Conserved Vector Current

Abhass Kumar
Abstract
Here we rst derive a two component spinor theory for all spin half fermions as was
done by Feynman and Gell-Mann which then yields a V-A theory for weak interactions
such that parity is violated. The equality of coupling constants in beta decay and muon
decay is used as an argument for a conserved vector current which makes the theory
universal at least in the vector part.
1 Introduction
Experimentally, the coupling constants of beta decay, muon decay and muon capture led
many people to suggest that there is a universal Fermi interaction. The same form of
interaction should explain all such weak interaction processes. This was substantiated by
the discovery of parity violation in beta decay and muon decay which meant a similarity in
the structure of the two interactions. Parity had long been considered a sacred symmetry
which had to always be conserved. Its violation in beta decay came like a shock to many.
This led people like Salam, Landau, Lee and Yang to propose a two component wavefunction
for neutrinos instead of the usual four component. This meant that they will have only one
spin state instead of two. Conventionaly, particles with negative helicity have been called
neutrinos while those with right helicity are anti-neutrinos. This theory requires the mass of
the neutrino to be zero. Such a theory existed before in the form of Weyl spinors for mass-less
particles but was always rejected because of its intrinsic parity violation. In such a theory,
the Dirac equation does not have a mass term and we are left with just three matrices which
anti-commute among themselves, square to unity and are hermitian and complete. These
conditions are met by the three 22 Pauli matrices. Hence the wavefunctions also contain
only two components one a particle and one an anti-particle.
However, this is not the only way that parity violation can be incorporated. In fact if the
interaction does not contain neutrinos then theory will conseve parity even if the experiments
say otherwise. If we have a two component theory for all spin half fermions, then all of them
will intrinsically have parity violating terms such that the weak interaction does not obey the
law of reection symmetry. Essentially, having a two component theory for such fermions
means having a V-A theory for four fermion interactions where V stands for vector and A
stands for axial vector currents. Such a combination of vector-axial vector violates parity
maximally. This two component theory for all spin 1/2 fermions was pursued by Feynman
and Gell-Mann to lead into the conserved vector current hypothesis.
One can obtain the V-A theory also from chirality invariance as done by Sudarshan and
Marshak or by the mass-reversal invariance proposed by Sakurai. In this report, we follow
1
the approach of Feynman and Gell-Mann that of a two component spinor for all spin half
fermions.
Before we proceed further, why was the conserved vector current hypothesis needed in the
rst place? The answer lies in experiments of weak interactions. In O
14
decay in the channel
O
1
4 N
1
4 +e +
e
, the vector coupling constant g
V
was found to be within one percent of
the Fermi coupling constant g

of the muon decay. This agreement was a puzzle because beta

decay is touched by strong interactions as well due to the proton and neutron involved in
the interaction. Neutrons can emit virtual pions so we may expect the coupling constants to
be renormalized. In the muon decay process, there are no such strong interaction eects and
the coupling constant is as it is. Still, we see an equality in the physical coupling constant
of the beta decay and the bare coupling constant of the muon decay. Hence the CVC theory
was proposed based on analogy with electromagnetism where the physical charge e is the
same as the bare charge e and forms a universal coupling constant for all particles. The CVC
theory is very fundamental to modern particle physics and due to its importance, various
experiments have been done which strongly favour this theory.
In the following sections, we give the theory with all its rigour.
2 The two component formalism
The Dirac equation was invented as a relativistic equation for quantum mechanical particles
which came out to be a many particle equation instead of a single particle equation that it
was intended to be. It describes both the particle and its anti-particle in one four component
spinor. The anti-particles correspond to negative energy solutions of the equation. Of the
total four components in the spinor, two are positice energy and two are negative energy
solutions. The two positive energy solutions describe the two spin states of the particle and
similarly for the negative energy solutions. One can ask why should we have four components
as the above explanation is not enough? One could have a second order dierential equation
instead of a rst order as the Dirac equation and then have two component spinors instead
of four.
A spin zero particle obeys the Klein-Gordon equation which is a second order equation which
yields an anti-particle as well as a negative energy solution. We can have a second order
equation for a spin half particle as well and the negative energy solution of such an equation
would be the anti-particle.
One can easily form such an equation if one substitutes (the four component spinor solution
to Dirac equation in the presence of electromagnetic eld) as
=
1
m
(i

