Академический Документы
Профессиональный Документы
Культура Документы
iv
TABLE OF CONTENTS
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii
1. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Standard Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Supersymmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.1 Minimal Supersymmetric Standard Model . . . . . . . . . . . . . . 9
1.2.2 Supergravity and the Minimal Supergravity Model . . . . . . . . . 17
1.3 SO(10) SUSY Grand Unication . . . . . . . . . . . . . . . . . . . . . . 18
2. Yukawa-unied SO(10) SUSY GUTs . . . . . . . . . . . . . . . . . . . . . . . 20
2.1 HS v. DT Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.2 Random Scan in HS model . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2.1 Random scan results . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2.2 Three proposals to reconcile Yukawa-unied models with dark mat-
ter relic density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Dark matter solution via neutralino decay to axino . . . . . . . . . . 30
Dark matter solution via non-universal gaugino masses . . . . . . . . 31
Dark matter solution via generational non-universality . . . . . . . . 32
2.3 Discussion of Markov Chain Monte Carlo analysis . . . . . . . . . . . . . 33
2.3.1 HS model: neutralino annihilation via h resonance . . . . . . . . . 35
2.3.2 Solutions using weak scale Higgs boundary conditions . . . . . . . 41
2.4 Yukawa-unied benchmark scenarios and LHC signatures . . . . . . . . . 46
3. SUSY Cosmology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.1 Dark Matter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.2 SO(10) SUSY GUTs and Yukawa unication . . . . . . . . . . . . . . . 52
3.2.1 The gravitino problem . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.2.2 Non-thermal leptogenesis . . . . . . . . . . . . . . . . . . . . . . . 56
3.2.3 Axino dark matter . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.2.4 A consistent cosmology for axino DM from SO(10) SUSY GUTs . 58
4. Collider Searches for New Physics . . . . . . . . . . . . . . . . . . . . . . . . . 60
v
4.1 Early SUSY Discovery using Multi-leptons . . . . . . . . . . . . . . . . . 61
4.2 Yukawa-unied SO(10) at the Cern LHC . . . . . . . . . . . . . . . . . . 71
4.2.1 Cross sections and branching fractions for sparticles in Yukawa-
unied models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.2.2 Gluino pair production signals at the LHC . . . . . . . . . . . . . . 75
4.2.3 Sparticle masses from gluino pair production . . . . . . . . . . . . 83
4.2.4 Trilepton signal from
W
1
0
2
production . . . . . . . . . . . . . . . . 90
5. CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
BIOGRAPHICAL SKETCH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
vi
LIST OF TABLES
1.1 One generation of the MSSM. . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.1 Masses and parameters in GeV units for ve benchmark Yukawa unied points
using Isajet 7.75 and m
t
= 171.0 GeV. The upper entry for the
0
1
h
2
etc.
come from IsaReD/Isatools, while the lower entry comes from micrOMEGAs;
(
0
1
p) is computed with Isatools. . . . . . . . . . . . . . . . . . . . . . . . 48
4.1 Events generated and cross sections for various SM background processes plus
the SPS1a
and E
miss
T
cuts are specied in
the text. The W + jets and Z + jets background has been computed within
the restriction p
T
(W, Z) > 100 GeV. . . . . . . . . . . . . . . . . . . . . . . . 76
4.4 Clean trilepton signal after cuts listed in the text. . . . . . . . . . . . . . . . 92
vii
LIST OF FIGURES
1.1 Evolution of the SU(3)
C
SU(2)
L
U(1)
Y
gauge coupling constants from the
weak scale to the GUT scale for the case of (a) the SM, (b) the MSSM with
two Higgs doublets, and (c) the MSSM with four Higgs doublets [63]. . . . . 15
2.1 Plot of R versus various input parameters for a wide (dark blue) and narrow
(light blue) random scan over the parameter ranges listed in Eq. (2.6). We
take > 0 and m
t
= 171 GeV. . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.2 Plot of R versus various sparticle masses for a random scan over the parameter
range listed in Eq. (2.6). We take > 0 and m
t
= 171 GeV. . . . . . . . . . 28
2.3 Plot of R versus various sparticle masses for a random scan over the parameter
range listed in Eq. (2.6). We take > 0 and m
t
= 171 GeV. . . . . . . . . . 29
2.4 Plot of
0
1
h
2
vs. R for a random scan over the parameter range listed in
Eq. (2.6). We take > 0 and m
t
= 171 GeV. . . . . . . . . . . . . . . . . . 29
2.5 Plot of variation in
0
1
h
2
versus non-universal GUT scale gaugino mass M
1
for benchmark point A in Table 2.1. . . . . . . . . . . . . . . . . . . . . . . . 32
2.6 Plot of variation in
0
1
h
2
versus non-universal GUT scale rst/second gener-
ation scalar mass m
16
(1, 2) for benchmark point C in Table 2.1. . . . . . . . 34
2.7 Plot of MCMC results in the m
16
vs. m
10
plane; the light-blue (dark-blue)
points have R < 1.1 (1.05), while for the orange (red) points R < 1.1 (1.05)
and
0
1
h
2
< 0.136. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.8 Plot of MCMC results in the m
16
vs. A
0
/m
16
plane; the light-blue (dark-blue)
points have R < 1.1 (1.05), while for the orange (red) points R < 1.1 (1.05)
and
0
1
h
2
< 0.136. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.9 Plot of MCMC results in the m
16
vs. m
1/2
plane; the light-blue (dark-blue)
points have R < 1.1 (1.05), while for the orange (red) points R < 1.1 (1.05)
and
0
1
h
2
< 0.136. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
viii
2.10 Plot of MCMC results in the m
16
vs. m
H
d,u
plane; the light-blue (dark-blue)
points have R < 1.1 (1.05), while for the orange (red) points R < 1.1 (1.05)
and
0
1
h
2
< 0.136. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.11 Plot of MCMC results in the m
h
2m
0
1
vs. m
A
2m
0
1
plane; the light-
blue (dark-blue) points have R < 1.1 (1.05), while for the orange (red) points
R < 1.1 (1.05) and
0
1
h
2
< 0.136. . . . . . . . . . . . . . . . . . . . . . . . . 39
2.12 Plot of MCMC results in the R vs.
0
1
h
2
plane; the light-blue (dark-blue)
points have R < 1.1 (1.05), while for the orange (red) points R < 1.1 (1.05)
and
0
1
h
2
< 0.136. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.13 Plot of MCMC results in the m
0
2
m
0
1
vs. m
h
plane; the light-blue
(dark-blue) points have R < 1.1 (1.05), while for the orange (red) points
R < 1.1 (1.05) and
0
1
h
2
< 0.136. . . . . . . . . . . . . . . . . . . . . . . . . 40
2.14 Plot of MCMC results using WSH boundary conditions in the m
16
vs. A
0
/m
16
plane; the light-blue (dark-blue) points have R < 1.1 (1.05), while for the
orange (red) points R < 1.1 (1.05) and
0
1
h
2
< 0.136. . . . . . . . . . . . . . 42
2.15 Plot of MCMC results using WSH boundary conditions in the m
A
vs.
plane; the light-blue (dark-blue) points have R < 1.1 (1.05), while for the
orange (red) points R < 1.1 (1.05) and
0
1
h
2
< 0.136. . . . . . . . . . . . . . 43
2.16 Plot of MCMC results using WSH boundary conditions in the m
h
2m
0
1
vs.
m
A
2m
0
1
plane; the light-blue (dark-blue) points have R < 1.1 (1.05), while
for the orange (red) points R < 1.1 (1.05) and
0
1
h
2
< 0.136. . . . . . . . . 44
2.17 Plot of MCMC results using WSH boundary conditions in the m
h
vs.
BF(B
s
+
TP
a
h
2
= 0.08 and f
a
/N = 5 10
11
GeV. . . . . . . . . . . . . . . . . . . . . . 59
ix
4.1 Plot of jet multiplicity from SUSY collider events from SPS1a
after cuts C1
.
We also plot the histograms of various SM backgrounds, plus the total SM
background (gray histogram). . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.2 Plot of augmented eective mass A
T
(without E
miss
T
) from SUSY collider events
from SPS1a
after cuts C1
after cuts
C1
after cuts C1
and n
3, for m
0
= 200 and 1000 GeV. We also take A
0
= 0, tan = 10,
> 0 and m
t
= 171 GeV. We also plot the 5 background level for 0.1 and 1
fb
1
of integrated luminosity. . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.6 Distribution in variable A
T
from SUSY events from SPS1a
after cuts C1
plus
3 plus 1 b-jet. We also plot the remaining SM backgrounds (gray
histogram). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.7 Plot of OS/SF dilepton invariant mass from SUSY events from SPS1a
after
cuts C1
after cuts C1
s = 14 TeV versus m
g
. We use Prospino with
scale choice Q = m
g
, and show LO (solid) and NLO (dashes) predictions in
the vicinity of point A (red) and point D (blue) from Table 4.2. . . . . . . . 74
4.10 Plot of various -ino pair production processes in fb at
s = 14 TeV versus
m
1
, for m
q
= 3 TeV and = m
g
, with tan = 49 and > 0. . . . . . . . 74
4.11 Plot of various sparticle branching fractions taken from Isajet for points A
and D from Table 4.2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4.12 Plot of distribution in transverse sphericity S
T
for events with cuts C1
, from benchmark
points A (full red line) and D (dashed blue line). . . . . . . . . . . . . . . . 79
4.15 Plot of jet multiplicity from benchmark points A (full red line) and D (dashed
blue line) after cuts C1
from benchmark
points A (full red line) and D (dashed blue line) along with SM backgrounds. 84
4.20 Same as Fig. 4.19 but for same-avor minus dierent-avor subtracted
invariant-mass. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.21 Plot of m(X
1,2
) from benchmark points A and D along with SM backgrounds
in events with cuts C1
)
min
from points A and D along with SM backgrounds. . . 89
4.24 Plot of m(X
1,2
)
min
from points A and D, minimizing m(X
1
X
2
) as
explained in the text, along with SM backgrounds. . . . . . . . . . . . . . . 89
4.25 Plot of m(
+
1
0
2
3
channel, and should exhibit the same dilepton mass edge as in the gluino cascade decay
signal.
A problem generic to all supergravity models comes from overproduction of gravitinos
in the early universe: if gravitinos are unstable, then their late decays may destroy the
predictions of Big Bang nucleosynthesis. We also present a Yukawa-unied SO(10) SUSY
GUT scenario which avoids the gravitino problem, gives rise to the correct matter-antimatter
asymmetry via non-thermal leptogenesis, and is consistent with the WMAP-measured
xii
abundance of cold dark matter due to the presence of an axino LSP. To maintain a consistent
cosmology for Yukawa-unied SUSY models, we require a re-heat temperature T
R
10
6
10
7
GeV, an axino mass around 0.1 10 MeV, and a Peccei-Quinn breaking scale f
a
10
12
GeV.
xiii
CHAPTER 1
INTRODUCTION
In order to understand the feasibility of SO(10) Supersymmetric Grand Unied Theories
(SUSY GUTs), a foundation must rst be laid. For our purposes, this foundation is built
upon the Standard Model (SM), the Minimal Supersymmetric Standard Model (MSSM), and
nally the gauge group SO(10). The theoretical basis must also be reconciled with known
cosmological and experimental constraints, such as Dark Matter (DM) relic abundance and
LEP2 ndings. This chapter covers the rudiments needed to justify the models covered in
the remainder of this dissertation.
1.1 Standard Model
Formulated in the 1970s, the Standard Model of particle physics has been undeniably
successful in describing and predicting the properties of matter. The gauge group of the SM
is SU(3)
C
SU(2)
L
U(1)
Y
, where the component symmetry groups represent strong and
electroweak interactions. The SU(2)
L
U(1)
Y
group unies the weak and electromagnetic
interactions; upon spontaneous symmetry breaking via the Higgs mechanism, it breaks down
to just U(1)
em
. Thus, the SM can be completely characterized by quantum chromodynamics
(QCD) and electroweak theory (EW). It is comprised of 19 free parameters: three lepton
masses, six quark masses, three Cabibbo-Kobayashi-Maskawa (CKM) mixing angles plus
one CP-violating phase , three gauge couplings, the QCD vacuum angle
QCD
, the Higgs
quadratic coupling , and the Higgs self-coupling strength .
The matter constituents of the SM are three generations of leptons and quarks with
1
left-handed SU(2) doublets and right-handed singlets:
l
1,L
=
_
e
e
_
L
, e
R
; q
1,L
=
_
u
d
_
L
, u
R
, d
R
;
l
2,L
=
_
_
L
,
R
; q
2,L
=
_
c
s
_
L
, c
R
, s
R
;
l
3,L
=
_
_
L
,
R
; q
3,L
=
_
t
b
_
L
, t
R
, b
R
.
The Higgs doublet eld
=
_
+
0
_
acquires a non-zero vacuum expectation value (vev) thereby giving mass to the fermions
1
and
gauge bosons. The force carriers (gauge bosons) complete the particle content of the SM:
gluons g (in 8 colors, massless, spin-0, mediate the strong force), W
and Z
0
(massive, spin-
1, mediate the weak force), and the photon (massless, spin-1, mediates the electromagnetic
force).
The SM Langrangian has the form
L
SM
= L
QCD
+L
EW
= L
QCD
+L
gauge
+L
scalar
+L
fermion
+L
Yukawa
. (1.1)
Starting with the QCD term, the components of this Lagrangian are as follows:
L
QCD
=
1
4
F
i
F
i
+
r
q
r
i , D
r
, (1.2)
where
F
i
G
i
G
i
g
s
f
ijk
G
j
G
k
(1.3)
is the eld strength tensor for the gluon elds G
i
(i = 1, ..., 8), g
s
is the QCD gauge coupling
constant, and the structure constants f
ijk
(i, j, k = 1, ..., 8) are dened by the commutators
of the Gell-Mann matrices as
_
i
2
,
j
2
_
= if
ijk
k
2
. (1.4)
The gauge covariant derivative in the second term of the QCD Lagrangian is
D
a
b
= (D
)
ab
=
a
b
+ig
s
G
i
T
ia
b
, (1.5)
1
However, the Higgs is only partially responsible for masses of neutrinos which are thought to get their
masses from the see-saw mechanism. Here, neutrinos are classied as Majorana particles with the light
left-handed neutrinos having heavy right-handed counterparts.
2
where the quarks transform according to the 3 3 representation matrices T
i
=
i
/2, q
r
indicates the r
th
quark avor, and , = 1, 2, 3 are color indices.
Secondly,
L
gauge
=
1
4
F
i
F
i
1
4
B
, (1.6)
with eld strength tensors
B
and
F
W
i
W
i
g
ijk
W
j
W
k
, (1.7)
where W
i
(i = 1, 2, 3) and B
) is the
SU(2) (U(1)) gauge coupling, and
ijk
is the totally antisymmetric Levi-Civita symbol.
The third term is for the Higgs
L
scalar
L
= (D
V () , (1.8)
where is the complex Higgs scalar already shown earlier, the gauge covariant derviative is
D
=
_
+ig
i
2
W
i
+
ig
2
B
_
, (1.9)
the
i
are the Pauli matrices, the Higgs potential V () has the form
V () = +
2
+(
)
2
. (1.10)
It is for
2
< 0 that spontaneous symmetry breaking occurs giving rise to the SM particle
masses.
Next is the fermionic term
L
fermion
L
f
=
3 generations
m=1
( q
mL
i , Dq
mL
+
l
mL
i , Dl
mL
+ u
mR
i , Du
mR
+
d
mR
i , Dd
mR
+ e
mR
i , De
mR
),
(1.11)
where m is generation or family index, and the L and R refer to the left and right chiral
projections of the elds
L
(1
5
)/2 and
R
(1 +
5
)/2.
Finally, we have the Yukawa term
L
Yukawa
=
3 generations
m,n=1
( q
mL
u
mn
u
nR
+ q
mL
d
mn
d
nR
+
l
mL
e
mn
e
nR
) +h.c., (1.12)
3
where the matrices
mn
describe the Yukawa couplings between the Higgs doublet and the
various avors m and n of quarks and leptons.
The SM is comprised of 19 free parameters: three lepton masses, six quark masses, three
Cabibbo-Kobayashi-Maskawa (CKM) mixing angles plus one CP-violating phase , three
gauge couplings, the QCD vacuum angle
QCD
, the Higgs quadratic coupling , and the
Higgs self-coupling strength .
Although a host of electroweak precision measurement tests such as the predicted
existence and form of the weak neutral current and the existence and masses of the W and
Z bosons have soundly reinforced the SM, other cosmological and theoretical observations
have not been resolved within the SM. Some remaining issues are how particles attain mass,
the presence of only three generations of quarks and leptons and their mass heirarchy, the
imbalance of matter-antimatter in the universe (a.k.a., the baryogenesis problem), neutrino
masses and mixing, cold dark matter (CDM), dark energy, and gravity (the postulated force
carrier, the graviton, a massless, spin-2 particle has not found a place in the SM). Also, scalar
masses are quadratically divergent in the SM giving rise to what is known as the ne-tuning
problem: at scales above the weak scale, the Higgs mass parameter may require a great deal
of ne tuning to provide the needed cancellation that will maintain a physical Higgs below
its unitarity limit.
Generally, the SM is accepted as an eective eld theory applicable to energies up to the
weak scale. Higher than this scale, there are currently a few mainstream theories, e.g., Little
Higgs Models, Universal Extra Dimensions (LED), Technicolor, and Supersymmetry. This
dissertation focuses on the last of these beyond-the-standard-model (BSM) theories.
1.2 Supersymmetry
Supersymmetry (SUSY) was conceived in the late 1960s and early 70s; it is the unique
symmetry that relates the properties of bosons to those of fermions. Supersymmetry requires
that for every boson, a fermion partner should exist, and vice versa. These supersymmetric
partners (or sparticles) serve as the new perturbatively coupled degrees of freedom that act
to cancel the quadratic divergences of the SM. There is a plethora of motivations for studying
SUSY, among them are
aesthetics of building a super-Poincare extension of the Poincare group,
4
stability of the scalar potential under radiative corrections (ultra-violet completeness)
making SUSY GUTs natural and extrapolation even to the Planck scale possible,
gravity is contained within the theory,
SUSY is an essential ingredient of superstring theories,
unication of gauge couplings,
suitable candidates for cold dark matter (CDM) are contained within SUSY,
can explain radiative breakdown of electroweak symmetry,
and provides better constraints for the light or SM Higgs boson mass.
