Short histories of the research on this spectrum have been published by Kiess, Humphreys and Laun (46.1), Diringer (65.1), Blaise and Radziemski (76.2) and H.M. Crosswhite (82.3). Uranium is of great interest because it is the last stable element of the periodic table and is used as a fuel for nuclear power plants. As a result, many investigators have contributed information on wavelengths in emission and in absorption, g-values, isotope shifts, hyperfine structures of 235U and 238U and the interpretation of the energy levels. It has therefore been difficult to trace with certainty the origin of the data collected in the following tables. For instance, the g values and isotope shifts of a number of possible odd levels had been determined by Ben Osman (66.7) from transitions involving the 5f47s2 5I4 level long before this level was located at 7020 cm-1 by Guyon (71.8) and the first author has been credited in the table with the discovery of these odd levels.
Spectrum observations: The original analysis of U I by Kiess et al. (46.1) was made with the aid of the Zeeman effect from a list of over 9000 lines in the region 290 to 1100 nm. Besides the ground level 5f36d7s2 5L6 and the lowest level 7M6, of 5f36d27s, established independently by Schuurmans (46.2), they found 16 other odd levels and 275 even levels. Of those levels, 35 were spurious and the J values of 24 others had to be corrected later on. The g values of 43 even levels were measured at the Zeeman Laboratorium, Amsterdam (47.2; 49.6).
MacNally's compilation (unp.23) of the uranium wavelengths, measured at M.I.T. and the National Bureau of Standards and a list of about 3000 lines, observed in emission and in absorption by Bovey and Wise (59.6) were used in the early stages of the analysis of U I, undertaken by Blaise and Diringer with the help of the 238U - 235U isotope shifts. Accurate measurements of wavenumbers were started by Steinhaus at Los Alamos in 1959. Blaise measured during a stay at Argonne, the isotope shifts of about 6000 lines of U I and U II between 340 and 660 nm, on plates taken by J.K. Brody with the 9.15 m Paschen-Runge spectroghraph. Since 1960 a collaborative effort of spectroscopists at Los Alamos, Argonne, Livermore, Berkeley and Laboratoire Aimé Cotton (LAC) has greatly extended the analysis of U I and U II, and recently that of U III.
The wavenumbers of U lines emitted by hollow-cathode lamps with various carrier gases were measured interferometrically by Steinhaus and were used as internal standards in the measurements of the spectra observed with grating instruments. The section 330-500 nm was reported (72.3) while an unpublished list, completed in 1974, contains over 92000 lines, measured with an accuracy of about 0.003 cm-1 from 310 to 900 nm. The majority of these lines belong to the first two spectra and a few hundreds to U III.
The wavelengths and isotope shifts of some 500 infrared lines emitted by electrodeless discharge tubes of 238U, 236U and 235U were measured at Harwell with a conventional grating spectrograph supplemented by a pressure-scanned Fabry-Perot interferometer (60.8; 60.10).
Extensive measurements of the infrared spectrum were made at LAC by Guelachvili with a SISAM spectrometer (65.5) and later on with a Fourier transform (FT) spectrometer under various excitation conditions of the electrodeless discharge tubes. These last spectra were investigated by Guyon (72.4). Morillon recorded the spectrum between 2.3 and 3.5 µm with a grid spectrometer (70.7). The FT spectra emitted from hollow cathode lamps were recorded from 3000 to 13900 cm-1 by Vergès in 1975 and from 15872 to 26792 cm-1 by Gerstenkorn et al. (77.1).
Since 1980 FT spectra have also been recorded at the Kitt Peak National Observatory. Palmer et al. list 4928 lines covering the region 11000-26000 cm-1, in the atlas (80.6) and they have now measured the whole spectrum from 1900 to 41700 cm-1. Conway et al. recorded again at Kitt Peak the spectrum between 1817 and 40000 cm-1 from a hollow cathode of depleted uranium with a flowing argon carrier gas and published a list of over 9000 infrared lines, half of them classified, between 1817 and 5600 cm-1 (84.1).
Absorption spectra have been produced from uranium-furnaces by Penkin and Frish (57.5) and Bovey and Wise (59.6). Radziemski et al. used the absorption spectra obtained by the flash-photolysis and flash-discharge techniques to differentiate between U I and U II lines (71.4). Using an inductively heated furnace, Tomkins recorded some 2300 lines in the region 115-650 nm with the Paschen-Runge spectrograph and with a 3 m normal-incidence vacuum spectrograph below 200 nm (75.1). One fourth of these lines were transitions between the lowest odd levels and 319 even levels located between 32834 and 49790 cm-1. Amongst those levels, 79 had already been found from the emission spectrum (75.3; 76.6). The J values of 66 other levels were determined subsequently in the emission spectrum. Finally, the g-values and isotope shifts for a number of them were measured by Blaise.
Zeeman effect: Zeeman spectrograms were taken at Argonne by Fred and Giacchetti on the Paschen-Runge spectrograph, between 9065 and 32300 cm-1, with a magnetic field of 2.4 T. These spectrograms were analyzed at LAC by Guelachvili (65.5), Ben Osman (66.7) and Blaise and Radziemski (76.2). The Zeeman structures of 27 infrared lines between 4831 and 9842 cm-1 were observed with a high resolution SISAM by Vergès (71.9).
Isotope shifts: Isotope shifts of uranium lines have been measured in several laboratories, either with grating spectrographs at Oak Ridge (49.1; 51.5) , Moscow (55.1), Harwell (59.6; 60.10) and Argonne, or interferometrically at Moscow (59.4), Harwell (60.8), LAC (60.3; 63.5; 64.1; 65.1), and Montreal (76.4; 77.4; 78.3). Despite the inaccuracy of many isotope shift determinations, these data have led to the discovery of a large number of energy levels and to the identification of new configurations. Inversely, extensive measurements of 238U - 234U shifts by Gerstenkorn et al. (78.5) Engleman and Palmer (80.2) and of 238U - 235U shifts in the infrared by Conway and Worden (84.2) resulted in the derivation of more accurate level isotope shifts. Whereby a zero shift was assigned to the ground state and the relationship DT(238-235) = DT(238-234)/1.194 was assumed (77.4).
These past years, using all the available data, Blaise has revised the analysis of U I and found a number of new levels. The energies of 534 odd levels and 1422 even levels are now known with an accuracy of about 0.003 cm-1. g and DT values are credited to the first investigators whose measurements differ from the revised values of columns 7 and 8 by less than 0.02 for g and 25.10-3 cm-1 for DT.
Theoretical interpretation:
Low odd levels: Preliminary calculations of the lowest levels of f3ds2 were carried out by Judd (62.5) and Spector (73.3). Guyon et al. (74.7) have used a truncated basis set f3ds2 + f3(4I)d2s and interpreted 68 levels with a rms uncertainty of 124 cm-1. Similar calculations have been performed by Rajnak (76.10; unp.28).
The designation of the levels in this table is from (82.3). H. Crosswhite and H.M. Crosswhite who have taken into account the influence of the f3d3 configuration. Rajnak suggested that the latter may begin between 12000 and 15000 cm-1, while the recent identification of f3d3 7M6 at 23084 cm-1 confirmed Brewer's prediction (21000 ± 4000 cm-1) (71.1). Only 30 levels have a first component greater than 50 % but the contribution of f3d3 to the eigenvectors is always less than 1 %. Labels within parentheses indicate a leading component percentage between 25 and 50 % and a blank entry, a percentage less than 25 %. The contributions of the configurations f3ds2 and f3d2s are given in the last two columns. They do not agree with the isotope shift measured for the 13149 cm-1 level by Engleman and Palmer (80.2).
Above 14200 cm-1 a few level designations are based on calculations but most of them are empirical. The 27920 cm-1 level was identified as 5f37d7s2 5L6 by Rajnak and Fred who have discussed the correlation between observed isotope shifts and electron density at the nucleus (77.3).
Low even levels: Calculations of f2d2s2 + f3s2p + f4s2 were carried out by Rajnak (unp.28) and Crosswhite (82.3). Configuration percentages for the lowest levels given in the table are from (82.3). The presence of levels belonging to the configurations f3dsp, f4ds and probably f2d3s makes it difficult to identify levels with certainty above 15000 cm-1.
High odd levels: About 125 levels discovered by means of multistep laser excitation between 32273 and 34804 cm-1 (76.3; 76.9; 79.1) and not identified with levels found by classical methods, are listed separately.
High even levels: Levels deduced only from absorption spectra (75.1; 75.3) most of them with uncertain J values, and levels above 49592 cm-1 which were obtained by three step dye laser excitation (79.1), are listed separately. Rydberg series and autoionizing series have also been reported near the first ionization limit 49958.4 ± 0.5 cm-1 (76.11; 82.2).
Forbidden lines: Two parity-forbidden transitions, f3ds2 5L6 - f3ds2 5L7 at 3800.829 cm-1 and f4s2 5I4 - f4s2 5I5 at 3030.605 cm-1, have been observed by Conway et al. (84.1). Eleven other forbidden transitions between levels of the same configuration have been identified by Blaise among the lines listed in (84.1) at the following wavenumbers (in cm-1):
3844.817 (5L7 - 5L8), 3662.508 (5L8 - 5L9), 3655.385 (5K5 - 5K6), 3050.412 (5K6 - 5K7), ~ 3021.216 (5K7 - 5K8), 3410.785 (5I4 - 5I5), 4526.538 (3I5 - 5I6) and 4263.687 (5H4 - 5H5) between levels of f3ds2 and 1869.601 (7M6 - 7M7), 2567.157 (7M7 - 7M8) and 2780.578 (7M8 - 7M9) between levels of f3d2s.
After the publication of the Tables of Actinides in 1992, two contributions to the interpretation of the low odd configurations were issued :
- A laser induced fluorescence experiment in an atomic beam led Avril et al. to hyperfine constants A and B for 28 low and 22 even levels.
- The parametric interpretation of 18 A and 16 B values pertaining to the 5f3 6d 7s2 configuration used the fine structure study by Guyon (72.4, 74.7).
On the way to a comprehensive description of the low odd levels, Petit (99.1) used an extended version of the Cowan codes to calculate 5f3 6d 7s2 + 5f3 6d2 7s with complete basis sets, the average deviation being 53 cm-1 for 155 experimental levels. Doubts on the energies on the lowest levels of 5f3 6d3, the lack of g-factors and some questionably isotope shifts prevent reliable identification above 17000 cm-1 (unp.29 ).