First Observations of Separated Atmospheric ?? and ??

1
_
First Observations of Separated Atmospheric ??
and ?? Events in the MINOS Detector.
A. S. T. Blake (for the MINOS collaboration) Ca
vendish Laboratory, University of Cambridge, J.J.
Thomson Avenue, Cambridge, CB3 0HE, United
Kingdom.
Overview
Studying Atmospheric Neutrinos at MINOS
Atmospheric neutrinos are incident on the MINOS
detector from all directions. The signature for
????? oscillations is a deficit in the observed
flux of up-going ?? events, since these events
correspond to large neutrino path lengths and
therefore large oscillation probabilities. In
order to perform an accurate measurement of these
oscillations, a clean sample of ?? CC events must
be selected. The signature used to identify these
events is a muon track whose interaction vertex
is contained inside the fiducial volume of the
detector. space
The complete MINOS far detector has been
collecting atmospheric neutrino data since August
2003. The detector is a 5.4 kT steel-scintillator
calorimeter located at a depth of 700m (2100 mwe)
in the Soudan Underground Laboratory, Minnesota.
It is the first massive underground detector to
possess a magnetic field, making the separation
of atmospheric ?? and ?? possible for the first
time. This poster presents the first MINOS
observations of separated ?? and ?? charged
current neutrino interactions based on a detector
exposure of 418 days (6.18 kT-Yrs). The ratio of
?? to ?? is calculated and compared to the Monte
Carlo expectation. An extended maximum likelihood
analysis of the observed L/E distribution is used
to fit the neutrino oscillation parameters ?m232
and sin22?23.
High energy cosmic ray interactions at the top of
the atmosphere produce an intense flux of
neutrinos that can be detected on the earth. In
recent years, the deficit observed in the
atmospheric ?? flux has become firmly
established, with the favoured interpretation
being ????? oscillations.
The MINOS far detector, located in the Soudan
mine, is a sampling calorimeter composed of many
planes of steel and plastic scintillator. The
detector is magnetized by a current-carrying coil
running through its centre. This allows the
charge of muons produced in neutrino interactions
to be identified, enabling the separation of
neutrinos and anti-neutrinos. A scintillator veto
shield is constructed above the main detector to
tag cosmic muons entering the detector. space
oscillations. The MINOS far detector can be used
to study atmospheric neutrino oscillations. Its
deep location 700m underground provides shielding
against the high flux of cosmic rays incident on
the surface of the earth, and its large 5.4 kT
mass ensures that a high rate of atmospheric
neutrinos can be observed in the detector. space
_
_
The contained vertex ?? CC events are divided
into three classes
_

• Fully Contained Events (FC).
• Downward-Going Partially
• Contained Events (PCDN).
• Upward-Going Partially
• Contained Events (PCUP).

The MINOS Far Detector
Selecting Atmospheric Neutrinos
Results of Event Selection
Oscillation Analysis
The selection of contained atmospheric ?? events
is hampered by the large background of cosmic
muon events appearing contained or up-going. This
background must be reduced by a factor of 106 in
order to obtain a clean neutrino signal. space
Events are reconstructed using software specially
designed for this analysis. A pre-selection is
applied requiring a clean muon track. The events
are then divided into the FC/PCDN and PCUP
classes and separate selection criteria are
applied.
From 418 days detector exposure, a total of 107
candidate events are selected from the data. This
compares with an expectation of 12713 events
assuming no oscillations, and 9610 events
assuming ?m2320.0024eV2 and sin22?231.0. The
background contribution from cosmic muons is
measured to be 4.40.5 events using the shield.
space
A total of 77 events are found to have a
well-measured direction, with 49 down-going and
28 up-going events. The up-down double-ratio is
This is 2 standard deviations away from
unity.
An extended maximum likelihood analysis of the
observed L/E distribution is used to fit the
parameters ?m232 and sin22?23 assuming
two-flavour ????? oscillations. The sensitivity
of the L/E distribution to neutrino oscillations
depends on the L/E resolution. Therefore a
Bayesian method is used to calculate a log(L/E)
probability density function for each event and
the resolution ?log(L/E) of each event is given
by the rms of this distribution. space

1. Fully Contained or Down-Going Partially
   Contained Events.

The planar structure of the detector allows
cosmic muons to enter the detector between planes
and appear contained. These events tend to travel
nearly parallel to the planes or to contain large
energy deposits at the track vertex. A series of
topology cuts are applied to the FC and PCDN
events to separate the signal and background.
space
The following topology cuts are applied

• The reconstructed track is first extrapolated
  back to
• the outside of the detector and a cut placed on
  the
• distance traversed perpendicular to the planes.
• For events travelling parallel to the planes a
  cut is
• placed on the energy deposited at the track
  vertex.
• The background is tagged using the veto shield.

above the reconstructed log(L/E) resolution
binned in regions of log(L/E) resolution (?).
right confidence limits calculated for the
oscillation parameters ?m232 and sin22?23.
The data is binned in terms of log(L/E) and
?log(L/E) and a likelihood fit to the
oscillations is performed. The maximum likelihood
occurs at
The 68 and 90 confidence limits are determined
using the Feldman-Cousins method. The data are
consistent with a wide range of oscillation
parameters, but disfavour the null oscillation
hypothesis at the 98 confidence level.
?m232 0.0013 eV2 , sin22?23 0.90
above The reconstructed zenith angle
distribution for the 77 events with
well-measured direction, compared with the MC
expectation.
above The reconstructed log10L/E distribution
for the 77 events with well-measured
direction, compared with the MC expectation.
Charge Ratio
Separating Neutrinos and Anti-Neutrinos
cosmic muon background
_
left illustration of cosmic muon background
event middle distribution of perpendicular
distance between track vertex and edge of
detector right energy deposited at track vertex
vs. reconstructed direction for MC signal and
background events, showing the cut regions.
The MINOS detector is magnetized with a mean
field of 1.3T. This allows the charge sign of
muons to be determined from the curvature of
their tracks. A track fitting algorithm is used
to determine the best fit value of Q/p and its
error ?Q/p, where Q is the muon charge and p is
the muon momentum. An event is classified as
having a well-measured charge sign if
(Q/p)/(?Q/p) lt 50. A total of 52 events are
found to have a well-measured charge, with 34 ??
and 18 ?? events. The ??-?? ratio is
The ratio of ?? to ?? events in the data compared
to the MC expectation (which assumes the same
oscillation parameters for both neutrinos and
anti-neutrinos) is calculated to be Although
the current statistics are limited, the data are
found to be consistent with the hypothesis that
neutrinos and anti-neutrinos oscillate with the
same parameters.
(b) Up-Going Partially Contained Events.
The dominant background in the PCUP selection
arises from cosmic muons reconstructed as
up-going rather than down-going. Up-going events
are identified using timing information. After
careful calibration a single-hit timing
resolution of 2.3 ns is achieved. Selection cuts
are applied to the timing information to ensure
that the events are up-going. space
The event direction is determined by applying
fits to the measured times along the muon track,
assuming that the muon is travelling upwards and
then downwards at the speed of light. The RMS of
each fit is calculated, and a cut is placed on
the quantities RMSup, RMSdown, and RMSup
RMSdown to select a clean sample of up-going
events.
_
_
above the up-down distributions of selected
neutrino and anti-neutrino events, showing the MC
expectations for three different neutrino
oscillation scenarios.
above the distribution of (Q/p)/?(Q/p) for
selected neutrino events. The two broad peaks
correspond to neutrinos and anti-neutrinos.
Example Events
_
_
Summary
FC up-going ??
PC down-going ??
PC up-going ??
_
The MINOS detector has been taking atmospheric
neutrino data since August 2003, and an analysis
of 418 days (6.18 kT-Yrs) data is presented. A
total of 107 atmospheric neutrino events are
observed compared with an expectation of 12713
events. The background is estimated to be
4.40.5 events. space
The ?? and ?? events are separated using the
curvature of muons in the MINOS magnetic field.
From the events containing a well-measured muon,
a total of 34 ?? and 18 ?? events are observed.
This represents the first observation of
separated atmospheric neutrino and anti-neutrino
interactions.
_
reference Phys. Rev. D73, 072002 (2006)
2
_
First Observations of Separated Atmospheric ??
and ?? Events in the MINOS Detector.
A. S. T. Blake (for the MINOS collaboration) Ca
vendish Laboratory, University of Cambridge, J.J.
Thomson Avenue, Cambridge, CB3 0HE, United
Kingdom.