Some of these motivations will be discussed in the following sections.
For a brief introduction to SUSY, we follow the Wess-Zumino (WZ) toy model formulated
in 1974. In this model, the Lagrangian takes the form
L = L
kinetic
+L
mass
,
=
1
2
(
A)
2
+
1
2
(
B)
2
+
i
2
, +
1
2
_
F
2
+G
2
_
m
_
1
2
GA FB
_
, (1.13)
where A and B are real scalar elds with dimensionality [A] = [B] = 1, is a Majorana
spinor with =
c
= C
T
and [] =
3
2
, and F and G are auxiliary (non-propagating) elds
with dimension [F] = [G] = 2. They can be eliminated by the Euler-Lagrange equations
F = mB and G = mA.
The spinorial eld expansions are:
D
(x) =
_
d
3
k
(2)
3
1
2E
k
s
_
c
ks
u
ks
e
ikx
+d
ks
v
ks
e
ikx
_
c
D
(x) =
_
d
3
k
(2)
3
1
2E
k
s
_
c
ks
v
ks
e
ikx
+d
ks
u
ks
e
ikx
_
M
(x) =
_
d
3
k
(2)
3
1
2E
k
s
_
c
ks
u
ks
e
ikx
+c
ks
v
ks
e
ikx
_
.
5
The innitesimal SUSY eld transformations in the WZ model go as
A = i
5
,
B = ,
= F +iG
5
+ ,
5
A i ,B,
F = i ,,
F = i
5
,,
with A A + A, etc. and where is a spacetime-independent anticommuting Majorana
spinor parameter with dimension [] = 1/2. Using Majorana bilinear re-arrangements
(e.g.,
= ) along with other algebraic manipulations, we can show that L L + L
with
L
kinetic
=
1
2
,B +
i
2
5
,A +
i
2
F
+
1
2
G
5
_
L
mass
=
(mA
5
+imB
) .
Since the Langrangian changes by a total derivative, the action is invariant, and thus a WZ
transformation is a symmetry of the action. It can also be shown that the action remains
invariant with the addition of interaction terms and that quadratic divergences all cancel
within this model.
The most general supersymmetry algebra includes anticommutators, and so is referred
to as a graded Lie algebra. Only theories with a single Majorana spinorial generator Q
a
,
known as N = 1 supersymmetry theories, allow chiral representations. Models with more
than one SUSY charge in the low energy theory do not lead to chiral fermions and so are
excluded for phenomenological reasons. Using the fact that the super-charges Q
a
are spin-
1
2
objects, a supersymmetric extension of the Poincare algebra or super-Poincare algebra can
6
be written as
[P
, P
] = 0, (1.14)
[M
, P
] = i(g
), (1.15)
[M
, M
] = i (g
gM
+g
) , (1.16)
[P
, Q
a
] = 0, (1.17)
[M
, Q
a
] =
_
1
2
_
ab
Q
b
, (1.18)
_
Q
a
,
Q
b
_
= 2 (
)
ab
P
, (1.19)
Q
a
, Q
b
= 2 (
C)
ab
P
, (1.20)
_
Q
a
,
Q
b
_
= 2
_
C
1
_
ab
P
. (1.21)
Since Q is a Majorana spinor charge, the last two anticommutators can be found from the
rst anticommutation relation Eq. (1.19).
In order to combine scalar and spinor elds into a single object, we move to superspace
x
(x
,
a
), where
a
(a = 1 4) are four anticommuting dimensions arranged as a
Majorana spinor. Superelds provide us with a convenient procedure of formulating theories
that are guaranteed to be supersymmetric and will help us in our ultimate goal of writing
down the simplest supersymmetric extension of the Standard Model. A supereld results
from combining all three members of an irreducible supermultiplet into a single entity. We
denote the three components o,
L
, and T and can be written in terms of complex elds
o =
1
2
(A +iB) ,
L
=
1
5
2
, (1.22)
T =
1
2
(F +iG) .
Since only one chiral component of the Majorana spinor enters the transformations,
such superelds are referred to as (left) chiral superelds. This is fully a left-chiral scalar
supereld, because the lowest spin component of the multiplet has spin zero.
A general supereld has the form
(x, ) = o i
5
i
2
_
_
/+
1
2
_
_
^ +
1
2
_
_
V
+i
_
_
_
_
+
i
2
,
__
1
4
_
_
2
_
T
1
2
o
_
. (1.23)
7
Left and right chiral scalar superelds appear as
o
L
( x, ) = o( x) +i
L
( x) +i
L
T( x), (1.24)
o
R
( x, ) = o( x
) +
R
( x
) i
R
T( x
), (1.25)
where x
= x
+
i
2
A
=
1
2
_
_
V
A
+i
1
4
_
_
2
T
A
(in WZ gauge) or (1.26)
W
A
( x, ) =
LA
( x) +
1
2
F
A
( x)
L
i
L
(, D
R
)
A
iT
A
( x)
L
, (1.27)
allows one to write a master formula for supersymmetric gauge theories as shown below in
Eq. (1.28).
L =
i
(D
o
i
)
(D
o
i
) +
i
2
i
, D
i
+
,A
_
i
2
,A
(, D)
,A
1
4
F
A
F
A
_
i,,A
_
o
i
g
t
A
A
1
5
2
i
+ h.c.
_
(1.28)
1
2
,A
_
i
o
i
g
t
A
o
i
+
A
_
2
o
i
S=S
1
2
i,j
i
_
_
_
2
f
o
i
o
j
_
S=S
1
5
2
+
_
2
f
o
i
o
j
_
S=S
1 +
5
2
_
_
j
,
where the covariant derivatives are given by,
D
o =
o +i
,A
g
t
A
V
A
o, (1.29)
D
+i
,A
g
(t
A
V
A
)
L
i
,A
g
(t
A
V
A
)
R
, (1.30)
(, D)
A
= ,
,A
+ig
_
t
B
,V
B
_
AC
C
, (1.31)
F
A
=
V
A
V
A
g
f
ABC
V
B
V
C
. (1.32)
8
Global SUSY may be spontaneously broken
0[T
i
[0) , = 0 (Ftype breaking), or
0[T
A
[0) , = 0 (Dtype breaking),
as well as explicitly broken by adding soft SUSY breaking (SSB) terms to the Langranian.
In the absence of knowledge about SUSY breaking dynamics, the best that we can do is to
parameterize the eects of SUSY breaking by adding to the Lagrangian all possible SUSY
breaking terms, consistent with all desired (unbroken) symmetries at the SUSY breaking
scale that do not lead to the re-appearance of quadratic divergences, i.e., softly break the
symmetries. Girardello and Grisaru have classied the forms of the soft breaking operators
in a general theory [1]. They have shown that to all orders in perturbation theory, the
following break supersymmetry softly:
linear terms in the scalar eld o
i
(relevant only for singlets of all symmetries),
scalar masses o
i
m
2
ij
o
j
,
and bilinear or trilinear operators of the form o
i
o
j
or o
i
o
j
o
k
(where
o
i
o
j
and
o
i
o
j
o
k
occur in the superpotential),
and nally, in gauge theories, gaugino masses, one for each factor of the gauge group.
1.2.1 Minimal Supersymmetric Standard Model
We are now ready to build a supersymmetric model. It is desirable to build upon that which
we already believe to describe nature at the weak scale, so a supersymmetric version of the
Standard Model would best serve our purposes. The simplest such model is known as the
Minimal Supersymmetric Standard Model (MSSM). It is a direct supersymmetrization of
the SM (except for the fact that one needs to introduce two Higgs doublet elds) and is
minimal in the sense that it contains the smallest number of new particle states and new
interactions consistent with phenomenology. To construct the MSSM, we follow this recipe
1. Choose the gauge symmetry (adopting appropriate gauge superelds for each gauge
symmetry).
2. Select matter and Higgs representations included as left-chiral scalar superelds.
9
3. Choose the superpotential
f as a gauge invariant analytic function of left-chiral scalar
superelds; the degree is 3 for a renormalizable theory.
4. Adopt all allowed gauge invariant soft SUSY breaking terms; these are generally chosen
to parameterize our ignorance of the SUSY-breaking mechanism.
5. Compute the supersymmetric Lagrangian using the master formula Eq. (1.28), aug-
mented by all possible soft SUSY breaking terms.
In the rst step, we choose the gauge symmetry of the Standard Model: SU(3)
C
SU(2)
L
U(1)
Y
. The gauge bosons of the SM are promoted to gauge superelds, so in the
Wess-Zumino gauge,
B
B (
0
, B
, T
B
),
W
A
W
A
(
A
, W
A
, T
WA
), A = 1, 2, 3, and
g
A
g
A
( g
A
, G
A
, T
gA
), A = 1, ..., 8.
Secondly, we stipulate the matter content to have three generations of quarks and leptons.
The fermion elds of the SM are promoted to chiral scalar superelds, with one supereld for
each chirality of every SM fermion. We use the left-handed charge conjugates of the right-
handed fermions, since the superpotential must be a function of just left-chiral superelds.
Then the matter superelds consist of
_
iL
e
iL
_
L
i
_
i
e
i
_
,
(e
R
)
c
E
c
i
,
_
u
iL
d
iL
_
Q
i
_
u
i
d
i
_
,
(u
R
)
c
U
c
i
,
(d
R
)
c
D
c
i
,
where, e.g.,
e = e
L
( x) +i
eL
( x) +i
L
T
e
( x), (1.33)
while
E
c
= e
R
( x) +i
E
c
L
( x) +i
L
T
E
c( x). (1.34)
10
SM Dirac fermions are constructed out of Majorana fermions via
e = P
L
e
+P
R
E
c, (1.35)
where in the chiral representation for matrices
e
=
_
_
_
_
e
1
e
2
e
2
e
1
_
_
_
_
and
E
c =
_
_
_
_
e
4
e
2
e
3
e
4
_
_
_
_
.
Next, we introduce the Higgs multiplets of the theory so that the SM Higgs doublet is
promoted to a doublet of left-chiral superelds:
=
_
+
0
_
H
u
=
_
h
+
u
h
0
u
_
(1.36)
These spin-
1
2
higgsinos with Y = 1 can circulate in triangle anomalies, thus it is necessary
to introduce a second left-chiral scalar doublet supereld with Y = 1,
H
d
=
_
h
h
0
d
_
(1.37)
The MSSM matter and Higgs supereld content along with their gauge transformation
properties and weak hypercharge assignments for a single generation is listed in Table 1.1
We now choose the MSSM superpotential to describe the interactions between the various
chiral superelds,
f =
H
a
u
H
da
+
i,j=1,3
_
(f
u
)
ij
ab
Q
a
i
H
b
u
U
c
j
+ (f
d
)
ij
Q
a
i
H
da
D
c
j
+ (f
e
)
ij
L
a
i
H
da
E
c
j
_
. (1.38)
We assume R-parity conservation, R = (1)
3(BL)+2s
, for all work done here, so baryon
and lepton number violating terms in the superpotential are excluded even though they are
gauge invariant and renomalizable.
Finally, we add into the Lagrangian all gauge invariant soft SUSY breaking terms,
L
soft
=
_
i
m
2
Q
ij
Q
j
+
d
Ri
m
2
Dij
d
Rj
+ u
Ri
m
2
Uij
u
Rj
+
L
i
m
2
Lij
L
j
+ e
Ri
m
2
Eij
e
Rj
+m
2
Hu
[H
u
[
2
+m
2
H
d
[H
d
[
2
_
1
2
_
M
1
0
+M
2
A
+M
3
g
B
g
B
i
2
_
M
0
+M
A
+M
g
B
5
g
B
+
_
(a
u
)
ij
ab
Q
a
i
H
b
u
u
Rj
+ (a
d
)
ij
Q
a
i
H
da
d
Rj
+ (a
e
)
ij
L
a
i
H
da
e
Rj
+ h.c.
_
+ [bH
a
u
H
da
+ h.c.] ,
11
Table 1.1: One generation of the MSSM.
Field SU(3)
C
SU(2)
L
U(1)
Y
L =
_
eL
e
L
_
1 2 1
E
c
1 1 2
Q =
_
u
L
d
L
_
3 2
1
3
U
c
3
1
4
3
D
c
3
1
2
3
H
u
=
_
h
+
u
h
0
u
_
1 2 1
H
d
=
_
h
h
0
d
_
1 2
-1
If we count the number of parameters in the MSSM, the number is daunting.
g
1
, g
2
, g
3
,
QCD
gaugino masses M
1
, M
1
, M
2
, M
2
, M
3
(M
3
absorbed into g)
m
2
Hu
, m
2
H
d
, , b (phase of b absorbed)
5 (6 + 3) = 45 in sfermion mass matrices
3 (3 3 2) = 54 in Yukawa matrices
3 (3 3 2) = 54 in a-term matrices
a global U(3)
5
transformation in matter allows 452 = 43 phases absorbed into matter
sfermions
total parameters = 9 + 5 + 45 + 54 + 54 43 = 124
However, we can simplify this number signicantly by neglecting all SUSY sources that are
CP-violating or lead to avor-changing neutral currents (FCNCs).
12
For a robust theory, the observed electroweak symmetry breaking that gives masses to
the W and Z bosons and fermions must be present. To investigate electroweak symmetry
breaking, we must examine the minima of the scalar potential in the MSSM. We can construct
this at tree-level with three parts
V
MSSM
= V
F
+V
D
+V
soft
, (1.39)
with minimization conditions
V
h
0
u
=
V
h
0
d
= 0,
and non-trivial, real solutions
h
0
u
) v
u
and h
0
d
) v
d
= tan
v
u
v
d
.
Thus, W
, Z
0
and SM fermions (e.g., m
e
= f
e
v
d
) all become massive as in the Standard
Model.
Since states with the same electric charge, color, and spin can mix, SUSY predicts many
new particle states. These predicted new matter states are:
spin
1
2
massive color octet: gluino g
spin
1
2
bino, wino, neutral higgsinos neutralinos
0
1
,
0
2
,
0
3
,
0
4
spin
1
2
charged wino, higgsinos charginos
W
1
,
W
2
spin-0 squarks: u
L
, u
R
,
d
L
,
d
R
, s
L
, s
R
, c
L
, c
R
,
b
1
,
b
2
,
t
1
,
t
2
spin-0 sleptons: e
L
, e
R
,
e
,
L
,
R
,
,
1
,
2
,
(h usually SM-like)
Since we are examining unication of couplings at the GUT scale, we would be remiss
to not give a discussion about renomalization group equations (RGEs). If the MSSM is to
be valid between vastly dierent mass scales, then we must be able to relate parameters
between these scales. RGEs govern the evolution of gauge couplings, Yukawa couplings, the
term, and soft breaking parameters. For gauge couplings, RGEs have the form
dg
i
dt
= (g
i
) with t = log Q, (1.40)
13
where Q is the renormalization scale. In the Standard Model,
(g) =
g
3
16
2
_
11
3
C(G)
2
3
n
F
S(R
F
)
1
3
n
H
S(R
H
)
_
, (1.41)
where C(G) is the quadratic Casimir operator for the adjoint representation of the associated
Lie algebra, S(R
F
) is the Dynkin index for representation R
F
of the fermion elds, S(R
H
)
is the Dynkin index for representation R
H
of the scalar elds, n
F
is the number of fermion
species, and n
H
is the number of complex scalars. In the MSSM, the gauginos, matter and
Higgs scalars also contribute:
(g) =
g
3
16
2
[3C(G) S(R)] . (1.42)
The precision values of g
1
, g
2
and g
3
measured at Q = M
Z
at LEP2 can be used as boundary
conditions and extrapolated to higher energies. Gauge coupling evolution from the weak to
GUT scales is displayed in Fig. 1.1
The one-loop RGEs for third generation Yukawa couplings of the MSSM are given by
df
t
dt
=
f
t
16
2
_
i=13
c
i
g
2
i
+ 6f
2
t
+f
2
b
_
, (1.43)
df
b
dt
=
f
b
16
2
_
i=13
c
i
g
2
i
+f
2
t
+ 6f
2
b
+f
2
_
, (1.44)
df
t
dt
=
f
t
16
2
_
i=13
c
i
g
2
i
+ 3f
2
b
+ 4f
2
_
, (1.45)
where c
i
= (13/15, 3, 16/3), c
i
= (7/15, 3, 16/3), c
i
= (9/5, 3, 0), and t = log(Q).
Like the gauge and Yukawa couplings, the various soft SUSY breaking parameters as
well as the superpotential Higgs mass , evolve with energy scale. The one-loop RGEs for
the soft SUSY breaking parameters, , and for the third generation sfermion masses and
A-parameters are as follows (the rst two generations are easily obtained by making the
14
Figure 1.1: Evolution of the SU(3)
C
SU(2)
L
U(1)
Y
gauge coupling constants from the
weak scale to the GUT scale for the case of (a) the SM, (b) the MSSM with two Higgs
doublets, and (c) the MSSM with four Higgs doublets [63].
