A Higgs of high mass 
1 TeV is very broad
and leads to strong scattering of gauge boson pairs, eventually
violating unitarity for the s-wave scattering amplitude.
In order to probe the regularization
mechanism of this cross-section, it is important to understand
the production of
gauge boson pairs at high mass. At this energy scale,
the longitudinally polarized 
play
the role of the Goldstone bosons of the Higgs symmetry-breaking
mechanism. Thus, WW fusion will be dominated by the longitudinal
components, and the rate of production of 
pairs will
provide information on the Higgs boson if it exists, or more generally
on the nature of the dynamical process responsible for the symmetry
breaking.
The study of same-sign W pairs produced in the WW-fusion process
pp 
has been suggested by many
authors [49,50], since
this channel does not suffer from gg or
fusion backgrounds. The expected rates are very low
and vary rapidly with the pp centre-of-mass energy.
The W-pair signal can only be observed above the background if
both W bosons decay to an electron or a muon and if two tag jets
are required in the forward regions (Section 
).
The dominant backgrounds [51] are:
in amplitude),
such as qq 
W
W
qq,
which produce mostly 
or
pairs, with a cross-section independent of the Higgs mass.
in amplitude.
production.
fusion,
with Z-boson decays into electrons or muons.
production.
or
W
W
production, with charge misidentification
of one of the leptons.
The signal from a Standard Model Higgs with 
= 1 TeV, and the
backgrounds from 
, WZ and ZZ were generated using
PYTHIA 5.7. The other background processes
were simulated at the parton level using the event generator
of [50]. The detector acceptance
and resolution for leptons and jets were included in the simulation.
In a first step, only events containing two same-sign isolated
leptons, with 
> 25 GeV and 
< 2.5, were retained
(see first column of Table ).
At this stage, the background dominates, with the largest
contribution coming from WZ/ZZ production (the 
background
is significantly reduced by the lepton isolation cuts).
If a third lepton was present within the acceptance,
the invariant dilepton masses, computed using
all of the selected leptons of same flavour
and opposite charge, were required to be outside 
15 GeV,
thus rejecting the dominant WZ/ZZ background.
Additional cuts, which
increase the signal-to-background ratio, required that the dilepton mass
be above 100 GeV, that the opening angle in the transverse plane
between the two leptons be larger than 90
,
and that their transverse momenta differ by less than 80 GeV.
The second column
of Table shows the expected rates for the signal
and various backgrounds after these additional lepton cuts.
Table: The expected numbers of events for the 
signal
and for the
various background processes with the expected significances
as a function of the cuts,
for an integrated luminosity of 
The next steps of the selection procedure were aimed at further
reducing the remaining backgrounds. The 
background
was greatly reduced by rejecting events with a
jet of 
> 40 GeV in the central region and 
< 2
(third column of Table ). Finally
two tag jets were required, one in each of the forward regions,
with 15 GeV < 
< 130 GeV (in a cone of size 
= 0.35).
The upper limit set on the tag-jet 
significantly
reduced the 
residual backgrounds.
The last column of
Table shows the expected rates after all
cuts. The rates for W
W
pairs would be about a factor of three
lower for the 
and the 
processes, but are expected to be the
same for the potentially more dangerous reducible backgrounds, thus
providing a useful check of the background estimates.
Charge misidentification is a negligible background,
given the charge identification capabilities of the
ATLAS
detector for 
-values between 25 and 500 GeV (see
Sections , ).
In particular, it is important to note
that, after cuts, the opposite-sign lepton pairs from W
W
and 
production contain no events with leptons
of 
> 200 GeV.
Figure shows the expected same-sign dilepton
distribution after all cuts. The total event rate is quite low
and the signal-to-background ratio does not vary much
as a function of the lepton 
. Since the 
signal
cannot be completely separated from the 
background,
accurate theoretical predictions will be necessary
to optimize the sensitivity to a possible signal in this channel.
Table shows an estimate of the event
rates predicted by alternate models for the regularization of the
WW scattering cross-section, assuming that the effects of cuts
and acceptance are the same for 
pairs in all scenarios.
The signal rates vary by a factor of 
2 around that
predicted for a SM Higgs with 
= 1 TeV. Clearly, several years of
running at high luminosity will be needed to establish
an excess of events with respect to the expected background processes
in this channel.
Figure: The 
spectrum expected for same-sign dileptons with
two tag jets, assuming an integrated luminosity of 
.
The signal corresponds to a Higgs with 
= 1 TeV,
and the various
backgrounds are discussed in the text.
Table: Expected numbers of events after cuts in the 
search
for an integrated luminosity of 
and for different
models [50]