1: %% neutron-antineutron oscillation paper
2: %% Last Revised: 05/13/2002
3:
4: %Section III.A \ref{sec:nuandrockevts}
5: %Section III.C \ref{sec:rockbkgrdindata}
6: %Section V \ref{sec:deteff}
7: %Section VIII.B \ref{sec:errors_bkgrds}
8:
9: %\documentstyle[preprint,epsfig,aps]{revtex}
10: %From Progress Report:
11: %\documentstyle[12pt,epsf,epsfig,graphicx]{report}
12: %
13: \documentstyle[fleqn,epsfig]{article}
14:
15: %\documentclass[10pt]{article}
16: %\usepackage{float,fleqn,epsfig}
17: %\documentstyle[twoside,fleqn,espcrc2]{article}
18: %\usepackage{epsfig}
19:
20: \textwidth 6.5in
21: \textheight 8.25in
22: \oddsidemargin 0.0in
23: \evensidemargin 0.0in
24:
25: \newcommand{\etal}{{\em et al.}\bs}
26: \newcommand{\h}{\hfill{}\\}
27: \newcommand{\hl}{\hline}
28: \newcommand{\hs}[1]{\hspace{#1}}
29: \newcommand{\bs}{\hspace{5.0pt}}
30: \newcommand{\ls}{\hspace*{50pt}}
31: \newcommand{\bl}{\hspace{3.0pt}}
32: \newcommand{\s}{\mb{s}}
33: \newcommand{\tnn}{\mbox{$\tau_{n\overline{n}}\ $}}
34: \newcommand{\Tnn}{\mbox{$T_{n\overline{n}}$}}
35: \newcommand{\TA}{\mbox{$T_{A}$}\bs}
36: \newcommand{\TAn}{\mbox{$T_{A}$}}
37: \newcommand{\Tnty}{\mbox{$T_{A, n_{90}}$}\bl}
38: \newcommand{\TR}{\mbox{$T_{R}$}\bs}
39: \newcommand{\TRn}{\mbox{$T_{R}$}}
40: \newcommand{\enn}{\mbox{$\epsilon_{n\overline{n}}$}}
41:
42: \newcommand{\lequ}[1]{\lefteqn{#1}}
43: \newcommand{\mulc}[3]{\multicolumn{#1}{#2}{#3}}
44: \newcommand{\rbx}[2]{\raisebox{#1}{#2}}
45: \newcommand{\mb}[1]{\mbox{#1}}
46: \newcommand{\cp}[1]{\caption{#1}}
47: \newcommand{\vs}{\vspace{-0.15in}}
48:
49: \newcommand{\frejus}{Fr\'{e}jus\hs{0.2cm}}
50: \newcommand{\frejusn}{Fr\'{e}jus}
51: \newcommand{\pnet}{\mbox{$P_{net}$}\bs}
52: \newcommand{\pnetn}{\mbox{$P_{net}$}}
53: \newcommand{\evis}{\mbox{$E_{vis}$}\bs}
54: \newcommand{\evisn}{\mbox{$E_{vis}$}}
55:
56: \newcommand{\n}{$\overline{\mb{\rm n}}$\hs{0.2cm}} % anti-neutron
57: \newcommand{\nn}{${\rm n}\overline{\rm n}$\hs{0.2cm}} % neutron,anti-neutron
58: \newcommand{\nN}{\mbox{$\rm {\overline{n}N}$}\hs{0.2cm}} % anti-neutron + nucleon
59: \newcommand{\nNn}{\mbox{$\rm {\overline{n}N}$}}
60: \newcommand{\NN}{\mbox{$\rm {\overline{N}N}$}\hs{0.2cm}} % anti-nucleon + nucleon
61: \newcommand{\pmm}{$\pm$} % plus and minus in math mode
62:
63: \newcommand{\neut}{\mbox{$\nu$}} % nu
64: \newcommand{\numu}{\mbox{$\nu_{\mu}$}} % nu_mu
65: \newcommand{\nue}{\mbox{$\nu_{e}$}} % nu_e
66: \newcommand{\nutau}{\mbox{$\nu_{\tau}$}} % nu_tau
67: \newcommand{\anu}{\mbox{$\overline{\nu}$}} % anu
68: \newcommand{\anue}{\mbox{$\overline{\nu}_{e}$}} % anu_e
69: \newcommand{\anumu}{\mbox{$\overline{\nu}_{\mu}$}} % anu_mu
70:
71: \newcommand{\pn}{\mbox{$\pi$}} % pion
72: \newcommand{\pip}{\mbox{$\pi^{+}$}} % pi plus
73: \newcommand{\pim}{\mbox{$\pi^{-}$}} % pi minus
74: \newcommand{\piz}{\mbox{$\pi^{0}$}} % pi zero
75: \newcommand{\spip}{\mbox{$\scriptscriptstyle\pi^{+}$}} % pi plus small
76: \newcommand{\spim}{\mbox{$\scriptscriptstyle\pi^{-}$}} % pi minus small
77: \newcommand{\m}{$\mu$} % muon
78: \newcommand{\mup}{\mbox{$\mu^{+}$}} % mu plus
79: \newcommand{\mum}{\mbox{$\mu^{-}$}} % mu minus
80: \newcommand{\del}{\delta} % delta
81: \newcommand{\g}{\mbox{$\gamma$}} % gamma
82:
83:
84: \title{
85: \begin{flushright}
86: \vskip -50pt
87: {\normalsize
88: PDK-788 \\
89: TUHEP-02-01 \\
90: \vskip -10pt
91: May 2002}
92: \end{flushright}
93: \vskip 50pt
94: {\bf Search for Neutron-Antineutron Oscillations \\
95: Using Multiprong Events in Soudan 2}
96: \vskip 30pt
97: }
98:
99: \author{
100: J.~Chung$^5$,
101: W.W.M.~Allison$^3$, G.J.~Alner$^4$, D.S.~Ayres$^1$, W.L.~Barrett$^6$,\\
102: P.M.~Border$^2$, J.H.~Cobb$^3$, H.~Courant$^2$, D.M.~Demuth$^2$\\
103: T.H.~Fields$^1$, H.R.~Gallagher$^5$, M.C.~Goodman$^1$,
104: R.~Gran$^2$, \\ T.~Joffe-Minor$^1$, T.~Kafka$^5$, S.M.S.~Kasahara$^2$,
105: P.J.~Litchfield$^4$, \\ W.A.~Mann$^5$, M.L.~Marshak$^2$,
106: R.H.~Milburn$^5$, W.H.~Miller$^2$, L.~Mualem$^2$, \\ A.~Napier$^5$,
107: W.P.~Oliver$^5$, G.F.~Pearce$^4$, E.A.~Peterson$^2$, D.A.~Petyt$^4$,\\
108: K.~Ruddick$^2$, M.~Sanchez$^5$, J.~Schneps$^5$, A.~Sousa$^5$,
109: \\ B.~Speakman$^2$, J.L.~Thron$^1$, S.P.~Wakely$^2$, N.~West$^3$\\
110: \\
111: $^1${\it Argonne National Laboratory, Argonne, IL 60439}\\
112: $^2${\it University of Minnesota, Minneapolis, MN 55455}\\
113: $^3${\it Department of Physics, University of Oxford, Oxford OX1 3RH, UK}\\
114: $^4${\it Rutherford Appleton Laboratory, Chilton, Didcot, Oxfordshire
115: OX11 0QX, UK}\\
116: $^5${\it Tufts University, Medford, MA 02155}\\
117: $^6${\it Western Washington University, Bellingham, WA 98225}\\
118: }
119:
120:
121: \begin{document}
122:
123: %\vskip -50pt
124:
125: \maketitle
126: \thispagestyle{empty}
127:
128: \begin{abstract}
129: \normalsize
130: We have searched for neutron-antineutron oscillations using
131: the 5.56 fiducial kiloton-year exposure of the Soudan 2
132: iron tracking calorimeter. We require candidate \nn occurrences
133: to have $\ge$ 4 prongs (tracks and showers) and to have kinematics
134: compatible with \nN annihilation within a nucleus.
135: We observe five candidate events,
136: with an estimated background from atmospheric neutrino and cosmic ray
137: induced events of 4.5 $\pm$ 1.2 events.
138: Previous experiments with smaller exposures observed no candidates,
139: with estimated background rates similar to this experiment.
140: We set a lifetime lower limit at 90\% CL for
141: the \nn oscillation time in iron:
142: $\TAn(Fe) > 7.2 \times 10^{31}$~years.
143: The corresponding lower limit for oscillation of free neutrons
144: is $\tnn > 1.3 \times 10^{8}$~seconds.
145:
146:
147: \vskip 20pt
148: \noindent PACS numbers: 11.30.Fs, 12.20.Fv, 12.60.Jv, 14.20.Dh
149:
150: \end{abstract}
151: \eject
152:
153: \large
154:
155: \section{Introduction}
156:
157: \subsection{Neutron-antineutron oscillations}
158:
159: An intriguing variation on the grand unification theme that nucleons
160: are likely to be unstable is the proposal that
161: neutrons can oscillate into antineutrons.
162: Neutron-antineutron oscillations were first predicted in 1970 by V. A. Kuzmin
163: in a model intended as a realization of requirements given earlier
164: by A.D. Sakharov for evolution of the universe to net baryon asymmetry
165: \cite{{Kuzmin_70},{Sakharov_67}}. Subsequently \nn
166: oscillations emerged as a predicted reaction in certain grand unification
167: theories \cite{GUTS_non-SUSY}.
168: More recently it has been shown that \nn oscillations can occur
169: in a large class of supersymmetric
170: $SU(2)_{L} \times SU(2)_{R} \times SU(4)_{c}$
171: models \cite{GUTS_SUSY}. In such models
172: the dominant baryon number violating process is a
173: ${\it \Delta} B = -2, {\it \Delta} L = 0$ nucleon transition
174: (e.g. \nn oscillations or p + n $\rightarrow$ pions)
175: rather than a ${\it \Delta} B = -1, {\it \Delta}L = -1
176: $ nucleon-antilepton transition
177: (e.g. p $\rightarrow e^{+} \pi^{0}$ or p $\rightarrow \overline{\nu} K^{+}$).
