1: %\include{symbols}

2: 

3: \newcommand{\zee}{$Z \rightarrow ee$}

4: \newcommand{\met}{\mbox{${\rm \not\! E}_{\rm T}$}}

5: \newcommand{\pT}{p$_{\rm T}$}

6: \newcommand{\Et}{{\rm E}_{\rm T}}

7: \newcommand{\ET}{E$_{\rm T}$}

8: \newcommand{\Pt}{{\rm p}_{\rm T}}

9: \newcommand{\wenu}{$W \rightarrow e\nu$}

10: \newcommand{\Z}{{ Z^0}}

11: \newcommand{\gravitino}{\tilde{G}}

12: 

13: \documentclass[prl,twocolumn,superscriptaddress,showpacs,floatfix]{revtex4}

14: %\documentclass[prl,preprint,superscriptaddress,showpacs,floatfix]{revtex4}

15: \usepackage[letterpaper]{geometry}

16: \usepackage{amsfonts}

17: \usepackage{graphicx}

18: \usepackage{psfrag}

19: \usepackage{times}

20: %\usepackage{doublespace}

21: \begin{document}\setlength{\unitlength}{1mm}

22: % You should use BibTeX and apsrev.bst for references

23: \bibliographystyle{apsrev}

24: 

25: % Use the \preprint command to place your local institutional report

26: % number on the title page in preprint mode.

27: % Multiple \preprint commands are allowed.

28: \preprint{CDF/PUB/EXOTIC/CDFR/5765}

29: \preprint{EFI 01-53}

30: \preprint{Version 1.3}

31: 

32: %Title of paper

33: \title{Limits on Extra Dimensions and New Particle Production in the

34: Exclusive Photon and Missing Energy

35: Signature in $p\bar p$ Collisions at $\sqrt{s} =$ 1.8 TeV}

36: 

37: % repeat the \author .. \affiliation  etc. as needed

38: % \email, \thanks, \homepage, \altaffiliation all apply to the current

39: % author. Explanatory text should go in the []'s, actual e-mail

40: % address or url should go in the {}'s for \email and \homepage.

41: % Please use the appropriate macro for the type of information

42: 

43: % \affiliation command applies to all authors since the last

44: % \affiliation command. The \affiliation command should follow the

45: % other information

46: 

47: %\author{A large group of people}

48: %\email[]{Your e-mail address}

49: %\homepage[]{Your web page}

50: %\thanks{}

51: %\altaffiliation{}

52: %\noaffiliation

53: %\author{P. Onyisi}

54: %\author{H. Frisch}

55: %\affiliation{University of Chicago}

56: 

57: \input{author_list_plusme.tex}

58: 

59: %Collaboration name if desired (requires use of superscriptaddress

60: %option in \documentclass). \noaffiliation is required (may also be

61: %used with the \author command).

62: %\collaboration{The CDF Collaboration}

63: %\noaffiliation

64: 

65: \date{\today}

66: 

67: \begin{abstract}

68: % insert abstract here

69: The exclusive $\gamma$\met{} signal has a small standard model 

70: cross-section and is thus

71: a good channel in which to look for signs of new physics.  This signature is 

72: predicted by models

73: with a superlight gravitino or with large extra spatial dimensions.  We search for such

74: signals at the CDF detector at the Tevatron, using 87 pb$^{-1}$ of data at 

75: $\sqrt{s} = 1.8$ TeV, and extract 95\% C.L.\ limits on these processes.

76: A limit of 221 GeV is set on the scale $|F|^{1/2}$ in supersymmetry 

77: models.  For 4, 6, and 8 extra dimensions, limits on the fundamental

78: mass scale $M_D$ of 549, 581, and 602 GeV, respectively, 

79: are found.  We also specify a `pseudo-model-independent' method of comparing 

80: the results to theoretical predictions.

81: \end{abstract}

82: % insert suggested PACS numbers in braces on next line

83: \pacs{13.85.Rm, 14.80.-j}

84: 

85: %\maketitle must follow title, authors, abstract and \pacs

86: \maketitle

87: 

88: % body of paper here - Use proper section commands

89: % References should be done using the \cite, \ref, and \label commands

90: 

91: Many extensions to the standard model predict the existence of

92: minimally-interacting particles, such as the gravitino in

93: supersymmetric models and Kaluza-Klein (KK) modes of the graviton in models

94: with large compact spatial dimensions \cite{ledsref}.  Such particles cannot be

95: directly observed in a detector, but their production can be inferred

96: from a transverse momentum imbalance (or ``missing transverse energy,''

97: $\met{}$

98: \cite{etdef}) among the visible particles in a high-energy collision.

