Helicity Zero Particles

W. A. Perkins
Perkins Advanced Computing Systems,
12303 Hidden Meadows Circle, Auburn, CA 95603, USA
E-mail: [email protected]
In this paper we consider the possibility that a vector particle with mass
might exist in only one helicity state, rather than the usual three states with
helicity equal to +1, -1, and 0. Massless particles, of course, need only have
one helicity state. (For invariance under parity, they need two.) We show
arXiv:hep-ph/0409166v1 14 Sep 2004

that a massive vector particle can exist only in the helicity-0 state, if it is
composed of a fermion-antifermion pair and they are massless. This requires
the mass to be generated by the interaction between the massless particles.
An interaction of the form Ψ† iγ4 γµ Ψ is attractive between particle and an-
tiparticle and preserves helicity. Methods for experimentally distinguishing
an helicity-0 vector particle from both a spin-0 pseudoscalar particle and a
spin-1 vector particle are discussed.
PACS numbers: 12.60.Rc

I. INTRODUCTION

The purpose of this paper is to investigate the possibility of a massive vector particle
having less than all three helicity states. We are all familiar with massless particles that
only exist in two helicity states (photon) and one helicity state (two-component neutrino).
Here we will attempt to formulate a theory of a massive vector particle that only exist in the
helicity-0 state. In Sec. III we will discuss a method for distinguishing an helicity-0 vector
particle from a pseudoscalar particle or from a spin-1 vector particle.
Our definition of a vector particle is: A vector particle is a particle whose field transforms
as a four-vector under Lorentz transformations. Notice that there is no mention of spin in
the definition. As Veltman noted (see Ref. [1], p. 169), “We can freely use scalar fields, vector
fields, spinor fields, as long as the theory gives rise to results agreeing with the observed data.
One of the required properties is Lorentz invariance, and we must take care that Lorentz
invariance is maintained. Our classification in terms of the fields mentioned is really done
that way to keep this invariance transparent.”
To determine the spin corresponding to a given field requires some assumptions and a
theoretical calculation. The result, as given by Roman (see Ref. [2], p. 99), is, “in the case of
a scalar field (j = j′ = 0) we need no supplementary condition; the field has the unique spin
1 1
value zero. The same holds for the elementary spinor representations D 2 0 and D 0 2 (and also
their direct sum), which give the unique spin value 12 . However, for a four-vector field, which
11
belongs to the D 2 2 representation, 4jj′ = 1; thus one supplementary condition is required
in order to have the unique spin value 1 and to exclude the value zero.” This supplementary
condition is the well-known Lorentz condition and applies to the photon field. However,

1
there are four-vector fields for which the Lorentz condition does not apply as illustrated
near the end of Sec. II.
The two obvious kinematic properties of elementary particles are spin and mass, and they
should be describable by quantities that are invariant under relativistic transformations. As
Wigner first showed [3] the relevant group in considering the spin of a particle is the Poincare
group as spin is an invariant under that group.
Since the Poincare group is of rank 2, there are just two Casimir invariants [4],

C1 = Pµ Pµ ,
C2 = Wµ Wµ , (1)

where Wµ is the Pauli-Lubanski vector and is defined by,

1
Wµ = ǫµνρσ Jνρ Pσ , (2)
2
and Pµ is the momentum operator and the Jνρ ’s are the generators for Lorentz transforma-
tions. The representations of the Poincare group depend upon whether M > 0 or M = 0,
where M the mass of the particle.
For M > 0, the momentum is time-like, and the little group (i.e. the subgroup of the
Poincare group that leaves pµ invariant) is the group of rotations. We will represent a state
by |pλ >, where p is the momentum and λ is the helicity. For a rest state, p = p0 ,

Pµ |p0 λ > = pµr |p0 λ >,

J 2 |p0 λ > = s(s + 1)|p0 λ >,
J3 |p0 λ > = λ|p0 λ >, (3)

where pµr = (0, 0, 0, iM) and s is the intrinsic spin.

