/home/www/ftp/data/hep-ph/dir_0308001/0308001.dvi


ANL-HEP-PR-03-057
EFI-03-39

Beautiful Mirrors, Unification of Couplings

and Collider Phenomenology

D.E. Morrisseya,b, and C.E.M. Wagnera,b

aHEP Division, Argonne National Laboratory, 9700 Cass Ave., Argonne, IL 60439, USA
bEnrico Fermi Institute, Univ. of Chicago, 5640 Ellis Ave., Chicago, IL 60637, USA

August 4, 2003

Abstract

The Standard Model provides an excellent description of the observables measured at
high energy lepton and hadron colliders. However, measurements of the forward-backward
asymmetry of the bottom quark at LEP suggest that the effective coupling of the right-
handed bottom quark to the neutral weak gauge boson is significantly different from the
value predicted by the Standard Model. Such a large discrepancy may be the result of a
mixing of the bottom quark with heavy mirror fermions with masses of the order of the
weak scale. To be consistent with the precision electroweak data, the minimal extension
of the Standard Model requires the presence of vector-like pairs of SU (2) doublet and
singlet quarks. In this article, we show that such an extension of the Standard Model is
consistent with the unification of gauge couplings and leads to a very rich phenomenology
at the Tevatron, the B-factories and the LHC. In particular, if the Higgs boson mass lies
in the range 120 GeV <∼ mh <∼ 180 GeV, we show that Run II of the Tevatron collider with
4–8 fb−1 of integrated luminosity will have the potential to discover the heavy quarks,
while observing a 3-σ evidence of the Higgs boson in most of the parameter space.



1 Introduction

In the absence of direct evidence for physics beyond the Standard Model (SM), precision elec-

troweak tests are the best way to get information about the scale and nature of a possible

breakdown of the SM description. While the SM has held firm in the face of a great number of

precision electroweak tests, the model has not emerged completely unscathed. Fits of the SM

to electroweak data show about a 2.5-σ deviation in the b-quark forward-backward asymmetry

(AbFB) [1], and this situation has not improved much in the last five years. This discrepancy is

important for two reasons. On one hand, it seems to indicate a significant deviation of the cou-

pling of the right-handed bottom quark to the Z-gauge boson (see, for example, Ref. [2]). On

the other, this measurement plays an important role in the present fits to the SM Higgs mass;

the removal of the heavy quark data from the electroweak fits would push the central values of

the Higgs mass to lower values, further inside the region excluded by the LEP2 searches [3].

There are two ways of solving this apparent discrepancy, and both of them seem to indicate

the presence of new physics. In Ref. [4] it was proposed to exclude the heavy quark data while

introducing new physics that raises the central value of the Higgs boson mass, and improves the

fit to the other observables. Such a task requires new physics that gives a negative contribution

to the S parameter, positive contributions to the U parameter and a moderate contribution

to the T parameter. At least two examples of this kind of physics have been presented in the

literature [4], [5]; the first within low energy supersymmetry and the second within a warped

extra-dimension scenario.

An alternative to this procedure is to take seriously the heavy quark data while introducing

new physics that modifies in a significant way the right-handed bottom quark coupling to the Z.

The Beautiful Mirror model of Ref. [6] accomplishes this by allowing the b-quark to mix strongly

with a set of exotic vector-like quarks. This model turns out to have several other interesting

features which we investigate in this paper. To be specific, we consider the unification of gauge

couplings, additional patterns of flavour mixing, the Higgs phenomenology, and searches for

the heavy vector quarks.

The model consists of the SM plus additional vector-like “mirror” quarks. These are a

pair of SU(2) doublets, ΨL,R =
(χ′L,R
ω′

L,R

)
, and a pair of SU(2) singlets, ξ′L,R. Here and in what

follows we use primed fields to denote gauge eigenstates, while mass eigenstates are written as

unprimed fields. The gauge group quantum numbers are the same as those of the analogous

SM particles: (3, 2, 1/6) for the doublets, and (3, 1,−1/3) for the singlets. Since the quarks are

1



added in vector-like pairs, these can have gauge-invariant Dirac masses, and the model is free

of anomalies. This is a minimal set of mirror quarks needed to improve the fit to electroweak

data.

The Yukawa and mass couplings of the mirror quarks are taken to be

L ⊃ −(ybQ̄′L + y2Ψ̄′L)b′RΦ − (ytQ̄′L + y4Ψ̄′L)t′RΦ̃ − M1Ψ̄′LΨ′R (1)
− (y3Q̄′L + y5Ψ̄′L)ξ′RΦ − M2ξ̄′Lξ′R + (h.c.)

where Q′L =
(
t′
L
b′
L

)
is the usual third generation SM quark doublet, and Φ =

(
φ+

φ0

)
is the Higgs

doublet. This is the most general set of renormalizable couplings provided the mirror quarks

couple only to each other and to the third SM generation.1 As pointed out in [6], the Yukawa

couplings yb, y3 and y4 are constrained to be much smaller than y2. Adjusting the ratio

( v√
2
y2)/M1 ' 0.7, where v = 246.22 GeV is the Higgs VEV, gives the best fit to precision

electroweak data while reducing the discrepancy in AbFB to about one standard deviation, and

keeping the left-right b-quark asymmetry measured at SLC within one standard deviation of

the measured value. This forces y2 to be O(1) since M1 & 200 GeV is needed to explain why
mirror quarks have not yet been observed. On the other hand, there are no strong constraints

on y5. As in [6], we will neglect it for simplicity.

This paper consists of seven sections. In Section 2 we examine the running of the gauge

couplings and their unification at a high scale. In section 3 we discuss the issue of flavour

mixing as well as the quark couplings to the neutral and charged weak gauge bosons, and the

Higgs. Section 4 consists of an investigation of the Higgs phenomenology in the model. In

Section 5 we review the current limits on exotic quarks and investigate the possibility of finding

mirror quarks at the Tevatron. In Section 6 we examine how the new types of flavour mixing

possible with mirror quarks can affect CP violation in B → φKs decays. Finally, Section 7 is
reserved for our conclusions.

2 Unification of Gauge Couplings

The idea that the low energy gauge forces proceed from a single grand unified description is a

very attractive one, and is supported by the apparent convergence of the weak, hypercharge and

strong couplings at short distances. The interest in low energy supersymmetry, for instance,

has been greatly enhanced by the discovery that the value of the strong coupling, αs(MZ),

1Note that couplings like Q̄′LΨ
′
R and ξ̄

′
Lb

′
R can be rotated away.

2



can be deduced if one assumes that the gauge couplings unify at a high scale. This prediction

depends on model-dependent threshold corrections at the GUT scale, but to within the natural

uncertainty in these corrections [18] the predicted value of αs(MZ) is perfectly consistent with

the values measured at low energies. In the Standard Model, instead, the assumption of gauge

coupling unification leads to a prediction for αs(MZ) that differs from the measured value by

an amount that is well beyond the natural uncertainties induced by threshold corrections.

In [6] it was noted that, to one-loop order, adding mirror quarks of the type considered

here to the SM greatly improves the prediction of αs(MZ) based on the assumption of gauge

coupling unification. We extend this analysis by including the two-loop contributions to the

gauge coupling beta functions and the low-scale threshold corrections. Since, for the consistency

of this study, the Higgs sector must remain weakly-coupled while the Higgs potential should

remain stable up to scales of the order of the unification scale, MG, we also investigate the

related issues of stability and perturbative consistency of the Higgs sector.

In extrapolating the low energy description of the theory to short distances, it is important

to remark that the Beautiful Mirror model [6] does not provide a solution to the hierarchy

problem. Therefore, a main assumption behind this extrapolation is that the physics that leads

to an explanation of the hierarchy problem does not affect the connection of the low energy

couplings to the fundamental ones. An example of such a theory construction is provided by

warped extra dimensions [7], and has been investigated by several authors [8]–[10]. In order

to preserve the good agreement with the precision electroweak data, the Kaluza-Klein modes

must be heavier than a few TeV in this case [11], and therefore the low energy physics analyzed

in the subsequent sections will not be affected. On the other hand, extra dimensions could

modify the proton decay rate in a significant way by introducing new baryon number violating

operators, and, in the case of warped extra dimensions with a Higgs field located in the infrared

brane, would make the issue of the running of the Higgs quartic coupling an irrelevant one. For

the rest of this section, we shall proceed with a pure four dimensional analysis of the evolution

of couplings and of the proton decay rate.

2.1 Renormalization Group Equations

Using the results of [12, 13], the two-loop (MS scheme) gauge coupling beta functions are

βl =
dgl
dt

= − 1
(4π)2

blg
3
l −

1

(4π)4

3∑
k=1

bklg
2
kg

2
l −

1

(4π)4
g3l Y

l
4 (F) (2)

3



where t = ln
(

µ
MZ

)
is the energy scale, and l = 1, 2, 3 refers to the U(1), SU(2), and SU(3)

gauge groups respectively. The first term is the one-loop contribution, while the other terms

come from two-loop corrections.

The coefficients bl and bkl are given by

b1 = −92, b2 = 76, b3 = 5, (3)

and

bkl = −



291
25

1 13
10

3 91
3

15
2

52
5

20 12


 . (4)

In the SM, the corresponding one-loop beta function coefficients are bSM1 = −41/10, bSM2 = 19/6
and bSM3 = 7. The variation of these coefficients are hence ∆b1 = 2/5 and ∆b2 = ∆b3 = 2.