+ m) (1)
We put this value for in the Dirac equation such that satises:
2
(i

m)
1
m
(i

+ m) = 0 (2)
(i

m)(i

+ m) = 0
(

m
uA

m
2
) = 0
(

+ A

iA

1
2

) = m
2

_
(i

)(i

)
1
2

_
= m
2
(3)
where

=
i
2
(

) and F

.
The above steps have given us a second order dierential equation but the spinor is still a
four component one. Consider the operator
5
. This operator commutes with

(as it is
independent of the coordinates), A

(as these are just numbers) and

as shown below:
[
5
,

] =
5

5
=
i
2
(
5
(

) (

)
5
)
=
i
2
(
5

5
+

5
)
=
i
2
(
5

+
5

)
= 0 (4)
When two operators commute, they have simultaneous eigen-vectors. There are two eigen-
values of
5
which are 1 which correspond to only two eigen-vectors. Hence the modied
Dirac equation for has only two independent solutions for .
What the above paragraph means is that we have only two independent components in the
, the other two being linear combinations.
Feynman and Gell-Mann dened
5
as
1

0
. We will dene
5
as
0

3
. Hence
what is i
5
for them is
5
for us in this report. We adopt this change for later convenience
when we will see that the form of the interaction is

(1
5
) to make the V-A prescription
more clear.
Since is an eigen-vector of
5
with eigen value 1, we have

5
= (5)
We can set the choice of once and for all from this relation. Let us set it to be such that

5
= . With this choice, we can invert eq. 1 to write in terms of . Multiplying eq.
1 by (1-
5
), we get
(1
5
) = 1/m(1
5
)(i

+ m)
(1
5
) = 1/m((i

+ m) (i

+ m)
5
)
(1
5
) = 1/m((i

+ m) + (i

+ m))
1
2
(1
5
) = (6)
3
Now it can be explicitly seen that will have only two independent component. Suppose
=
_
a
b
_
(7)
Then with the denitions of the matrices as:

0
=
_
1 0
0 1
_

i
=
_
0
i

i
0
_

5
=
_
0 1
1 0
_
(8)
we obtain for =
1
2
(1
5
)
=
_
1
2
(a b)

1
2
(a b)
_
=
_

_
(9)
Thus we see that has only one independent two-component spinor. The other part is
just the additive inverse of the rst part and stands for the negative energy solution of the
corresponding equation.
Using the explicit form for , one may convert eq. 3 into an equation for which is a two
component spinor as was what we aimed for. The explicit form of the elctromagnetic eld
tensor is:
F

=
_
_
_
_
0 E
x
E
y
E
z
E
x
0 B
z
B
y
E
y
B
z
0 B
x
E
z
B
y
B
x
0
_
_
_
_
(10)
Contracting this with
1
2

