Table Of ContentUH-511-1185-12
WZ plus missing-E signal
T
from gaugino pair production at LHC7
2
1
0
2 Howard Baer1 , Vernon Barger2 , Sabine Kraml3 , Andre Lessa4 ,
∗ † † ‡
n
Warintorn Sreethawong1 and Xerxes Tata5
a § ¶
J
5
1Dept. of Physics and Astronomy, University of Oklahoma, Norman, OK 73019, USA
2
2Dept. of Physics, University of Wisconsin, Madison, WI 53706, USA
h] 3Laboratoire de Physique Subatomique et de Cosmologie, UJF Grenoble 1, CNRS/IN2P3,
p
INPG, 53 Avenue des Martyrs, F-38026 Grenoble, France
-
p 4Instituto de F´ısica, Universidade de S˜ao Paulo, S˜ao Paulo-SP, Brazil
e
5Dept. of Physics and Astronomy, University of Hawaii, Honolulu, HI 96822, USA
h
[
1
v Abstract
2
8 LHC searches for supersymmetry currently focus on strongly produced sparticles, which
3 arecopiouslyproducedifgluinosandsquarkshavemassesofafewhundredGeV.However,
5
in supersymmetric models with heavy scalars, as favored by the decoupling solution to
.
1 >
120 rtheseulStUs,ScYharflgaivnoor–naenudtrCalPinop(rWfob1±leZme2s),parnoddumctg˜ion∼is50th0eGdeoVminasanitndcricoasstesdecbtiyonrefcoernmt WLe1H∼C
< <
v: mZe2 < mg˜/3 at LHC with √s = 7 TeV (LHC7). Furthermore, if mZe1 +mZ ∼ mZe2 ∼
e e e f f e
Xi mZe1 +mh, then Z2 dominantly decays via Z2 → Z1Z, while W1 decays via W1 → Z1W.
We investigate the LHC7 reach in the WZ+E channel (for both leptonic and hadronic
r 6 T
a decays of the W boson) in models with and without the assumption of gaugino mass
universality. In the case of the mSUGRA/CMSSM model with heavy squark masses, the
LHC7 discovery reach in the WZ+E channel becomes competetive with the reach in
T
the canonical E + jets channel for6 integrated luminosities 30 fb−1. We also present
T
6 ∼
the LHC7 reach for a simplified model with arbitrary me and me me . Here, we
Z1 W1 ∼ Z2
find a reach of up to me 200 (250) GeV for 10 (30) fb−1.
W1 ∼
PACS numbers: 14.80.Ly, 12.60.Jv, 11.30.Pb
∗Email: [email protected]
†Email: [email protected]
†Email: [email protected]
‡Email: [email protected]
§Email: [email protected]
¶Email: [email protected]
1 Introduction
A major goal of the CERN Large Hadron Collider (LHC) is to test the idea of weak scale
supersymmetry (SUSY) [1], wherein superpartners of the Standard Model (SM) particles have
masses of the order of 1 TeV. The SUSY searches by the ATLAS and CMS collaborations have
reported no signal beyond SM expectations [2,3] in 1 fb−1 of data. Interpreting their results
∼
within the mSUGRA/CMSSM model [4], ATLAS and CMS exclude roughly the mass range
m m < 1 TeV for m m , and m < 550 GeV in the case where m m .1 This reach
q˜ g˜ q˜ g˜ g˜ q˜ g˜
∼ ∼ ≃ ∼ ≫
will soon be extended since each experiment now has 5 fb−1 of data collected. Analysis of
∼
this extended data sample is eagerly anticipated by the HEP community.
Within a large class of SUSY models, it is expected that pair production of strongly inter-
acting sparticles—g˜g˜, g˜q˜and q˜q˜production—constitutes the dominant SUSY production cross
sections [6,7]. The gluinos and squarks are then expected to decay through a (possibly lengthy)
cascade to lighter sparticles plus SM particles, until the decay chain terminates in the (stable)
lightest SUSY particle (LSP) [8]. The LSP is expected from cosmological arguments to be a
e
massive, neutral, weakly interacting particle (such as the lightest neutralino Z ) and so does
1
not deposit energy in the experimental apparatus, giving rise to the classic missing transverse
energy (E ) signature. Thus, gluino and squark pair production followed by cascade decays
T
6
is expected to give rise to final states containing multiple isolated leptons, multiple jets and
E [9].
