Atomic Structure Calculations with Spectroscopic Accuracy : Implications for Laboratory Work

The observation of atomic spectra constitutes an important tool for diagnostics of astrophysical plasmas, and there is a boom of activity involving several new and planned multibillion-dollar telescopes. However, to correctly interpret observed spectra, the atomic lines must be known and identified from laboratory work. Laboratory work is hard and time-consuming, and present efforts do not in any way match the needs for data, partly due to lack of funding [1] and partly due to experimental limitations. One goal of atomic structure calculations is to provide energy differences with ”spectroscopic accuracy” to aid laboratory work. Using highly accurate calculated energy differences it should be possible to directly validate or rule out experimental energy level and line identifications. New and efficient methods for solving the Dirac-equation for many electron systems, together with today’s fast computers, indeed make it possible to perform calculations with spectroscopic accuracy for ions of medium complexity. We give a number of examples of calculations based on the relativistic configuration interaction (RCI) method in B-, C-, N-, O-, and Ne-like systems, where energies levels far up in the spectrum have been predicted with uncertainties of 0.05 % or less [2,3,4]. Depending on the spectral range, these uncertainties are in many cases close to what can be experimentally obtained. The above mentioned calculations reveal that many experimental energy levels given in the literature and in data bases are wrong and based on misidentifications. We finally show how the accuracy of atomic structure calculations can be further improved, and results extended to more complex systems, by using the novel partitioned configuration function interaction (PCFI) method [5]. Some practical consequences of the recent advances in computational methodology for laboratory work are discussed