JetEtmissApproved2013HighMuEtmiss

E_T^miss linearity as a function of the true E_T^miss in events with different number of interactions per bunch crossing (<μ>), in Z'->ttbar (m_Z’=2 TeV) events for 25 ns bunch spacing. E_T^miss is calculated taking into account only the calorimeter deposits. The E_T^miss linearity is expected to be zero if the reconstructed E_T^miss has the correct scale. The positive bias at low true E_T^miss is due to the finite resolution of the E_T^miss measurement, and is highly dependent on the topology of the process studied. For true E_T^miss above 150 GeV (250 GeV) and <μ>=40 (<μ>=140), the linearity is within ~1%. The ETmiss value at which the linearity reaches a constant value increases with <μ>, due to the poorer resolution in high <μ> events. The evolution of the resolution as a function of <μ> was also studied and is consistent with the results shown here.

Dependence of the pile-up contribution to ∑E_T on μ in Z'->ttbar (m_Z’=2 TeV) events with different numbers of interactions per bunch crossing (<μ>) and 25 ns bunch spacing. E_T^miss is calculated taking into account only the calorimeter deposits. The pileup noise (σ_noise^pile-up) used in the topoclustering is optimized for each value of pileup. The pileup contribution to ∑E_T is defined as ∑E_T^PU = ∑E_T – ∑E_T^Truth where ∑E_T^Truth is due to all the interacting particles from the hard scatter, excluding muons. The points are fitted by ∑E_T^PU=c√μ. The resolution worsens as <μ> increases, but the effect is mitigated by use of the optimized σ_noise^pile-up value.

E_x,y^miss resolution as a function of ∑E_T in Z'-> ttbar (m_Z’=2 TeV) events and 25 ns bunch spacing, in events with different numbers of interactions per bunch crossing (<μ>). E_T^miss is calculated taking into account only the calorimeter deposits. The pileup noise (σ_noise^pile-up) used in the topoclustering is optimized for each value of <μ>. The resolution curves are fitted with s+k √ ∑E_T  . For events with higher levels of pile-up the distribution gets shifted upwards but does not change its slope as a function of ∑E_T. This effectively means that in the resolution curves the <μ> dependence is absorbed by the s-term, whereas the k-term remains approximately constant.

E_x,y^miss resolution as a function of the number of truth vertices (N_vtx) in Z'-> ttbar (m_Z’=2 TeV) events and 25 ns bunch spacing, in events with different numbers of interactions per bunch crossing (<μ>). E_T^miss is calculated taking into account only the calorimeter deposits. The pileup noise (σ_noise^pile-up) used in the topoclustering is optimized for each value of <μ>. The resolution is fitted with a straight line to show that the E_x,y^miss resolution increases linearly with the number of truth vertices, and independently of <μ>, when the optimal pileup noise values are used.

E_x,y^miss resolution as a function of the number of truth vertices (N_vtx) in events with an average of 40 interactions per bunch crossing (<μ>) for different values of pileup noise (σ_noise^pile-up) in dijet events at 25 ns bunch spacing. E_T^miss is calculated taking into account only the calorimeter deposits. In general, tightening σ_noise^pile-up results in poorer E_x,y^miss resolution.