172
5 Direct Searches for New Physics
optimised for W
→ Bt and W
→ T b, which both result in the t Hb final state for the
decays B → Hb and T → Ht [935]. The analysis targets high m W and m VLQ , such
that the H and t are produced with large boost and can be reconstructed using large-R
jets with p T > 300 and 400 GeV, respectively. Even for the smallest mass differences
considered in this search, m W − m VLQ = 200 GeV, the b quark from the W
decay
receives large enough momentum to be reconstructed with a b-tagged small-R jet with
p T > 200 GeV. The situation is different for the W
→ Bt decay, where small mass
differences lead to a t quark produced nearly at rest, and therefore not reconstructible
with a single large-R jet. In this regime, the analysis loses sensitivity because the two
decays W
→ T b and W
→ Bt are assumed to happen with the same frequency. The
H and t tagging algorithms select jets with a soft drop mass in the range 105–135
and 105–210 GeV, respectively. In addition, H jets have to pass a selection based
on the discriminator from the double-b tagger and t jets have to have τ 32 < 0.8 and
a subjet b tag. The signal region is defined by events with a H -, t- and b-tagged jet.
The distribution in the reconstructed W
is used to search for a signal, obtained from
the four-vector sum of the three identified jets in the event. Sideband and validation
regions are used to estimate the multijet background. These are obtained by inverting
the b-tagging, τ 32 or jet mass requirements. A transfer function of the H tagger is
derived as a function of p T and η, describing the ratio of probabilities to pass the H -
and inverted H -tagging selections. This transfer function is obtained from events with
an inverted t tag. It is used to derive an event weight to construct a template for the
reconstructed W
distribution in a control region with inverted H tag, which is used to
predict the background in the signal region. This approach is validated in simulation
and in a dedicated validation region in data. In both tests good agreement between
the predicted and expected background distributions is observed. This method has
the distinct advantage that it provides the background estimation for any distribution,
such that the background model can be checked thoroughly. In the signal region, the
multijet background constitutes about 70% of the total background, the remaining
part originating from tt production. No significant deviation from the background
prediction is observed in the data and cross section upper limits on W
production in
the t Hb decay mode are reported as a function of m W , for several m VLQ hypotheses.
This is the only search to date targeting the heavy-light decay of a W
boson.
No dedicated search for the decay modes B → W t, T → W b, B → Zb or T →
Zt exists, which would result in tt W , bbZ and t Zb final states, respectively. While
SM measurements in these final states have been carried out [936–943], these do not
extend into the high energy tails of the production cross sections, which are relevant
for BSM models. The best sensitivity to BSM effects in these final states is obtained
by dedicated searches for resonant VLQ production, which will hopefully be carried
out in the future.
5 Direct Searches for New Physics
optimised for W
→ Bt and W
→ T b, which both result in the t Hb final state for the
decays B → Hb and T → Ht [935]. The analysis targets high m W and m VLQ , such
that the H and t are produced with large boost and can be reconstructed using large-R
jets with p T > 300 and 400 GeV, respectively. Even for the smallest mass differences
considered in this search, m W − m VLQ = 200 GeV, the b quark from the W
decay
receives large enough momentum to be reconstructed with a b-tagged small-R jet with
p T > 200 GeV. The situation is different for the W
→ Bt decay, where small mass
differences lead to a t quark produced nearly at rest, and therefore not reconstructible
with a single large-R jet. In this regime, the analysis loses sensitivity because the two
decays W
→ T b and W
→ Bt are assumed to happen with the same frequency. The
H and t tagging algorithms select jets with a soft drop mass in the range 105–135
and 105–210 GeV, respectively. In addition, H jets have to pass a selection based
on the discriminator from the double-b tagger and t jets have to have τ 32 < 0.8 and
a subjet b tag. The signal region is defined by events with a H -, t- and b-tagged jet.
The distribution in the reconstructed W
is used to search for a signal, obtained from
the four-vector sum of the three identified jets in the event. Sideband and validation
regions are used to estimate the multijet background. These are obtained by inverting
the b-tagging, τ 32 or jet mass requirements. A transfer function of the H tagger is
derived as a function of p T and η, describing the ratio of probabilities to pass the H -
and inverted H -tagging selections. This transfer function is obtained from events with
an inverted t tag. It is used to derive an event weight to construct a template for the
reconstructed W
distribution in a control region with inverted H tag, which is used to
predict the background in the signal region. This approach is validated in simulation
and in a dedicated validation region in data. In both tests good agreement between
the predicted and expected background distributions is observed. This method has
the distinct advantage that it provides the background estimation for any distribution,
such that the background model can be checked thoroughly. In the signal region, the
multijet background constitutes about 70% of the total background, the remaining
part originating from tt production. No significant deviation from the background
prediction is observed in the data and cross section upper limits on W
production in
the t Hb decay mode are reported as a function of m W , for several m VLQ hypotheses.
This is the only search to date targeting the heavy-light decay of a W
boson.
No dedicated search for the decay modes B → W t, T → W b, B → Zb or T →
Zt exists, which would result in tt W , bbZ and t Zb final states, respectively. While
SM measurements in these final states have been carried out [936–943], these do not
extend into the high energy tails of the production cross sections, which are relevant
for BSM models. The best sensitivity to BSM effects in these final states is obtained
by dedicated searches for resonant VLQ production, which will hopefully be carried
out in the future.