99: Photons can be

100: emitted in such hard-scattering processes due to the presence of charged

101: quarks in the $p\bar p$ initial state; many models also predict the 

102: production of photons from the decay of

103: final-state particles \cite{Ambrosiano}.

104: A search for the $\gamma\met$ signature thus explores a wide range of

105: models and complements searches in the single jet$+\met$ channel 

106: \cite{castro}.  Here we present the results of a search in the

107: exclusive $\gamma$\met{} signature, i.e.\ where only a photon and invisible

108: particles are produced.

109: 

110: The data used for this analysis were collected with the Collider Detector 

111: at Fermilab (CDF)

112: during Run 1b of the Tevatron, with an integrated luminosity of 

113: 87 $\pm$ 4 pb$^{-1}$ of $p\bar p$ collisions at $\sqrt{s} = 1.8$ TeV.

114: The CDF detector has been described in detail elsewhere \cite{detector}; 

115: subsystems most important to this search are summarized here.

116: A system of time projection chambers around the beampipe allows the

117: determination of the event vertex position.  Surrounding these chambers is

118: the central tracking chamber (CTC), a

119: cylindrical drift chamber inside a 1.4 T superconducting solenoid, which

120: is fully efficient for track reconstruction in the pseudorapidity region 

121: $|\eta| < 1.0$ \cite{cdf_coo}.

122: The central electromagnetic calorimeter (CEM) covers the region

123: $|\eta| < 1.1$.  Strip chambers (the CES system) are embedded in the CEM at

124: the depth of shower maximum to

125: allow the measurement of the 2-dimensional transverse profile

126: of electromagnetic showers.

127: The central hadronic calorimeter

128: covers the range $|\eta| < 1.3$ and is instrumented with 

129: time-to-digital converters which associate times to shower signals. 

130: The calorimeter modules are arranged in ``towers,'' with hadronic modules

131: behind the electromagnetic modules, in a projective geometry pointing

132: at the center of the detector.  High energy electromagnetic

133: showers frequently leak from the

134: electromagnetic modules into the hadronic modules behind them;

135: when sufficient leakage occurs timing can be associated with the 

136: electromagnetic 

137: shower.

138: Outside the calorimeters, drift chambers provide muon identification in

139: the region $|\eta| < 1.0$.

140: 

141: To select events with high-\pT{} photons during data-taking, we use

142: the CDF 3-level trigger system \cite{trigger}.  Level 1 requires a

143: central calorimeter tower with $\Et^{EM} > 8$ GeV \cite{cdf_coo}.  The Level 2

144: system forms clusters of towers and then requires the event to pass an

145: OR of several triggers, including: a) a very loose trigger requiring only

146: an electromagnetic cluster \cite{top_evidence} with $\Et^{EM} >$ 50 GeV and

147: b) a trigger requiring \met{} $>$ 35 GeV.  Level 3 requires that

148: the photon candidate satisfy \ET{} $>$ 50 GeV and have a CES cluster

149: within the fiducial region \cite{jeff_prd}.

150: 

151: The offline photon candidate identification (``Photon ID'') criteria 

152: \cite{toback_prd,ray_prd,jeff_prd} are

153: a) an electromagnetic cluster 

154: in the CEM with $|\eta^\gamma| <$ 1 \cite{eta_note}, a ratio 

155: $\mathrm{E}^{HAD}/\mathrm{E}^{EM}$ less than $0.055 + 0.00045 

156: \times \mathrm{E}^{SUM}$, a centroid 

157: within the 

158: fiducial region of the CES, and shower evolution measured by the CES 

159: consistent with expectation;

160: b) no second energetic object in the same CES wire chamber as the cluster;

161: c) at most one CTC track, and none

162: with \pT{} $>$ 1 GeV \cite{units}, pointing at the cluster;

163: d) within a radius of 0.4 in $\eta$-$\phi$

164: space around the cluster centroid, 

165: \ET{} (summed over towers excluding those in the photon cluster) $<$ 2 GeV 

166: and a sum of track

167: \pT{} $<$ 5 GeV;

168: e) $\Et^\gamma >$ 55 GeV \cite{correctionsnote}; 

169: and 

170: f) an event vertex within 60 cm of the center of the detector along the

171: beamline.

