December 1, 2009

arXiv:0912.0364v1 [hep-ph] 2 Dec 2009

A Soluble Model of ”Higgs boson” as a

Composite∗†
R. Friedberg1 and T. D. Lee1,2
1. Physics Department, Columbia University, New York, New York 10027, USA

2. China Center of Advanced Science and Technology

(CCAST/World Lab.), P. O. Box 8730, Beijing 100190, China

Abstract
Higgs boson may turn out to be a composite. The theoretical description of such a composite is
illustrated by an example of a soluble model.

∗ to be included in ”Selected Papers and a Glimpse into the life of Gunnar Källen”, Edited
by C. Jarlskog and A. C. T. Wu.
† This research was supported in part by the U.S. Department of Energy (grant no. DE-
FG02-92-ER40699).
 1. Introduction

The σ-model [1] is a successful phenomenological theory of low energy particle

physics. Yet, the 0+ σ-particle itself has never been identified experimentally
[2]. One of the possible reasons for this failure might be that the spin-parity
transformation of σ is the same as that of

σ2, σ3, · · · , σn, · · · . (1.1)

Thus, what is a σ-field in the idealized theoretical model may appear experimen-
tally as a ”composite” due to the possible mixture of (1.1). Today, a major focus
of high energy physics is to search for Higgs boson. It might be that for similar
and other reasons, Higgs boson [3] could also turn out to be a ”composite”
[4-8]. Neither the experimental identification of σ nor that of Higgs boson would
correspond to the usual simple theoretical description of a single pole in the com-
plex energy plane. In this paper, we explore the theoretical description of such a
composite from a more elementary perspective, by examining the generalization
of a soluble model [9]. The structure of the model is given in Section 2, and its
solution in Sections 3 and 4.
This paper is dedicated to the memory of Gunard Källen, who (besides others)
has made important contributions [10-15] to the original soluble model.

2. A Generalized Soluble Model

We generalize the original V ⇄ N θ model by retaining the same fixed

Fermion states V and N , but replacing the single θ(r) field by three boson
fields A(r), B(r) and C(r). The Hamiltonian H in the new model is given by

H = H0 + H1 + H2 (2.1)
where
X
†
H0 = m0 V V + (λk a†k ak + µk b†k bk + νk c†k ck ). (2.2)
k
For convenience, the entire system is enclosed within a sphere of large radius
R. The s-wave part of the annihilation field operators A(r), B(r) and C(r)
are given in terms of their annihilation operators ak , bk and ck by

P − 21
A(r) = k (4πRλk ) uk r−1 (sin kr)ak ,
P (2.3)
− 21 −1
B(r) = k (4πRµk ) vk r (sin kr)bk

and
P −2 1
−1
C(r) = k (4πRνk ) wk r (sin kr)ck

with
1
λk = (k 2 + α2 ) 2 ,
1
µk = (k 2 + β 2) 2 , (2.4)
1
νk = (k 2 + γ 2) 2

and α, β, γ the masses of bosons a, b and c. The functions uk , vk , wk

are convergence factors, which may all be chosen to be 1 for k < kmax and 0
otherwise. In (2.3) all summations extend over

k = nπ/R (2.5)

with n = 1, 2, 3, · · · . At equal time, we have the anti-commutation relations

{V, V † } = {N, N † } = 1 (2.6)

and the commutation relations

[ak , a†k′ ] = [bk , b†k′ ] = [ck , c†k′ ] = δkk′ . (2.7)

If one wishes, (2.6) can also be changed into commutation relations, and V and
N would then be bosons.
In (2.2), we set the mass of N to be zero, and the ”bare” mass of V to be
m0 . The interaction Hamiltonians H1 and H2 are given by

H1 = g(V † N C(0) + N † V C †(0)) (2.8)

and
H2 = f (V † N B † (0)A(0) + N † V A† (0)B(0)) . (2.9)