15
requisite replacements in the appropriate formulas):
dM
i
dt
=
2
16
2
b
i
g
2
i
M
i
, (1.46)
dA
t
dt
=
2
16
2
_
i
c
i
g
2
i
M
i
+ 6f
2
t
A
t
+f
2
b
A
b
_
, (1.47)
dA
b
dt
=
2
16
2
_
i
c
i
g
2
i
M
i
+ 6f
2
b
A
b
+f
2
t
A
t
+f
2
_
, (1.48)
dA
dt
=
2
16
2
_
i
c
i
g
2
i
M
i
+ 3f
2
b
A
b
+ 4f
2
_
, (1.49)
dB
dt
=
2
16
2
_
3
5
g
2
1
M
1
3g
2
2
M
2
+ 3f
2
b
A
b
+ 3f
2
t
A
t
+f
2
_
, (1.50)
d
dt
=
16
2
_
3
5
g
2
1
3g
2
2
+ 3f
2
t
+ 3f
2
b
+f
2
_
, (1.51)
dm
2
Q
3
dt
=
2
16
2
_
1
15
g
2
1
M
2
1
3g
2
2
M
2
2
16
3
g
2
3
M
2
3
+
1
10
g
2
1
S +f
2
t
X
t
+f
2
b
X
b
_
,
dm
2
t
R
dt
=
2
16
2
_
16
15
g
2
1
M
2
1
16
3
g
2
3
M
2
3
2
5
g
2
1
S + 2f
2
t
X
t
_
, (1.52)
dm
2
b
R
dt
=
2
16
2
_
4
15
g
2
1
M
2
1
16
3
g
2
3
M
2
3
+
1
5
g
2
1
S + 2f
2
b
X
b
_
, (1.53)
dm
2
L
3
dt
=
2
16
2
_
3
5
g
2
1
M
2
1
3g
2
2
M
2
2
3
10
g
2
1
S +f
2
_
, (1.54)
dm
2
R
dt
=
2
16
2
_
12
5
g
2
1
M
2
1
+
3
5
g
2
1
S + 2f
2
_
, (1.55)
dm
2
H
d
dt
=
2
16
2
_
3
5
g
2
1
M
2
1
3g
2
2
M
2
2
3
10
g
2
1
S + 3f
2
b
X
b
+f
2
_
, (1.56)
dm
2
Hu
dt
=
2
16
2
_
3
5
g
2
1
M
2
1
3g
2
2
M
2
2
+
3
10
g
2
1
S + 3f
2
t
X
t
_
, (1.57)
where m
Q
3
and m
L
3
denote the mass term for the third generation SU(2) squark and slepton
doublet respectively, and
X
t
= m
2
Q
3
+m
2
t
R
+m
2
Hu
+A
2
t
, (1.58)
X
b
= m
2
Q
3
+m
2
b
R
+m
2
H
d
+A
2
b
, (1.59)
X
= m
2
L
3
+m
2
R
+m
2
H
d
+A
2
, and (1.60)
S = m
2
Hu
m
2
H
d
+Tr
_
m
2
Q
m
2
L
2m
2
U
+m
2
D
+m
2
E
. (1.61)
16
1.2.2 Supergravity and the Minimal Supergravity Model
If we allow the parameters in SUSY transformations to become spacetime dependent, i.e.,
(x) in e
i Q
,
then SUSY becomes a local symmetry. Local SUSY transformations are called supergravity
transformations for reasons that will soon become clear. Just as for gauge theories, we need
to introduce a gauge eld to maintain covariance:
(x), a spin
3
2
vector-spinor (Rarita-
Schwinger) eld. In order to maintain local SUSY, we must also introduce a bosonic partner,
a spin 2 eld g
(x), where g
= e
a
e
b
ab
and
ab
is the Minkowski metric.
Supergravity (SUGRA) is inherently non-renormalizable, since gravity itself is non-
renomalizable. Although the Lagrangian for a general non-renormalizable SUSY theory
depends on three independent functions, the Kahler potential K(
o
,
o), the superpotential
f(
o), and the gauge kinetic function
f
AB
(
o), SUGRA depends only on the gauge kinetic
function and just one combination of the Kahler potential and superpotential called the
Kahler function,
G(
o
,
o) = K(
o
,
o) + log [
f(
o)[
2
. (1.62)
Other notable features of SUGRA are
it can be spontaneously broken just as we saw for SUSY,
being a local SUSY theory, a super-Higgs mechanism exists wherein the gravitino eld
gains a mass m
3/2
while graviton remains massless,
the MSSM can be embedded into a SUGRA theory along with gauge singlet eld(s)
h
m
and superpotential such that SUGRA is spontaneously broken (hidden sector),
SUGRA breaking communicated from the hidden sector to the visible sector via gravity
induces soft SUSY breaking terms of order m
3/2
.
17
The above features are used to build the Minimal Supergravity model (mSUGRA)
by assuming the MSSM is embedded in a SUGRA theory and SUSY is broken in the
hidden sector with m
3/2
M
weak
1 TeV. In the simplest models (e.g. using a Polonyi
superpotential in the hidden sector) universal scalar masses m
0
, gaugino masses m
1/2
and
trilinear terms A
)
are induced as soft SUSY breaking terms. Since this model is inspired by
gauge coupling unication, these universal values are usually taken at Q = M
GUT
210
16
GeV. The couplings and soft parameters are evolved from M
GUT
to M
weak
causing m
2
Hu
to
become negative and thereby breaking EW symmetry. All sparticle masses and mixings are
calculated at Q = M
weak
in terms of a small parameter set
m
0
, m
1/2
, A
0
, tan , sign() (1.63)
Although the mSUGRA model may be too simplistic to be a complete theory for beyond
weak scale physics, it has thus far been the paradigm SUSY model for phenomenological
analysis and is convenient with only ve parameters.
1.3 SO(10) SUSY Grand Unication
Grand unied theories (GUTs) based upon the gauge group SO(10) and augmented by
supersymmetry (SUSY) is currently one of the most promising concepts in particle physics.
In addition to gauge group unication, matter unication of each generation occurs within
the SO(10) 16-dimensional spinorial representation (16). Furthermore, even the simplest
SO(10) GUTs allow for Yukawa coupling unication, especially for the third generation.
Triangle anomaly cancellation is automatic in SO(10) theories, thus explaining the ad-hoc
triangle anomaly cancellation in SU(5) GUTs or in the SM. The combination with softly
broken N=1 SUSY allows for stabilization of the weak scale to GUT scale gauge hierarchy
and is experimentally supported by the fact that the measured weak scale gauge couplings
meet at M
GUT
under MSSM renormalization group evolution. SUSY SO(10) also elegantly
addresses the neutrino mass problem, since one only has matter unication found within the
supereld
(16) provided one adds to the set of supermultiplets a SM gauge singlet supereld
N
c
i
containing a right-handed neutrino state. Upon breaking of SO(10), a superpotential
term
f
1
2
M
N
i
N
c
i
N
c
i
leading to a Majorana neutrino mass M
N
i
is induced in the Lagrangian.
This term is required for implementing the see-saw mechanism for neutrino masses.
18
In the past, GUTs (including SUSY GUTs) formulated in 4-d spacetime have been
plagued with a variety of problems mainly associated with GUT gauge symmetry breaking via
the Higgs mechanism. These include the doublet-triplet splitting problem, lack of observation
of proton decay, and the frequently awkward implementation of GUT symmetry breaking
via at least one large and unwieldy Higgs representation. With the onset of model building
utilizing extra dimensions, it has been shown to be possible to formulate SUSY GUTs in
ve or more spacetime dimensions. Then, the GUT gauge symmetry can be broken via
compactication of the extra dimensions on a suitable sub-space, such as an orbifold. In these
5-d and 6-d SUSY GUT models, the large GUT-scale Higgs representations can be dispensed
with, the doublet-triplet problem can be solved, and the proton can be made longer-lived
than current limits or even absolutely stable [75]. The extra-dimensional SUSY GUT models
act as a sort of proof of principle of what might be possible in more complicated set-ups
where the SUSY GUT model might arise from compactication of superstring models.
The work presented in this dissertation is chiey concerned with Yukawa unication
within SO(10) SUSY GUTs. This implies additional restrictions on the general model,
especially splitting of the GUT scale Higgs masses. Further characterization of the specic
models used here and results obtained within these models will be saved for the upcoming
chapters.
19
CHAPTER 2
Yukawa-unied SO(10) SUSY GUTs
To avoid dealing with the unknown physics above the GUT scale, we will assume that nature
is described by an SO(10) SUSY GUT theory at energy scales Q > M
GUT
2 10
16
GeV
and that the model breaks (either via the Higgs mechanism or via compactication of extra
dimensions) to the MSSM (or MSSM plus right-handed neutrino states) at Q = M
GUT
.
Thus, below M
GUT
the MSSM is the correct eective eld theory which describes nature.
We will further assume that the superpotential above M
GUT
is of the form
f f
16
16
10
+ (2.1)
so that the third generation Yukawa couplings f
t
, f
b
and f
are unied at M
GUT
. It is
simple in this context to include as well the eect of a third generation neutrino Yukawa
coupling f
; this eect has been shown to be small, although it can help improve Yukawa
coupling unication by a few percent if the neutrino Majorana mass scale is within a few
orders of magnitude of M
GUT
. Within this ansatz, the GUT scale soft SUSY breaking (SSB)
terms are constrained by the SO(10) gauge symmetry so that matter scalar SSB terms have a
common mass m
16
, Higgs scalar SSB terms have a common mass m
10
, and there is a common
trilinear soft breaking parameter A
0
. As usual, the bilinear soft term B can be traded for
tan , the ratio of Higgs eld vevs, while the magnitude of the superpotential Higgs mass is
determined in terms of M
2
Z
via the electroweak symmetry breaking minimization conditions.
Here, electroweak symmetry is broken radiatively (REWSB) due to the large top quark mass.
In order to accomodate REWSB, it is well-known that in Yukawa-unied models, the
GUT scale Higgs soft masses must be split such that m
2
Hu
< m
2
H
d
in order to fulll the
EWSB minimization conditions; this eectively gives m
2
Hu
a head start over m
2
H
d
in running
towards negative values at or around the weak scale. We parametrize the Higgs splitting
20
as m
2
H
u,d
= m
2
10
2M
2
D
. The Higgs mass splitting might originate via a large near-GUT-
scale threshold correction arising from the neutrino Yukawa coupling: see the Appendix
to Ref. [46] for discussion. Thus, the Yukawa unied SUSY model is determined by the
parameter space
m
16
, m
10
, M
2
D
, m
1/2
, A
0
, tan , sign() (2.2)
along with the top quark mass. We will take m
t
= 171 GeV, in accord with recent
measurements from CDF and D0 [3].
In particle phenomenology, we continually strive for agreement with established and/or
forthcoming experimental data in our cutting-edge physics models. With this in mind, one
primary concern is to determine how SO(10) SUSY GUT theories could manifest themselves
in the environment of the LHC detectors while enforcing Yukawa coupling unication and
dark matter (DM) abundance of the universe constraints. This chapter is dedicated to
addressing this concern by means of analyzing complementary Random Scan (RS) and
Markov Chain Monte Carlo (MCMC) scan methods.
2.1 HS v. DT Models
The necessity of splitting the Higgs soft masses at the GUT scale for Yukawa coupling
unication has already been discussed. However, there are two known ways in which to do
this: i.) through D-term contributions to all scalar masses[7] (the DT model), or ii.) via
splitting of only the Higgs soft terms[46] (the HS model). Ensuing is an explanation of why
we choose the HS model method of splitting the Higgs terms followed by our results within
this model.
In Ref. [8], it was found using the Isajet sparticle mass spectrum generator [83] Isasugra
that Yukawa coupling unication to 5% could be achieved in the MSSM using D-term
splitting, but only for < 0; for > 0, the Yukawa coupling unication was much worse, of
order 3050%. These parameter space scans allowed m
16
values of up to only 1.5 TeV and
used a GUT scale Yukawa unication quantity
R =
max(f
t
, f
b
, f
)
min(f
t
, f
b
, f
)
, (2.3)
so that, e.g., R = 1.1 would correspond to 10% Yukawa unication. The < 0 Yukawa
unication solutions were examined in more detail in Ref. [9], where dark matter allowed
21
solutions were found, and the neutralino A-annihilation funnel was displayed for the rst
time.
With the announcement from BNL experiment E-821 that there was a 3 deviation from
SM predictions on the muon anomalous magnetic moment a
(g 2)
t
1
1 TeV.
1
In a long follow-up study using Isajet, Auto et al. [76] found that Yukawa-unied solutions
good to less than a few percent could be found in the > 0 case using the HS model of
BDR, but only for very large values of m
16
>
510 TeV and low values of m
1/2
<
100 GeV,
again using Bagger et al. boundary conditions. Yukawa unication in the DT model was
at best good to 10% (for this reason, in this chapter, we focus only on the HS model). The
spectra were characterized by three mass scales:
1. 515 TeV rst and second generation scalars,
2. 1 TeV third generations scalars, term and m
A
and
3. chargino masses m
f
W
1
100200 GeV and gluino masses m
g
350450 GeV.
These Yukawa-unied solutions owing to very large values of scalar masses, m
A
and
predicted dark matter relic density values
0
1
h
2
far beyond the WMAP-measured result [49]
of
DM
h
2
= 0.111
+0.011
0.015
(2). (2.5)
Meanwhile, the spectra generated using the BDR program could easily generate
0
1
h
2
values
close to 0.1 since their allowed and m
A
values were far lower, so that mixed higgsino dark
matter or A-funnel annihilation solutions could easily be found. In follow-up papers to the
BDR program [13, 14], the neutralino relic density and branching fraction B
s
+
were
evaluated. To avoid constraints on BF(B
s
+
t
L
m
t
R
, while non-mixing
SSB terms are frozen out at their own mass scale [84]. The scalar potential is minimized
using the RG-improved one-loop MSSM eective potential evaluated at an optimized scale
Q = M
SUSY
which accounts for leading two-loop eects [19]. Once the tree-level sparticle
mass spectrum is computed, full one-loop radiative corrections are calculated for all sparticle
and Higgs boson masses, including complete one-loop weak-scale threshold corrections for
the top, bottom, and tau masses at scale Q = M
SUSY
. These fermion self-energy terms are
critical to evaluating whether or not Yukawa couplings do indeed unify. Since the GUT-scale
Yukawa couplings are modied by the threshold corrections, the Isajet RGE solution must
be imposed iteratively with successive updown running until a convergent solution for the
spectrum is found. For most of parameter space, there is very good agreement between Isajet
2
As inputs, we take the top quark pole mass m
t
= 171 GeV. We also take m
DR
b
(M
Z
) = 2.83 GeV [18]
and m
DR
(M
Z
) = 1.7463 GeV. The paper Ref. [76] addresses consequences of varying the values of m
t
and
m
b
.
24
and the other public spectrum codes SoftSusy, SuSpect, and SPheno, although at the edges
of parameter space agreement between the four codes typically diminishes [20].
We rst adopt a wide parameter range scan, and then once the best Yukawa-unied
regions are found, we adopt a narrow scan to try to hone in on the best unied solutions.
The parameter range we adopt for the wide (narrow) scan is
m
16
: 0 20 TeV (1 20 TeV),
m
10
/m
16
: 0 1.5 (0.8 1.4),
m
1/2
: 0 5 TeV (0 1 TeV),
A
0
/m
16
3 3 (2.5 1.9),
M
D
/m
16
: 0 0.8 (0.25 0.8),
tan : 40 60 (46 53).
(2.6)
For the random scan, we evaluate
0
1
h
2
, BF(b s), a
and BF(B
S
+
) using
Isatools (a sub-package of Isajet). We plot only solutions for which m
f
W
1
> 103.5 GeV, in
accord with LEP2 searches, and for the moment implement no other constraints, such as
relic density, Higgs mass, etc..
2.2.1 Random scan results
Our rst results are shown in Fig. 2.1, where we show points from the wide scan (dark blue)
and points from the narrow scan (light blue) in the parameter versus R plane. From frame
a), we see that Yukawa unication to better than 30% (R < 1.3) cannot be achieved for
m
16
< 1 TeV, while Yukawa coupling unication becomes much more likely at multi-TeV
values of m
16
. Frame b) shows that Yukawa-unied models prefer m
10
1 1.3 m
16
, while
frame c) shows that a positive value of M
D
(0.25 0.5) m
16
which yields m
2
Hu
< m
2
H
d
is preferred. In frame d), we see that the best Yukawa-unied solutions are found for
the lowest possible values of m
1/2
. We note here that using 1-loop RGEs along with the
LEP2 constraint m
f
W
1
> 103.5 GeV one would expect from models with gaugino mass
unication that since m
f
W
1
M
2
(weak) 0.8m
1/2
that we would have m
1/2
>
125 GeV
always. However, the very large values of m
16
we probe alter the simple 1-loop gaugino mass
unication condition (that
M
1
1
=
M
2
2
=
M
3
3
) via 2-loop RGE eects. Thus, values of m
1/2
much lower than 125 GeV are possible if m
16
is large.
In frame e), we see a sharp dependence that Yukawa-unied solutions can only be obtained
for A
0
2m
16
, while frame f) shows that tan must indeed be large: in the range 4753.
Bagger et al. had shown in Ref. [82] that a radiatively-driven inverted scalar mass hierarchy
25
with m(third generation) m(first/second generation) could be derived provided one
starts with unied Yukawa couplings, the boundary conditions
4 m
2
16
= 2 m
2
10
= A
2
0
, (2.7)
and one neglects the eect of gaugino masses. Our results in Fig. 2.1 show the inverse
eect: that Yukawa coupling unication can only be achieved if one imposes the boundary
conditions (2.7) along with m
16
m
1/2
. This result holds only in our numerical calculations
for > 0 and A
0
< 0 and of course m
2
Hu
< m
2
H
d
. The results shown in Fig. 2.1 also verify that
the results obtained in Ref. [76] still hold, even with updated spectra code and a lower value
of m
t
= 171 GeV. In Fig. 2.2, we show various -ino masses
3
versus R as generated from our
random scan. In frame a), we see that owing to the preference of Yukawa-unied solutions
to have m
1/2
as small as possible the chargino mass m
f
W
1
is preferred to be quite light, as
close to the LEP2 limit as possible, with m
f
W
1
100200 GeV. Likewise, in frame b), the
gluino mass should be relatively light, with m
g
350500 GeV. The lightest neutralino
0
1
mass is shown in frame c), and is preferred in the range m
0
1
50100 GeV. Meanwhile, the
mass dierence m
0
2
m
0
1
is shown in frame d), and is also in the range 50100 GeV. This
latter quantity is important because if m
0
2
m
0
1
< M
Z
, two body spoiler decay modes such
as
0
2
0
1
Z will be kinematically closed, and the three body decays
0
2
0
1
( = e or )
should occur at a suciently large rate at the LHC that an edge should be visible in the m(
)
invariant mass distribution at m
0
2
m
0
1
[21]. This measureable mass edge can serve as the
starting point for sparticle mass reconstruction in SUSY particle cascade decay events at the
LHC [22]. Thus, in Yukawa-unied models, this mass edge is highly likely to be visible. In
Fig. 2.3, we show the expected masses of a) u
L
-squark, b) the
t
1
-squark, c), the pseudoscalar
Higgs boson A and d) the superpotential Higgs parameter . Frame a) shows that Yukawa-
unied solutions prefer rst/second generation squarks and sleptons with masses in the 520
TeV range far higher than values typically examined in phenomenological SUSY studies!