T
6
While weak scale supersymmetric models are theoretically very compelling, they do suffer
from a variety of problems, including 1. the SUSY flavor problem, 2. the SUSY CP problem,
3. the gravitino problem, and 4. the danger of too rapid proton decay in SUSY grand unified
theories (GUTs). All four of these problems are greatly ameliorated if not solved by the
decoupling solution, wherein first and second generation sfermion masses are pushed into the
multi-TeV regime or even beyond. Naturalness may be maintained in models wherein sparticles
that couple directly to the Higgs sector—the third generation scalars and electroweak-inos—
remain at or below the TeV scale [10,11]. Also, in many SUSY models, it is expected that
gauginomass parametersunifyattheGUTscale, inparallelwithunificationofgaugecouplings.
Renormalization group running effects result in weak scale gaugino masses occurring in the
approximate ratio M : M : M 1 : 2 : 7. We would thus expect the physical gluino g˜,
1 2 3
f ∼ e
the wino-like chargino W and the bino-like neutralino Z to be found with roughly the same
1 1
mass ratio, provided the superpotential µ-parameter µ M . Consequently, in models with
2
| | ≫
gauginomassunification, theexperimental boundsonthegluinomassimposesevere constraints
on chargino and neutralino masses. Current analyses do not put independent constraints on
the electroweak-ino masses if the gaugino mass unification condition is dropped [12]. Moreover,
the relative strengths of signals in various multilepton topologies (as well as the gluino mass
reach if the parent-daughter mass difference is sufficiently small) depend sensitively on the
e f
g˜ Z and/or g˜ W mass differences. Finally, an independent discovery of directly produced
1 1
− −
charginos andneutralinos is essential to elucidate the supersymmetry origin of any excess in the
well-studied multilepton plus multijet plus E channel at the LHC. It is therefore interesting
T
6
1Tobeprecise,inthemSUGRA/CMSSMinterpretation,squarkmassesarevarieduptomq˜ < 2TeV,giving
∼
>
a gluino mass limit of about mg˜ 700 GeV; this limit suffers further weakening for decoupling scalars: see [5].
∼
1
and relevant to find ways to discover charginos and neutralinos independently of gluinos.
Another point is important to note: as we push the gluino mass to larger values, convolution
of the g˜g˜subprocess cross sections with parton distribution functions (PDFs) requires sampling
higher and higher values of parton fractional momentum x . For such high values of x ,
F F
the parton-parton luminosity is sharply falling. At some point we expect that, despite being
strongly-produced, gluino pair production will no longer dominate over electroweak-ino pair
production, since these latter reactions will sample the PDFs at much lower values of x if
F
electroweak-inos are significantly lighter than gluinos.
To illustrate this, we plot in Fig. 1 the g˜g˜, Wf±Ze and Wf+Wf− production cross sections in
1 2 1 1
pb at LHC with pp collisions at √s = 7 TeV. Our results are in NLO QCD from the program
Prospino [13]. We take m 15 TeV for the first and second generations, in accord with a
q˜
≃
decoupling solution to the above-mentioned pathologies and, for simplicity, assume universal
gaugino masses at the GUT scale. From Fig. 1, we see that gluino-pair production is dominant
for m < 500 GeV. For higher values of m , Wf±Ze production is dominant, followed by Wf+Wf−
g˜ g˜ 1 2 1 1
product∼ion (the reaction Wf±Ze has lower cross section,2 as can be seen e.g. in Fig. 12.23 of
1 1
Ref. [1]). For LHC with √s = 14 TeV, g˜g˜ production remains dominant up to m 1 TeV if
g˜
∼
<
squarks are very heavy. Since ATLAS and CMS already exclude m 550 GeV when m is
g˜ q˜
∼
large it may prove fruitful to probe electroweak gaugino pair production in the 2011 data but
most of all in the 2012 LHC run. This was recognized early on in [6,7] and also more recently in
in [14,15]. Recognizing that the stability of the Higgs sector also requires sub-TeV top squarks,
we also show the cross section for top squark pair production for m = m by the dotted line3
t˜1 g˜
in Fig. 1. We see that this cross section also drops off rapidly with the top squark mass. Unless
top squarks are exceptionally light (with masses of order me or smaller, and certainly much
W1
smaller than m ), electroweak-ino production remains the dominant mechanism.
g˜
f e
Letusnext examinethesignaturesresultingfromW1Z2 production. IfmZe2 < MZ+mZe1, the
well-knowntrileptonsignalprovidesagoldensignatureforchargino-neutralinoproduction[7,16]
provided only that the branching fraction for neutralino decay is not unduly suppressed [17].
f e >
The two-body chargino decay W1 → Z1W is expected to dominate for mWe1 ∼ MW +mZe1, while
e e < <
the two-body decay Z2 → Z1Z dominates for MZ +mZe1 ∼ mZe2 ∼ mh +mZe1. For even higher
> e e
values of mZe2, i.e. mZe2 ∼ mZe1 +mh, the decay mode Z2 → Z1h turns on and dominates.
e
This is illustrated in Fig. 2, where we show the Z2 branching fractions versus me for a
Z2
mSUGRA model line with m = 10 TeV, A = 2m , tanβ = 25 and µ > 0. We vary
0 0 0
− f e e e
m1/2 to obtain the variation in mZe2. In this case, W1Z2 → WZ + Z1Z1 is kinematically
< <
allowed for 175 GeV me 250 GeV, which corresponds to gluino masses in the interval
∼ Z2 ∼
< < f e
600 GeV m 800 GeV. Thus, in this mass range, we expect the single reaction pp W Z
g˜ 1 2
∼ ∼ →
2For the wino-like Wf1 and Ze2, Wf1Ze2 production occurs via the unsuppressed isotriplet WWf1Ze2 gauge
f e
coupling, whereas the WW1Z1 coupling is strongly suppressed because it arises only due to the subdominant
higgsinocontentofthewino-likecharginoandthebino-likeneutralino—theW-bino-winocouplingisforbidden
by gauge invariance.
3TheLOtopsquarkpairproductioncrosssectionisdeterminedbyQCDandisindependentofmg˜. Inother
words,forthedottedline,the graphis plottedversusmt˜1. Ifotherthirdgenerationsquarksarealsolight,their
pair production cross sections are also given by the dotted line with the understanding that the label on the
horizontal axis is the corresponding squark mass.
2
1 g~ g~
~ ~+ ~ ~-
Z W + Z W
2 1 2 1
10-1 ~+ ~-
W W
1 1
~~
) t t
b 1 1
p 10-2
(
O
L
N
σ
10-3
10-4 m ≈ m ≈ m /3
W~ Z~ g~
m 1= m 2
~t g~
1
10-5
500 1000 1500 2000 2500
m (GeV)
g~
Figure 1: Total NLO cross sections (from Prospino) for g˜g˜, Wf±Ze and Wf+Wf− production at
1 2 1 1
LHC7 versus mg˜, where mq˜ = 15 TeV and mWe1 ≈ mZe2 ≈ mg˜/3. The dotted line shows the cross
section for t˜t˜¯ production with m = m and neglecting intra-generational squark mixing.
1 1 t˜1 g˜
m = 10 TeV, A = -2m, tanβ = 25, µ >0
10 0 0 0
Z~ → Z~ + qq Z~ → Z~ + Z Z~ → Z~ + h
1 2 1 2 1 2 1
Z~ → Z~ + l+l-
10-1 2 1
10-2
R
B
10-3
10-4
~ ~
W → Z + W
1 1
10-5
10-6
140 160 180 200 220 240 260 280 300
m (GeV)
~
Z
2
e
Figure 2: Some prominent branching fractions for Z decay in the mSUGRA model with pa-
2
f e
rameters m = 10 TeV, A = 2m , tanβ = 25 and µ > 0. We also show the W W +Z
0 0 0 1 1
− →
branching fraction (dotted line).
f e e e
followed by W Z W andZ Z Z tobe thedominant SUSYproduction anddecay process
1 1 2 1
→ →
at LHC7 for models with full gaugino mass unification. The endpoints of this interval can shift
up or down in non-universal mass scenarios.
3
2 Trilepton+ E channel
T
6
f e
We begin by examining the viability of the reaction pp W Z WZ+ E for SUSY
1 2 T
→ → 6
discovery at LHC7, focusing on the case where both Z and W decay leptonically, resulting
in clean trilepton events. It is worth mentioning that the trilepton signal from the decay
e e
Z Z Z where a pair of opposite-sign same-flavor (OS/SF) dileptons reconstruct the Z
2 1
→
mass has generally been regarded as unobservable because of large SM background from WZ
production. The case where the W decays hadronically will be discussed in Section 3.
For our LHC7 event generation, we use the event generator Isajet 7.79 [18] for signal re-
actions, while for the simulation of the background events, we use AlpGen [19] and Mad-
Graph [20] to compute the hard scattering events and Pythia [21] for the subsequent showering
and hadronization. In our simulation, we include the following backgrounds for the WZ+ E
T
signal: tt¯, W(ℓν)W(ℓν), W(ℓν)Z(ℓℓ), ZZ, W(ℓν)+tb, Z(ℓℓ)+jets, W(ℓν)+jets, Z(ℓℓ)+6b¯b,
Z(ℓℓ)+tt¯and W +tt¯. For tt¯, Z +jets, W +jets, Z +b¯b and Z +tt¯we include the full matrix
elements for at least two real parton emissions and use the MLM matching algorithm to avoid
double counting. For WZ production we include the full matrix elements for the 2 4 process
pp WZ ℓ+ℓ−ℓ′ν′. K-factors for both signal and background4 (BG) are inclu→ded and are
→ →
computed using Prospino [13] and MCFM [22], respectively.
Inourcalculations, weemployatoydetectorsimulationwithcalorimetercellsize∆η ∆φ =
×
0.05 0.05 and 5 < η < 5 . The HCAL (hadronic calorimetry) energy resolution is taken to
× −
be 80%/√E 3% for η < 2.6 and FCAL (forward calorimetry) is 100%/√E 5% for η > 2.6,
⊕ | | ⊕ | |
where the two terms are combined in quadrature. The ECAL (electromagnetic calorimetry)
energy resolution is assumed to be 3%/√E 0.5%. In all these, E is the energy in GeV units.
⊕
We use the cone-type Isajet [18] jet-finding algorithm to group the hadronic final states into
jets. Jets and isolated lepton are defined as follows:
Jets are hadronic clusters with η < 3.0, R √∆η2 +∆φ2 0.4 and E (jet) > 40 GeV.
T
• | | ≡ ≤
Electrons and muons are considered isolated if they have η < 2.5, p (l) > 10 GeV
T
• | |
with visible activity within a cone of ∆R < 0.2 about the lepton direction, ΣEcells <
T
min[5,0.15p (l)] GeV.
T
We identify hadronic clusters as b-jets if they contain a B hadron with E (B) > 15 GeV,
T
•
η(B) < 3.0 and ∆R(B,jet) < 0.5. We assume a tagging efficiency of 60% and light
| |
quark and gluon jets can be mis-tagged as a b-jet with a probability 1/150 for E 100
T
≤
GeV, 1/50 for E 250 GeV, with a linear interpolation for 100 GeV E 250 GeV.
T T
≥ ≤ ≤
Next, we invoke the following pre-selection cuts on our signal and background event samples
to extract those with a ℓ+ℓ−ℓ′+E topology:
T
6
Pre-Selection Cuts:
n(b jets) = 0 (to aid in vetoing tt¯background),
• −
3 isolated leptons with p (ℓ) > 20 GeV and
T
•
4For the background processes where the NLO cross section is not known we take the K-factor to be 1.
4
m(ℓ+ℓ−) M < 10 GeV,
Z
• | − |
where two of the leptons in the event must form an OS/SF pair. If more than one OS/SF
pairing is possible, the pair which minimizes m(ℓ+ℓ−) M is chosen. The remaining lepton
Z
′ | − |
is labeled ℓ .
′
In Fig. 3 we show the E and transverse mass (m (ℓ , E )) distributions for the signal
T T T
6 6
and the SM BG after the pre-selection cuts have been applied. The signal point has me =
W1
f e
189.3 GeV, me = 187.3 GeV and me = 89.4 GeV and we only consider W1Z2 production.
Z2 Z1
Due to its relatively light parent mass scale, the signal presents a soft E spectrum, barely
T
6
visible above the SM background. This is in strong contrast with events from production of
the much heavier gluinos or squarks, where the cascade decays to the LSP result in a usually
much harder E spectrum. Therefore, the usual E plus jets/leptons searches (optimized to
T T
6 6 f e
look for strongly produced gluinos and squarks) are insensitive to the W Z signal.
1 2
As seen in the upper frame of Fig. 3, after the pre-selection cuts the BG is dominated by ZZ
production at lowE andby WZ production forE > 20 GeV. The transverse mass m (ℓ′,E )
T T T T
6 6 ∼ 6
from W ℓ′νℓ′, shown in the lower frame of Fig. 3, falls sharply beyond the expected Jacobian
→ f e
peak at m = M . In constrast, the corresponding signal distribution from W Z production
T W 1 2
extends to considerably larger values due to the presence of the two neutralinos in the final
state. Therefore, a m cut is extremely efficient to suppress the WZ background. This is seen
T
in the lower frame of Fig. 3, where the signal distribution clearly stands out for m > 100 GeV.
T
However, since a precise prediction for the m tail from WZ events requires a full detector
T
simulation or data-driven estimates, we define a conservative signal region requiring:
E > 50 GeV,
T
• 6
′
m (ℓ ,E ) > 125 GeV .
T T
• 6
The BG cross sections from the dominant SM processes after each of the cuts mentioned
above, together with the corresponding cross sections for the representative signal point with
me = 189.3 GeV, me = 187.3 GeV and me = 89.4 GeV, are shown in Table 1. We stress
W1 Z2 Z1
f e
that the signal shown in Fig. 3 and listed in Table 1 comes exclusively from W Z production.
1 2
Depending on the sparticle spectrum, the actual signal may be larger if heavier electroweak-
inos are also accessible, or if gluino and/or squark pair production followed by their cascade
decays to the WZ final state is sizeable. Nonetheless, a trilepton signal would be visible with
an integrated luminosity of 10 fb−1 at LHC7 even if light electroweak-inos are the only SUSY
∼
particles being produced.
2.1 LHC7 Reach
As shown in Table 1 and Fig. 3, for me = 189.3 GeV, me = 187.3 GeV and me = 89.4 GeV,
W1 Z2 Z1
only an excess of 2 events in the trilepton channel (after cuts) would be expected for lumi-
∼
nosity of 5 fb−1. Thus larger integrated luminosities are required in order to claim a signal.
∼
In Fig. 4, we show the signal significance for various integrated luminosities versus me (solid
W1
lines). For now we use a mSUGRA model line with m = 10 TeV, A = 2m , tanβ = 25
0 0 0
f e −
and µ > 0, and we consider the signal only from W Z production. To allow for the low signal
1 2
5
LHC7 ( 5 fb-1)
102 tWtZ
ZZ
Z+tt
W+tt
10 Signal+BG
s
ent 1
v
E
10-1
10-2
0 50 100 150 200 250 300 350 400 450 500
Emiss (GeV)
T
tt
102 WZ
ZZ
Z+tt
10 W+tt
Signal+BG
s
nt
e 1
v
E
10-1
10-2
0 50 100 150 200 250 300 350 400 450 500
m (GeV)
T
′
Figure3: E and transversemass(m (ℓ ),E ) distributions in3ℓ+E events afterpre-selection
T T T T
6 6 6
cuts, for an integrated luminosity of 5 fb−1. The summed SM backgrounds are shaded while
the signal plus background is shown by the dashed histogram. Only the dominant background
processes are shown. The signal point has me = 189.3 GeV, me = 187.3 GeV and me =
W1 Z2 Z1
89.4 GeV.
<
rates, the significance is computed using Poisson statistics. For me 170 GeV, the decay into
W1 ∼
real Zs is kinematically forbidden– as shown in Fig. 2– and the signal significance (solid lines)
sharply drops in this region. In this case, however, the well-studied trilepton signal mentioned
earlier from Wf Ze 3ℓ+ E where m(ℓ+ℓ−) < M is observable. To illustrate this, we show
1 2 T Z
→ 6
by dashed lines the signal significance, where the same cuts listed in Table 1 are applied, except
for the m and m(ℓ+ℓ−) cuts. Since in this region Ze and Wf can decay to off-shell Zs and Ws
T 2 1
we require instead:
m > 0, m(ℓ+ℓ−) < M 10 GeV.
T Z
• −
6
tt¯ WZ ZZ Z +tt¯ W +tt¯ Total BG Signal
Events Generated 5.1M 100K 194K 451K 9.5M 200K
Total σ (fb) 1.6 105 5.1 102 5.4 103 22.3 183 7.8 106 1.1 104
× × × × ×
n(b) = 0,n(l) = 3 1.6 85.1 9.2 0.9 0.4 97.5 6.7
OS/SF pair 1.1 84.9 9.2 0.9 0.3 96.6 6.7
m(ℓ+ℓ−) cut 0.3 79.1 9.1 0.66 0.06 89.5 6.6
m > 125 GeV 0.03 0.20 0.03 0.03 0.02 0.31 0.67
T
E > 50 GeV 0.03 0.17 0 0.03 0.02 0.25 0.64
T
6
Table 1: Number of events generated, total cross section and cross section after cuts for the
dominant backgrounds in the trilepton channel and for the signal. All cross sections are in fb
f e
and the signal is from just W1Z2 production with me = 189.3 GeV, me = 187.3 GeV and
W1 Z2
me = 89.4 GeV. The Total BG values include all processes listed in the text, including the
Z1
subdominant ones not shown in the Table.
<
As seen from Fig. 4, we confirm that the signal in the low me region ( 170 GeV) is readily
W1 ∼
f e
observable via this “golden” trilepton channel, due to the large W Z production cross sections
1 2
and small background.5
e e
As me me increases so that the Z2 Z1Z decay turns on, the significance for our
W1 ≃ Z2 →
WZ → 3ℓ+E6 T signal increases, reaching its maximum for mWe1 ∼ 220 GeV. This is due to the
< < e e
fact that, for mWe1 ∼ 200 GeV, mZe2−mZe1−MZ ∼ 15 GeV and the Z1’s coming from Z2 decays
f
(and to some extent also those from W decay) are rather soft and so contribute relatively
1
little to both E and to m . As a result, the E > 50 GeV and m > 125 GeV requirements
T T T T
6 6 f e
significantly reduce the signal in this region. As me increases beyond 220 GeV, the W1Z2
W1
production cross section (after cuts) decreases, and so does the signal significance. Finally,
once mZe2 > mZe1 + mh (at mWe1 ∼ 255 GeV), the Ze2 → Ze1h decay turns on and dominates6
causing the signal to drop sharply.
We remark that for 5 fb−1 of data, we would expect a 2σ effect over essentially the entire
e e
region where the decay Z Z Z dominates. Therefore, the LHC experiments already have
2 1
→
accumulated enough luminosity to probe this entire region at 95% C.L.! However, in the
∼
happy circumstance that some excess is seen in the data, 20 30 fb−1 of data will be required
∼ −
in order to establish a 5σ discovery. This may indeed be achieved in the 2012 run of LHC7.
We note further that the SUSY signal events will contain a distinctive asymmetry of trilepton
charges +(+ ) vs. (+ ) (where the (+ ) pair reconstructs m ) that originates from the
Z
− − − −
PDFs since LHC is a pp collider. In contrast, SM backgrounds from tt¯and Ztt¯(but not WZ)
should have the number of +(+ ) events equal to (+ ) events, up to statistical fluctuations.
− − −
In addition, should a large enough data sample be accrued, the p (Z) distribution should be
T
5ThevalleyattheintersectionofthesolidanddashedlinesinFig.4arisesbecausewehavedifferentanalysis
e
cuts for the two-body andthree-body decays of Z2. This valley would be smoothed out (and partially filled in)
in a treatment that treats Z as a resonance rather than a particle with a definite mass.
6This decay occurs via the Higgs-higgsino-gauginocoupling and so is suppressed by the higgsino content of
just one of the two neutralinos. In contrast, the decay to Z occurs via the doubly suppressed higgsino content
of both neutralinos.
7
LHC7
5 fb-1
12 10 fb-1
20 fb-1
10 30 fb-1
σ) no mT cut, m(l-,l+) < 81 GeV
e ( 8
c
n
a
c
nifi 6
g
Si
4
2
0
160 180 200 220 240 260 280
m~ (GeV)
W
1
f e
Figure4: Significance of W Z WZ+E 3ℓ+E signal for various integrated luminosities
1 2 T T
→ 6 → 6
at LHC7. The solid lines have all the trilepton cuts listed in Table 1, while the dashed lines do
not include the m cut and require M m(ℓ+ℓ−) > 10 GeV instead.
T Z
−
e
well-suited for a Z mass extraction since the production and decay modes are single channel.
2
f e
In Fig. 5, we generalize our results to models with unrelated W and Z masses, i.e. models
1 1
without gaugino mass universality, taking me = me and µ M2. In this figure, we show
Z2 W1 ≫
the discovery regions for several integrated luminosities. We require the following discovery
criteria:
significance > 5σ,
•
signal/BG> 0.2 and
•
at least 5 signal events.
•
The mSUGRA model line with m = 10 TeV, A = 2m , tanβ = 25 and µ > 0, assumed in
0 0 0
−
Fig. 4, is shown as the dashed orange line. The purple band shows the kinematically allowed
region, where MZ < mZe2 − mZe1 < mh. As can be seen, chargino masses up to ∼ 170 GeV
can already be probed with 5 fb−1, if me < 50 GeV. As discussed above, for heavier Ze1,
Z1 ∼
the mZe2 −mZe1 mass gap reduces, resulting in softer mT and 6ET distributions. Therefore the
signal efficiency is reduced, requiring higher luminosities in order to achieve 5σ significance.
This effect is seen throughout the me vs. me plane, rendering the narrow region close to
W1 Z1
mZe2 − mZe1 ∼ MZ, where the Ze1 is produced at low pT, inaccessible even for L = 30 fb−1.
< e
On the other hand, the region where mZe2 − mZe1 ∼ mh results in boosted Z1s and can be
easily probed until the decay Ze Ze + h turns on, c.f. Fig. 4. The 30 fb−1 reach extends
2 1
<→
up to me 250 GeV, for me 130 GeV, covering almost all of the kinematically allowed
W1 ∼ Z1 ∼
region for the mSUGRA line with m = 10 TeV, A = 2m . We also show in Fig. 5 a
0 0 0
−
second mSUGRA line with m = 1.5 TeV, A = 0, tanβ = 45 and µ > 0. For these choice
0 0
of parameters the me me mass difference is reduced, due to a small (positive) A0 value
Z2 − Z1
8
300
280
260
V) 240
e 5 fb-1
G 220
(~W1200 10 fb-1
m 20 fb-1
180
30 fb-1
160
mSUGRA (m = 10 TeV)
0
140
mSUGRA (m = 1.5 TeV)
0
120
50 100 150 200 250
m~ (GeV)
Z
1
Figure 5: 5σ discovery regions for various integrated luminosities at LHC7 in the me me
W1 − Z1
f e
plane. We assume me = me and consider only W1Z2 production. The Higgs boson mass is
Z2 W1
assumed to be 128.5 GeV throughout the plane. The orange (red) line shows the mSUGRA line
with m = 10 TeV, A = 2m , tanβ = 25 and µ > 0 (m = 1.5 TeV, A = 0, tanβ = 45 and
0 0 0 0 0
−
µ > 0).
and smaller squark masses. As a result, all of the region where the WZ+ E channel is open
T
6
falls into the inaccessible region at high me . However, values of A0 0 now seem excluded in
Z1 ∼
mSUGRA if indeed m turns out to be 125 GeV [23].
h
∼f e
Up to now we have only considered W Z production. Despite having subdominant pro-
1 2
f e
duction cross sections, production of heavier chargino W and neutralinos Z usually leads to
2 3,4
a harder E spectrum due to their cascade decay, possibly enhancing the signal. Furthermore,
T
6
for low m (m ) squark (gluino) production and cascade decay can also enhance the trilepton
0 1/2
signal. In order to clearly see these effects we choose the A and tanβ values from the red curve
0
f e
in Fig. 5 (A = 0 and tanβ = 45), where we do not expect the W Z signal to be visible for
0 1 2
any value of m , even for 30 fb−1. However, now we perform a scan over the m m plane
1/2 0 1/2
−
and include the production from all SUSY particles, including squarks and gluinos. For each
point in parameter space, we apply the trilepton cuts shown in Table 1 and take the point to
be visible if the discovery criteria listed above are satisfied.
The results are shown in Fig. 6, again for four values of integrated luminosities. All points
shown are deemed visible for the corresponding integrated luminosity. The gray regions show
the parts of the m m plane excluded by theoretical considerations or by experimental
0 1/2
−
constraints. The purple band across the middle of the plot shows the region in parameter space
where MZ < mZe2 −mZe1 < mh, while the pink area at low values of m0 and m1/2 corresponds
to the region where at least 50% of the signal comes from gluino and/or squark production.
From Fig. 6 we see that, for heavy squarks (m > 800 GeV), the signal mostly comes from
0
electroweakly produced inos. For an integrated luminosity of 5 fb−1 no points are visible.
However, for an integrated luminosity of 10 fb−1, the enhancement of the signal from gluino
and squark production renders a few points at low m and low m accessible. For 20 fb−1 the
0 1/2
9