178: Neutron-antineutron oscillations have also been indicated as viable
179: by recent GUT models which invoke the existence
180: of extra spacetime dimensions \cite{GUTS_higherD}.
181:
182: If indeed a neutron can evolve into an antineutron,
183: the experimental signatures for the metamorphosis
184: should be distinctive. The resulting antineutron will
185: annihilate with a baryon of the surrounding environment, producing
186: multiple mesons ($B=0$) whose visible energy and net momentum
187: are approximately those of two nucleon masses having nuclear Fermi motion.
188:
189: From the phenomenology of neutron-antineutron oscillations it can
190: be shown \cite{ORNL-6910_Alberico} that the oscillation time $T_A$
191: of a neutron bound within a nucleus of atomic mass $A$ is related to
192: the neutron oscillation time in vacuum \tnn ~according to
193:
194: \begin{equation}
195: T_A = (\tnn)^2 \cdot T_R .
196: \label{equ:bound_neutron_time}
197: \end{equation}
198:
199: \noindent Here $T_R$, which has units of inverse time,
200: is the suppression factor representing the effect
201: of the nuclear environment which substantially prolongs the effective
202: oscillation time.
203:
204: Detailed calculations of the suppression factor $T_R$ for parent nuclei
205: of experimental interest, including deuterium, oxygen, argon, and iron,
206: have been reported in literature.
207: The calculations utilize phenomenological frameworks provided by
208: nuclear potential theory
209: \cite{{Dover_83_85_89},{Huefner_98}}, and by S-matrix theory
210: \cite{{Chetyrkin_81},{Alberico_82_84},{Alberico_91},{Kondratyuk_96}}.
211: In the analysis of Dover, Gal, and Richard
212: \cite{Dover_83_85_89},
213: it is proposed that neutron-antineutron oscillations
214: will occur mostly in outer nuclear shells
215: and near the nuclear surface.
216: However reservations concerning this picture
217: have been expressed and
218: in a number of calculations
219: the entire nuclear volume contributes to
220: \nn oscillations \cite{Huefner_98}.
221:
222:
223: \subsection{ Previous experimental searches }
224:
225: Two types of experiments have been used to search
226: for neutron-antineutron oscillations.
227: In one approach, slow neutrons from a fission reactor
228: are channeled through a magnetically shielded vacuum pipe towards a
229: target region. An antineutron produced during the flight
230: will annihilate in the target and the annihilation products are
231: registered by detectors surrounding the target.
232: Experiments of this type have been carried out at Pavia
233: \cite{Bressi_89} and at Grenoble \cite{{Fidecaro_85},{Baldo-Ceolin_90_94}}.
234: The Grenoble reactor experiment obtained
235: \tnn \bs $\ge 0.86 \times 10^8$ s at 90\% confidence level (CL). This is
236: the most stringent oscillation time lower limit reported to date using
237: free neutrons.
238:
239: The alternate approach, used in this experiment, is to continuously
240: monitor neutrons bound in nuclei, usually as part of an
241: ongoing nucleon decay search. Searches of this type have been reported
242: by the underground experiments Homestake \cite{Homestake_83},
243: NUSEX \cite{NUSEX_83}, KOLAR \cite{KOLAR_86},
244: IMB \cite{IMB_84}, Kamiokande \cite{KAM_86}, and Fr\'{e}jus \cite{Frejus_90}.
245: The searches by Kamiokande and Fr\'{e}jus
246: obtained the most stringent \nn oscillation time lower limits.
247:
248: In a search based upon a 1.11 kiloton-year (kty) exposure of the
249: Kamiokande-I water Cherenkov detector \cite{KAM_86},
250: no candidate \nN event was observed. An oscillation time lower limit
251: of $T_{A} >$ 4.3 $\times 10^{31}$ years at 90\% CL was set.
252: Using the suppression factor $T_{R} = 1 \times 10^{23}$ s$^{-1}$
253: calculated by Dover {\it et al.} for oxygen \cite{Dover_83_85_89},
254: Kamiokande obtained an oscillation time limit for free neutrons of
255: $\tnn > 1.2 \times 10^8$ seconds at 90\% CL.
256: The Fr\'{e}jus collaboration, in a search using a 1.56 fiducial kty exposure
257: of the experiment's planar iron tracking calorimeter,
258: also reported zero \nn oscillation candidates.
259: The oscillation time lower limit for $T_{A}$ in iron thereby obtained was
260: $6.5 \times 10^{31}$ years at 90\% CL.
261: Using $T_{R} = 1.4 \times 10^{23}$ s$^{-1}$
262: as calculated by Dover {\it et al.}
263: for iron, Fr\'{e}jus determined the free neutron limit to be also
264: $\tau_{n\overline{n}}$ $> 1.2 \times 10^8$ seconds
265: at 90\% CL \cite{Frejus_90}.
266:
267:
268: In the Kamiokande analysis an enhanced probability for \nn oscillations
269: to occur in the nuclear periphery as postulated by Dover {\it et al.} was
270: assumed. In the Monte Carlo simulations of the experiment, the effect of
271: this assumption is to reduce distortion of the final state meson spectrum
272: arising from intranuclear absorption and inelastic scattering processes.
273: As a result experimental detection efficiencies are enhanced
274: relative to expectations for the case where oscillations may occur
275: throughout the entire volume of parent nuclei.
276: The search reported here follows the more conservative approach
277: adopted previously by \frejusn. For our primary \nN simulation on
278: which our detection efficiency is based, we assume
279: \nn oscillation to occur throughout the nuclear volume.
280: However the effects of peripheral predominance are also described.
281:
282:
283: \section{Detecting \nn Oscillations in Soudan 2}
284:
285: \subsection{Detector and data exposure}
286:
287: Soudan 2 is a 963 metric ton (770 tons fiducial)
288: iron tracking calorimeter
289: of honeycomb lattice geometry which operates as a slow--drift
290: time projection chamber \cite{S2:NIM_A376_A381}. The
291: tracking elements are one-meter-long drift tubes
292: filled with an Argon--CO$_{2}$ gas mixture. Electrons liberated by
293: throughgoing charged particles drift to the tube ends under the action
294: of a voltage gradient applied along the tubes. The drift charge is
295: registered by vertical anode wires, while horizontal
296: cathode pad strips register the image charges. The third coordinate is
297: obtained from the drift--time. The amount of charge measures the
298: deposited ionization. The drift tubes
299: are laid onto corrugated steel sheets, and the sheets are stacked to form
300: 1$\times$1$\times$2.5 m, 4.3 ton modules
301: from which the calorimeter is assembled in building-block fashion.
302: Surrounding the tracking calorimeter on all sides but mounted on the
303: cavern walls and well separated from calorimeter surfaces is
304: a 1700 m$^{2}$ active shield array of two or three layers of
305: proportional tubes \cite{S2:NIM_A276}. The shield facilitates
306: identification of events which are not contained within the calorimeter.
307: In particular, it provides tagging of background events initiated by
308: cosmic-ray-induced neutrons.
309:
310: The detector is located at a depth of 2070 meters--water--equivalent
311: on the 27th level of the Soudan Underground Mine State Park
312: in northern Minnesota.
313: The modular design enabled data taking to commence
314: in April 1989 when the detector was one-quarter of its final size;
315: routine operation with the fully--deployed detector got underway
316: in November 1993. The fiducial (total) exposure analyzed here,
317: obtained from data--taking through December 2000,
318: is 5.56 (6.96) kiloton--years.
319:
320: Calibration of calorimeter module response
321: was carried out at the Rutherford ISIS spallation
322: neutron facility using test beams of positive and negative
323: pions, electrons, muons, and protons \cite{Wall_PRD62_00}.
324: Spatial resolutions for track reconstruction and for vertex
325: placement in anode, cathode, and drift time coordinates are of the
326: same scale as the drift tube radii, $\approx$ 0.7 cm.
327: In Soudan 2, ionizing particles having non-relativistic as well
328: as relativistic momenta are imaged with $dE/dx$ sampling in a
329: fine--grained honeycomb lattice geometry.
330: Protons can be distinguished from pions and muons
331: via ionization and ranging, energetic
332: muons discriminated from pions via absence of secondary scatters,
333: and prompt e$^{\pm}$ showers distinguished from photon showers
334: on basis of proximity to primary vertices.
335: These event imaging capabilities offer advantages, in comparison to
336: water Cherenkov detectors and to planar iron calorimeters,
337: for analysis of the complicated multiprong
338: topologies that would arise with \nN annihilations initiated by
339: \nn oscillations.
340:
341:
342: \subsection{Simulation of \nN annihilation arising from \nn oscillations}
343:
344: We have developed realistic simulations of \nn oscillations yielding
345: \nN annihilations as they would occur in the iron nuclei which
346: comprise the bulk of the calorimeter mass.
347: Generation of \nN events is carried out as follows:
348:
349: Momenta of the initial state antineutron and nucleon
350: are assigned according to a distribution based upon a Fermi-gas model
351: parameterization of quasi-elastic electron nucleus scattering \cite{Bodek_81}.
352: Final state particle four-momenta were
353: constructed in accordance with N-body phase space \cite{Chaffee_SAGE}.
354: Assignment of \nN reactions to generated events is weighted according to
355: cross section data for $\overline{\rm p}$p annihilation at rest
356: \cite{{Backenstoss_83},{pbar_p_Refs}}.
357: For the purpose of channel selection, $\overline{\rm n}$n annihilations
358: are assigned the same cross sections as observed for
359: $\overline{\rm p}$p;
360: $\overline{\rm n}$p is assumed to have the
361: same total rate as $\overline{\rm p}$p, and cross sections
362: were inferred from $\overline{\rm p}$p as allowed by charge conservation. We
363: restricted our reaction compilation to cross sections
364: exceeding 2\% of the total $\rm {\overline{N}N}$ cross section. Consequently
365: production of $\rho$, $\omega$, charged
366: and neutral pions and kaons is represented,
367: however final states with $\eta$ and $\eta'$ mesons were neglected.
368:
369:
370: Provision was made to include intranuclear rescattering (INS)
371: of final state pions in the simulations. Our treatment follows
372: the approach utilized previously in simulations of
373: atmospheric neutrinos and of nucleon decay \cite{Wall_PRD62_00}.
