Supersymmetry Dileptons and Trileptons at the Tevatron

HEP-PH-9412346 4 y y y h C) vit vit CT-23/94 A AMU-66/94 hargino h should text of a and harginos in hep-ph/9412346 e also discuss CTP-T , whic CERN-TH.7535/94 hes in c 1 , and A. ZICHICHI (1) sup ergra 2 ; U ). W 1 fb 1 ch Center (HAR exas A&M University atron ar fb (5) y via hadronically quiet ANG ese and 1 ev (1 SU 1 1 dR T tal sensitivities for dileptons t deca e pb pb ,XUW e nd that the dilepton mo de is 3 ; 2 ; d lands, TX 77381, USA dileptons dvanc 1 o for 100 the y mo del, and eV atron has completed its short-term searc vit G artment of Physics, T ev Abstract at hing ratios. The estimated reac e p erform our computations in the con W oup, Houston A (5) sup ergra ge Station, TX 77843{4242, USA ory Division, 1211 Geneva 23, Switzerland atron, and their subsequen o- and one-parameter mo duli and dilaton scenarios. Our y mo dels, including generic four-parameter sup ergra w t to the trilepton signal when the latter is suppressed b SU ev tegrated luminosities of 100 CERN, 1211 Geneva 23, Switzerland al Physics, Dep vit 4 Col le , D. V. NANOPOULOS 2 etic ; trasted with estimated exp erimen 1 or CERN The 3 The Mitchel l Campus, The Wo article Physics Gr trileptons e consider the pro duction of sup ersymmetry neutralinos and c y of sup ergra ailable in the short and long terms. W v op W collisions at the T AMU-66/94 p b er 1994 ariet Sup ersymmetry for dilepton and trilepton pro duction. masses can b e asthe large task as left 100 (150)for LEPI I once the T small neutralino leptonic branc and trileptons for in be a a needed complemen with string inspired t results are con mo dels, the minimal p dileptons and trileptons. v Astr 2 GE L. LOPEZ Center for The 1 CT-23/94 CERN-TH.7535/94 CTP-T A Decem JOR

Exp erimental searches for sup ersymmetric particles have come a long way since

the commissioning of the Tevatron pp collider at Fermilab (1988) and the LEP e e

collider at CERN (1989). The strengths and weaknesses of these twotyp es of colliders

are well known. A hadron collider is b est suited for searching the highest accessible

mass scales since a sharp kinematical limit do es not exist, but discoverability dep ends

on the event rate and the cleanliness of the signal. An e e collider is capable of

discovery essentially up to the kinematical limit, but this is muchlower than what

would b e accessible in a hadron collider. In fact, it has b ecome apparent that an e e

linear collider with a center-of-mass energy in the multi-hundred GeV range would b e

ideal for what has b een termed \sparticle sp ectroscopy". At present though, in the

search for new physics wehave to make the b est p ossible use of existing facilities, since

information gathered there would illuminate the path towards higher energy machines.

One such e ort is b eing conducted at the Tevatron, where the search for weakly

interacting sparticles (charginos and neutralinos) has b ecome quite topical, in view of

the fact that the reach of the machine for the traditional strongly interacting sparticles

has b een nearly reached. This e ort will b ene t from an integrated luminosityin

1 1

excess of 100 pb by the end of the ongoing Run IB, and p ossibly 1{2 fb during

the Main Injector era around the year 2000.

In this pap er we study the prosp ects for sup ersymmetry discovery at the Teva-

tron via the hadronically quiet dilepton and trilepton signals which o ccur in the pro-

duction and decayofcharginos and neutralinos in pp collisions [1, 2 , 3 ]. Our previous

study [3] considered the trilepton signal with an estimated 100 pb of accumulated

data. Here we up date this analysis by incorp orating the latest exp erimental informa-

tion on the trilepton signal and prosp ects for its detection with O (1 fb ) data. We

also include for the rst time the dilepton signal, which has b een recently shown to

b e exp erimentally extractable [4], and as we discuss, has the advantage of allowing a

signi cant exploration of the parameter space for chargino masses in the LEPI I acces-

sible range. Our calculations are p erformed in the context of a broad class of uni ed

sup ergravity mo dels, whichhave the virtue of having the least numb er of free pa-

rameters and are therefore highly predictive and straightforwardly testable through a

variety of correlated phenomena at di erent exp erimental facilities. Our study should

give a go o d idea of the range of p ossibilities op en to exp erimental investigation, and

allow quantitativechecks of sp eci c mo dels which yield the largest rates.