The helicity is transformed among the 2s + 1 possible states for this massive case. The
eigenvalue of the first Casimir operator, c1 = −M 2 , and the eigenvalue of the second Casimir
operator, c2 = −M 2 s(s + 1).
For M = 0, the momentum is light-like, and spin is no longer described by the rotational
group. The little group in this case is a combinations of boosts and rotations that surprisingly
depend upon the boost parameters. The state of a massless particle is characterized by the
helicity, λ. Unlike the massive case, the helicity is invariant under Lorentz transformations.
For a particle moving along the third axis,

Pµ |pℓ λ > = pℓµ |pℓ λ >,

J3 |pℓ λ > = λ|pℓ λ >,
Wi |pℓ λ > = 0, (4)

where pℓµ = (0, 0, p, ip) and i = 1, 2. Here the eigenvalues of the Casimir operators, c1 = 0,
and c2 = 0.
We are interested is seeing if a massive particle can exist in only one helicity state as
massless particles can. However, a particle with mass can be transformed to a coordinate

2
system in which it is at rest. In that system there is no helicity axis unless the particle has
some internal structure.
Suppose this special massive particle is composed of two spin- 12 particles with the same
mass, but oppositely directed helicities. We can visualize the different types of particles as
shown in Fig. 1. In the coordinate system for which the massive particle is at rest, the two
fermions would have momentum, p and −p. Thus, the opposing helicities form an axis, even
though the massive composite particle is as rest. If the massive particle is given a boost
along any axis other than the one formed by the opposing helicities, the internal axis will
not coincide with the momentum of the massive particle.

spin-0
scalar

spin-1
vector

helicity-0
vector

FIG. 1. Illustration of spin-0 scalar particle which is spherically symmetric, spin-1 vector particle
formed of two spin-1/2 particles, and helicity-0 vector particle formed of two spin-1/2 particles with
its two states. The arrows indicate the fermion spin directions.

3
 Let us assume this special massive particle is in the helicity-0 state given by (↑↓ + ↓↑),
where we have used up and down arrows to denote the z-component of spin for the spin- 12
particles. Normally, we would assume that this state could be transformed into the other
helicity ±1 states. However, if the two fermions (that compose the massive particle) are
massless, this helicity state will be just as invariant as the helicity states of the photon
and neutrino. Mechanisms for generating mass have been suggested [5,6], and glueballs [7],
in quantum chromodynamics are an example of massive particles created by combining
massless particles.
The usual argument that a massive particle must have three polarization states does not
consider the possibilities allowed by composite particles. For example, with a vector particle
of mass M at rest, the polarization states are (see Ref. [1], p. 176) described by the three
four-vectors, ejµ ,

1
e1µ = √ (1, 0, 0, 0),
2
1
e2µ = √ (0, 1, 0, 0),
2
1
e3µ = √ (0, 0, 1, 0). (5)
2
By rotating about a suitable axis, any of these can be transformed into the others. For
example, if we rotate about the first axis,

1 0 0 0 0 0
    

1  0 cos θ sin θ 0   0 
   1  sin θ 

√    = √ 

(6)
2 0 − sin θ cos θ 0 1 2 cos θ 
 
    
0 0 0 1 0 0

and for θ = 90 degrees, e3µ is transformed into e2µ .

However, there are good reasons to looks at the polarization states of composite particles.
The author has suggested [8] that all integral spin particles are composite particles formed of
fermions. Also Varlamov [9] has shown recently that the problems encountered in quantizing
the electromagnetic field disappear if the photon is composed of fermion pairs. Thus, it seems
very reasonable to consider four-vector fields constructed from fermions.