The g-coupling governs the transition

V ⇄ Nc . (2.10)

and the f -coupling gives rise to the scattering

Na ⇄ V b . (2.11)

Thus, when f = 0 the Hamiltonian is identical to that of the original V ⇄ N θ

model, with θ replaced by c.
Throughout the paper, we assume V to be unstable through V → N c when
R → ∞. Since the mass of N is set to be zero in the model, V is unstable if its
physical mass m is larger than γ, the mass of c; i.e.,

m>γ . (2.12)

Thus, in a collision of N a, beside the elastic scattering

Na → Na , (2.13)

we also have the inelastic process

N a → V b → N bc (2.14)
provided that the total energy E satisfies

E >β+γ , (2.15)

the threshold energy of the channel N bc.

We shall assume
α<β+γ (2.16)

Hence, in the N a channel at low energy when

α<E <β+γ , (2.17)

there is only the elastic scattering (2.13); at higher energy when E > β + γ, we
have both (2.13) and the inelastic process (2.14).
Consider first the process

Nc ⇄ V ⇄ Nc . (2.18)

Denote the corresponding state vector by

" #
X − 21
− 21
|N c i ∝ V † + g(4πR) νk kwk (E − νk )−1N † c†k |vac i . (2.19)
k

One can readily verify that it satisfies

H|N c i = E|N c i . (2.20)

At a finite R, E satisfies the eigenvalue equation

X k2 w2  1 
k
hR (E) ≡ E − m0 − g 2 =0 , (2.21)
4πνk R E − νk
k

with its derivative

X k 2 w 2  1 2
h′R (E) = 1 + g 2 k
(2.22)
4πνk R E − νk
k
always positive.
When R → ∞, hR (E) becomes
Z ∞
2 k 2 wk2 dk
h∞ (E) = E − m0 − g . (2.23)
0 4π 2 νk (E − νk )

The condition for V being unstable is that when E = γ,

h∞ (γ) < 0 . (2.24)

In this case, we introduce a cut along the real axis from

E=γ to ∞ (2.25)

where γ is the mass of the c-meson. The derivative of h∞ (E) is

Z ∞ 2 2
′ 2 k wk dk
h∞ (E) = 1 + g . (2.26)
0 4π 2 νk (E − νk )2

which is positive ≥ 1, when E is real < γ. For E = νk > γ but just above the
cut along the real axis, we have from (2.23)
 2
g
Imh∞ (νk + io+) = i kwk2 . (2.27)
4π

Thus, on the second sheet near and below the cut (2.25), there is a zero of
h∞ (E), which corresponds to the V -resonance. It can be shown that the phase
shift δ for N c scattering (2.18) is related to h∞ (νk + io+) and its complex
conjugate by
h∞ (νk + io+)
e−2iδ = (2.28)
[h∞ (νk + io+)]∗

3. Na Sector (General Discussion)

As discussed in the Introduction, assume the idealized case that there does
exist a fundamental spin 0 field φ which is the origin of masses of spin nonzero
particles. In any physical process, there are bound to be effective couplings
between φ and some of its higher power products, such as

φ2 , φ3, · · · , φn , · · · .

Thus, the physical Higgs channel becomes connected to not only a complex
pole, but also to a cut in the complex energy plane, or other more complicated
analytical structure.
In this section, the N a channel that we shall analyze represents a highly
idealized model of ”Higgs” as a composite. From reactions (2.10) and (2.11),
we see that a state vector | i in the N a sector must also have components in
V b and N bc channels as well. Thus for R finite, we may write
" #
X X X
|i= ψ(k)a†k N † + φ(p)b†pV † + χ(p, q)b†pc†q N † |0 i . (3.1)
k p p,q

