PHYSICAL REVIEW D VOLUME 45, NUMBER 5 1 MARCH 1992

Feynman rules for Majorana-neutrino interactions

J. Gluza and M. Zrafek*

Department of Field Theory and Particle Physics, Institute of Physics, Uniuersity of Silesia,
40-007 Katomice, ul. Unimersytecka 4, Poland
(Received 25 July 1991)
Simple Feynman rules for Majorana neutrinos and Dirac fermions interacting with spin-1 or spin-0
bosons are presented. Several examples using these rules are given.

PACS number(s): 13.15.—f, 11.15.Bt, 13.10.+ q

I. INTRODUCTION Sec. III the Feynman technique is presented. At the be-

ginning of this section we consider a very simple example
In perturbation field theory the Feynman-diagram which explains why it is possible to formulate explicit
technique is usually applied. This technique is very well Feynman rules in spite of the multiplicity of vertices and
known for neutral and charged spin-0 and spin-1 bosons propagators. In Sec. IV we give certain examples using
and for charged spin- —,' fermions while for neutral spin- —,' our rules.
fermions the diagrammatic methods are not fully
developed. We know of only three papers which touch II. MA JORANA-NEUTRINO INTERACTIONS
on the problem [ I —3]. However, the presented Feynman
methods are complicated and in many papers which de- We will consider only that part of any gauge theory
which describes the Majorana-neutrino interactions. The
scribe Majorana-neutrino interactions the authors prefer
to use the Wick theorem [4]. The Feynman rules for Ma- other parts of the gauge theory Lagrangian are not im-
jorana particles differ from those of Dirac fermions due portant for our purpose, and give Feynman rules which
to the presence of several different propagators, a related are well known. We take the interactions of Majorana
neutrinos (N) with charged leptons (1) and charged ( W~)
multiplicity of Feynman vertices, and the problem in-
volved in the relative signatures of the various diagrams
or neutral (Z ) gauge bosons in the form
contributing to a given amplitude [5]. The self-conjugacy L~~ =NI I IS'„+ +l I I NJY„ (2. l)
of Majorana fermions is responsible for all these
differences. On the other hand, the Majorana neutrinos L o=(NI "N+lI", l)Z„, (2.2)
appear in a number of proposed extensions of the stan-
dard model so it becomes important to have simple di- where
agrammatic rules to calculate cross sections. The num- f'~~. ~=r"(PL AL +PR Az"'), x =l, N, lN,
ber of necessary propagators and related Feynman ver-
tices was reduced already in Ref. [3] but the applications and
of their Feynman rules are still complicated. First of all,
the authors of Ref. [3] give the rule for an absolute signa- f'i = xof'i'—
ro=r"(Pi AL +Pg AR
ture of each Feynman diagram which depends on the or-
der of fermion operators in the Green's function. In
PL, (it)= 2(i+l's) .
Refs. [l —3], the Majorana representation for the Dirac The parameters AL z are real for x =N, lN and gen-
y" matrices and spinors is used; therefore, it is not obvi- erally complex for x = I and depend on the theory under
ous which simplifications originate from the proposed investigation. In the case of mixing between generations
method and which from the spin representation. the AL R quantities are matrices, but this does not com-
In this paper we propose a Feynman-diagram tech- plicate our approach to the Feynman rules.
nique for Majorana neutrinos which is independent of the In a similar way we consider the Majorana neutrino in-
spinor representation and does not attempt to give a rule teraction (N) with the charged (H —) and neutral (H )
for the absolute sign of each diagram. We found a simple spin-0 Higgs particles:
method which makes it possible to predict the relative
signs between different diagrams, which are important in L + =NI, IH++I I,NH (2.3)
practical calculations. The proposed Feynman-diagram
technique for neutral spin- —, fermions is very simple and L =(Nl N+lI i l)H (2.4)
we hope it will find applications in practice. where
In the next section we describe the Majorana-neutrino
interactions for which we have found Feynman rules. In I )=PLBL"'+P~B~ ' for x =I, N, IN,
(

I = y' I Iyo—
I— = P~BI(I)* +P B (I)*
'Electronic address: zralek Nj plktus 11. and