172: 

173: The selection on missing transverse energy is

174: \met{} $>$ 45 GeV.  This threshold is lower than the 

175: $\Et^\gamma$ threshold to keep this requirement fully efficient for signal 

176: processes, taking into consideration the \met{} resolution and 

177: the intrinsic parton \pT{} in the $p$ and $\bar p$ initial states.

178: 

179: Backgrounds to the $\gamma$\met{} signal include:

180: a) $q\bar q \rightarrow Z\gamma \rightarrow \nu\bar\nu\gamma$;

181: b) cosmic ray muons that undergo bremsstrahlung in the CEM but for which 

182: no track is found;

183: c) $W \rightarrow e\nu$ with the electron misidentified as a photon;

184: d) $W\gamma$ production where the charged lepton in a leptonic $W$

185: decay is lost;

186: e) prompt $\gamma\gamma$ production where a photon is lost; and 

187: f) dijet and photon + jet production.  

188: %The expected numbers of events from these background sources are given in 

189: %Table~\ref{backgrounds}.

190: 

191: To reject cosmic ray muons, we require 

192: a timing signal in the hadronic calorimeter which is 

193: in-time with the collision within a window 55 ns wide

194: for at least one tower 

195: in the cluster, and no evidence of a muon

196: in the central muon systems 

197: within $30^\circ$ in $\phi$ of the photon.  The efficiency of requiring that

198: timing information be present

199: rises with $\Et^\gamma$ from 78\% at 55 GeV to over 98\% above 100

200: GeV.  The efficiency of these two cuts is measured with a sample of 

201: isolated electrons.

202: 

203: To remove the $W\gamma$ background as well as events in which 

204: mismeasurement of jet energy produces fake

205: \met{},

206: we require 

207: no jets \cite{top_evidence} with \ET{} $>$ 15 GeV,

208: no jets with \ET{} $>$ 8 GeV within 0.5 radians in $\phi$ of the photon,

209: and

210: no tracks in the event with $\Pt >$ 5 GeV.  

211: 

212: \begin{figure}

213: \includegraphics[width=\linewidth]{et_spectrum_staged_prl_grey}

214: \caption{Photon \ET{} spectrum for events remaining after 

215: each stage of cuts.  

216: %The

217: %``Photon ID'' cuts include the event vertex requirement.

218: }

219: \label{fig:staged_spectrum}

220: \end{figure}

221: 

222: \begin{figure}

223: \includegraphics[width=\linewidth]{pet_vs_met}

224: \caption{Photon \ET{} versus \met{} for the 11 events passing the selection

225: criteria. The tight correlation of $\Et^\gamma$ and \met{} reflects the

226: detector resolution for unclustered energy.}

227: \label{fig:pet_vs_met_spectrum}

228: \end{figure}

229: 

230: Trigger and background considerations drive the choice of the $\Et^\gamma$

231: threshold. The Level 3 trigger becomes fully 

232: efficient ($>$ 99\%) at 55 GeV. In addition, 

233: below 45 GeV the background from \wenu{} with

234: a misidentified electron 

235: is very large; as the $\Et^\gamma$ threshold is increased beyond

236: the kinematic limit for electrons from $W$ decay at rest, the

237: $W$ must recoil against another object, and the event is then rejected by the

238: jet and track vetoes.

239: 

240: For an exclusive photon and invisible particle process,

241: the overall efficiency for all cuts is found to vary from 0.45 at 

242: \ET{} = 55 GeV to 0.56 for \ET{} $>$ 100 GeV, with a $\pm$10\% uncertainty.

243: The cumulative effect of each cut is shown in

244: Fig.~\ref{fig:staged_spectrum}.  The number of events surviving the

245: photon ID, \met{}, cosmic ray rejection, and jet and track cuts are

246: 15,046, 1,475, 94, and 11, respectively.  The

247: $\Et^\gamma$ and \met{} in the 11 events in the final sample

248: are shown in Fig.~\ref{fig:pet_vs_met_spectrum}. 