gives us .(

B + i

E) such that the equation satised by is

_
(i

)
2
+.(

B + i

E)
_
= 0 (11)
Electrodynamics done using eq. 11 gives the same results as those done using the orig-
inal Dirac equation in the presence of electromagnetic eld as both of the equations are
equivalent.
Due to the equivalence of both the four component spinor equation and two component
spinor equation, everything that could be written in terms of one can be written in terms
of the other.
The beta decay interaction contains terms written in the four component fermionic eld as
they are what take part in it. With our two component theory, that interaction can be
written in terms of the new spinor. We use for the calculations but substitute it with
1
2
(1
5
) so that we avoid any derivatives of the eld. Avoiding derivatives of the eld
is not an arbitrary technique that we are using. It has experimental backing. We know
that the four Fermi interaction formulated in terms of is linear in the elds and does not
contain derivatives. One reason to avoid derivatives is that in a point interaction such as
the beta decay at low energies, the interaction expression does not have any propagators as
no particle is being exchanged.
Writing in terms of
1
2
(1
5
) has an added advantage that this does not conserve parity
which is what we actually wanted with our leptons so that in the weak interaction, even if
4
the neutrino is not present, the interaction violates parity. Another fact that comes out of
the theory is that only left handed electrons will take part in the beta decay. This is seen
by acting the operator
5
on the two component eld when expressed in terms of

5
1
2
(1
5
) =
1
2
(
5
1) =
1
2
(1
5
) (12)
This observation made Feynman suggest that the electron in beta decay processes is coupled
through which in the four-component fermion spinor then takes the form

i
C[(a
n
)

O
i
a
p
][(a

O
i
a
e
] (13)
where
n
,
p
,

and
e
are the spinors corresponding to neutron, proton, neutrino and
electron respectively and a =
1
2
(1
5
), a

=
1
2
(1 +
5
). Here O
i
is the operator coupling
the elds.
If our initial choice had been
5
= , everything here would have been reversed. We would
have electrons and neutrinos which are right handed. But that is an equivalent case and
depends on taste. If we dene the elctron and the neutrino, the proton and the neutron to
be the anti-particles and the conventional anti-particles the particles then we would again
have left handed leptons and right handed anti-leptons. As it is, this ambiguity is always
there and we follow some conventions to distinguish the two cases by dening the left handed
lepton as the particle and the right handed lepton as the anti-particle. (with the neutrino
already having a parity non-conserving form)
Consider eq. 13. The operator O
i
can be either scalar (I, the identity matrix), vector (

),
pseudo-scalar (
5
), axial vector (

5
) or tensor (

). Of these, only the vector and the ax-

ial vector survive. When O
i
is scalar, tensor or pseudo-scalar, then: [(a
n
)

O
i
a
p
][(a

O
i
a
e
] =
[

n
a

O
i
a
p
][

O
i
a
e
]. The operator O
i
comes in the form a

O
i
a. When O
i
commutes
with a, a

O
i
a = O
i
a

a = 0 as a

a =
1
4
(1 +
5
)(1
5
) = 0. When O
i
= I, or O
i
=
5
or
O
i
=

, it commutes with a. When O

i
=

,
a

a =
1
2

(1
5
) (14)
and for the axial vector,
a

5
a =
1
2

(
5
1) (15)
Thus the form of the coupling is determined from here to be composed of the vector and
the axial vector only. Notice that apart from the sign, the vector and the axial vector give
the same coupling.
Since for these two components, a

O
i
a = O
i
a, the eq. 13 can be written as
C[

a
p
][

n
u

a
e
] (16)
In more usual terms with the Fermi coupling constant appearing in the expression, eq. 16
in the most general form for the beta decay is

8g
V
[

a
p
][

n
u

a
e
] (17)
5
Notice however that to get V-A, we have assumed that the proton and the neutron are also
left handed. Otherwise, it would just have been an interaction having V and A terms without
the actual connection of V-A form. The coupling constants could have been arbitrary if we
don not assume the proton and the neutron to be left-handed as well.
3 Other methods to obtain V-A
3.1 Chirality invariance
The chirality transformation is dened as
5
. For a zero mass particle like the
neutrino, two components of the Dirac 4-spinor have positive chirality and two have negative
chirality. For non-zero mass particles, is not an eigen-state of
5
but we may expand it in
terms of such states:
=
+
+

