Abstract
Three-times ionized xenon Xe IV spectrum in the 1070–6400 Å region was analyzed using a pulsed discharge light source. A set of 163 transitions was classified for the first time, and 36 new energy levels belonging to the 5s25p26d and 5s25p27s even configurations were determined. The relativistic Hartree–Fock method, including core-polarization effects, were used. In these calculations, the electrostatic parameters were optimized by a least-square procedure in order to improve the adjustment to experimental energy levels. We also present a calculation based on a relativistic multiconfigurational Dirac–Fock approach.
1. Introduction
There is great interest in spectroscopy data of Xenon due to their applications in collision physics, astrophysics, and laser physics. Various atomic parameters, such as energy levels, oscillator strengths, transition probabilities, and radiative lifetimes, have many important astrophysical applications. Transition probabilities are needed for calculating the energy transport through the star in model atmospheres [1] and for direct analysis of stellar chemical compositions [2]. Xenon was observed in chemically peculiar stars [3] and planetary nebulae [4]. The spectrum analysis of planetary nebula NGC7027 by Péquignot and Baluteau [5] has stimulated the calculation of transition probabilities for some forbidden lines of astrophysical interest [6]. The Xe VI and Xe VII lines were observed in the ultraviolet spectrum of the hot DO-type white dwarf RE 0503-289 [7,8]. In particular, the Xe IV spectrum was detected in the spectrum of NGC 7027 together with a variety of ionic species, providing a unique opportunity to study the chemical composition of the nebula at a level normally unachievable in another emission line nebulae [9,10].
Saloman [11] published a revised compilation of energy levels and observed spectral lines of all ionization stages of Xe, referring to studies published to date [12,13,14,15,16]. Light sources include direct-current hollow cathode discharge, theta-pinch discharge, and pulsed capillary discharge. Most of the information is from two studies: Tauheed et al. [13] classified 114 Xe IV lines in VUV using a modified triggered spark initiated by a xenon gas blast as spectral source, and Gallardo et al. [14], who analyzed the 5s25p26p, 5s25p24f, 5s5p4, 5s25p25d, and 5s25p26s configurations, providing the wavelengths for 618 classified lines in their list, using a capillary discharge as light source.
More recently the study by Raineri et al. [15] reported the weighted oscillator strengths and cancellation factor (CF), calculated from fitted values of the energy parameters of all 769 dipole electric lines belonging to the Xe IV spectrum reported in the compilation [11], including 49 new classified lines. Hartree–Fock relativistic (HFR) calculations and parametric fits were used. In addition, the results presented in their study were compared to those from Bertuccelli et al. [16].
In order to proceed withthe study of the threetimes ionized xenon spectrum, a new spectral analysis of this ion is presented in this paper. New 36 energy levels for 5s25p2 (6d+7s) configurations and 163 new transitions in the 1070–6400 Å region are reported. The relativistic Hartree–Fock method based on the code of Cowan [17] was used. The energy matrix was calculated using energy parameters adjusted to fit the experimental energy levels. Core polarization effects were taken into account in our calculations [18]. We also present a multiconfigurational relativistic approach for the Dirac equation (MCDF), by using the general relativistic atomic structure package (GRASP) [19].
2. Experimental Methods
The spectral source used in this study is based on the pulsed discharge tube built at the Centro de Investigaciones Opticas to study highly ionized noble gases [20]. It consists of a Pyrex tube of about 100 cm with inner diameter of 0.5 cm. The electrodes, placed 80 cm apart, are made of tungsten covered with indium to avoid the impurities coming from the electrodes. The gas excitation was produced by discharging a bank of low-inductance capacitors ranging from 20 to 280 nF, charged with voltages up to 20 kV. The VUV region of the spectrum was recorded using a 3m normal incidence spectrograph equipped with 1200 lines/mm concave diffraction grating and with a plate factor of 2.77 Å/mm in the first diffraction order. Internal wavelength standards are from C, N, O, and known lines of xenon. The wavelength range above 2000 Å was recorded using a 3.4 m Ebert plane-grating spectrograph with 600 lines/mm and a plate factor of 5 Å/mm in the first diffraction order. Thorium lines from an electrodeless discharge were superimposed on the spectrograms and served as reference lines. A photoelectric semiautomatic Grant comparator was used to measure the spectrograms. The uncertainty of the wavelength values of lines was estimated to be correct to ±0.01Å above 2000 Å and ±0.02 Å in the VUV region.
