New signature for color octet pseudoscalars at the CERN LHC

Alfonso R. Zerwekh*

Centro de Estudios Subatómicos and Instituto de Fı́sica, Facultad de Ciencias, Universidad Austral de Chile,
Casilla 567, Valdivia, Chile

Claudio O. Dib+

Centro de Estudios Subatómicos and Departamento de Fı́sica, Universidad Técnica Federico Santa Marı́a, Valparaı́so, Chile

Rogerio Rosenfeld‡

Instituto de Fı́sica Teórica - UNESP, Rua Pamplona, 145, 01405-900, São Paulo, SP, Brazil
(Received 11 March 2008; published 29 May 2008)

Color octet (pseudo)scalars, if they exist, will be copiously produced at the CERN Large Hadron

Collider (LHC). However, their detection can become a very challenging task. In particular, if their decay

into a pair of top quarks is kinematically forbidden, the main decay channel would be into two jets, with a

very large background. In this brief report we explore the possibility of using anomaly-induced decays of

the color octet pseudoscalars into gauge bosons to find them at the LHC.

DOI: 10.1103/PhysRevD.77.097703 PACS numbers: 12.60.Nz, 13.85.Qk

I. INTRODUCTION

In spite of its great experimental successes [1], the
standard model (SM) of the electroweak interactions is
still widely regarded as an incomplete theory. The reasons
are manifold, including the existence of nonbaryonic dark
matter and nonzero neutrino masses in addition to the
theoretical problems of triviality and naturalness related
to the scalar Higgs sector responsible for the mechanism of
electroweak symmetry breaking.

There are many extensions of the SM that require the
existence of color octet scalar particles, such as the extra
component of the gluon field in models with extra dimen-
sions [2], supersymmetric models with an adjoint chiral
supermultiplet [3], or models with an extended color sym-
metry [4]. The existence of color octet scalars has also been
used to explain the accelerated expansion of the Universe
[5]. Extended scalar sectors of the SM with color octet
scalars that respect the principle of minimal flavor viola-
tion were also recently considered [6,7]. Color octet scalars
may also have important effects in the Higgs boson pro-
duction via gluon fusion [8].

Early studies on the existence of color octet scalars were
done in the context of one-family technicolor models [9].

Both electroweak triplets (P�;0
8 ) and singlets (P00

8 , some-

times denoted also as technieta �T8, a notation which we
will adopt in this paper) are present in the spectrum of the
pseudo-Nambu-Goldstone boson (PNGB) arising from the
global SUð8ÞL � SUð8ÞR ! SUð8ÞV spontaneous symme-
try breaking [10]. The masses of the color octet PNGB
arise mainly from QCD contributions and are expected to
be of the order of 300 GeV [11].

The cross section for pair production of the color octet
scalars is dominated by gluon fusion, which in turn is
determined by gauge invariance [12] and consequently,
except for the gluon parton densities, is mostly model
independent, being fixed by the masses of the particles.
There could also be a model dependent enhancement due
to the coupling of the scalars to color octet vector reso-
nances such as a technirho �T8 [13], which in turn couples
to quarks and gluons. However, as shown by two of the
present authors [12], a proper gauge invariant treatment of
the �T8 results in a vanishing �T8- g- g coupling [14].
Hence, only the quark initial state can cause this
enhancement.
While models of a strongly interacting sector respon-

sible for electroweak symmetry breaking fell in disfavor in
the mid-1990’s due to tight bounds from electroweak
precision measurements, they have experienced a recent
resurrection due to the correspondence with weakly inter-
acting models in extra dimensions [15]. In this case, the
�T8 could be interpreted as the Kaluza-Klein excitation of
the gluon. Tevatron searches have excluded the existence
of a �T8 in the mass range 260<M�T8

< 480 GeV decay-

ing primarily into two jets [16]. However, this limit should
be used with caution since it was obtained from a naive
vector meson dominance estimate and model uncertainties
can suppress the production cross section [17–19]. Also, in
our case the main decay mode of �T8 is into a pair of color
octet pseudoscalars.
Pair production of color octet scalars at the LHC has

recently been analyzed [6,7,20,21]. Their detection was
discussed using the b �bb �b, b �bt�t, and t�tt�t channels [20–22]
for a scalar mass of the order of a TeV. However, the low
mass values suggested by technicolor models, typically
below the t�t threshold, can be more challenging to observe.
One may think that the Tevatron limits on the �T8 from

*alfonsozerwekh@uach.cl
+claudio.dib@usm.cl
‡rosenfel@ift.unesp.br

PHYSICAL REVIEW D 77, 097703 (2008)

1550-7998=2008=77(9)=097703(4) 097703-1 � 2008 The American Physical Society

http://dx.doi.org/10.1103/PhysRevD.77.097703


dijets can also be applied to the color octet scalars.
However, since the single color octet scalar production
cross section is much smaller (in the model discussed
below it goes through the anomaly only), the limit is not
applicable [23].