The top squark mass and the A, H and H
0
1
h
2
0.1, then we generate Yukawa-unied solutions to 40% unication with the random
scan. This plot underscores the diculty of nding sparticle mass spectra solutions which
are compatible with both the measured dark matter abundance and tb Yukawa coupling
unication.
28
Figure 2.3: Plot of R versus various sparticle masses for a random scan over the parameter
range listed in Eq. (2.6). We take > 0 and m
t
= 171 GeV.
Figure 2.4: Plot of
0
1
h
2
vs. R for a random scan over the parameter range listed in Eq. (2.6).
We take > 0 and m
t
= 171 GeV.
29
2.2.2 Three proposals to reconcile Yukawa-unied models with
dark matter relic density
Dark matter solution via neutralino decay to axino
We see from Fig. 2.4 that models generated from the random scan with R 1.0 all have
0
1
h
2
30 30, 000 far beyond the WMAP-measured result of
CDM
h
2
0.1. One
possible solution to reconcile the predicted and measured dark matter density is to assume
that the lightest neutralino
0
1
is in fact not the LSP but is unstable. Some alternative LSP
candidates consist of the gravitino
G or the axino a. In gravity-mediated SUSY breaking
models, the gravitino mass m
3/2
arises due to the superHiggs mechanism and is expected
to set the scale for all the soft SUSY breaking terms. Usually it is assumed the gravitino
is heavier than the lightest neutralino m
3/2
> m
0
1
, in which case the gravitino essentially
decouples from phenomenology. However, if m
3/2
< m
0
1
, then the
0
1
becomes unstable and
can decay via modes such as
0
1
G. The
0
1
lifetime is expected to be very long of
order 10
4
10
12
sec so the neutralino still escapes detection at collider experiments, but is
susceptible to constraints from Big Bang nucleosynthesis (BBN) and CMB anisotropies [23].
The relic density of gravitinos is expected to be simply
G
h
2
=
m
3/2
m
0
1
0
1
h
2
, since the
gravitinos inherit the thermally produced neutralino relic number density. Thus, a scenario
with a
G superWIMP as LSP in SUGRA-type models can reduce the relic density by typically
factors of a few which is not enough in the case of Yukawa-unied models, where relic
density suppression factors of 10
2
10
5
are needed.
A better option occurs if we hypothesize an axino a LSP. If indeed there is a Peccei-Quinn
solution to the strong CP problem, then one expects the existence of axions, typically with
mass below the eV scale. While axions can themselves form cold dark matter, it is also easily
possible that they contribute little to the CDM relic density. However, in models with SUSY
and axions, then the axion is just one element of an axion supereld, the superpartner of
the axion being a spin-
1
2
axino a. The axino mass can be far dierent from the typical soft
SUSY breaking scale, and the range m
a
eVGeV is allowed.
Axinos can be produced in the early universe both thermally or non-thermally from NLSP
decay. From the latter source, we expect roughly [24]
a
h
2
m
a
m
0
1
0
1
h
2
. (2.8)
30
Thus, for
0
1
h
2
10
3
and with m
0
1
50 GeV as in Yukawa-unied models, an axino mass
of m
a
<
5 MeV is required.
In this mass range, the axinos from
0
1
decay are expected to give a hot/warm component
to the dark matter [92]. However, thermally produced axinos in this mass range could yield
the required cold dark matter. Thus, if an unstable neutralino decay
0
1
a is to reconcile
Yukawa-unied models with the relic density, then we would expect the dark matter to be
predominantly cold axinos produced thermally, but with a re-heat temperature T
R
< T
f
,
where T
f
is the temperature where axinos decouple from the thermal plasma in the early
universe. This scenario admits a dark matter abundance that can be in accord with WMAP
measurements, and would be primarily CDM, but with a warm dark matter component
arising non-thermally from
0
1
decays. For a bino-like neutralino, as in Yukawa-unied
models, the
0
1
lifetime is given by [26]
3.3 10
2
sec
1
C
2
aY Y
_
f
a
/N
10
11
GeV
_
2
_
50 GeV
m
0
1
_
3
, (2.9)
where the model-dependent constant C
aY Y
is of order 1, f
a
is the Peccei-Quinn breaking
scale, and N is a model dependent factor (N = 1(6) for the KSVZ (DFSZ) axion model).
Thus, for reasonable choices of model parameters, we expect the neutralino lifetime to be
of order 3 10
2
sec. This is short enough so that photon injection into the early universe
from
0
1
a decay occurs before nucleosynthesis, thus avoiding constraints from BBN.
For illustration, we adopt a point A listed in Table 2.1 of Yukawa-unied benchmark
models. The point has m
16
= 9202.9 GeV, m
10
= 10966.1 GeV, M
D
= 3504.4 GeV,
m
1/2
= 62.5 GeV, A
0
= 19964.5 GeV, tan = 49.1 GeV with > 0 and m
t
= 171
GeV. It has m
0
1
= 55.6 GeV and
0
1
h
2
= 423 (IsaReD result). Thus,
0
1
a with m
a
<
1
MeV would allow for a mixed warm/cold axino dark matter solution to the problem of relic
density in Yukawa-unied models.
Dark matter solution via non-universal gaugino masses
An alternative solution to reconciling the dark matter abundance with Yukawa-unied
models is to consider the possibility of non-universal gaugino masses. If we adopt any of the
Yukawa unied models from the random scan and vary the SU(2) (SU(3)) gaugino masses
M
2
(M
3
), then the Yukawa coupling unication will be destroyed via the eect of
t
i
W
j
( g q)
31
loops. However, if M
1
is varied, Yukawa coupling unication is preserved since contributions
to fermion masses from loops containing
0
1
are small.
By raising the GUT scale value of M
1
to values higher than m
1/2
, the weak scale value
of M
1
is also increased. If M
1
is increased enough, then m
0
1
(which is nearly equal to M
1
since
0
1
is largely bino-like) becomes close to m
f
W
1
. When this happens, the
0
1
becomes
more wino-like, with an increased annihilation cross section to WW pairs if m
0
1
> M
W
[27].
In our case, usually m
0
1
< M
W
. Then raising M
1
still lowers the relic density, but now via
bino-wino co-annihilation (BWCA) [17].
In Fig. 2.5, we show the variation in
0
1
h
2
versus M
1
(M
GUT
) for benchmark point A
in Table 2.1. The location of M
1
for point A is marked by the arrow. The double dips at
low M
1
are due to neutralino annihilation through the Z and h poles. Once M
1
(M
GUT
) is
increased to 195 GeV, then we reach a relic density in accord with WMAP measurements.
Since m
f
W
1
m
0
1
, and m
f
W
1
m
0
1
, the
0
2
0
1
mass gap is small, of order 1020 GeV. We
list the raised M
1
(M
GUT
) = 195 GeV point as point B in benchmark Table 2.1.
Figure 2.5: Plot of variation in
0
1
h
2
versus non-universal GUT scale gaugino mass M
1
for
benchmark point A in Table 2.1.
Dark matter solution via generational non-universality
Another possibility for reconciling the neutralino relic density with the measured value is
to lower the rst/second generation scalar masses m
16
(1, 2), while keeping m
16
(3) xed
32
at m
16
. The Bagger et al. inverted hierarchy solution depends only on third generation
scalar masses, while the eects of the rst two generations decouple. Ordinarily, solutions
with m
16
(1, 2) = m
16
(3) are taken to enforce the super-GIM mechanism for suppression of
avor changing neutral current (FCNC) processes. Limits from FCNCs mainly require near
degeneracy between the rst two generations, while limits on third generation universality
are much less severe [28]. Lowering m
16
(1, 2) works to lower the relic density because of the
large S term in the scalar mass RGEs:
S = m
2
Hu
m
2
H
d
+Tr
_
m
2
Q
m
2
L
2m
2
U
+m
2
D
+m
2
E
. (2.10)
In models with universality, like mSUGRA, S = 0 to one-loop at all energy scales; in models
with non-universal Higgs scalars, like the HS model, this term can be large and have a major
inuence on scalar mass running. The large S term helps suppress right-squark masses. If
m
16
(1, 2) is taken light enough, then m
u
R
m
c
R
m
0
1
, and neutralino-pair annihilation
into quarks and neutralino-squark co-annihilation can act to reduce the relic density.
In Fig. 2.6, we show the variation in
0
1
h
2
versus m
16
(1, 2) where we take m
16
(3) = 5018.8
GeV, m
1/2
= 160 GeV, A
0
= 10624.2 GeV, tan = 47.8 and > 0. When m
16
(1, 2) is
lowered to 603.8 GeV, then m
u
R
m
c
R
= 98.3 GeV, and we have neutralino annihilation
via light t-channel squark exchange and also neutralino-squark co-annihilation.
4
IsaReD
and Micromegas give
0
1
h
2
0.1 at this point, which we adopt as benchmark point C in
Table 2.1. The two light squarks are just at the limit of LEP2 exclusion. They may possibly
be excludable by Tevatron analyses, but the squark neutralino mass gap is quite small, so
the energy release from u
R
u
0
1
is low. So far, no such study has been made, and so the
possibility cannot yet be denitively excluded.
2.3 Discussion of Markov Chain Monte Carlo analysis
The Markov Chain Monte Carlo (MCMC) technique is an improvement over Random
Scanning (RS) in that it searches more eciently for parameter space regions of good Yukawa
unication and WMAP-compatible DM relic density. A Markov Chain [29] is a discrete-time,
random process having the Markov property, which is dened such that given the present
state, the future state only depends on the present state, but not on the past states. That
4
A bug x is needed in the Isajet 7.75 IsaReD subroutine in order to obtain the correct relic density.
This bug has been rectied in all later versions of Isajet.
33
Figure 2.6: Plot of variation in
0
1
h
2
versus non-universal GUT scale rst/second generation
scalar mass m
16
(1, 2) for benchmark point C in Table 2.1.
is:
P(X
t+1
= x[X
t
= x
t
, ..., X
1
= x
1
) = P(X
t+1
= x[X
t
= x
t
). (2.11)
An MCMC constructs a Markov Chain through sampling from a parameter space with
the help of a specied algorithm. In this study, we have applied the Metropolis-Hastings
algorithm [30], which generates a candidate state x
c
from the present state x
t
using a proposal
density Q(x
t
; x
c
). The candidate state is accepted to be the next state x
t+1
if the ratio
p =
P(x
c
)Q(x
t
; x
c
)
P(x
t
)Q(x
c
; x
t
)
, (2.12)
(where P(x) is the probability calculated for the state x) is greater than a uniform random
number a = U(0, 1). If the candidate is not accepted, the present state x
t
is retained and a
new candidate state is generated. For the proposal density we use a Gaussian distribution
that is centered at x
t
and has a width . This simplies the p ratio to P(x
c
)/P(x
t
).
Once taking o from a starting point, Markov chains are aimed to converge at a target
distribution P(x) around a point with the highest probability. The time needed for a Markov
chain to converge depends on the width of the Gaussian distribution used as the proposal
density. This width can be adjusted during the run to achieve a more ecient convergence.
In our search in the SO(10) parameter space, we assume at priors and we approximate
34
the likelihood of a state to be e
2
(x)
. We dene the
2
for R as
2
R
=
_
R(x) R
unification
R
_
2
(2.13)
where R
unification
= 1 and
R
is the discrepancy we allow from absolute Yukawa unication
which, in this case, we take to be 0.05. On the other hand, for h
2
we dene
2
h
2 =
_
1, (0.094 h
2
0.136)
_
h
2
(x)h
2
mean
h
2
_
2
, (h
2
< 0.094 or h
2
> 0.136)
(2.14)
where h
2
mean
= 0.115 is the mean value of the range 0.094 < h
2
< 0.136 proposed
in [31], and
h
2 = 0.021. This way, the MCMC primarily searches for regions of Yukawa-
unications, and within these regions for solutions with a good relic density.
For each search, we select a set of 10 starting points in order to ensure a more thorough
investigation of the parameter space. Then we run the MCMC, aiming to maximize the
likelihood of either R alone, or R and h
2
simultaneously. For the case of simultaneous
maximization, we compute the p ratios for R and h
2
individually, requiring both p
R
> a and
p
h
2 > a separately. We do not strictly seek convergence to an absolute maximal likelihood,
but we rather use the MCMC as a tool to reach compatible regions and to investigate the
amount of their extension in the SO(10) parameter space.
2.3.1 HS model: neutralino annihilation via h resonance
The MCMC scans were initiated by selecting 10 starting points pseudorandomly that
is, selecting them from dierent m
16
regions to cover a wider range of the parameter space
and imposing some loose limits (dened by previous works and random scans) on the rest of
their parameters to achieve a more ecient convergence. Our initial scan is directed to look
for points only with R as close to 1.0 as possible by maximizing solely the likelihood of R.
Based on the results of the rst MCMC scan, we then pick a new set of 10 starting points
with low R and also low
0
1
h
2
, and direct the second scan to look for points with both
R = 1.0 and
0
1
h
2
< 0.136 by maximizing the likelihoods of R and h
2
simultaneously. For
MCMC scans, the code is interfaced to the micrOMEGAs [32] package to evaluate the relic
density and low-energy constraints.
Figure 2.7 shows the Yukawa-unied region found by the MCMC results as a projection
in the plane of m
16
versus m
10
. The light-blue dots are points which have R < 1.1, while
35
dark blue dots have R < 1.05. In addition, we show in orange (red) the points which satisfy
R < 1.1 (1.05) and
0
1
h
2
< 0.136. The points with low R are narrowly correlated along the
line m
10
1.2m
16
. While the low R points range over m
16
values from 3 to over 12 TeV (in
agreement with the results from the random scans) the MCMC has also identied a range
of points with both R 1 and
0
1
h
2
< 0.136, but only for m
16
values of about 34 TeV!
Figure 2.7: Plot of MCMC results in the m
16
vs. m
10
plane; the light-blue (dark-blue) points
have R < 1.1 (1.05), while for the orange (red) points R < 1.1 (1.05) and
0
1
h
2
< 0.136.
Fig. 2.8 shows the MCMC scan results in the m
16
vs. A
0
/m
16
plane. Again, we see that
points with low R populate the region with A
0
(22.1)m
16
over a wide range of m
16
values.
The plot includes the
0
1
h
2
< 0.136 points around m
16
34 TeV.
In Fig. 2.9, we show MCMC results in the m
16
vs. m
1/2
plane. Here, we see the very
lowest R points select out the lowest possible m
1/2
values allowed for a given value of m
16
, and
that the minimum m
1/2
value allowed steadily decreases with increasing m
16
the boundary
being determined by the LEP2 limit on chargino masses. The points with a good relic
density are clustered around m
1/2
100 GeV.
We also show in Fig. 2.10 the individual GUT-scale values of Higgs soft terms m
Hu
(lower branch) and m
H
d
(upper branch). This plot displays the required Higgs splitting and
conrms that m
H
d
> m
Hu
.
36
Figure 2.8: Plot of MCMC results in the m
16
vs. A
0
/m
16
plane; the light-blue (dark-
blue) points have R < 1.1 (1.05), while for the orange (red) points R < 1.1 (1.05) and
0
1
h
2
< 0.136.
Figure 2.9: Plot of MCMC results in the m
16
vs. m
1/2
plane; the light-blue (dark-blue) points
have R < 1.1 (1.05), while for the orange (red) points R < 1.1 (1.05) and
0
1
h
2
< 0.136.
37
Figure 2.10: Plot of MCMC results in the m
16
vs. m
H
d,u
plane; the light-blue (dark-
blue) points have R < 1.1 (1.05), while for the orange (red) points R < 1.1 (1.05) and
0
1
h
2
< 0.136.
In Fig. 2.11, we show points with low R in the m
h
2m
0
1
vs. m
A
2m
0
1
plane. In
these solutions, m
A
is usually far greater than 2m
0
1
, indicating the neutralino annihilation
through the A-resonance is not the cause of the reduced relic density orange and red points.
However, the low
0
1
h
2
points all do lie along the m
h
2m
0
1
line, indicating that h-
resonance annihilation is the mechanism at work to reduce the relic density in the early
universe. In Fig. 2.12, we show R vs.
0
1
h
2
for the MCMC scan. In this frame, we see that
the points with high relic density extend down to R = 1, while the low relic density points
reach below R = 1.05 but can reach no lower than R = 1.03.
In summary, what we learn from this set of scans is that the search for low R pushes
m
16
to very high, multi-TeV values. Meanwhile, in order for h-resonance annihilation to
reduce the relic density to the WMAP-allowed range, m
16
cannot be too large. The region
around m
16
34 TeV oers a compromise between these two tendencies: for m
16
not too
large, the dip in relic density due to the h-resonance annihilation is sucient to bring the
relic density into the desired range. But since m
16
cant be too large, the Yukawa unication
38
Figure 2.11: Plot of MCMC results in the m
h
2m
0
1
vs. m
A
2m
0
1
plane; the light-blue
(dark-blue) points have R < 1.1 (1.05), while for the orange (red) points R < 1.1 (1.05) and
0
1
h
2
< 0.136.
Figure 2.12: Plot of MCMC results in the R vs.
0
1
h
2
plane; the light-blue (dark-blue) points
have R < 1.1 (1.05), while for the orange (red) points R < 1.1 (1.05) and
0
1
h
2
< 0.136.
39
is limited to a couple of percent at best. This new class of solutions was dicult to reach
using a random scan, since the h-resonance is so narrow. The necessary value of m
0
1
has to
be just right with 2m
0
1
slightly below m
h
so that the thermal averaging of neutralino
energies convolutes with the resonant cross section with enough strength to give substantial
neutralino annihilation in the early universe.
We adopt point D in Table 2.1 as being representative of the light Higgs h-resonance
annihilation compromise solutions. The relic density computed with micrOMEGAs (
0
1
h
2
=
0.06) is below the preferred range, while IsaReD gives
0
1
h
2
= 0.1. Yukawa couplings are
unied at the 9% level. We note here that we could have adopted a solution with even better
Yukawa coupling unication at the 45% level. These solutions tend to give light Higgs mass
m
h
<
110 GeV (as can be seen by the red dots in Fig. 2.13) which are more likely to be
excluded by LEP2 Higgs search results.
