CH-5232 Villigen PSI, Switzerland
Search for a Neutral Particle of Mass 33.9 MeV in Pion Decay
M. Daum1, M. Janousch2,
P.-R. Kettle1,
J. Koglin1,3,
D. Pocanic3, J. Schottmüller1,
C. Wigger1, Z. G. Zhao4
1PSI, Paul-Scherrer-Institut,
CH-5232 Villigen-PSI, Switzerland
2IPP, Institut für
Teilchenphysik der ETHZ, CH-5232 Villigen-PSI, Switzerland
3
Physics Department, University of
Virginia, Charlottesville, Virginia 22901,USA.
4Institute of High
Energy Physics, Chinese Academy of Science, Beijing 100039, The People's
Republic of China
Abstract
We have searched for the rare pion-decay p+
® m+ X where X is a neutral
particle of mass 33.9 MeV, a process suggested by the KARMEN Collaboration
[1] to explain an anomaly seen in their time spectrum of neutrino induced
reactions. By measuring the muon momentum spectrum of charged pions decaying
in flight using the pE1 high intensity pion
channel at PSI, we were able to place an upper limit on the branching fraction
h
of such decays, at the level of h < 6.0·10-10
with a 95% confidence level.
Motivation
In 1995, the KARMEN
Collaboration originally reported [1] on a `long-standing discrepancy'
between their measured and expected time distributions of neutrino induced
reactions originating from neutrinos produced from p+
and m+ -decays
at rest in the primary target of the ISIS facility at RAL. Further evidence
of an excess of events over the expected exponential time distribution
characterized by the muon lifetime, and clustered around 3.6 ms
after beam-on-target has since been reported [2,3]. The speculative
explanation given in ref. [1] was that the anomalous events could originate
from the rare pion decay process,
p+ ® m+
X
where X is a heavy neutral particle with a rest mass of
mX= 33.905 MeV.
The mass of the hypothetical particle is very close to the mass difference
between the charged pion and the muon,
mp - mm
= 33.91157 ± 0.00067 MeV,
and thus, the Q-value of this reaction is very small which can be used
to ones advantage in a decay-in-flight experiment, but is prohibitive to
a decay at rest experiment. Soon after the first KARMEN anamoly publication,
we conducted a search for the hypothetical decay using the pE1
high intensity pion channel at PSI. We set an upper limit on the
on the branching fraction h of such decays at
the level of h < 2.6·10-8
with a 95% confidence level [4]. In 1997, we began a new series of
experiments using a similar method but with a completely new setup.
Experimental Implications
In a decay-in-flight experiment using 150 MeV/c pions, the emission angles
of muons from the hypothetical decay with respect to the parent pions in
the laboratory system are less then 20 mrad; whereas, over ther momentum
range of interest (100-125 MeV/c), the emission angles of muons from the
normal pion decay mode, p+
® m+ nm,
are greater then 250 mrad, and the radiative pion decay mode, p+
® m+ nm
g, is highly suppressed, especially for small decay angles [c.f.
Fig.
1]. Also, the velocity of the muon is very close to the velocity
of the original pion which means that the specific energy loss of the muon
is amost the same as that of its parent pion. Thus, the beam-line
itself can be used as a spectormeter to separate the muons from the hypothetical
and normal pion decay modes, and the electronics timing and counter thresholds
can be setup for muons using beam pions.
Experimental Setup
The experimental setup consists of three parts: the pion transport
region which defines the pion beam phase space; the 4.5 m long field-free
pion decay region; and, the momentum analysis spectrometer. An overhead
photograph of the pE1 area is shown in Fig.
2. and a schematic of the entire setup is shown in Fig.
3. The triple coincidence of the S1, S2, and S3 beam counters
was used to define our trigger events, and time-of-flight (TOF) information
was used for particle identification. A muon telescope in the decay
region and a proton monitor in the primary proton channel were used for
normalization purposes. Active and passive collimators were placed
in the vacuum system to suppress background from scattered particles.
Experimental Procedure
The entire beamline was optimized using 150 MeV/c pions. 'X-particle
scans' were performed by varrying the spectrometer magnets in steps of
0.5 MeV/c around 113 MeV/c, the expected momentum for muons from the hypothetical
decay. At each momentum setting, pulse-height and timing information
for the beam and active veto counters were recorded on an event-by-event
basis for a given amount of integrated proton charge. The shape and
normalization of the peak which would be produced from the hypothetical
decay were determined from computer simulations. 'Pion scans' performed
around 150 MeV/c and'muon scans' with the initial beamline scaled to 113
MeV/c and the spectrometer magnets scanned around this same momentum were
used to calibrate the simulations.
To improve on the limit set in 1995, we made initial beam tests in 1997
on a new experiment designed to have better beam definition and background
suppression. After a detailed analysis of the 1997 data and
further simulation studies, we concluded that a significant fraction of
our background events could be attributed to scattered muons originating
from pions decaying through the normal decay mode within the pion decay
region and also in the first dipole magnet of the spectrometer (ASL52)
which serves to separate out the beam pions from the decay muons.
In the fall of 1998, we conducted a full experiment using the same basic
1997 setup with a widened ASL52 pole gap and the addition of several strategically
placed active veto counters. After optimising our setup, we
were able to significantly reduce the background with much of the remaining
background thought to originate from muons decaying by through the normal
pion decay mode within the first bending magnet of the spectrometer.
In the spring of 1999, we used the full setup shown in Fig.
3, where the ASL54 dipole magnet was added to the 1997-98 setup to
improve our spectrometer momentum resolution. We succeded in reducing
our background to a level near that which is expected from radiative pion
decay, p+ ®
m+ nm
g. The results from the 1997, 1998, and 1999 measurements are compared
in Fig. 4, along with the expected level of
radiative background determined from simulation.
Data Analysis
From TOF information, one can clearly distinguish beam pions, muons, and
electrons in the calibration 'pion runs' [cf., Fig.5(a)].
X-candidate muon events can be cleanly isolated from scattered beam particles
since they have the same timing as beam pions [compare the timing cut windows
in Figs. 5(a) and 5(b)]. In addition
to timing cuts, veto cuts were applied to the data for those events with
assosiated active veto counter triggers, thus rejecting over 90% of the
candidate events.
The final momentum spectrum of X-candidate muon events is obtained by
summing the events within the timing cuts at each momentum and normalizing
them appropriately. Each scan was fitted with a gaussian funtion
simulating the expected distribution for muons from the hypothetical decay
together with a hyperbolic function with five degrees of freedom simulating
the background. For normalization purposes, the data was divided into four
data sets based on modifications we made to our setup during our beam-time.
To illustrate the sensitivity of the experiment, the combined data the
nine scans in Data Set E are shown in Fig. 6.
Results
The fit results of each individual scan are shown in Fig.
7. . No indication for the existance for the hypothetical decay is evident
in our data. The weighted mean of the 28 scan measurements results in a
combined branching fraction of
h = (1.3 ± 2.3) · 10-10,
with an estimated overall systematic uncertainty of 5%. Using the Feldman-Cousins
frequentest approach, we impose the following upper limit on the branching
fraction of this hypothetical decay process:
h < 6.0 · 10-10
(95% c.l.),
Conclusions
The correlation between the production branching fraction h
and the mean lifetime t of the X-particle required
to explain the KARMEN anomaly is shown in Fig.
8. A suppersymmetric solution proposed by Choudhury and Sarkar
in 1996 [5] assumes X is a light photino or zino which decays via the two-body
decay mode X ® m g. This solution was reported
to be consistent with the KARMEN hypothesis only for the region emphasized
in bold red in Fig. 8, and it can now be ruled out by our new upper limit.
Other theoretical solutions were X is some sort of issosinglet neutrino
(a sterile neutrino) have been reported to be consistent within different
regions of this plot for h > 10-13
[6, 7].
References
[1] KARMEN Collaboration, B. Armbruster et. al., Phys. Lett. B
348 (1995) 19-28.
[2] KARMEN Collaboration, B. Seligmann, Proc. Int. Europhysics
Conf. on High Energy Physics, Brussels, 27 July-2 Aug. 1995, eds. J. Lemonne,
C. Vander Velde and F. Verbeure (World Scientific, Singapur, 1996) 526-527.
[3] Ch. Oehler, Dissertation, Universität Karlsruhe, April
1999
[4] M. Daum et. al., Phys. Lett. B 361 (1995) 179-183.
[5] D. Choudary and S. Sarkar,
Phys. Lett B 374 (1996) 87-92.
[6] V. Barger et. al., Phys.
Lett. B 352 (1995) 365-371.
V.
Barger et. al., Phys. Lett. B 365 (1995) 617.
[7] J. Govaerts et. al., Phys.
Lett. B 389 (1996) 700.
Created by J. Koglin.
Last updated on 17th December 2001.