We consider uni ed sup ergravity mo dels with universal soft sup ersymmetry

breaking at the uni cation scale, and radiative electroweak symmetry breaking (en-

forced using the one-lo op e ective p otential) at the weak scale [5]. These constraints

reduce the numb er of parameters needed to describ e the mo dels to four, which can

b e taken to b e m ; m =m ; A=m ; tan , with a sp eci ed value for

0 0 A

1=2 1=2

pole

the top-quark mass (m ). In what follows we take m = 160 GeV whichisthe

central value obtained in ts to all electroweak and Tevatron data in the context of

sup ersymmetric mo dels [6]. Of relevance to our discussion, we note that in all mo dels

These signals contain no hadronic activity, except for initial state radiation e ects, and are thus

distinct from the usual multilepton signals in squark and gluino pro duction. 1

considered the following relation holds to various degrees of approximation

0 0

m m 2m : (1)

2 1

Among these four-parameter sup ersymmetric mo dels we consider generic mo dels with

continuous values of m and discrete choices for the other three parameters:

tan =2;10 ; =0;1;2;5; =0 : (2)

0 A

The choices of tan are representative; higher values of tan are likely to yield

values of B (b ! s ) in con ict with present exp erimental limits [7]. The choices

of corresp ond to m (0:8; 0:9; 1:1; 1:9)m . The choice of A has little impact

0 q~ g~

on the results. We also consider the case of minimal SU (5) sup ergravity, where in

addition we imp ose the constraints from proton decay and cosmology (a not to o young

Universe). The parameter space in this case is still four-dimensional, but restricted

< > <

to tan 10, 4, and m 120 GeV [8].

We also consider the case of no-scale SU (5)U (1) sup ergravity [5]. In this class

of mo dels the sup ersymmetry breaking parameters are related in a string-inspired

way. In the two-parameter moduli scenario = =0 [9], whereas in the dilaton

0 A

scenario = ; =1 [10]. We also compute the rates in the one-parameter

0 A

mo duli (B (M ) = 0) and dilaton (B (M )=2m ) scenarios, where M is the string

U U 0 U

uni cation scale, and with this extra condition tan is determined as a function

of m , which is the only free parameter in the mo del. A series of exp erimental

constraints and predictions for these mo dels have b een given in Refs. [11 ] and [12],

resp ectively.

The pro cesses of interest are

Trileptons: pp ! , where the next-to-lightest neutralino decays leptoni-

0 0 + 0

cally ( ! ` ` ), and so do es the lightest chargino ( ! ` ). The

2 1 1

cross section pro ceeds via s-channel exchange of an o -shell W and (small) t-

M , and otherwise falls channel squark exchange, and thus p eaks at m

o smo othly with increasing chargino masses with a small tan dep endence.

Dileptons: pp ! , where b oth charginos decay leptonically. The cross

1 1

section pro ceeds via s-channel exchange of o -shell Z and and t-channel

M . Dileptons could also come from squark exchange, and p eaks for m

0 0 0 0 0

pp ! ; , with the appropriate leptonic or invisible decays of . Both

1 2 2 2 2

of these pro cesses are negligible [2] b ecause the couplings of the Z and to

neutralinos are highly suppressed when the neutralinos have a high gaugino

content, as is the case when Eq. (1) holds. Yet another source of dileptons

via pp ! e~ e~ su ers from small rates for selectron masses b eyond the LEP

R R

limit [13 ]. 2

The more imp ortant factors in the dilepton and trilepton yields are the leptonic

branching fractions which can vary widely throughout the parameter space [3]. If all

sparticles are fairly heavy, the decay amplitude is dominated by W or Z exchange. In

this case the branching fractions into electrons plus muons are B ( ! ` ) 2=9

0 0 +

and B ( ! ` ` ) 6%. On the other hand, if some of the sparticles are rela-

2 1

tively light, most likely the sleptons, the branching fractions are altered. The extreme,

although not unusual, case o ccurs when the sleptons are on-shell. These two-body