249: 

250: \begin{table}

251: \caption{Background sources.  The uncertainty in the QCD background is

252: unknown, and this background is not considered when setting limits.  The

253: numbers do not total due to rounding.}

254: \label{backgrounds}

255: \begin{ruledtabular}

256: \begin{tabular}{lc}

257: Cosmic rays & 6.3 $\pm$ 2.0 \\

258: $Z\gamma \rightarrow \nu\bar\nu\gamma$ & 3.2 $\pm$ 1.0 \\

259: $W \rightarrow e\nu$ & 0.9 $\pm$ 0.1 \\

260: Prompt $\gamma\gamma$ & 0.4 $\pm$ 0.1\\

261: $W\gamma$ & 0.3 $\pm$ 0.1\\

262: \hline

263: Total non-QCD background & 11.0 $\pm$ 2.2 \\

264: QCD background & $\sim$ 1\\

265: \hline

266: Total observed & 11 \\

267: \end{tabular}

268: \end{ruledtabular}

269: \end{table}

270: 

271: To estimate the number of cosmic ray events in the signal sample, we use

272: the events which have a timing signal

273: outside the in-time window but which pass all other cuts.  We then

274: extrapolate into the signal region, assuming a flat distribution in time.

275: 

276: The Monte Carlo simulations of both signal

277: processes and the

278: $Z\gamma$, $W\gamma$ and prompt $\gamma\gamma$ backgrounds

279: use the \textsc{Pythia} event generator \cite{torbjorn} with

280: the CTEQ5L parton distribution functions (PDFs) \cite{cteq}, 

281: followed by a parametrized simulation of the CDF detector.

282: The simulations are then corrected for deficiencies in the detector model

283: and the $\pm$10\% efficiency uncertainty applied.

284: We turn off initial state radiation (ISR) to obtain

285: leading-order (LO) cross-sections and efficiencies.

286: For the background processes, the resulting cross-sections are corrected 

287: by the 

288: ratio of the LO cross-section to the next-to-leading-order 

289: ``zero-jet'' cross-section, obtained from theoretical calculations and Monte 

290: Carlo estimates.

291: This allows the correct estimation of the acceptance $\times$ efficiency 

292: $\times$ cross-section for the exclusive process. 

293: We obtain correction factors of 0.95 $\pm$ 0.3

294: for $Z\gamma$ 

295: \cite{ZgammaKfactor}, 0.9 $\pm$ 0.2 for $W\gamma$

296: \cite{WgammaKfactor}, and 1.0 $\pm$ 0.3 for prompt $\gamma\gamma$

297: \cite{ggKfactor};

298: the systematic uncertainties considered

299: are $Q^2$ choice and acceptance variations due to modeling of

300: ISR in the Monte Carlo simulations.  These uncertainties are added in

301: quadrature with the efficiency uncertainty.

302: 

303: The background from \wenu{} arises either from hard bremsstrahlung by the 

304: electron before it enters the tracking chamber or inefficiency in the track

305: reconstruction.  As a radiated photon tends to be collinear with 

306: the electron,

307: the \ET{} of the identified electromagnetic object will, in

308: either case, be close to the initial energy of the electron.

309: Let $\cal{P}$ be the ratio between the number of electrons

310: faking photons and the number of electrons passing standard electron 

311: identification cuts \cite{jeff_prd} in the region $|\eta^e| < 1$; we

312: estimate $\cal{P}$ 

313: by assuming that ``$e\gamma$'' events with invariant masses

314: within 10 GeV of the $\Z$ mass are actually $\Z \rightarrow 

315: ee$ events.  We obtain $\cal{P}$ $=$ (0.8 $\pm$ 0.1)\%.  The background estimate

316: is $\cal{P}$ times the number of \wenu{} events that have $|\eta^e| < 1$,

317: $\Et^e >$ 55 GeV,

318: \met{} $>$ 45 GeV, and pass the jet and 

319: track vetoes (discounting the electron track).

320: 

321: We have investigated QCD backgrounds which involve the

322: mismeasurement of jet energy leading to apparent \met{} or

323: misidentification of a jet as a photon.  

324: The most likely

325: contributors to fakes are events with one high-energy object and many

326: low-energy jets.

327: With the \met{}, jet, and track requirements, these events are rare.

328: To estimate these backgrounds one must use data; however all control

329: samples have small statistics and estimates range from 0.1 to 2 

330: events.

331: We take the conservative approach of not including this background

332: source in the total background used in the limit calculations.  This can

333: only make the limits less stringent \cite{omit_QCD}.

334: 

335: We study two hypothetical signal processes in detail.  One is predicted by a 

336: supersymmetric model and the other by a model with large compact 

337: extra dimensions.