374: Pions, either directly produced in annihilation or in
375: $\rho$ decays, are propagated through a model nucleus.
376: The nuclear radius and density are parameters of the model;
377: the radius is scaled according to $A^{1/3}$.
378: Scattering in the nuclear medium is
379: characterized using a momentum-dependent pion interaction
380: length. The nuclear parameters were set by requiring the model
381: to reproduce single and multiple pion
382: production rates observed in bubble chamber $\nu_\mu$-deuteron ($A$=2)
383: and $\nu_\mu$-neon ($A$=20) interactions \cite{Merenyi_92}.
384: For pions from vertices placed at random within an iron nucleus,
385: our model predicts $\sim 45\%$ to emerge either unscattered or
386: to have undergone only small angle elastic scattering. About 30\% are
387: predicted to undergo total absorption,
388: while the remainder undergo charge
389: exchange (9\%) or emerge having undergone inelastic scattering
390: (16\%). Our treatment of nuclear rescattering does not include
391: proton secondaries. While it is expected that protons would
392: occasionally be ejected from parent nuclei
393: as the result of rescattering, their momenta would almost always fall
394: below the effective threshold (approx. 450 MeV/$c$) for creation of
395: distinct tracks in the detector.
396:
397: For simulations with intranuclear rescattering included,
398: the average multiplicity per event for mesons emerging from an
399: iron nucleus is 3.8 with rms deviation 1.2.
400: The average track plus shower multiplicity emitted from an
401: iron nucleus (``prong multiplicity") per event is 5.0 with
402: an rms deviation of 2.1 prongs. (Note these multiplicities are
403: for ``perfect detection".)
404:
405: The \nN event generation produces four-momenta of all particles
406: which exit the parent nuclei. This information is fed,
407: event-by-event, into the Soudan 2 Monte Carlo (MC)
408: which provides a realistic detector response.
409: The simulation includes detector background ``noise" arising from
410: natural radioactivity and from the electronics. Event records
411: are generated with format identical to that of data events, allowing
412: \nn oscillation events to be processed using the same codes and
413: procedures as for data and for events of the atmospheric
414: neutrino Monte Carlo.
415:
416:
417: \subsection{Properties of processed \nN samples}
418: \label{sec:processed_samples}
419:
420: Three separate $\overline {\rm n}$N simulation samples
421: were generated. For each sample, events were generated
422: at random locations throughout the tracking calorimeter as it evolved
423: during the experiment's eleven years of data-taking.
424: In our ``primary" simulation, \nN annihilations were started
425: at random points throughout their parent nuclei and pion intranuclear
426: rescattering was implemented. In a second simulation which also
427: included intranuclear rescattering, \nN annihilations were
428: restricted to the nuclear periphery ($0.75 R \le r \le R$, where $R$ is
429: the nuclear radius).
430: A third simulation was carried out in which no pion intranuclear
431: rescattering was included.
432:
433:
434: Each of the three oscillation samples was processed using
435: a sequence of selections and procedures very similar to that
436: routinely used in reduction of data events in Soudan 2.
437: Each sample was subjected (in software) to the
438: hardware trigger. Events which passed the hardware trigger
439: requirements were subjected to two different software ``Filter" codes.
440: The Filters impose event containment criteria, e.g. that no track from
441: the event approaches closer than 20 cm to the detector
442: outer surfaces; they also mitigate against backgrounds arising
443: from cosmic ray muons, natural radioactivity, and detector noise.
444: Events which survived the Filters were then subjected to two
445: separate scans by physicists. Scanning was carried out
446: using interactive color graphics workstations.
447: The scan rules provided refinements to the Filter selections
448: and introduced requirements on imaging quality, e.g. an event was
449: rejected if ({\it i}) its primary vertex occurred
450: in material interior to the detector but not instrumented,
451: or if ({\it ii}) its proximity to inter-module gaps
452: compromised the reconstruction of the event.
453:
454: Events which survived successive application of the
455: hardware trigger, software
456: filters, and physicists' scans are tallied in the upper four rows
457: of Table \ref{tbl:survival_rates} for each of the three simulations.
458: These entries are input to the calculation of the detection
459: efficiency for neutron-antineutron oscillations in Soudan 2 as will be
460: described in Sect.~\ref{sec:deteff}.
461:
462: %Table 1
463: \begin{table}[ht]
464: \centering
465: {\normalsize
466: \begin{tabular}{|l||c|c|c|}
467: \hline
468: \it Event Processing and & \it \nN annihilation & \it \nN annihilation in
469: & \it \nN annihilation in \\ %\vspace{-10pt}\\
470: \it Selection Stages & \it without INS & \it nuclear periphery
471: & \it nuclear volume \\ % \vspace{-10pt} \\
472: & & \it with INS &\it with INS \\
473: \hline \hline
474: Initial sample & 491 & 491 & 491 \\ \hline
475: Hardware trigger & 490 & 469 & 451 \\ \hline
476: Containment and quality filters & 288 & 301 & 286 \\ \hline
477: Two physicist scans & 214 & 229 & 205 \\ \hline
478: Multiplicity $\ge$ 4; no proton events & 172 & 148 & 135 \\ \hline
479: Exclude events with ``muons" & 146 & 130 & 123 \\ \hline
480: Kinematic selection on $E_{vis}$, $P/E$ & 137 & 102 & 86 \\
481: \hline
482: \end{tabular}
483: }
484: \caption{Survival through successive processing and selection stages
485: for events of three different simulations of \nn oscillations
486: yielding \nN annihilations in Soudan 2.}
487: \label{tbl:survival_rates}
488: \end{table}
489:
490: Preliminary to kinematic reconstruction, the topology of each event
491: was characterized in terms of track and shower prongs.
492: A track prong results when an ionizing, non-showering charged particle
493: (e.g. a pion, muon, or proton) traverses drift tubes of
494: the calorimeter's honeycomb lattice leaving a continuous trail
495: of tube ``hits".
496: A shower prong on the other hand is created by photon conversion or by
497: a primary electron or positron. The electromagnetic shower
498: consists of many distinct particles and exhibits a
499: cone-like pattern of hits, interspersed with gaps, which is aligned with
500: the direction of the initiating photon or electron. The mean conversion
501: length for photons in Soudan 2 is approximately 15 cm.
502: As a consequence, photon-induced showers generally appear in the vicinity of
503: but moderately displaced from event primary vertices.
504:
505:
506: Images of two \nn oscillation events simulated with full
507: detector response are displayed in Fig. \ref{fig:nnbmcevents.eps},
508: where the anode ($X$) versus drift time ($Z$)
509: view of each event is shown. For event scanning, three views
510: are always used, including cathode-time and anode-cathode as well
511: as anode-time projections. (The anode-cathode view is automatically
512: assembled for all events, MC as well as data, using demultiplexing
513: and hit-matching algorithms.)
514: As suggested by Fig. \ref{fig:nnbmcevents.eps},
515: $\overline {\rm n}$N events appear to be energetic
516: yet isotropic to a degree uncommon for data events and for
517: events of the atmospheric neutrino Monte Carlo.
518:
519:
520: %Figure 1: Two illustrative n-nbar simulated events.
521: \begin{figure}[hbt]
522: \centerline{\epsfig{figure=nnbfin01.eps,width=15.0cm}}
523: \caption{Anode versus drift time views (magnified) of two simulated
524: annihilation events within iron nuclei following \nn oscillation.
525: The multiprong hadronic final states of these events are isotropic
526: to a degree unusual for neutrino interactions.}
527: \label{fig:nnbmcevents.eps}
528: \end{figure}
529:
530:
531: The occurrence of relatively high multiplicity per event
532: of track plus shower prongs is a signature feature of \nN annihilation.
533: From simulation of \nn oscillation events in the absence of
534: intranuclear rescattering, we find the prong multiplicity distribution
535: - after triggering, filtering, scanning, and reconstruction
536: - to have a mean of 6.3 prongs per event with an rms width of 2.8 prongs.
537: Here the prong multiplicity is relatively high (compared to multiplicities
538: as generated, see Sect. 2.2), due to depletion of low multiplicities
539: by the processing.
540: With intranuclear rescattering included, the prong multiplicity
541: distributions of processed samples are shifted lower. For both
542: simulations with INS included, the multiplicity distributions have
543: mean values of 4.8 prongs with rms widths of 3.2 prongs.
544: Thus a sizable contribution from multiplicities
545: exceeding three-prongs survives in the latter simulations. Events of
546: this kind are relatively uncommon among contained atmospheric neutrino
547: events.
548:
549: In addition to event topology, the reconstructed kinematic quantities
550: $E_{vis}$, the visible energy, and $P_{net}$, the net momentum,
551: provide discrimination between \nN and background events. These quantities
552: and their correlation have been used previously in our searches for nucleon
553: decay \cite{Wall_PRD61_00}.
554: $E_{vis}$ is calculated as the sum of the relativistic energies
555: for each of the final state tracks (using the pion mass)
556: and showers. There occur a few short, heavily ionizing tracks which
557: satisfy our identification criteria for protons \cite{Allison_98}.
558: For these, only the kinetic energy is added into the calculation of $E_{vis}$.
559: Distributions of $E_{vis}$ versus $P_{net}$
560: for \nN events in the absence of and including
561: intranuclear rescattering, are plotted in
562: Fig. \ref{fig:nnb_evis_p.eps}a and Fig. \ref{fig:nnb_evis_p.eps}b
563: respectively. Reconstructed events, even in the
564: absence of INS as in Fig. \ref{fig:nnb_evis_p.eps}a, exhibit large
565: energy losses arising from low energy prongs and secondary hadronic
566: scatters which are unresolved.
567: Consequently the events cluster well below $E_{vis}$ of 1.88 GeV.
568: Comparison of Fig. \ref{fig:nnb_evis_p.eps}a with Fig.
569: \ref{fig:nnb_evis_p.eps}b shows that a further large degradation
570: arises with INS. The INS effects substantially increase
571: the kinematic overlap of
572: \nN events with atmospheric neutrino events.
573:
574:
575: %Figure 2: E_vis versus P_net for two different n-nbar simulations.