decays then dominate and the chargino leptonic branching fraction is maximized,

i.e., B ( ! ` ) =2=3. Light sleptons also a ect the neutralino leptonic

l max

branching ratio. When the sneutrino is on-shell and is lighter than the corresp ond-

ing right-handed charged slepton (e ~ ; ~ ), the channel ! ~ dominates the

R R l l

amplitude, and the neutralino leptonic branching ratio is suppressed. This situation

is reversed when the charged slepton is on-shell and is lighter than the sneutrino,

which leads to an enhancement of the neutralino leptonic branching ratio. For su-

ciently high neutralino and chargino masses, b oth leptonic branching ratios decrease

b ecause the W and Z go on-shell and dominate the decay amplitudes. In the case

0 0

of the neutralino, the sp oiler mo de ! h also b ecomes kinematically allowed.

2 1

These high-mass suppressions do not kickinuntil chargino and neutralino masses

m m 2M ; 2m 200 GeV .

Z h

The results of our computations for the various mo dels are shown in Figs. 1,2,

3,4,5,6. The various curves in the gures terminate at the low end b ecause of various

parameter space constraints, whereas at the high end they are cuto when the yields

fall b elow the foreseeable sensitivity. In most cases we note that the rates are higher

for <0. This is a consequence of suppressed branching fractions for >0, but also

a generally smaller allowed parameter space which requires minimum values of the

chargino mass whichmay exceed signi cantly the present exp erimental lower limit.

M , whereas the We can also observe that the dilepton rates indeed p eak near

trilepton rates are not as large for lightchargino masses since they p eak at M .It

is also evident in the gures that for chargino masses b elow 100 GeV , the trilepton

rates can b e highly suppressed, while the dilepton rates are not, thus pro ducing

a rather complementary e ect. This \threshold" phenomenon is most evidentin

Figs. 4,5,6 and, as dicussed ab ove, corresp onds to a suppression of the neutralino

leptonic branching ratio for light sleptons, i.e., when ! ~ is allowed.

The signi cance of our results is quanti ed by the horizontal dashed lines in

the gures, which represent estimates of the exp erimental sensitivity to b e reached

1 1

with 100 pb (upp er lines) and 1 fb (lower lines). The lesser sensitivity should

be achievable at the end of Run IB during 1995 (i.e., prior to the LEPI I upgrade),

whereas the higher sensitivity should b e available with the Main Injector upgrade in

1999 (i.e., after the LEPI I shutdown but b efore the LHC commissioning).

The trilepton sensitivity with 100 pb (i.e.,0:4pb ) has b een estimated by

. =m

~ e~ ~ e~

L L R R

4. This is why For = 5, radiative electroweak symmetry breaking is only p ossible for tan

there is no curve for = 5 in Fig. 2 (tan = 10), whereas there is such a curve in Fig. 1 (tan = 2).

0 3

Table 1: Estimated chargino mass reaches in various sup ergravity mo dels for chargino-

neutralino pro duction in pp collisions at the Tevatron via dilepton and trilepton mo des

1 1

for integrated luminosities of 100 pb and 1 fb . All masses in GeV. Dashes ({)

indicate negligible sensitivity.

generic >0 <0

1 1 1 1

tan 100 pb 1 fb 100 pb 1 fb

2 0 { 120 100 145

1 { 125 75 115

2 { 100 65 100

5 { 80 55 80

10 0 { 105 70 135

1 70 95 65 100

2 { 70 { 70

Mo del >0 <0

1 1 1 1

tan 100 pb 1 fb 100 pb 1 fb

mo duli 2 { 115 100 150

(2-par) 6 75 160 100 150

10 75 160 70 150

dilaton 2 95 135 80 120

(2-par) 6 80 130 80 120

10 80 125 80 120

mo duli (1-par) N/A N/A 70 150

dilaton (1-par) N/A N/A 80 125

minimal SU (5) 50 80 50 75

simply scaling down by a factor of 5 the present exp erimental limit of 2 pb obtained

with 20 pb of recorded data [14]. The factor of 5 is the exp ected increase in recorded

luminosity, and a simple L scaling is appropriate assuming the trilepton signal has

no Standard Mo del backgrounds at this level of sensitivity. The sensitivityat1fb

requires a study of the background since small Standard Mo del pro cesses and detector-

dep endent instrumental backgrounds b ecome imp ortant at this level of sensitivity [15].

