Stops and MET: the shape of things to come

Abstract

LHC experiments have placed strong bounds on the production of supersymmetric colored particles (squarks and gluinos), under the assumption that all flavors of squarks are nearly degenerate. However, the current experimental constraints on stop squarks are much weaker, due to the smaller production cross section and difficult backgrounds. While light stops are motivated by naturalness arguments, it has been suggested that such particles become nearly impossible to detect near the limit where their mass is degenerate with the sum of the masses of their decay products. We show that this is not the case, and that searches based on missing transverse energy (MET) have significant reach for stop masses above 175 GeV, even in the degenerate limit. We consider direct pair production of stops, decaying to invisible LSPs and tops with either hadronic or semi-leptonic final states. Modest intrinsic differences in MET are magnified by boosted kinematics and by shape analyses of MET or suitably-chosen observables related to MET. For these observables we show that the distributions of the relevant backgrounds and signals are well-described by simple analytic functions, in the kinematic regime where signal is enhanced. Shape analyses of MET-related distributions will allow the LHC experiments to place significantly improved bounds on stop squarks, even in scenarios where the stop-LSP mass difference is degenerate with the top mass. Assuming 20/fb of luminosity at 8 TeV, we conservatively estimate that experiments can exclude or discover degenerate stops with mass as large as ~ 360 GeV and 560 GeV for massless LSPs.

Figures (11)

• Figure 1: Stop pair production at $\sqrt{s}= 8$ TeV, calculated at NLO using ProspinoBeenakker:1996ed.
• Figure 2: Left: Differential distribution of events for 20 fb$^{-1}$ with respect to $\slashed {E}_{T}$ of QCD (blue) and $t\bar{t}$ (green), and the total background (black) passing the hadronic trigger. The analytic fits to Eq. \ref{['eq:METfit']} using the parameters in Table \ref{['tab:bestfit']} are shown in red for QCD (dashed), $t\bar{t}$ (dotted) and their sum (solid). Right: Differential distribution of events corresponding to 20 fb$^{-1}$ with respect to $\slashed {E}_{T}$ for signal $\tilde{t}\bar{\tilde{t}}\to t\bar{t}\chi\chi$ passing the hadronic trigger for a range of stop and LSP masses $(m_{\tilde{t}},m_\chi)$.
• Figure 3: Differential distribution of $t\bar{t}$ events with respect to $M_T^W$ (black). The analytic fit (Eq. \ref{['eq:METfit']} using the parameters of Table \ref{['tab:bestfit_lep']}) is shown in red. Also shown are the differential distributions of stop signal events with respect to $M_T^W$ for a range of stop and LSP masses. The semi-leptonic event selection is described in Section \ref{['sec:semilep']}.
• Figure 4: Left: Signal trigger efficiency as a function of stop and LSP masses for hadronic event selection. Right: Signal cross section times trigger efficiencies as a function of stop and LSP masses. Like all such plots in this paper, the contours are extrapolated from a grid of Monte Carlo results with $5-25$ GeV spacing in $m_{\tilde{t}}$ and $m_\chi$. The degeneracy line ($m_{\tilde{t}}=m_t+m\chi$) is shown in black.
• Figure 5: Expected sensitivity, in standard deviations, for the hadronic $\slashed {E}_{T}$ shape analysis as a function of the stop and LSP masses. The test statistic is computed from 200 pseudo-experiments of 20 fb$^{-1}$. In the left-hand plot the uncertainty on the background $\slashed {E}_{T}$ shape are as shown in Table \ref{['tab:bestfit']} and in the right-hand plot these errors have been inflated by a factor of 3.