Figure 2.13: Plot of MCMC results in the m
0
2
m
0
1
vs. m
h
plane; the light-blue (dark-
blue) points have R < 1.1 (1.05), while for the orange (red) points R < 1.1 (1.05) and
0
1
h
2
< 0.136.
The SO(10) model parameters leading to low R and good relic density occur only over
a very narrow range of m
1/2
100 GeV and m
16
3 TeV. This means the Yukawa-unied
40
h-resonance annihilation points have very specic mass spectra predictions. Implications for
collider searches within this model is discussed in Chapter 4 of this thesis.
2.3.2 Solutions using weak scale Higgs boundary conditions
In the analysis put forth by BDR [46], Yukawa-unied solutions are found with low values
of both and m
A
in the 100200 GeV range, while m
16
and m
10
are typically at the 23
TeV scale. We have seen from our results so far that and m
A
are typically in the TeV
regime. Some low solutions were generated using Isajet in Table 2 of Ref. [76], but these
had R 1.25.
We nd here that we can generate small and small m
A
solutions using Isajet by using
the pre-programmed non-universal Higgs model (NUHM).
5
The approach is to start with a
set of GSH soft term boundary conditions and evolve the soft SUSY breaking Higgs masses
m
2
Hu
and m
2
H
d
down to the weak scale M
SUSY
. At Q = M
SUSY
, re-calculate what m
2
Hu
and
m
2
H
d
should have been in order to get the input values of m
A
and using the two electroweak
symmetry breaking minimization conditions (in practice, we use 1-loop relations):
B =
(m
2
Hu
+m
2
H
d
+ 2
2
) sin 2
2
and (2.15)
2
=
m
2
H
d
m
2
Hu
tan
2
(tan
2
1)
M
2
Z
2
, (2.16)
then run back up to the GUT scale using these new WSH boundary conditions. At each
iteration, the weak scale values of m
2
Hu
and m
2
H
d
have to be re-computed so as to maintain
the input value of and m
A
; in this case, the GUT scale values of m
2
Hu
and m
2
H
d
are outputs
instead of inputs. For this class of solutions, both GSH and WSH boundary conditions must
be used in Isajet. The GSH boundary conditions are needed just to get an acceptable EWSB
on the rst iteration so that a spectrum can be computed, then later modied to yield the
input values of m
A
and . Using default universal GSH soft terms will usually fail to give
appropriate EWSB on any iteration where Yukawa couplings are unied.
We implement an MCMC scan over the modied parameter space
m
16
, m
1/2
, A
0
, tan , m
A
, (2.17)
(eectively trading the GUT scale inputs m
2
Hu
and m
2
H
d
(or alternatively m
10
and M
2
D
) for
weak scale inputs m
A
and ). We begin with 10 starting points selected pseudorandomly
5
This is model line 8 of the Isajet non-universal supergravity models (NUSUG).
41
from dierent regions of the above parameter space, and implement two MCMC scans on
them, one searching for points with lowest R values by maximizing the likelihood of R and
the other for solutions with R = 1 and
0
1
h
2
< 0.136 by maximizing likelihoods of R and
h
2
simultaneously.
Our rst results are shown in Fig. 2.14 for the m
16
vs. A
0
/m
16
plane, where we plot points
with R < 1.1 (1.05) using dark blue (light blue) dots, and solutions with
0
1
h
2
< 0.136 for
R < 1.1 (1.05) using orange (red) dots. While we again get good Yukawa-unied solutions
over a wide range of multi-TeV values of m
16
, this time we pick up additional dark matter
allowed solutions for m
16
: 36 TeV. The solutions again respect the Bagger et al. boundary
condition A
0
2m
16
.
Figure 2.14: Plot of MCMC results using WSH boundary conditions in the m
16
vs. A
0
/m
16
plane; the light-blue (dark-blue) points have R < 1.1 (1.05), while for the orange (red) points
R < 1.1 (1.05) and
0
1
h
2
< 0.136.
Using these boundary conditions, while we again get good Yukawa-unied solutions over
a wide range of multi-TeV values of m
16
, this time we pick up additional dark matter allowed
solutions for m
16
: 36 TeV. The solutions again respect the Bagger et al. boundary condition
A
0
2m
16
. An additional scan (not included here) shows that the minimum in allowed
42
m
1/2
values again decreases with increasing m
16
. We see that for the WSH class of solutions,
much larger values of m
1/2
ranging up to 300 500 GeV are DM-allowed.
In Fig. 2.15, the bulk of the DM-allowed solutions occur at relatively low values of
m
A
130250 GeV. These low m
A
solutions were extremely dicult to generate with the
top-down approach, and indicate that they have a high degree of ne-tuning.
6
A scattering of
DM-allowed dots occur with high m
A
values. These turn out to be the h-resonance solutions
as generated with the GSH boundary conditions in Sec. 2.3.1. This is seen more clearly by
plotting in the m
h
2m
0
1
vs. m
A
2m
0
1
plane, Fig. 2.16 where we see a narrow strip at
m
h
2m
0
1
= 0 corresponding to h-resonance annihilation solutions, while we also have a
wider band of solutions at m
A
2m
0
1
= 0, which indicate neutralino annihilation through
the A-resonance. The width of the latter band is due to the fact that the A width can be
quite wide typically a few GeV, while the h-width is much narrower, of order 50 MeV.
Figure 2.15: Plot of MCMC results using WSH boundary conditions in the m
A
vs. plane;
the light-blue (dark-blue) points have R < 1.1 (1.05), while for the orange (red) points
R < 1.1 (1.05) and
0
1
h
2
< 0.136.
The A-resonance solutions occur at tan 50 and relatively low m
A
values. This can
6
The TW paper (Ref. [11]) remarks that there must be considerable ne-tuning as well to reconcile
BF(b s) with Yukawa unication and the dark matter relic abundance.
43
Figure 2.16: Plot of MCMC results using WSH boundary conditions in the m
h
2m
0
1
vs.
m
A
2m
0
1
plane; the light-blue (dark-blue) points have R < 1.1 (1.05), while for the orange
(red) points R < 1.1 (1.05) and
0
1
h
2
< 0.136.
signal dangerously high branching fractions for B
s
+
) vs. m
h
in Fig. 2.17. The recent
experimental limit from the CDF collaboration is that BF(B
s
+
) < 5.8 10
8
[34].
Thus, the entire band of A-resonance annihilation solutions becomes excluded! The
smattering of DM-allowed dots below the CDF limit all occur with DM annihilation via
the h-resonance. However, we may still want to consider A-resonance solutions in case they
are somehow allowed perhaps by additional avor-violating soft terms. This was done by
Baer et al [79].
At this point, it is useful to compare the Isajet SUSY spectral solutions to those generated
by Dermisek et al. in Ref. [13] and [14]. In Fig. 2.18, we plot the Isajet 7.75 solutions in
the m
1/2
vs. plane for m
16
= 3 TeV, m
10
/m
16
= 1.3, A
0
/m
16
= 1.85, tan = 50.9 and
m
2
H
= 0.14, with m
A
= 500 GeV: i.e. corresponding closely to Fig. 1 of [13]. We plot
contours of R from 1.15 to 1.3. Also, the green-shaded regions give the WMAP-measured
relic density, while white-shaded regions give
0
1
h
2
< 0.095, and pink-shaded regions give
44
Figure 2.17: Plot of MCMC results using WSH boundary conditions in the m
h
vs.
BF(B
s
+
) plane; the light-blue (dark-blue) points have R < 1.1 (1.05), while for
the orange (red) points R < 1.1 (1.05) and
0
1
h
2
< 0.136.
0
1
h
2
> 0.13 (as in Dermisek et al.). The LEP2 constraint on m
f
W
1
is indicated by the solid
contour at low m
1/2
and low . We see qualitatively the same shape to the DM-allowed
regions as generated by Dermisek et al.: the thick green regions are DM-allowed either by
A-resonance annihilation at large , or by mixed higgsino DM annihilation at low . There
is also a light Higgs h-resonance solution at m
1/2
120 GeV.
A notable feature of Fig. 2.18 is that over much of the DM-allowed region, the Yukawa
unication has R > 1.2.
7
As we move to larger values and lower m
1/2
values, the Yukawa
unication gets better and better. Most of the region with R < 1.15 is DM-forbidden,
save for the upper part of the light h-resonance solution. In fact, now we can see why
our compromise solution (point D) works and why it is so hard to nd using a top-down
approach: only the very narrow upper tip is both DM-allowed, and has a low R value.
7
Note that although the general features in Fig. 2.18 here and Fig. 1 of [13] are similar, the latter results
were obtained in a top-down t to low energy obervables assuming exact Yukawa unication, which is a
dierent approach then the one followed here. Moreover, there are several important dierences in the level
of sophistication of the spectrum computations between Ref. [46, 13, 14] and the study presented here. For
instance, Ref. [46, 13, 14] has only 1-loop RGE running of the SUSY-breaking parameters, takes sparticle
masses to be running masses at scale Q = M
Z
; ISAJET 7.75 applies full 2-loop running plus 1-loop threshold
corrections.
45
Figure 2.18: Contours of R and DM-allowed regions in the m
1/2
vs. parameter space for
m
16
= 3 TeV, m
10
/m
16
= 1.3, A
0
/m
16
= 1.85, m
H
= 0.14, tan = 50.9, m
A
= 500 GeV
and m
t
= 173.9 GeV, as in Dermisek et al., but using Isajet 7.75 for mass spectra generation.
2.4 Yukawa-unied benchmark scenarios and LHC
signatures
We have assembled in Table 2.1 ve Yukawa-unied benchmark scenarios that yield the
correct relic abundance of dark matter in ve dierent ways. With the LHC turn-on being
imminent, it is fruitful to examine what each of these ve scenarios implies for new physics
signatures.
At the bottom of Table 2.1 we list
0
1
h
2
, BF(b s), BF(B
s
+
), a
and spin-
independent neutralino-proton direct DM detection cross section (
0
1
p). For the rst four
of these numbers, we list output from IsaReD/Isatools (upper) and micrOMEGAs (lower).
While the results for the low-energy constraints agree fairly well, there is almost a factor of
2 dierence in the relic density when the neutralino dominantly annihilates through h or A
exchange (points A, D, E). This is due to dierences in the treatment of the Higgs resonance.
For example, IsaReD in Isajet 7.75 uses Yukawa couplings evaluated at scale Q =
t
L
m
t
R
for annihilation through the A resonance and for evaluation of the heavy Higgs widths, while
micrOMEGAs uses an eective Lagrangian approach and Q = 2m
0
1
.
8
This section serves
8
A complete discussion of the details of the calculations in the two programs is beyond the scope of this
46
as a discussion of a few various scenarios. Actual collider studies of these scenarios will be
saved for Chapter 4
Point A
Point A of Table 2.1 is a generic Yukawa-unied model with rst and second generation
scalar masses 9 TeV, so they essentially decouple from LHC physics. Third generation
and heavy Higgs scalars have masses at the 23 TeV level, while the lightest charginos,
neutralinos and gluinos all have masses in the range 100400 GeV. Since
0
1
h
2
400, we
postulate that the neutralino
0
1
is in fact an NLSP, decaying to a with a lifetime of order
0.03 seconds. In this case, the mean decay distance of a
0
1
will be of order 10
4
km. Thus,
the
0
1
will still escape the LHC detectors, leading to missing energy signatures (although it
is conceivable some may decay occassionally within the detector).
The LHC SUSY events will consist of a hard and soft component [70]. The hard
component comes from pair production of 400 GeV gluinos. The gluinos decay via 3-
body modes dominantly via g tb
W
1
, b
b
0
1
and especially b
b
0
2
[35]. The g g production
cross section is of order 10
5
fb at LHC, so we might expect 10
7
gluino pair events per 100
fb
1
of integrated luminosity. After cascade decays, we expect an assortment of events
with high jet and b-jet multiplicity, plus an assortment of isolated leptons. The
0
2
0
1
e e
branching fraction is at 2.2% , which should be enough to reconstruct the dilepton mass
edge at m
0
2
m
0
1
73 GeV. Correct pairing of b-jets and/or b-jets with isolated leptons,
plus the total event rate, should allow for an extraction of the gluino mass.
The soft component of signal will come from
+
1
1
,
W
1
0
2
and
W
1
0
1
production. These
events will be followed by 3-body decays to various nal states, but since the visible
components of the signal are much softer than that from gluino pair production, these events
will be harder to see above SM background levels. With judicious cuts, the soft component
might also be visible at some level (e.g.
W
1
0
2
3 +E
miss
T
) [36].
Point B
Point B is the same as point A, except that in this case the gaugino mass M
1
has been raised
to 195 GeV so that the
W
1
0
1
mass gap shrinks to only 13 GeV. Since is quite large,
dissertation; we refer the interested reader to the respective manuals.
47
Table 2.1: Masses and parameters in GeV units for ve benchmark Yukawa unied points
using Isajet 7.75 and m
t
= 171.0 GeV. The upper entry for the
0
1
h
2
etc. come from
IsaReD/Isatools, while the lower entry comes from micrOMEGAs; (
0
1
p) is computed with
Isatools.
parameter A B C D E
m
16
9202.9 9202.9 5018.8 2976.5 5877.3
m
1/2
62.5 62.5 160 107.0 113.6
A
0
19964.5 19964.5 10624.2 6060.3 12052.6
m
10
10966.1 10966.1 6082.1 3787.9
tan 49.1 49.1 47.8 49.05 47.4
M
D
3504.4 3504.4 1530.1 1020.8
M
1
195
m
16
(1, 2) 603.8
f
t
0.51 0.51 0.49 0.48 0.49
f
b
0.51 0.51 0.41 0.47 0.49
f
t
1
2315.1 2310.5 1048.4 434.5 944.7
m
b
1
2723.1 2714.9 1894.0 849.3 1452.7
m
e
L
9131.9 9132.0 311.9 2955.8 5833.6
m
e
R
9323.7 9323.9 891.8 3009.0 5945.8
m
f
W
1
128.8 128.8 165.7 105.7 141.3
m
0
2
128.6 128.1 165.1 105.1 140.9
m
0
1
55.6 115.9 80.2 52.6 65.7
m
A
3273.6 3266.0 1939.9 776.8 177.8
m
h
125.4 125.4 123.2 111.1 113.4
0
1
h
2 423
220
0.09
0.08
0.11
0.11
0.10
0.06
0.15
0.08
BF(b s)
3.010
4
3.310
4
3.010
4
3.310
4
6.210
4
3.710
4
1.910
4
4.010
4
2.510
4
2.210
4
a
5.010
12
5.110
12
5.010
12
5.010
12
3.010
10
2.810
10
2.210
10
2.210
10
4.110
11
4.110
11
BF(B
s
+
)
5.010
9
4.410
9
5.010
9
4.410
9
11.810
9
6.910
9
5.810
8
6.210
8
2.010
5
2.010
5
sc
(
0
1
p) [pb] 1.3 10
15
1.9 10
17
1.5 10
6
2.7 10
9
5.3 10
8
48
the
0
1
remains nearly pure bino-like, but the relic density problem is solved via bino-wino
co-annihilation. This case will again give a hard component to the LHC new physics signal
from gluino pair production, but this time the m(
+
W
1
, we note that
0
2
u u
R
and c c
R
is also large, leading again to relatively soft jet activity.
In spite of the soft jet activity, the scenario should be easily seen at LHC, since q
L
q
W
1
occurs at a large rate, and
W
1
e
e
0
1
occurs at 43% branching fraction. This can lead to a
large same-sign dilepton rate from pp u
L
u
L
production, along with a large asymmetry in
++ SS dileptons over SS dileptons (which occur from
d
L
d
L
production). This scenario
may also be subject to exclusion by analysis of Fermilab Tevatron data. We further note that
point C is naively excluded by direct dark matter search limits. These latter limits depend
on an assumed standard local relic density mass and velocity distribution, so that the limits
can be avoided if one postulates that we live in a local underdensity of dark matter.
49
Point D
Point D is an example of a compromise solution, where we allow m
16
as low as 3 TeV at
some expense to Yukawa unication (here, Yukawa unication is good to only 10%) in
order to allow for neutralino annihilation through the light Higgs h-resonance (neutralinos
can still annihilate through the light h resonance for higher m
16
values; it is just that the relic
density cant be pushed as low as
0
1
h
2
0.1). This scenario is extremely predictive, with
gluinos around 350450 GeV, so again we expect LHC events to be dominated by gluino pair
production. As in the case of point A, the g g events will be followed by 3-body decays to
b-jet rich nal states. A dilepton mass edge at m
0
2
m
0
1
53 GeV should be visible since
0
2
0
1
e
+
e
W
1
with a 50% branching fraction, and also has signicant
branching fractions to t
0
1
, t
0
2
and b
W
2
nal states. The
b
1
dominantly decays to b g and
W
t
1
nal states. Moreover, the heavy Higgs bosons A
0
, H
0
and H
and b
CDM
h
2
CDM
/
c
= 0.111
+0.011
0.015
(2), (3.1)
where h = 0.740.03 is the scaled Hubble constant. While the density of CDM is becoming
precisely known, the identity of the CDM particle (or particles) is still a complete mystery.
Although numerous candidate CDM particles populate the theoretical literature, the WIMPs
(weakly interacting massive particles) stand out in that their thermal abundance can be
calculated, and is found to be in rough accord with Eq. (3.1) provided the WIMP mass is
of order 1001000 GeV. Of the numerous WIMP candidates in the literature, the lightest
neutralino of supersymmetric (SUSY) theories is especially popular because SUSY solves a
host of theoretical problems associated with the SM, and also receives some (albeit indirect)
support from data (in the form of the measured gauge couplings unifying at Q = M
GUT
under MSSM RG evolution and also from other precision electroweak measurements[?]).
There is at present a multi-pronged eort aimed at identifying WIMP dark matter
particles and measuring their properties[?]. The most direct approach is to try to detect relic
WIMPs left over from the Big Bang by observing WIMP-nucleon collisions in experiments
located deep underground. Limits from the CDMS[?] experiment and more recently from
the Xenon-10[?] experiment have begun probing the upper limits of SUSY model parameter
51
space.
WIMP particles can also be searched for at collider experiments such as those at the
CERN LHC, especially if the dark matter particle is but one of a whole family of particles,
some of which can be produced via strong and electromagnetic interactions. The dark matter
particle would then be produced by cascade decays of heavier particles, and would lead to
missing transverse energy in collider events. Such is the case of theories such as R-parity
conserving supersymmetry[63], KK-parity conserving universal extra dimensions (UED)[?]
and little Higgs models with T-parity[?].
Dark matter may also be searched for indirectly. For instance, the sun can sweep up
WIMP particles as it traverses its galactic orbit, so that WIMPs accumulate at a high
density in the solar core. Then WIMP-WIMP annihilation to SM particles can occur at high
rates in the solar core. While most SM particles would be absorbed by the surrounding solar
medium, multi-GeV scale
G
ends up setting the mass scale for all
the soft breaking terms, so then all SSB terms end up also being of order the weak scale.
The coupling of the gravitino to matter is strongly suppressed by the Planck mass, so the
G is never in thermal equilibrium with the thermal bath in the early universe. Nonetheless, it
does get produced by scatterings of particles that do partake of thermal equilibrium. Thermal
production of gravitinos in the early universe has been calculated in Refs. [53], where the
abundance is found to depend naturally on m
G
and on the re-heat temperature T
R
at the
end of ination. Once produced, the
Gs decay into all varieties of particle-sparticle pairs,
but with a lifetime that can exceed 1 sec, the time scale where Big Bang nucleosynthesis
(BBN) begins. The energy injection from
G decays is a threat to dis-associate the light
54
element nuclei which are created in BBN. Thus, the long-lived
Gs can destroy the successful
predictions of the light element abundances as calculated by nuclear thermodynamics.
The BBN constraints on gravitino production in the early universe have been calculated
by several groups [54]. The recent results from Ref. [55] give an upper limit on the re-heat
temperature as a function of m
G
. See Fig. 3.1 by Kohri et al. displaying the gravitino lifetime
and reheating temperature in the mSUGRA model space. The results depend on how long-
lived the
G is (at what stage of BBN the energy is injected), and what its dominant decay
modes are. Qualitatively, for m
G
<
5 TeV, T
R
<
10
6
GeV is required; if this is violated, then
too many
G are produced in the early universe, which detroy the
3
He,
6
Li and D abundance
calculations. For m
G
550 TeV, the re-heat upper bound is much less: T
R
<
510
7
10
9
GeV (depending on the
4
He abundance) due to overproduction of
4
He arising from n p
conversions. For m
G
>
50 TeV, there is an upper bound of T
R
<
5 10
9
GeV due to
overproduction of
0
1
LSPs due to
G decays.
Figure 3.1: The gravitino problem in generic SUGRA models: an overproduction of
gravitinos followed by late gravitino decay can destroy successful BBN predictions upper
bound on reheating temperature.
Solutions to the gravitino BBN problem then include: (1) having m
G
>
50 TeV but with
an unstable
0
1
(no T
R
bound), (2) having a gravitino LSP so that
G is stable or (3) keep
the re-heat temperature below the BBN bounds. We will here adopt solution number (3).
55
In the case of SO(10) SUSY GUT models, with m
G
m
16
5 20 TeV, this means we
need a re-heat temperature T
R
<
10
8
10
9
GeV.
3.2.2 Non-thermal leptogenesis
The data gleaned on neutrino masses during the past decade has lead credence to a particular
mechanism of generating the baryon asymmetry of the universe known as leptogenesis[56].
Leptogenesis requires the presence of heavy gauge singlet Majorana right handed neutrino
states
N
c
i
( N
i
) with mass M
N
i
(i = 1 3 is a generation index). The N
i
states may be
produced thermally in the early universe, or perhaps non-thermally, as suggested in Ref.
[57] via inaton N
i
N
i
decay. The N
i
may then decay asymmetrically to elements of the
doublets for instance (N
1
h
+
u
e
) ,= (N
1
h
u
e
+
) owing to the contribution of CP
violating phases in the tree/loop decay interference terms. Focussing on just one species of
heavy neutrino N
1
, the asymmetry is calculated to be[58]
(N
1
+
) (N
1
N
1
3
8
M
N
1
v
2
u
m
eff
, (3.2)
where m
3
is the heaviest active neutrino, v
u
is the up-Higgs vev, and
eff
is an eective
CP-violating phase factor which may be of order 1. The ultimate baryon asymmetry of the
universe is proportional to , so larger values of M
N
1
lead to a higher baryon asymmetry.
To nd the baryon asymmetry, one may rst assume that the N
1
is thermally produced
in the early universe, and then solve the Boltzmann equations for the BL asymmetry. The
ultimate baryon asymmetry of the universe arises from the lepton asymmetry via sphaleron
eects. The nal answer[59], compared against the WMAP-measured result
n
B
s
0.910
10
for the baryon-to-entropy ratio, requires M
N
1
>
10
10
GeV, and thus a re-heat temperature
T
R
>
10
10
GeV. This high a value of reheat temperature is in conict with the upper bound
on T
R
discussed in Sec. 3.2.1. In this way, it is found that generic SUGRA models are
apparently in conict with leptogenesis as a means to generate the baryon asymmetry of the
universe.
If one instead looks to non-thermal leptogenesis, then it is possible to have lower reheat
temperatures, since the N
1
may be generated via inaton decay. The Boltzmann equations
for the BL asymmetry have been solved numerically in Ref. [60]. The BL asymmetry is
then converted to a baryon asymmetry via sphaleron eects as usual. The baryon-to-entropy
56
ratio is calculated in [60], wherein it is found
n
B
s
8.2 10
11
_
T
R
10
6
GeV
__
2M
N
1
m
_
_
m
3
0.05 eV
_
eff
, (3.3)
where m
NTP
a
h
2
=
m
a
m
0
1
0
1
h
2
, (3.4)
since in this case the axinos inherit the thermally produced neutralino number density. Notice
that neutralino-to-axino decay oers a mechanism to shed large factors of relic density. For
a case where m
0
1
50 GeV and
0
1
h
2
1000, as can occur in SO(10) SUSY GUTs, an
axino mass of less than 5 MeV reduces the DM abundance to below WMAP-measured levels.
The lifetime for these decays has been calculated, and it is typically in the range of
(
0
1
a) 0.03 s [52]. The photon energy injection from
0
1
a decay into the cosmic
soup occurs well before BBN, thus avoiding the constraints that plague the case of a gravitino
57
LSP [61]. The axino DM arising from neutralino decay is generally considered warm or even
hot dark matter for cases with m
a
<
1 GeV [92].
Even though they are not in thermal equilibrium, axinos can still be produced thermally
in the early universe via scattering processes. The axino thermally produced (TP) relic
abundance has been calculated in Refs. [52, 62], and is given by
TP
a
h
2
5.5g
6
s
ln
_
1.108
g
s
__
10
11
GeV
f
a
/N
_
2
_
m
a
0.1 GeV
_
_
T
R
10
4
GeV
_
, (3.5)
where g
s
is the strong coupling evaluated at Q = T
R
and N is the model dependent color
anomaly of the PQ symmetry, of order 1. The thermally produced axinos qualify as cold
dark matter as long as m
a
>
0.1 MeV [52, 62].
3.2.4 A consistent cosmology for axino DM from SO(10) SUSY
GUTs
At this point, we are able to check if we can implement a consistent cosmology for SO(10)
SUSY GUTs with axino dark matter. Our rst step is to select points from the SO(10)
parameter space Eq. 2.2 that are very nearly Yukawa-unied. In Ref. [79], Yukawa unied
solutions were searched for by looking for R values as close to 1 as possible, where recall
R =
max(f
t
, f
b
, f
)
min(f
t
, f
b
, f
)
(3.6)
and the f
t
, f
b
and f
NTP
a
h
2
= 0.01 according to Eq. 3.4. In step 3, we plug m
a
into Eq. 3.5, where we also
take g
s
= 0.915 (the running g
s
value at 10
6
GeV), and PQ scale f
a
/N = 10
12
GeV. By
insisting that
TP
a
= 0.1, we may calculate the value of T
R
that is needed.
Our results are plotted in the m
a
vs. T
R
plane in Fig. 3.2 and occupy the upper band
of solutions. In this plane, solutions with T
R
<
3 10
7
5 10
8
GeV are allowed by the
gravitino constraint (with m
G
5 20 TeV) and BBN. Solutions with T
R
>
10
6
GeV can
generate the matter-antimatter asymmetry correctly via non-thermal leptogenesis. Solutions
58
with m
a
>
10
4
GeV give dominantly cold DM from TP of axinos. Solutions with m
16
> 15
TeV are denoted by lled (turquoise) symbols, while solutions with m
16
< 15 TeV have open
(dark blue) symbols.
1e-05 0.0001 0.001 0.01
m
a
~
(GeV)
1e+06
1e+07
1e+08
1e+09
T
R
(
G
e
V
)
BBN/gravitino
NT leptogenesis
warm a
~
DM
Figure 3.2: Plot of Yukawa unied solutions with R < 1.05 and 5 TeV < m
16
< 20 TeV in the
m
a
vs.T
R
plane. The upper band of solutions has
NTP
a
h
2
= 0.01,
TP
a
h
2
= 0.10 and f
a
/N = 10
12
GeV, while the lower band of solutions has
NTP
a
h
2
= 0.03,
TP
a
h
2
= 0.08 and f
a
/N = 5 10
11
GeV.
We see that a variety of points fall in the allowed region. These points give rise to a
consistent cosmology for SO(10) SUSY GUT models! Of course, there is some uncertainty
in these results. We can take higher or lower values of the PQ breaking scale, higher or
lower fractions of
NTP
a
, and the T
R
upper (and lower) bounds have some variability built
into them. As an example, the lower band of solutions is obtained with
NTP
a
= 0.03,
TP
a
h
2
= 0.08 and f
a
/N = 5 10
11
GeV. In this case, some of the previously excluded
solutions migrate into the allowed region to give a consistent cosmology with somewhat
dierent parameters.
59
CHAPTER 4
Collider Searches for New Physics
It is expected that the CERN Large Hadron Collider (LHC), a
s = 14 TeV pp collider,
will begin operation in late 2008 or early 2009. It is not unreasonable to expect of order 0.1
fb
1
of integrated luminosity in the rst full year of running. One of the main goals of the
LHC is to either discover or exclude the existence of weak scale supersymmetry[63].
Most theories of weak scale SUSY have an added R-parity invariance which is necessary
to stabilize the proton against rapid decay through R-violating interactions. A consequence
of R-parity conservation is that superpartners of SM particles must decay to other superpart-
ners. In this case, the lightest SUSY particle (LSP) must be absolutely stable. If produced
in the early universe, then there should exist relic LSPs in the universe today, and in fact it
is popular to conjecture that these might make up the required cold dark matter (CDM) in
the universe. Null searches for massive charged or colored relics from the Big Bang indicate
that the LSP must be electrically and color neutral. In many models, the lightest neutralino
(
0
1
or
0
1
) turns out to be the LSP, and is an excellent candidate CDM particle. A neutralino
LSP, if produced in a collider experiment, would escape detection and thus provide a signal
characterized by an apparent non-conservation of (transverse) energy.
It was recognized early on that perhaps the classic signature for production of SUSY
particles in collider events is the presence of an excess of E
miss
T
+ jets events above SM
background
1
. Thus, most studies of sparticle discovery at collider experiments rely on the
presence of large E
miss
T
in the events in order to reject SM backgrounds such as multi-jet
production in QCD. At LHC, many analyses require for instance E
miss
T
>
100 GeV as a
minimum requirement[65].
1
Indeed, it is suggested in Ref. [64] that SUSY gives rise to so-called zen events: jets balanced by
E
miss
T
, which correspond to the sound of one hand clapping.
60
Even if the neutralino is not the LSP but is instead perhaps the axino, the SUSY
signatures would still be characterized by missing E
T
, and the above description still holds.
If the neutralino in the neutralino decay to axino plus photon were not so long lived, there
would be a possibility of the electromagnetic calorimeters of the LHC to capture these and
an implied discovery via direct detection could be made.
4.1 Early SUSY Discovery using Multi-leptons
From the experimental side, the requirement of large E
miss
T
can be problematic, especially if
an early discovery of SUSY is desired. Missing transverse energy can arise not only from the
presence of weakly interacting neutral particles such as neutrinos or the lightest neutralinos,
but also from a variety of other sources, including:
energy loss from cracks and un-instrumented regions of the detector,
energy loss from dead cells,
hot cells in the calorimeter that report an energy deposition even if there isnt one,
mis-measurement in the electromagnetic calorimeters, hadronic calorimeters or muon
detectors and
the presence of mis-identied cosmic rays in events.
Thus, in order to have a solid grasp of expected E
miss
T
from SM background processes, it will
be necessary to have detailed knowledge of the complete detector performance. As experience
at the Tevatron suggests, this complicated task may well take some time to complete. The
same is likely to be true at the LHC, as many SM processes will have to be scrutinized rst
in order to properly calibrate the detector. For this reason, SUSY searches using E
miss
T
as a
crucial requirement may well take rather longer than a year to provide reliable results.
On the other hand, if SUSY particles are relatively light, then production cross sections
can be huge, and many new physics events may be generated in the rst few months of
running. For instance, for m
g
400 GeV and heavy squarks, the expected gluino pair cross
sections are in the 10
5
fb range. If m
g
m
q
400 GeV, then production cross sections are
even higher: of order 10
6
fb! Thus, with just 0.1 fb
1
of integrated luminosity, we might
61
expect of order 10
4
10
5
new physics events to be recorded on tape if the gluino is in the
400 GeV range.
In this chapter, we wish to examine if an early SUSY discovery might be made without
using E
miss
T
cuts. The key is to take advantage of the large production cross sections of
strongly interacting SUSY particles (the gluinos and squarks) and their complex cascade
decays. Gluinos and squarks generally decay through a multi-step cascade of decays[98]
until the LSP state is reached, so that SUSY signal events are expected to be rich in
jet multiplicity, b-jet multiplicity, isolated lepton multiplicity and sometimes large tau-jet
multiplicity. In addition, gauge mediated SUSY can lead to collider events with high isolated
photon multiplicity. Thus, we would like to be able to use detected objects such as jets, b-jets
and isolated leptons to maximize signal over background, rather than inferred objects like
E
miss
T
which requires a complete detector knowledge. Our main result in this chapter is that
we nd a substantial reach for SUSY at the LHC by requiring multi-jet plus multi-lepton
events, without requiring the presence of E
miss
T
.
2
By searching in this channel, one may be
able to discover SUSY even before the detectors are fully calibrated such that E
miss
T
is a
useful variable for background rejection.
We use Isajet 7.76[83] for the simulation of signal and background events at the LHC.
A toy detector simulation is employed with calorimeter cell size = 0.05 0.05
and 5 < < 5. The HCAL energy resolution is taken to be 80%/
E + 0.5%. We use a UA1-like jet nding algorithm with jet cone size R = 0.4 and
require that E
T
(jet) > 50 GeV and [(jet)[ < 3.0. Leptons are considered isolated if they
have p
T
(e or ) > 20 GeV and [[ < 2.5 with visible activity within a cone of R < 0.2 of
E
cells
T
< 5 GeV. The strict isolation criterion helps reduce multi-lepton backgrounds from
heavy quark (c c and b
b) production.
We identify a hadronic cluster with E
T
> 50 GeV and [(j)[ < 1.5 as a b-jet if it contains
a B hadron with p
T
(B) > 15 GeV and [(B)[ < 3 within a cone of R < 0.5 about the jet
axis. We adopt a b-jet tagging eciency of 60%, and assume that light quark and gluon jets
can be mis-tagged as b-jets with a probability 1/150 for E
T
100 GeV, 1/50 for E
T
250
GeV, with a linear interpolation for 100 GeV< E
T
< 250 GeV[101].
2
Similar signal calculations for models with a charged stable LSP have been performed in Ref. [67].
62
For our initial analysis, we adopt the well-studied SPS1a
t(171) and WW, WZ, ZZ channels at the rates shown in Table 4.3. The W + jets and
Z + jets backgrounds use exact matrix elements for one parton emission, but rely on the
parton shower for subsequent emissions.
We begin by applying a set of pre-cuts to our event samples. These cuts, known as set
C1 in Ref. [70], were used for studying gluino mass determination in the focus point region
of mSUGRA. Here, we abandon the E
miss
T
> (100 GeV, 0.2M
eff
) cut and call the new set of
cuts C1
.
C1
cuts:
n(jets) 4, (4.1)
E
T
(j1, j2, j3, j4) 100, 50, 50, 50 GeV, (4.2)
S
T
0.2. (4.3)
We will also make use of the augmented eective mass A
T
= E
miss
T
+
jets
E
T
(j) +
leptons
E
T
(). Here, stands for either e or . If we remove E
miss
T
from A
T
, we will
63
Table 4.1: Events generated and cross sections for various SM background processes plus
the SPS1a
+ 3 (fb)
QCD (p
T
: 50 100 GeV) 10
6
2.6 10
10
QCD (p
T
: 100 200 GeV) 10
6
1.5 10
9
QCD (p
T
: 200 400 GeV) 10
6
7.3 10
7
QCD (p
T
: 400 1000 GeV) 10
6
2.7 10
6
QCD (p
T
: 1000 2400 GeV) 10
6
1.5 10
4
W +jets; W e, , (p
T
(W) : 100 4000 GeV) 5 10
5
3.9 10
5
0.8
Z +jets; Z , s (p
T
(Z) : 100 3000 GeV) 5 10
5
1.4 10
5
0.3
t
t 3 10
6
5.1 10
5
5.1
WW, ZZ, WZ 5 10
5
8.0 10
4
signal (SPS1a
: m
g
= 608 GeV) 2.5 10
5
4.7 10
4
46.6
call the new variable A
T
. S
T
is transverse sphericity
3
. The event rates in fb are listed before
cuts in column 3 of Table 4.3.
In Fig. 4.1, we plot the resulting jet multiplicity n
j
after cuts C1
benchmark (orange histogram) along with the various SM backgrounds. The gray histogram
gives the sum of all SM backgrounds. We see immediately that SM background, dominated
by QCD multi-jet production, dominates out to very high jet multiplicity.