II. POLARIZATION STATES OF COMPOSITE PARTICLES

Now let us look at the helicity states of a composite particle formed of a fermion-
antifermion pair, each particle with mass m. Using the function Λ(helicity) and up and
down arrows to denote the z-component of spin for the spins 12 particles, we have the usual
four combinations with well-defined exchange symmetry,

4
 Λ(1) =↑↑
Λ(−1) =↓↓
1
Λ(0) = √ (↑↓ + ↓↑)
2
1
Λ(0) = √ (↑↓ − ↓↑) (7)
2
To determine the polarization vectors for the different combinations of (7), we consider
the vector field created from the fermion-antifermion pair,

Ψ† iγ4 γµ Ψ. (8)

This assumes that the pair are bound by a local interaction that does not involve some
quanta [10]. Unlike the pseudoscalar interaction, for example, this vector interaction has
the required property of being attractive between particle and antiparticle and repulsive
between particles [5,10]. The Feynman diagram for this interaction is chain-like as shown
in Fig. 2. Nambu and Jona-Lasinio [5] have used the Bethe-Salpeter equation in the chain
approximation to form a pseudoscalar particle as a bound fermion-antifermion pair, and
Ohanian [11] has used the Bethe-Salpeter equation in the chain approximation to form
gravitons from bound fermion-antifermion pairs.

FIG. 2. Bound states as a chain of constituent fermion-antifermion bubbles.

Following Kane [6] we can derive an important relation about the helicity of Ψ† iγ4 γµ Ψ.
First we separate the upper and lower parts of the wave function,
!
ΨR
Ψ= , (9)
ΨL

where L represents a left-handed positive energy solution and R represents a right-handed

one.
We will use Weyl realization of the γ matricies with,

0 →
! ! !
→ σ 0 1 −1 0
γ =i → , γ4 = , γ5 = . (10)
−σ 0 1 0 0 1

where the σ’s are the Pauli spin matricies.

5
 We can define projection operators,
! !
1 + γ5 0 0 1 − γ5 1 0
PL = = , PR = = , (11)
2 0 1 2 0 0
whose effect on Ψ is,
ΨL = PL Ψ,
ΨR = PR Ψ. (12)
We can write,
Ψ† γ4 γµ Ψ = Ψ† γ4 (PL + PR )γµ (PL + PR )Ψ
= Ψ† γ4 PL γµ PL Ψ + Ψ† γ4 PR γµ PL Ψ + Ψ† γ4 PL γµ PR Ψ + Ψ† γ4 PR γµ PR Ψ. (13)
Since
PL γ µ = γ µ PR ,
PR γ µ = γ µ PL , (14)
the first and fourth terms of Eq. (13) are zero and we also have,
Ψ†L γ4 = Ψ† γ4 PR ,
Ψ†R γ4 = Ψ† γ4 PL . (15)
Substituting Eq. (15) into Eq. (13), we obtain,
Ψ† γ4 γµ Ψ = Ψ†L γ4 γµ ΨL + Ψ†R γ4 γµ ΨR , (16)
which shows that the helicity is preserved for an interaction of the form Ψ† iγ4 γµ Ψ.
For determining the polarization vectors, we will separate out the space-time dependence
by putting Ψ = uei(p·x−Et) and use plane wave spinors that are solutions of the Dirac
equation,
 
