PHYSICAL REVIEW A 85, 042508 (2012) Heavy-ion storage-ring-lifetime measurement of metastable levels in the C-, N-, and O-like ions of Si, P, and S E. Tr¨ abert Astronomisches Institut, Ruhr-Universit¨ at Bochum, D-44780 Bochum, Germany M. Grieser, J. Hoffmann, C. Krantz, R. Repnow, and A. Wolf Max-Planck-Institut f¨ ur Kernphysik, D-69117 Heidelberg, Germany (Received 8 March 2012; published 6 April 2012) In a quest for benchmarking transition rate data on electric-dipole (E1) forbidden transitions to be used in collisional-radiative models of plasma spectra, the radiative decay rates of the metastable levels 2s 2 2p 21 D 2 , 2s 2 2p 32 P o 1/2,3/2 , and 2s 2 2p 41 D 2 in C-, N-, and O-like ions, respectively, have been measured for the elements Si, P, and S. Results with uncertainties mostly well below 1% are obtained using ions circulating in a storage ring. Precision results for Si 7+ ,P 7+ ,P 8+ ,P 9+ ,S 8+ , and S 10+ obtained here complete the isoelectronic sequence data sets for these four metastable levels in all three elements. DOI: 10.1103/PhysRevA.85.042508 PACS number(s): 32.70.Cs, 32.30.Jc, 34.50.Fa I. INTRODUCTION The ground configurations of many-electron atomic sys- tems, beginning with the doublet ground term of B-like (five-electron) ions, harbor a number of levels. The transition rates between these levels are low, because the transition energies are low and because the usually dominant electric dipole (E1) transitions are ruled out by the same parity of initial and final states. Hence magnetic dipole (M1) and electric quadrupole (E2) transitions dominate. Representing the very first excited levels of a given ion, such transitions have long been used as indicators that a given ionization stage of a given elemental species is present in a plasma [1,2], and thus the plasma (electron) temperature must be sufficiently high and can be estimated. However, the optical spectra carry more information [3,4]. In particular in a dilute plasma, the collision frequency competes with the radiative decay rate, and then the relative population of levels within the ground configuration (notably fine-structure levels of the ground term, for example, in C- and O-like ions) may depend on the electron density. On the other hand, if the first-level interval is relatively large (as, for example, in N-like ions), the level populations of the upper levels of the ground configuration compared to that of the true ground level depend significantly on the plasma temperature. One approach in plasma diagnostics measures the line ratios of lines that depend on the level populations of low-lying levels. For this procedure, collisional-radiative models have been developed that include thousands of levels and tens of thousands of transitions between mostly high-lying levels; such models are used to provide synthetic spectra and to learn about their dependence on density and temperature. Most of the atomic data used in such models are obtained by calculation only, since experiment cannot practically provide the vast amount of information. However, the excitation energies of low-lying and resonance levels are usually better obtained from spectroscopic experiment, and the calculated energy scale is regularly corrected for experimental data. Also, a number of E1 transition rates are available from beam-foil spectroscopy, in most cases at moderate precision (level lifetimes in the range from a few picoseconds to many nanoseconds, typically measured with a precision of 5% to 10%). The measurement of the much-longer-lived levels in the ground configurations (lifetimes in the millisecond to second range, for low to moderate charge states) requires ion trapping techniques. In about the last decade, experiments at heavy-ion storage rings and at electron beam ion traps have provided such lifetime data [5–8] with accuracies of usually between 3% and 0.3% (and in some cases even better than that, see examples in Ref. [9]). In the early measurements, priority was given to specific cases of particular interest (plasma diagnostics, astro- physics); by now the data extend over sections of isoelectronic ranges and begin to permit one to look into further questions: With the growing experience and confidence in highly accurate measurements, does the systematic error scenario need to be reassessed? Do measurements on several ion species reveal new problems that may have been overlooked initially; does theory hold up with the best experiments? Do data on isoelectronic sequences perhaps reveal shortcomings in individual measurements or calculations? What obstacles can be identified on the way to testing the accuracy limits of calculations of transition rates, recognizing that two of the most accurate lifetime measurements [10–12] disagree with the results of extensive calculations at a level larger than the QED correction to the magnetic-dipole (M1) transition probability [10,13]? Not all of these questions can be answered in a single experiment. Ions with an open L shell may be more conducive to high accuracy studies, since recent work on M shell ions— though using the very same experimental techniques—has found complications due to the level structure that appear to limit the accuracy of lifetime measurements there [9,14]. About a decade ago we have measured the radiative lifetimes of the 2s 2 2p 21 D 2 level in the C-like ion Si 8+ , of the 2s 2 2p 3 2 P o 1/2,3/2 level in the N-like ion S 9+ , and of the 2s 2 2p 41 D 2 level in the O-like ion Si 6+ [15,16], with an uncertainty of about 1%. The level structure and the transitions of interest are sketched in Fig. 1. We now have extended these heavy-ion storage-ring measurements to the same levels in the isoelectronic ions of Si, P, and S, with even better statistical reliability of the data and with the consistency check provided by isoelectronic data 042508-1 1050-2947/2012/85(4)/042508(8) ©2012 American Physical Society