(18)
where

is
1
2
(1
5
).
These spinors are themselves made of two component spinors

. Sudarshan and Marshak

conjectured that the total 4-fermion interaction should remain invariant under a chhirality
transformation of any of the 4-component spinor elds. From such conditions, it was required
that the operator coupling two dierent s should anti-commute with
5
. And of the ve
possible couplings scalar (S), pseudo-scalar (P), tensor (T), vector (V) and axial vector
(A), only V and A anti-commute. Hence the operator could only be a linear combination of
V and A.
3.2 Mass-reversal invariance
Here, we demand that the chirality transformation should be accompanied by a change in
sign of the mass for each fermion simultaneously. The Dirac equation is

= m (19)
Clearly,
5
is not an eigen-function of this equation unless the mass term goes to m i.e.
m m simultaneously.
4 The Conserved Vector Current
As mentioned in the introduction, the beta decay coupling constant is not renormaized even
though loosely speaking, it should be since proton and neutron are not point particles. A
neutron stays in its bare form for a very short time. Rest of the time, it lives like a proton
surrounded by a negatively charged cloud of pions or it lives like a neutron surrounded by a
charge neutral cloud of pions. In the Fermi beta decay theory, only the bare neutron decays
into electron, proton and anti-neutrino of electron type. Hence the coupling constant should
be reduced proportionately. But the experiments tell us otherwise. The coupling constant
does not reduce. In fact it is within 2% of the value of the constant for muon decay.
6
Now muon decay does not involve any hadrons so there is no chance of any pion coming into
the picture and spoiling the party. The muon decays into an electron, a mu-neutrino and
an anti-neutrino of electron type. Thus its coupling constant is free of any renormalization
eects.
Something must be amiss as the beta decay coupling is not renormalized by this logic. To
overcome this problem, at least for the Fermi part of the interaction i.e. the vector part
of the interaction, Feynman and Gell-Mann proposed a conserved vector current theory
taking analogy from electrodynamics where again we expect the charge of proton to be
renormalized to something else because the strong interactions. However, it remains the
same as the positron charge. This means that all the other interactions are arranged in such
a manner that the total charge remains conserved and the physical coupling constant which
is the charge itself e remains the same as the bare charge value.
This analogy of electrodynamics is used to construct a vector current for beta decay such
that there is no need for renormalization.
We imagine that there is very massive vector particle mediating the weak interaction. A
massive particle is a good guess because the range of the interactions is very small. Higher
the mass of the particle being exchanged, smaller is the range of interaction. Let its mass
be M. When we compared beta decay and muon decay, the electron-electron anti-neutrino
was the common link. Dierence was in the proton-neutron and muon-mu neutrino parts.
Hence let this imaginary particle be coupled to these currents (proton-neutron and muon-mu
neutrino)

a
p
and

a
e
.
The total current is made up of proton-neutron, electron-neutrino, muon-neutrino currents
etc. and should be such that renormalization is not required. However, the electron-neutrino
and the muon-neutrino currents do not need any renormalization. Hence we must concen-
trate on the nucleon current
J

a
p
=
1
2

1
2

p
=
1
2
(J
V

+ J
A

) (20)
Proton and neutron form an isospin doublet for =
1
2
with
z
= 1/2 for protons and

z
= 1/2 for neutrons. Also, there exists the isospin raising and lowering operators

such that

n
=
x
+ i
y
=

2
p
(21)

p
= 0 (22)

p
=
x
i
y
=

2
n
(23)

n
= 0 (24)
In terms of these raising and lowering operators, we can write the vector part of the nucleon
current J
V

as as sum of isovector currents in the isospin notation.

J
V

N
(25)
7
where
N
is the nucleon spinor. It is easy to see that the current in eq. 25 is not conserved:

V
=

N
) (26)
= (

N
)

N
+

+
(

N
)
= = 0 (27)
The eq. 25 has an analogous term for electromagnetic currents in the isospin space
i

(1/2 +
z
)
N
(28)
The eq. 26 lends itself easily to separation into an isoscalar and an isovector part. The
isoscalar part which contains only

is a conserved current which implies that the number

of nucleons is conserved. However, the isovector part of the current is not conserved because
of the presence of the
z
term in it. This isovector current is conserved only when we add
the pion currents to it i.e.
J
el,conserved
V
=
N

N
+ i[ T
z

T
z

] (29)
The T
z
in the above equation is the third component of isospin for the pions. Note that
addition of vectors is done component wise hence if the isovector current of nucleons is
coming with the third component of isospin, the pion current will also come with the third
component of isospin. Hence eq. 27 is consistent.
As remarked earlier, this isovector electromagnetic current is analogous to the isovector
nucleon-nucleon current of weak interactions with
z
replaced by

. The eect of

is to
change neutron to proton.
The isovector conserved current of beta decay is therefore:
J
V

N
+ i[ T
+

T
+

] (30)
Again notice that we have used T
+
in the pion current (T
+
= T
x
+ iT
y
).
Equation 28 is the conserved vector current that we were talking of. We included only pions
in this conserved current because we assumed only pions aecting the beta decay interaction.
If we include the eects of other particles like the K particles, particles, particles then we
will have to include those currents as well. Since eq. 28 is a vector current and is conserved
as well, we replace the vector part of the beta decay current in eq. 20 by this current. The
physical meaning of this CVC is that the beta decay interaction strength or the coupling
constant magnitude is same as the muon decay coupling constant. In more general terms,
the baryon-lepton vertex has as much strength as the pion-lepton vertex.
After such a replacement, the Fermi part of the coupling will become universal i.e. applicable
to all weak interactions i.e. The Fermi coupling constant g

of muon decay which is exactly

of the V-A form will also be the exact coupling for at least the vector part of beta decay.
When the old beta decay theory was follows i.e. when the Fermi part was believed to
be scalar interaction, Gershtein and Zeldovich wrote It is of no practical signicance but
only of theoretical interest if the interaction is vector type; then g
V
(bare) = g
V
(eective).
No renormalization can be foreseen by analogy with Wards identity for the interaction of
a charged particle with the electromagnetic eld; in this case, virtual processes involving
particle do not lead to charge renormalization of the particle
8
The analogy of electromagentism and beta decay can be summarized as follows:
Table 1: Beta decay and electromagnetism analogy
Electromagnetism Beta decay vector part
Coupling constant e
1

2
g
V
Current J
el,conserved

J
V

Field potential A

a
e
Interaction Hamiltonian eJ
el,conserved

A
1

2
g
V
J
V

a
e
Since we have an extra term in the form of the pion current, more weak interactions can be
predicted. For example with the pion term, we can have an interaction by which a

can
go to
0
, and e. The interaction term for such a process will be
g
V
i[ T
+

T
+

](

a
e
) (31)
If we include more terms in the CVC, more such weak interactions willbe predicted. In
fact many such interactions have been experimentally observed, giving credence to the CVC
hypothesis.
5 After eects of the CVC hypothesis
The main consequences of the CVC hypothesis are :
Equality of the Fermi coupling constant of muon decay g

and the vector part (Fermi

part) g
V
of the beta decay.
The electromagnetic form-factors and weak vector form factors are equivalent
The absolute decay rate of the reaction

0
+ e +
e
can be determined
The expeiments that are done to verify these facts corroborate very strongly the CVC
hypothesis. In the following paragraphs, we consider these consequences one by one.
5.1 Equality of g
V
and g

Well, this fact was what propelled the very idea of CVC hypothesis. Hence prediction of equal
coupling constants just means that the hypothesis is on the right track. The experimental
accuracy achieved at the time that Feynman and Gell-Mann proposed the theory was within
2% of one another. The main expeiment for this was the usual beta decay of O
1
4 into N
1
4,
electron and the electron anti-neutrino and the muon decay into an electron, a mu neutrino
and an electron anti-neutrino.
9
5.2 Equivalence of the form factors
The amplitude of the electromagnetic interaction between two nucleons is
M = e u
2
_
F
1
(q
2
)