s 1
p1 +ip2
E + p3
 
u+1 E+p3
 
+1 (p) =  ,
m
2E 
 E+p3


0
−p1 +ip2
 
s E+p3
E + p3 1
 
u−1
 
−1 (p) =  ,
2E 
 0 

−m
E+p3
−m
 
s E+p3
E + p3 0
 
u−1
 
+1 (p) =  ,
2E 
 1 

p1 +ip2
E+p3
 
s 0
m
E + p3
 
u+1 E+p3
 
−1 (p) = , (17)
−p1 +ip2

2E 
 E+p3


1

6
where pµ = (p, iE), and the superscripts and subscripts on u refer to the energy and helicity
states respectively.
Taking the composite particle to be at rest, one fermion will have momentum p and the
other momentum −p. The spinors for negative momenta are related to those of positive
momenta by,
u+1 −1
+1 (−p) = u−1 (p),
+1
u−1
−1 (−p) = u+1 (p),
u+1 −1
−1 (−p) = u+1 (p),
+1
u−1
+1 (−p) = u−1 (p). (18)
The polarization vectors corresponding to the different combinations of (7) are,
1
ǫ1µ (p) = √ [u−1 † +1
−1 (p)] iγ4 γµ u+1 (p),
2
1
ǫ2µ (p) = √ [u+1 † −1
+1 (p)] iγ4 γµ u−1 (p),
2
1  
ǫ3µ (p) = √ [u+1
+1 (p)]†
iγ 4 γ µ u +1
+1 (p) + [u +1
−1 (p)]†
iγ 4 γ µ u +1
−1 (p) ,
2 2
1  
ǫ4µ (p) = √ [u+1 † +1 +1 †
+1 (p)] iγ4 γµ u+1 (p) − [u−1 (p)] iγ4 γµ u−1 (p) .
+1
(19)
2 2
Carrying out the matrix multiplications results in,
1 −ip1 p2 +E 2 +p3 E −p21 −p1 p2 +iE 2 +ip3 E −ip22 −p1 −ip2
!
ǫ1µ (p) = √ , , ,0 ,
2 E(E + p3 ) E(E + p3 ) E
2 2 2 2
!
1 ip 1 p 2 +E +p 3 E −p 1 −p 1 p 2 −iE −ip 3 E +ip 2 −p 1 +ip 2
ǫ2µ (p) = √ , , ,0 ,
2 E(E + p3 ) E(E + p3 ) E
1 p1 p2 p3
 
3
ǫµ (p) = √ , , ,i ,
2 E E E
ǫ4µ (p) = (0, 0, 0, 0) . (20)
As expected ǫ4µ (p) vanishes, showing that there is no vector state for the last combination
of Eq. (7).
There are several problems with these polarization vectors. First of all, there is the
normalization problem,
m2 (m2 +p23 −E 2 )
ǫ1µ (p) · ǫ1∗
µ (p) = 1 + ,
2E 2 (E + p3 )2
m2 (m2 +p23 −E 2 )
ǫ2µ (p) · ǫ2∗
µ (p) = 1 + ,
2E 2 (E + p3 )2
3 3∗ m2
ǫµ (p) · ǫµ (p) = 1 − , (21)
(2E 2 )
instead of ǫjµ (p) · ǫj∗
µ (p) = 1. Another problem is that the different polarization vectors are
not orthogonal,

7
 m2 (−m2 −2ip1 p2 − 2p21 − 2p23 +E 2)
ǫ1µ (p) · ǫ2∗
µ (p) = ,
2E 2 (E + p3 )2
m2 (p1 +ip2 )
ǫ1µ (p) · ǫ3∗
µ (p) = ,
2E 2 (E + p3 )
m2 (p1 −ip2 )
ǫ2µ (p) · ǫ3∗
µ (p) = , (22)
2E 2 (E + p3 )
instead of ǫjµ (p) · ǫk∗
µ (p) = 0 for j 6= k. Furthermore, the dot product of each polarization
vector with the internal four-momentum pµ is not equal to zero.
All of these problems can be overcome immediately if we assume that the massive com-
posite particle is formed of two massless fermions (i.e., set m = 0). The composite particle
could obtain mass as in the Nambu-Jona-Lasino dynamic model [5] or by the Higgs mech-
anism [6]. If the fermions are massless, the polarization vectors depend only upon the
direction of p (the internal momentum), n = p/|p|. Equation (20) now becomes,
1 −in1 n2 +1+n3 −n21 −n1 n2 +in21 +in23 +in3
!
ǫ1µ (n) =√ , ,−n1 −in2 , 0 ,
2 1 + n3 1 + n3
1 in1 n2 +1+n3 −n21 −n1 n2 −in21 −in23 −in3
!
2
ǫµ (n) = √ , ,−n1 +in2 , 0 ,
2 1 + n3 1 + n3
1
ǫ3µ (n) = √ (n1 , n2 , n3 , i) . (23)
2
These polarization vectors satisfy the normalization relation,
ǫjµ (n) · ǫj∗
µ (n) = 1,