45 1693 1992 The American Physical Society

1694 J. GLUZA AND M. ZRAKEK 45

B'"'=B'
L R
' for x =N, lN, I ~„~
—Cl ~„~C
T '= "(P„A'"'+PI AR"'),
y— (2.7}

and as previously BL"l'„] are theory-dependent numbers and

(or inatrices). The Majorana fields
have a plane-wave decomposition,
are self-conjugate and
r, .„=—cr,'. c-'= r, „,
, . (2. 8)

d k
N (x)= g n f (2n. ) 2E [u (k, )(. )a(k, A, )e
III. THE FEYNMAN RULES
+ U (k, A, )a (k, A, )e'""], We will present our Feynman rules for the Majorana
(2.5) fermion interactions given by (2. 1) —(2.4). On Feynman
diagrams the Dirac fermion will be depicted by a double
N (x):—CN (x)C '=C &[N (x)]&=N (x), continuous line ( ), the Majorana fermion by a
single line ( and other particles will be depicted
&

where C & is the 4X4 charge-conjugate matrix in Dirac traditionally.

space The problems arising with the Feynman rules for Ma-
c~"c '= —y&,T ct=c ', c = —c . (2.6) jorana neutrinos are connected with the fact that Majora-
na fields satisfy the self-conjugate condition (2.5). The
The Feynman vertices will be described by the I ~„] or basic propagator for Majorana neutrinos is the same as
I ] matrices and their charge-conjugate quantities
~
for Dirac fermions:

x, a
= iS p(x
&OIT[N (x)Np(y)]IO&— y)= (3. 1)

where
"4x
d x —r'(x —yg k +m
f (2~) k —m +i@
and describe the fermion created at y and annihilated at x. Using the relations N =NC and N =C 'N valid for
Majorana particles (2.5} we can define the other three propagators,
ygP
= —i[S(x —y)C]
cx

(O~T[N (x)N&(y)]~0) &= (3.2}

describing the Majorana neutrino annihilated at x and y,

x, a yiP
(0~T[N (x)N&(y)]~0): i[C —
'S(x —y)] &= (3.3}

where the particle is created at x and y, and finally

X, a yiP
(O~T[N (x —
(x)N&(y)]~0) = i[C 'S— y)C] &= (3.4)

which describes the particle created at x and annihilated at y. For a given interaction between fermions and a gauge
bo-
son there are also four types of vertices. To show this, let us take the neutrino-charged-lepton interaction given by
(2.1). Using the relations N =CN and N= NC ', the La—
grangian can be written in the form

L = ,'[l&r~& N W„+N I i—'&1&W„++1&(rlC)& NrW„+Nr( —C 'li') &l&W„+] . (3.5)

Hence four types of vertices which are connected successively with the terms of the Lagrangian (3.5) can be defined as

(3.6)

(3.7)
 FEYNMAN RULES FOR MAJORANA-NEU'M. INO INTERACTIONS 1695

(3.8)

and the last term

(3.9)

At first sight this appears to be complicated because it seems to be necessary to remember the particle-antiparticle
direction of fiow on the fermion lines. Fortunately, we shall see that the additional C matrices which appear in the ver-
tices and propagator all cancel and hence we are able to formulate simple Feynman rules for Majorana neutrinos. To
present the basic idea we will consider one simple example. The lepton-number-violating process e 8'+ ~e+ 8' is
described by one diagram which can be depicted in four different versions:

=[il "] [iC 'S(x —y)] p[ir "]ps

Xy
= [iI "]T[ic 'S(x —y)][iI'"], (3.10)

x y =[ —iC 'I "] [ iS(x —y)C][ iC I "],

— (3.11)

W y =[ir"]'[—ic 'S(x —y)C][ ic 'r" ],—- (3.12)

and

x& Xy
= —[iC 'I "] [iS(x —y)][iI "] . (3.13)

We may see that in all four cases we end up with the same amplitude (C = —C):
+ +~- ——iC '[rc]S(x —y)[r"] . (3. 14)

After appropriate attribution of spinors to external fermion lines the C matrix can be eliminated from the amplitude.
1696 J. GLUZA AND M. ZRAKEK 45

In the full amplitude the C matrix (3. 14) is sandwiched with spinors, u for the incoming electron and v for the outgo-
ing positron; then using the relation — u C '=U we get

T = —
iu (e )C '[I c]S(k)[I "]u(e+)=iv(e )[I c]S(k)[1"]u(e+) . (3. 15)

Hence, irrespective of which propagator (3. 1) —(3.4) is used, with correct choice of vertices (3.6) —(3.9) and suitable attri-
bution of spinors to the external lines, we obtained for the amplitude one simple expression (3. 15). Now we will formu-
late the rules for Majorana-neutrino interactions (2. 1)—(2.4) which can be used for a rapid calculation of the amplitude
for any process.
At the beginning of our procedure we need to attribute spinors to fermion lines on the given Feynman diagram (not
only to external lines). The propagators for the internal fermion will be built from spinors as described below.