Since b2 and b3 are shifted by an equal amount, they tend to unify at the same scale as in the

SM, about a few times 1016 GeV. Interestingly enough, the shift in b1 is much smaller than

that of b2 and b3, leading to, as we shall see, a successful unification of the three couplings.

The coefficients Y l4 (F) involve the Yukawa couplings. Neglecting the small Yukawa couplings

yb,y3,y4, and y5, they are

Y l4 (F) = Cltyt
2 + Cl2y2

2 (5)

where

Clf =




17
10

1
2

3
2

3
2

2 2


 (6)

and f = t,2.

The Yukawa couplings evolve according to

(4π)2
dyf
dt

= βfyf (7)

where f = t, 2. The one-loop and leading two-loop contributions to βf were calculated following

[14]. Of the two-loop terms, we include only those involving g3 or the Higgs self coupling λ; the

g3 terms are enhanced by large colour factors while the λ terms can become important when

investigating the stability of this coupling. The one-loop contributions are

(4π)2β
(1)
t =

9

2
y2t + 3y

2
2 −

(
17

20
g21 +

9

4
g22 + 8g

2
3

)
,

(4π)2β
(1)
2 =

9

2
y22 + 3y

2
t −

(
1

4
g21 +

9

4
g22 + 8g

2
3

)
. (8)

4



The two-loop contributions that we have included are

(4π)4β
(2)
t =

3

2
λ2 − 6y2t λ + g23(46y2t + 20y22) −

284

3
g43,

(4π)4β
(2)
2 =

3

2
λ2 − 6y22λ + g23(20y2t + 46y22) −

284

3
g43. (9)

The total beta-function is the sum of these pieces: βf = β
(1)
f +β

(2)
f . Aside from the modifications

due to the mirror quarks, these are in agreement with the results of [12].

The Higgs self-coupling λ is taken to be

L ⊃ µ2Φ†Φ − 1
2
λ(Φ†Φ)2. (10)

With this definition, the tree-level Higgs mass is mh =
√
λv, where v = 246.22 GeV =√

2 〈φ0〉. λ evolves according to dλ
dt

= βλ. We have calculated the one-loop and leading two-loop

contributions to βλ using the results of [15, 16]. As for the Yukawas, only the largest two-loop

terms involving g3 or λ were included. For the one-loop part, we obtain

(4π)2β
(1)
λ = 12λ

2 −
(

9

5
g21 + 9g

2
2

)
λ +

9

4

(
3

25
g41 +

2

5
g21g

2
2 + g

4
2

)
+ 12λ

(
y2t + y

2
2

)
− 12

(
y4t + y

4
2

)
. (11)

The two loop part is given by

(4π)4β
(2)
λ = −78λ3 − 72(y2t + y22)λ2 − 3(y4t + y42)λ + 60(y6t + y62)

+ 18(
3

5
g21 + 3g

2
2)λ

2 + 80g23(y
2
t + y

2
2)λ − 64g23(y4t + y42). (12)

Again, the total beta-function is the sum of the one- and two-loop parts.

2.2 Input Parameters and Threshold Corrections

We have investigated the running of these couplings numerically. The initial values MS scheme

values were taken from [17]:

α−1(MZ) = 127.922 ± 0.027 sin2 θw(MZ) = 0.23113 ± 0.00015
MZ = 91.1876 ± 0.0021 v = 246.22 GeV

m̄t(mt) = 165 ± 5 GeV.
(13)

These parameters correspond to the effective SU(3)c×U(1)em theory with five quarks obtained
by integrating out the heavy gauge bosons and quarks in the full SU(3)c×SU(2)×U(1)Y theory

5



at scale MZ. Threshold corrections to the gauge couplings arise in the process of matching the

theories. We define

α̃−11 =
3

5
(1 − sin2 θw)α−1, (14)

α̃−12 = sin
2 θwα

−1,

α̃−13 = α
−1
s .

Then the gauge couplings at MZ are given by

α−1l = α̃
−1
l + ρl, (15)

where the ρl terms represent threshold corrections. To one-loop order, these are[18]

ρ3 =
1

3π

∑
ζ

ln(
mζ
Mz

),

ρ2 = sin
2 θw

[
1

6π
(1 − 21 ln(MW

MZ
)) +

2

π

∑
ζ

Q2ζ ln(
mζ
MZ

)

]
,

ρ1 =
3

5

(
1 − sin2 θw

sin2 θw

)
ρ2, (16)

where the sums run over ζ = t,χ,ω,ξ. As shown in [6], the tree-level masses of the mirror

quarks are given by

mχ = M1, mω = (M
2
1 + Y

2
2 )

1/2
,

mξ = M2,
(17)

where Y2 =
v√
2
y2. These parameters are not completely independent. As explained above,

Y2 ' 0.7 M1 gives the best fit to electroweak data, while M2,M1 & 200 GeV are needed to
explain why these exotics have not yet been observed at the Tevatron[35] (see section 5).

2.3 Numerical Evolution

The unification of gauge couplings was investigated by fixing sin2 θw(MZ) and αem(MZ) accord-

ing to their measured values, and varying αs(MZ) until the gauge couplings unified to within

1%. GUT-scale threshold corrections were not considered. Figure 1 shows the range of αs(MZ)

obtained in this way for 250 GeV ≤ M2 ≤ 1000 GeV and all values of λ(MZ) and y2(MZ)
consistent with unification. (See the following section.) The range is plotted against the uni-

fication scale. In general, the unification is quite insensitive to the input values of M2,λ, and

6



α
s

Z
M(

)

16
(10   GeV)MG

 0.115

 0.1155

 0.116

 0.1165

 0.117

 0.1175

 0.118

 0.1185

 0.119

 0.1195

 0.12

 2.5  2.55  2.6  2.65  2.7  2.75  2.8  2.85  2.9  2.95  3

Figure 1: Range of αs consistent with a 1% unification of the gauge couplings plotted against
the unification scale MG.

y2. The scale of unification is quite high, MG = (2.80 ± 0.15) × 1016 GeV, depending on the
input values, at which point the unified gauge coupling constant has value α−1G = 35.11 ± 0.05.

The predicted range of the strong gauge coupling is in excellent agreement with the values

measured experimentally. This agreement is quite intriguing since the particular completion

of the Standard Model considered in this work is motivated by data and not by any model

building consideration. We shall not attempt to construct a grand unified model leading to

the appearance of the mirror quarks considered here in the low energy spectrum. Instead,

we will concentrate on additional issues regarding the renormalization group evolution of the

dimensionless couplings of the theory, as well as on exploring general features of the low energy

phenomenology of this particular extension of the Standard Model.

2.4 Stability and Non-Triviality of the Higgs

If the extrapolation of the model up to high scales is to be self-consistent, it should remain stable

and weakly-coupled up to the unification scale. The only source of trouble in this regard is the

Higgs self-coupling λ. Stability of the Higgs sector requires λ(Q) > 0, for all Q < MG while

7



perturbative consistency means that λ must not be too large. For concreteness we demand 2

that 0 < λ < 2 up to 1017 GeV. This is sufficient to guarantee that the effective potential is

similarly well behaved [19]. The evolution of λ is largely determined by the initial values of y2

and λ. Only for a small subset of initial values does λ remain well-behaved (i.e. 0 < λ < 2) up

to MG. This subset is shown in Figure 2, where we have written λ in terms of the (tree-level)

Higgs mass, and Y2 in terms of mχ assuming Y2 = 0.7M1.

We compare the allowed region with the region favoured by precision electroweak data found

in [6]. There is a small overlap between the allowed band found here and the 1-σ allowed region

of [6] corresponding to mh ∼ 170 GeV and mχ ∼ 200 GeV (mω ∼ 250 GeV).

m   (GeV)h

m
   

(G
eV

)
χ 2−σ

1−σ

190

200

210

220

230

240

250

260

270

280

150 160 170 180 190 200 210 220

Figure 2: Region in the mh-mχ plane that is consistent with unification (shaded region) and
precision electroweak data (below the dashed and solid lines).

2.5 Proton Decay

Grand unified models induce baryon number violating operators that lead to a proton decay

rate that may be observable in the next generation of proton decay experiments. The present

bounds on the proton lifetime [20] already put relevant bounds on grand unified scenarios. In

four dimensional supersymmetric grand unified models, for instance, dimension five operators

2This upper limit on λ is somewhat arbitrary but fairly unimportant in the present case since λ grows very
quickly when it becomes larger than unity.

8



may easily induce a proton lifetime shorter than the present experimental bound [21]. This

situation may be avoided by a suitable choice of the low energy spectrum [22]. Heavy first and

second generation sfermions and light gauginos are preferred from these considerations.