576: \begin{figure}[hbt]
577: %\centerline{\epsfig{figure=nnbfin02.eps,width=9.0cm}}
578: \vspace {320.0pt}
579: \special{psfile=nnbfin02.eps angle=0 hoffset=120 voffset=0 hscale=50 vscale=50}
580: \caption{Distributions of visible energy versus magnitude of
581: vector net momentum for reconstructed \nN annihilation events
582: of two independent full detector simulations.
583: The simulation plotted in \ref{fig:nnb_evis_p.eps}a
584: (\ref{fig:nnb_evis_p.eps}b)
585: omits (includes) intranuclear rescattering of final state
586: pions. The shift to lower $E_{vis}$ in simulation of b) versus a)
587: arises from pion absorption and inelastic scattering within
588: iron nuclei.}
589: \label{fig:nnb_evis_p.eps}
590: \end{figure}
591:
592:
593: \section{Data and Backgrounds for the \nN Search}
594:
595: \subsection{Neutrino, rock, and atmospheric $\nu$ MC events}
596: \label{sec:nuandrockevts}
597:
598: Since \nN events generally have high multiplicities we restrict our
599: analysis to events having a ``multiprong" topology, that is, having
600: two or more produced particles emerging from primary vertices.
601: Quasi-elastic neutrino interactions which produce a charged lepton plus
602: a recoil proton are readily distinguished and removed.
603: As was done for the simulated \nn events, we require
604: all events under consideration to be fully contained in an interior
605: volume which is everywhere 20 cm from calorimeter outer surfaces.
606:
607: Three multiprong event samples, which are in addition to
608: the simulated \nn samples discussed above,
609: have been isolated for this search:
610:
611: First and foremost are data events for which the cavern-liner
612: active shield array was quiescent during the allowed time window.
613: These events comprise our {\it shield-quiet data sample}. They
614: originate mostly with atmospheric neutrino interactions
615: but would also include \nn oscillation events.
616: There are 188 multiprong events
617: in the shield-quiet data of this analysis.
618:
619: There is a second category of data events which comprises a background
620: to neutrino multiprongs as well as to \nn oscillations. Events of this
621: category are usually produced by energetic neutrons released
622: in inelastic cosmic ray muon interactions with the cavern rock
623: surrounding the tracking calorimeter.
624: Most of these {\it rock events} are accompanied by charged
625: particles which give coincident hits in the
626: active shield array; they constitute our {\it shield-tagged rock
627: event sample}. However, a few rock events are not
628: accompanied by shield hits, either because of shield inefficiency
629: (detection efficiency is $94\%$) or because no charged particles
630: enter the cavern along with a rock neutron. The latter {\it shield-quiet
631: rock events} may end up as background in shield-quiet data.
632: Fortunately the shield-tagged rock events provide a control sample from
633: which the amount of residual rock background in shield-quiet data may be
634: inferred.
635:
636: The predominant background for \nn oscillations arises from
637: the atmospheric neutrino reactions. Their contribution is evaluated
638: using the Soudan 2 {\it atmospheric neutrino Monte Carlo sample}.
639: Our neutrino MC simulation utilizes the atmospheric flux calculation
640: of the Bartol group for the Soudan site \cite{Agrawal_96}.
641: Details of the neutrino event generation and comparisons with low
642: energy $\nu N$ data are given elsewhere \cite{Gallagher_thesis}.
643: MC atmospheric neutrino events were inserted into the
644: experiment's data stream during data-taking. The MC events were
645: introduced at a rate 6.06 times higher than the data rate expected
646: (for null neutrino oscillations) from the atmospheric neutrino flux.
647: The events were processed identically to real data events,
648: with their identity revealed only at the final analysis stage.
649: The atmospheric $\nu$ MC multiprong sample finally extracted
650: contains 1267 events.
651:
652: Properties of our multiprong data, rock, and neutrino MC samples
653: germane to an \nn oscillation search are reported below.
654: Further details can be found in previous publications
655: \cite{Wall_PRD62_00,Wall_PRD61_00,Allison_98}.
656:
657:
658: \subsection{Multiprong data compared to atmospheric $\nu$ MC}
659:
660: The prong multiplicity distribution of the shield-quiet data
661: sample is summarized using tallies in the grid displayed in
662: Table \ref{tbl:trk_vs_shw}. In the Table, the topology of each
663: event is represented by the number of tracks, $n_{track}$,
664: and number of showers, $n_{shower}$, which comprise the visible
665: final state. The number of data events which have a particular
666: $(n_{track},n_{shower})$ combination are given by the upper entries
667: at the grid location having the appropriate integer coordinates.
668: The multiprong data topologies having highest rates are seen to be
669: combinations of lowest multiplicity, namely
670: $(n_{track}, n_{shower}) = (2,0)$ and (1,1).
671: For comparison, the corresponding prong distribution
672: for the atmospheric neutrino MC event sample
673: is given by the lower entries in each of the
674: track-shower grid locations of Table \ref{tbl:trk_vs_shw}.
675: The MC sample (for null neutrino oscillations)
676: has been normalized to the same
677: fiducial exposure as for the data (1267/6.06 = 209.1 events).
678: The two distributions are seen
679: to match well, e.g. the $\chi^2$/bin averages to below 1.0, with
680: the exception of the (1,2) topology.
681: Since most shield-quiet data multiprongs are initiated by
682: atmospheric neutrinos, the agreement is evidence that
683: the neutrino MC provides a good general representation of our data.
684:
685:
686: %Table 2
687: \begin{table}[ht]
688: \centering
689: \begin{tabular}{|c|c|c|c|c|c|c|c|c}
690: \cline{1-1}
691: Number of &\mulc{7}{}{} \\
692: Tracks &\mulc{7}{}{} \\ \cline{1-8}
693: \rule[0.4cm]{0mm}{0cm}
694: \raisebox{-1.0ex}{5}
695: & {\it Data} & 0 & 0 & 0 & 2 & 1 & 1 & \\ % \cline{2-8}
696: & {\it MC} & 0.3 & 0.3 & 0.3 & 0 & 0.2 & 0.5 & \\ \cline{1-8}
697: \raisebox{-1.0ex}{4}
698: & {\it Data} & 3 & 1 & 3 & 0 & 1 & 0 & \\ % \cline{2-8}
699: & {\it MC} & 2.1 & 0.3 & 0.2 & 1.6 & 0.2 & 0.2 & \\ \cline{1-8}
700: \raisebox{-1.0ex}{3}
701: & {\it Data} & 16 & 7 & 2 & 2 & 0 & 0 & \\ % \cline{2-8}
702: & {\it MC} & 14.5 & 5.1 & 3.1 & 1.4 & 1.6 & 0.3 & \\ \cline{1-8}
703: \raisebox{-1.0ex}{2}
704: & {\it Data} & 26 & 20 & 15 & 4 & 0 & 2 & \\ % \cline{2-8}
705: & {\it MC} & 26.8 & 21.1 & 14.6 & 5.9 & 2.7 & 1.3 & \\ \cline{1-8}
706: \raisebox{-1.0ex}{1}
707: & {\it Data} & & 25 & 10 & 5 & 3 & 3 & \\ % \cline{2-8}
708: & {\it MC} & & 28.3 & 23.6 & 9.3 & 5.0 & 1.6 & \\ \cline{1-8}
709: \raisebox{-1.0ex}{0}
710: & {\it Data} & & & 14 & 13 & 4 & 1 & \\ % \cline{2-8}
711: & {\it MC} & & & 12.4 & 9.0 & 2.9 & 2.1 & \\ \hline
712: & & & & & & & & \mulc{1}{l|} {Number} \\
713: & & 0 & 1 & 2 & 3 & 4 & 5 & \mulc{1}{c|} {of} \\
714: & & & & & & & & \mulc{1}{c|}{Showers} \\
715: \hline
716: \end{tabular}
717: \caption{Event topology distributions: Event counts are tabulated according to
718: $(n_{track},n_{shower})$ combination per event, for multiprong events
719: of the data sample (upper entries) and of the atmospheric
720: neutrino MC sample (lower entries).
721: Event tallies of the latter sample represent expectations for
722: an atmospheric flux with null $\numu$ oscillations
723: normalized (factor of 6.06) to the data exposure.}
724: \label{tbl:trk_vs_shw}
725: \end{table}
726:
727: Kinematics for the shield-quiet multiprong data
728: is shown by the $E_{vis}$ versus vector net
729: momentum diplot of Fig. \ref{fig:evis_p_data_numc.eps}a.
730: The events populate a broad, correlated region extending from threshold
731: to 1.5 GeV (1.5 GeV/$c$) in visible energy (net momentum).
732: Portions of this data distribution overlap the distribution
733: predicted for \nn events shown in Fig. \ref{fig:nnb_evis_p.eps}b.
734:
735: A direct comparison with kinematics for
736: the atmospheric neutrino MC sample is provided by
737: Fig. \ref{fig:evis_p_data_numc.eps}b. Note that the
738: latter sample has an exposure-equivalent of 33.7 fiducial kton-years.
739: The distribution of the MC sample, normalized to the exposure, is very
740: similar to that of the data.
741:
742: In Fig. \ref{fig:evis_p_data_numc.eps}, events which have
743: prompt protons are denoted by solid circles. As previously
744: noted, such events are highly improbable as \nn oscillations.
745: However our atmospheric $\nu $ MC indicates that detectable
746: recoil protons are to be
747: expected in 24\% of neutrino multiprongs, and this expectation
748: is born out by the frequency of recoil proton events observed in the
749: data of Fig. \ref{fig:evis_p_data_numc.eps}a.
750: In Fig. \ref{fig:evis_p_data_numc.eps}a there
751: is a tendency for data events having recoil protons to occur in a region
752: parallel to and below the more populated kinematic band.
753: This trend is reproduced in the MC distribution
754: of Fig. \ref{fig:evis_p_data_numc.eps}b.
755:
756:
757: %Figure 3: E_vis versus P_net for Data and for Atm. neutrino MC samples.
758: \begin{figure}[hbt]
759: %\centerline{\epsfig{figure=nnbfin03.eps,width=9.0cm}}
760: \vspace {300.0pt}
761: \special{psfile=nnbfin03.eps angle=0 hoffset=120 voffset=0 hscale=45 vscale=45}
762: \caption{$E_{vis}$ versus $P_{net}$ distributions for shield-quiet
763: data events and for
764: events of the atmospheric neutrino MC sample.