ǫjµ (n) · ǫk∗

µ (n) = 0 for k 6= j. (24)

The Lorentz-invariant dot products of the internal four-momentum pµ = |p|(n1 , n2 , n3 , i)

with the polarization vectors are,
pµ ǫ1µ (n) = 0,
pµ ǫ2µ (n) = 0,
pµ ǫ3µ (n) = 0. (25)
The helicity operator acting upon the polarization vectors gives,
(J · n)ǫ1µ (n) = ǫ1µ (n),
(J · n)ǫ2µ (n) = −ǫ2µ (n),
(J · n)ǫ3µ (n) = 0, (26)
where the components of J are,
0 0 0 0 0 0 i 0 0 −i 0 0
     
0 0 −i 0  0 0 0 0 i 0 0 0
J1 =  , J2 =  , J3 =  , (27)
     
0 i 0 0  −i 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0

8
 Under a rotation about n by an angle θ, they change as follows,

ǫ1µ (n) → eiθ ǫ1µ (n),

ǫ2µ (n) → e−iθ ǫ2µ (n),
ǫ3µ (n) → ǫ3µ (n). (28)

Thus, ǫ1µ (n), ǫ2µ (n), and ǫ3µ (n) correspond to states with helicity +1, −1, and 0 respectively.
If the momentum is along the third axis, the polarization vectors reduce to,
1
ǫ1µ (n) = √ (1, i, 0, 0),
2
1
ǫ2µ (n) = √ (1, −i, 0, 0),
2
1
ǫ3µ (n) = √ (0, 0, 1, i). (29)
2

Thus, ǫ1µ (n) and ǫ2µ (n) reduce to the usual polarization vectors for right and left circular-
polarized photons respectively while ǫ3µ (n) corresponds to longitudinal polarization. The
polarization vectors cannot be transformed into each other under a Lorentz transformation
because they are orthogonal (see Eq. 24).
It has been noted that one could consider each transverse polarization as an independent
particle (see Ref. [1], p. 173, 180-2). This is because these polarization vectors are inde-
pendent degrees of freedom and under a Lorentz transformation change into themselves.
(For invariance under parity, one needs both transverse polarizations.) The longitudinal
polarization causes problems if one considers it to be part of the photon field [12]. We will
take this longitudinal polarization as belonging to an independent particle with helicity-0.
One does not usually think of the longitudinal polarization of a massive particle as being
invariant, but as shown in Sec. II the helicity is preserved for an interaction of the form
Ψ† iγ4 γµ Ψ.
As discussed in Sec. I, if one can impose the Lorentz condition on a four-vector field,
this reduces the spin possibilities to spin-1 and eliminates spin-0. We will now apply the
Lorentz condition to the above four-vector fields. The photon field can be constructed [13]
from ǫ1µ (n) and ǫ2µ (n),

1 nh i
γR (P)ǫ1µ (n) + γL (P)ǫ2µ (n) eiP R
X
Aµ (R) = q
P 2V ωp
h i o
+ γR† (P)ǫ1∗ † 2∗
µ (n) + γL (P)ǫµ (n) e
−iP R
, (30)

This four-vector field satisfies the Lorentz condition,

(∂Aµ /∂xµ ) = 0 (31)

which follows from (25).