We find it timely to extend the work done previously in
[12], where rarer decay modes of the color octet (pseudo)
scalar are used, namely, decays induced by the chiral
anomaly into massless gauge bosons (photons and gluons),
to investigate the possibility of detecting these new states at
the LHC.

In Sec. II we describe the details of our model, in
particular, the description of �T8 pair production and de-
cay. In Sec. III we present the details and results of our
simulations of �T8 pairs, produced in p� p collisions at
LHC energies and detected as �þ 3 jet events; we include
a study of the background and the cuts required for a
statistically significant signal. Finally, in Sec. IV we draw
our conclusions.

II. MODEL

When introducing massive vector bosons in a theory that
contains gauge fields, care must be taken not to spoil the
gauge symmetry. In our case, we need to describe QCD
with the addition of color octet vector bosons. Following
the prescription of Ref. [12], we introduce two massive

vector fields, ~G� and ~��, each transforming as an octet
under their respective SUð3ÞG and SUð3Þ� symmetry

groups, with a mixing term which is fixed in such a way
that the resulting mass matrix has a zero eigenvalue. The
corresponding eigenvector is identified with the physical
gluon field. The orthogonal combination is identified with
the physical color octet vector resonance �T8. The physical
SUð3ÞQCD gauge symmetry is the linear combination of the

generators of SUð3ÞG and SUð3Þ� corresponding to the

massless gluon. The color octet pseudoscalar �T8 is intro-
duced as a matter field that transforms purely under gauged
SUð3Þ�. In contrast, quarks are introduced in the funda-

mental representation of the gauged-SUð3ÞG symmetry.
Hence, in the mass eigenbasis there is a direct coupling
of quarks to the physical color octet resonance �T8.

At this point the free parameters of the model can be
taken as the masses of the color octet vector and scalar
particles, M�T8

and M�T8
, and a coupling constant g� that

controls the strong interaction decay of the technirho into
technietas, �T8

! �T8
�T8

.

One extra parameter, the PNGB decay constant FQ, is

necessary to describe the possible �T8 decays. The ampli-
tude for �T8 ! q �q decay is given by

M ð�a
T8 ! q �qÞ ¼ mq

FQ

�q�5

�a

2
q (1)

where �a are the Gell-Mann SUð3Þ matrices. The color
octet technieta couples to gauge bosons through the Adler-

Bell-Jackiw anomaly with an amplitude given by [24]

M ð�T8 ! B1B2Þ ¼
S�T8B1B2

4
ffiffiffi
2

p
�2FQ

���	
"
�
1 "

�
2k

	
1 k



2 (2)

where "i and ki denote the polarization vector and momen-
tum of the vector boson i. For the cases in which we are
interested, one has

S�a
T8
gbgc ¼ g2sdabcNTC (3)

and

S�a
T8g

b� ¼ gse

3
�abNTC (4)

where NTC is the number of technicolors, which we take as
NTC ¼ 2 for definiteness.
Hence, our model is completely determined by M�T8

,

M�T8
, g�, and FQ. However, we should stress that our

results are very insensitive to M�T8
, g�, and FQ, since the

pure QCD process dominates over the resonance contribu-
tion and the decay branching ratios of the �T8 are inde-
pendent of FQ. In the next section we perform a realistic

simulation of the possibility to detect a pair of color octet
technietas produced in the process pp ! �T8�T8 ! �ggg
at the LHC.

III. SIMULATION

The model was implemented in COMPHEP [25] using
LANHEP [26]. We used COMPHEP for generating events for

the double technieta production. Subsequently, they were
processed by a FORTRAN codewewrote in order to generate
the �T8 decay products. All the simulations were done for
M�T8

¼ 320 GeV. For definiteness we fixed M�T8
¼

640 GeV and FQ ¼ 80 GeV (as mentioned before, our

results are insensitive to these values). We restricted our-
selves to consider onlyM�T8

below the top quark threshold

because above that point the technieta width is strongly
dependent on the existence of a topcolor interaction.
We considered two sources of standard model back-

ground: the production of a photon and three jets and the
production of four jets with the possibility that a jet can be
misinterpreted as a photon. We assume that the probability
that such misidentification occurs is about 10�3. The back-
ground was generated using MADGRAPH/MADEVENT with
CTEQ5L as a parton distribution function. We considered
gluons and all the (anti)quarks of the first two generations
in the initial state and as sources of jets. All the possible
tree-level partonic subprocesses were taken into account in
the calculation of pp ! �jjj and pp ! jjjj.
In order to introduce some degree of realism in our

calculations, we took into account the smearing of the final
momenta. We used the following parametrization for the
detector resolution:

BRIEF REPORTS PHYSICAL REVIEW D 77, 097703 (2008)

097703-2



�ðEÞ
E

¼ 0:20
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
EðGeVÞp for photons; (5)

�ðEÞ
E

¼ 0:80
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
EðGeVÞp for jets: (6)

In fact, the resolution at ATLAS is expected to be better
than what we have used here, so in that sense our estimate
is on the conservative side.

The reconstruction of the technieta was done by study-
ing the photon-jet invariant mass (M�j) and the two-jet

invariant mass (Mjj). We addressed the combinatorics

problem by taking the M�j closer to M�T8
as a ‘‘technieta

candidate.’’
The background was reduced imposing the following set

of kinematical cuts:

PT� > 80 GeV; (7)

PTj > 80 GeV; (8)

jM�j �Mjjj< 0:15M�; (9)

jM�j �M�T8
j< 20 GeV: (10)

In Fig. 1, we show the reconstructed technieta mass
obtained by summing up the M�j and Mjj distributions

after the cuts were applied. The dashed line represents the
direct background, while the double-dotted-dashed line is
the result obtained when a jet is misidentified as a photon.
Our signal is shown by the dotted-dashed line.

Assuming an integrated luminosity of L ¼ 10 fb�1, we
expect to observe about 21 400 events with 4200 of them
corresponding to our signal. That would correspond to a

deviation from the SM with a statistical significance of
32�.
In this analysis, we have used some cuts that depend on

the technieta mass. This procedure may be uncomfortable
from the experimental point of view since the mass of the
searched particle is not known a priori and the whole
possible range must be scanned. Fortunately, in our case
the mass interval is limited because, due to QCD contri-
butions, the technieta cannot be lighter than 300 GeV and
we do not expect the channel considered in this work to be
useful for discovery if the technieta is heavier than
350 GeV.
However, it is possible to devise a search strategy which

is independent of the technieta mass. Consider the follow-
ing set of cuts:

PT� > 80 GeV; (11)

PTj > 80 GeV; (12)

jM�j �Mjjj< 10 GeV; (13)

jminðcos
�jÞj< 0:6; (14)

where 
�j is the angle formed by a photon and a jet. The

last cut comes from the fact that the technieta is a spin 0
particle and its decay is isotropic in its rest system, while
the background tends to have peaks at cos
�j ¼ �1.

Figure 2 shows the invariant mass distribution obtained
with the new set of cuts. We expect to observe, integrating
over the whole mass range, 12 600 background events and
1100 events coming from the technietas with L ¼
10 fb�1, corresponding to a 10� signal.

Invariant Mass (GeV)
260 280 300 320 340 360 380

E
ve

n
ts

/b
in

1000

2000

3000

4000

5000

FIG. 1 (color online). Reconstructed technieta invariant mass
distribution The dashed line represents the direct background,
while the double-dotted-dashed line is the result obtained when a
jet is misidentified as a photon. Our signal is shown by the
dotted-dashed line.

Invariant Mass (GeV)
300 305 310 315 320 325 330 335 340 345 350

E
ve

n
ts

/b
in

200

250

300

350

400

450

500

550

600

650

FIG. 2 (color online). Reconstructed technieta invariant mass
distribution using a set of cuts independent of the technieta mass.
The dashed line represents the direct background (we neglect the
small indirect background due to misidentification). Our signal is
shown by the dotted-dashed line.

BRIEF REPORTS PHYSICAL REVIEW D 77, 097703 (2008)

097703-3



IV. CONCLUSIONS

We studied the pair production and detection of color
octet pseudoscalar bosons at the LHC, which could be
present in models of electroweak symmetry breaking in-
duced by new strong interactions. We restricted our analy-
sis to m�T8

(the color octet pseudoscalar mass) below 2mt,

in which case the decay into t- �t is forbidden and conse-
quently the detection is more challenging. In such cases,
the color octet pseudoscalar decays mainly into two glu-
ons, induced by the anomaly, while the direct decay into b-
�b has a fraction less than 20%. We perform simulations for
the production and decay of pseudoscalar pairs, including
background. We did not look for the dominant 4-jet mode,
but for the suppressed 3-jetþ photon mode, which has a
much lower background. We used two different methods of
analysis. In one method we assume m�T8

to be known in

our cuts, and obtain a number of events 32� above the
expected background, for an integrated luminosity of
10 fb�1. In our second method, we did not include any

value of m�T8
in our cuts, but relied only on the invariant

mass reconstruction of the two decaying pseudoscalars, in
which case the number of events resulted in a statistical
significance of 10� above the expected background. These
results show that the LHC has the potential to detect or
exclude the existence of such pseudoscalar colored bosons.

ACKNOWLEDGMENTS

The work of A. R. Z. is partially supported by Grant
No. DID-UACH S-2006-28 and by Fondecyt Grant
No. 1070880; C. O.D. is partially supported by Fondecyt,
Chile, Grant No. 1070227; R. R. is partially supported by a
CNPq, Brazil, Research Grant No. 309158/2006-0. C. D.
and R. R. also acknowledge the support of Fondecyt, Chile,
Grant No. 7070098 for international cooperation. This
work is also supported by the CNPq PROSUL Program
No. 490157/2006-8 and Conicyt Grant No. PBCT-
ACT028.

[1] For an updated overview, see LEP Electroweak Working
Group, lepewwg.web.cern.ch/LEPEWWG/.

[2] See, e.g., G. Burdman, B. A. Dobrescu, and E. Ponton,
Phys. Rev. D 74, 075008 (2006).

[3] Y. Cui, Phys. Rev. D 74, 075010 (2006).
[4] P. Yu. Popov, A.V. Povarov, and A.D. Smirnov, Mod.

Phys. Lett. A 20, 3003 (2005).
[5] D. Stojkovic, G. Starkman, and R. Matsuo, Phys. Rev. D

77, 063006 (2008).
[6] A. V. Manohar and M. B. Wise, Phys. Rev. D 74, 035009

(2006).
[7] M. I. Gresham and M.B. Wise, Phys. Rev. D 76, 075003

(2007).
[8] R. Bonciani, G. Degrassi, and A. Vicini, J. High Energy

Phys. 11 (2007) 095.
[9] E. Farhi and L. Susskind, Phys. Rev. D 20, 3404 (1979); S.

Dimopoulos, Nucl. Phys. B168, 69 (1980).
[10] For reviews see, e.g., R. S. Chivukula, R. Rosenfeld, E. H.

Simmons, and J. Terning, in Electroweak Symmetry
Breaking and New Physics at the TeV Scale, edited by
T. L. Barklow, S. Dawson, H. E. Haber, and J. L. Siegrist
(World Scientific, Singapore, 1996); C. T. Hill and E.H.
Simmons, Phys. Rep. 381, 235 (2003); 390, 553(E)
(2004).

[11] M. E. Peskin, Nucl. Phys. B175, 197 (1980); J. Preskill,
Nucl. Phys. B177, 21 (1981).

[12] A. R. Zerwekh and R. Rosenfeld, Phys. Lett. B 503, 325
(2001).

[13] E. Eichten, I. Hinchliffe, K. D. Lane, and C. Quigg, Rev.
Mod. Phys. 56, 579 (1984); 58, 1065 (1986).

[14] In this sense, our work reproduces the colored sector of the
model by R. Casalbuoni, S. De Curtis, A. Deandrea, N. Di
Bartolomeo, R. Gatto, D. Dominici, and F. Feruglio, Nucl.
Phys. B409, 257 (1993).

[15] See, e.g., K. Agashe, R. Contino, and A. Pomarol, Nucl.
Phys. B719, 165 (2005).

[16] F. Abe et al. (CDF Collaboration), Phys. Rev. D 55, R5263
(1997).

[17] R. S. Chivukula, A. Grant, and E.H. Simmons, Phys. Lett.
B 521, 239 (2001).

[18] A. R. Zerwekh, Int. J. Mod. Phys. A 19, 4387 (2004).
[19] A. R. Zerwekh, Eur. Phys. J. C 49, 1077 (2007).
[20] B. A. Dobrescu, K. Kong, and R. Mahbubani,

arXiv:0709.2378.
[21] M. Gerbush, T. J. Khoo, D. J. Phalen, A. Pierce, and D.

Tucker-Smith, arXiv:0710.3133 [Phys. Rev. D (to be pub-
lished)].

[22] B. Lillie, J. Shu, and T.M. P. Tait, arXiv:0712.3057.
[23] C. Kilic, T. Okui, and R. Sundrum, arXiv:0802.2568.
[24] J. Ellis, M. K. Gaillard, D.V. Nanopoulos, and P. Sikivie,

Nucl. Phys. B182, 529 (1981).
[25] E. Boos et al. (CompHEP Collaboration), Nucl. Instrum.

Methods Phys. Res., Sect. A 534, 250 (2004).
[26] A. Semenov, Nucl. Instrum. Methods Phys. Res., Sect. A

389, 293 (1997); arXiv:hep-ph/0208011; arXiv:0805.
0555.

BRIEF REPORTS PHYSICAL REVIEW D 77, 097703 (2008)

097703-4