The sensitivity in the gures (i.e.,0:07 pb ) is obtained by scaling up by L the value

given in Table I I of Ref. [15].

The dilepton (plus p= ) signal su ers from several Standard Mo del backgrounds,

most notably Z ! and WW pro duction. A study based on the D0 detector [4]

reveals that with suitable cuts, in 100 pb an estimated background of 8 events is

exp ected, whichwould require 8 signal events at 3 signi cance. The eciencies

for dilepton detection have also b een studied [4], and they improve with increasing

chargino masses, 8% is a typical value. All this implies a sensitivityof1pb for dilepton 4

detection. With 1 fb one can scale down the sensitivity with L, obtaining a

sensitivityof0:3pb .

The reaches in chargino masses in the various mo dels, can b e readily obtained

from the gures by considering b oth dilepton and trilepton signals, and are summa-

rized in Table 1 for the twointegrated luminosity scenarios. The reaches in Table 1

translate into indirect reaches in every other sparticle mass, since they are all related.

In particular, m 0:3m and m (m =2:9) 6+ (in the SU (5) U (1) mo dels

g~ q~ g~

the numerical co ecients in this relation are slightly di erent, implying m m ).

q~ g~

It is also interesting to p oint out that the pattern of yields for the various mo dels

is quite di erent, therefore observation of a signal will disprove many of the mo dels,

while supp orting a small subset of them.

It has b een p ointed out that the dilepton and trilepton data sample maybe

enhanced by considering presumed trilepton events where one of the leptons is either

missed or has a p b elow 5 GeV (\2-out-of-3") [16]. Such enhancements would alter

our reach estimates ab ove, making them even more promising.

From Table 1 it is clear that in some regions of parameter space, the reach of the

Tevatron for chargino masses is quite signi cant. With 100 pb it should b e p ossible

to prob e chargino masses as high as 100 GeV in the generic mo dels for tan =2,

=0,and<0, and in the two-parameter SU (5) U (1) mo duli scenario for

tan 10. More generally, the accessible region of parameter space should overlap

with that within the reach of LEPI I, although it would have b een explored b efore

LEPI I turns on. However, LEPI I has an imp ortant task: chargino searches at LEPI I

will not b e hindered by small branching fractions, and thus a more mo del-indep endent

lower limit on the chargino mass should b e achievable, i.e., m s. Wewould

like to conclude with Fig. 6, where we show the predictions for the dilepton and

trilepton rates in our chosen one-parameter SU (5) U (1) mo dels, which are the

most predictive sup ersymmetric mo dels to date. It is interesting to note that in the

mo duli scenario, the mass reach for charginos could b e as high as 150 GeV with an

integrated luminosityof1fb .Further prop osed increases in luminosity or center-

of-mass energy of the Tevatron collider have the p otential of probing even deep er into

the parameter space [15 ]. We conclude that detection of weakly interacting sparticles

at the Tevatron maywell bring the rst direct signal for sup ersymmetry.

Acknowledgments

Wewould like to thank James White for motivating this study and for providing us

with valuable information ab out the sensitivity of the dilepton signal. This work has

b een supp orted in part by DOE grant DE-FG05-91-ER-4 06 33. The work of X. W.

has b een supp orted by the World Lab oratory.

At LEPI I it might b e p ossible to extend the indirect reach for charginos by studying the pro cess

+ 0 0 0 0

e e ! with ! +2j. Equation (1) implies a kinematical reachofm s.

1 2 2 1

1 5

References

[1] J. Ellis, J. Hagelin, D. V. Nanop oulos, and M. Srednicki, Phys. Lett. B 127 (1983)

233; P. Nath and R. Arnowitt, Mo d. Phys. Lett. A 2 (1987) 331; R. Barbieri, F.

Caravaglios, M. Frigeni, and M. Mangano, Nucl. Phys. B 367 (1991) 28.

[2] H. Baer and X. Tata, Phys. Rev. D 47 (1993) 2739; H. Baer, C. Kao, and X.

Tata, Phys. Rev. D 48 (1993) 5175.

[3] J. L. Lop ez, D. V. Nanop oulos, X. Wang, and A. Zichichi, Phys. Rev. D 48

(1993) 2062.

