullfac(expand(denom(Mfi)),-M)
(%o112) -(1-cos(th)2
*
v2)
*
M
(%i113) Mfi:Mfi_n/Mfi_d
(%o113) 4
*
m
*
(1-v)/((1-cos(th)2
*
v2)
*
M)
(%i114) (display2d:true,
disp("CASE: e(-,p1,+1) e(+,p2,+1) --> gamma(k1,-1) gamma(k2,-1) "),
display(Mfi),display2d:false)
CASE: e(-,p1,+1) e(+,p2,+1) --> gamma(k1,-1) gamma(k2,-1)
4 m (1 - v)
Mfi = -------------------
2 2
(1 - cos (th) v ) M
71
(%i115) " which agrees with Schwinger, PSF I, (3-13-98)"
(%i116) "----------------------------------------------"
(%i117) "CASE: e(-,p1,+1) e(+,p2,+1) --> gamma(k1,1) gamma(k2,-1) "
(%i118) " -------------------------------------------------"
(%i119) comp_def(e1_cc(0,(-1)/sqrt(2),%i/sqrt(2),0))
(%i120) listarray(e1_cc)
(%o120) [0,-1/sqrt(2),%i/sqrt(2),0]
(%i121) " the spinors and denominators are the same "
(%i122) " recompute the g1 and g2 matrices "
(%i123) g1:sL(e2_cc) . (m
*
I4-sL(k1)+sL(p1)) . sL(e1_cc)
(%o123) matrix([0,0,0,0],[0,0,sin(th)
*
v
*
M,0],[0,0,0,0],[-sin(th)
*
v
*
M,0,0,0])
(%i124) g2:sL(e1_cc) . (m
*
I4-sL(k2)+sL(p1)) . sL(e2_cc)
(%o124) matrix([0,0,0,0],[0,0,sin(th)
*
v
*
M,0],[0,0,0,0],[-sin(th)
*
v
*
M,0,0,0])
(%i125) " ------------------------------------"
(%i126) " M1 = Mtn/Mtd "
(%i127) Mtn:vp2b . g1 . up1
(%o127) -2
*
cos((%pi-th)/2)
*
cos(th/2)
*
sin(th)
*
v
*
M
*
sqrt(-(v-1)
*
M/2)
*
sqrt((v+1)
*
M/2)
(%i128) expand(ratsubst(sin(th/2),cos((%pi-th)/2),%))
(%o128) -cos(th/2)
*
sin(th/2)
*
sin(th)
*
sqrt(1-v)
*
v
*
sqrt(v+1)
*
M2
(%i129) expand(ratsubst(2
*
m/M,sqrt(1-v2),rootscontract(%)))
(%o129) -2
*
m
*
cos(th/2)
*
sin(th/2)
*
sin(th)
*
v
*
M
(%i130) Mtn:ratsubst(sin(th)/2,cos(th/2)
*
sin(th/2),%)
(%o130) -m
*
sin(th)2
*
v
*
M
(%i131) M1:Mtn/Mtd
(%o131) 2
*
m
*
sin(th)2
*
v/((1-cos(th)
*
v)
*
M)
(%i132) "------------------------------"
(%i133) " M2 = Mun/Mud "
(%i134) Mun:vp2b . g2 . up1
(%o134) -2
*
cos((%pi-th)/2)
*
cos(th/2)
*
sin(th)
*
v
*
M
*
sqrt(-(v-1)
*
M/2)
*
sqrt((v+1)
*
M/2)
(%i135) expand(ratsubst(sin(th/2),cos((%pi-th)/2),%))
(%o135) -cos(th/2)
*
sin(th/2)
*
sin(th)
*
sqrt(1-v)
*
v
*
sqrt(v+1)
*
M2
(%i136) expand(ratsubst(2
*
m/M,sqrt(1-v2),rootscontract(%)))
(%o136) -2
*
m
*
cos(th/2)
*
sin(th/2)
*
sin(th)
*
v
*
M
(%i137) Mun:ratsubst(sin(th)/2,cos(th/2)
*
sin(th/2),%)
(%o137) -m
*
sin(th)2
*
v
*
M
(%i138) M2:Mun/Mud
(%o138) 2
*
m
*
sin(th)2
*
v/((cos(th)
*
v+1)
*
M)
(%i139) " ----------------------------------------"
(%i140) " Add amplitudes and simplify trig functions "
(%i141) Mfi:trigsimp(M2+M1)
(%o141) -4
*
m
*
sin(th)2
*
v/((cos(th)2
*
v2-1)
*
M)
(%i142) Mfi:num(Mfi)/pullfac(expand(denom(Mfi)),-M)
(%o142) 4
*
m
*
sin(th)2
*
v/((1-cos(th)2
*
v2)
*
M)
(%i143) (display2d:true,
disp("CASE: e(-,p1,+1) e(+,p2,+1) --> gamma(k1,+1) gamma(k2,-1) "),
display(Mfi),display2d:false)
CASE: e(-,p1,+1) e(+,p2,+1) --> gamma(k1,+1) gamma(k2,-1)
2
4 m sin (th) v
Mfi = -------------------
2 2
(1 - cos (th) v ) M
(%i144) " ------------------------------------------------"
(%i145) "----------------------------------------------"
(%i146) "CASE: e(-,p1,+1) e(+,p2,+1) --> gamma(k1,-1) gamma(k2,+1) "
(%i147) " -------------------------------------------------"
(%i148) comp_def(e1_cc(0,1/sqrt(2),%i/sqrt(2),0),
e2_cc(0,1/sqrt(2),%i/sqrt(2),0))
72
(%i149) listarray(e1_cc)
(%o149) [0,1/sqrt(2),%i/sqrt(2),0]
(%i150) listarray(e2_cc)
(%o150) [0,1/sqrt(2),%i/sqrt(2),0]
(%i151) " the spinors and denominators are the same "
(%i152) " recompute the g1 and g2 matrices "
(%i153) g1:sL(e2_cc) . (m
*
I4-sL(k1)+sL(p1)) . sL(e1_cc)
(%o153) matrix([0,0,0,sin(th)
*
v
*
M],[0,0,0,0],[0,-sin(th)
*
v
*
M,0,0],[0,0,0,0])
(%i154) g2:sL(e1_cc) . (m
*
I4-sL(k2)+sL(p1)) . sL(e2_cc)
(%o154) matrix([0,0,0,sin(th)
*
v
*
M],[0,0,0,0],[0,-sin(th)
*
v
*
M,0,0],[0,0,0,0])
(%i155) " ------------------------------------"
(%i156) " ------------------------------------"
(%i157) " M1 = Mtn/Mtd "
(%i158) Mtn:vp2b . g1 . up1
(%o158) -2
*
sin((%pi-th)/2)
*
sin(th/2)
*
sin(th)
*
v
*
M
*
sqrt(-(v-1)
*
M/2)
*
sqrt((v+1)
*
M/2)
(%i159) expand(ratsubst(cos(th/2),sin((%pi-th)/2),%))
(%o159) -cos(th/2)
*
sin(th/2)
*
sin(th)
*
sqrt(1-v)
*
v
*
sqrt(v+1)
*
M2
(%i160) expand(ratsubst(2
*
m/M,sqrt(1-v2),rootscontract(%)))
(%o160) -2
*
m
*
cos(th/2)
*
sin(th/2)
*
sin(th)
*
v
*
M
(%i161) Mtn:ratsubst(sin(th)/2,cos(th/2)
*
sin(th/2),%)
(%o161) -m
*
sin(th)2
*
v
*
M
(%i162) M1:Mtn/Mtd
(%o162) 2
*
m
*
sin(th)2
*
v/((1-cos(th)
*
v)
*
M)
(%i163) "------------------------------"
(%i164) " M2 = Mun/Mud "
(%i165) Mun:vp2b . g2 . up1
(%o165) -2
*
sin((%pi-th)/2)
*
sin(th/2)
*
sin(th)
*
v
*
M
*
sqrt(-(v-1)
*
M/2)
*
sqrt((v+1)
*
M/2)
(%i166) expand(ratsubst(cos(th/2),sin((%pi-th)/2),%))
(%o166) -cos(th/2)
*
sin(th/2)
*
sin(th)
*
sqrt(1-v)
*
v
*
sqrt(v+1)
*
M2
(%i167) expand(ratsubst(2
*
m/M,sqrt(1-v2),rootscontract(%)))
(%o167) -2
*
m
*
cos(th/2)
*
sin(th/2)
*
sin(th)
*
v
*
M
(%i168) Mun:ratsubst(sin(th)/2,cos(th/2)
*
sin(th/2),%)
(%o168) -m
*
sin(th)2
*
v
*
M
(%i169) M2:Mun/Mud
(%o169) 2
*
m
*
sin(th)2
*
v/((cos(th)
*
v+1)
*
M)
(%i170) " --------------------------------"
(%i171) Mfi:trigsimp(M2+M1)
(%o171) -4
*
m
*
sin(th)2
*
v/((cos(th)2
*
v2-1)
*
M)
(%i172) Mfi:num(Mfi)/pullfac(expand(denom(Mfi)),-M)
(%o172) 4
*
m
*
sin(th)2
*
v/((1-cos(th)2
*
v2)
*
M)
(%i173) (display2d:true,
disp("CASE: e(-,p1,+1) e(+,p2,+1) --> gamma(k1,-1) gamma(k2,+1) "),
display(Mfi),display2d:false)
CASE: e(-,p1,+1) e(+,p2,+1) --> gamma(k1,-1) gamma(k2,+1)
2
4 m sin (th) v
Mfi = -------------------
2 2
(1 - cos (th) v ) M
(%i174) " ------------------------------------------------"
(%i175) " Schwingers amplitude for these last two cases differs by a minus sign "
(%i176) " ----------------------------------------------- "
(%i177) " CASES LEPTONS HAVE OPPOSITE HELICITIES: see Schwinger (3-13-97). "
(%i178) "----------------------------------------------"
(%i179) "CASE: e(-,p1,+1) e(+,p2,-1) --> gamma(k1,+1) gamma(k2,-1) "
(%i180) " -------------------------------------------------"
73
(%i181) " recompute vp2b "
(%i182) vp2b:sbar(VV(M/2,M
*
v/2,%pi-th,%pi,-1))
(%i183) " denominators Mtd, Mud remain the same "
(%i184) " recompute g1 and g2 "
(%i185) comp_def(e1_cc(0,(-1)/sqrt(2),%i/sqrt(2),0),
e2_cc(0,(-1)/sqrt(2),%i/sqrt(2),0))
(%i186) listarray(e1_cc)
(%o186) [0,-1/sqrt(2),%i/sqrt(2),0]
(%i187) listarray(e2_cc)
(%o187) [0,-1/sqrt(2),%i/sqrt(2),0]
(%i188) g1:sL(e2_cc) . (m
*
I4-sL(k1)+sL(p1)) . sL(e1_cc)
(%o188) matrix([0,0,0,0],[0,0,sin(th)
*
v
*
M,0],[0,0,0,0],[-sin(th)
*
v
*
M,0,0,0])
(%i189) g2:sL(e1_cc) . (m
*
I4-sL(k2)+sL(p1)) . sL(e2_cc)
(%o189) matrix([0,0,0,0],[0,0,sin(th)
*
v
*
M,0],[0,0,0,0],[-sin(th)
*
v
*
M,0,0,0])
(%i190) " ------------------------------------"
(%i191) " ------------------------------------"
(%i192) " M1 = Mtn/Mtd "
(%i193) Mtn:vp2b . g1 . up1
(%o193) sin((%pi-th)/2)
*
cos(th/2)
*
sin(th)
*
v
*
(v+1)
*
M2/2
-sin((%pi-th)/2)
*
cos(th/2)
*
sin(th)
*
(v-1)
*
v
*
M2/2
(%i194) Mtn:expand(ratsubst(cos(th/2),sin((%pi-th)/2),%))
(%o194) cos(th/2)2
*
sin(th)
*
v
*
M2
(%i195) M1:Mtn/Mtd
(%o195) -2
*
cos(th/2)2
*
sin(th)
*
v/(1-cos(th)
*
v)
(%i196) "------------------------------"
(%i197) " M2 = Mun/Mud "
(%i198) Mun:vp2b . g2 . up1
(%o198) sin((%pi-th)/2)
*
cos(th/2)
*
sin(th)
*
v
*
(v+1)
*
M2/2
-sin((%pi-th)/2)
*
cos(th/2)
*
sin(th)
*
(v-1)
*
v
*
M2/2
(%i199) Mun:expand(ratsubst(cos(th/2),sin((%pi-th)/2),%))
(%o199) cos(th/2)2
*
sin(th)
*
v
*
M2
(%i200) M2:Mun/Mud
(%o200) -2
*
cos(th/2)2
*
sin(th)
*
v/(cos(th)
*
v+1)
(%i201) " -----------------------------"
(%i202) trigsimp(M2+M1)
(%o202) 4
*
cos(th/2)2
*
sin(th)
*
v/(cos(th)2
*
v2-1)
(%i203) Mfi:(-num(%))/(-denom(%))
(%o203) -4
*
cos(th/2)2
*
sin(th)
*
v/(1-cos(th)2
*
v2)
(%i204) (display2d:true,
disp("CASE: e(-,p1,+1) e(+,p2,-1) --> gamma(k1,+1) gamma(k2,-1) "),
display(Mfi),display2d:false)
CASE: e(-,p1,+1) e(+,p2,-1) --> gamma(k1,+1) gamma(k2,-1)
2 th
4 cos (--) sin(th) v
2
Mfi = - --------------------
2 2
1 - cos (th) v
(%i205) " -------------------------------------------------"
(%i206) " This differs by a sign compared to Schwinger. "
(%i207) " -----------------------------------------------"
(%i208) "CASE: e(-,p1,+1) e(+,p2,-1) --> gamma(k1,-1) gamma(k2,+1) "
(%i209) " -------------------------------------------------"
(%i210) " spinors and denominators are the same , recompute g1 and g2"
(%i211) comp_def(e1_cc(0,1/sqrt(2),%i/sqrt(2),0),
e2_cc(0,1/sqrt(2),%i/sqrt(2),0))
(%i212) listarray(e1_cc)
(%o212) [0,1/sqrt(2),%i/sqrt(2),0]
(%i213) listarray(e2_cc)
(%o213) [0,1/sqrt(2),%i/sqrt(2),0]
74
(%i214) g1:sL(e2_cc) . (m
*
I4-sL(k1)+sL(p1)) . sL(e1_cc)
(%o214) matrix([0,0,0,sin(th)
*
v
*
M],[0,0,0,0],[0,-sin(th)
*
v
*
M,0,0],[0,0,0,0])
(%i215) g2:sL(e1_cc) . (m
*
I4-sL(k2)+sL(p1)) . sL(e2_cc)
(%o215) matrix([0,0,0,sin(th)
*
v
*
M],[0,0,0,0],[0,-sin(th)
*
v
*
M,0,0],[0,0,0,0])
(%i216) " M1 = Mtn/Mtd "
(%i217) Mtn:vp2b . g1 . up1
(%o217) cos((%pi-th)/2)
*
sin(th/2)
*
sin(th)
*
(v-1)
*
v
*
M2/2
-cos((%pi-th)/2)
*
sin(th/2)
*
sin(th)
*
v
*
(v+1)
*
M2/2
(%i218) Mtn:expand(ratsubst(sin(th/2),cos((%pi-th)/2),%))
(%o218) -sin(th/2)2
*
sin(th)
*
v
*
M2
(%i219) M1:Mtn/Mtd
(%o219) 2
*
sin(th/2)2
*
sin(th)
*
v/(1-cos(th)
*
v)
(%i220) Mun:vp2b . g2 . up1
(%o220) cos((%pi-th)/2)
*
sin(th/2)
*
sin(th)
*
(v-1)
*
v
*
M2/2
-cos((%pi-th)/2)
*
sin(th/2)
*
sin(th)
*
v
*
(v+1)
*
M2/2
(%i221) Mun:expand(ratsubst(sin(th/2),cos((%pi-th)/2),%))
(%o221) -sin(th/2)2
*
sin(th)
*
v
*
M2
(%i222) M2:Mun/Mud
(%o222) 2
*
sin(th/2)2
*
sin(th)
*
v/(cos(th)
*
v+1)
(%i223) " -----------------------------"
(%i224) trigsimp(M2+M1)
(%o224) -4
*
sin(th/2)2
*
sin(th)
*
v/(cos(th)2
*
v2-1)
(%i225) Mfi:(-num(%))/(-denom(%))
(%o225) 4
*
sin(th/2)2
*
sin(th)
*
v/(1-cos(th)2
*
v2)
(%i226) (display2d:true,
disp("CASE: e(-,p1,+1) e(+,p2,-1) --> gamma(k1,-1) gamma(k2,+1) "),
display(Mfi),display2d:false)
CASE: e(-,p1,+1) e(+,p2,-1) --> gamma(k1,-1) gamma(k2,+1)
2 th
4 sin (--) sin(th) v
2
Mfi = --------------------
2 2
1 - cos (th) v
(%i227) " -------------------------------------------------"
(%i228) " This differs by a sign compared to Schwinger. "
(%i229) " -------------------------------------------------"
(%i230) "CASE: e(-,p1,+1) e(+,p2,-1) --> gamma(k1,+1) gamma(k2,+1) "
(%i231) " -------------------------------------------------"
(%i232) comp_def(e1_cc(0,(-1)/sqrt(2),%i/sqrt(2),0),
e2_cc(0,1/sqrt(2),%i/sqrt(2),0))
(%i233) listarray(e1_cc)
(%o233) [0,-1/sqrt(2),%i/sqrt(2),0]
(%i234) listarray(e2_cc)
(%o234) [0,1/sqrt(2),%i/sqrt(2),0]
(%i235) g1:sL(e2_cc) . (m
*
I4-sL(k1)+sL(p1)) . sL(e1_cc)
(%o235) matrix([2
*
m,0,-2
*
(M/2-cos(th)
*
v
*
M/2),0],[0,0,0,0],
[-2
*
(cos(th)
*
v
*
M/2-M/2),0,2
*
m,0],[0,0,0,0])
(%i236) g2:sL(e1_cc) . (m
*
I4-sL(k2)+sL(p1)) . sL(e2_cc)
(%o236) matrix([0,0,0,0],[0,2
*
m,0,-2
*
(cos(th)
*
v
*
M/2+M/2)],[0,0,0,0],
[0,-2
*
(-cos(th)
*
v
*
M/2-M/2),0,2
*
m])
(%i237) " M1 = Mtn/Mtd "
(%i238) Mtn:vp2b . g1 . up1
(%o238) cos((%pi-th)/2)
*
sqrt(-(v-1)
*
M/2)
*
(2
*
m
*
cos(th/2)
*
sqrt((v+1)
*
M/2)
-2
*
cos(th/2)
*
sqrt(-(v-1)
*
M/2)
*
(cos(th)
*
v
*
M/2-M/2))
-cos((%pi-th)/2)
*
sqrt((v+1)
*
M/2)
*
(2
*
m
*
cos(th/2)
*
sqrt(-(v-1)
*
M/2)
-2
*
cos(th/2)
*
sqrt((v+1)
*
M/2)
*
(M/2-cos(th)
*
v
*
M/2))
75
(%i239) expand(ratsubst(sin(th/2),cos((%pi-th)/2),%))
(%o239) cos(th/2)
*
sin(th/2)
*
M2-cos(th/2)
*
sin(th/2)
*
cos(th)
*
v
*
M2
(%i240) expand(ratsubst(sin(th)/2,cos(th/2)
*
sin(th/2),%))
(%o240) sin(th)
*
M2/2-cos(th)
*
sin(th)
*
v
*
M2/2
(%i241) Mtn:pullfac(%,sin(th)
*
M2/2)
(%o241) sin(th)
*
(1-cos(th)
*
v)
*
M2/2
(%i242) M1:Mtn/Mtd
(%o242) -sin(th)
(%i243) Mun:vp2b . g2 . up1
(%o243) sin((%pi-th)/2)
*
sqrt((v+1)
*
M/2)
*
(2
*
m
*
sin(th/2)
*
sqrt(-(v-1)
*
M/2)
-2
*
sin(th/2)
*
sqrt((v+1)
*
M/2)
*
(cos(th)
*
v
*
M/2+M/2))
-sin((%pi-th)/2)
*
sqrt(-(v-1)
*
M/2)
*
(2
*
m
*
sin(th/2)
*
sqrt((v+1)
*
M/2)
-2
*
sin(th/2)
*
sqrt(-(v-1)
*
M/2)
*
(-cos(th)
*
v
*
M/2-M/2))
(%i244) expand(ratsubst(cos(th/2),sin((%pi-th)/2),%))
(%o244) -cos(th/2)
*
sin(th/2)
*
cos(th)
*
v
*
M2-cos(th/2)
*
sin(th/2)
*
M2
(%i245) expand(ratsubst(sin(th)/2,cos(th/2)
*
sin(th/2),%))
(%o245) -cos(th)
*
sin(th)
*
v
*
M2/2-sin(th)
*
M2/2
(%i246) Mun:factor(%)
(%o246) -sin(th)
*
(cos(th)
*
v+1)
*
M2/2
(%i247) M2:Mun/Mud
(%o247) sin(th)
(%i248) Mfi:M2+M1
(%o248) 0
(%i249) (display2d:true,
disp("CASE: e(-,p1,+1) e(+,p2,-1) --> gamma(k1,+1) gamma(k2,+1) "),
display(Mfi),display2d:false)
CASE: e(-,p1,+1) e(+,p2,-1) --> gamma(k1,+1) gamma(k2,+1)
Mfi = 0
(%i250) " The amplitude is zero for this case. "
(%o250) "pair2.mac"
12.16 moller3.mac: Squared Polarized Amplitudes Using Symbolic or Explicit Matrix Trace Methods
We provide here examples of using trace methods for the square of a polarized amplitude in the case of arbitrary energy
(nite mass) Moller scattering. This detailed verication using both explicit matrix traces and the symbolic traces takes
several minutes to carry out for each case, with the symbolic nc_tr method (equivalent to noncov (tr(..))) re-
quiring about twice the computation time as the explicit matrix method m_tr.
The new feature is the tr or m_tr argument S(sv, Sp), in which either sv = 1 or sv = -1, and Sp is a symbol
for the particle spin 4-vector associated with the particles 4-momentum.
In the explicit matrix method, such an argument inserts a matrix factor P(sv,Sp), which turns into the matrix
(I4 + sv
*
Gam[5].sL(Sp))/2. This is the matrix version of the nite mass spin projection operator, and requires
a frame dependent denition (using comp_def) of the components of the spin 4-vector corresponding to a particles
4-momentum. ( See our section on high energy physics notation.)
When used with the symbolic tr method, the result is a factor
1
2
(1 +
5
s /), with = 1 being the value of sv and
s / being (Sp)

.
76
Here is the batch le moller3.mac:
/
*
file moller3.mac
m_tr and nc_tr comparisons of
the square of polarized amplitudes
compared with dirac spinor results
found in moller2.mac
*
/
" moller3.mac "$
"
*********************************************
"$
print (" ver: ",_binfo%," date: ",mydate )$
" Maxima by Example, Ch. 12 "$
" Dirac Algebra and Quantum Electrodynamics "$
" Edwin L. Woollett "$
" http://www.csulb.edu/woollett "$
" woollett@charter.net "$
"SQUARED POLARIZED AMPLITUDES VIA SYMBOLIC AND MATRIX TRACE METHODS"$
" FOR ARBITRARY ENERGY MOLLER SCATTERING "$
" e(-,p1,sv1) + e(-,p2,sv2) --> e(-,p3,sv3) + e(-,p4,sv4) "$
" --------------------------------------------------------"$
invar (D(p1,p1) = m2,
D(p2,p2) = m2,
D(p3,p3) = m2,
D(p4,p4) = m2,
D(p1,Sp1) = 0,
D(Sp1,Sp1) = -1,
D(p2,Sp2) = 0,
D(Sp2,Sp2) = -1,
D(p3,Sp3) = 0,
D(Sp3,Sp3) = -1,
D(p4,Sp4) = 0,
D(Sp4,Sp4) = -1 )$
comp_def ( p1( E,0,0,p),
Sp1 (p/m,0,0,E/m),
p2( E,0,0,-p),
Sp2 (p/m,0,0,-E/m),
p3 (E,p
*
sin(th),0,p
*
cos(th)),
Sp3 (p/m,E
*
sin(th)/m,0,E
*
cos(th)/m),
p4 (E,-p
*
sin(th),0,-p
*
cos(th)),
Sp4 (p/m,-E
*
sin(th)/m,0,-E
*
cos(th)/m))$
p_Em (expr) := expand (ratsubst(E2-m2,p2,expr))$
E_pm (expr) := expand (ratsubst (p2 + m2,E2,expr))$
77
s_th : VP (p1+p2,p1+p2);
t_th : VP (p1-p3,p1-p3);
u_th : VP (p1-p4,p1-p4);
t_th2 : to_ao2 (t_th,th);
u_th2 : to_ao2 (u_th,th);
t_thE : p_Em(t_th);
u_thE : p_Em(u_th);
load("MSQcomp.mac")$
and here is the automated comparison code which Moller3.mac loads:
/
*
file MSQcomp.mac
*
/
/
*
comparison code for moller3.mac
*
/
disp (" MSQcomp() to compare matrix trace, symbolic trace
methods for square of polarized amplitudes
with square of spinor amplitudes. This is an
automated comparison, assuming A_spinor is
globally defined and also global sv1,sv2,sv3,sv4.
To see progress details, set details to true.")$
details:false$
/
************************************************************
/
MSQcomp () :=
block(
if not details then (disp (" this comparison could take 3 plus minutes ")),
A_spinor_th : fr_ao2 (A_spinor,th),
A_spSQ : E_pm (A_spinor_th2),
if details then display (A_spSQ),
M1n_m : trigsimp (E_pm (mcon ( m_tr (S(sv3,Sp3),p3+m,mu,S(sv1,Sp1),p1+m,nu)
*
m_tr (S(sv4,Sp4),p4+m,mu,S(sv2,Sp2),p2+m,nu), mu,nu))),
if details then (disp ("M1n_m"), display (M1n_m)),
78
M1n_s : trigsimp (E_pm (mcon ( nc_tr (S(sv3,Sp3),p3+m,mu,S(sv1,Sp1),p1+m,nu)
*
nc_tr (S(sv4,Sp4),p4+m,mu,S(sv2,Sp2),p2+m,nu), mu,nu))),
if details then (disp ("M1n_s"),display (M1n_s)),
M1n_diff : M1n_s - M1n_m,
if details then print (" M1n_s - M1n_m = ",M1n_diff),
if M1n_diff # 0 then (
disp ("M1n symbolic method # matrix method disagree "),
return (M1n_diff)),
M2n_m : trigsimp (E_pm (mcon ( m_tr (S(sv4,Sp4),p4+m,mu,S(sv1,Sp1),p1+m,nu)
*
m_tr (S(sv3,Sp3),p3+m,mu,S(sv2,Sp2),p2+m,nu), mu,nu))),
if details then disp ("M2n_m"),
M2n_s : trigsimp (E_pm (mcon ( nc_tr (S(sv4,Sp4),p4+m,mu,S(sv1,Sp1),p1+m,nu)
*
nc_tr (S(sv3,Sp3),p3+m,mu,S(sv2,Sp2),p2+m,nu), mu,nu))),
if details then disp ("M2n_s"),
M2n_diff : M2n_s - M2n_m,
if details then print (" M2n_s - M2n_m = ",M2n_diff),
if M2n_diff # 0 then (
disp ("M2n symbolic method # matrix method disagree "),
return (M2n_diff)),
M12n_m : trigsimp (E_pm (mcon ( m_tr (S(sv3,Sp3),p3+m,mu,S(sv1,Sp1),p1+m,
nu,S(sv4,Sp4),p4+m,mu,S(sv2,Sp2),p2+m,nu),mu,nu))),
if details then disp ("M12n_m"),
M12n_s : trigsimp (E_pm (nc_tr (S(sv3,Sp3),p3+m,mu,S(sv1,Sp1),p1+m,
nu,S(sv4,Sp4),p4+m,mu,S(sv2,Sp2),p2+m,nu))),
if details then disp ("M12n_s"),
M12n_diff : M12n_s - M12n_m,
if details then print (" M12n_s - M12n_m = ",M12n_diff),
79
if M12n_diff # 0 then (
disp ("M21n symbolic method # matrix method disagree "),
return (M21n_diff)),
M21n_m : trigsimp (E_pm (mcon ( m_tr (S(sv4,Sp4),p4+m,mu,S(sv1,Sp1),p1+m,
nu,S(sv3,Sp3),p3+m,mu,S(sv2,Sp2),p2+m,nu),mu,nu))),
if details then disp ("M21n_m"),
M21n_s : trigsimp (E_pm (nc_tr (S(sv4,Sp4),p4+m,mu,S(sv1,Sp1),p1+m,
nu,S(sv3,Sp3),p3+m,mu,S(sv2,Sp2),p2+m,nu))),
if details then disp ("M21n_s"),
if details then print (" M21n_s - M21n_m = ",M21n_s - M21n_m),
M21n_diff : M21n_s - M21n_m,
if details then print (" M21n_s - M21n_m = ",M21n_diff),
if M21n_diff # 0 then (
disp ("M21n symbolic method # matrix method disagree "),
return (M21n_diff)),
MfiSQ : expand (M1n_m/t_th2 + M2n_m/u_th2 - M12n_m/(t_th
*
u_th)
- M21n_m/(t_th
*
u_th)),
MfiSQ : trigsimp (MfiSQ),
MSQ_diff : trigsimp (MfiSQ - A_spSQ),
if MSQ_diff = 0 then disp (" agreement ")
else disp (" no agreement "),
display ([sv1,sv2,sv3,sv4]))$
and here are some cases worked out. After loading dirac2.mac and batching moller3.mac (which loads also
MSQcomp.mac), interactive work begins with input (%i28) below:
(%i35) load(dirac2);
(%o1) "c:/work5/dirac2.mac"
(%i2) batch ("moller3.mac");
read and interpret file: #pc:/work5/moller3.mac
(%i3) " moller3.mac "
(%i4) "
*********************************************
"
80
(%i5) print(" ver: ",_binfo%," date: ",mydate)
ver: Maxima 5.23.2 date: 2011-04-13
(%i6) " Maxima by Example, Ch. 12 "
(%i7) " Dirac Algebra and Quantum Electrodynamics "
(%i8) " Edwin L. Woollett "
(%i9) " http://www.csulb.edu/woollett "
(%i10) " woollett@charter.net "
(%i11) " SQUARED POLARIZED AMPLITUDES VIA SYMBOLIC AND MATRIX TRACE METHODS"
(%i12) " FOR ARBITRARY ENERGY MOLLER SCATTERING "
(%i13) " e(-,p1,sv1) + e(-,p2,sv2) --> e(-,p3,sv3) + e(-,p4,sv4) "
(%i14) " --------------------------------------------------------"
(%i15) invar(D(p1,p1) = m2,D(p2,p2) = m2,D(p3,p3) = m2,D(p4,p4) = m2,
D(p1,Sp1) = 0,D(Sp1,Sp1) = -1,D(p2,Sp2) = 0,D(Sp2,Sp2) = -1,
D(p3,Sp3) = 0,D(Sp3,Sp3) = -1,D(p4,Sp4) = 0,D(Sp4,Sp4) = -1)
(%i16) comp_def(p1(E,0,0,p),Sp1(p/m,0,0,E/m),p2(E,0,0,-p),Sp2(p/m,0,0,(-E)/m),
p3(E,p
*
sin(th),0,p
*
cos(th)),
Sp3(p/m,E
*
sin(th)/m,0,E
*
cos(th)/m),
p4(E,-p
*
sin(th),0,-p
*
cos(th)),
Sp4(p/m,(-E
*
sin(th))/m,0,(-E
*
cos(th))/m))
(%i17) p_Em(expr):=expand(ratsubst(E2-m2,p2,expr))
(%i18) E_pm(expr):=expand(ratsubst(m2+p2,E2,expr))
(%i19) s_th:VP(p2+p1,p2+p1)
(%o19) 4
*
E2
(%i20) t_th:VP(p1-p3,p1-p3)
(%o20) 2
*
p2
*
cos(th)-2
*
p2
(%i21) u_th:VP(p1-p4,p1-p4)
(%o21) -2
*
p2
*
cos(th)-2
*
p2
(%i22) t_th2:to_ao2(t_th,th)
(%o22) -4
*
p2
*
sin(th/2)2
(%i23) u_th2:to_ao2(u_th,th)
(%o23) 4
*
p2
*
sin(th/2)2-4
*
p2
(%i24) t_thE:p_Em(t_th)
(%o24) 2
*
cos(th)
*
E2-2
*
E2-2
*
m2
*
cos(th)+2
*
m2
(%i25) u_thE:p_Em(u_th)
(%o25) -2
*
cos(th)
*
E2-2
*
E2+2
*
m2
*
cos(th)+2
*
m2
(%i26) load("MSQcomp.mac")
" MSQcomp() to compare matrix trace, symbolic trace
methods for square of polarized amplitudes
with square of spinor amplitudes. This is an
automated comparison, assuming A_spinor is
globally defined and also global sv1,sv2,sv3,sv4.
To see progress details, set details to true."
(%o27) "moller3.mac"
(%i28) /
*
case RR --> LL
*
/
[sv1,sv2,sv3,sv4]:[1,1,-1,-1]$
(%i29) A_spinor : 2
*
m2/p2$
(%i30) MSQcomp();
" each case comparison could take 2 plus minutes "
" agreement "
[sv1,sv2,sv3,sv4] = [1,1,-1,-1]
(%o30) done
(%i31) time(%);
(%o31) [115.91]
(%i32) /
*
case RR --> RL
*
/
[sv1,sv2,sv3,sv4]:[1,1,1,-1]$
(%i33) A_spinor : m
*
sin(th/2)
*
E/(p2
*
cos(th/2))-m
*
cos(th/2)
*
E/(p2
*
sin(th/2))$
(%i34) MSQcomp();
" each case comparison could take 2 plus minutes "
" agreement "
[sv1,sv2,sv3,sv4] = [1,1,1,-1]
(%o34) done
81
(%i35) time(%);
(%o35) [120.71]
(%i36) /
*
case RR --> RR
*
/
[sv1,sv2,sv3,sv4]:[1,1,1,1]$
(%i37) A_spinor : m2
*
sin(th/2)2/(p2
*
cos(th/2)2)+2
*
sin(th/2)2/cos(th/2)2
+m2
*
cos(th/2)2/(p2
*
sin(th/2)2)
+2
*
cos(th/2)2/sin(th/2)2+4$
(%i38) MSQcomp();
" each case comparison could take 2 plus minutes "
" agreement "
[sv1,sv2,sv3,sv4] = [1,1,1,1]
(%o38) done
(%i39) time(%);
(%o39) [122.34]
(%i40) /
*
case RL --> RL
*
/
[sv1,sv2,sv3,sv4]:[1,-1,1,-1]$
(%i41) A_spinor : m2
*
cos(th/2)2/(p2
*
sin(th/2)2)+
2
*
cos(th/2)2/sin(th/2)2-m2/p2$
(%i42) MSQcomp();
" each case comparison could take 2 plus minutes "
" agreement "
[sv1,sv2,sv3,sv4] = [1,-1,1,-1]
(%o42) done
(%i43) time(%);
(%o43) [122.5]
(%i44) /
*
case RL --> LR
*
/
[sv1,sv2,sv3,sv4]:[1,-1,-1,1]$
(%i45) A_spinor : m2
*
sin(th/2)2/(p2
*
cos(th/2)2)+
2
*
sin(th/2)2/cos(th/2)2-m2/p2$
(%i46) MSQcomp();
" each case comparison could take 2 plus minutes "
" agreement "
[sv1,sv2,sv3,sv4] = [1,-1,-1,1]
(%o46) done
(%i47) time(%);
(%o47) [123.52]
12.17 List of Dirac Package Files and Example Batch Files
The Dirac package consists of
dirac2.mac: This le loads in the other les and includes some utility functions and program startup settings.
simplifying-new.lisp: This lisp code le has been modied slightly (by Maxima developer Barton Willis)
from the le simplifying.lisp which appears in .../share/contrib. This code denes a recursive
method to progressively simplify expressions. We list the functions which make use of this recursive simplication
method in the last section of this chapter.
dgcon2.mac: This le has code for symbolic Lorentz index contraction.
dgtrace2.mac: This le has code for symbolic traces of products of gamma matrices.
dgeval2: This le has code for frame dependent evaluation, such as noncov.
dgmatrix2.mac: This le has most of the code used for explicit Dirac spinor and matrix calculations.
The le dgfunctions2.txt has an semi-alphabetical listing of most of the Dirac package functions, together with a
reference to the le in which the code can be found. There are eleven example batch les discussed in this chapter. In
alphabetical order these are:
bhabha1.mac: High energy electron-positron scattering.
82
bhabha2.mac: Arbitrary energy electron-positron scattering.
compton0.mac: Photon plus scalar charged particle scattering.
compton1.mac: Photon plus Lepton scattering.
moller0.mac: Scalar plus scalar charged particle scattering.
moller1.mac: High energy limit of electron plus electron scattering.
moller2.mac: Arbitrary energy electron plus electron scattering.
moller3.mac: Use of explicit matrix trace methods and symbolic trace methods for the squared polarized am-
plitudes for arbitrary energy electron plus electron scattering.
pair1.mac: Unpolarized two photon annihilation of an electron-positron pair.
pair2.mac: Polarized two photon annihilation of an electron-positron pair.
photon1.mac: Sum identities for photon polarization 3-vectors.
In addition to the above les, there is the le MSQcomp.mac which denes a function used by (and loaded by)
moller3.mac.
12.18 Test Suite Files
There are sixteen test suite les, and once dirac2.mac has been loaded, there are convenient abbreviations available to
run each of these sixteen test les separately.
dgintro2-test.mac: run using tin(),
dgtrace2-test.mac: run using tt1(),
dgtrace2_test.mac: run using tt2(),
dgcon2-test.mac: run using tcon(),
dgmcon-test.mac: run using tmcon(),
dgeval2-test.mac: run using teval(),
bhabha1-test.mac: run using tb1(),
bhabha2-test.mac: run using tb2(),
compton0-test.mac: run using tc0(),
compton1-test.mac: run using tc1(),
moller0-test.mac: run using tm0(),
moller1-test.mac: run using tm1(),
moller2-test.mac: run using tm2(),
moller3-test.mac: run using tm3(),
pair1-test.mac: run using tp1(),
pair2-test.mac: run using tp2()
83
When you make your own modications to the Dirac package (perhaps to t your own research goals), you can assure
yourself that your new code has not broken anything signicant by running these sixteen test les. Once you have loaded
dirac2.mac, you can continue to use the above abbreviations to run one test le at a time, and you do not need to
manually reload dirac2.mac.
To run a test le manually (without using the abbreviations dened in dirac2.mac), you would need two lines for each
test le. Here is an example:
load ("dirac2.mac");
batch ("dgintro2-test.mac", test );
12.19 Using simplifying-new.lisp
The following functions are based on the use of simplifying-new.lisp:
simp_tr1: used when tr1 is called by TR1 to symbolically evaluate the trace of a product of Dirac matrices,
simp_Gexpand1: used when Gexpand1 is called by Gexpand,
simp_Dexpand1: used when Dexpand1 is called by Dexpand,
simp_scon1: used when scon1 is called by scon,
simp_VP1: used when VP calls VP1.
After the denition of, for example, simp_VP1, there is a line of code which uses a Maxima level function dened in
the code le simplifying-new.lisp:
simplifying (VP1, simp_VP1)
This causes a call to VP1 to be taken over by simp_VP1, which is an ordinary Maxima function which can call VP1
itself, leading to a recursive process. Each new entry into simp_VP1 is then able to apply different sorts of massaging
of the supplied expression based on the current state of that simplication.
At the top of the le simplifying-new.lisp is a simple example of using this code.
I am grateful to Maxima developer Barton Willis for suggesting the use of this method and for providing valuable advice
with details.
The Maxima code le ...share/contrib/simplifying.lisp was written by Maxima developer Stavros Macrakis,
and will eventually be updated to the form of our package le simplifying-new.lisp, according to my sources.