3. Results and Discussion
In this study, we used the modified version of Cowan’s atomic calculation package [17], described in our paper [18], for the inclusion of the polarization potentials as a modification in the Hartree–Fock equations. In addition, we considered the corrections of the reduced matrix element used in our previous papers [21], which is the same modification used by Quinet et al. [22] to correct transition matrix elements when including CP and core penetration effects. These methods demand knowledge on the polarizability andcore cut-off radius. The value of αd for Xe IV core, that is, for Xe 8+ is given by Koch [23] in 0.81130 a03 and the rc value in 1.16 a0, defines the boundaries of the atomic core.
We adjusted the values of energy parameters to the experimental energy levels of the Xe IV through a least-squares calculation. With the adjusted values, we calculated the composition of the 5s25p2 (6d+7s) energy levels presented in Table 1, where we included lifetimes calculated using HFR and HFR+CP with adjusted energy parameters (here named HFRa and HFR+CPa, respectively) and using multiconfigurational Dirac Fock (MCDF). The MCDF approach was carried out with the extended average level assuming a uniform charge distribution in the nucleus, with a xenon atomic weight of 131.3. The values presented in this work for lifetimes in the MCDF calculation are in Babushkin gauge since this one, in the non-relativistic limits (length), has been found to be the most stable value in many situations, in the sense that it converges smoothly as more correlation is included [24].
Table 1.
Energy levels, composition, and lifetimes of Xe IV.
In the analysis of spectroscopic data, we take into account isoelectronic trends, Ritz combinations, least-squares adjustment, and relative line intensities in order to identify 36 energy levels belonging to 5s25p2(6d+7s) configurations for the first time.
As for the isoelectronic sequence calculations used to produce the plots for observed minus calculated (“obs.-calc.”) trends along the six first elements of the Sb sequence, we used the configurations 5s25p3, 5s25p24f, 5s25p26p, 5s5p36s, 5s5p37s, 5s5p35d, 5s5p3 6d, 5p5 for odd parity and 5s5p4, 5s25p25d, 5s5p35f, 5s25p25g, 5s25p26s, 5s25p26d, 5s25p27s for even parity. The calculations included core polarization effects (HFR+CP), with the values of αd and rc taken from Table 2.
Table 2.
Values for polarizability αd and cut-off radius rc, used in antimony isoelectronic sequence calculations (HFR+CP). Here, is the Bohr radius.
It must be noted that we implemented the modifications suggested by Kramida [25,26] to correct an error in Cowan’s package in order to perform the calculations presented here.
Data for isoelectronic analysis are from NIST [27] for Sb I, Te II, I III and from Sharman, Tauheed, and Rahimullah for Ba VI [28]. Our analysis is synthesized in Figure 1, Figure 2 and Figure 3. Surely the LS coupling scheme is not the most appropriate to describe the 6d and 7s configurations, which we concluded after glancing over configuration purities; intermediate couplings provide better descriptions for these levels. We observed a strong eigenvector mixing for all elements studied. However, most of the isoelectronic data available for comparisons are described in the LS scheme, and that was the reason why we chose it.
Figure 1.
Isoelectronic trend for the multiplet (3P) 4F energy levels of the 5s25p2 6d configuration.
Figure 2.
Isoelectronic trend for the multiplet (3P) 4P energy levels of the 5s25p2 6d configuration.
Figure 3.
Isoelectronic trend for the multiplet (3P) 4P energy levels of the 5s25p2 7s configuration.
There is no absolute scale for experimental intensity and therefore we only test its proportionality with the theoretical intensity. We do not include corrections due to the variation of plate reflectivity as a function of wavelength—there is no precise model for this. Our criterion for statistical correlation is to obtain a positive value as close as possible to the unit. Therefore, having a good statistical correlation supports our analysis, but it is just one of the analysis criteria.