338: 

339: The first process ($q\bar q \rightarrow \gravitino \gravitino \gamma$) is 

340: described in \cite{brignole}.  It presumes 

341: that the gravitino $\gravitino$ is the lightest

342: supersymmetric particle, with the other superpartners too 

343: heavy to produce

344: on-mass-shell at the Tevatron.  Since the gravitino coupling is very small, 

345: being able to produce

346: other supersymmetric particles increases the cross-section;

347: we therefore set an absolute lower limit on the gravitino mass $m_{3/2}$

348: or, equivalently, the supersymmetry breaking scale $|F|^{1/2}$ (the two are 

349: related by

350: $|F| = \sqrt{3} m_{3/2} M_P$, with $M_P$ being the Planck mass).

351: The cross-section for this process scales as $1/|F|^4$; the kinematic

352: distributions are independent of $|F|$.

353: 

354: The second process ($q\bar q \rightarrow \gamma G_{KK}$) is described 

355: in \cite{giudice}:

356: $n$ extra spatial dimensions are assumed to be compactified with radius $R$.

357: The fundamental mass scale $M_D$ and $R$ are related

358: to Newton's constant and the number of extra dimensions by $G_N^{-1} = 

359: 8\pi R^n M_D^{2+n}$ \cite{convnote}.  The standard

360: model fields propagate only on a 3+1 dimensional subspace, while gravitons

361: propagate in the whole space.  The graviton modes which propagate in the extra

362: dimensions appear to four-dimensional observers as massive states of the 

363: graviton.

364: A large value of $R$ results in a large phase space for graviton

365: production, canceling the weakness of the coupling to standard 

366: model fields.  For a given $n$, the cross-section

367: scales as $1/M_D^{n+2}$ \cite{efftheory}; for fixed $n$, the kinematic

368: distributions are independent of $M_D$.

369: 

370: 

371: The two signal processes are simulated with modified versions of 

372: \textsc{Pythia}.  The $q\bar q \rightarrow \gravitino \gravitino\gamma$ 

373: process is simulated with $|F|^{1/2} = 100$ GeV, and the $q\bar q 

374: \rightarrow \gamma G_{KK}$ process is simulated with 

375: $M_D$ = 1 TeV for $n=$ 4, 6, and 8 extra dimensions.

376: 

377: We consider three sources of theoretical systematic uncertainty in the 

378: cross-section

379: and acceptance predictions: uncertainty in the choice of $Q^2$ scale,

380: the choice of parton distribution function, and the modeling of ISR.  

381: We obtain uncertainty

382: estimates by varying $Q^2$ by a factor of 4 both up and down,

383: by using the GRV98 LO PDFs \cite{GRV} instead of the CTEQ5L PDFs,

384: and by turning the modeling of ISR on and off. 

385: The uncertainty due to ISR  

386: includes order-$\alpha_s$ effects and acceptance

387: changes due to the jet and track vetoes.

388: For $q\bar q \rightarrow \gravitino \gravitino\gamma$,

389: the dominant uncertainty is the $Q^2$ choice ($^{+26}_{-18}$\%), followed

390: by ISR ($\pm$14\%) and PDF choice ($\pm$10\%).

391: For $q\bar q \rightarrow \gamma G_{KK}$, the dominant uncertainty comes from

392: ISR ($\pm$34\%), followed by $Q^2$ choice ($^{+18}

393: _{-16}$\%) and PDF choice ($\pm$8\%).  The overall uncertainty in the

394: $q\bar q \rightarrow \gravitino \gravitino\gamma$ acceptance $\times$

395: efficiency $\times$ cross-section, which includes the $\pm$10\% efficiency

396: uncertainty,

397: is $^{+33}_{-27}$\%.  For $q\bar q \rightarrow \gamma G_{KK}$,

398: the corresponding figure is $^{+41}_{-40}$\%.

399: 

400: %To set limits, we use a method derived from that of the Particle Data

401: %Group \cite{oldpdg}.  A 95\% C.L.\ limit on a parameter $\alpha$

402: %is defined as that value of $\alpha$ for which, in 95\% of an ensemble of

403: %experiments, the number of events from background is less than

404: %the observed 11 and the total number of events from background and

405: %signal processes is greater than 11.

406: 

407: The method we use to set limits is outlined in \cite{poilim}.

