A New Measurement of the Positron Fractionat High

We present measurements of the cosmic ray positron fraction from 4.5 { 50 GeV obtained from the rst balloon ight of the High Energy Antimatter Telescope (HEAT). A superconducting magnet spectrometer combined with a transition radiation detector, an electromagnetic calorimeter, and time-of-ight counters were used to achieve powerful hadron rejection at these energies. We see no evidence for the signiicant rise in the positron fraction at energies above 10 GeV reported by previous experiments. Cosmic ray electrons may either be produced in primary acceleration sites or can be secondaries from nuclear interactions of cosmic rays in interstellar space. The secondary component is characterized by about equal proportions of e + and e ?. In the 1 to 10 GeV range, the measured positron fraction is < 10%%1, 2] which implies a dominant primary e ? component in this region. Above 10 GeV an unexpected rise in the positron fraction was observedd3, 4, 5, 6, 7] and has been attributed to a depletion of the primary electron source at high energy or to the appearance of new sources of e pairs. The present experiment was designed to clarify these issues. The HEAT-e instrument, shown in Fig. 1, employs multiple techniques for electron identiication and hadron background rejection. The magnet spectrometer, consisting of a split-coil, superconducting magnet and a precision drift tube hodoscope (DTH) measures the particle rigidity and charge sign. Particle identiication is accomplished with a transition radiation detector (TRD) and an electromagnetic calorimeter (EC). The time-of-ight (TOF) counter in conjunction with the top three layers of the EC is used to measure the charge, speed, and direction of the particle. Details about these detectors and their performance are given elsewhere in these proceedingss8]. HEAT was own on May 3{5, 1994 from Ft. Sumner, New Mexico. The data presented here were recorded in a 29.5 hour period at oat altitudes of 3.8{7.4 g/cm 2. A clean sample of e events is obtained by a succession of selections. The analysis is helped greatly by the fact that diierent components of the instrumentation provide mutually independent electron identiication and background rejection. First, events with a well deened trajectory, measured consistently in both the TRD and the DTH, and with resolved momentum, are selected, and the various counter signals are corrected for position-and time-dependent eeects. Fig. 2 shows a histogram of