In the model under study there are no dimension-five operators, so the dominant decay

mode is expected to be p → π0e+. The high unification scale obtained above means that the
proton will be long-lived regardless of the details of the unification mechanism. If proton decay

proceeds via SU(5) gauge bosons, the decay rate is given by [23]

Γ(p → π0e+) = πmpα
2
G

8f2πM
4
G

(1 + D + F)2α2N
[
A2R + (1 + |Vud|2)2A2L

]
(18)

where fπ = 0.131 GeV is the pion decay constant, D = 0.81 and F = 0.44 are chiral Lagrangian

factors, αN is a coefficient related to the π
0p operator matrix element, and AL and AR are cor-

rection factors due to the running of the couplings. A recent lattice-QCD calculation gives [24]

|αN| = 0.015(1) GeV3, where the uncertainty is purely statistical. The systematic uncertainty
is probably much larger; we take it to be ∼ 50% [22]. The correction factors AL,R split into
long and short distance pieces: AL,R = Al

∏3
i=1 A

L,R
i , where Al comes from the renormalization

group evolution below MZ and A
L,R
i from that above. Here, Al ' 1.3 is identical to the SM

value, while the short distance pieces, to one-loop order, are [25]

AL3 =
[
α3(MZ )
αG

]6/(33−4ng−6)
' 3.15 = AR3 ,

AL2 =
[
α2(MZ )
αG

]27/(86−16ng−24)
' 1.39 = AR2 ,

AL1 =
[
α1(MZ )
αG

]−69/(6+80ng +24)
' 1.14,

AR1 =
[
α1(MZ )
αG

]−33/(6+80ng +24)
' 1.07.

(19)

Using MG = 2.8 × 1016 GeV, and α−1G = 35.1, we find

τ(p → π0e+) = 3 × 1036±1yrs, (20)

well in excess of the Super-Kamiokande bound of τ(p → π0e+) = 5.3 × 1033 yrs [20].

3 Flavour Mixing

Extending the matter content of the SM also introduces new sources of flavour mixing. With

mirror quarks, this pattern can be quite complicated, involving right-handed couplings to the

9



W, and tree-level flavour-changing couplings to the Z and the Higgs. We consider first the

generic case, taking the most general set of Yukawa and mass terms possible. Next, we simplify

our results making use of the fact that, in the model under study, the mirror quarks couple

only to the third generation quarks and calculate explicitly the couplings of the heavy quarks

to the weak gauge bosons and the Higgs. In subsequent sections we shall use these results to

investigate the Higgs phenomenology, the collider searches for mirror quarks, and CP violation

in B → φKs decays.
Let λu and λd be the flavour-space mass matrices describing the flavour mixing between

gauge eigenstates. These matrices will be 4×4 and 5×5 respectively, and will have contributions
from Yukawa couplings and Dirac mass terms. Both matrices can be diagonalized by bi-unitary

transformations:

λu = UuDuW
†
u, λd = UdDdW

†
d, (21)

where the U’s and W ’s are unitary, and Du and Dd are the diagonalized mass matrices. The

corresponding (unprimed) mass eigenstates are then related to the (primed) gauge eigenstates

by the unitary transformations

u′AL = U
AB
u u

B
L, u

′A
R = W

AB
u u

B
R,

d′PL = U
PQ
d d

Q
L, d

′P
R = W

PQ
d d

Q
R.

(22)

Here, the indices A,B = 1, . . . 4 correspond to {u,c,t,χ}, while P,Q = 1, . . . 5 refer to
{d,s,b,ω,ξ} respectively.

In terms of the physical (mass eigenstate) fields, the charged currents become

√
2J

µ
W +

= ūBLγ
µV

BQ
L d

Q
L + ū

B
Rγ

µV
BQ
R d

Q
R,

J
µ
W− = J

µ
W +

†
, (23)

where the 4 × 5 flavour-mixing matrices are given by

V
BQ
L =

4∑
i=1

UiBu
∗
U
iQ
d ,

V
BQ
R = W

B4
u

∗
W

4Q
d . (24)

The matrix VL is analogous to the CKM matrix VCKM of the SM. It is nearly unitary in the

sense that VLVL
† = I4×4 and V

†
LVL = I5×5 − Vd, where the matrix Vd is defined below. The

matrix VR describes right-handed couplings, and has no analogue in the SM.

10



Similarly, the hadronic neutral current is

cos θwJ
µ
Z = ū

A
Lγ

µ(
1

2
− 2

3
sin2 θw)u

A
L + ū

A
Rγ

µ(−2
3

sin2 θw)u
A
R

+d̄PLγ
µ(−1

2
+

1

3
sin2 θw)d

P
L + d̄

P
Rγ

µ(
1

3
sin2 θw)d

A
R

+
1

2

(
ūARγ

µV ABu u
B
R − d̄PRγµṼ PQd dQR + d̄PLγµV PQd dQL

)
(25)

where the matrices Vu,Vd, Ṽd are given by

V ABu = W
4A
u

∗
W 4Bu , (26)

V
PQ
d = U

5P
d

∗
U

5Q
d ,

Ṽ
PQ
d = W

4P
d

∗
W

4Q
d .

The off-diagonal elements of these matrices describe FCNCs. Each is Hermitian and satisfies

V 2 = V .

3.1 Heavy Quark Neutral and Charged Currents

The expressions above can be simplified considerably by using the fact that, in the model under

study, the (gauge eigenstate) mirror quarks couple only to the quarks of the third generation (see

Eq. (1)). The mixing between the mirror quarks and the first and second generation quarks is

thus very small, and so will be neglected. Moreover, the mixing between the SM quarks is given

approximately by the usual CKM description. The flavour violating effects among the heavy

quarks are then related to the mixing of the right-handed SU(2) quark-doublet with the third

generation right-handed quarks, as well as the mixing of the left-handed SU(2) quark-singlet

with the left-handed bottom-quark.

The mixing in the top sector must be very small to avoid a conflict with the B(b → sγ)
predictions [26]. We shall therefore assume, for simplicity, vanishing mixing in the top sector

(y4 = 0). The top-sector mass matrix is then diagonal, with mt =
v√
2
yt and mχ = M1.

Following [6] we take y5 = 0 as well. The bottom-sector mass matrix, in the basis (b
′,ω′,ξ′),

is then given by

λd =


 Yb 0 Y3Y2 M1 0

0 0 M2


 (27)

where Yi =
v√
2
yi, i = b, 2, 3. The phenomenologically interesting regime is Yb,Y3 � Y2,M1,M2 [6].

Working to linear order in the small quantities Yb/M1 and Y3/M2, the left- and right-handed

11



mixing matrices are

Ud =


 cLc̃L sL cLs̃L−sL cL 0
−s̃L 0 c̃L


 (28)

and

Wd =


 cR sR 0−sR cR 0

0 0 1


 (29)

where sR = sin θR, sL = sin θL, and s̃L = sin θ̃L are given by

sR =
Y2

(Y 22 + M
2
1 )

1/2
,

sL =
YbY2

(M21 + Y
2
2 )
,

s̃L =
Y3
M2

. (30)

Applying the mixing matrices to λd, the b-sector masses are

mb = Yb

(
1 +

Y 22
M21

)−1/2
,

mω = (M
2
1 + Y

2
2 )

1/2, mξ = M2.
(31)

To obtain a good agreement between the predictions of the model and precision electroweak

data the angle in the right-handed sector must be sizeable,

tan θR = Y2/M1 ' 0.7, (32)
while s̃L should be small,

sin θ̃L ' 0.09. (33)
Note that Eq. (32) fixes sL in terms of the b mass.

In this approximation, the relevant neutral currents read

cos θwJ
µ
Z = b̄Rγ

µbR

(
−s

2
R

2
+

sin2 θW
3

)

+ ω̄Rγ
µωR

(
−c

2
R

2
+

sin2 θW
3

)
−

(
b̄Rγ

µωR + h.c.
) sRcR

2

+ b̄Lγ
µbL

(
−c̃

2
R

2
+

sin2 θW
3

)

+ ξ̄Lγ
µξL

(
−s̃

2
L

2
+

sin2 θW
3

)
+

(
b̄Lγ

µξL + h.c.
) s̃Lc̃L

2

+ ω̄Lγ
µωL

(
−1

2
+

sin2 θW
3

)
+ ξ̄Rγ

µξR
sin2 θW

3
. (34)

12



Within the same approximation, the charged currents read

J
µ
W +

= t̄Lγ
µ (c̃LbL + sLωL + s̃LξL)

+ χ̄Lγ
µ (ωL − sLbL)

+ χ̄Rγ
µ (cRωR − sRbR) , (35)

where we have neglected terms of order m2b/M
2
1 .

3.2 Higgs Couplings

One of the most important immediate goals of high energy physics is to understand the mech-

anism of electroweak symmetry breaking. In the Standard Model and its supersymmetric

extensions, this symmetry is broken spontaneously through the vacuum expectation value of

one or more scalar Higgs bosons. The same is true for the model under study and it is therefore

quite relevant to understand the possible modification of the Higgs boson search channels at

the Tevatron and the LHC.

In addition to introducing new sources of flavour mixing, mirror quarks also modify the

couplings to the Higgs. The Dirac mass terms for the mirror quarks mean that the Higgs-

quark couplings need no longer be flavour diagonal in the basis of mass eigenstates, nor be

proportional to the quark masses. We find that the coupling of the Higgs to the b quark is

reduced relative to the SM. This, along with the contribution of the heavy quarks in loops, has

interesting consequences for the detection of the Higgs.