From
H| i = E| i , (3.2)

we find
X
(E − λk )ψ(k) = f Uk Vp φ(p) , (3.3)
p
X X
(E − m0 − µp )φ(p) = g Wq χ(p, q) + f Vp Uk ψ(k) (3.4)
q k
and
(E − µp − νq )χ(p, q) = gWq φ(q) (3.5)

where k, p, q are all given by (2.5) and Uk , Vp , Wq are related to the uk , vp and
wq of (2.3) by
1
Uk = (4πRλk )− 2 kuk
(3.6)
− 21
Vp = (4πRµp ) pvp
and
1
Wq = (4πRνq )− 2 qwq
 Substituting (3.5) into (3.4), we have
X X
(E − m0 − µp )φ(p) = g 2 Wq2(E − µp − νq )−1φ(p) + f Vp Uk ψ(k) (3.7)
q k

and therefore
X
φ(p) = [Dp (E)]−1f Vp Uk ψ(k) (3.8)
k
where
X
Dp (E) = E − m0 − µp − g 2
Wq2(E − µp − νq )−1 . (3.9)
q

From (3.8), we also have

" #
X X Vp2 X
Vp φ(p) = f Uk ψ(k) .
p p
Dp (E)
k

Thus, (3.3) becomes

" #
X Vp2 X
(E − λk )ψ(k) = f 2Uk Uk′ ψ(k ′ ) (3.10)
p
Dp (E) ′ k

Multiplying both sides by Uk /(E − λk ) and summing over k, we find that the
eigenvalue E satisfies
X Vp2
2
1 = f F (E) (3.11)
p
Dp (E)
in which
X
F (E) = Uk2 (E − λk )−1 . (3.12)
k
Next, we study the continuum limit. When R → ∞, the sum
X X 1
Uk Vp = (4πR)−1 (λk µp )− 2 kuk pvp (3.13)
p p

becomes
Z
1
(4π 2 )−1 (λk µp )− 2 kuk pvpdp . (3.14)

Thus, from (3.3) we have

Z
1
2 −1
(E − λk )ψ(k) = (4π ) f kuk (λk µp )− 2 pvp φ(p)dp . (3.15)
Likewise, (3.7) leads to
Z
(E − m0 − µp )φ(p) = (4π 2)−1g 2 φ(p) νq−1(E − µp − νq )−1q 2wq2 dq
Z (3.16)
− 21
+ (4π 2)−1f pvp (µp λk ) kuk ψ(k)dk ,

which gives
Z
1
φ(p) = (4π 2 Dp(E))−1f pvp (µp λk )− 2 kuk ψ(k)dk (3.17)

where
Z
2 −1 2
Dp(E) = E − m0 − µp − (4π ) g [νq (E − µp − νq )]−1q 2 wq2 dq (3.18)

In a collision of N a, in order to describe reactions (2.13) and (2.14), we write

ψ(k) and φ(p) as
e
ψ(k) = δ(k − k0) + ψ(k) (3.19)

and
e
φ(p) = φ(p) (3.20)

e
in which ψ(k) e
and φ(p) denote the scattered wave amplitudes. Thus, (3.15)
remains valid if we replace ψ, φ simply by ψe and φ.
e Hence
Z
1
e
(E − λk )ψ(k) 2 −1 e
= (4π ) f kuk (λk µp )− 2 pvp φ(p)dp . (3.21)

On the other hand, (3.17) yields

 Z 
1 1
φ(p) e
e = (4π 2 Dp(E))−1f pvp (µpλ0 )− 2 k0u0 + (µp λk )− 2 kuk ψ(k)dk (3.22)

with
λ0 = λk and u0 = uk at k = k0 . (3.23)

Define
Z
1
A = (4π ) 2 −1 e
(vpp/µp )φ(p)dp
2
, (3.24)
 Z
1
B = (4π 2 )−1 e
(uk k/λk2 )ψ(k)dk (3.25)

and
1
C = (4π 2)−1(u0k0/λ02 ) . (3.26)

