| 1 |
|
\section{Limits on New Physics} |
| 2 |
|
\label{sec:limit} |
| 3 |
|
|
| 4 |
< |
%As discussed in Section~\ref{sec:results}, we see one event |
| 5 |
< |
%in the signal region. |
| 6 |
< |
%The background prediction from the SM MC is |
| 7 |
< |
%1.3 events. |
| 8 |
< |
%The data driven background predictions from the ABCD method |
| 9 |
< |
%and the $\pt(\ell\ell)$ method are $1.3 \pm 0.8({\rm stat}) \pm 0.3({\rm syst})$ and |
| 10 |
< |
%$2.1 \pm 2.1({\rm stat}) \pm 0.6({\rm syst})$ events, respectively. |
| 4 |
> |
We first set an upper limit on the signal yield extracted by the fit to the dilepton mass |
| 5 |
> |
distribution, assuming the LM1 shape. The 95\% confidence level (CL) upper limit (UL) |
| 6 |
> |
is extracted using a profile likelihood technique, giving an UL of X events. The expected |
| 7 |
> |
LM1 yield is X $\pm$ X events; hence LM1 is excluded by these results. |
| 8 |
> |
|
| 9 |
> |
We proceed to set upper limits on the non-SM contributions to the high \MET\ and high \Ht\ signal regions. |
| 10 |
> |
For both regions, we find reasonable agreement between the observed yields and the predictions from MC and from the 2 |
| 11 |
> |
data-driven methods. We choose here to extract the upper limits using the MC prediction for the |
| 12 |
> |
background estimate. The 95\% CL upper limit is extracted using a Bayesian technique~\cite{ref:cl95cms}, |
| 13 |
> |
using a log-normal model of nuissance parameter integration. The results are summarized in |
| 14 |
> |
Table~\ref{tab:results}. Based on these results, we exclude LM1. |
| 15 |
|
|
| 16 |
< |
%These three predictions are in good agreement with each other |
| 13 |
< |
%and with the observation of one event in the signal region. |
| 14 |
< |
%We calculate a Bayesian 95\% CL upper limit~\cite{ref:cl95cms} |
| 15 |
< |
%on the number of non SM events in the signal region to be 4.1, |
| 16 |
< |
%using a background prediction of $N_\textrm{BG}=1.4 \pm 0.8$ |
| 17 |
< |
%events and a Gaussian model of nuissance parameter integration. |
| 18 |
< |
%The upper limit is not very sensitive to $N_\textrm{BG}$ and its uncertainty. |
| 19 |
< |
|
| 20 |
< |
|
| 21 |
< |
The three background predictions for the high-\pt\ lepton trigger search |
| 22 |
< |
discussed in Section~\ref{sec:results} are in good agreement with each other and |
| 23 |
< |
with the observation of one event in the signal region. |
| 24 |
< |
A Bayesian 95\%~confidence level (CL) upper limit~\cite{ref:cl95cms} on the number of |
| 25 |
< |
non-SM events in the signal region is determined to be 4.0, |
| 26 |
< |
using a background prediction of $N_\textrm{BG} = 1.4 \pm 0.8$ |
| 27 |
< |
events and a log-normal model of nuisance parameter integration. |
| 28 |
< |
The upper limit is not very sensitive to $N_\textrm{BG}$ and its uncertainty. |
| 29 |
< |
This generic upper limit is not corrected for the possibility |
| 30 |
< |
of signal contamination in the control regions. This is justified because |
| 31 |
< |
the two independent background estimation methods based on data agree |
| 32 |
< |
and are also consistent with the SM MC prediction. |
| 33 |
< |
Moreover, no evidence for non-SM contributions in |
| 34 |
< |
the control regions is observed (Table~\ref{tab:datayield} and Figure~\ref{fig:victory}). |
| 35 |
< |
This bound rules out the benchmark SUSY scenario LM0, for which the |
| 36 |
< |
number of expected signal events is $8.6 \pm 1.6$, while the LM1 scenario predicts |
| 37 |
< |
$3.6 \pm 0.5$ events. |
| 38 |
< |
The uncertainties in the LM0 and LM1 event yields arise from energy scale, luminosity, |
| 39 |
< |
and lepton efficiency, as discussed in Section~\ref{sec:systematics}. |
| 40 |
< |
|
| 41 |
< |
For the same-flavour search using hadronic activity triggers discussed in Section~\ref{sec:HT}, |
| 42 |
< |
no same-flavour events are observed and the corresponding Bayesian 95\% CL upper limit on the |
| 43 |
< |
non-SM yield is 3.0 events. |
| 44 |
< |
This bound rules out the benchmark SUSY scenarios LM0 and LM1, for which the |
| 45 |
< |
numbers of expected signal events are $7.3 \pm 1.6$ and $3.6 \pm 0.7$, respectively. |
| 46 |
< |
|
| 47 |
< |
%As an example of the sensitivity of this search, we |
| 48 |
< |
%remind the reader that the number of expected signal events in the benchmark |
| 49 |
< |
%SUSY scenarios LM0 and LM1 are $8.6 \pm 1.6$ and $3.6 \pm 0.5$, respectively, |
| 50 |
< |
%where the uncertainties are from energy scale, luminosity, and lepton efficiency. |
| 51 |
< |
|
| 52 |
< |
|
| 53 |
< |
%Following text on CMSSM scan taken from RA1 paper |
| 54 |
< |
We also quote the result more generally in the context of the CMSSM. |
| 55 |
< |
The Bayesian 95\% CL limit in the $(m_0,m_{1/2})$ plane, for $\tan\beta=3$, |
| 56 |
< |
$A_0 = 0$ and $\mu > 0$ is shown in Figure~\ref{fig:msugra}. |
| 57 |
< |
The high-\pt\ lepton and hadronic trigger searches have similar sensitivity to the CMSSM; |
| 58 |
< |
here we choose to show results based on the high-\pt\ lepton trigger search. The |
| 59 |
< |
SUSY particle spectrum is calculated using SoftSUSY~\cite{Allanach:2002uq}, and the |
| 60 |
< |
signal events are generated at leading order (LO) with \PYTHIA6.4.22. |
| 61 |
< |
NLO cross sections, obtained with the |
| 62 |
< |
program Prospino~\cite{Beenakker:1997kx}, are used to calculate the observed |
| 63 |
< |
exclusion contour. |
| 64 |
< |
At each point in the $(m_0,m_{1/2})$ plane, the acceptance uncertainty is calculated by |
| 65 |
< |
summing in quadrature the uncertainties from jet and \MET\ energy scale using the |
| 66 |
< |
procedure discussed in Section~\ref{sec:systematics}, the uncertainty in the |
| 67 |
< |
NLO cross section due to the choice of factorization and renormalization scale, |
| 68 |
< |
and the uncertainty from the parton distribution function (PDF) for CTEQ6.6~\cite{Nadolsky:2008fk}, |
| 69 |
< |
estimated from the envelope provided by the CTEQ6.6 error sets. |
| 70 |
< |
The luminosity uncertainty and dilepton |
| 71 |
< |
selection efficiency uncertainty are also included, giving a total relative acceptance uncertainty which varies |
| 72 |
< |
in the range 0.2--0.3. |
| 73 |
< |
%We consider a point excluded if the NLO yield exceeds the 95\% CL Bayesian upper limit |
| 74 |
< |
%calculated with this acceptance uncertainty, using a log-normal model for the nuisance |
| 75 |
< |
%parameter integration. |
| 76 |
< |
A point is considered to be excluded if the NLO yield exceeds the 95\% CL |
| 77 |
< |
Bayesian upper limit calculated with this acceptance uncertainty, using |
| 78 |
< |
a log-normal model for the nuisance parameter integration. |
| 79 |
< |
The limit curves do not include the effect of signal |
| 80 |
< |
contamination in the control regions. We have verified that this |
| 81 |
< |
has a negligible impact on the excluded regions in Figure~\ref{fig:msugra}. |
| 82 |
< |
|
| 83 |
< |
\begin{figure}[tbh] |
| 16 |
> |
\begin{table}[hbt] |
| 17 |
|
\begin{center} |
| 18 |
< |
\includegraphics[width=1\linewidth]{plots_final/RA6_ExclusionLimit_tanb3_LO.pdf} |
| 19 |
< |
\caption{\label{fig:msugra}\protect |
| 87 |
< |
%Text taken from RA1 paper |
| 88 |
< |
The observed 95\% CL exclusion contour at NLO (solid red line) and LO (dashed blue line) |
| 89 |
< |
in the CMSSM $(m_0,m_{1/2})$ plane for $\tan\beta=3$, $A_0 = 0$ and $\mu > 0$. |
| 90 |
< |
The area below the curve is excluded by this measurement. Exclusion limits obtained from |
| 91 |
< |
previous experiments are presented as filled areas in the plot. Thin grey lines correspond to |
| 92 |
< |
constant squark and gluino masses. |
| 93 |
< |
%The plot also shows the two benchmark points LM0 and LM1 for comparison. |
| 94 |
< |
%Exclusion curve in the CMSSM parameter space, |
| 95 |
< |
%assuming $\tan\beta=3$, sign of $\mu = +$ and $A_{0}=0\GeVcc$. |
| 18 |
> |
\caption{\label{tab:results} |
| 19 |
> |
Summary of the observed and predicted yields in the 2 signal regions. MC errors are statistical only. |
| 20 |
|
} |
| 21 |
+ |
\begin{tabular}{|l|c|c|c} |
| 22 |
+ |
\hline |
| 23 |
+ |
& high \met\ signal region & high \Ht\ signal region \\ |
| 24 |
+ |
\hline |
| 25 |
+ |
Observed yield & 4 & 3 \\ |
| 26 |
+ |
\hline |
| 27 |
+ |
MC prediction & 2.6 $\pm$ 0.8 & 2.5 $\pm$ 0.8 \\ |
| 28 |
+ |
ABCD' prediction & 1.2 $\pm$ 0.4 (stat) $\pm$ 0.5 (syst) & 0.0 $\pm$ 0.6 (stat) $\pm$ 0.3 (syst) \\ |
| 29 |
+ |
\ptll\ prediction & 5.4 $\pm$ 3.8 (stat) $\pm$ 2.2 (syst) & 1.7 $\pm$ 1.7 (stat) $\pm$ 0.6 (syst) \\ |
| 30 |
+ |
non-SM yield UL & 5.2 & 4.1 \\ |
| 31 |
+ |
LM1 yield & 17 $\pm$ 3.1 & 14 $\pm$ 3.1 \\ |
| 32 |
+ |
LM3 yield & 4.3 $\pm$ 0.9 & 4.3 $\pm$ 1.0 \\ |
| 33 |
+ |
\hline |
| 34 |
+ |
OF subtraction ($\Delta$) & 1.3 $\pm$ 1.9 (stat) $\pm$ 0.1 (syst) & 0.1 $\pm$ 1.5 (stat) $\pm$ 0.0 (syst) \\ |
| 35 |
+ |
\hline |
| 36 |
+ |
\end{tabular} |
| 37 |
|
\end{center} |
| 38 |
< |
\end{figure} |
| 99 |
< |
|
| 100 |
< |
The excluded regions for the CDF |
| 101 |
< |
search for jets + missing energy final states~\cite{PhysRevLett.102.121801} were |
| 102 |
< |
obtained for $\tan\beta=5$, while those from D0~\cite{Abazov2008449} were obtained for |
| 103 |
< |
$\tan\beta=3$, each with approximately 2~fb$^{-1}$ of data and for $\mu < 0$. |
| 104 |
< |
The LEP-excluded |
| 105 |
< |
regions are based on searches for sleptons and charginos~\cite{LEPSUSY}. |
| 106 |
< |
%A comparison of the exclusion limit for tan b = 3 to that for tan b = 10 |
| 107 |
< |
%for fixed values of A0 = 0 and sign(m) > 0 indicates that |
| 108 |
< |
%the exclusion reach is only weakly dependent on the value of tan b; |
| 109 |
< |
%the limit shifts by less than 20\GeV in m0 and by less than 10\GeV in |
| 110 |
< |
%m1/2. |
| 111 |
< |
The D0 exclusion limit, valid for $\tan\beta=3$ and obtained from |
| 112 |
< |
a search for associated production of charginos $\chi_{1}^{\pm}$ and |
| 113 |
< |
neutralinos $\chi_2^0$ in trilepton final states~\cite{Abazov200934}, is also |
| 114 |
< |
included in Figure~\ref{fig:msugra}. In contrast to the other limits presented in |
| 115 |
< |
Figure~\ref{fig:msugra}, the results of our search and of the trilepton search are strongly dependent on |
| 116 |
< |
the choice of $\tan\beta$ and they reach the highest sensitivity in the |
| 117 |
< |
CMSSM for $\tan\beta$ values below 10. |
| 118 |
< |
|
| 119 |
< |
%We also performed a scan of the CMSSM parameter space. We set $\tan\beta=3$, |
| 120 |
< |
%sign of $\mu = +$, $A_{0}=0\GeVcc$, and scan the $m_{0}$ and $m_{1/2}$ parameters |
| 121 |
< |
%in steps of 10\GeVcc. At each point we calculate the acceptance uncertainty by |
| 122 |
< |
%summing in quadrature the uncertainties from jet and \MET\ energy scale using the |
| 123 |
< |
%procedure discussed in Section~\ref{sec:systematics} ($2-21$\%), the uncertainty in the |
| 124 |
< |
%NLO cross-section assessed by varying the factorization and renormalization scale |
| 125 |
< |
%by a factor of 2 ($12-15$\%), and the uncertainty from the parton density function (PDF) |
| 126 |
< |
%assessed by () ($X-X$\%). We also include the luminosity uncertainty (11\%) and dilepton |
| 127 |
< |
%selection efficiency uncertainty (5\%), giving a total acceptance uncertainty of $X-X$\%. |
| 128 |
< |
%For each scan point, we calculate a 95\% CL upper limit using a Bayesian technique |
| 129 |
< |
%using the observed signal yield (1 event), the acceptance uncertainty as calculated above, and |
| 130 |
< |
%the predicted background yield and uncertainty $N_\textrm{BG}=1.4 \pm 0.8$. The point is considered |
| 131 |
< |
%excluded if the NLO event yield exceeds this upper limit. The results are shown in Figure~\ref{fig:msugra}. |
| 132 |
< |
|
| 38 |
> |
\end{table} |