As in the previous section, we will assume that the mirror quarks mix almost exclusively

with the third generation quarks and take y4,y5 ≈ 0. This implies that the only Higgs-quark
coupling that is significantly modified from the SM is that of the b-quark. The relevant terms

in the Lagrangian are therefore

L ⊃ −(ybQ̄′L + y2Ψ̄′L)b′RΦ − y3Q̄′Lξ′RΦ − M1Ψ̄′LΨ′R − M1ξ̄′Lξ′R + (h.c.). (36)
After symmetry breaking Φ = 1√

2

(
0

v+h

)
in the unitary gauge. These couplings can then be

written as

L ⊃ −d̄′L (Mb + hNb) d′R + (h.c.) (37)
with d′L,R =

(
b′L,R,ω

′
L,R,ξ

′
L,R

)t
, and

Nb =


 Yb 0 Y3Y2 0 0

0 0 0


 . (38)

13



Transforming to the mass eigenstate basis, the Higgs couplings become

L ⊃ −c2R
mb
v
hb̄LbR − s2R

mω
v
hω̄LωR (39)

− sRcRh
mb
v
b̄LωR − sRcRh

mω
v
ω̄LbR

− mξs̃L
v

h
(
b̄L + s̃Lξ̄L

)
ξR + (h.c.).

Using the value tan θR = Y2/M1 ' 0.7 favoured by the model, we find that the hbb coupling is
reduced by a factor of c2R ∼ 2/3.

4 Higgs Phenomenology

4.1 Higgs Production and Decay

This scenario modifies the phenomenology of the Higgs in two ways. First, by reducing the

hbb coupling by a factor of c2R, the partial width Γ(h → b̄b) is attenuated by the square of
this factor, c4R ∼ 4/9. Since this channel is dominant for Higgs masses below mh ' 130 GeV,
the reduction of the partial width for this mode increases the branching fractions of the other

modes in this range. Secondly, ω quark loop effects increase the partial width Γ(h → gg). This
increases both the branching fraction of this mode, and the Higgs production cross-section by

gluon fusion.

Let us consider the effect of the ω quark in a bit more detail. The presence of this quark

in a loop connecting the Higgs to two gluons modifies the h → gg partial width. Neglecting
light quark contributions, and keeping only the effects of the dominant Yukawa coupling y2,

the partial width becomes

Γ(h → gg) = αα
2
s

128π2 sin2 θw

(
m3h
M2W

) ∣∣F1/2(τt) + s2RF1/2(τω)∣∣2 (40)
where the function F1/2(τq) is given by [27]

F1/2 = −2τq[1 + (1 − τq)f(τ)], (41)

with τq = 4(
mq
mh

)2 and

f(τ) =

{
[sin−1(1/

√
τ)]2; τ ≥ 1

−1
4
[ln(η+/η−) − iπ]2; τ < 1; η± = 1 ±

√
1 − τ. (42)

14



The corresponding expression in the SM is obtained by setting sR = 0. Since the new term

interferes constructively, the effect is to increase the decay width. While the h → gg mode isn’t
directly observable at hadron colliders, this decay width is nevertheless important because it

determines the cross-section for Higgs production by gluon fusion; σ(gg → h) ∝ Γ(h → gg),
up to soft gluonic effects. The h → γγ decay width is similarly modified by an ω loop. In this
case, the new contribution interferes destructively with the SM terms, the dominant parts of

which come from W and Goldstone boson loops. However, the change in Γ(h → γγ) is very
small for any reasonable input parameter values. Figure 3 shows the dominant Higgs decay

branching ratios in the model under study. Additional NLO corrections to the h → gg mode
were included as well, following [28].

  m   (GeV)h

lo
g 

  (
B

ra
nc

hi
ng

 F
ra

ct
io

n)
10

ω ω

   b b

  c c
t t

bω bω+

γγ

ZZ

WW

gg

ττ

−4

−3.5

−3

−2.5

−2

−1.5

−1

−0.5

0

100 150 200 250 300 350 400 450 500 550

Figure 3: Higgs Branching Ratios with Mirror Quarks, for mω = 250 GeV and Y2/M1 = 0.7.

Figures 4 and 5 show the enhancement of a few Higgs discovery modes relative to the SM,

which is mostly due to the increase in the gluon fusion cross-section. For low Higgs masses,

mh . 150 GeV, there is an additional enhancement of certain modes as a result of the decrease
in the h → b̄b branching ratio. It should be noted, however, that such low Higgs masses worsen
the fit to the precision electroweak data in this model.

If the Higgs mass exceeds 150 GeV the process h → V V , where V is a real or virtual vector
boson, becomes the primary Higgs discovery mode at both the Tevatron and the LHC [29, 30].

The inclusive modes are enhanced by the increase in gluon fusion, while modes in which the

15



σ (gg−>h)
σ

SM
(gg−>h)

σ
SM

(tth)
B(tth)
B

SM

σ (h−> bb)
(h−>bb)

m   (GeV)h

σ (VVh)
σ

SM
(VVh)

(h−>      )ττB

(h−>      )ττB
SM

(h−>     )γγB
(h−>     )γγB

SM

σ (gg−>h)
σ

SM
(gg−>h)

 0

 0.5

 1

 1.5

 2

 2.5

 3

 100  105  110  115  120  125  130  135  140  145  150

Figure 4: Enhancement of low mass Higgs
production and detection modes for mω =
250 GeV, Y2/M1 = 0.7.

m   (GeV)h

σ (gg−>h)
σ

SM
(gg−>h)

(h −> VV)Bσ
(h −> VV)

SMBσSM

(VBF)

(VBF)

σ (gg−>h)

σ
SM

(gg−>h)

(h −> VV)B
(h −> VV)

SMB

0

0.5

1

1.5

2

2.5

3

150 200 250 300 350 400 450 500 550

Figure 5: Enhancement of intermediate mass
Higgs modes for mω = 250 GeV, Y2/M1 = 0.7.

Higgs is produced by other means, such as vector boson fusion, are very slightly attenuated

due to Higgs decays into mirror quarks.

4.2 Higgs Searches at the Tevatron and the LHC

The enhancement of Higgs detection signals decreases the collider luminosity needed to find

the Higgs. We have translated the above results into collider units using detector simulation

results from [31, 32, 33] for the Tevatron, and [33, 34] for the LHC. Figure 6 shows the minimum

luminosity per detector (with CDF and D∅ data combined) needed for a 3-σ signal at the
Tevatron. Figure 7 shows the luminosity needed for a 5-σ discovery at the LHC. All channels

displayed in this plot are for CMS alone except for the WW and ZZ modes, which are for

ATLAS. These plots correspond to inclusive searches unless stated otherwise: V V → h →
X denotes weak boson fusion channels, while t̄th → X and W/Zh → X denote associated
production channels. For the inclusive channels, we have assumed that gluon fusion accounts

for 80% of the total Higgs production.

The model significantly improves the likelihood of observing the Higgs at the Tevatron for

a Higgs mass between 120 and 180 GeV. Note that gg → h → τ+τ− becomes the dominant
discovery channel at the Tevatron collider in the low Higgs mass region.

The predictions of the model for the LHC are less dramatic, although the improvement

in the gg → h → γγ process will make a light Higgs much easier to see. For intermediate
Higgs masses, the inclusive h → WW channel becomes competitive with the V V → h → WW

16



L
m

in
(f

b 
  )−1

MQ

SM

hm   (GeV)

h −> ττ

W/Zh −> bb

50

h −> WW

 1

 10

 100  120  140  160  180  200

Figure 6: Minimum luminosity needed for a 3-σ Higgs signal at the Tevatron for mω = 250 GeV,
Y2/M1 = 0.7.

channel. For masses larger than the ones displayed in the Figure 7, searches for the Higgs at

the LHC can proceed via the golden mode h → ZZ.

MQ

SM

L
m

in
(f

b 
  )−1

hm   (GeV)

VV−>h−> ττ h−> γγ

h−> γγ

VV−>h−> ττ

h−> WW

h −> ZZ

h −> ZZ

tth−> ttbb

tth−> ttbb

 0

 10

 20

 30

 40

 50

 60

 100  120  140  160  180  200

Figure 7: Minimum luminosity needed for a 5-σ Higgs discovery at the LHC for mω = 250 GeV,
Y2/M1 = 0.7.

17



5 Mirror Quark Collider Signals

If the model is to improve the electroweak fits, the mirror quarks must not be too heavy. In

particular, this requires mχ
<∼ 250 GeV, which (using (17)) implies mω <∼ 300 GeV as well.

On the other hand, mξ is largely unconstrained. These relatively low masses suggest that the

Tevatron may be able to see mirror quarks by the end of Run II. Previous searches for exotic

quarks have concentrated on a possible fourth generation b′ quark. In the most recent of these,

CDF has put a lower bound on the b′ mass of mb′ > 199 GeV[35], provided the branching ratio

b′ → bZ is 100%. This bound is relevant to our model as well.
At the Tevatron, mirror quarks are produced mostly by q̄q annihilation, with a smaller

contribution from gluon fusion. Previous calculations of the top-quark pair-production cross-

section apply to mirror quarks as well. These indicate σq̄q ' 3.0−0.5 pb, for mq = 200−300 GeV
at the centre of mass energy

√
s = 2.0 TeV[36, 37]. This is small, but comparable to the top

production cross section in Run I, σt̄t = 6.1 ± 1.1 pb, where we have averaged the results of D∅
and CDF[17].