In Fig. 4.2, we plot the augmented eective mass (minus the E
miss
T
component) A
T
. When
the E
miss
T
cut is used in cut set C1, then signal generally emerges from background at some
large value of A
T
which is somewhat correlated with the values of m
g
and m
q
[71]. In this
case, with no E
miss
T
cut, the signal is hopelessly below the summed BG distribution.
3
Sphericity is dened, e.g. in Collider Physics, V. Barger and R. J. N. Phillips (Addison Wesley, 1987).
Here, we restrict its construction to using only transverse quantities, as is appropriate for a hadron collider.
64
5 10 15
n
j
0.01
0.1
1
10
100
1000
10000
1e+05
1e+06
1e+07
(
f
b
)
SPS1a(70,250,-300,10,1,171)
QCD Jets
tt
W+jets
Z+jets
WW, WZ, ZZ
Sum of Backgrounds
No. of Jets
Cuts C1
Figure 4.1: Plot of jet multiplicity from SUSY collider events from SPS1a
after cuts C1
.
We also plot the histograms of various SM backgrounds, plus the total SM background (gray
histogram).
0 1000 2000 3000
A
T
(GeV)
0.01
0.1
1
10
100
1000
10000
1e+05
d
/
d
A
T
(
f
b
/
G
e
V
)
SPS1a(70,250,-300,10,1,171)
QCD jets
tt
W+jets
Z+jets
WW, WZ, ZZ
Sum of Backgrounds
Augmented Transverse M
eff
Cuts C1
Figure 4.2: Plot of augmented eective mass A
T
(without E
miss
T
) from SUSY collider events
from SPS1a
after cuts C1
(
f
b
)
SPS1a(70,250,-300,10,1,171)
QCD Jets
tt
W+jets
Z+jets
WW, WZ, ZZ
Sum of Backgrounds
No. of b-jets
Cuts C1
Figure 4.3: Plot of b-jet multiplicity n
b
from LHC SUSY events from SPS1a
after cuts C1
.
We also plot the histograms of various SM backgrounds, plus the total SM background (gray
histogram).
In Fig. 4.3, we plot the multiplicity of tagged b-jets n
b
in events after cuts C1
. We see
that out to n
b
= 5, SM background from QCD jet production including both b
b production,
parton shower production from g b
= 0 or 1, signal is dominated by
BG. However, at n
= 3, SM background is well below signal. In this case, it is clear that we can gain
good BG rejection by requiring the cut set C1
, plus n
t
production via its one and two lepton signatures should allow for a solid calibration of this
most important background.
To gain an estimate of the LHC reach using cuts C1
(
f
b
)
SPS1a(70,250,-300,10,1,171)
QCD Jets
tt
W+jets
Z+jets
WW, WZ, ZZ
Sum of Backgrounds
No. of Isolated Leptons
Cuts C1
Figure 4.4: Plot of isolated lepton multiplicity n
after cuts C1
is suppressed
due to destructive interference in the Z and slepton mediated decay processes. In the high
m
0
case, 3-body decay of
0
2
via the Z
3
l
(
f
b
)
m
0
= 200 GeV
m
0
= 1000 GeV
m
0
= 70, A
0
= -300
Cuts C1a
m
t
=171.0, A
0
=0, tan=10, sgn>0
5 level at 0.1 fb
-1
5 level at 1 fb
-1
Figure 4.5: Plot of signal cross section from mSUGRA model versus m
g
after cuts C1
and
n
3, for m
0
= 200 and 1000 GeV. We also take A
0
= 0, tan = 10, > 0 and m
t
= 171
GeV. We also plot the 5 background level for 0.1 and 1 fb
1
of integrated luminosity.
One possible criticism of our results so far is that we use only leading order cross sections
as calculated by Isajet. However, we expect that NLO total cross sections for both signal
and background to be somewhat enhanced beyond the LO Isajet results, so we would expect
our overall conclusion to remain valid qualitatively. Indeed, it is expected that the major
SM processes will be measured to high accuracy at LHC already at low luminosity, so that a
good background calibration should be at hand. A second criticism could be that there are
additional background processes to be checked. These would include 2 n processes such as
t
tt
t, t
tV , t
T
will allow signal to emerge
68
from t
(GeV)
0.001
0.01
d
/
d
A
T
(
f
b
/
G
e
V
)
SPS1a(70,250,-300,10,1,171)
QCD jets
tt
W+jets
Z+jets
WW, WZ, ZZ
Sum of Backgrounds
Augmented Transverse M
eff
Cuts C1 + 3l + 1b-jet
Figure 4.6: Distribution in variable A
T
from SUSY events from SPS1a
after cuts C1
plus
3 plus 1 b-jet. We also plot the remaining SM backgrounds (gray histogram).
We also note here that if a SUSY signal is found in the 4 jets plus 3 sample,
then the resulting event sample may be used for precision sparticle mass measurements
just as in the case where one requires jets +E
miss
T
. As an example, we examine all events
passing cuts C1
) = m
0
2
_
1
m
2
m
2
0
2
_
1
m
2
0
1
m
2
R
= 123.3 GeV
and m
0
1
= 97.8 GeV). The mass edge is evident from the plot, and serves as a starting point
for further sparticle mass reconstruction.
69
50 100 150
m
ll
(GeV)
0
1
2
3
4
d
/
d
m
l
l
(
f
b
/
G
e
V
)
SPS1a(70,250,-300,10,1,171)
QCD Jets
tt
W+jets
Z+jets
WW, WZ, ZZ
Sum of Backgrounds
Cuts C1 + 3l
m
ll
max
Figure 4.7: Plot of OS/SF dilepton invariant mass from SUSY events from SPS1a
after cuts
C1
/
d
m
l
l
(
f
b
/
G
e
V
)
SPS1a + BG
QCD Jets
tt
W+jets
Z+jets
WW, WZ, ZZ
Sum of Backgrounds
Cuts C1a
Figure 4.8: Plot of OS/SF dilepton invariant mass from SUSY events from benchmark SPS1a
after cuts C1
0
1
. In this case, we require cuts C1
and SM
BG. We see a continuum of background, along with a Z peak. The BG Z peak arises because
Isajet includes W and Z radiation in its parton shower algorithm. The orange histogram
shows the sum of signal plus BG, and the OS/SF dilepton mass edge clearly stands out
below the Z peak. A further example comes from SO(10) benchmark point A suggested in
Ref. [79], which includes a 400 GeV gluino. I n this case, the large g g cross section allows
signal to stand out even more abruptly from SM background.
4.2 Yukawa-unied SO(10) at the Cern LHC
In this section, we wish the explore consequences of Yukawa-unied SUSY models for
sparticle detection at the LHC. We focus most of our attention on the two cases presented
in Table 1 of Ref. [79]: 1. point A with m
16
9 GeV and an axino LSP, and 2. point D with
m
16
3 TeV and a neutralino LSP. While our studies here focus on just two cases, we feel
the qualitative features of the LHC signatures should be rather similar to these two cases.
In fact, the collider phenomenology of these cases is rather similar between the two, since in
the rst case the neutralino decays to an axino far beyond the detector boundaries. Thus,
in both cases the lightest neutralino
0
1
leads to missing E
T
at collider experiments.
For the benet of the reader, we present in Table 4.2 the two case studies we examine. We
present the parameter space values, sparticle mass spectrum, and in addition the total tree-
level LHC sparticle production cross section. We also list as percentages some contributing
2 2 subprocess reactions.
4.2.1 Cross sections and branching fractions for sparticles in
Yukawa-unied models
Given the characteristic spectrum of superpartners obtained in Yukawa-unied SUSY models,
it is useful to examine what sort of new physics signals we would expect at the LHC.
Obviously, rst/second generation squarks and sleptons in the multi-TeV mass range will
71
Table 4.2: Masses and parameters in GeV units for two cases studies points A and D of
Ref. [79] using Isajet 7.75 with m
t
= 171.0 GeV. We also list the total tree level sparticle
production cross section in fb at the LHC, plus the percent for several two-body nal states.
parameter Pt. A Pt. D
m
16
9202.9 2976.5
m
1/2
62.5 107.0
A
0
19964.5 6060.3
m
10
10966.1 3787.9
tan 49.1 49.05
M
D
3504.4 1020.8
f
t
0.51 0.48
f
b
0.51 0.47
f
0.52 0.52
4179.8 331.0
m
g
395.6 387.7
m
u
L
9185.4 2970.8
m
t
1
2315.1 434.5
m
b
1
2723.1 849.3
m
e
L
9131.9 2955.8
m
1
128.8 105.7
m
0
2
128.6 105.1
m
0
1
55.6 52.6
m
A
3273.6 776.8
m
h
125.4 111.1
[fb] 75579.1 89666.1
% ( g g) 86.8 80.5
% (
1
0
2
) 8.8 12.8
% (
t
1
t
1
) 0 1.1
essentially decouple from LHC physics. Gluinos in the 350-500 GeV range will be
produced in abundance via q q and gg fusion subprocesses. Charginos and neutralinos, being
in the 100160 GeV range, may also be produced with observable cross sections.
As noted above, we list the tree-level total sparticle production cross sections obtained
from Isajet for cases A and D in Table 4.2. In case A, we nd (tot) 810
4
fb, so that 8000
sparticle pair events are expected at LHC with just 0.1 fb
1
of integrated luminosity. Of this
total, 86.7% comes from gluino pair production, while 8.8% comes from
W
1
0
2
production
and 4.5% comes from
W
+
1
1
production. In case D, (tot) 9.6 10
4
fb, with 80.4%
72
from g g production, 12.8% from
W
1
0
2
production, 6.4% from
W
+
1
1
production while
top-squark pair production yields just 1.1% of the total. Given these production cross
sections, we expect Yukawa-unied SUSY to yield primarily g g events at the LHC. Gluino
pair production typically leads to events with hard jets, hard E
miss
T
and isolated leptons from
the gluino cascade decays[98]. We also expect a soft component coming from
W
+
1
1
and
1
0
2
production. While both these reactions lead to events with rather soft jets, leptons
and E
miss
T
, the latter reaction can also yield clean trilepton events[?], which might be visible
at LHC above SM backgrounds.
For the case of gluino masses other than those listed in Table 4.2, we show in Fig. 4.9
the total gluino pair production rate versus m
g
at the LHC at tree level (solid) and next-
to-leading-order (NLO) using the Prospino program[100]. The scale choice is taken to be
Q = m
g
. We take m
q
to be 3 TeV (blue) and 9 TeV (red). As can be seen, the results
hardly vary between this range of squark masses. The tree level results agree well with
Isajet, but the NLO results typically show an enhancement by a factor 1.6. Thus, we
expect Yukawa-unied SUSY models to yield pp g gX events at a 30-150 pb level at LHC.
In Fig. 4.10, we show the total -ino pair production cross sections versus chargino mass
m
f
W
1
. While
1
0
2
and
W
+
1
1
production dominate, and have rates around 10
3
10
4
fb over
the range of interest, there exists a sub-dominant rate for
W
1
0
1
and also
0
1
0
2
production.
Now that we see that Yukawa-unied SUSY will yield dominantly gluino pair production
events at the LHC, we next turn to the gluino branching fractions in order to understand their
event signatures. All sparticle branching fractions are calculated with Isajet 7.75. In Fig.
4.11, we show various gluino branching fractions for points A and D. We see immediately that
in both cases, BF( g b
b
0
2
) dominates at around 56%. This is followed by BF( g b
b
0
1
) at
16%, and BF( g b
W
+
1
) and BF( g t
1
) each at 10%. Decays to rst and second
generation quarks are much suppressed due to the large rst and second generation squark
masses. From these results, we expect gluino pair production events to be rich in b-jets,
E
miss
T
and occassional isolated leptons from the leptonic decays
W
1
0
1
and
0
2
0
1
,
where = e or .
73
360 380 400 420 440 460 480 500
m
g
50
100
150
200
p
p
p
b
~
1
(GeV)
10
-2
10
-1
10
0
10
1
10
2
10
3
10
4
(
f
b
)
~
1
+
~
1
~
1
~
1
0
~
1
~
2
0
~
1
0
~
2
0
Figure 4.10: Plot of various -ino pair production processes in fb at
s = 14 TeV versus
m
1
, for m
q
= 3 TeV and = m
g
, with tan = 49 and > 0.
74
0.0001
0.001
0.01
0.1
1
g ~
B
r
a
n
c
h
i
n
g
F
r
a
c
t
i
o
n
s
Point A Point D
bb
2
0
g
2
0
uu
2
0
bb
1
0
uu
1
0
dd
1
0
g
1
0
bt
du
bb
2
0
bb
1
0
bt
2
0
g
1
0
uu
1
0
dd
1
0
uu
2
0
dd
2
0
du
3
0
g
4
0
Figure 4.11: Plot of various sparticle branching fractions taken from Isajet for points A and
D from Table 4.2.
4.2.2 Gluino pair production signals at the LHC
To examine collider signals from Yukawa-unied SUSY at the LHC in more detail, we
generate 10
6
sparticle pair production events for points A and D, corresponding to 13 and
11 fb
1
of integrated luminosities. We use Isajet 7.75[83] for the simulation of signal and
background events at the LHC. A toy detector simulation is employed with calorimeter cell
size = 0.05 0.05 and 5 < < 5. The HCAL energy resolution is taken to be
80%/
and E
miss
T
cuts are specied in the text. The W+jets
and Z +jets background has been computed within the restriction p
T
(W, Z) > 100 GeV.
process events (fb) C1
C1
+E
miss
T
QCD (p
T
: 0.05 0.1 TeV) 10
6
2.6 10
10
4.1 10
5
QCD (p
T
: 0.1 0.2 TeV) 10
6
1.5 10
9
1.4 10
7
QCD (p
T
: 0.2 0.4 TeV) 10
6
7.3 10
7
6.5 10
6
2199
QCD (p
T
: 0.4 1.0 TeV) 10
6
2.7 10
6
2.8 10
5
1157
QCD (p
T
: 1 2.4 TeV) 10
6
1.5 10
4
1082 25
W
+jets 5 10
5
3.9 10
5
3850 1275
Z +jets 5 10
5
1.4 10
5
1358 652
t
t 3 10
6
4.9 10
5
8.2 10
4
2873
WW, ZZ, WZ 5 10
5
8.0 10
4
197 7
Total BG 9.5 10
6
2.76 10
10
2.13 10
7
8188
Point A: 10
6
7.6 10
4
3.6 10
4
8914
S/B 0.002 1.09
S/
S +B (1 fb
1
) 68
Point D: 10
6
9.0 10
4
3.7 10
4
10843
S/B 0.002 1.32
S/
S +B (1 fb
1
) 78
leptons with p
T
() > 50 GeV.
We identify a hadronic cluster with E
T
> 50 GeV and [(j)[ < 1.5 as a b-jet if it contains
a B hadron with p
T
(B) > 15 GeV and [(B)[ < 3 within a cone of R < 0.5 about the jet
axis. We adopt a b-jet tagging eciency of 60%, and assume that light quark and gluon jets
can be mis-tagged as b-jets with a probability 1/150 for E
T
100 GeV, 1/50 for E
T
250
GeV, with a linear interpolation for 100 GeV< E
T
< 250 GeV[101].
In addition to signal, we have generated background events using Isajet for QCD jet
production (jet-types include g, u, d, s, c and b quarks) over ve p
T
ranges as shown in
Table 4.3. Additional jets are generated via parton showering from the initial and nal state
hard scattering subprocesses. We have also generated backgrounds in the W +jets, Z+jets,
t
t (with m
t
= 171 GeV) and WW, WZ, ZZ channels at the rates shown in Table 4.3. The
W +jets and Z +jets backgrounds use exact matrix elements for one parton emission, but
rely on the parton shower for subsequent emissions.
First we require modest cuts: n(jets) 4. Also, SUSY events are expected to spray large
76
E
T
throughout the calorimeter, while QCD dijet events are expected to be typically back-to-
back. Thus, we expect QCD background to be peaked at transverse sphericity S
T
0, while
SUSY events have larger values of S
T
.
4
The actual S
T
distribution for point A is shown in
Fig. 4.12 (the S
T
distribution for point D is almost identical to that of point A). Motivated
by this, we require S
T
> 0.2 to reject QCD-like events.
0 0.2 0.4 0.6 0.8 1
S
T
1e+03
1e+04
1e+05
1e+06
1e+07
1e+08
d
/
d
S
T
(
f
b
)
Point A
Cuts C1
Figure 4.12: Plot of distribution in transverse sphericity S
T
for events with cuts C1
from
benchmark point A and the summed SM background; point D leads to practically the same
distribution.
We plot the jet E
T
distributions of the four highest E
T
jets from Pt. A (in color) and
the total SM background (gray histogram) in Fig. 4.13, ordered from highest to lowest E
T
,
with jets labelled as j1 j4. The histograms are normalized to unity in order to clearly
see the dierences in distribution shapes. Again, the distributions for point D just look the
same. We nd that the highest E
T
jet distribution peaks around E
T
150 GeV with a
long tail extending to higher E
T
values, while for the background it peaks at a lower value
4
Here, S
T
is the usual sphericity variable, restricted to the transverse plane, as is appropriate for hadron
colliders. Sphericity matrix is given as
S =
_
p
2
x
p
x
p
y
p
x
p
y
p
2
y
_
(4.4)
from which S
T
is dened as 2
1
/(
1
+
2
), where
1,2
are the larger and smaller eigenvalues of S.
77
of E
T
100 GeV. Jet 2 and jet 3 have peak distributions around 100 GeV both for the
signal and backgrounds, while the jet 4 distribution backs up against the minimum jet E
T
requirement that E
T
(jet) > 50 GeV. Thus, at little cost to signal but with large background
(BG) rejection, we require E
T
(j1) > 100 GeV.
The collection of cuts so far is dubbed C1
[102]:
C1
cuts:
n(jets) 4, (4.5)
E
T
(j1, j2, j3, j4) 100, 50, 50, 50 GeV, (4.6)
S
T
0.2. (4.7)
0
0.002
0.004
0.006
0.008
0.01
Jet 1
BG
0
0.002
0.004
0.006
0.008
0.01
Jet 2
BG
100 200 300 400 500
0
0.004
0.008
0.012
0.016
Jet 3
BG
100 200 300 400 500
0
0.006
0.012
0.018
0.024
Jet 4
BG
E
T
(j
i
) (GeV)
1
/
/
E
T
(
j
i
)
(
f
b
/
G
e
V
)
Figure 4.13: Plot of jet E
T
distributions for events with 4 jets after requiring just S
T
> 0.2,
from benchmark point A; distributions for point D are the same.