The formula from Cowan’s book [17], page 403, tells us that line intensity is proportional to wavenumber and weighted transition probability. We analyzed the statistical correlation of the logarithm related to this quantity with the experimental line intensities, which is a visual estimate of the plate blackening (hence the logarithm), obtaining 0.20 for the array 0.32 for and 0.34 for . These values were acquired by the HFR+CPa calculation, which is close to HFRa and much better thanab initioHFR and HFR+CP calculations. We also performed a MCDF calculation for gA values. Its agreement with the experimental line intensity shows a poor correlation when compared with HFRa and HFR+CPa for , that is, 0.06 for the line array, 0.14 for , and 0.18 for It is important to note that our MCDF calculations were performed using a non-current version of the GRASP code where more configurations could not be included. By using a newer version of Grasp codes it would be possible to expand the number of configurations to get better results, which could be more competitive to HFRa and HFR + CPa methods
To understand thesignificance of these values, we compared our values of gA with the experimental values that are in the paper by Bertuccelli et al. [16]. Similarly to them, only 25% of our gA values (HFR+CPa) are within the experimental error. However, a statistical correlation of 0.94 indicates that our values are very linearly proportional to their experiment. When considering the same lines of [16], but substituting their experimental gA values by our estimates for line intensity, correlation with HFR+CPa results in 0.33 for the line array, 0.48 for , and 0.50 for Therefore, we can conclude that the calculated gA values support our line classification with reasonable agreement.
It is important to note that in this spectral analysis all new levels but two are classified on the basis of two or more lines. The level 4F5/2 is a classification attempt based on the only possible line in our spectrograms at 1801.53 Å, a transition with 4f:4G5/2, the strongest spontaneous emission from this level. However, this value does not fit the isoelectronic “obs.-calc.” curve. We remove this problem by switching the positions of levels 6d:4P5/2 and 6d:4F5/2 for Xe IV in the isoelectronic analysis. An intense mixing for 6d:4P5/2, 4D5/2, and 4F5/2 makes the components for the eigenvectors exchange their intensity along with the four first elements, and our choice grouped the energy of the respective multiplets.
Due to similar reasons, we also switched 4D5/2 and 4P5/2 energy levels for Te III and I III in the respective isoelectronic sequences.
The other level that only has one observed transition is (1S)6d: 2D3/2 that we confirm by our isoelectronic analysis and considering the good agreement in the least squares fit calculation.
There is not much data available for isoelectronic analysis. The lack of information on Cesium and the composition mixing makes level designation a challenge. However, the isoelectronic sequences agree reasonably well with our designations.
Table 3 shows 163 Xe IV lines classified for the first time for transitions involving 5s25p2(6d+7s) energy levels. We also calculated the weighted transition probability rate gA, where g is the statistical weight 2J+1 of the upper level. We presented gA values obtained from the four methods studied: With and without optimized parameters obtained from least-squares calculations, and with and without core polarization effects for wavefunctions and reduced matrix elements calculations.In these methods, we used the same configuration sets as in [15], that is, 5s25p3, 525p2 (4f+6p), 5s 5p35d, 5p5 and 5s5p4, 5s25p2 (6s+7s+5d+6d) configurations for odd and even parities, respectively.
Table 3.
Transitions and weighted transition rates for Xe IV.
Table 4 shows the result of least squares adjustment for even parity levels, where 6d and 7s configurations are included. All single configuration parameters, the Rk integrals for 5s5p4-5s25p26s, 5s5p4-5s25p25d, 5s25p26s-5s25p2 5d interactions, and the R1(5p,5d;6d,5p)of the 5s25p2 5d-5s25p26d interaction were left free during the final calculation. The rest of the configuration interaction integrals remained fixed at 85% of their Hartree–Fock values. We found a standard deviation of 138 cm−1 for this adjustment.
Table 4.
Least-squares parameters for even parity of Xe IV. Standard deviation is 138 cm−1.