The up-type χ quark is most strongly constrained in the model. It decays almost entirely

by χR → bRW due to a large tree-level right-handed W coupling, Eq. (35). This will produce
a signature very similar to that of the top quark. Indeed, top quark decays present a nearly

irreducible background. Searching for the χ therefore reduces largely to a counting experiment

in which one compares the number of measured top events to the number expected. Searches

at Run I of the Tevatron have already put interesting limits on mχ since the top production

cross-section measured there agrees well with SM predictions[37] (See, however, the comments

in [38].)

In order to make a quantitative estimate of the present bound on a possible sequential

top quark, we conservatively assume equal acceptance and detection rates for both χ and top

events. The fraction of χ events in the top sample will be 40-10% for 200 < mχ < 250 GeV.

(In practice, the detection efficiency for χ’s will be slightly better since the b-tagging efficiency

improves with jet PT .) Comparing the Run I value for the top production cross-section with

the theoretical prediction [37], we get that the cross-section for any new sequential top quark

should be lower than 2.9 pb, at the 2-σ level. This method, based just on counting, leads to a

bound of about

mχ
>∼ 200 GeV. (43)

This in turn implies mω
>∼ 250 GeV, well in excess of the b′ bound from [35].

18



At Run II, the goal is to determine the top production cross section to an accuracy of

7–9% [39] with a few fb−1 of data. Assuming that this goal is achieved, and considering that

the uncertainty is mostly due to systematic effects (in particular, the proper determination of

the b-tagging efficiency) and therefore weakly dependent on small luminosity variations, the

Tevatron Run II will be sensitive enough to rule out the presence of a sequential top quark

with a mass smaller than 230 GeV. This will imply an indirect bound on mω > 285 GeV. On

the other hand, the Tevatron might see evidence of a χ quark with mass smaller than 220 GeV

at the 3-σ level.

That the χ → bW vertex is (V + A), Eq. (35), makes the χ somewhat easier to find. This
is because the W +’s emitted in χR → bRW + have positive or zero helicity, whereas those from
tL → bLW + have negative or zero helicity 3. Leptons emitted by positive helicity W ’s tend
to be harder than those from longitudinal or negative helicity W ’s. Thus, a slightly higher

lepton PT cut will increase the relative acceptance of χ’s, although the improvement will be

small since the majority of W ’s emitted are longitudinal. CDF has looked for positive helicity

W ’s in top decays. They find a positive helicity fraction of F+ = 0.11 ± 0.15[40], consistent
with both the SM, and a χ of mass above about 200–250 GeV, for which we predict a value of

F+ . 0.08–0.02 [41].
Run II at the Tevatron will also cover part of the mass range of the ω quark by direct searches

for this particle. The strong ωb mixing leads to tree-level bZω and bhω vertices, Eqs. (34),(39),

with the same O(1) flavour-mixing factors. The dominant decay modes are thus ωR → bRZ,
and ωR → bLh provided the Higgs isn’t heavier than the ω. Other modes are suppressed by
loops, small flavour-mixing factors, and in the case of ω → χW , phase space. Indeed, this decay
is forbidden for almost all of the model parameter space consistent with precision electroweak

data[6]. If the Higgs is heavier than the ω, the CDF bound applies directly to ω̄ω → b̄bZZ
modes and constrains the ω mass to be greater than 199 GeV.

Things are more interesting if the Higgs is lighter than the ω. In this case, the ratio of the

decay widths of the Higgs and Z modes is [42]

Γ(ω → bh)
Γ(ω → bZ) =

(1 − rh)2
(1 − rZ)2(1 + 2rZ)

(44)

where rh = (mh/mω)
2, rZ = (MZ/mω)

2. Figure 8 shows the branching ratios for these modes

for a Higgs mass of 170 GeV. In their b′ search, CDF looked for b̄′b′ → b̄bZZ events in which
3Correspondingly, the W −’s emitted in the charge conjugate decays have negative or zero helicity for a

(V + A) vertex, and positive or zero helicity for a (V − A) vertex. We shall refer to both positive helicity W +’s
and negative helicity W −’s as “positive helicity” W ’s, and so on.

19



one Z decayed into jets while the other decayed into a pair of high transverse momentum (PT )

leptons[35]. They only accepted events in which the reconstructed mass of the lepton pair lay

within the range 75-105 GeV and at least two jets were tagged as b’s. For low Higgs masses,

below 150 GeV, this search strategy is sensitive to both ωω → bZbZ and ωω → bZbh events,
since in the latter, the Higgs decays predominantly into a b̄b pair which mimics the hadronic

decay of a Z. This also lends itself to modifying the search strategy to include four b-tags [42].

Even more search strategies become possible if the Higgs mass exceeds 150 GeV as favoured

by the model. Such massive Higgs bosons decay mostly into WW pairs (see Figure 3) so for

instance, ωω → bZbh could be distinguished by looking for events with four jets, at least two
of which are b-tagged, accompanied by a pair of high PT leptons and large missing ET .

In order to estimate the possibility of observing an ω quark at the Tevatron using the CDF

b′ search strategy, we shall assume that the Higgs mass is mh = 170 GeV, and that this search

strategy has no sensitivity to the ω → bh modes and a detection efficiency of 13% (as in Run I
for large mb′). The most important background comes from Z

0 events associated with hadronic

jets. To reduce the background, one can impose a cut on the total transverse energy of the

jets [35]. For mω > 250 GeV, a cut of
∑
ET > 150 GeV will eliminate most of the background

without reducing the signal in a significant way. The number of observable signal events scales

with the luminosity and is approximately equal to

Nωω̄ ' 1.5 − 7.0 L[fb−1] , (45)

for an ω mass varying from 300 GeV to 250 GeV. Due to the smallness of the background, a

simple requirement for evidence of a signal is that at least five events be observed. Therefore

the Tevatron Run II should cover the whole mass range of the model, mω < 300 GeV, if the

luminosity is above 4 fb −1. For a luminosity of 2 fb−1, a signal may be observed up to masses

of about 280 GeV.

The isosinglet ξ decays predominantly by ξL → bLZ, ξL → tLW , and ξR → bLh (see
Eqs. (34),(35) and (39)). All three of these go at tree level. Figure 9 shows the corresponding

branching ratios for these decays. The ξ → bZ mode is dominant for mξ . 400 GeV, although
the ξ → tW quickly becomes important above the tW-production threshold[43]. For relatively
light ξ’s, of mass less than 250 GeV, the search strategies are similar to those for the ω. The

CDF b′ search is thus sensitive to a light ξ. Their b′ mass bound also applies in this case,

although this limit may be weakened somewhat depending on the mass of the Higgs.

20



m   (GeV)ω

ω −> bZ

ω −> bh

B
ra

nc
hi

ng
 F

ra
ct

io
n

 0

 0.2

 0.4

 0.6

 0.8

 1

 200  250  300  350  400

Figure 8: Branching ratios for decays of the ω
quark with mh = 170 GeV.

m   (GeV)ξ

ξ −> bZ

ξ −> tW

ξ −> hZ

B
ra

nc
hi

ng
 F

ra
ct

io
n

 0

 1

 200  250  300  350  400

 0.2

 0.4

 0.6

 0.8

Figure 9: Branching ratios for decays of the ξ
quark with mh = 170 GeV.

6 CP-violation in B0 → φKs Decays
The value of the CP-violation parameter sin(2β) measured in B0 → φKs decays appears to
disagree with the value extracted from B0 → J/ψKs decays: [44, 45, 46]

sin(2β) =

{
+0.735 ± 0.054; B0 → J/ψKs
−0.39 ± 0.41; B0 → φKs (46)

This discrepancy is particularly interesting because the B → φKs mode is loop mediated,
making it much more sensitive to new physics than the B → J/ψKs mode, which goes at tree-
level. We investigate whether this discrepancy can be explained by the FCNC’s which arise in

the model.

It is not sin(2β) that is measured directly, but rather the time-dependent CP asymmetry

aCP (t). For decays of the B
0(= b̄d) meson into a CP eigenstate f, this is defined to be [47]

a
f
CP (t) :=

Γ(B0(t) → f) − Γ(B̄0(t) → f)
Γ(B0(t) → f) + Γ(B̄0(t) → f) (47)

= Cf cos(∆Mt) − Sf sin(∆Mt)

where ∆M is the mass difference between the mass eigenstates, and Sf , Cf are given by

Cf = 1−|ξf |
2

1+|ξf |2 , Sf = −
2Imξf
1+|ξf |2 , (48)

with ξf given by (for the B
0
d system)

ξf ≡ e−2iβ
A(B̄0 → f̄)
A(B0 → f). (49)

21



In the SM, the B̄ → φK̄s amplitude has the form Ā = λtA0, where A0 is CP-invariant and
λt = V

ts∗
CKMV

tb
CKM. The phase of λt is very small [47], so to a good approximation CφK = 0,

and SφK = sin(2β). This result applies to the B → J/ψKs mode as well. For this reason, the
values of Sf measured in the φKs and J/ψKs modes are sometimes quoted as sin(2β), as we
have done in (46).