+
F
2
(q
2
)
2M
i

_
u
1
(32)
where F
1
(q
2
) is the electric form factor and F
2
(q
2
) is the magnetic form factor and is com-
posed of the magnetic moments of proton and neutron. Analogously, one can dene the
amplitude for the vector (Fermi) part of the weak interaction with the following replace-
ments: e

g
V

2 as follows:
M
beta
=
g
V

2
u
2
_
f
1
(q
2
)

+ f
2
(q
2
)i

u
1
(33)
Experimental evidence for this equivalence comes from the beta decay modes of B
12
and
N
12
.
5.3 Determination of the absolute decay rates of

decay
The CVC hypothesis as mentioned earlier predicts the weak decay of

to
0
, e and
e
.
Also we noted from the form of the CVC that the beta decay interaction strength for the
pion is same as that of nucleon beta decay. This means that this decay is of similar nature
to the O
14
beta decay.
The pion decay interaction is given in eq, 29. With such an interaction term and the coupling
constant being equal to g
V
= g

, the amplitude for the process is :

M = 4g
V
(p

+ p
0

)(

a
e
) (34)
with p

and p
0

being the four-momenta of the negatively charged pion and the neutral pion
respectively. In this reaction the energies involved are very low which makes the experiment
very dicult to perform. Since the vector part of the beta decay of which this pion current
is a part, does not require any renormalization it means that the interaction is free of any
loop structures that could have caused divergences in the theory. Nor is there any stron
coupling in the CVC. Thus whatever we need to calculate from the CVC for the pion decay
is direct.
Some considerations required in this process are that T
z
changes from -1 to 0. for the isospin
T = 1 for pions. This change is similar to beta decay where the third component of isospin

z
changes from
1
2
to
1
2
. In both cases, the change in isospin third component is 1. From
the formula for the half-life time period , we have:
1

=
mc
2

g
2
V
2
3
|M|
2
.f (35)
Since M is just the plain one vertex amplitude without any loops, propagators etc, |M|
2
= 2.
The value that one nds for this is 0.43 sec
1
10
6 What about the axial vector part
In the theory of weak interactions, the vector part did not require any renormalization
because of the conserved isovector current. However, there is no such thing for the axial part
of the current. It is not clear whether the axial vector current will require renormalization
or not. In the muon decay case, the coupling is exactly V-A which means that the coupling
constants for the vector part as well as the axial vector part are equal. However, when we
see the same thing in beta decay, the two coupling constants are not equal. The vector
coupling constant g
V
is equal to that of the muon decay but the axial vector part is not
equal to g
V
. The ratio of the axial vector coupling which is the same as Gamow-Teller
coupling to the vector coupling for beta decay comes out to be 1.3 0.1. The departure
from 1 which means that the axial vector part being bigger than the vector part may be
due to the renormalization eects of this part of the beta decay interaction. However, some
experiments do suggest within error ranges that the ratio of the two coupling constants is
1. The determination whether the axial vector part can be arranged in a manner that does
not require renormalization poses an interesting problem.
7 Conclusion
We derived the V-A form for the weak interaction current and from it by analogy to electro-
magnetism obtained a conserved isovector current through the CVC hypothesis. The CVC
theory has tasted many successes. In fact Feynman and Gell-Mann were so convinced of its
correctness that when the theory did not agree with the He
6
recoil experiment, they said
that the experiment needed to be redone as it was. It was found that the rst time, the
experiment was performed wrongly. When performed again, the results were congruent with
the CVC hypothesis.
References
1. Theory of Fermi Interaction, Feynman and Gell-Mann, Phys. Rev. 109, 1 (193,198)
(1958)
2. The Universal Fermi Interaction and the Conserved Vector Current in Bet Decay, C.S.
Wu, Rev. of Mod. Phys., April 1964.
11