CNRS Université Bordeaux

Accueil du site > ANGLAIS > Research > Exotic Nuclei > Research topics > Experiments > JYFL (Jyväskylä) > Half-life measurement of 30S - 2009

Half-life measurement of 30S - 2009

last update: may 2014

Precision Half-Life of the Superallowed \beta+ emitter 30S

Date: summer 2009


  1. Centre d’Etudes Nucléaires de Bordeaux Gradignan (France)
  2. Instituto Estructura de la Materia, CSIC (Spain)
  3. Department of Physics, University of Jyväskylä (Finland)
  4. Physics Division, Los Alamos National Laboratory (USA)
  5. Grand Accélérateur National d’Ions Lourds (France)

contact@CENBG: T. Kurtukian-Nieto


  1. P. Ascher, B. Blank, J. Giovinazzo, T. Kurtukian Nieto and J. Souin.
  2. J. Äystö, V.-V. Elomaa, T. Eronen, J. Hakala, A. Jokinen, V. S. Kolhinen, P. Karvonen, I. D. Moore, J. Rissanen, A. Saastamoinen
  3. S. Rahaman
  4. J.-C. Thomas


The experiment was performed at the IGISOL facility at the Accelerator Laboratory of the University of Jyväskylä. The \beta-decay half-life of 30S was measured with a relative precision of 14 in 104. The half-life measurement yields a value of 1175.9 (17) ms which is in agreement with previous measurements but has a precision that is better by a factor of three.

Experimental setup:

The experiment was performed at the Accelerator Laboratory of the University of Jyväskylä. We used the IGISOL technique to prepare samples of 30S. A 35 MeV proton beam impinged on a 2 mg/cm2 Ni2P target with a 2.4 mg/cm2 nickel backing foil. The 30S evaporation residues recoiled out of the target and were stopped in the helium gas of the ion guide. From there, the ions were extracted and accelerated to 30 keV. After mass separation (M/\DeltaM = 500), they were sent to the experimental setup where they were implanted on a tape for the half-life measurement.

The experimental setup used for the half-life measurement was similar to the one used in our recently reported measurement of the half-life of 26Si and 42Ti (

The measurements were structured in cycles. Each half-life measurement started with a 2 s measurement of the background. Then, after a 4 s accumulation period of the A = 30 products, the proton beam was switched off, the IGISOL beam was deflected and a 12 s decay period started. During this period, the decay time distribution was measured by detecting \beta-particles in a close to 4\pi plastic scintillator coupled to two photomultipliers with an electronic coincidence condition and \gamma radiation was detected in three 60-70% co-axial germanium detectors positioned in a horizontal plane at -90°, 0°, and +90° with respect to the beam axis at a distance of 15 cm. After the decay period, the tape was moved and the remaining activity was removed. This protocol of background measurement, accumulation, decay and tape move formed a cycle. Between 40 and 872 cycles with identical experimental conditions formed the different runs. Between the runs, experimental parameters like the photomultiplier high voltages and the trigger threshold were changed to search for an experimental bias of the results. The event trigger was a coincident signal from the two photomultipliers. We used one "cycle-by-cycle" acquisition system to register the time of a decay event as detected by the plastic scintillator with a fixed dead time of 2 \mus yielding data set 1. The time step was 10 ms. A listmode data acquisition allowed the registration of the time of an event and also the energy signal from the photomultipliers and from the germanium detectors. This data acquisition had a fixed dead time of 100 \mus and generated the data set 2.

Analysis procedure and results:

The first step of the analysis was a cycle selection. We have selected all the cycles having a number of counts larger than 50. The accepted cycles were then corrected for the dead time. The next step was a cycle-by-cycle fit. For this fit, the grow-in was excluded and only the constant background at the beginning of a cycle and the 12 s decay were fitted. The fit function was defined to take into account the decay of 30S, of 30P from the decay of 30S, but also the contribution from 30P accumulated as part of the beam from IGISOL. Four free parameters were used: the number of 30S at the beginning of the decay cycle, the half-life of 30S, a constant background, and the amount of 30P at the beginning of the cycle. The half-life of 30P (T1/2 = 149.88 (24) s) was a fixed parameter which was varied during the analysis within its error bars. The effect of this uncertainty will be taken into account for the final result. During the fit, we imposed the condition that the normalized \chi2 has to be two or better in order to accept the cycle.