Physics beyond the SM can change the values of SφKs and SJ/ψKs by adding additional
CP-violating terms to the decay amplitudes. We have investigated whether the new tree-level

FCNC Z-couplings

J
µ
Z ⊃

1

2 cos θw

(
s̄Lγ

µV sbd bL − s̄RγµṼ sbd bR
)

+ (h.c.) (50)

can explain the apparent difference between SφK and SJ/ψKs . To simplify the analysis, we have
assumed that the mixing of the mirror quarks with the SM quarks (other than the b) is very

small. In particular, we have neglected all right-handed W couplings and all FCNC couplings

other than those connecting the b and s quarks. |V sb| and |Ṽ sb| will also be treated as small
parameters whose size we will bound below. These assumptions imply that the three generation

CKM description of flavour mixing in the SM is correct up to small modifications, and that we

can ignore the loop contributions of mirror quarks to the effective sb vertex.

6.1 Constraints from Semi-Leptonic Decays

The strongest constraints on these couplings come from semileptonic b → s modes [48]. At
energies much below MZ, the SM decay amplitudes can be written as the matrix elements of

an effective Hamiltonian;

Heff = Heff(b → sγ) −
GF√

2
λt(C

SM
9V

Q9V + C
SM
10A

Q10A ), (51)

where Q9V = (s̄b)(V −A)(l̄l)V , and Q10A = (s̄b)(V −A)(l̄l)A are four-quark operators, λt = V
ts∗
CKMV

tb
CKM,

and l = µ,e. (See [49] for a definition of Heff(b → sγ).) The FCNC couplings contribute to
Q9V and Q10A , and generate the new (V + A) operators Q

′
9V

= (s̄b)(V +A)(l̄l)V , and Q
′
10A

=

(s̄b)(V +A)(l̄l)A.

We neglect the contribution of mirror quarks to Heff (b → sγ), since these only arise from
loops, that become negligible for y4 = 0, contrary to the dominant tree-level effects included in

our analysis. The effective Hamiltonian thus becomes

Heff = Heff (b → sγ) −
GF√

2
λt

(
C9Q9V + C10Q10A + C

′
9Q

′
9V

+ C′10Q
′
10A

)
. (52)

22



In terms of η :=
V sb

d

λt
and η′:= − Ṽ

sb
d

λt
, the Wilson coefficients are now given by

C9V = C
SM
9V

+ ( 1
2
− 2 sin2 θw)η, C′9V = ( 12 − 2 sin2 θw)η′,

C10A = C
SM
10A

− 1
2
η, C′10A = −12η′.

(53)

Since 1
2
− 2 sin2 θw ' 0.05, to a good approximation we need only consider the shift in

the C10A coefficients. We have examined the effect of modifying the C10A coefficients on the

branching ratios of the inclusive B → Xsl+l− mode, as well as the two exclusive B → Kl+l−
and B → K∗l+l− decays.

The constraint on b → s FCNC’s from the B → Xsl+l− mode has been considered previously
(e.g.[49, 50, 51]). We repeat the analysis for this particular model using updated input values.

From [49], modified to include the new (V + A) operators and neglecting lepton masses, the

shift in the branching ratio relative to the SM is

∆BXs = B − BSM (54)

=
α2

8π2f(z)κ(z)

∣∣∣∣VtsVcb
∣∣∣∣
2 (
|C̃10A|2 + |C̃′10A|2 − |C̃SM10A |2

)
B(b → ceν̄),

where C̃i :=
2π
α
Ci, f(z) = 0.54 ± 0.04 is a phase space factor, κ = 0.879 ± 0.002 is a QCD

correction, and B(b → ceν̄) = 0.109 ± 0.005. We have also taken CSM10A = −
Y0(xt)

sin2 θw
' −4.2,

α−1(mb) = 129, and
∣∣∣VtsVcb

∣∣∣2 = 1 in our analysis. The measured branching ratio and the SM
prediction for this mode are listed in Table 6.1, as is the 2-σ allowed shift in the branching

ratio based on these values.

Mode Bexp(10−6) BSM(10−6) 2-σ Allowed Range (10−6)
B → Xsl+l− 6.1 ± 1.4+1.4−1.1 [53] 5.5 ± 0.6 [54] -3.6 < ∆BXs < 4.8
B → Kl+l− 0.76+0.19−0.18 [55, 56] 0.35 ± 0.12 [54] −0.09 < ∆BK < 0.91
B → K∗l+l− 1.68+0.68−0.58 ± 0.28 [56] 1.39 ± 0.31 [54] −1.3 < ∆BK

∗
< 2.1

Table 1: Experimental inputs and SM predictions. All errors were combined in quadrature,
and the SM predictions were averaged over µ and e modes.

The inclusive modes B → Kl+l− and B → K∗l+l− have been considered in [51, 52]. The
result for the latter mode, neglecting lepton masses, is [51]

∆BK∗ = (4.1+1.0−0.7) × 10−8
(
|C̃10A − C̃′10A|2 − |C̃SM10A |2

)
(55)

+ (0.9+0.4−0.2) × 10−8
(
|C̃10A + C̃′10A|2 − |C̃SM10A |2

)
.

23



Table 6.1 lists the experimental values, the theoretical SM prediction, and the corresponding

2-σ allowed range for ∆BK∗.
For the B → Kl+l− mode, the shift in the branching ratio is [51]

∆BK = G
2
Fα

2m5B
1536π5

τB|λt|2I
(
|C̃10A + C̃′10A|2 − |C̃SM10A |2

)
, (56)

where τB = 1.60±0.05 ps is the total B lifetime, and I is an integral of form factors. Explicitly,
I =

∫ ŝ1
ŝ0
dŝλ

3/2
K (ŝ)f

2
+(ŝ), where 0 '

4m2
l

m2
B
≤ ŝ ≤ (mB−mK )2

m2
B

, and λK(ŝ) = 1 + r
2
K + ŝ

2 − 2ŝ − 2rKŝ
with rK = (mK/mB)

2
. We have evaluated the integral I numerically using the form factors

in [52], and find I = (0.056+0.015−0.09 ). Again, table 6.1 lists the relevant input data.

Figure 10 shows the η and η′ values consistent with all three semileptonic decay mode

constraints taken at the 2-σ level. Note that points within the two regions are correlated.

η‘

η‘Re    , η

η‘Im    ,η

η

−0.02

−0.015

−0.01

−0.005

 0

 0.005

 0.01

 0.015

 0.02

−0.03 −0.025 −0.02 −0.015 −0.01 −0.005  0  0.005  0.01  0.015  0.02

Figure 10: 2-σ allowed ranges of η and η′.

6.2 Range of SφK
As for the semi-leptonic modes, the non-leptonic B → φKs decay amplitude can be written in
terms of an effective Hamiltonian. In the SM, this is given by [57]

HSMeff =
GF√

2

[
λu(C1Q

u
1 + C2Q

u
2 ) + λc(C1Q

c
1 + C2Q

c
2) − λt

10∑
i=3

CiQi

]
, (57)

24



where λq = V
qs∗
CKMV

qb
CKM are products of CKM factors, Qi(µ) are four-fermion operators, Ci(µ)

are the Wilson coefficients, and µ is the renormalization scale.4 The operators Q1,Q2, Q3, . . .Q6,

and Q7 . . .Q10 are the usual SM current-current, QCD penguin, and electroweak (EW) penguin

operators respectively, as defined in [57]. Note that the four-quark operators Q9 and Q10 are

different from the semi-leptonic operators Q9V and Q10A considered above.

If we include the FCNC couplings from the mirror quarks, the tree-level contribution to

Heff at scale MW is

∆Heff = +
GF√

2
λt

[
η(s̄b)(V −A)

∑
q

g
q
R,L(q̄q)(V ±A) + η

′(s̄b)(V +A)
∑
q

g
q
R,L(q̄q)(V ±A)

]
(58)

where the sum runs over q = u,d,s,c,b, g
q
R,L = (T3 − eq sin2 θw) is the Z(q̄q)L,R coupling, and

η =
V sb

d

λt
and η′ = −Ṽ sbd

λt
are the same as above. The first operator, multiplied by η, can be

written as a linear combination of Q3, . . . ,Q10. The second operator, multiplied by η
′, has

no SM counterpart. We introduce a new “(V + A)” operator basis Q′1, . . . ,Q
′
10 related to

Q1, . . . ,Q10 by the interchange (V − A) ↔ (V + A) wherever these appear.
We incorporate the new (V −A) operator contribution by modifying the Wilson coefficients

at scale MW . The changes are

CSM3 (MW ) → CSM3 (MW ) +
1

6
η, (59)

CSM7 (MW ) → CSM7 (MW ) +
2

3
sin2 θwη,

CSM9 (MW ) → CSM9 (MW ) −
2

3
(1 − sin2 θw)η.

For the (V + A) operators,
(
CSMi

)′
(MW ) = 0, while the FCNC contribution gives

C′5(MW ) =
1

6
η′, (60)

C′7(MW ) = −
2

3
(1 − sin2 θw)η′,

C′9(MW ) =
2

3
sin2 θwη

′,

with all others zero.

The RG evolution of these operators proceeds much like in the SM since the (V − A) and
(V + A) operators evolve independently. The anomalous dimension matrix that determines

4In writing Hef f in this form, we have made use of λu + λc + λt ' 1. This is also approximately true when
mirror quarks are included.

