Abstract. The rate coefficient for radiative and dielectronic recombination of beryllium-like magnesium ions was measured with high resolution at the Heidelberg heavy-ion storage ring TSR. In the electron-ion collision energy range 0-207 eV resonances due to 2s → 2p (∆N = 0) and 2s → 3l (∆N = 1) core excitations were detected. At low energies below 0.15 eV the recombination rate coefficient is dominated by strong 1s 2 (2s 2p 3 P) 7l resonances with the strongest one occuring at an energy of only 21 meV. These resonances decisively influence the Mg  recombination rate coefficient in a low temperature plasma. The experimentally derived Mg  dielectronic recombination rate coefficient (±15% systematical uncertainty) is compared with the recommendation by Mazzotta et al. (1998, A&AS, 133, 403) and the recent calculations by Gu (2003, ApJ, 590, 1131 and by Colgan et al. (2003, A&A, 412, 597). These results deviate from the experimental rate coefficient by 130%, 82% and 25%, respectively, at the temperature where the fractional abundance of Mg  is expected to peak in a photoionized plasma. At this temperature a theoretical uncertainty in the 1s 2 (2s 2p 3 P) 7l resonance positions of only 100 meV would translate into an uncertainty of the plasma rate coefficient of almost a factor 3. This finding emphasizes that an accurate theoretical calculation of the Mg  recombination rate coefficient from first principles is challenging.

Absolute total electron-ion recombination rate coefficients of argonlike Sc 3+ (3s 2 3p 6 ) ions have been measured for relative energies between electrons and ions ranging from 0 to 45 eV. This energy range comprises all dielectronic recombination resonances attached to 3p→3d and 3p→4s excitations. A broad resonance with an experimental width of 0.89 ± 0.07 eV due to the 3p 5 3d 2 2 F intermediate state is found at 12.31 ± 0.03 eV with a small experimental evidence for an asymmetric line shape. From R-Matrix and perturbative calculations we infer that the asymmetric line shape may not only be due to quantum mechanical interference between direct and resonant recombination channels as predicted by Gorczyca et al. [Phys. Rev. A 56, 4742 (1997)], but may partly also be due to the interaction with an adjacent overlapping DR resonance of the same symmetry. The overall agreement between theory and experiment is poor. Differences between our experimental and our theoretical resonance positions are as large as 1.4 eV. This illustrates the difficulty to accurately describe the structure of an atomic system with an open 3d-shell with state-of-the-art theoretical methods. Furthermore, we find that a relativistic theoretical treatment of the system under study is mandatory since the existence of experimentally observed strong 3p 5 3d 2 2D and 3p 5 3d4s 2 D resonances can only be explained when calculations beyond LS-coupling are carried out.

We have measured resonance strengths and energies for dielectronic recombination (DR) of beryllium-like Fe xxiii forming boron-like Fe xxii via N ¼ 2 ! N 0 ¼ 2 and N ¼ 2 ! N 0 ¼ 3 core excitations. All measurements were carried out using the heavy-ion Test Storage Ring at the Max Planck Institute for Nuclear Physics (MPI-K) in Heidelberg, Germany. We have also calculated these resonance strengths and energies using three independent, perturbative, state-of-the-art theoretical techniques: the multiconfiguration Breit-Pauli (MCBP) method, the multiconfiguration Dirac-Fock (MCDF ) method, and the Flexible Atomic Code (FAC). Overall reasonable agreement is found between our experimental results and these theoretical calculations. We have used our measurements to produce a Maxwellian-averaged DR rate coefficient for Fe xxiii. Our experimentally derived rate coefficient is estimated to be accurate to better that %20%. At temperatures where Fe xxiii is predicted to form in both photoionized and electron-ionized gas, we find mixed agreement between our experimental rate coefficient and previously published rate coefficients. We find good agreement at these temperatures between the experimentally derived rate coefficient and our MCBP, MCDF, and FAC results.

Recombination involving the core excitation of two electrons, which may be termed trielectronic recombination, has been experimentally identified for the first time. Using Cl13+ ions circulating in the TSR heavy-ion storage ring, we have observed surprisingly strong low-energy trielectronic recombination resonances, comparable to the dielectronic process. At higher electron-ion collision energies, trielectronic recombination is suppressed due to the autoionization of the triply excited intermediate state into excited final states. The formation of the intermediate state depends sensitively on configuration mixing, making trielectronic recombination a challenge to atomic-structure calculations.

Be-like ions have been investigated employing the resonant electron–ion collision process of dielectronic recombination (DR) as a spectroscopic tool. The experiments were performed at the experimental storage ring in Darmstadt, Germany, using its electron cooler as a target for free electrons. DR Rydberg resonance series for the associated intra-L-shell transitions and were observed with high resolution. In addition to these excitations from the ground state we determined resonances associated with excitations of ions initially in the metastable state. The corresponding excitation energies were determined to be and and . These excitation energies are compared with previous measurements and with recent state-of-the-art atomic structure calculations.