78
The classic signature for SUSY collider events is the presence of jets plus large E
miss
T
[103].
In Fig. 4.14, we show the expected distribution of E
miss
T
from points A and D, along with
SM BG. We do see that signal becomes comparable to BG around E
miss
T
150 GeV. We list
cross sections from the two signal cases plus SM backgrounds in Table 4.3 after cuts C1
plus
E
miss
T
> 150 GeV. While signal S is somewhat higher than the summed BG B, the signal
and BG rates are rather comparable in this case: S/B = 1.09 for pt. A while S/B = 1.32
for pt. D.
0 100 200 300 400
E
T
miss
(GeV)
1
10
100
1000
10000
1e+05
1e+06
d
/
d
E
T m
i
s
s
(
f
b
/
G
e
V
)
Point A
Point D
Background
Cuts C1
Figure 4.14: Plot of missing E
T
for events with 4 jets after cuts C1
, from benchmark
points A (full red line) and D (dashed blue line).
Even so, it has been noted in Ref. [102] that E
miss
T
may be a dicult variable to reliably
construct during the early stages of LHC running. The reason is that missing transverse
energy can arise not only from the presence of weakly interacting neutral particles such as
neutrinos or the lightest neutralinos, but also from a variety of other sources, including:
energy loss from cracks and un-instrumented regions of the detector,
energy loss from dead cells,
hot cells in the calorimeter that report an energy deposition even if there is not one,
mis-measurement in the electromagnetic calorimeters, hadronic calorimeters or muon
detectors,
79
real missing transverse energy produced in jets due to semi-leptonic decays of heavy
avors,
muons and
the presence of mis-identied cosmic rays in events.
Thus, in order to have a solid grasp of expected E
miss
T
from SM background processes, it will
be necessary to have detailed knowledge of the complete detector performance. As experience
from the Tevatron suggests, this complicated task may well take some time to complete. The
same may also be true at the LHC, as many SM processes will have to be scrutinized rst
in order to properly calibrate the detector[104]. For this reason, SUSY searches using E
miss
T
as a crucial requirement may well take rather longer than a year to provide reliable results.
For this reason, Ref. [102] advocated to look for SUSY signal events by searching for a
high multiplicity of detected objects, rather than inferred undetected objects, such as E
miss
T
.
In this vein, we show in Fig. 4.15 the jet multiplicity from SUSY signal (Pts. A and D) along
with SM BG after cuts C1
, i.e with no E
miss
T
cut. We see that at low jet multiplicity, SM
BG dominates the SUSY signal. However, signal/background increases with n(jets) until at
n(jets) 15 nally signal overtakes BG in raw rate.
One can do better in detected b-jet multiplicity, n
b
, as shown in Fig. 4.16. Since each
gluino is expected to decay to two b-jets, we expect a high n
b
multiplicity in signal. In this
case, BG dominates signal at low n
b
, but signal overtakes BG around n
b
4.
The isolated lepton multiplicity n
. In this case, isolated leptons should be relatively common in gluino cascade decays. We
see that signal exceeds BG already at n
= 3. In fact, high
isolated lepton multiplicity was advocated in Ref. [102] in lieu of an E
miss
T
cut to search for
SUSY with integrated luminosities of around 1 fb
1
at LHC.
We also point out here that g g production can lead to large rates for same-sign (SS)
isolated dilepton production[105], while SM BG for this topology is expected to be small.
We plot in Fig. 4.18 the rate of events from signal and SM BG for cuts C1
plus a pair
of isolated SS dileptons, versus jet multiplicity. While BG is large at low n(jets), signal
emerges from and dominates BG at higher jet multiplicities.
80
4 5 6 7 8 9 10 11 12 13 14 15 16 17
n
j
0.1
1
10
100
1000
10000
1e+05
1e+06
1e+07
(
f
b
)
Point A
Point D
Background
Cuts C1
Figure 4.15: Plot of jet multiplicity from benchmark points A (full red line) and D (dashed
blue line) after cuts C1
(
f
b
)
Point A
Point D
Background
Cuts C1
Figure 4.16: Plot of b-jet multiplicity from benchmark points A (full red line) and D (dashed
blue line) after cuts C1
(
f
b
)
Point A
Point D
Background
Cuts C1
Figure 4.17: Plot of isolated lepton multiplicity from benchmark points A (full red line) and
D (dashed blue line) after cuts C1
/
d
n
j
(
f
b
/
c
o
u
n
t
)
Point A
Point D
QCD Jets
tt
W+jets
Z+jets
WW,WZ,ZZ
Sum of Backgrounds
Cuts S
T
0.2 + 2 SS dileptons
Figure 4.18: Plot of jet multiplicity in events with isolated SS dileptons from benchmark
points A (full red line) and D (dashed blue line) after cut S
T
> 0.2 along with SM
backgrounds.
82
4.2.3 Sparticle masses from gluino pair production
There exists good prospects for sparticle mass measurements in Yukawa-unied SUSY models
at the LHC. One reason is that sparticle pair production is dominated by a single reaction:
gluino pair production. The other propitious circumstance is that the mass dierence
m
0
2
m
0
1
is highly favored to be bounded by M
Z
. This means that
0
2
decays dominantly
into three body modes such as
0
2
0
1
) which is
bounded by m
0
2
m
0
1
: this kinematic edge can serve as the starting point for sparticle mass
reconstruction in cascade decay events[21, 99].
As an example, we require cuts set C1
) d/dm(
+
and
e
+
pairs from processes like chargino pair production in cascade decay events. A clear
peak at m(
) = M
Z
is seen in the BG distribution. This comes mainly from QCD jet
production events, since Isajet includes W and Z radiation in its parton shower algorithm
(in the eective W approximation). The signal displays a histogram easily visible above SM
BG with a distinct cut-o at m
0
2
m
0
1
= 73 GeV. Isajet contains the exact decay matrix
elements in 3-body decay processes, and in this case we see a distribution that diers from
pure phase space, and yields a distribution skewed to higher m(
b
0
2
) dominates at
around 56%. If one can identify events with a clean
0
2
0
1
decay, then one might also
try to extract the invariant mass of the associated two b-jets coming from the gluino decay,
which should have a kinematic upper edge at m
g
m
0
2
267 (283) GeV for point A (D). A
second, less pronounced endpoint is expected at m
g
m
0
1
340 (335) GeV due to g b
b
0
1
83
0 50 100 150
m(ll) (GeV)
0
5
10
15
20
25
d
/
d
m
(
l
l
)
(
f
b
/
G
e
V
)
Point A
Point D
Background
Cuts C1 + 2 SF/OS
m
2
0
m
1
0
m
2
0
m
1
0
Figure 4.19: SF/OS dilepton invariant mass distribution after cuts C1
from benchmark
points A (full red line) and D (dashed blue line) along with SM backgrounds.
decays which have 16% branching ratio. A third endpoint can also occur from
0
2
0
1
b
b
decay where m
0
2
m
0
1
= 73 (52.5) GeV, respectively.
The high multiplicity of b-jets (typically two from each gluino decay), however, poses
a serious combinatorics problem in extracting the b
b invariant-mass distribution.
84
0 50 100 150
m(ll) (GeV)
0
4
8
12
16
20
24
28
d
/
d
m
(
l
l
)
/
d
m
(
l
l
)
(
f
b
/
G
e
V
)
Point A
Point D
BG
Cuts C1 + 2 OS dileptons
m
2
0
m
1
0
m
2
0
m
1
0
Figure 4.20: Same as Fig. 4.19 but for same-avor minus dierent-avor subtracted invariant-
mass.
85
Parton-level Monte Carlo simulations revealed that the two hardest (highest E
T
) b-jets
almost always originated from dierent gluinos. Thus, we require events with at least
four tagged b-jets (along with cuts C1
b decays, and of
0
2
0
1
b
) < m
0
2
m
0
1
), we reconstruct m(bb
/
d
m
(
X
i
)
(
m
(
X
1
X
2
)
)
m
i
n
(
f
b
/
G
e
V
)
m
g
~
- m
2
0
m
g
~
- m
2
0
m
g
~
- m
1
0
m
g
~
- m
1
0
Point A
Point D
Background
Cuts C1 + 4 b-jets
Figure 4.21: Plot of m(X
1,2
) from benchmark points A and D along with SM backgrounds
in events with cuts C1
/
d
m
(
X
i
)
(
m
(
X
1
X
2
)
)
m
i
n
(
f
b
/
G
e
V
)
m
g
~
- m
1
0
m
g
~
- m
1
0
m
g
~
- m
2
0
m
g
~
- m
2
0
Point A
Point D
Background
Cuts C1 + 4 b-jets + 2 SF/OS leptons
Figure 4.22: Same as Fig. 4.21 but requiring in addition a pair of SF/OS leptons.
87
According to [107], measurements of hadronic mass edges can be made with a precision
of roughly 10%. Nevertheless, from the kinematic distributions discussed above we can only
determine mass dierences. There is still not enough information to extract absolute masses,
i.e., each of m
g
, m
0
2
and m
0
1
. However, it is pointed out in Ref. [70] that in cases (such as
the focus point region of minimal supergravity) where sparticle pair production occurs nearly
purely from g g production, and when the dominant g branching fractions are known (from
a combination of theory and experiment), then the total g g production cross section after
cuts allows for an absolute measurement of m
g
to about an 8% accuracy. These conditions
should apply to our Yukawa-unied SUSY cases, if we assume the 56% branching fraction
for g b
b
0
2
decay (from theory). The study of Ref. [70] required that one fulll the cuts
C2 which gave robust gluino pair production signal along with small SM backgrounds:
88
0 200 400 600 800 1000
m(ll+2b-jets)
min
(GeV)
0
0.2
0.4
0.6
0.8
d
/
d
m
(
l
l
+
2
b
-
j
e
t
s
)
m
i
n
(
f
b
/
G
e
V
)
Point A
Point D
Background
Cuts C1 + 2 SF/OS
m
g
~
- m
1
0
m
g
~
- m
1
0
Figure 4.23: Plot of m(bb
+
)
min
from points A and D along with SM backgrounds.
0 100 200 300 400 500 600 700 800
m(X
i
ll)
(m(X
1
X
2
))
min
(GeV)
0
0.02
0.04
0.06
d
/
d
m
(
X
i
l
l
)
(
m
(
X
1
X
2
)
)
m
i
n
(
f
b
/
G
e
V
)
m
g
~
- m
1
0
m
g
~
- m
1
0
Point A
Point D
Background
Cuts C1 + 4 b-jets + 2 SF/OS leptons
Figure 4.24: Plot of m(X
1,2
)
min
from points A and D, minimizing m(X
1
X
2
) as
explained in the text, along with SM backgrounds.
89
C2 cuts:
E
miss
T
> (max(100 GeV, 0.2M
eff
), (4.8)
n(jets) 7, (4.9)
n(b jets) 2, (4.10)
E
T
(j1, j2 j7) > 100, 50 GeV, (4.11)
A
T
> 1400 GeV, (4.12)
S
T
0.2 , (4.13)
where A
T
is the augmented eective mass A
T
= E
miss
T
+
leptons
E
T
+
jets
E
T
. In this
case, the summed SM background was about 1.6 fb, while signal rate for Point A (D) is 57.3
(66.2) fb. The total cross section after cuts varies strongly with m
g
, allowing an extraction
of m
g
to about 8% for 100 fb
1
integrated luminosity, after factoring in QCD and branching
fraction uncertainties in the total rate. Once an absolute value of m
g
is known, then m
0
2
and m
0
1
can be extracted to about 10% accuracy from the invariant mass edge information.
4.2.4 Trilepton signal from
W
1
0
2
production
While the signal from gluino pair production at the LHC from Yukawa-unied SUSY models
will be very robust, it will be useful to have a conrming SUSY signal in an alternative
channel. From Fig. 4.10, we see that there also exists substantial cross sections for
W
1
0
1
,
W
+
1
1
and
W
1
0
2
production. The
1
0
1
f
f
and
0
2
0
1
f
f decays (here f stands
for any of the SM fermions) are dominated by W and Z exchange, respectively, so that
in this case the branching fractions BF(
1
0
1
f
f
f
f
) and
BF(
0
2
0
1
f
f) is similar to Z f
f.
The
W
1
0
1
0
1
q q
+
0
1
process will be dicult to observe at LHC since the nal state
jets and E
miss
T
will be relatively soft, and likely buried under SM background. Likewise, the
1
0
1
0
1
0
1
signal will be buried under a huge BG from W
production. The
W
+
1
1
production reaction will also be dicult to see at LHC. The purely hadronic nal
state will likely be buried under QCD and Z +jets BG, while the lepton plus jets nal state
will be buried under W + jets BG. The dilepton nal state will be dicult to extract from
W
+
W
and t
t production.
90
The remaining reaction,
W
1
0
2
production, yields a trilepton nal state from
1
0
1
and
0
2
0
1
decays which in many cases is observable above SM BG. The LHC reach for
1
0
2
3 + E
miss
T
production was mapped out in Ref. [108], and the reach was extended
into the hyperbolic branch/focus point (HB/FP) region in Ref. [109]. The method was to
use the cut set SC2 from Ref. [106] but as applied to the LHC. For the clean trilepton signal
from
W
1
0
2
3 +E
miss
T
production, we require:
three isolated leptons with p
T
() > 20 GeV and [
[ < 2.5,
SF/OS dilepton mass 20 GeV < m(
+
,
a transverse mass veto 60 GeV < M
T
(, E
miss
T
) < 85 GeV to reject on-shell W
contributions,
E
miss
T
> 25 GeV and,
veto events with n(jets) 1.
The resulting BG levels and signal rates for points A and D are listed in Table 4.4. The
2 2 processes are calculated with Isajet, while the 2 4 processes are calculated at parton
level using Madgraph1[110]. The combination of hard lepton p
T
cuts and the requirement
that n(jets) = 0 leaves us with no 2 2 background, while the parton level 2 4 BG
remains at 0.7 fb. Here, we see that signal from the two Yukawa-unied points well exceeds
background.
In the clean 3 channel, since two of the leptons ought to come from
0
2
0
1
decay,
they should display a conrmatory dilepton mass edge at m
0
2
m
0
1
as is evident in the
gluino pair production events, where the dileptons are accompanied by high jet multiplicity.
The distribution in m(
+
t 3 10
6
4.9 10
5
WW, ZZ, WZ 5 10
5
8.0 10
4
, W
10
6
0.7
Total BG 4.5 10
5
0.7
Point A: 10
6
7.6 10
4
3.4
S/B 4.86
S/
S +B (10 fb
1
) 5.31
Point D: 10
6
9.0 10
4
4.1
S/B 5.86
S/
S +B (10 fb
1
) 5.92
0 20 40 60 80 100
m(ll) (GeV)
0
0.05
0.1
0.15
0.2
d
/
d
m
(
l
l
)
(
f
b
/
G
e
V
)
Point A
Point D
Cuts SC2 Clean Trilepton
Zero Jets Only
Figure 4.25: Plot of m(
+
G
(2) a WMAP-allowed relic density of CDM that
consists dominantly of thermally produced axinos, and (3) the re-heat temperature needed
to fulll the relic density falls above the lower bound required by non-thermal leptogenesis,
and below the upper bound coming from gravitino/BBN constraints.
We feel that the fact that Yukawa unied SO(10) SUSY GUT models pass these several
cosmological tests makes them even more compelling than they were based on pure particle
physics reasons. In any case, with a spectrum of light gluinos, charginos and neutralinos,
they should soon be tested by experiments at the CERN LHC[102].
Finally, in Chapter 4, we make an argument for using the multi-lepton channel instead of
missing E
T
and show collider results for Yukawa-unied SO(10) SUSY GUT models. In the
very early run of the LHC pp collider, it may not be possible to use E
miss
T
as a discrimination
variable due to detector calibration issues. We show here that a substantial reach for gluino
and squark production followed by cascade decays can be gained by requiring events with
large jet and isolated lepton multiplicity, but with no requirement on E
miss
T
. In the mSUGRA
model with a low and high value of m
0
, an LHC reach for m
g
of 750 (1000) GeV is found
with 0.1 (1) fb
1
of integrated luminosity by requiring 4 jets plus 3 isolated leptons.
If enough signal events are found, then some kinematic reconstruction of sparticle masses
should be possible as in the cases where large E
miss
T
is required. SUSY signal can also be
seen above SM BG if just two OS/SF leptons are required, especially in the case where there
is a distinctive kinematic dilepton invariant mass edge.
Simple SUSY grand unied models based on the gauge group SO(10) may have t b
Yukawa coupling unication in addition to gauge group and matter unication. By assuming
the MSSM is the eective eld theory valid below M
GUT
, we can, starting with weak scale
fermion masses as boundary conditions, check whether or not these third generation Yukawa
couplings actually unify. The calculation depends sensitively on the entire SUSY particle
mass spectrum, mainly through radiative corrections to the b, t and masses. It was
found in previous works that t b Yukawa coupling unication can occur, but only for
very restrictive soft SUSY breaking parameter boundary conditions valid at the GUT scale,
leading to a radiatively induced inverted mass hierarchy amongst the sfermion masses. While
squarks and sleptons are expected to be quite heavy, gluinos, winos and binos are expected
to be quite light, and will be produced at large rates at the CERN LHC.
95
We expect LHC collider events from Yukawa-unied SUSY models to be dominated by
gluino pair production at rates of (30 150) 10
3
fb. The gs decay via 3-body modes into
b
b
0
2
, b
b
0
1
and tb
1
, followed by leptonic or hadronic 3-body decays of the
0
2
and
1
. A
detailed simulation of signal and SM BG processes shows that signal should be easily visible
above SM BG in the 4 jets plus 3 channel, even without using the E
miss
T
variable, with
about 1 fb
1
of integrated luminosity.
If Yukawa-unied signals from g g production are present, then at higher integrated
luminosities, mass edges in the m(
+
), m(b
b) and m(b
b
+