Hence, (3.21) and (3.22) can be written as

1
e
ψ(k) = f A uk k/λk2 (E − λk ) (3.27)

and
1
e = f (B + C )vpp/µp2 Dp(E)
φ(p) (3.28)

with Dp (E) given by (3.18). In above expressions, all integrations over p and k
extend from 0 to ∞.
Substituting (3.28) into (3.24), we find

I ≡ A /(B + C ) (3.29)

is given by
Z ∞
f (vp2p2/µp )dp
I= 2 Z ∞ . (3.30)
4π 0 g2 q 2 wq2dq
E − m0 − µp − 2
4π 0 νq (E − µp − νq )
Likewise,
J ≡ B/A (3.31)

becomes
Z ∞
f (u2k k 2/λk )dk
J= 2 . (3.32)
4π 0 E − λk
Thus
I
A = C (3.33)
1 − IJ
and
1
B+C = C (3.34)
1 − IJ
From (3.26), (3.30) and (3.32), we have the explicit expressions for C , I and
J. Hence A and B are also known. Equation (3.27) and (3.28) then give
e
scattering amplitudes ψ(k) e
and φ(p).

4. Na Sector (Critical f 2 )

We shall show that when R → ∞ and f 2 greater than a critical strength fc2,
there exists a bound state in the N a sector. Write the R → ∞ limit of (3.11)
as
1 = f 2 F (E)G (E) (4.1)

in which
F (E) = lim F (E)
R→∞
Z ∞ (4.2)
= (4π 2 )−1 [k 2u2k /λk (E − λk )]dk
0
with uk , λk given by (2.3) and (2.4). The function G (E) is similarly related to
the last summation in (3.11) by
X Vp2
G (E) = lim
R→∞ Dp (E)
p
Z ∞ (4.3)
= (4π 2)−1 [p2vp2 /µp Dp(E)]dp
0
where Dp (E) is given by (3.18), Since Dp(E) is related to h∞ (E) of (2.23) by

Dp(E) = h∞ (E − µp ) , (4.4)

we have from (2.24)

Dp (µp + γ) = h∞ (γ) < 0 . (4.5)

From (2.4),
νq = (q 2 + γ 2) > γ . (4.6)

Thus, (3.18) and (4.4) - (4.5) imply that Dp(E) and its derivative
∂
Dp′ (E) = Dp (E) (4.7)
∂E
are continuous and satisfy

Dp(E) < 0 and Dp′ (E) > 0 (4.8)

over the range

E < µp + γ , (4.9)

which includes the range

E<α (4.10)

in accordance with (2.4) and (2.16). Thus, both F (E) and G (E) are negative,
with negative derivatives; their product is positive and varies from 0 to a finite
value as E increases from −∞ to α, the mass of a-meson.
Define a critical f 2-coupling by

fc2 = [F (α)G (α)]−1 (4.11)

It then follows that there exists a bound state energy E0 in the N a sector with

1 = f 2F (E0)G (E0) (4.12)

provided
f 2 > fc2 (4.13)

For f 2 < fc2, the state turns into a resonance with a complex E0 . In this case
e
the scattering amplitudes ψ(k) e
and φ(k) have besides the cuts given by (3.30)
and (3.32), also a complex pole at E = E0.
It is of interest to note the difference between this pole in the N a sector and
the V -pole in the N c sector. The V -pole becomes stable in the weak coupling
limit when g 2 → 0, whereas in the N a sector the boundstate E0 becomes stable
only in the strong coupling region when f 2 > fc2.
We note that when g 2 = 0, G (E) of (4.3) becomes
Z ∞
2 −1
 2 2 
G0(E) = (4π ) p vp /µp (E − m0 − µp ) dp . (4.14)
0
Correspondingly, (4.12) becomes

1 = f 2F (E0 )G0(E0) (4.15)

with the same F (E) of (4.2). Thus, the existence of the pole at E = E0 does
not depend sensitively on g 2 ; instead, it is closely related to the second (and
higher) order attractive potential between N a due to the f -coupling transitions