765: Events without (with) proton
766: tracks are depicted using open (solid) circles.
767: The $\nu$ MC sample (\ref{fig:evis_p_data_numc.eps}a),
768: which corresponds to an
769: exposure (for null $\nu$ oscillatiions) of
770: 6.06 times the data, reproduces the general
771: kinematic trends of multiprong data
772: (\ref{fig:evis_p_data_numc.eps}b).}
773: \label{fig:evis_p_data_numc.eps}
774: \end{figure}
775:
776:
777: \subsection{Background from rock events}
778: \label{sec:rockbkgrdindata}
779:
780: Rock data events which are flagged by hits in the surrounding
781: shield array, have a distribution of vertex depth into the detector
782: which reflects the mean hadronic interaction length of neutrons
783: in the Soudan detector medium, approximately 80 cm.
784: A rock background component in our shield-quiet multiprong
785: data will have a corresponding vertex depth distribution in the detector.
786: Neutrino and \nn oscillation events on the other hand, will
787: distribute uniformly throughout the detector volume.
788: Thus by fitting the distribution of vertex depths observed
789: in our multiprong data to the sum of two component distributions
790: representing vertex depths of shield-tagged rock and of neutrino MC
791: events, the amount of shield-quiet rock contamination
792: can be determined \cite{Allison_98}. On the
793: basis of a maximum likelihood fit, we estimate that a fraction
794: $0.06^{+0.11}_{-0.06}$ of shield-quiet data multiprongs
795: are produced by rock neutrons.
796:
797:
798: \section{Selection of \nn Oscillation Events}
799: \label{sec:Selection_of_Events}
800:
801: Four selection criteria for neutron oscillation candidates
802: (Table~\ref{tbl:criteria})
803: have been defined and applied to all event samples. The
804: selections are designed to maximize \nn detection in the shield-quiet
805: multiprong data sample while minimizing backgrounds, the principal one
806: arising from inelastic interactions of atmospheric neutrinos.
807:
808: %Table 2
809: \begin{table}[ht]
810: \centering
811: {\normalsize
812: \begin{tabular}{|c|l|}
813: \hline
814: \it Cut & \it Definition \\
815: \hline \hline
816: 1. & $N_{prong} \ge 4$ \\ \hline
817: 2. & No prompt ``proton" tracks \\ \hline
818: 3. & No prompt ``muon" tracks longer than 150 cm \\ \hline
819: 4. & Evt kinematics: $P/E < 0.6, 700 < E_{vis} < 1800$ MeV \\
820: \hline
821: \end{tabular}
822: }
823: \caption{Selection cuts for \nn oscillation events in Soudan 2.}
824: \label{tbl:criteria}
825: \end{table}
826:
827:
828: \newcounter{bean}
829: \begin{list}
830: {\it\roman{bean})}{\usecounter{bean}
831: \setlength{\rightmargin}{\leftmargin}}
832:
833: \item[1.]
834: {\it Prong multiplicity}:
835: We require the sum of track and shower prongs
836: directly associated with the primary
837: vertices to be greater than or equal to four. This value is below
838: but near to the mean multiplicities for processed \nn samples as
839: discussed in Sect. \ref{sec:processed_samples}.
840: Lowering our selection to include three-prongs would increase
841: the efficiency for \nN detection by only 3\%, however
842: backgrounds arising from atmospheric neutrinos
843: and rock events would increase by 21\% and 32\% respectively.
844:
845: \item[2.]
846: {\it Primary proton tracks}:
847: Events having proton tracks emerging from their primary vertices are
848: removed from consideration as \nn oscillation candidates. The rate
849: of prompt protons is negligible for \nN annihilation events, however
850: $\sim 24\%$ of atmospheric neutrino events and $\sim 33\%$ of rock
851: events have visible recoil protons.
852:
853:
854: \item[3.]
855: {\it Primary muon tracks}: Pions produced in \nN annihilations usually
856: scatter, become absorbed, or range-to-stopping over distances comparable
857: to the calorimeter's hadronic interaction length of 80 cm. In order to
858: mitigate against $\numu$ charged current background, we regard
859: any non-scattering track which has pion/muon ionization
860: and range-to-stopping $\ge$ 150 cm (240 g/cm$^{2}$) to be a muon track.
861: Any event having such a track is rejected. This selection eliminates 36\%
862: of the atmospheric neutrino sample while cutting less than 10\% of \nn
863: events of the primary simulation.
864:
865: \item[4.]
866: {\it Event kinematics}:
867: As our final criterion we require the kinematics of
868: \nn candidates to be compatible with those of \nN annihilation.
869: The event net momentum fraction $P_{net}/E_{vis}$
870: (hereafter $P/E$) has previously been shown to be a useful variable for
871: separating \nn oscillation events from
872: atmospheric neutrino reactions \cite{Frejus_90}, and its utility
873: is confirmed in our analysis.
874: Figures \ref{fig:nnb_sims_final.eps}a and
875: \ref{fig:nnb_sims_final.eps}b show $P/E$ versus $E_{vis}$ for \nN events
876: after topology cuts (1 through 3)
877: from simulations in which oscillations occur
878: throughout the parent nuclear volumes. For the case of no
879: intranuclear rescattering (Fig. \ref{fig:nnb_sims_final.eps}a),
880: approximately 94\% of \nn oscillation events occur in a region
881: having $P/E$ less than 0.6, with $E_{vis}$ ranging
882: from one to two nucleon masses. With intranuclear rescattering included,
883: Fig. \ref{fig:nnb_sims_final.eps}b, the \nn event
884: population shifts to lower values of $E_{vis}$ and to higher
885: values of $P/E$. Nevertheless a degree of clustering remains which allows a
886: useful search region in the ($E_{vis}$, $P/E$) plane to be defined.
887: Our selection for \nn kinematics is depicted by
888: the rectangular region (dotted-line boundary) shown in
889: the Figs \ref{fig:nnb_sims_final.eps}-\ref{fig:MPDATA-FINAL.PS}.
890: Candidate \nn events are required to have
891: $P/E$ $<$ 0.6 and $700 < \evisn < 1800$ MeV.
892: These cuts are designed to minimize the atmospheric neutrino
893: background shown in Fig. \ref{fig:MPMC-FINAL.PS}
894: while maximizing the acceptance for \nN events.
895:
896: \end{list}
897:
898: %Figure 4: Kinematic distributions for two n-nbar simulations after cuts.
899: \begin{figure}[hbt]
900: %\centerline{\epsfig{figure=nnbfin04.eps,width=9.0cm}}
901: \vspace {300.0pt}
902: \special{psfile=nnbfin04.eps angle=0 hoffset=120 voffset=0 hscale=50 vscale=50}
903: \caption{Momentum fraction versus visible energy for events from
904: \nn simulation after topology cuts,
905: with intranuclear rescattering omitted
906: (\ref{fig:nnb_sims_final.eps}a) versus included
907: (\ref{fig:nnb_sims_final.eps}b). The kinematic search
908: region of this experiment enclosed by the dotted-line
909: boundary, contains 94\% of events which pass topology
910: selections. In the more realistic simulation of
911: \ref{fig:nnb_sims_final.eps}b however, INS degrades \evis
912: and increases momentum anisotropy on average,
913: leaving 70\% of events in the search region.}
914: \label{fig:nnb_sims_final.eps}
915: \end{figure}
916:
917: %Figure 5: Kinematic distribution of atm. neutrino MC events after cuts.
918: \begin{figure}[hbt]
919: %\centerline{\epsfig{figure=nnbfin05.eps,width=9.0cm}}
920: \vspace {150.0pt}
921: \special{psfile=nnbfin05.eps angle=0 hoffset=120 voffset=0 hscale=50 vscale=50}
922: \caption{Kinematic distribution of atmospheric neutrino MC events
923: which survive topology selections for \nn candidates. Events
924: which occur within the bordered region represent the background
925: expected from an exposure (assuming null neutrino oscillations)
926: which is 6.06 times the data exposure.}
927: \label{fig:MPMC-FINAL.PS}
928: \end{figure}
929:
930:
931: \section{\nn Candidates and Backgrounds}
932: \label{sec:deteff}
933:
934: The reductions in each \nN annihilation sample upon
935: successive \nn candidate selections are summarized
936: in the bottom three rows of Table \ref{tbl:survival_rates}.
937: It can be seen that intranuclear rescattering significantly
938: lowers the survival rates.
939: For our primary simulation, with INS operative and with
940: annihilations occurring throughout the volume of the parent nucleus,
941: 86 events from an initial 491 event sample survive all selections,
942: giving a detection efficiency of 0.18 $\pm$ 0.02.
943: Detection efficiencies for the comparison simulations are
944: higher, $(28 \pm 3)\%$ for the simulation without INS, and
945: $(21 \pm 2)\%$ for the simulation which includes INS but
946: restricts annihilations to the nuclear periphery.
947:
948: Shield-quiet data events which satisfy \nn selections 1--3 of
949: Table~\ref{tbl:criteria}
950: distribute in the $P/E$ versus \evis plane as shown
951: in Fig. \ref{fig:MPDATA-FINAL.PS}. Of these sixteen events,
952: five occur within the kinematically allowed region and are
953: candidate \nn oscillation events of this search. Three of the
954: candidates have four-prong topologies, while the remaining two are
955: five-prong events.
956:
957: Two views of one of the candidates consisting of four pion-or-muon tracks
958: emerging cleanly from a reaction vertex,
959: are shown in Fig. \ref{fig:R68882.PS}. This event is also consistent with
960: multipion production by atmospheric neutrinos,
961: e.g. $^(\bar\nu^) + n \rightarrow \mu^\pm \pi^\mp \pi^+ \pi^- (N)$
962: \cite{Derrick_84}.
963:
964:
965: %Figure 6: Kinematic distribution of data events after cuts.
966: \begin{figure}[hbt]
967: %\centerline{\epsfig{figure=nnbfin06.eps,width=9.0cm}}
968: \vspace {150.0pt}
969: \special{psfile=nnbfin06.eps angle=0 hoffset=120 voffset=0 hscale=50 vscale=50}
970: \caption{Distribution of data multiprong events after
971: topology selections, in the P/\evis versus \evis plane.
