Revised and extended analysis of five times ionized argon (Ar VI)

The spectrum of five times ionized argon, (Ar VI), has been observed in the 280-2100 Å wavelength range. Eighty-seven lines have been identified as transitions between levels of the 3s23p, 3s3p2, 3s23d, 3p3, 3s3p3d, 3s24s, 3s24d and 3s3p4s configurations. For 33 of the lines the classification is new. Forty-one energy level values belonging to these configurations were analyzed and we propose 9 new energy level values for levels corresponding to odd parity configurations. The configurations are interpreted by fitting the theoretical energy expressions to the observed energy levels using leastsquares techniques. The parameter values are compared with results from Hartree-Fock calculations.


Introduction
The ground-state configuration of five times ionized argon, (A?'), is 3s23p with the term 'P. Ar VI belongs to the A1 I isoelectronic sequence. Excited states either belong to simple one-electron configuration of the type 3s'nl or to threeelectron configurations such as 3s3p2, 3p3 and 3~3p('*~P)nl etc., giving both doublets and quartets.
The spectra of the first, second and third elements in this sequence are presented in Atomic Energy Levels (AEL), Ref.
[l]. Subsequent to this tabulation, the A1 I spectrum was investigated by Eriksson and Isberg [2]. Results about Si I1 were published by Shenstone [3] and the P I11 spectrum was studied by Magnusson [20] who have studied all argon spectra from Ar I to Ar IX. Some anomalies in resonance transitions in the A1 I isoelectronic sequence were observed by Engstrom et al. [21] and a new work about energy levels and lifetimes of Ar VI was recently published by Pinnington et al. [22]. An extended analysis of spectra and term systems in Al-like Ca VIII to Ni XVI was published by Redfors and Litztn [23] and lifetimes of the 3s24s2S states for Al-like ions from S IV to Fe XIV were published by Thornbury et al. C24). Transitions in spectra of highly ionized Kr and MO belonging to the A1 I isoelectronic sequence were recently reported by Jupkn et al. [25]. In the present work we report a revised and extended analysis of Ar VI that includes 33 newly classified lines and 9 new energy level values.

Experiment
The light source used in the present work is a discharge tube built at the Centro de Invesitgaciones Opticas, (CIOp), to study highly ionized gases [26]. It is a 30cm long Pyrex tube with an inner diameter of 3".
Gas excitation was produced by discharging a bank of low inductance capacitors varying between 2.5 and lOOnF and charged up to 19 kV through the tube. Light radiation emitted axially was analyzed using a 3 m normal incidence vacuum spectro-Revised and Extended Analysis of Five Times Ionized Argon (Ar VI) graph with a concave diffraction grating of 1200 lines/", blazed for 12008,. The plate factor in the first order is 2.77 A/ ". Ilford 4-2 plates were used to record the spectra between 280-2100A. C I11 and N I1 [27], 0 I11 [28] and lines of Ar 111-Ar V [27] were recorded as internal wavelength standards. Exposing the plates with lo4 shots we were able to obtain good spectra of argon below 3008,. The lines were observed in the first, second, and in some cases, third order of diffraction.
The gas pressure, the discharge voltage, and the capacitance were varied to distinguish between different stages of ionization. A well developed Ar VI spectrum was obtained with the following parameters: 120mTorr, 18 kV and 20 nF. The positions of spectral lines on the plates were determined with a rotating prism photoelectric automatic Grant comparator whose precision is lpm. The uncertainty in determining the wavelength of unperturbed lines by this procedure is estimated to be kO.018, in the first diffraction order.

Analysis
The Ar VI lines observed in the present work are given in Table I, 33 of them being without previous classification. The intensities of the lines given in the table are based on visual estimates and the values given in the calculated column are deduced from the optimized level values. The optimized energy values derived from the observed lines are given in Table I1 and the general structure of the term system is shown in Fig. 1 configurations made at the Department of Physics, University of Lund, were also used in the analysis. The predictions were obtained by diagonalizing the energy matrices with appropriately scaled Hartree-Fock (HF) values for the energy parameters. The computer code developed by Cowan [29] was used. All levels designations in Table I1 are in LS notation, and in the same table we present the percentage composition of the levels.
A comparison with the level system given by Pinnington et al. [22] shows that six of their level values have been confirmed, although the accuracy has been considerably improved. However, for seven levels we propose new identifications as discussed below.
The observed structure of the configurations 3p3, 3s3p3d and 3~3~4 s is shown in Fig. 2. For the level 3s3d3P)3d 'F,,, we propose the new value 344 307.9 cm-'. The level is established by five new lines that are given in Table I. For the level 3s3d3P)3d 'F,/, we propose the new value 346073.4cm-' determined by two new lines, see Table I. Both level values fit well with the extrapolation that can be done from the isoelectronic data published in Ref. [23].
For the level 3s3d3P)3d 'P3/2 we propose the new value 375 657.6cm-'. This level is determined by four new lines, see Table I. The level value fits very well with the isoelectronic trend from Ref. [23]. The isoelectronic graph gives for the 3~3p(~P)3d'P,~, level a probable value near 376300~x11-'. We were not able to find the transitions establishing this level, but according to the extrapolated values mentioned above and our theoretical predictions we reject the experimental value published in the work of Pinnington et al. [22] for this level.

i ( A ) Observed
Calculated Transition

Theoretical Interpretation
The level structure was theoretically interpreted by a leastsquares fit of the energy parameters to the experimental level values. For this purpose the computer code developed by Cowan [29] was used.
The scaled Hartree-Fock factor was 0.85 for all parameters, except for where the scaled factor was 0.95 and for E,, where the scaled factor was 1.00. These scaled factors were taken in this form because the computed energy-level intervals agree better with the experimental ones.
In order to obtain a better interpretation of the levels it was necessary to introduce the 3s24p configuration. The results of the parametric calculations are presented in Table  111.
The c1 parameter was kept free because all the levels of the 3p3 configuration are known. The first three configuration interaction integrals were held fixed in the calculation scaled at 0.75, 0.95 and 0.95 of their Hartree-Fock values. The rest of the configuration interaction integrals were held fixed at  Energy parameters (cm-') for the 3s23p, 3s24p, 3p3, 3s3p3dY 3s3p4s