Spinors

For the Majorana-Dirac coupling with charged bosons ( W , H ) —the — spinors' attribution to the Majorana line de-
pends on the nature of the Dirac line.
(a) For the incoming Dirac particle (antiparticle) the outgoing Majorana fermion must also be treated as a particle
(antiparticle):
Dirac particle (antiparticle) u(v)

W
+
or H
+ (3.16)

(b) For the outgoing Dirac fermion the fermion "nature" of the vertex is also conserved and for the outgoing Dirac
particle (antiparticle) the incoming Majorana fermion must be treated as a particle (antiparticle):
u(u) Dirac particle (antiparticle)

W
+
or H
+ (3. 17)

e see that for the Dirac-Majorana transition the attribution of spinors to Majorana lines is de6nitive the Dirac —
"particle nature" is remembered in the vertex. This is not so in the case of Majorana-Majorana coupling with the neu-
tral bosons (Zo, Ho). ~e can treat the Majorana fermion as a particle or an antiparticle, and the final result will be in-
dependent of this choice.
(c)
u(u)
Z or H (3. 18)
0 0

The situation sometimes arises where the "fermion flow" is opposite to the momentum flow on a line. In such a case
we use the standard relation for spinors:
(d)
u(+k)=v(+k) . (3. 19)

Vertices

Contrary to previous approaches, our vertices are independent of the direction of the "ferrnion flow" on the line. We
have only one vertex for the given Lagrangian as in the Dirac fermion case.
—
(a) For Dirac-Majorana fermion coupling with the bosons with spin 1 ( W ) or spin 0 (H )we have—
45 FEYNMAN RULES FOR MAJORANA-NEUTRINO INTERACTIONS 1697

r,
(
iI ~& for outgoing W or incoming 8'+,
+
8'
(

W or H
for outgoing 8'+ or incoming
(3.20)
iI &
for outgoing H or incoming H+,
for outgoing H+or incoming H

(b) For two Majorana fermions' coupling with the neutral bosons Z or H we obtain

p ., H i(I ~+I'~Nc) for Zo,

I'

(3.21)
i(I ~+I ~c) for H

As we see, there is no "particle flow" indication on the operators inside the Green's function [3]. However, the
vertices' lines. This means that the vertices are the same amplitude is normally used to calculate a cross section;
for particles and antiparticles, and are independent of the then we need only the relative sign between various Feyn-
direction of the "particle flow" if they flow in the same man diagrams. This relative sign problem can be
direction. When the relation (3.19) is used, the vertices
with opposite "particle fiow" (e.g. ,
) are not needed.
= or : resolved as follows.
(a) We choose any Feynman diagram which we call the
reference diagram. In its amplitude, Dirac and Majorana
fermions appear in the fermion chain in a given order.
(b) We compare all the other diagrams with the refer-
Propagators ence one. We permute the fermions in their "fermion
chains" to obtain the same order as in the reference dia-
In the proposed Feynman technique we need only one gram.
propagator for internal Majorana particles, the same as (c} The signature of each diagram in the full amplitude
for the Dirac fermion. To eliminate the problem con- depends on the fermion order. If parity of the permuta-
nected with the sign ambiguity [3] the propagators taken tion is even (odd), the sign of the diagram in the full am-
will be made of spinors. plitude remains unchanged (is changed). In this way we
(a) For the internal Majorana line, we may choose any have resolved the relative signature problem.
direction of momentum k. The final result will be in- The Lagrangians (2. 1)-(2.4) are usually a part of some
dependent of the direction of k (taking into account the gauge-independent theory. The Feynman rules for bo-
sign convention). sons remain unmodified. For Dirac fermions the same
(b) We use the relation of type rules as described above for the Majorana particles can be
v20u& = —v&O&u2, u20u& = —v&O&v2, (3.22) applied.
We now apply the presented Feynman technique in or-
for 0 = ( I'"„,I „, . . . ) and Oz =—CO C der to calculate amplitudes in a few simple examples.
Then, using relations (3.22) we transform the ampli-
tude in order to get an expression of u (k)u(k) type and
instead we put the normal Dirac propagator IV. EXAMPLES
u(k, A, )u(k, A, ) i(0+m)
k m+ie kz —
m+ie—
z A. Electron-neutrino scattering e N ~e N
, ;„~z~
(3.23} Three diagrams given in Fig. 1 describe the process on
the tree level. If we compare the process with neutrino
(c) If we use relations (3.22) n times, in order to find the (v, ) or antineutrino (v, ) scattering with electrons, in the
propagators (3.23) for Majorana or Dirac particles, we former case the A and A2 diagrams contribute, and in
multiply the amplitudes by ( —)", which is equivalent to
&