972: Five events are observed to have kinematics compatible
973: with \nn oscillations in the Soudan 2 detector.}
974: \label{fig:MPDATA-FINAL.PS}
975: \end{figure}
976:
977:
978: %Figure 7: Neutron oscillation candidate with four tracks.
979: \begin{figure}[hbt]
980: %\centerline{\epsfig{figure=nnbfin07.eps,width=8.0cm}}
981: \vspace {300.0pt}
982: \special{psfile=nnbfin07.eps angle=0 hoffset=150 voffset=0 hscale=75 vscale=75}
983: \caption{Anode-time and cathode-time views of
984: a four-track \nn oscillation candidate.}
985: \label{fig:R68882.PS}
986: \end{figure}
987:
988:
989: Fig. \ref{fig:MPMC-FINAL.PS} shows that, among
990: atmospheric $\nu$ MC events which satisfy topology
991: selections 1--3 of Table~\ref{tbl:criteria}, there are
992: 28 events (11 $\numu$~CC, 7 $\nue$~CC, and
993: 10 NC) which occur in the selected kinematic region.
994: That is, 28 neutrino-induced background events would occur in a
995: 33.7 fiducial kiloton-year exposure to an atmospheric neutrino
996: flux having null neutrino oscillations.
997: However strong evidence for atmospheric neutrino oscillations
998: now exists \cite{SuperK_OSC}, \cite{MACRO_OSC},
999: \cite{{Allison_98},{Mann_Sudbury}}. We allow for neutrino
1000: oscillations in our background estimate by weighting our
1001: MC $\numu$~CC prediction by the atmospheric neutrino flavor
1002: ratio-of-ratios: $R_\nu = [(\numu + \mbox{$\overline{\nu}_{\mu}$})/(\nue +
1003: \mbox{$\overline{\nu}_{e}$})]_{Data}/[(\numu + \mbox{$\overline{\nu}_{\mu}$})/
1004: (\nue + \mbox{$\overline{\nu}_{e}$})]_{MC}$.
1005: We have used the value of $R_\nu$ measured in this experiment,
1006: based upon a 5.14 fiducial kton-year exposure:
1007: $ R_\nu = 0.68 \pm 0.11 \pm 0.06$~\cite{Mann_Sudbury}.
1008: This correction reduces the neutrino background sample
1009: to 24.5 events.
1010: After scaling to the data exposure and accounting for
1011: statistical error plus uncertainties arising from our atmospheric neutrino
1012: MC including the neutrino oscillation correction (see
1013: Sect.~\ref{sec:errors_bkgrds}), we obtain
1014: $4.0 \pm 1.0 \bs $ events as the estimated neutrino-induced
1015: background of our \nn oscillation search.
1016:
1017:
1018: As discussed in Sect.~\ref{sec:nuandrockevts}, neutron-induced
1019: rock events which are background for \nn oscillations
1020: are those which elude tagging by the active shield.
1021: As described in Sect.~\ref{sec:rockbkgrdindata},
1022: we estimate that a fraction $0.06^{+0.11}_{-0.06}$ of data multiprongs
1023: are rock events. Thus of the 188 shield-quiet data multiprongs,
1024: we estimate $11.3^{+20.6}_{-11.3}$ to be rock events.
1025: However, among the 375 shield-tagged rock multiprong events,
1026: only fifteen pass our four \nn selection criteria.
1027: Thus, assuming that the selection efficiency of shield-quiet rock events
1028: is the same as that of shield-tagged rock events, the background of rock events
1029: which pass all \nn selections is estimated to be $0.5^{+0.8}_{-0.5}$ events.
1030:
1031: In summary, we estimate the number of background events to be
1032: $4.0 \pm 1.0 \bs (atm. \bs \nu)$ plus $0.5^{+0.8}_{-0.5} \bs (rock) =
1033: 4.5 \pm 1.2$ events.
1034:
1035:
1036: \section{Neutron Oscillation Time Lower Limits }
1037:
1038: Having observed five candidate \nn oscillation events and
1039: having calculated background neutrino and cosmic ray processes to
1040: contribute 4.5 events, we set a lower limit
1041: for the \nn oscillation time $T_A$ in iron nuclei.
1042: We follow previous experiments in using our best estimate of the background
1043: in determining the limit. Additional uncertainties in the limit due to the
1044: uncertainties in the background calculation are discussed in
1045: Sect. \ref{sec:stat_and_syst_errors}. We use the approach
1046: to confidence level construction as formulated by Feldman
1047: and Cousins \cite{{Feldman_Cousins},{PDG_00}}.
1048: For five candidate events with 4.5 background events the
1049: signal limit at 90\% CL is $n_{90} = 5.5$ events.
1050: The oscillation time lower limit is then calculated using
1051:
1052: \begin{equation}
1053: \centering
1054: T_{A} > \frac{N_{n} \cdot T_{f} \cdot \epsilon}{n_{90}}
1055: \label{equ:bound_neutron_limit}
1056: \end{equation}
1057:
1058: Here, $N_n = 3.15 \times 10^{32}$ neutrons in a kiloton
1059: of the Soudan 2 detector, $T_f = 6.96$ kiloton years
1060: is the full detector exposure, and $\epsilon$ = 0.18 is
1061: the efficiency for detection of \nn oscillations in the full detector mass,
1062: calculated using our primary \nn simulation (see Sect.~\ref{sec:deteff}).
1063: At 90\% CL we obtain $T_{A} > 7.2 \times 10^{31}$ years.
1064:
1065: Equation (\ref{equ:bound_neutron_time}) shows that the free \nn
1066: oscillation time \tnn \bs can be inferred from oscillation time
1067: \TA for neutrons of nucleus $A$.
1068: For the suppression factor of iron we use the value
1069: $T_{R} = 1.4 \times 10^{23} \bl {\rm s}^{-1}$
1070: obtained by Dover, Gal, and Richard upon
1071: averaging over calculations using different nuclear optical potentials
1072: \cite{Dover_83_85_89}.
1073: Similar or moderately higher \TR values have been reported from
1074: other calculations\cite{{Huefner_98},{Alberico_91}}.
1075: At 90\% CL we obtain $\tnn > 1.3 \times 10^{8}$ seconds.
1076:
1077:
1078: Oscillation time lower limits based upon
1079: detection efficiencies of each of our three \nn simulations
1080: are summarized in Table \ref{tbl:S2_search_limits}.
1081: The range of values which results from the variation in $\epsilon$
1082: illustrates the important role of the intranuclear rescattering treatment.
1083: The alternative simulations serve as estimators of
1084: sensitivity to systematic uncertainties arising
1085: from modelling of the nuclear environment.
1086:
1087:
1088: %Table 2
1089: \begin{table}
1090: \centering
1091: \begin{tabular}{|l||c|c|}
1092: \hline
1093: \it \nn Simulation & \it $T_A$ lower limit & \it \tnn\ lower limit \\
1094: \hline \hline
1095: \nn without INS &$ 11. \times 10^{31}$ y &$ 1.6 \times 10^8$ s \\ \hline
1096: \nn in Nuclear Periphery, with INS & $ 8.4 \times 10^{31}$ y &
1097: $ 1.4 \times 10^8 $ s \\ \hline
1098: \nn throughout Nucleus, with INS &$ 7.2 \times 10^{31}$ y
1099: &$ 1.3 \times 10^8 $ s \\
1100: \hline
1101: \end{tabular}
1102: \caption{Neutron-antineutron oscillation time lower limits (90\% CL)
1103: arising from three different treatments of pion
1104: intranuclear scattering in \nN annihilations. Simulations of
1105: the second and third rows are regarded to be realistic; the
1106: one with more conservative nuclear modelling (bottom row)
1107: is the basis for the limits for this search.}
1108: \label{tbl:S2_search_limits}
1109: \end{table}
1110:
1111:
1112: \section{Statistical and Systematic Uncertainties}
1113: \label{sec:stat_and_syst_errors}
1114:
1115: Our lower limits for \TA and \tnn ~contain
1116: uncertainties arising from finite statistics of data and
1117: simulation samples and from systematic errors inherent in
1118: our analysis techniques. To quantify these uncertainties,
1119: we have evaluated the error contributions which enter
1120: via each of the factors of Eq. (\ref{equ:bound_neutron_limit}).
1121:
1122: \subsection{ Exposure and \nn detection efficiency}
1123:
1124: The $N_n$ value \cite{S2:NIM_A376_A381} and
1125: the detector exposure are known accurately,
1126: with a consequent fractional error
1127: for the factor $N_n \times T_f$ of order 2\%.
1128:
1129:
1130: The \nn detection efficiency $\epsilon$
1131: has statistical error due to the finite sample size of our
1132: primary simulation at generation and especially after cuts.
1133: Furthermore, systematic uncertainties may arise due to inaccuracies
1134: in the simulation. Based upon trial variations of
1135: relative rates among dominant \nN channels (cross sections known to
1136: $\leq 20\%$), we estimate channel rate uncertainties to
1137: contribute $\sim 5\%$ uncertainty to $\epsilon$.
1138: A systematic error in $\epsilon$ may arise from our placement of
1139: annihilation sites within parent nuclei. Based upon our alternative
1140: simulations, we assign a 17\% fractional uncertainty to modelling details.
1141: There are uncertainties inherent with our treatment of pion
1142: intranuclear rescattering. Compared to the unphysical case wherein
1143: INS is neglected, our treatment gives rise to an efficiency decrease
1144: $\Delta \epsilon = 0.10$. Uncertainty in the amount of this
1145: decrease arises from finite statistics of bubble chamber
1146: neutrino samples \cite{Merenyi_92} and with
1147: extrapolation from A = 20 to 56 of the component INS processes,
1148: namely pion absorption, charge
1149: exchange, inelastic and elastic scattering. We estimate the
1150: latter uncertainties to total 30\% \cite{Wall_PRD61_00}.
1151: We then infer the fractional error contribution to $\epsilon$
1152: arising from our INS treatment to be $17\%$. Combining the above
1153: uncertainties in quadrature, we estimate the total fractional
1154: uncertainty $\delta \epsilon/\epsilon$ to be 27\%.