All in all, 86 cycles out of a total of 6614 were rejected because the fit did not converge, the \chi2 was higher than 2 or the minimum number of counts was not reached. The accepted cycles were further grouped into runs and the cumulated decay spectra were fitted again run-by-run with the same procedure. The time distribution obtained for a single run (full circles). The full line is the result of the fit with contributions from 30S, 30P from accumulation and from the decay of 30S, and the background represented separately. A half-life of 1.1758(47) s was obtained for this run. Figure 1 shows the experimental decay time spectrum decomposed into its different contributions from the decay of 30S, 30P and of the background for one run.

Figure 1: The time distribution obtained for a single run (full circles). The full line is the result of the fit with contributions from 30S, 30P from accumulation and from the decay of 30S, and the background represented separately. A half-life of 1.1758 (47) s was obtained for this run.


The fit results of the two data sets and the associated normalized \chi2 as a function of the run number are presented in Figure 2. The half-lives obtained for the different runs are in mutual agreement and the averaging procedure yields a normalized \chi2 of 1.2 and 1.1 for data sets 1 and 2. The half-lives from the two data sets are 1176.0 (16) ms and 1175.7 (17) ms. The error bars were increased to take the non-unity \chi2 into account. The average of the two values is our final result with its statistical error: T1/2 = 1175.9 (17) ms. This value is the weighted mean of the two data sets and the statistical error is chosen to be the biggest one since the data sets are not independent measurements.

Figure 2: Upper part: experimental half-life as a function of the run number for the two data sets. The error-weighted average values are 1176.0(16) ms and 1175.7(17) ms (we cite here only the statistical error). Lower part: the normalized \chi2 obtained from the fit of the experimental data as a function of the run number for the two data sets.


Half-life measurements should scatter around a central value like a Gaussian distribution. We have investigated this behaviour of our data for the individual cycles and for the runs. The result of this analysis is presented in Figure 3. Both the cycle data and the run data scatter in a statistical manner around the central value. In addition, this procedure yields half-lives of 1176.5(16) ms (2 \mus data) and 1176.0(16) ms (100 \mus data) for the central value of the Gaussian for the cycles, very close to the value determined by averaging the runs. It should be kept in mind that such an analysis does not include the uncertainty of the half-life of the cycles or runs.

Figure 3: The half-life distributions obtained for all accepted cycles (top) and runs (bottom) are compared to a Gaussian distribution. Reasonable agreement is obtained in both cases. The data are from the 2 \mus data set.


Final experimental result for the half-life:

The final result for the half-life of the 30S ground state is obtained by combining the statistical error of 1.7 ms with the systematic uncertainties. We tested the influence of the fixed parameters in the fit function, systematic errors due to experimental conditions, etc. These systematic errors were found rather small, and the final error is again 1.7 ms. Therefore, our final experimental result is 1175.9 (17) ms.


We have performed a high-precision measurement of the half-life of 30S. The nuclei of interest were produced at IGISOL in fusion-evaporation reactions and were mass separated. To measure the half-life, the mass A=30 beam was implanted in the center of a high-efficiency \beta plastic scintillator, read out by two photomultipliers. Germanium detectors were used to search for contaminants which could influence the experimental results. No unexpected activity was observed.

The result of T1/2 = 1175.9 (17) ms obtained in this work is in agreement with older half-life values from the literature. The present result is a factor of 3 more precise than the previous experimental average. With the half-life precision of 14 parts in 104 and the f value precision of 2 parts in 104, 30S could contribute to testing the CVC hypothesis as soon as the super-allowed branching ratio is known to similar precision.

This work was published at J. Souin et al., Eur. Phys. J. A (2011) 47: 40, DOI 10.1140/epja/i2011-11040-5