the latter, only A, and A3 are present. Our rules tell us

disregarding the minus sign in formulas (3.22). If both
that in the Az diagram the Majorana neutrino should be
spinors 1 and 2 in (3.22) describe the external particles,
considered as a particle but in A 3 as an antiparticle and
we have to take into account the minus sign.
hence we have

A, =u(3)[iI ~i~]u (1)iD„„(k,—k3)u(4)

Sign convention
X [i(r„"+r„",)]u(2), (4. 1)
Using the proposed rules for each Feynman diagram
we obtain some analytical expression which contributes A =u(4)[iI ~&]u(1)iD„(k, —k )u(3)
to the full amplitude with an unknown sign. To find the
absolute sign we need to know the order of fermion X[iI (]u(2), (4.2)
1698 J. GLUZA AND M. ZRAKEK 45

A
3

FIG. 1. Feynman diagrams for electron-Majorana-neutrino interaction e N ~e N.

A =u( —k )[iI", ]u(1)iD„„(k,+k )u(3) B2=u(1)[iI ~1]iS (kz —k3)[iI tc]u(2)e„"(4)p„*(3).

X[iI (]u( —k~) (4.7)
=V(2)[il ", ]u (1)iD„„(k,+k )u(3) And for B2 we have

X [iI,"]u (4) . (4.3) B =u(1)[iI ", ]iS (k —k )[iI'", ]u(2)e„'(4)e„'(3) .

To find the relative sign between the amplitudes A; As in both amplitudes, the external fermions ordering is
(i =1,2, 3) letus analyze the order of the fermions in the identical and we have
"ferrnion chains. " For diagram A, the order is (3,1,4, 2),
for Az-(4, 1, 3, 2) and for A3-(2, 1, 3, 4}. Let us take di-
M + + ~+~+ —Bj, +B (4.8)

agram A, as the reference point. To find ordering as in

As previously stressed, our method is independent of the
A, we have to make an odd number of transpositions in direction of internal Majorana fermion momentum. Let
A2 and an even number in A3, so that the full amplitude us choose the momentum in the B2 diagram in the oppo-
for the process is
site direction (Fig. 3}. Then we have
-~ —A& —A2+ A3 . (4.4) B' e„'(4)e,*(3)v(1)[il'~(]v (k)u(2)[iI I ]v (
—k) .

In diagram A, we treat the Majorana neutrinos as par-

ticles. Our rules teil us that these neutrinos can also be and hence
considered as antiparticles; then the neutral vertex will be B' =e„*(4)e„'( 3u}(2)[il,"]i S(k, —k )[iI ]u(1} ",
described by
=e„'(4)e„'(3)u (l)C 'C[il " ]C 'CiS (k, —k„)
V(2) [i( I /+ I /c ) ]v (4)
XC-'C[ir "]@4;Wu'(2)= —B, . (4.9)
=v (4)C C [l(r„'+r„"',)]C-'Cu '(2)
= —u(4)[i(I" +1" )]u(2) . (4.5) But also the fermion order in Bz(1,2) is opposite to
that in B2(2, 1); therefore it is in agreement with our sign
If we consider the Majorana neutrino as an antiparticle convention
then the amplitude for the A diagram will have the op-
M B2 =8~+B2
&

posite sign. Taking the same A diagram as a reference + + + + B~

&

we obtain and the final amplitude is independent of the direction of

M' = —A i+ A2 —A3, (4 6) momentum.

which differs in sign from the previous one and has no

meaning for practical calculations.

B. The process e+e+ ~ W+ IY+ k=k -k

k=lt
2 3
2 4
Only two diagrams describe the process in the tree ap-
proximation (Fig. 2}. Let us choose the internal neutrino
momentum from the positron 2~1, as in Fig. 2. Ac-
cording to our rules we have
B
B, u(2) [iI ", ]v (k}v(1 }[iI "]u (
—k) B
1 2

V( l)[il &']u (k)u(k)[iI ~&c]u (2), FIG. 2. Feynman diagrams for two- W+ production process
e+e+~ W+W . We changed the neutrino rnomenturn (from
and hence the positron 2 to the positron 1).
45 FEYNMAN RULES FOR MAJORANA-NEUTRINO INTERACTIONS 1699

k=k k=k -k
1

k -k -k'
1 3

rr k k -k
1 3
B2

FIG. 3. Diagram B2 from previous figure where we change

the direction of the neutrino momentum (k =k& —
D
k4). 1

FIG. 5. Diagram describing the virtual one-loop correction

to the Majorana neutrino propagators in the process
C. Process e e —+e H N e e —+W 8'

Let us consider a more complicated process which, in C r =u(5)[il i]iS'(k +k )[il't' ]u (1)iD„„r(k —k )
the framework of our Lagrangian (2. 1) —(2.4), is described
X u (3 }[il (~ ]u (2) . (4. 15)
by six diagrams given in Fig. 4. Using our rules we get

C, =v(1)[iI I ]iS (k —k, )[iI, ]u(2)iD (k +k )

Taking into account the order of fermions C, (1,2, 3, 5),
C2(3, 1, 5, 2), C3(3, 2, 5, 1), C&(1,2, 3, 5), C~(5, 2, 3, 1), and
X u (3)[iI I ]v (5), (4. 10) C6(5, 1, 3, 2) and C, as the reference diagram, we have
C =u(3)[iI', ]iS (k, —k )[iI, ]u(1)iD (k —k ) M — -H-N = C1 —C2+ C3+ C4
X u(5)[il'(]u (2), (4. 11) —C —C$+C +Cr (4. 16)

C =u(3)[iI I]iS (k —k )[iI', ]u(2)iD (k, —k ) in agreement with the Wick theorem. To reach agree-
ment it is important to use relation (3.22) without the
X u(5)[iI, ]u (1} (4. 12) minus sign in the amplitudes C, and C4.
C =v(1)[il r ]iS (k —k )[iI, ]u(2)iD (k +k )
D. Virtual corrections
Xu(3)[iI', ]v(5), (4. 13) Our rules can be applied in any order of perturbation
C r =u(5)[iI, ]iS'(k +k )[il'r,' ]u(2)iD„, r(k —k, ) theory. As an example, let us calculate only one diagram
(Fig. 5) which presents the one-loop correction to the
X u(3)[il,"~]u ( I ), (4. 14) neutrino propagator in the process e e W 8' ~
H 4 H 4 N 5

q =k
2 1 5

q =k -k q =k -k &(q =k-k
1 4 1 1 1 4 1 2 4

3
III

N
2 e H 4

c1 c2 c3

1 e e 3

q =k -k r&
1 2 4

H 4 e '1 S 2 e e 3

c5 c6
FIG. 4. Six Feynman diagrams which describe the three-body process e e ~e H N. All momenta inside the propagators are
specified.
1700 J. GLUZA AND M. ZRAKEK

I
— )]iS (k, —k, )
D =E (4)e„*(3) f (2') U(2)[iI'", ]iS (k k )[i(l ~ +I ~ )]iS (k')[i(I +I
X [i I ", ]u (1)D iI(k ( —k3 —k') . (4. 17)

The same convention as for charged leptons, known from QED, is applied now for the Majorana fermion; e.g. , a fac-
tor of——,' must be associated vrith each closed Majorana fermion loop.
V. CONCLUSIONS
We present here very simple Feynman rules for Majorana fermions. We have used only one vertex for each Majorana
fermion interaction, independent of the particle or antiparticle Bow. We have also used one Majorana fermion propaga-
tor, identical to the Dirac one. Our rules are so simple that it appears to be advantageous to use them instead of the
Wick theorem.

[I] S. K. Jones and C. H. Llewellyn-Smith, Nucl. Phys. B217, [4] B. Kayser, F. Gibrat-Debu, and F. Perrier, The Physics of
145 (1983). Massi Ue Neutrinos (World Scientific, Singapore, 1989).
[2] H. E. Haber and G. I. Kane, Phys. Rep. 117, 76 (1985). [5] M. Doi, T. Kotani, and E. Takasugi, Prog. Theor. Phys.
[3] E. I. Gates and K. L. Kowalski, Phys. Rev. D 37, 938 Suppl. 83, 1 (1985); S. Bilenky and S. Petcov, Rev. Mod.
(1988). Phys. 59, 671 (1987).