1155:
1156:
1157: \subsection{ Background estimation }
1158: \label{sec:errors_bkgrds}
1159:
1160: The value of $n_{90}$ relies upon our estimation of rates for
1161: background events which satisfy the \nn oscillation search criteria.
1162:
1163: Our estimate of atmospheric neutrino backgrounds is susceptible
1164: to errors from three sources: ({\it i}) Uncertainty arises from
1165: event statistics (after cuts) of the atmospheric
1166: neutrino simulation (19\%); ({\it ii}) The normalization of the
1167: $\nu$ MC to the experimental exposure, which predicts 183
1168: neutrino events will be observed, is uncertain. From our data we
1169: observe (data-rock) = 177 $\pm$ 20 events, from which we infer an
1170: error (11\%). ({\it iii}) The atmospheric $\nu$ MC may not fully
1171: represent aspects of neutrino data which feature in the
1172: selections of Sect. \ref{sec:Selection_of_Events}. However our data
1173: imply limits on the extent of MC mis-representation.
1174: For data events, and with respect to the cuts applied in
1175: Sect. \ref{sec:Selection_of_Events}, we observe 34 $\pm$ 5\% to have
1176: $n_{prongs}$ $\ge$ 4, 78 $\pm$ 9\% to be devoid of protons, 67 $\pm$ 8\%
1177: to be devoid of ``muons", and 43 $\pm$ 6\% to have kinematics in the
1178: vicinity (500 $\le$ $\evis$ $\le$ 1800 MeV, $P/E$ $\le$ 0.8)
1179: of our search region. The corresponding fractions for the $\nu$ MC
1180: sample are 35\%, 76\%, 64\%, and 39\% respectively. The agreement
1181: is good, and we take the quadrature sum of the fractional differences
1182: $\mid Data - MC\mid/Data$ in the four ratios as our estimator of uncertainty
1183: arising from MC representation of atmospheric neutrino physics (11\%).
1184: The total fractional error for our atmospheric neutrino background
1185: is then $25\%$, which corresponds to an error of $\pm$ 1.0 events
1186: on our estimate of 4.0 neutrino-induced background events.
1187:
1188: Concerning our estimate of background from rock events, the uncertainty
1189: arising from the determination of rock-event contribution to the vertex depth
1190: distribution of shield-quiet multiprongs outweighs any other systematic
1191: uncertainty.
1192:
1193: The total absolute error assigned to our
1194: 4.5 background events from atmospheric neutrinos and rock
1195: amounts to $\pm$ 1.2 events.
1196: Variation of our background estimate by $\pm$1 $\sigma$ yields
1197: $\delta n_{90}/n_{90} = 21\%$. The total uncertainty in our
1198: oscillation time lower limit value for \TA thereby implied by
1199: Eq. (\ref{equ:bound_neutron_limit}) is
1200: $\delta \TA/\TA = 34\%$.
1201:
1202:
1203: Our limit for the oscillation time of free neutrons \tnn
1204: depends upon the nuclear suppression
1205: factor \TR as implied by Eq. (\ref{equ:bound_neutron_time}).
1206: For \TR of iron, we used the value of Dover \etal \cite{Dover_83_85_89},
1207: $\TRn = 1.4 \times 10^{23} \bl \mb{s}^{-1}$.
1208: However from the range of \TR($A$=56) values indicated by
1209: the calculation of Alberico \etal \cite{Alberico_91},
1210: a theoretical uncertainty as large as 100\% may be inferred.
1211: Then the fractional error in \tnn is $\del \tnn/\tnn \approx 53\%.$
1212:
1213: We conclude that the uncertainties $\delta \TAn/\TA$ and $\delta \tnn/\tnn$
1214: on the oscillation time lower limits obtained in this work may be as large
1215: as 34\% and 53\% respectively. Of course, comparable uncertainties also
1216: apply to other published limits on \nn oscillation times.
1217:
1218:
1219: \section{Comparison with Previous Experiments}
1220:
1221: Table \ref{tbl:Experiments_and_Results} compares our results with those
1222: obtained by the three most recent of previous searches. The numbers of
1223: candidate events and corresponding estimates for background rates are
1224: shown in columns four and five. Background estimates for the underground
1225: experiments (1.1, 2.5, 4.5 events for Kamiokande, \frejusn, and Soudan 2
1226: respectively) are based upon simulations of atmospheric neutrino interactions
1227: in the detectors. For the \frejus experiment, an alternative estimate
1228: of 2.1 background events was obtained based upon interactions recorded
1229: using planar spark chambers exposed to beams of neutrinos and antineutrinos
1230: at the CERN-PS~\cite{Frejus:NIM_A302}.
1231: Soudan 2 is the underground experiment with the largest
1232: exposure. It is also the only experiment to observe candidate events, although
1233: all of the underground experiments estimated about one background event per
1234: kiloton year.
1235:
1236:
1237: %\eject
1238:
1239: %Table 4
1240: \begin{table}
1241: \centering
1242: %\footnotesize
1243: {\normalsize
1244: \tabcolsep 0.01cm
1245: \begin{tabular}{|l|c|c|c|c|c|c|c|}
1246: \hline
1247: & \it Source of & \it Exposure & \it Cand.
1248: & \it Est. & \TR & \TA &\tnn \\
1249: \it Experiments & \it neutrons & \it fiducial (total) & \it Events &
1250: \it Bkgrd & & & \\
1251: & &kton-years & & &$10^{23}$\ s$^{-1}$ &$10^{31}$\ y &10$^8$\ s \\
1252: \hline \hline
1253: Grenoble('90) \cite{Baldo-Ceolin_90_94}&reactor beam &- &0 &0 &- &- & 0.86 \\ \hline
1254: Kamiokande('86) \cite{KAM_86}&$^{16}$O
1255: & 1.11 & 0 & 1.2 & 1 & 4.3 & 1.2 \\ \hline
1256: Fr\'{e}jus('90) \cite{Frejus_90}&$^{56}$Fe
1257: &1.56 (2.56) &0 &2.5,2.1 & 1.4 & 6.5 & 1.2 \\ \hline
1258: Soudan 2 [this study]&$^{56}$Fe
1259: &5.56 (6.96) &5 &4.5 &1.4 & 7.2 & 1.3 \\
1260: \hline
1261: \end{tabular}
1262: }
1263: \caption{Experimental lower limits at 90\% CL for \nn oscillation times
1264: \TA and \tnn\ of bound and free neutrons respectively. For results
1265: of this search (bottom row), the background estimate is corrected
1266: for $\nu_\mu \rightarrow \nu_\tau$ oscillations and limits
1267: are calculated using the Feldman-Cousins method.}
1268: \label{tbl:Experiments_and_Results}
1269: \end{table}
1270:
1271:
1272: A degree of caution is warranted when comparing
1273: oscillation time lower limits for free neutrons as calculated by
1274: underground experiments versus a reactor neutron beam
1275: experiment (last column of Table \ref{tbl:Experiments_and_Results}).
1276: The \tnn limits of Soudan 2, \frejusn, and Kamiokande
1277: are based upon values for the neutron suppression factor
1278: \TR as calculated by Dover {\it et al.} \cite{Dover_83_85_89}.
1279: However if \TR values for iron and oxygen are considered which are
1280: at the high end of ranges obtained by Alberico {\it et al.}
1281: \cite{Alberico_91}, then limits at or below the Grenoble value
1282: are implied for the underground experiments.
1283:
1284:
1285: \section{Summary And Conclusion}
1286:
1287: The fine-grained Soudan 2 tracking calorimeter has been used in a
1288: search for neutron-to-antineutron oscillations occurring with bound
1289: neutrons. Based upon a fiducial exposure of 5.56 kiloton-years of the
1290: underground detector, a new oscillation time lower limit for \nn
1291: oscillations in iron nuclei has been determined at 90\% CL:
1292: \begin{eqnarray}
1293: \TA(Fe) > 7.2 \times 10^{31} \bs \mb{y}.
1294: \end{eqnarray}
1295:
1296: Assuming the suppression factor for iron is
1297: $\TR = 1.4 \times 10^{23} \bs \mb{s}^{-1}$~\cite{Dover_83_85_89},
1298: the corresponding limit at 90\% CL for n to \n oscillations
1299: of free neutrons is
1300: \begin{eqnarray}
1301: \tnn > 1.3 \times 10^{8} \bs \mb{s}.
1302: \end{eqnarray}
1303:
1304: The search reported here is background-limited.
1305: Candidate events are observed to occur at a rate similar to that
1306: predicted for backgrounds. Since the predominant background arises
1307: from kinematic overlap with multiprong reactions
1308: initiated by atmospheric neutrinos,
1309: it seems unlikely that future underground experiments
1310: can avoid also becoming background-limited
1311: in larger exposures.
1312: Thus reactor neutron beam experiments,
1313: rather than underground experiments monitoring bound neutrons,
1314: may offer a more promising route for future improvements in sensitivity
1315: to \tnn (see \cite{Kamyshkov_99} and references therein).
1316:
1317:
1318: \section*{Acknowledgements}
1319:
1320: This work was supported by the U.S. Department of Energy, the U.K. Particle
1321: Physics and Astronomy Research Council, and the State and University of
1322: Minnesota. We gratefully acknowledge the Minnesota Department of Natural
1323: Resources for allowing us to use the facilities of the Soudan Underground
1324: Mine State Park.
1325:
1326:
1327: \begin{thebibliography}{99}
1328:
1329: \bibitem{Kuzmin_70} V.A. Kuzmin, JETP Lett. {\bf 12}, 228 (1970).
1330:
1331: \bibitem{Sakharov_67} A.D. Sakharov, JETP Lett. {\bf 5}, 24 (1967).
1332:
1333: \bibitem{GUTS_non-SUSY} R.N. Mohapatra and R.E. Marshak,
1334: Phys. Rev. Lett. {\bf 44}, 1316 (1980);
1335: Phys. Lett. {\bf 94B}, 183 (1980);
1336: L.-N. Chang and N.-P. Chang, Phys. Lett. {\bf 92B}, 103 (1980);
1337: T.K. Kuo and S. Love, Phys. Rev. Lett. {\bf 45}, 93 (1980);
1338: R. Cowsik and S. Nussinov, Phys. Lett. {\bf 101B}, 237 (1981);
1339: S. Rao and R. Shrock, Phys. Lett. {\bf 116B}, 238 (1982); Nucl.
1340: Phys. {\bf B232}, 143 (1984).
1341:
1342: \bibitem{GUTS_SUSY} Z. Chacko and R.N. Mohapatra,
1343: Phys. Rev. D {\bf 59}, 055004 (1999);
1344: K. S. Babu and R. N. Mohapatra,
1345: Phys. Lett. B {\bf 518}, 269 (2001).
1346:
1347: \bibitem{GUTS_higherD} Stephan J. Huber and Qaisar Shafi,
1348: Phys. Lett. B {\bf 512}, 365 (2001); S. Nussinov and R. Shrock,
1349: Phys. Rev. Lett {\bf 88}, 171601 (2002).
1350:
1351: \bibitem{ORNL-6910_Alberico} W. M. Alberico, {\it in:}
1352: Proceedings of the Workshop on Future Prospects
1353: of Baryon Instability Search in p-Decay and N-\={N} Oscillation Experiments,
1354: ed. S.J. Ball and Y. A. Kamyshkov, ORNL-6910, March 1996; pp. 221-234.
1355:
1356: \bibitem{Dover_83_85_89} C.B. Dover, A. Gal and J.M. Richard,
1357: Phys. Rev. D {\bf 27}, 1090 (1983); Phys. Rev. C {\bf 31}, 1423 (1985);
1358: Nucl. Instr. Meth. A {\bf 284}, 13 (1989).
1359:
1360: \bibitem{Huefner_98} J. Huefner and B.Z. Kopeliovich,
1361: Mod. Phys. Lett. A {\bf 13}, 2385 (1998).
1362:
1363: \bibitem{Chetyrkin_81} K.G. Chetyrkin, M.V. Kazarnovsky, V.A. Kuzmin, and
1364: M.E. Shaposhnikov, Phys. Lett. {\bf 99B}, 358 (1981).
1365:
1366: \bibitem{Alberico_82_84} W.M. Alberico, A. Bottino and A. Molinari, Phys. Lett.
1367: {\bf 114}B, 266 (1982);
1368: W.M. Alberico, J. Bernabeu, A. Bottino and A. Molinari,
1369: Nucl. Phys. A {\bf 429}, 445 (1984).
1370:
1371: \bibitem{Alberico_91} W.M. Alberico, A. De Pace and M. Pignone, Nucl. Phys.
1372: A {\bf 523}, 488 (1991).
1373:
1374: \bibitem{Kondratyuk_96} L.A. Kondratyuk, JETP Lett. {\bf 64}, 495 (1996).
1375:
1376: \bibitem{Bressi_89} G. Bressi {\em et al.}, Z. Phys. C {\bf 43}, 175 (1989).
1377:
1378: \bibitem{Fidecaro_85} G. Fidecaro {\em et al.}, Phys. Lett. {\bf 156}B,
1379: 122 (1985).
1380:
1381: \bibitem{Baldo-Ceolin_90_94} M. Baldo-Ceolin {\em et al.},
1382: Phys. Lett. B {\bf 236}, 95 (1990); Z. Phys. C {\bf 63}, 409 (1994).
1383:
1384: \bibitem{Homestake_83} Homestake Collaboration, M. L. Cherry {\em et al.},
1385: Phys. Rev. Lett. {\bf 50}, 1354 (1983).
1386:
1387: \bibitem{NUSEX_83} NUSEX Collaboration, G. Battistoni {\em et al.},
1388: Phys. Lett. B {\bf 133}, 454 (1983).
1389:
1390: \bibitem{KOLAR_86} KOLAR Collaboration, M.R. Krishnaswamy {\em et al.},
1391: Nuovo Cim. C {\bf 9}, 167 (1986).
1392:
1393: \bibitem{IMB_84} IMB Collaboration, T.W. Jones {\em et al.}, Phys. Rev. Lett.
1394: {\bf 52}, 720 (1984).
1395:
1396: \bibitem{KAM_86} Kamiokande Collaboration, M. Takita {\em et al.},
1397: Phys. Rev. D {\bf 34}, 902 (1986).
1398:
1399: \bibitem{Frejus_90} Fr\'{e}jus Collaboration, Ch. Berger {\it et al.},
1400: Phys. Lett. B {\bf 240}, 237 (1990).
1401:
1402: \bibitem{S2:NIM_A376_A381} Soudan 2 Collaboration, W.W.M. Allison
1403: {\it et al.}, Nucl. Instr. Meth. A {\bf 376}, 36 (1996);
1404: Nucl. Instr. Meth. A {\bf 381}, 385 (1996).
1405:
1406: \bibitem{S2:NIM_A276} Soudan 2 Collaboration, W.P. Oliver
1407: {\it et al.}, Nucl. Instrum. Methods Phys. Res. A {\bf 276}, 371 (1989).
1408:
1409: \bibitem{Wall_PRD62_00} Soudan 2 Collaboration, D. Wall {\it et al.},
1410: Phys. Rev. D {\bf 62}, 092003 (2000).
1411:
1412: \bibitem{Bodek_81} A. Bodek and J.L. Ritchie, Phys. Rev. D {\bf 23},
1413: 1070 (1981).
1414:
1415: \bibitem{Chaffee_SAGE} R.B. Chaffee, SLAC Computation Group Technical Memo
1416: No. 195, 1979.
1417:
1418: \bibitem{Backenstoss_83} G. Backenstoss {\it et al.}, Nucl. Phys. B {\bf 228},
1419: 424 (1983). A convenient compilation of channel rates based on this work
1420: appears in Yu.A. Golubkov and M.Yu. Khlopov,
1421: astro-ph/0005419; submitted to Yad. Fiz.
1422:
1423: \bibitem{pbar_p_Refs} N. Horwitz {\it et al.}, Phys. Rev. {\bf 115},
1424: 472 (1959); R. Armenteros and B. French, {\it in:}
1425: High Energy Physics, edited by E.H.S. Burhop
1426: (Academic, New York, 1969), Vol. 4, p. 237;
1427: C. Ghesquiere, {\it in:} Proceedings of
1428: the Symposium on Antinucleon-nucleon Interactions,
1429: Liblice-Prague, Edited by L. Montanet, CERN 74-18, (1974), p. 436;
1430: P. Pavlopoulous {\it et al.}, {\it in:}
1431: Nucleon-Nucleon Interactions 1977, Proc. of the Second Int. Conf.,
1432: Vancouver, Edited by H. Fearing, D. Measday, and A. Strathdee
1433: (AIP Conf. Proc. No. 41), (AIP, New York, 1978), p. 340.
1434:
1435: \bibitem{Merenyi_92} R. Merenyi {\it et al.}, Phys. Rev D {\bf 45}, 743 (1992).
1436:
1437: \bibitem{Wall_PRD61_00} Soudan 2 Collaboration, D. Wall {\it et al.},
1438: Phys. Rev. D {\bf 61}, 072004 (2000).
1439:
1440: \bibitem{Allison_98} Soudan 2 Collaboration, W.W.M. Allison {\it et al.},
1441: Phys. Lett. B {\bf 427}, 217 (1998); see Sect. 3.2.
1442:
1443: \bibitem{Agrawal_96} V. Agrawal, T.K. Gaisser, P. Lipari, and T. Stanev,
1444: Phys. Rev. D {\bf 53}, 1313 (1996).
1445:
1446: \bibitem{Gallagher_thesis} H.R. Gallagher, Ph.D. thesis,
1447: University of Minnesota, 1996; H. Gallagher and M. Goodman, Soudan 2
1448: internal note PDK-626 (MINOS note NuMI-L-112), 1995(unpublished).
1449:
1450: \bibitem{Derrick_84} M. Derrick {\it et al.}, Phys. Rev. D {\bf 30},
1451: 1605 (1984).
1452:
1453: \bibitem{SuperK_OSC} Super-Kamiokande Collaboration, Y. Fukuda
1454: {\it et al.}, Phys. Lett. B {\bf 433}, 9 (1998); Phys. Rev. Lett. {\bf 81},
1455: 1562 (1998); Phys. Lett. B {\bf 436}, 33 (1998); Phys. Rev. Lett. {\bf 85},
1456: 3999 (2000).
1457:
1458: \bibitem{MACRO_OSC} MACRO Collaboration, S. Ahlen {\it et al.}, Phys. Lett.
1459: B {\bf 357}, 481 (1995); M. Ambrosio {\it al.}, Phys. Lett. B {\bf 434},
1460: 451 (1998); Phys. Lett. B {\bf 478}, 5(2000); Phys. Lett. B {\bf 517},
1461: 59 (2001).
1462:
1463: \bibitem{Mann_Sudbury} Soudan 2 Collaboration, W.A. Mann,
1464: Nucl. Phys. B (Proc. Suppl.) {\bf 91}, 134 (2001).
1465:
1466: \bibitem{Feldman_Cousins} G. J. Feldman and R. D. Cousins,
1467: Phys. Rev. D {\bf 57}, 3873 (1998).
1468:
1469: \bibitem{PDG_00} Particle Data Group, Eur. Phys. J. C {\bf 15}, 201 (2000).
1470:
1471: %\bibitem{Chung_thesis} J. Chung, Ph.D. thesis, Tufts University, 2001\hfill\\
1472: % (http://hepunx.rl.ac.uk/soudan2/pubs/theses.html).
1473:
1474: \bibitem{Frejus:NIM_A302} Fr\'{e}jus Collaboration, Ch. Berger {\it et al.},
1475: Nucl. Instr. Meth. A {\bf 302}, 406 (1991).
1476:
1477: \bibitem{Kamyshkov_99} Yu. Kamyshkov,
1478: ``Nucleon Instability and ($B-L$) Non-Conservation",
1479: {\it in:} Proc. NNN99 Workshop, AIP Conf. Proc. 533, Edited by M.V. Diwan
1480: and C.K. Jung, Stony Brook, New York, September 1999; p. 84.
1481:
1482: \end{thebibliography}
1483:
1484: \end{document}
1485:
1486: