\documentclass[hyperref=pdftex, presentation]{beamer}
% Replacing 'presentation' with 'handout' in the above line
% will produce 4 slides per page.
% The hyperref option makes it possible to include hyperlinks.

\mode<presentation> {
\usetheme{Boadilla}
\setbeamercovered{transparent}
}

\usepackage[english]{babel}
\usepackage[latin1]{inputenc}
\usepackage{times}
\usepackage[T1]{fontenc}

\mode<handout>{
\usepackage{pgfpages}
\pgfpagesuselayout{4 on 1}[letterpaper,landscape,border shrink=5mm]
\setbeamercolor{background canvas}{bg=black!10} }

\setbeamertemplate{footline}
{
\leavevmode%
\hbox{%
\begin{beamercolorbox}[wd=.333333\paperwidth,ht=2.25ex,dp=1ex,center]{author
	in head/foot}%
	\usebeamerfont{author in head/ foot}\insertshortauthor%&\approx& (\insertshorti
\end{beamercolorbox}%
\begin{beamercolorbox}[wd=.333333\paperwidth,ht=2.25ex,dp=1ex,center]{title
	in head/foot}%
	\usebeamerfont{title in head/foot}\insertshorttitle
\end{beamercolorbox}%
\begin{beamercolorbox}[wd=.333333\paperwidth,ht=2.25ex,dp=1ex,right]{date in
	head/foot}%
	\usebeamerfont{date in head/foot}\insertshortdate\hspace*{2em}
	\insertframenumber / \inserttotalframenumber\hspace*{2ex}
\end{beamercolorbox}}%
\vskip0pt%
}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%


\title[Heavy Photon Search]{Searching for Heavy Photons in the HPS Experiment}

\author{Sho Uemura}
\institute{SLAC}
\date[September 23, 2014]

\begin{document}

\begin{frame}
	\titlepage
\end{frame}

\begin{frame}{Reminder of review format}
	\begin{itemize}
		\item Fourth-year students are required to give a 45-minute oral presentation to their Ph.D. 
			reading committees. %Generally no other people besides the student, advisor and 
			%reading committee members are present at the oral presentation. 
		%\item All students must complete this requirement in the fourth year of study. Experience 
			%has proven this is an extremely reliable tool to help students stay on track to degree 
			%completion.
		\item The purpose of the requirement is to increase contact between students and faculty 
			members, to help students organize their thoughts, to give students practice in giving 
			oral presentations, and most importantly to obtain feedback on the development of the 
			thesis, approximate date of thesis completion and future plans.
		%\item These are informal meetings, and no grades are given. Students schedule the 
			%presentations themselves. By end of winter quarter of the fourth-year students should 
			%have a set date for the oral presentation.
		\item The sessions should consist of a half-hour presentation by the student, 15 minutes of 
			discussion between the student, research advisors and readers, and then a closed door 
			discussion by the committee.
	\end{itemize}
\end{frame}

\begin{frame}{Outline}
	\begin{itemize}
		\item Background on HPS
		\item What I've been doing for HPS
		\item Plans for the analysis
	\end{itemize}
\end{frame}

\begin{frame}{Motivation for heavy photon}
	\begin{columns}
	\column{0.6\textwidth}
		\begin{itemize}
			\item A new massive $U(1)$ boson with no (direct) coupling to SM
			\begin{itemize}
				\item Kinetic mixing with the photon $\to$ weak coupling to electric charge [Holdom 1986]
			\end{itemize}
			\item One possible ``portal'' between SM and a dark sector
			\item Two relevant parameters: mass $m_{A'}$, relative coupling strength $\alpha'/\alpha$
			\item Useful in explaining PAMELA/Fermi/AMS positron excess, muon g-2 anomaly
		\end{itemize}
	\column{0.4\textwidth}
	\begin{center}
		\includegraphics[width=\textwidth]{sm_dark}

		\includegraphics[width=\textwidth]{kinetic_mixing}
	\end{center}
	\end{columns}
\end{frame}

\begin{frame}{Producing heavy photons}
	\begin{columns}
	\column{0.65\textwidth}
	\begin{itemize}
		\item Similar to bremsstrahlung: $e^-$ (1.1, 2.2 and 6.6 GeV) on high-Z fixed target
	\end{itemize}
	\column{0.35\textwidth}
	\includegraphics[width=0.8\textwidth]{prod_diagram}
	\end{columns}
	\begin{center}
		\includegraphics[width=0.65\textwidth]{production}
	\end{center}
	\begin{itemize}
		\item $A'$ carries most of incident $e^-$ energy (unlike $\gamma$ bremsstrahlung)
		\item Pairs from $A'$ decay are produced along beam with some decay length and small opening angle
		% \begin{itemize}
		% 	\item Track pairs, find vertex and invariant mass
		% \end{itemize}
	\end{itemize}
\end{frame}

\begin{frame}{Reach}
	\begin{columns}
	\column{0.5\textwidth}
	\begin{itemize}
		\item Bump hunt: look for a peak in pair invariant mass
		\item Vertexing (my thesis topic): look for pairs originating downstream of the target
			\begin{itemize}
				\item Requires a tracker close to the target for $\sim$mm vertex resolution
			\end{itemize}
		\item HPS probes a large unexplored region of the parameter space
			\begin{itemize}
				\item Bump-hunt region is under pressure (Babar result not shown in plot); vertexing is our strength
			\end{itemize}
	\end{itemize}
	\column{0.5\textwidth}
		\includegraphics[width=\textwidth]{reach.pdf}
	\end{columns}
\end{frame}

\begin{frame}{The HPS detector}
	\begin{center}
		\includegraphics[width=0.8\textwidth]{svt.pdf}
	\end{center}
	\begin{columns}
	\column{0.5\textwidth}
		\begin{itemize}
			\item Thin (0.125\% or 0.25\% $X_0$) tungsten target
			\item Silicon microstrip tracker in dipole magnet for measurement
			\item PbWO$_4$ calorimeter for trigger
		\end{itemize}
	\column{0.5\textwidth}
		\includegraphics[width=\textwidth]{hps_setup.pdf}
	\end{columns}
\end{frame}

\begin{frame}{HPS schedule}
	\begin{itemize}
		\item Schedule has been controlled by CEBAF schedule:
			\begin{itemize}
				\item Test run in May 2012 before shutdown for 12 GeV upgrade: photon beam only, good for testing detector performance but no physics
				\item Commissioning November-December 2014, physics running in spring 2015
				\item Further running planned, but thesis will be written on 2015 data
			\end{itemize}
		\item We are pushing to publish a result within a year after we get data
	\end{itemize}
\end{frame}

\begin{frame}{My role in HPS}
	\begin{itemize}
		\item SVT mechanics
			\begin{itemize}
				\item Testing, disassembly and re-QA of test run SVT
				\item Assembly and survey of new SVT
			\end{itemize}
		\item Software
			\begin{itemize}
				\item Readout simulation for ECal and SVT
				\item Trigger studies and simulation
				\item Hit time reconstruction in SVT
				\item Bulk production of MC for mock data challenge
			\end{itemize}
		\item Taking shifts
	\end{itemize}
\end{frame}

%\begin{frame}{Work with test run SVT}
	%\begin{columns}
	%\column{0.6\textwidth}
		%\begin{itemize}
			%\item Developed calibrations for test run SVT
			%\item Helped set up power/cooling/dry air for DAQ tests of the SVT last year
			%\item Disassembled SVT; repeating QA for all half-modules
		%\end{itemize}
	%\column{0.4\textwidth}
		%\includegraphics[width=\textwidth]{svt_tim.jpg}
%
		%\includegraphics[width=\textwidth]{disassembled.jpg}
	%\end{columns}
%\end{frame}

\begin{frame}{Assembly and survey of new SVT}
	\begin{itemize}
		\item Pair half-modules onto support structure at 50/100 mrad stereo angle
		\item Survey assembled modules using optical/touch-probe CMM: get sensor positions and curvature
		\item Modules (in sets of 3) mount onto U-channels; survey U-channels to relate sensors to each other and beamline survey
	\end{itemize}
	\begin{centering}
		\includegraphics[width=0.28\textwidth]{support_proto.jpg}
		\includegraphics[width=0.22\textwidth]{pairing_l456.jpg}
		\includegraphics[width=0.28\textwidth]{l123_front.PNG}
		\includegraphics[width=0.22\textwidth]{uchannel-l123.jpg}
	\end{centering}
\end{frame}

\begin{frame}{SVT timing}
	\begin{columns}
	\column{0.6\textwidth}
	\begin{itemize}
		\item SVT hit occupancies are very high; reject hits outside a time window defined by the ECal event time
		\item HPS is designed for a maximum occupancy of 1\% (L1 next to the beam gap) in 8 ns window; to run at planned current with good tracking efficiency, hit time recon must be good within 8 ns
		\item Use APV25 readout chip to take 6 samples at 24 ns spacing; fit the pulse coming out of the preamp to get hit time resolution $\sim$2 ns
			\begin{itemize}
				\item Minimize effect of pulse pileup on the time fit
			\end{itemize}
		\item Use fitted hit times in track finding
	\end{itemize}
	\column{0.4\textwidth}
		\includegraphics[width=\textwidth]{nakahama_pileup.pdf}
	\end{columns}
\end{frame}

\begin{frame}{MC production}
	\begin{itemize}
		\item Mock data challenge: 1 week equivalent of simulated data with secret A' signals mixed in
		\item Run on JLab compute farm
		\item Set up and exercise all our MC, data handling and recon tools
	\end{itemize}
	\begin{center}
		\includegraphics[width=0.6\textwidth]{flowchart.png}
	\end{center}
\end{frame}

\begin{frame}{Analysis!}
	\begin{itemize}
		\item Reconstruction overview
		\item Event sample: $e^+e^-$ pairs with 2-track vertex and invariant mass
		\item Basic event selection
		\item Bump hunt, ref. APEX
		\item Blinding
	\end{itemize}
\end{frame}

\begin{frame}{Reconstruction, in brief (SVT only)}
	\begin{itemize}
		\item Hit time reconstruction of individual strip hits
		\item Clustering to form strip clusters
		\item Track finding to form tracks with 5 or 6 stereo hits
		\item Track refit using GBL
		\item Vertexing using Billoir algorithm
		\item Result: $e^+e^-$ pairs with 2-track vertex and invariant mass
	\end{itemize}
\end{frame}

\begin{frame}{Vertex search}
	\begin{itemize}
		\item Signal: small number of events with detached vertices and a single mass
			\begin{itemize}
				\item Parameters: mass, production cross-section, lifetime
				\item In minimal model (100\% branching ratio to $e^+e^-$), cross-section and lifetime are both determined by $\alpha'/\alpha$
			\end{itemize}
		\item Background: QED tridents with finite vertex resolution (with non-Gaussian tails) and smooth mass distribution
			\begin{itemize}
				\item Mix of Bethe-Heitler and radiative tridents
				\item Beam-gas interactions possible but expected to be rare
			\end{itemize}
	\end{itemize}
\end{frame}

\begin{frame}{Resolutions and tails}
	\begin{itemize}
		\item We care about mass resolution and vertex resolution
		\item Mass resolution depends on angular and momentum resolution
			\begin{itemize}
				\item Position resolution at target tells us about angular resolution
				\item e-W elastic scattering gives momentum resolution near beam energy
				\item M\o ller scattering gives angular and momentum resolution across momentum range
			\end{itemize}
		\item Vertex resolution depends on angular resolution, but non-Gaussian tails of the distribution are a big deal
			\begin{itemize}
				\item Mishits: wrong L1 hit is added to the track (MS makes the ```bad'' hit a better fit than the ``good'' one), pulls track
				\item Kinks: large-angle scatter in early layer
			\end{itemize}
	\end{itemize}
\end{frame}

\begin{frame}{Event cuts}
	\begin{itemize}
		\item Basic event cuts:
			\begin{itemize}
				\item Track quality (track fit and time $\chi^2$, ECal track match)
				\item Pair quality: (vertex fit $\chi^2$, $E_++E_-\approx E_{beam}$)
				\item Kinematic cuts to reduce Bethe-Heitler tridents
			\end{itemize}
		\item Important to reduce vertex tails (may or may not be useful for other analyses):
			\begin{itemize}
				\item Track quality: layer 1 isolation cut (mishits), kink cut (large-angle scatters)
			\end{itemize}
	\end{itemize}
\end{frame}

\begin{frame}{Blinding}
	\begin{itemize}
		\item Hidden signal box: cut out events with vertex Z above some limit
			\begin{itemize}
				\item Box should be looser than expected signal region, but tight enough that background and resolutions can be characterized before unblinding
				\item Must be able to kill most backgrounds without looking in the box
			\end{itemize}
		\item Subsample: develop analysis on a fraction of the data
			\begin{itemize}
				\item Bump hunt will probably use this approach
			\end{itemize}
	\end{itemize}
\end{frame}

\begin{frame}{Analysis}
	\begin{center}
		\includegraphics[width=0.5\textwidth]{toy}
	\end{center}
	\begin{itemize}
		\item Moving mass window
		\item Profile likelihood
		\item Discovery/setting limits
	\end{itemize}
\end{frame}


\begin{frame}{Fitting the signal}
	\begin{itemize}
		\item A complete fit of the background is hard and unnecessary; use a scanning window
		\item Scan through the signal parameters (mass, lifetime); only fit data within a mass window and above some cutoff Z
		\item Signal shape is completely determined by mass, lifetime, and the known mass resolution; background distribution can be fit with an ansatz
		\item Use a mass window larger than mass resolution (to capture all signal events), but small enough that background (i.e. trident rate and vertex resolution as function of mass) can be fit well
		\item Use unbinned likelihood fit to use full event-by-event vertex resolution information
	\end{itemize}
\end{frame}

\begin{frame}{Profile likelihood}
	$$\lambda(\mu)=\frac{L(\mu,\hat{\hat{\theta}})}{L(\hat{\mu},\hat{\theta})}$$
	\begin{itemize}
		\item Do a likelihood fit to the data in the scanning window, optimizing the background parameters $\theta$ as a function of the signal strength (production rate) parameter $\mu$
		\item Profile likelihood ratio: ratio of the likelihood at each $\mu$ to the likelihood at the best fit $\hat{\mu}$
		\item $\lambda(0)$ gives the local p-value for discovery (correct for look-elsewhere effect to get global p-value)
		\item Setting limits: we can calculate the p-value for the $\mu$ corresponding to 100\% branching ratio
	\end{itemize}
\end{frame}

\end{document}