4. Conclusions
In this study we extended the knowledge of the Xe IV spectrum to the 5s25p27s and 5s25p2 6d configuration, from a set of 163 new line classifications. To produce this new information, we used a set of different analysis tools, including calculations from three models (HFR, HFR+CP, and MCDF), least-squares adjustment, line intensity comparisons, and isoelectronic analysis, which makes us very confident in our results.
Author Contributions
All authors contributed equally to this work.
Funding
This research received no external funding.
Acknowledgments
This research was supported by the Consejo Nacional de Investigaciones Científicas y Tecnicas (CONICET), Argentina, and by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), Brazil, Finance Code 001. The authors thank Espaço da Escrita–Pró-Reitoria de Pesquisa–UNICAMP-for the language services provided. Support of the Comision de Investigaciones Científicas de la Província de Buenos Aires (CIC), where M.R. is a researcher, is also gratefully acknowledged.
Conflicts of Interest
The authors declare no conflicts of interest.
References
- Gustafsson, B. The future of stellar spectroscopy and its dependence on YOU. Phys. Scr. 1991, 34, 14–19. [Google Scholar] [CrossRef]
- Biémont, E.; Blagoev, K.; Campos, J.; Mayo, R.; Malcheva, G.; Ortíz, M.; Quinet, P. Radiative parameters for some transitions in Cu(II) and Ag(II) spectrum. Spectrosc. Relat. Phenom. 2005, 144, 27–28. [Google Scholar] [CrossRef]
- Cowley, C.R.; Hubrig, S.; Palmeri, P.; Quinet, P.; Biémont, É.; Wahlgren, G.M.; Schütz, O.; González, J.F. HD 65949: Rosetta stone or red herring. Mon. Not. R. Astron. Soc. 2010, 405, 1271–1284. [Google Scholar] [CrossRef]
- Otsuka, M.; Tajitsu, A. Chemical abundances in the extremely carbon-rich and xenon-rich halo planetary nebula H4-1. Astrophys. J. 2013, 778, 146. [Google Scholar] [CrossRef]
- Péquignot, D.; Baluteau, J.-P. The identification of krypton, xenon, and other elements of rows 4, 5 and 6 of the periodic table in the planetary nebula NGC 7027. Astron. Astrophys. 1994, 283, 593–625. [Google Scholar]
- Biémont, E.; Hansen, J.E.; Quinet, P.; Zeippen, C.J. Forbidden transitions of astrophysical interest in the 5pk (k=1–5) configurations. Astron. Astrophys. Suppl. Ser. 1995, 111, 333–346. [Google Scholar]
- Werner, K.; Rauch, T.; Ringat, E.; Kruk, J.W. First detection of krypton and xenon in a white dwarf. Astrophys. J. 2012, 753, L7. [Google Scholar] [CrossRef]
- Rauch, T.; Hoyer, D.; Quinet, P.; Gallardo, M.; Raineri, M. The Xe VI ultraviolet spectrum and the xenon abundance in the hot do-type white dwarf RE 0503−289. Astron. Astrophys. 2015, 577, A88. [Google Scholar] [CrossRef]
- Zhang, Y.; Liu, X.-W.; Luo, S.-G.; Péquignot, D.; Barlow, M.J. Integrated spectrum of the planetary nebula NGC 7027. Astron. Astrophys. 2005, 442, 249–262. [Google Scholar] [CrossRef]
- Zhang, Y.; Williams, R.; Pellegrini, E.; Cavagnolo, K.; Baldwin, J.A.; Sharpee, B.; Phillips, M.; Liu, X.-W. Abundances of s-process elements in planetary nebulae: Br, Kr &Xe in Planetary Nebulae in Our Galaxy and Beyond. Proc. IAU Symp. 2006, 234, 549–550. [Google Scholar] [CrossRef]
- Saloman, E.B. Energy levels and observed spectral lines of xenon, Xe I through Xe LIV. J. Phys. Chem. Ref. Data 2004, 33, 765–921. [Google Scholar] [CrossRef]
- Gallardo, M.; ReynaAlmandos, J.G. XenonLines in the Range from 2000 Å to 7000 Å; Serie “Monografias Cientificas” No. 1; Centro de Investigaciones Opticas: LaPlata, Argentina, 1981.