408: %From this procedure, 

409: We find the following limits at 95\% C.L.: for the

410: supersymmetric model, $|F|^{1/2} \ge 221$ GeV (equivalently,

411: $m_{3/2} \ge 1.17\times 10^{-5}$ eV); for large extra dimensions,

412: $M_D \ge$ 549, 581, and 602 GeV for $n=$ 4, 6, and 8 extra dimensions

413: (equivalently, $R \le$ 24 nm, 55 fm, and 2.6 fm, respectively)

414: \cite{limitsnote}.  The previous best limit published for $|F|^{1/2}$ is 

415: 217 GeV,

416: from a CDF jet+\met{} search \cite{castro}; the previous best published 

417: $M_D$ limits set from direct production of

418: gravitons are 0.68 TeV, 0.51 TeV, and 411 GeV for $n=$ 4, 6, and 8 extra dimensions,

419: the first two set by DELPHI \cite{delphilimit} and the third by 

420: L3 \cite{l3limit}.

421: 

422: \begin{figure}

423: \includegraphics[width=\linewidth]{acc_eff_combined}

424: \caption{Plots of (a) acceptance vs.\ $\eta^\gamma$ and (b) efficiency 

425: vs.\ $\Et^\gamma$ for the analysis selection.

426: These plots are valid for any exclusive photon

427: and invisible particle process.  The error bars are statistical only.

428: The falloff in acceptance at $|\eta|\simeq 0$ and  $|\eta|\simeq 1$

429: is due to the folding of the fiducial region of the calorimeter with

430: the  longitudinal spread ($\sigma \simeq 30$ cm) of the $p\bar p$

431: collisions.}

432: \label{fig:modind}

433: \end{figure}

434: 

435: The results of this analysis can be presented in a `pseudo-model-independent'

436: manner.  In both the above models,

437: the uncertainties in the predicted numbers of signal events have been 

438: dominated by theoretical factors.

439: It can be useful to derive a limit 

440: which considers only the uncertainties in the detector simulation

441: of the processes and so can easily be compared across models \cite{ray_prd}

442: (keeping in mind that such a limit is not a substitute for the rigorous

443: extraction of a limit noting theoretical uncertainties).

444: To obtain this limit, we compute a 95\% C.L.\ upper limit on the number of

445: events from new physics that would be detected, using only the $\pm$10\% uncertainty

446: in efficiency as the uncertainty in the acceptance $\times$ efficiency $\times$

447: cross-section for the new process.

448: This limit is 9.8 events, which for this integrated luminosity corresponds

449: to a cross-section of 112 fb.

450: 

451: The plots in Fig.~\ref{fig:modind} allow a comparison of  

452: models to the acceptance $\times$ efficiency $\times$ 

453: cross-section limit.  These curves are obtained by

454: studying the acceptance and efficiency curves for simulated events and

455: correcting for deficiencies in the detector simulation.  These plots are

456: valid for both the $\gravitino \gravitino \gamma$ and $\gamma G_{KK}$ processes

457: studied above, and for any process producing an exclusive photon and 

458: invisible particle signature.

459: One can estimate the acceptance $\times$ efficiency $\times$ cross-section for

460: such a process by convolving the theoretical photon $\eta$ and \ET{} spectra

461: with the acceptance and efficiency curves.

462: 

463: In conclusion, we have performed a search for new physics in the exclusive

464: $\gamma$\met{} channel.  We have found no departure from the

465: expected Standard Model cross-section

466: and have set limits on two specific models of new physics, one a 

467: supersymmetric model in which

468: the photon is produced in association with two gravitinos, the second a

469: model with large extra dimensions in which the photon is produced in

470: association with a KK mode of the graviton.

471: We have also presented the limit in a `pseudo-model-independent' manner.

472: 

473: %\begin{acknowledgements}

474: We thank the Fermilab staff and the technical staffs of the

475: participating institutions for their vital contributions.  We would

476: also like to thank J. Lykken, K. Matchev, and D. Rainwater for their

477: help.  This work was supported by the U.S. Department of Energy and

478: National Science Foundation; the Italian Istituto Nazionale di Fisica

479: Nucleare; the Ministry of Education, Culture, Sports, Science, and

480: Technology of Japan; the Natural Sciences and Engineering Research

481: Council of Canada; the National Science Council of the Republic of

482: China; the Swiss National Science Foundation; the A. P. Sloan

483: Foundation; the Bundesministerium f\"ur Bildung und Forschung,

484: Germany; the Korea Research Foundation and the Korea Science and

485: Engineering Foundation (KoSEF); and the Comision Interministerial de

486: Ciencia y Tecnologia, Spain.

487: %\end{acknowledgements}

488: 

489: 

490: 

491: % Create the reference section using BibTeX:

492: %\bibliography{your bib file}

493: \bibliography{prl}

494: \end{document}

495: %

496: % ****** End of file template.aps ******

497: 

498: