arX

iv:0

912.

3799

v1 [

hep-

ph]

20

Dec

200

9

Combined analysis of Z ′ → tt and Z ′ → ttj produ tion

for ve tor resonan e sear hes at LHC

F. del Aguila

a, J. A. Aguilar�Saavedra

a, M. Moretti

b, F. Pi inini

c, R. Pittau

a,

M. Tre ani

a

aDepartamento de Físi a Teóri a y del Cosmos and CAFPE,

Universidad de Granada, E-18071 Granada, Spain

bDipartimento di Fisi a, Università di Ferrara, and INFN, 44100 Ferrara, Italy

cINFN, Sezione di Pavia, v. A. Bassi 6, I 27100, Pavia, Italy

Abstra t

We have implemented a ode for Z ′+ n jets produ tion in ALPGEN, with

Z ′de ays into several �nal states, in luding ℓ+ℓ− and tt. The MLM pres ription

is used for mat hing the matrix element with the parton shower, in luding in

this way the leading soft and ollinear orre tions. In order to demonstrate its

apabilities, we perform a ombined analysis of Z ′ → tt and Z ′ → ttj produ tion

for a heavy leptophobi gauge boson. It is found that the e�e t of the extra jet

annot only be a ounted for by a K fa tor multiplying the leading-order ross

se tion. In fa t, the ombined analysis for Z ′ → tt and Z ′ → ttj presented

improves the statisti al signi� an e of the signal by 25% (8.55σ versus 6.77σ for

a Z ′mass of 1 TeV), ompared with the results of an in lusive analysis arried

out on the same sample of tt+ ttj events.

1 Introdu tion

Sear hing for neutral ve tor resonan es is one important task in the Large Hadron

Collider (LHC) programme, being the Drell-Yan pro ess pp→ Z ′ → ℓ+ℓ− , ℓ = e, µ, the

preferred one in these studies [1�4℄. This signal with harged leptons in the �nal state

has smaller ba kgrounds than those with only hadrons, but it requires that the new

boson ouples to both quarks and leptons, being this pro ess suppressed when either

of these types of ouplings vanishes. The tt hannel is an alternative if the ouplings

to leptons are the ones whi h are negligible. Among the hadroni �nal states, tt is

very interesting by itself be ause, being the top very heavy, it allows for a relatively

easy identi� ation and re onstru tion, and for this reason its ba kgrounds are relatively

smaller. Even more, Z ′ → tt is not only an alternative to the Drell-Yan pro ess when

the new resonan e is leptophobi , but it is always omplementary to determine the

1

model be ause it involves a di�erent ombination of ouplings, and it also allows for

asymmetry measurements in the semileptoni de ay [5℄.

New gauge bosons are predi ted in many of the best motivated Standard Model

(SM) extensions. For instan e, parity restoration, whi h an be at the TeV s ale,

requires extending the SM gauge symmetry to SU(2)L × SU(2)R ×U(1)B−L, with new

neutral and harged gauge bosons at this s ale [6℄. Grand uni�ed models always predi t

new gauge bosons, and some of them may survive at lower energies [7℄. In general,

models addressing the hierar hy problem whi h are not based on supersymmetry, su h

as Little Higgs models [8℄ or models with large extra dimensions [9℄, also introdu e (a

plethora of) new ve tor bosons, with some of them at the LHC rea h. Many of them

ouple to quarks and leptons but, as already stressed, even in this ase (and obviously

in the leptophobi limit) it is important to perform a detailed resonan e sear h in the

tt hannel.

With this purpose in mind, we have extended ALPGEN [10℄ with a generator for Z ′

produ tion, in luding real emissions mat hed with leading logarithmi (LL) orre tions

(both real and virtual ones). This is presented in the next se tion, and applied to the

sear h of ve tor resonan es de aying into tt in the following ones. For the sake of

illustration, we on entrate on the ase of a leptophobi Z ′

λ, whi h also de ays into

new heavy neutrinos Z ′

λ → NN if mN < MZ′

λ/2 [11℄. The model is des ribed in

se tion 3, where we also give details of the simulation of top pair produ tion mediated

by neutral gauge bosons. Then, in se tion 4 we show the relevan e of using a generator

in luding these ontributions: a ombined analysis of Z ′

λ → tt and Z ′

λ → ttj improves

the LHC sensitivity to neutral ve tor resonan es, raising the statisti al signi� an e

of the signal by 25% ompared with an in lusive analysis. Pre ise simulations of Z ′

produ tion in di�erent hannels are not only essential for dis overy, but to determinine

the model by measuring all the Z ′ ouplings with a pre ision as high as possible. The

last se tion is devoted to our on lusions.