[4] J. White and D. Norman, D0 Note 2395.

[5] For a recent review see e.g., J. L. Lop ez, D. V. Nanop oulos, and A. Zichichi,

Prog. Part. Nucl. Phys. 33 (1994) 303.

[6] J. Ellis, G.L. Fogli and E. Lisi, Phys. Lett. B 333 (1994) 118; J. Erler and P.

Langacker, UPR{0632T (hep-ph/9411203 ).

[7] J. L. Lop ez, D. V. Nanop oulos, X. Wang, and A. Zichichi, Phys. Rev. D 51

(1995) 147.

[8] J. L. Lop ez, D. V. Nanop oulos, and H. Pois, Phys. Rev. D 47 (1993) 2468; J.

L. Lop ez, D. V. Nanop oulos, H. Pois, and A. Zichichi, Phys. Lett. B 299 (1993)

262; J. L. Lop ez, D. V. Nanop oulos, and K. Yuan, Phys. Rev. D 48 (1993) 2766.

[9] J. L. Lop ez, D. V. Nanop oulos, and A. Zichichi, Phys. Rev. D 49 (1994) 343.

[10] J. L. Lop ez, D. V. Nanop oulos, and A. Zichichi, Phys. Lett. B 319 (1993) 451.

[11] J. L. Lop ez, D. V. Nanop oulos, G. Park, X. Wang, and A. Zichichi, Phys. Rev.

D 50 (1994) 2164.

[12] J. L. Lop ez, D. V. Nanop oulos, and A. Zichichi, TexasA&MUniversity preprint

CTP-TAMU-27/94 (hep-ph/9408345).

[13] H. Baer, C. Chen, F. Paige, and X. Tata, Phys. Rev. D 49 (1994) 3283.

[14] CDF Collab oration, Y. Kato, in \Pro ceedings of the 9th Topical Workshop on

Proton-Antiproton Collider Physics," Tsukuba, Japan, Octob er 1993, edited by

K. Kondo and S. Kim (Universal Academy Press, Tokyo), p. 291; CDF Collab-

oration, Fermilab-Conf-94/149-E (1994), to app ear in Pro ceedings of the 27th

International Conference on High Energy Physics, Glasgow, July 20{27, 1994.

[15] T. Kamon, J. L. Lop ez, P. McIntyre, and J. White, Phys. Rev. D 50 (1994)

5676.

[16] J. White, private communication. 6

Figure 1: The dilepton and trilepton rates at the Tevatron versus the chargino mass

in a generic uni ed sup ergravity mo del with tan =2, =0;1;2;5 (as indicated),

and A = 0. The upp er (lower) dashed lines represent estimated reaches with 100 pb

(1 fb ) of data. 7

Figure 2: The dilepton and trilepton rates at the Tevatron versus the chargino mass

in a generic uni ed sup ergravity mo del with tan = 10, =0;1;2 (as indicated),

and A = 0. The upp er (lower) dashed lines represent estimated reaches with 100 pb

(1 fb ) of data. 8

Figure 3: The dilepton and trilepton rates at the Tevatron versus the chargino mass

in the minimal SU (5) sup ergravity mo del (where tan <10, > 4). The upp er

1 1

(lower) dashed lines represent estimated reaches with 100 pb (1 fb ) of data. 9

Figure 4: The dilepton and trilepton rates at the Tevatron versus the chargino mass

in two-parameter SU (5) U (1) sup ergravity { mo duli scenario ( = = 0) for the

0 A

indicated values of tan . The upp er (lower) dashed lines represent estimated reaches

1 1

with 100 pb (1 fb ) of data. 10

Figure 5: The dilepton and trilepton rates at the Tevatron versus the chargino mass

; = 1) in two-parameter SU (5) U (1) sup ergravity { dilaton scenario ( =

A 0

for the indicated values of tan . The upp er (lower) dashed lines represent estimated

1 1

reaches with 100 pb (1 fb ) of data. 11

Figure 6: The dilepton and trilepton rates at the Tevatron versus the chargino mass

in one-parameter SU (5) U (1) sup ergravity { mo duli and dilaton scenarios (<0in

b oth cases). The upp er (lower) dashed lines represent estimated reaches with 100 pb

(1 fb ) of data. 12