Na ⇄ V b ⇄ Na . (4.16)

Its physical origin is quite different from the V -pole in the N c channel of (2.18).
 5. Remarks

Consider the case

f 2 < fc2 . (5.1)

The composite V b is unstable, and may serve as a highly simplified model of

either the σ-meson or the Higgs boson. Besides the elastic process (4.16), there
is also the inelastic reaction

N a ⇄ V b ⇄ N bc . (5.2)

In order to detect the composite V b as a resonance, we require in (4.12) the

corresponding pole at
E = E0 (5.3)

to be not too far from the real axis; hence f 2 cannot be too small. The amplitude
for the continuum background must then also be relatively large.
In any composite model, we may regard the amplitudes N a and V b as the
idealized representations of its low and high frequency components of the same
composite state vector. A second order transition between these two components
would always depress N a and elevate V b, as in (4.16). A resonance thus formed
would require a strong coupling, and therefore also a large continuum background
as in the model. This could be the reason why the σ-meson does not appear
as a sharp resonance, and it might also be difficult to isolate the Higgs boson
resonance.
We wish to thank N. Christ and E. Ponton for discussions.
 Appendix

In the special case when

wk2 = k/νk3 , (A.1)

the integral
Z ∞
k 2 wk2 dk
Fγ (E) = (A.2)
0 4π 2 νk (E − νk )
in (2.23) is given by
 
1 1 2 1 1
Fγ (E) = 2 (1 − z )ln − (z + 2) (A.3)
4π γ z 3 1 − z 2z 2

with
z = E/γ . (A.4)

At E = νk + i0+, we have
 
−i 1 2
ImFγ (E) = (z − 1) (A.5)
4πγ z 3

in agreement with (2.27).

 References

[1] J. Schwinger, Ann. Phys. (N.Y.) 2, 407(1957). J. C. Polkinghorne, Nuovo Cimento 8,

179(1958); 8, 781(1958). M. Gell-mann and M. Levy, Nuovo Cimento 16, 705(1960).
[2] C. Amsler et al. (Particle Data G Group), Physics Letters B667, 1(2008) and 2009 partial
update for the 2010 edition listed f0 (600) of mass 400−1200 Mev and full width 600−1000 Mev,
which closely resembles the appearance of a composite, similar to Na → Nbc of (2.14), that
will be analyzed in this paper.
[3] P. W. Higgs, Phys. Rev. Lett 12, 132(1964); Phys. Rev. 145, 1156(1966).
[4] S. Weinberg, Phys. Rev. D13, 974(1976).
[5] L. Susskind, Phys. Rev. D20, 2619(1979).
[6] C. T. Hill and E. H. Simmons, Phys. Report. 381, 235(2003).
[7] R. Contino, Y. Nomura and A. Pomarol, Nucl. Phys. B671, 148(2003).
[8] K. Agashe, R. Contino and A. Pomarol, Nucl. Phys. B717, 165(2005).
[9] T. D. Lee, Phys. Rev. 95, 1329(1954).
[10] G. Källen and W. Pauli, Dan. Mat. Fys. Medd. 30, 7(1955).
[11] S. Weinberg, Phys. Rev. 102, 285(1955).
[12] V. Glaser and G. Källen, Nucl. Phys. 2, 706(1956/57).
[13] G. Källen, Renormalization Theory in ”Lectures in Theoretical Physics”, W. A. Benjamin,
New York(1962), Vol. 1, pp. 169-256.
[14] W. Heisenberg, Proceedings of the Symposium on Basic Questions in Elementary Particle
Physics, June 8-18, 1971. Munich: Max-Planck Institute for Physics and Astrophysics: 1-10.
[15] T. D. Lee and G. C. Wick, Nucl. Phys. B9, 209-243(1969); Nucl. Phys. B10, 1-10(1969).