2 The Z ′generator

In this se tion we brie�y des ribe the main features of the new ode. It evaluates the

matrix elements for Z ′ + n jets produ tion, with Z ′de aying into several �nal states

in luding ℓ+ℓ− and tt. A tually �Z ′ + n jets� stands for the sum of the three possible

intermediate ve tor bosons, namely Z, γ∗ and Z ′; their interferen es as well as their

2

widths and all spin orrelations are taken into a ount.

1

Detailed information an be

found in the README �le available at http://mlm.home. ern. h/mlm/alpgen/. The

SM parameters are ontrolled, as it happens for all the other pro esses implemented

in ALPGEN [10℄, by the variable iewopt, whose default value is iewopt = 3. The

possible �nal state is sele ted with the parameter ifs (e.g. ifs = 0 → e+e−, ifs =

2 → tt), and its left-handed (LH) and right-handed (RH) ouplings to Z ′with glzpe,

grzpe and glzptop, grzptop respe tively. In addition, the new gauge boson (LH

and RH) ouplings to the initial quarks, and its mass and width an be arbitrarily de-

�ned through the variables glzpup, grzpup, glzpdown, grzpdown, glzp , grzp ,

glzps, grzps, and masszp and zpwid, respe tively, the lagrangian being

L = −1

2masszp (Z ′)2+

1

2glzpup uγµ(1−γ5)uZ ′

µ+1

2grzpup uγµ(1+γ5)uZ

′

µ+. . . , (1)

and zpwid the width of the Z ′boson. Finally, the MLM mat hing pres ription [12,13℄

an be applied by taking the parameter i kkw = 1. In Table 1 we gather the parton

subpro esses evaluated for any given jet multipli ity. Sin e our aim is studying QCD

Subpro esses

uu→ f f dd→ f f gu→ f fu gd→ f fd

ug → f fu dg → f fd gg → f fuu gg → f fdd

Table 1: Parton subpro esses (plus their harge onjugate) taken into a ount in the

omputation. Quarks u and d represent generi quarks of type up and down, respe -

tively, and f f stands for the fermion pair sele ted in the Z ′de ay through the variable

ifs. If the jet multipli ity ex eeds the number of light quarks in the �nal state, �nal

state gluons are added up to rea h the desired multipli ity.

orre tions to the produ tion of a single s- hannel Z ′boson, rather than the produ tion

of two Z ′bosons, we negle t ontributions with two light quark pairs, be ause in our

approa h the former ex ludes the latter.

In order to illustrate the apabilities of this ode we will evaluate top pair produ tion

for the leptophobi model des ribed in the next se tion. The full analysis is presented

in se tion 4. As it an readily be seen from Fig. 1, taking MZ′

λ= 1 TeV with a total

width ΓZ′

λ= 6.9 GeV, a proper al ulation of the real radiation ontributions to Z ′j is

1

The light jet(s) an also result from radiation o� the Z ′de ay produ ts, as for example in qq →

Z ′ → ttg, but for simpli ity we still denote these pro esses as Z ′ + n jets. Obviously, by Z ′ → ttj we

only refer to events in whi h the extra jet is radiated o� the Z ′de ay produ ts. The radiation from

top de ay produ ts is not in luded at the matrix element level.

3

0 500 1000 1500 2000m

tt (GeV)

0

1000

2000

3000

4000

5000m

ttj (

GeV

)qq

0 500 1000 1500 2000m

tt (GeV)

0

1000

2000

3000

4000

5000

mtt

j (G

eV)

gq

Figure 1: tt versus ttj invariant mass distributions at the generator level, for Z ′

λj events

originated in qq (left) and gq (right) ollisions at LHC. The number of points in the

plots (8172 for qq and 1332 for gq) orresponds to the expe ted number of events at

the LHC for MZ′

λ= 1 TeV with a luminosity of 30 fb

−1.

ompulsory to a ount for the relatively large number of Z ′ → ttj events produ ed in

qq ollisions.2 (In both ases the SM ontributions from Z and γ∗ are turned o�, but

they are in luded in the analyses below.) In parti ular, as it an learly be seen in the

left plot, 37% of the qq → ttj events orrespond to a ttj resonan e, being the extra

jet from �nal state radiation (FSR) from one top quark. The rest orresponds to a tt

resonan e with the jet from initial state radiation (ISR).

3 Des ription of the model and the simulation

As explained in the previous se tion, the ode allows for arbitrary Z ′ ouplings to

fermions (models). For illustration we hoose a model with vanishing ouplings to

ordinary leptons, but non-zero ouplings to SM quarks (in parti ular to the top) and to

new heavy neutrinos. Being the new boson leptophobi , the lower bound on its mass is

rather weak, and sizeable signals are already possible at Tevatron [11℄. There are many

models with extra leptophobi gauge bosons, originally studied to interpret the initial

disagreement between the LEP data on Z → bb, cc [15℄ and the SM predi tions [16,17℄.