- Tauheed, A.; Joshi, Y.N.; Pinnington, E.H. Revised and extended analysis of the 5s25p3, 5s5p4, 5s25p25d and 5s25p26s configurations of trebly ionized xenon (Xe IV). Phys. Scr. 1993, 47, 555–560. [Google Scholar] [CrossRef]
- Gallardo, M.; Raineri, M.; Reyna Almandos, J.G.; Di Rocco, H.O.; Bertuccelli, D.; Trigueiros, A.G. 5s25p2(6p + 4f) configurations in triply ionized xenon (Xe IV). Phys. Scr. 1995, 51, 737–751. [Google Scholar] [CrossRef]
- Raineri, M.; Lagorio, C.; Padilla, S.; Gallardo, M.; Reyna Almandos, J. Weighted oscillator strengths for the Xe IV spectrum. At. Data Nucl. Data Tables 2008, 94, 140–159. [Google Scholar] [CrossRef]
- Bertuccelli, G.; Di Rocco, H.O.; Iriarte, D.I.; Pomarico, J.A. Experimental Determination of Transition Probabilities of Xe IV; Comparison with Semiempirical Calculations. Phys. Scr. 2000, 62, 277–281. [Google Scholar] [CrossRef]
- Cowan, R.D. The Theory of Atomic Structure and Spectra; University of California Press: Berkeley, CA, USA, 1981. [Google Scholar]
- Pagan, C.J.B.; Raineri, M.; Gallardo, M.; Reyna Almandos, J. Spectral Analysis and New Visible and Ultraviolet Lines of ArV. Astron. Astrophys. Suppl. Ser. 2019, 242, 24. [Google Scholar] [CrossRef]
- Dyal, K.G.; Grant, I.; Johnson, C.T.; Parpia, F.A.; Plummer, E.P. GRASP: A general-purpose relativistic atomic structure program. Comput. Phys. Commun. 1989, 55, 425–456. [Google Scholar] [CrossRef]
- Reyna Almandos, J.; Bredice, F.; Raineri, M.; Gallardo, M. Spectral analysis of ionized noble gases and implications for astronomy and laser studies. Phys. Scr. 2009, T134, 014018. [Google Scholar] [CrossRef]
- Raineri, M.; Gallardo, M.; Pagan, C.J.B.; Trigueiros, A.G.; ReynaAlmandos, J. Lifetimes and transition probabilities in KrV. J. Quant. Spectrosc. Radiat. Transf. 2012, 113, 1612–1627. [Google Scholar] [CrossRef]
- Quinet, P.; Palmeri, P.; Biémont, E.; McCurdy, M.M.; Rieger, G.; Pinnington, E.H.; Wickliffe, M.E.; Lawler, J.E. Experimental and theoretical radiative lifetimes, branching fractions and oscillator strengths in Lu II. Mon. Not. R. Astron. Soc. 1999, 307, 934–940. [Google Scholar] [CrossRef]
- Koch, V.; Andrae, D. Static Electric DipolePolarizabilities for Isoelectronic Sequences. Int. J. Quantum Chem. 2011, 111, 891–903. [Google Scholar] [CrossRef]
- Grant, I. Relativistic Quantum Theory of Atoms and Molecules: Theory and Computation; Springer: Oxford, UK, 2007. [Google Scholar]
- Kramida, A. Configuration interactions of class 11: Na error in Cowan’s atomic structure theory. Comput. Phys.Commun. 2017, 215, 47–48. [Google Scholar] [CrossRef]
- Kramida, A. Corrigendum to “Configuration interactions of class 11: Na error in Cowan’s atomic structure theory”. Comput. Phys. Commun. 2018, 232, 266–267. [Google Scholar] [CrossRef]
- NIST Standard Reference Database 78. Version 5.6. Available online: https://www.nist.gov/pml/atomic-spectra-database (accessed on 31 October 2018).
- Sharman, M.K.; Tauheed, A.; Rahimullah, K. Spectral analysis of 5s25p2 (6p+6d+7s) configurations of Ba VI. J. Quant. Spectrosc. Radiat. Transf. 2014, 142, 37–48. [Google Scholar] [CrossRef]
© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).