Precise Half-Life Measurement of Ga

In last year’s Progress Report we reported on measurements of the β-decay of Ga, including a preliminary value for its halflife [1]. We report here on an additional measurement of the half-life focused on exposing systematic effects that could affect the result. We obtained Ga from the MARS recoil spectrometer by bombarding a cooled hydrogen gas target with Zn at 42A MeV. Special care was taken for the optimization of the Ga beam from MARS in order to minimize the impurities. At the exit of MARS, the beam was degraded and then implanted in the aluminized mylar tape of our fast tape-transport system. The short halflife of Ga (t1/2 = 116 ms) made it impractical to use our normal reel-to-reel tape system; instead, we employed a closed single-loop path design, similar to that described in [1], which removed the need for frequent tape rewinds. The Ga was collected on tape for 0.25 s; then the beam was turned off and the collected activity was moved within 135 ms in the center of a 4π proportional gas counter placed in a low background region. The counter signals were then multiscaled for 3.00 s and a decay spectrum consisting of 250 channels was recorded. About 10 separate decay spectra, one from each collect/move/detect cycle, were recorded. To allow for tests of possible systematic errors, the data were split into separate runs typically of over a thousand cycles each. The detector bias, discriminator threshold, time interval per channel and the setting for the dominant deadtime were changed sequentially from run to run. In analysis, each run underwent a presorting procedure that removed the cycles in which less than 25 decay events were recorded in the 4π detector, and the cycles for which the ratio of the recorded β-decays versus the number of deposited nuclei was anomalous, indicating either high noise (too many events) or bad tape positioning (too few events). Runs passing the pre-sort procedure with less than 50 valid cycles were nevertheless removed entirely as not having statistical significance. About 3 million events, in total, survived this procedure. The spectrum for the data subset with a timing of 12 ms per channel is presented in Fig. 1. To achieve a high-precision half-life measurement we require a very good knowledge