25



running of both the (V −A) and the (V + A) operators is the same as in the SM. This follows
from our definition of the (V +A) operators, and the fact that these are renormalized by parity-

invariant gauge interactions. We calculated the Wilson coefficients at scale µ = 2.5, 5.0 GeV at

one-loop order in both the QCD and QED corrections using the results of [58]. To this order,

the corresponding initial values of the Wilson coefficients are taken at tree-level in the QCD

corrections, although we have included the one-loop electroweak corrections which give a large

contribution to C9(MZ). (This agrees with the conventions of [57, 58, 59].)

Hadronic matrix elements for the B → φKs transition at scale µ = 2.5, 5.0 GeV were
estimated using factorization. Following [60, 61], the amplitude is given by

A(B̄ → φK̄) = A0λt
[
a3 + a4 + a5 −

1

2
(a7 + a9 + a10) −

1

2
(a′7 + a

′
9 + a

′
10)

]
(61)

where A0 = −
√

2GFfφmφF
BK
1 (m

2
φ)(�

∗ · pK) is a CP-invariant product of form factors and
constants, and the ai are functions of the Wilson coefficients. To leading order, they are [60]:

a2i−1 = C2i−1 +
1

Nef f
C2i, and a2i = C2i +

1
Nef f

C2i−1, where Neff is an effective number of colours.

In writing (61), we have neglected annihilation contributions which may be significant [61].

The ai coefficients were calculated numerically by taking as input the 2-σ allowed values of

η and η′ from the previous section, and running the Wilson coefficients down to µ = 2.5, 5.0

GeV. From these we calculate SφK, and the B → φKs branching ratio. This helps to reduce
the theoretical uncertainty due to the sensitivity of the amplitude to variations of Neff . Using

(61) and the input parameters fφ = 0.233 GeV, F
BK
1 = 0.39 ± 0.03, mφ = 1019.4 MeV,

mK = 497.7 MeV, mB = 5279.3 MeV, τB0 = 1.54 ± 0.02 ps, |λt| = 0.040 ± 0.003, the branching
ratio is

B(B0 → φK0s ) = (5.13 ± 0.15) × 10−3
∣∣∣∣a3 + a4 + a5 − 12(a7 + a9 + a10) −

1

2
(a′7 + a

′
9 + a

′
10)

∣∣∣∣
2

.

(62)

The range of SφK obtained for Neff = 2, 3, . . . 10 is shown in Figure 11, and is plotted against
the branching ratio. CLEO, BABAR, and BELLE have recently measured this branching

ratio[62, 63], and the average of their results is B(B0 → φK0) = (8.7 ± 1.3) × 10−6.
We find that a range 0.2 . SφK . 1.0 can be explained by FCNC’s in this model while

simultaneously accommodating semi-leptonic B decay and B → φK branching fraction data at
the 2-σ level. While the shift in SφK from FCNC’s is not large enough to completely explain
the current experimental value, it is still significant, and reduces the discrepancy to below 2-σ.

A strong phase, δ, in the new physics relative to the SM would only decrease the range of

26



S φΚS φΚ

µ = 5.0 GeV

µ = 2.5 GeV

B (10    )
−6

−1

−0.5

0

0.5

1

0 2 4 6 8 10 12 14 16

Figure 11: Range of SφK accessible by the model plotted against the branching ratio for two
choices of the renormalization scale. The vertical dotted lines indicate the 2-σ allowed region
for the branching ratio.

SφK [63]. Setting δ = π gives a result similar to that displayed in Fig. 11. The result shows
a strong dependence on the renormalization scale, µ, as well as the effective number colours,

Neff due to sensitive cancellations between terms in the amplitude. While this situation would

be improved by adding higher order corrections, we do not expect such terms to change these

general conclusions.

The above result is more constraining than the one obtained in Ref. [64], in which the

effect of a vector-like pair of singlet down quarks on SφK was considered. As we have done
here, these authors investigate the range of SφK that can be obtained from the sZb vertex
that is induced by vector-like down quarks. Our results should reduce to theirs in the limit

η′ → 0, which corresponds to considering only the flavour mixing effects due to the singlets.
By the same reasoning, the range of SφK should be greater in the present model since even
more flavour mixing is possible. Instead, these authors find a larger range for SφK than we
have obtained. One of the possible sources of discrepancy between our analysis and the one

in Ref. [64] resides in our use of a more stringent quantitative analysis of the constraints

27



coming from the semileptonic B-decays. Another source appears to be their inclusion of the

“colour-suppressed” operators (s̄βbα)(V −A)(s̄αsβ)(V ±A) at tree-level. While such operators do

arise from QCD corrections to the sZb vertex, all tree-level effects may be described by the

“colour allowed” operators (s̄βbβ)(V −A)(s̄αsα)(V ±A).

7 Conclusion

We have investigated the phenomenological properties of Beautiful mirrors, an extension of the

Standard Model consisting of additional vector-like “mirror” quarks with the same quantum

numbers as the SU(2) quark doublet and down quark singlet in the Standard Model. These

exotic quarks mix with the bottom quark resulting in a modified value of the right-handed

bottom quark coupling to the Z gauge boson, in agreement with indications coming from the

precision electroweak data. A good fit to the precision electroweak data also demands that the

additional quarks have masses lower than about 300 GeV implying a rich phenomenology at

the Tevatron and LHC Colliders, as well as a possible impact on the CP-violating observables

measured at the B-factories. In addition, the unification of gauge couplings is greatly improved

within the model.

In this article we have provided a detailed analysis of the question of gauge coupling unifi-

cation. We find that the gauge couplings unify at MG = (2.80±0.15)×1016 GeV. Perturbative
consistency and stability of the model restrict the possible values of the masses of the Higgs

and the mirror quarks. The allowed range, mh = 170 ± 10 GeV, overlaps with the range of
values of these parameters which give the best fit to precision electroweak data [6].

Flavour mixing due to the mirror quarks leads to right-handed Z couplings, a very small loss

of unitarity of the CKM matrix, and FCNCs. The flavour mixing also modifies the coupling

of the b and ω quarks to the Higgs, while the couplings of the other quarks to the Higgs are

not changed significantly. This has some interesting implications for Higgs searches at the

Tevatron. In particular, the required luminosity for a Tevatron or LHC Higgs discovery in the

WW decay mode, as well as in the τ+τ− mode at the Tevatron and the γγ mode at the LHC, is

greatly reduced within this model. We have analyzed the search for mirror vector quarks at the

Tevatron collider, and have found that Run II with a total integrated luminosity of about 4 fb−1

will be able to test all of the mirror quark mass range consistent with electroweak precision

data. Finally, the b → s FCNCs which arise in this model can help explain the discrepancy
between the values of sin(2β) measured in the B → φK and B → J/ψK decays.

28



Acknowledgements The authors would like to thank M. Carena, Z. Chacko, C.W. Chiang,

D. Choudhury, D.E. Kaplan. T. Le Compte, R. Sundrum and T.M.P. Tait for interesting

comments and suggestions. Work supported in part by the US DOE, Div. of HEP, Contract

W-31-109-ENG-38.

Note added : As this work was being completed, Ref. [65] appeared, in which the effect of

tree-level (V ± A) sZb couplings on SφK was calculated. The modifications to the electroweak
scale Wilson coefficients as well as the final result in this paper agree with our analysis.

References

[1] LEP Electroweak Working Group, LEPEWWG/2003-01.

[2] H. E. Haber and H. E. Logan, Phys. Rev. D 62, 015011 (2000) [hep-ph/9909335].

[3] M. S. Chanowitz, Phys. Rev. Lett. 87, 231802 (2001) [arXiv:hep-ph/0104024].

[4] G. Altarelli, F. Caravaglios, G. F. Giudice, P. Gambino and G. Ridolfi, JHEP 0106, 018

(2001) [arXiv:hep-ph/0106029].

[5] M. Carena, E. Ponton, T. M. Tait and C. E. M. Wagner, arXiv:hep-ph/0212307.

[6] D. Choudhury, T. M. Tait and C. E. Wagner, Phys. Rev. D 65, 053002 (2002) [arXiv:hep-

ph/0109097].

[7] L. Randall and R. Sundrum, Phys. Rev. Lett. 83, 3370 (1999) [arXiv:hep-ph/9905221].

[8] A. Pomarol, Phys. Rev. Lett. 85, 4004 (2000) [arXiv:hep-ph/0005293]. L. Randall and

M. D. Schwartz, JHEP 0111, 003 (2001) [arXiv:hep-th/0108114]; Phys. Rev. Lett. 88,

081801 (2002) [arXiv:hep-th/0108115]; K. w. Choi, H. D. Kim and Y. W. Kim, JHEP

0211, 033 (2002) [arXiv:hep-ph/0202257]; arXiv:hep-ph/0207013; K. Agashe, A. Del-

gado and R. Sundrum, Nucl. Phys. B 643, 172 (2002) [arXiv:hep-ph/0206099]; R. Con-

tino, P. Creminelli and E. Trincherini, JHEP 0210, 029 (2002) [arXiv:hep-th/0208002];

A. Falkowski and H. D. Kim, JHEP 0208, 052 (2002) [arXiv:hep-ph/0208058]; K. w. Choi

and I. W. Kim, Phys. Rev. D 67, 045005 (2003) [arXiv:hep-th/0208071]; L. Ran-

dall, Y. Shadmi and N. Weiner, JHEP 0301, 055 (2003) [arXiv:hep-th/0208120];

29



A. Lewandowski, M. J. May and R. Sundrum, Phys. Rev. D 67, 024036 (2003) [arXiv:hep-

th/0209050]; K. Agashe and A. Delgado, Phys. Rev. D 67, 046003 (2003) [arXiv:hep-

th/0209212].

[9] W. D. Goldberger and I. Z. Rothstein, Phys. Rev. Lett. 89, 131601 (2002) [arXiv:hep-

th/0204160]; arXiv:hep-th/0208060; arXiv:hep-ph/0303158.

[10] K. Agashe, A. Delgado and R. Sundrum, Annals Phys. 304, 145 (2003) [arXiv:hep-

ph/0212028].

[11] M. Carena, A. Delgado, E. Ponton, T. M. Tait and C. E. Wagner, arXiv:hep-ph/0305188.

[12] H. Arason, D. J. Castano, B. Keszthelyi, S. Mikaelian, E. J. Piard, P. Ramond and

B. D. Wright, Phys. Rev. D 46, 3945 (1992).

[13] M. E. Machacek and M. T. Vaughn, Nucl. Phys. B 222, 83 (1983).

[14] M. E. Machacek and M. T. Vaughn, Nucl. Phys. B 236, 221 (1984).

[15] M. E. Machacek and M. T. Vaughn, Nucl. Phys. B 249, 70 (1985).

[16] G. Cvetic, C. S. Kim and S. S. Hwang, Phys. Rev. D 58, 116003 (1998) [arXiv:hep-

ph/9806282].

[17] K. Hagiwara et al. [Particle Data Group Collaboration], Phys. Rev. D 66, 010001 (2002).

[18] L. J. Hall, Nucl. Phys. B 178, 75 (1981);

P. Langacker and N. Polonsky, Phys. Rev. D 47, 4028 (1993) [arXiv:hep-ph/9210235].

M. Carena, S. Pokorski and C. E. Wagner, Nucl. Phys. B 406, 59 (1993) [arXiv:hep-

ph/9303202].

[19] P. H. Frampton, P. Q. Hung and M. Sher, Phys. Rept. 330, 263 (2000) [arXiv:hep-

ph/9903387].

[20] Y. Suzuki et al. [TITAND Working Group Collaboration], arXiv:hep-ex/0110005.

[21] See, for example, P. Nath and R. Arnowitt, Phys. Rev. Lett. 70, 3696 (1993) [arXiv:hep-

ph/9302318]; R. Dermisek, A. Mafi and S. Raby, Phys. Rev. D 63, 035001 (2001)

[arXiv:hep-ph/0007213]; H. Murayama and A. Pierce, Phys. Rev. D 65, 055009 (2002)

[arXiv:hep-ph/0108104]; and references therein.

30



[22] S. Raby, arXiv:hep-ph/0211024.

[23] D. Emmanuel-Costa and S. Wiesenfeldt, arXiv:hep-ph/0302272; P. Langacker, Phys.

Rept. 72, 185 (1981).

[24] S. Aoki et al. [JLQCD Collaboration], Phys. Rev. D 62, 014506 (2000) [arXiv:hep-

lat/9911026].

[25] F. Wilczek and A. Zee, Phys. Rev. Lett. 43, 1571 (1979).

[26] F. Larios, M. A. Perez and C. P. Yuan, Phys. Lett. B 457, 334 (1999) [hep-ph/9903394].

[27] J.F. Gunion, H.E. Haber, G. Kane, S. Dawson, The Higgs Hunter’s Guide, Perseus Books,

1990. J. F. Gunion, H. E. Haber, G. L. Kane and S. Dawson, SCIPP-89/13

[28] A. Djouadi, M. Spira and P. M. Zerwas, Phys. Lett. B 264, 440 (1991).

[29] D. Rainwater, M. Spira and D. Zeppenfeld, arXiv:hep-ph/0203187;

[30] M. Carena and H. E. Haber, arXiv:hep-ph/0208209.

[31] A. Belyaev, T. Han and R. Rosenfeld, arXiv:hep-ph/0204210.

[32] M. Carena et al. [Higgs Working Group Collaboration], arXiv:hep-ph/0010338.

[33] M. Carena, J. R. Ellis, S. Mrenna, A. Pilaftsis and C. E. Wagner, arXiv:hep-ph/0211467.

[34] K.Jakobs, T. Trefzger, ATL-PHYS-2000-015

[35] T. Affolder et al. [CDF Collaboration], Phys. Rev. Lett. 84, 835 (2000) [arXiv:hep-

ex/9909027].

CDF Collaboration, T.Affolder et al., Phys. Rev. Lett. 84, 835 (2000) [hep-ex/9909027].

[36] E. L. Berger and H. Contopanagos, Phys. Rev. D 57, 253 (1998) [arXiv:hep-ph/9706206].

[37] N. Kidonakis, in Proc. of the APS/DPF/DPB Summer Study on the Future of Parti-

cle Physics (Snowmass 2001) ed. N. Graf, eConf C010630, P505 (2001) [arXiv:hep-

ph/0110145].

[38] K. Sliwa [CDF Collaboration], Acta Phys. Polon. B 33, 3861 (2002).

31



[39] P. Savard [CDF and D0 collaborations], eConf C020620, SABT05 (2002) [arXiv:hep-

ex/0209061].

[40] T. Affolder et al. [CDF Collaboration], Phys. Rev. Lett. 84, 216 (2000) [arXiv:hep-

ex/9909042].

[41] G. L. Kane, G. A. Ladinsky and C. P. Yuan, Phys. Rev. D 45, 124 (1992).

[42] M. Sher, Phys. Rev. D 61, 057303 (2000) [arXiv:hep-ph/9908238]; M. Sher, Phys. Rev.

D61, 057303 (2000) [hep-ph/9908238].

[43] I. I. Bigi, Y. L. Dokshitzer, V. A. Khoze, J. H. Kuhn and P. M. Zerwas, Phys. Lett. B

181, 157 (1986).

[44] G. Hiller, Phys. Rev. D 66, 071502 (2002) [arXiv:hep-ph/0207356].

[45] K. Abe et al. [Belle Collaboration], arXiv:hep-ex/0207098.

[46] B. Aubert et al. [BABAR Collaboration], arXiv:hep-ex/0207070.

[47] R. Fleischer, Phys. Rept. 370, 537 (2002) [arXiv:hep-ph/0207108].

R. Fleischer and T. Mannel, Phys. Lett. B 511, 240 (2001) [arXiv:hep-ph/0103121].

[48] G. Barenboim, F. J. Botella and O. Vives, Phys. Rev. D 64, 015007 (2001) [arXiv:hep-

ph/0012197];

J. A. Aguilar-Saavedra, Phys. Rev. D 67, 035003 (2003) [arXiv:hep-ph/0210112].

[49] A. J. Buras and R. Fleischer, Adv. Ser. Direct. High Energy Phys. 15, 65 (1998)

[arXiv:hep-ph/9704376].

[50] T. Yanir, JHEP 0206, 044 (2002) [arXiv:hep-ph/0205073].

[51] G. Buchalla, G. Hiller and G. Isidori, Phys. Rev. D 63, 014015 (2001) [arXiv:hep-

ph/0006136].

[52] A. Ali, P. Ball, L. T. Handoko and G. Hiller, Phys. Rev. D 61, 074024 (2000) [arXiv:hep-

ph/9910221].

32



[53] J. Kaneko et al. [Belle Collaboration], Phys. Rev. Lett. 90, 021801 (2003) [arXiv:hep-

ex/0208029].

[54] A. Ali, E. Lunghi, C. Greub and G. Hiller, Phys. Rev. D 66, 034002 (2002) [arXiv:hep-

ph/0112300].

[55] K. Abe et al. [BELLE Collaboration], Phys. Rev. Lett. 88, 021801 (2002) [arXiv:hep-

ex/0109026].

[56] B. Aubert et al. [BABAR Collaboration], arXiv:hep-ex/0207082.

[57] A. J. Buras, arXiv:hep-ph/9806471.

[58] A. J. Buras, M. Jamin and M. E. Lautenbacher, Nucl. Phys. B 400, 75 (1993) [arXiv:hep-

ph/9211321].

[59] M. Beneke, G. Buchalla, M. Neubert and C. T. Sachrajda, Nucl. Phys. B 606, 245 (2001)

[arXiv:hep-ph/0104110].

[60] A. Ali, G. Kramer and C. D. Lu, Phys. Rev. D 58, 094009 (1998) [arXiv:hep-ph/9804363].

[61] H. Y. Cheng and K. C. Yang, Phys. Rev. D 64, 074004 (2001) [arXiv:hep-ph/0012152].

[62] R. A. Briere et al. [CLEO Collaboration], Phys. Rev. Lett. 86, 3718 (2001) [arXiv:hep-

ex/0101032];

B. Aubert et al. [BABAR Collaboration], arXiv:hep-ex/0303029.

[63] C. W. Chiang and J. L. Rosner, arXiv:hep-ph/0302094.

[64] A. K. Giri and R. Mohanta, arXiv:hep-ph/0306041.

[65] D. Atwood and G. Hiller, arXiv:hep-ph/0307251.

33