9
 In a similar manner the field Vµ (R) for the particle with longitudinal polarization is,

1 n
ǫ3µ (n) [χ(P) − η(P)] eiP R
X
Vµ (R) = q
P 2V ωp
h i o
+ χ† (P) − η † (P) e−iP R , (32)

with χ(P)√ and η(P) being annihilation operators for the composite particle. Here P R =
P · R − P2 + M 2 t for a particle with mass M and momentum P. The vector n refers to
the internal momentum direction and not P. Since in the rest frame Pµ = (0, 0, 0, iM), we
now have,

Pµ ǫ3µ (n) = −M. (33)

The field equations for this particle are,

(2 − M 2 )Vµ = 0,
(∂Vµ /∂xµ ) = −MfV . (34)

where
1 n
[χ(P) − η(P)] eiP R
X
fV (R) = i q
P 2V ωp
h i o
− χ† (P) − η † (P) e−iP R . (35)

From the last equation of (34) we see that this four-vector field does not satisfy the Lorentz
condition. Note that the composite-particle four-vector field of Eq. (32) is not the four-
gradient of a scalar field, such as that discussed in [14], since the polarization vectors are
functions of n, not P.

III. CONCLUSION

The conditions under which a massive vector particle could exist in a single helicity state
have been examined. The requirements are that the particle must be composed of a massless
fermion-antifermion pair, which are bound by an interaction of the form Ψ† iγ4 γµ Ψ.
If a massive helicity-0 vector particle exists, it will have characteristics similar to those of
a pseudoscalar. However, since it has an axis with a direction, a helicity-0 particle could be
identified experimentally by detecting an asymmetry in its decay products. The longitudinal
polarization can lead to forward-backward asymmetries. Unlike a charged particle with spin
1/2 or spin 1, a charged helicity-0 particle will have zero magnetic moment and its direction
of polarization will not be altered by a magnetic field.
It has been noted [15] that nature uses simple but odd constructions. The longitudinally
polarized vector field, constructed in Sec. II is about the simplest vector field imaginable.
The Proca model [16] of a spin-1 vector field for a massive particle with transverse and
longitudinal polarizations is obviously much more complex.

10
 ACKNOWLEDGMENTS

Many helpful discussions with Prof. J. E. Kiskis are gratefully acknowledged.

[1] M. Veltman, Diagrammatica, The Path to Feynman Rules (University Press, Cambridge)
(1994).
[2] P. Roman, Theory of Elementary Particles (North-Holland, Amsterdam) 1960.
[3] E. Wigner, Ann. Math. 40, 149 (1939).
[4] W. Tung, Group Theory in Physics (World Scientific, Philadelphia) 1985.
[5] Y. Nambu and G. Jano-Lasinio, Phys. Rev. 122, 345 (1961); 124, 246 (1961).
[6] G. Kane, Modern Elementary Particle Physics (Addison-Wesley, Redwood City) (1987).
[7] C. J. Morningstar and M. Peardon, Phys. Rev. D 60, 034509 (1999).
[8] W. A. Perkins, Int. J. Theor. Phys. 41, 823 (2002).
[9] V. V. Varlamov, “About Algebraic Foundation of Majorana-Oppenheimer Quantum Electro-
dynamics and de Brogie-Jordan Neutrino Theory of Light,” math-ph/0109024 (Sept. 2001).
[10] E. Fermi and C. N. Yang, Phys. Rev. 76, 1739 (1949).
[11] H. C. Ohanian, Phys. Rev. 184, 1305 (1969).
[12] M. W. Evans, Physica A 214, 605 (1995).
[13] W. A. Perkins, “Interpreted History of Neutrino Theory of Light and Its Future,” in Lorentz
Group, CPT and Neutrinos, Eds. A. E. Chubykalo, V. V. Dvoeglazov, D. J. Ernst, V. G. Kady-
shevsky, and Y. S. Kim (World Scientific, Singapore) (2000) 115-126.
[14] S. Weinberg, The Quantum Theory of Fields, Volume I, Foundations (University Press, Cam-
bridge) Sec. 5.3 (1995).
[15] T. D. Lee and C. N. Yang, Phys. Rev. 105, 1671 (1957).
[16] L. H. Ryder, Quantum Field Theory (University Press, Cambridge) (1996).

11