| Recent reports of the possible existence of an Axion with mass = 9.5 MeV and lifetime less than 10−14 sec do not contradict any negative results of the 20-year long search. The present work aims at confirming or disproving these reports. An Axion may compete with M1 nuclear transitions and decay into a e+-e− pair, producing events with high angle separation, thus being detectable in the presence of internal pair conversion which favors small separation angles. In the present experiment the M1 transitions from two discrete states in 12C were produced using a (d,n) reaction. A hermetic array of plastic scintillator detectors for e+-e− pairs from nuclear transitions was upgraded to 65 elements covering 50% of 4π. A target chamber made of carbon fiber/epoxy resin, with wall thickness 0.8 mm, was introduced which absorbs only 172. 2 keV of the kinetic energy of minimum ionizing e+/e−. A neutron detector with total efficiency of 3% was constructed to measure the time of flight of neutrons. The detectors and chamber were installed on the beam line of the Stony Brook heavy ion LINAC. A test run was conducted using the reaction 11B(p,e+e−)12C (Ep = 7.2 MeV) to populate the Giant Dipole Resonance of 12C. The observation of the IPC from the 22.6-MeV E1 transition to the ground state of 12C established the pair-energy line-shape and produced an absolute pairenergy calibration. The angular correlation distribution of the pairs was found to be in agreement with the Born and point nucleus approximation of E1 angular correlations. A data run with the stripping reaction 11B(d,n)12C* (Ed = 7.2 MeV) populated the (Iπ,T) = (1+,1) 15.11-MeV and the (1 +,0) 12.7-MeV states of 12C. Detected pair events without neutron coincidence required showed a clear and strong peak of the 15.11 MeV to ground state transition. Analysis of these data agreed with angular correlations of M1 internal pair conversion. These did not support the earlier work and showed no evidence of an Axion emitted in the transition. A second test with neutron coincidence required, demonstrated that the full experiment is working as designed. However, for the kinematically complete experiment including neutron coincidence the rate of good events was found to be only 4–6 events/hr. A statistically decisive test will take data runs of about six months. |