We will restri t ourselves to an E6 based model [18℄. The neutral gauge intera tions

are des ribed by the Lagrangian [19℄

LNC = −ψγµ[

T3gWµ3 +

√

5

3Y gYB

µ +Q′g′Z′µλ

]

ψ , (2)

2

Noti e that in the SM, when usual isolation and transverse momentum uts are applied, the largest

ontribution to Zj omes from gq ollisions [14℄.

4

where a sum over the Weyl fermions of the fundamental E6 representation 27 and

the three families must be understood. Y is the SM hyper harge properly normalised,

and the extra Z ′

λ harges Q′ orrespond to the only leptophobi ombination within

E6 [17,20℄, Q′ = 3/

√10 (Yη+Y/3) , with Yη the extra U(1) de�ned by �ux breaking [21℄.

For the SM (LH) �elds

2Q′

u = 2Q′

d = Q′

uc = −2Q′

dc = − 1√6,

Q′

ν = Q′

e = Q′

ec = 0 , (3)

reading the ode harges for ilep = 2

glzptop = glzpup = glzpc = − g′

2√6,

grzptop = grzpup = grzpc =g′√6,

glzpdown = glzps = grzpdown = grzps = − g′

2√6. (4)

A detailed dis ussion of the phenomenology of this SM extension an be found in

Ref. [17℄, where a nearly-leptophobi model with Q′ ∼ Yη + 0.29 Y is studied together

with several other alternatives. In our phenomenologi al study we will assume for

simpli ity that the extra ve tor-like lepton doublets and quark singlets of harge −1/3

are heavier than MZ′

λ/2, as they are the heavy neutrinos.

3

Possible supersymmetri

partners are taken to be heavier as well. Otherwise, the total Z ′

λ width would be

larger, de reasing the ross se tions into SM �nal states. The total leading order (LO)

Z ′

λ produ tion ross se tion at LHC is plotted in Fig. 2 as a fun tion of MZ′

λ, together

with the maximum (i.e. when Z ′

λ de ays only to SM fermions) ross se tion for tt

�nal states. The oupling onstant of the new U(1)′ has been �xed for referen e as

g′ =√

5/3 gY =√

5/3 g sW/cW , and ross se tions are al ulated using CTEQ5L

parton distribution fun tions [23℄.

For the simulations in the next se tion we also take a Z ′

λ mass of 1 TeV, above

the Tevatron ex lusion limits for this model [24℄. The signal is generated with Monte

Carlo statisti s of 300 fb

−1and res aled to 30 fb

−1. SM ba kgrounds in lude ttnj (with

nj standing for n jets at the parton level), single top, W/Znj, W/Zttnj, W/Zbbnj,

W/Zccnj, diboson and triboson produ tion. They are generated with a Monte Carlo

statisti s of 30 fb

−1using ALPGEN, taking mt = 175 GeV, MH = 115 GeV. (The

omplete list of pro esses and numbers of events generated an be found in Ref. [25℄.)

3

If Z ′

λ an de ay into heavy neutrino pairs NN (with Z ′

λ harge Q′

N= −3/2

√6), this ould be

also observed in multi-lepton hannels [11, 22℄.

5

500 1000 1500 2000 2500 3000 3500M

Zλ′ (GeV)

10-3

10-2

10-1

1

10

102

103

σ (p

b)

Z′λ total (LO)

Z′λ → tt (LO)

Figure 2: Cross se tions for Z ′

λ produ tion (solid) and for Z ′

λ → tt (dashed) at LHC as

a fun tion of the Z ′

λ mass.

Events are interfa ed to Pythia 6.4 [26℄ to add soft ISR and FSR and pile-up, and

perform hadronisation. The MLM pres ription is also used to perform the mat hing

for the ba kgrounds, with default values for the mat hing parameters. A detailed

investigation of the un ertainties related to the mat hing pro edure with ALPGEN

for the main ba kground pro esses has been presented in Refs. [13, 27℄. In order to

simulate a real dete tor environment we use the fast simulation A erDET [28℄ whi h

is a generi LHC dete tor simulation, neither of ATLAS nor of CMS, with standard

settings. In parti ular, the lepton isolation riteria require a separation ∆R > 0.4

from other lusters and a maximum energy deposition ΣET = 10 GeV in a one of

∆R = 0.2 around the re onstru ted ele tron or muon. Jets are re onstru ted using

a one algorithm with ∆R = 0.4. In this analysis we only fo us on entral jets with

pseudo-rapidity |η| < 2.5. For entral jets, a simple b tagging is performed with

probabilities of 60% for b jets, 10% for harm and 1% for light jets.

Our estimate of the signal relies on the LO approximation. In the absen e of a

proper next-to-leading order (NLO) al ulation, we estimate the impa t of hard NLO

orre tions using MCFM v5.6 NLO ode [29℄. We have res aled Z and W masses

(and a ordingly the GF oupling onstant) in su h a way that the Z mass is pushed

to 500 GeV and 1 TeV, while keeping the SM ouplings, in order to simulate the

sequential Z ′

SM . We sele ted the bb �nal state among the possible ones, hoosing a

window MZ′

SM− 7ΓZ′

SM< mbb < MZ′

SM+ 7ΓZ′

SM. As a result the K-fa tor, de�ned

as σ(NLO)/σ(LO) with renormalization and fa torization s ales �xed to MZ′

SM, turns

out to be of the order of 20% for both MZ′

SM= 500 GeV and MZ′

SM= 1 TeV.

6

4 Z ′ → tt and Z ′ → ttj observation at LHC

The presen e of a Z ′resonan e is dete ted as a bump in the tt invariant mass spe -

trum, whi h an eventually be normalised from real data with a good pre ision. For

this analysis we restri t ourselves to the tt semileptoni de ay hannel, in whi h the

kinemati s an be fully re onstru ted. The pre-sele tion riteria are: (i) one harged

lepton with pT > 30 GeV; (ii) two b-tagged jets with pT > 20 GeV; (iii) at least two

light jets with pT > 20 GeV; (iv) total transverse energy HT > 750 GeV. The latter

ut is implemented in order to redu e the QCD ttnj ba kground, as well as to remove

from the signal pro esses the SM omponent mediated by Z and γ∗. Note that these

uts an be optimised to enhan e the statisti al signi� an e of the signal but, on the

other hand, redu ing the ba kground too mu h makes its normalisation from real data

di� ult. The number of events ful�lling these requirements are olle ted in Table 2

(left) for the signal and main ba kgrounds (the rest of ba kgrounds are not expli itly

shown but we keep them in the al ulations). On the right panel we show the de om-

position of the ttnj ba kground in n = 0, . . . , 5 multipli ities. The dominan e of the

n = 1, 2, 3 subsamples results from the HT ut, whi h has a mu h higher reje tion for

lower jet multipli ities.

Pre. Pre.

Z ′

λ (0j) 493.2 Z ′

λ (1j) 1213.9

ttnj 64688 ttbb 657

tW 2425 Wnj 915

tj 445 Wbbnj 2202

Pre. a . HT

tt0j 7512 2.7%

tt1j 16665 9.1%

tt2j 16392 22.7%

tt3j 10299 43.6%

tt4j 4580 66.6%

tt5j 1601 84.6%

Table 2: Left: number of signal and ba kground events at the pre-sele tion level. Right:

de omposition of the ttnj ba kground in subsamples with n jets at the partoni level

and a eptan e of the HT ut. The luminosity is 30 fb

−1.

The �rst step to sear h for the Z ′signal is the re onstru tion of the top quark

pair. These are found by hoosing the best pairing between b jets and re onstru ted

W bosons:

1. The hadroni W is obtained with the two jets (among the three ones with largest

pT ) having an invariant mass losest to MW .

2. The leptoni W is obtained from the harged lepton and the missing energy,

7

0 50 100 150 200 250 300 350 400 450 500

mt

had

0

50

100

150

200E

vent

s / 5

GeV

Z′λSM bkg / 40

0 50 100 150 200 250 300 350 400 450 500

mt

lep

0

50

100

150

200

250

300

Eve

nts

/ 5 G

eV

Z′λSM bkg / 40

Figure 3: Re onstru ted top quark masses. The luminosity is 30 fb

−1.

identifying (pν)T = pT6 , requiring (pℓ+pν)2 =M2

W and solving for the longitudinal

omponent of the neutrino momentum. If no real solution exists, the neutrino

transverse momentum is de reased in steps of 1% and the pro edure is repeated.

If no solution is still found after 100 iterations, the dis riminant of the quadrati

equation is set to zero. Both solutions for the neutrino momentum are kept, and

the one giving best re onstru ted masses is sele ted.

3. The two top quarks are ea h re onstru ted with one of the W bosons and one

of the b jets, and are labelled as `hadroni ' and `leptoni ' orresponding to the

hadroni and leptoni W , respe tively.

4. The ombination minimising

(mhad

W −MW )2

σ2W

+(mlep

W −MW )2

σ2W

+(mhad

t −mt)2

σ2t

(mlep

t −mt)2

σ2t

(5)

is sele ted, with σW = 10 GeV, σt = 14 GeV [2℄.

We present the re onstru ted mass distributions at the pre-sele tion level in Fig. 3.

With the top quark pair identi�ed, one an onsider several variables to dis riminate

the signal from the ba kground. The most interesting ones are the tt, ttj invariant

masses and the transverse momentum of the tt pair. They are presented in Fig. 4 for

the two signal subsamples with zero (0j) and one (1j) jet at the partoni level and

for the SM ba kground. For ompleteness, we also show the transverse momentum

distribution for the top quark de aying leptoni ally (for the hadroni top quark it is

similar), whi h is not used to enhan e the signal signi� an e.

4

The most remarkable

features, whi h guide our subsequent analysis, are:

4

Although at pre-sele tion this variable seems to have a good dis riminating power for Z ′(0j)

events, after event sele tion riteria based on pttT

and other variables the signal and ba kground

8

0 500 1000 1500 2000m

tt

0

50

100E

vent

s / 2

5 G

eV

Z′λ (0j)

Z′λ (1j)

SM bkg / 40

0 500 1000 1500 2000m

ttj

0

20

40

60

80

Eve

nts

/ 25

GeV

Z′λ (0j)

Z′λ (1j)

SM bkg / 40

0 50 100 150 200 250 300 350 400

pT

tt

0

50

100

150

Eve

nts

/ 5 G

eV

Z′λ (0j)

Z′λ (1j)

SM bkg / 40

0 100 200 300 400 500 600 700 800

pT

t,lep

0

25

50

75

100

125

150

Eve

nts

/ 20

GeV

Z′λ (0j)

Z′λ (1j)

SM bkg / 40

Figure 4: Kinemati distributions at pre-sele tion. Up, left: tt invariant mass; up,

right: ttj invariant mass (requiring at least three light jets); down, left: transverse

momentum of the tt pair; down, right: transverse momentum of the top quark de aying

leptoni ally (not used for event sele tion). The luminosity is 30 fb

−1.

1. The mtt distribution has a large o�-peak omponent from the Z ′(1j) sample. If

the SM ba kground annot be predi ted with very good a ura y (as it seems the

ase for tt plus several hard jets), and must be normalised from o�-peak data,

this tail will behave as a ombinatorial ba kground redu ing the peak signi� an e

unless a spe i� analysis is arried out with a di�erent re onstru tion for these

events.

2. The Z ′(1j) sample exhibits a good peak in the ttj invariant mass distribution,

although slightly shifter towards mttj values larger than 1 TeV.

3. We also observe that the tt transverse momentum is typi ally mu h smaller for the

distributions be ome very similar. Nevertheless, this and other distributions, as for example the tt

pair rapidity, an be implemented in a likelihood fun tion to a hieve a better dis riminating power

than with the ut-based analysis implemented here. Su h optimisation is beyond the s ope of the

present work.

9

Z ′(0j) signal than for the ba kground. Although tt pairs are typi ally produ ed

with low transverse momentum, the requirement HT > 750 GeV has a higher

suppression for lower hard jet multipli ities, as seen in Table 2. As a result of

this ut, the pttT distribution is shifted towards large values, as it an observed in

Fig. 5. This higher jet multipli ity is also the reason for the worse re onstru tion

of the hadroni top in Fig. 3.

0 50 100 150 200 250 300 350 400

pT

tt

0

0,05

0,1

0,15

0,2E

vent

s (n

orm

alis

ed)

ttnj (no cut)

ttnj (HT > 750 GeV)

Figure 5: tt transverse momentum distribution for the ttnj ba kground at the pre-

sele tion level with and without the HT ut. The luminosity is 30 fb

−1.

The best strategy to maximise the signal signi� an e is to develop two analyses. The

�rst one aims to re onstru t the tt peak in the region of low transverse momenta

of the tt pair setting pttT < 50 GeV, whi h eliminates a large fra tion of ba kground

and the o�-peak signal ontribution whi h would otherwise onstitute a ombinatorial

ba kground. The se ond analysis will sear h for the ttj peak in the omplementary

region pttT > 50 GeV. We present these two analyses in turn. Then, for omparison, we

will perform an in lusive analysis without separate re onstru tions.

4.1 Analysis I

As sele tion riteria we impose some loose quality uts on re onstru ted top quark

masses and, more importantly, small transverse momentum for the tt pair,

125 GeV < mhad

t , mlep

t < 225 GeV ,

pttT < 50 GeV . (6)

The number of signal and ba kground events with these uts an be read in Table 3.

We observe that the pttT ut signi� antly redu es the ttnj ba kground and removes 90%

10

Sel. Peak Sel. Peak

Z ′

λ (0j) 334.5 299.2 Z ′

λ (1j) 149.0 114.4

ttnj 6581 1652 ttbb 15 3

tW 85 17 Wnj 39 13

tj 9 4 Wbbnj 15 4

Table 3: Number of signal and ba kground events at the sele tion level (analysis I) and

at the tt resonan e peak. The luminosity is 30 fb

−1.

0 500 1000 1500 2000m

tt

0

100

200

300

400

500

Eve

nts

/ 25

GeV

SM bkg + Z′λSM bkg

Figure 6: tt invariant mass for the SM ba kground and the ba kground plus the Z ′

signal at the sele tion level (analysis I). The luminosity is 30 fb

−1.

of the Z ′signal with n = 1, in parti ular the o�-peak ontribution. The presen e of

the Z ′resonan e an be spotted with the analysis of the tt invariant mass distribution,

presented in Fig. 6 for the ba kground alone and the ba kground plus the signal at the

sele tion level. The number of events at the peak

900 GeV < mtt < 1100 GeV (7)

an be read in Table 3. We will assume that we normalise the ba kground by o� peak

measurements, obtaining a s aling fa tor κ = 1.05 determined from the omparison of

the distributions for signal plus ba kground and ba kground alone. Then, the statisti al

signi� an e of the peak is S ′/√B′, where

B′ = κB ⇒ S ′ = (S +B)− B′(8)

and S, B are the true numbers of signal and ba kground events. For the peak region

in Eq. (7), the ex ess of events amounts to 7.78σ.

11

4.2 Analysis II

The main motivation for this analysis is the fa t that many of the Z ′(1j) events do not

exhibit a peak in the tt invariant mass distribution and would fall o� the peak region

in Eq. (7). This is seen learly in Fig. 4 (up, left). When the invariant mass of the tt

pair plus the hardest additional jet is onsidered (requiring in this ase a minimum of

three jets) the distribution exhibits a lear peak, although slightly displa ed, as shown

on the upper right panel of that �gure. We then on entrate on the omplementary

event sample for this analysis, requiring at least three light jets and setting the uts

125 GeV < mhad

t , mlep

t < 225 GeV ,

pttT > 50 GeV ,

min(∆Rj,thad

,∆Rj,tlep

) < 1.6 (9)

as sele tion riteria. The last one is implemented to redu e the ba kground, sin e the

signal events peaking at mttj ∼ MZ′

λare produ ed by FSR in Z ′ → tt. The number

of events after these uts are given in Table 4. The ttj distribution for the signal plus

Sel. Peak Sel. Peak

Z ′

λ (0j) 52.2 41.7 Z ′

λ (1j) 412.8 293.5

ttnj 14226 4432 ttbb 77 18

tW 229 60 Wnj 43 17

tj 20 9 Wbbnj 77 25

Table 4: Number of signal and ba kground events at the sele tion level (analysis II)

and at the ttj resonan e peak. The luminosity is 30 fb

−1.

ba kground and ba kground alone is presented in Fig. 7. The number of events at the

peak

900 GeV < mttj < 1200 GeV (10)

an be read in Table 4. The ba kground s aling fa tor in this ase is κ = 1.022, and the

ex ess of events over the (normalised) SM ba kground expe tation amounts to 3.56σ.

4.3 In lusive analysis

In order to see the advantage of separate, dedi ated analyses for Z ′ → tt, Z ′ → ttj, we

also perform the in lusive analysis for the two Z ′signal omponents, sear hing for a

12

0 500 1000 1500 2000m

ttj

0

100

200

300

400

500

600

700

800

900

Eve

nts

/ 25

GeV

SM bkg + Z′λSM bkg

Figure 7: ttj invariant mass for the SM ba kground and the ba kground plus the Z ′

signal at the sele tion level (analysis II). The luminosity is 30 fb

−1.

peak in the mtt distribution. For event sele tion we only require a good re onstru tion

of the top quark pair,

125 GeV < mhad

t , mlep

t < 225 GeV , (11)

and drop the other kinemati uts. The number of signal and ba kground events an

be read in Table 5.

Sel. Peak Sel. Peak

Z ′

λ (0j) 400.1 332.6 Z ′

λ (1j) 727.1 361.0

ttnj 35203 4196 ttbb 163 9

tW 567 40 Wnj 170 20

tj 109 12 Wbbnj 263 13

Table 5: Number of signal and ba kground events at the sele tion level (in lusive

analysis) and at the tt resonan e peak. The luminosity is 30 fb

−1.

The tt distribution for the signal plus ba kground and ba kground alone is presented

in Fig. 8. The number of events at the peak

900 GeV < mtt < 1100 GeV (12)

is also given in Table 5. As expe ted, in this ase the ba kground s aling fa tor is

larger than for the other analyses, κ = 1.055, and the peak signi� an e is of 6.77σ.

This number is 25% smaller than the ombined signi� an es of the other two analyses,

8.55σ.

13

0 500 1000 1500 2000m

tt

0

500

1000

1500

2000

Eve

nts

/ 25

GeV

SM bkg + Z′λSM bkg

Figure 8: tt invariant mass for the SM ba kground and the ba kground plus the Z ′

signal at the sele tion level (in lusive analysis). The luminosity is 30 fb

−1.

5 Con lusions

We provide a new ode implemented in ALPGEN for neutral ve tor resonan e pro-

du tion, in luding the higher order orre tions from real emission and the virtual LL

ontributions. In order to illustrate its features we have generated tt and ttj events by

the ex hange of a leptophobi gauge boson based on E6 at LHC. The analysis of the

Z ′

λ → tt and Z ′

λ → ttj samples show that su h a program is ne essary to a ount for

∼ 23% of the ttX events. Indeed, for these events the resonan e is not found in the

tt invariant mass distribution but on the ttj invariant mass. Therefore, their presen e

annot be a ounted for by a K fa tor multiplying the LO Z ′

λ → tt ross se tion but

has to be properly simulated at the generator level.

We have shown that the LHC dis overy potential for neutral ve tor resonan es

in the ttX �nal state bene�ts from a separate analysis for tt and ttj. For the ase

examined (with a Z ′

λ mass of 1 TeV), the enhan ement over an in lusive sear h for tt

resonan es is of a 25% in the statisti al signi� an e. This improvement is expe ted to

be maintained for larger masses, and of ourse for other Z ′models than the leptophobi

Z ′

λ one used in the simulations. For other tt resonan es the trend is expe ted to be the

same as well.

Finally, the in lusion of orre tions to Z ′ → tt not only translates into a better

dis overy potential but also into a proper des ription of the kinemati al distributions.

A ode implementing them is ne essary to establish the model on e a new ve tor

resonan e is dis overed, dis riminating among models by a more pre ise measurement

of the angular distributions [30℄ and the di�erent Z ′ ouplings [8, 19℄.

14

A knowledgements

This work has been partially supported by MICINN (FPA2006-05294), by Junta de

Andalu ía (FQM 101, FQM 437 and FQM 03048), and by the European Community's

Marie-Curie Resear h Training Network under ontra t MRTN-CT-2006-035505 �Tools

and Pre ision Cal ulations for Physi s Dis overies at Colliders�. The work of J.A.A.S.

and M.T. has been supported by MICINN Ramón y Cajal and Juan de la Cierva on-

tra ts, respe tively. F.P. and M.M. a knowledge warm hospitality and partial support

from CERN, Physi s Department, TH Unit, during the ompletion of this work.

Referen es

[1℄ See for a re ent review P. Langa ker, arXiv:0801.1345 [hep-ph℄; and referen es

there in. See also A. Leike, Phys. Rept. 317 (1999) 143 [arXiv:hep-ph/9805494℄;

C. Amsler et al. [Parti le Data Group℄, Phys. Lett. B 667 (2008) 1.

[2℄ G. Aad et al. [The ATLAS Collaboration℄, arXiv:0901.0512 [hep-ex℄.

[3℄ G. L. Bayatian et al. [CMS Collaboration℄, J. Phys. G 34 (2007) 995.

[4℄ For a omparison between Z → ℓ+ℓ− generators and a study of QCD and ele -

troweak orre tions see C. Buttar et al., arXiv:0803.0678 [hep-ph℄.

[5℄ V. M. Abazov et al. [D0 Collaboration℄, Phys. Rev. Lett. 100 (2008) 142002

[arXiv:0712.0851 [hep-ex℄℄; T. Aaltonen et al. [CDF Collaboration℄, Phys. Rev.

Lett. 101 (2008) 202001 [arXiv:0806.2472 [hep-ex℄℄.

[6℄ P. Langa ker, R. W. Robinett and J. L. Rosner, Phys. Rev. D 30 (1984) 1470.

[7℄ J. L. Hewett and T. G. Rizzo, Phys. Rept. 183 (1989) 193.

[8℄ See for a re ent review R. Diener, S. Godfrey and T. A. W. Martin,

arXiv:0910.1334 [hep-ph℄; and referen es there in.

[9℄ I. Antoniadis, K. Benakli and M. Quiros, Phys. Lett. B 331 (1994) 313

[arXiv:hep-ph/9403290℄; T. G. Rizzo, Phys. Rev. D 61 (2000) 055005

[arXiv:hep-ph/9909232℄; A. Djouadi, G. Moreau and R. K. Singh, Nu l. Phys.

B 797 (2008) 1 [arXiv:0706.4191 [hep-ph℄℄.

15

[10℄ M. L. Mangano, M. Moretti, F. Pi inini, R. Pittau and A. D. Polosa, JHEP 0307

(2003) 001 [arXiv:hep-ph/0206293℄; F. Caravaglios, M. L. Mangano, M. Moretti

and R. Pittau, Nu l. Phys. B 539 (1999) 215 [arXiv:hep-ph/9807570℄.

[11℄ F. del Aguila and J. A. Aguilar-Saavedra, JHEP 0711 (2007) 072 [arXiv:0705.4117

[hep-ph℄℄.

[12℄ M. L. Mangano, �The so- alled MLM pres ription for ME/PS mat hing�, talk

given at the Fermilab ME/MC Tuning Workshop, O tober, 2004.

[13℄ M. L. Mangano, M. Moretti, F. Pi inini and M. Tre ani, JHEP 0701, 013 (2007)

[arXiv:hep-ph/0611129℄.

[14℄ F. del Aguila, L. Ametller and R. Pittau, Phys. Lett. B 628 (2005) 40

[arXiv:hep-ph/0507262℄; PoS HEP2005 (2006) 311 [arXiv:hep-ph/0511243℄.

[15℄ ALEPH, DELPHI, L3 and OPAL Collaborations and LEP ele troweak working

group, CERN-PPE-95-172.

[16℄ P. Chiappetta, J. Layssa , F. M. Renard and C. Verzegnassi, Phys. Rev.

D 54 (1996) 789 [arXiv:hep-ph/9601306℄; G. Altarelli, N. Di Bartolomeo,

F. Feruglio, R. Gatto and M. L. Mangano, Phys. Lett. B 375 (1996) 292

[arXiv:hep-ph/9601324℄; V. D. Barger, K. m. Cheung and P. Langa ker, Phys.

Lett. B 381 (1996) 226 [arXiv:hep-ph/9604298℄; G. Montagna, F. Pi inini,

J. Layssa , F. M. Renard and C. Verzegnassi, Z. Phys. C 75 (1997) 641

[arXiv:hep-ph/9609347℄.

[17℄ K. S. Babu, C. F. Kolda and J. Mar h-Russell, Phys. Rev. D 54 (1996) 4635

[arXiv:hep-ph/9603212℄.

[18℄ F. Gursey, P. Ramond and P. Sikivie, Phys. Lett. B 60 (1976) 177.

[19℄ F. del Aguila, M. Cveti and P. Langa ker, Phys. Rev. D 52 (1995) 37

[arXiv:hep-ph/9501390℄; see also F. del Aguila, A ta Phys. Polon. B 25 (1994)

1317 [arXiv:hep-ph/9404323℄; S. Godfrey and T. A. W. Martin, Phys. Rev. Lett.

101 (2008) 151803 [arXiv:0807.1080 [hep-ph℄℄.

[20℄ F. del Aguila, G. A. Blair, M. Daniel and G. G. Ross, Nu l. Phys. B 283 (1987)

50; F. del Aguila, M. Quiros and F. Zwirner, Nu l. Phys. B 287 (1987) 419.

[21℄ Y. Hosotani, Phys. Lett. B 126 (1983) 309; Y. Hosotani, Phys. Lett. B 129 (1983)

193; E. Witten, Phys. Lett. B 149 (1984) 351.

16

[22℄ J. A. Aguilar-Saavedra, Nu l. Phys. B 828 (2010) 289 [arXiv:0905.2221 [hep-ph℄℄.

[23℄ H. L. Lai et al. [CTEQ Collaboration℄, Eur. Phys. J. C 12 (2000) 375

[arXiv:hep-ph/9903282℄.

[24℄ D0 Collaboration, D0 Note 5882-CONF

[25℄ J. A. Aguilar-Saavedra, JHEP 0911 (2009) 030 [arXiv:0907.3155 [hep-ph℄℄.

[26℄ T. Sjostrand, S. Mrenna and P. Skands, JHEP 0605 (2006) 026 [hep-ph/0603175℄.

[27℄ J. Alwall et al., Eur. Phys. J. C 53, 473 (2008) [arXiv:0706.2569 [hep-ph℄℄.

[28℄ E. Ri hter-Was, hep-ph/0207355.

[29℄ J. M. Campbell and R. K. Ellis, Phys. Rev. D 65, 113007 (2002)

[arXiv:hep-ph/0202176℄; Phys. Rev. D 62, 114012 (2000) [arXiv:hep-ph/0006304℄;

Phys. Rev. D 60, 113006 (1999) [arXiv:hep-ph/9905386℄.

[30℄ See, for example, W. Bernreuther, M. Fles h and P. Haberl, Phys. Rev. D 58

(1998) 114031 [arXiv:hep-ph/9709284℄; W. Bernreuther, A. Brandenburg and

M. Fles h, arXiv:hep-ph/9812387; B. Lillie, J. Shu and T. M. P. Tait, Phys. Rev.

D 76 (2007) 115016 [arXiv:0706.3960 [hep-ph℄℄; R. Frederix and F. Maltoni, JHEP

0901 (2009) 047 [arXiv:0712.2355 [hep-ph℄℄.

17