Spectrum and Energy Levels of Four-times Ionized Yttrium (y V)

The analysis of the spectrum of four-times-ionized yttrium, Y V, was extended to provide a large number of new spectrum lines and energy levels. The new analysis is based on spectrograms made with sliding-spark discharges on 10.7 m normal-and grazing-incidence spectrographs. The measurements cover the region 184–2549 Å. The results revise levels for this spectrum by Zahid-Ali et al. (1975) and by Ateqad et al. (1984). Five hundred and seventy lines were classified as transitions between 23 odd-parity and 90 even-parity levels. The 4s 2 4p 5 , 4s4p 6 , 4s 2 4p 4 4d, 5s, 5p, 5d, 6s configurations are now complete. Results for the 4s 2 4p 4 6d and 7s configurations are tentative. Ritz-type wavelengths were determined from the optimized energy levels, with uncertainties as low as ±0.0004 Å. The observed configurations were interpreted with Hartree-Fock calculations and least-squares fits of the energy parameters to the observed levels. Oscillator strengths for all classified lines were calculated with the fitted parameters. The results are compared with values for the level energies, percentage compositions, and transition probabilities from recent ab initio theoretical calculations. The ionization energy was revised to 607,760 ± 300 cm −1 (75.353 ± 0.037 eV).


Introduction
The four-times ionized yttrium atom, Y V, has a Br-like electronic structure with ground configuration 4s 2 4p 5 and excited states 4s4p 6 and 4s 2 4p 4 nl.The spectrum has a somewhat checkered past.It was first analyzed in 1939 by Paul and Rense [1], who, from a set of transitions to the 4s 2 4p 5 2 P ground term, determined levels of the 4s4p 6 2 S 1/2 , 4s 2 4p 4 4d, and 4s 2 4p 4 5s configurations.Unfortunately, an isoelectronic plot published by Edlén [2] in 1964 showed that the 4s 2 4p 5 2 P 3/2 -2 P 1/2 interval of Paul and Rense [1] (12,068 cm −1 ) was inconsistent with the known intervals for the rest of the isoelectronic sequence.From his plot, Edlén predicted an interval of 12,470 ± 20 cm −1 .Since essentially all of their levels were based on transitions to the 4s 2 4p 5 2 P term, Edlén concluded that the analysis would have to be completely revised.A start on this revision came in 1970, when Reader and Epstein [3] observed the true 4s 2 4p 5 2 P 1/2,3/2 -4s4p 6 2 S 1/2 transitions, thus obtaining the position of 4s4p 6 2 S 1/2 and a revised value for the 2 P term splitting.Their splitting of 12,459.9± 3.0 cm −1 was indeed close to the value predicted by Edlén.In 1972 Reader and Epstein [4] observed further transitions to the 4s 2 4p 5 2 P ground term and established nearly all levels of the 4s 2 4p 4 4d and 5s configurations.Only the levels of 4p 4 4d with J = 7/2 and 9/2, which do not combine with 4p 5 2 P 1/2,3/2 , and the 4p 4 4d ( 3 P) 4 D 1/2 level could not be located.
In the present work we observed the spectrum of Y V in the ultraviolet and determined a new set of energy levels.About half the 4s 2 4p 4 5d levels of [5] were found to be spurious.Several of the 4s 2 4p 4 6s levels in this paper had incorrect J-values and in fact belong to 4s 2 4p 4 5d.Nearly all of the 4s 2 4p 4 5p levels of [6] were spurious, as were all of the reported J = 7/2, 9/2 levels of 4s 2 4p 4 4d and 5d.

Experiment
The observations were the same as used for earlier work in our laboratory on yttrium [4,7,8].Briefly, the light source was a low-voltage sliding-spark with metallic yttrium electrodes.The source was operated as described by Reader et al. [9].From 500 to 2549 Å we used the NIST 10.7-m normal-incidence vacuum spectrograph; from 184 to 500 Å we used the NIST 10.7-m grazing-incidence spectrograph.Both instruments had gratings with 1200 lines/mm.The plate factor for the normal-incidence spectrograph was about 0.78 Å/mm.The plate factor for the grazing-incidence spectrograph at 350 Å was 0.25 Å/mm.From 600 to 2549 Å the spectra were calibrated by spectra of Cu II excited in a hollow cathode discharge.Below 600 Å calibration was obtained from lines of Y in various stages of ionization.Shifts between the reference spectra and the yttrium spectra were removed by use of impurity lines of oxygen, nitrogen, carbon, and silicon.Complete references for the calibration spectra are given in Reference [8].
Ionization stages were distinguished by comparing the intensities of the lines at various peak currents in the spark.The spectra of Y V were relatively enhanced at a peak current of about 2000 A.
The wavelengths, intensities, and classifications of the observed lines of Y V are given in Table 1.All wavelengths are in vacuum.The intensities are estimates of photographic plate blackening.The intensities range from 1 to 5,000,000.The system used to obtain this extensive scale of intensities is described in a recent paper on Mo VI [10].No attempt was made to account for spectrograph or plate emulsion response.The strongest lines in the spectrum appear as a group of 4p 4 5p-5d transitions around 1350 Å.
The general uncertainty of the wavelengths is ±0.007Å. Hazy lines (h) were given an uncertainty of ±0.010 Å; perturbed (p), complex (c), or asymmetric lines (s, l) an uncertainty of ±0.020 Å; unresolved (u) or doubly classified (dc) lines an uncertainty of ±0.030 Å.All uncertainties are reported at the level of one standard deviation.

Spectrum Analysis and Level Values
The analysis was carried out in a manner similar to that used for the recent analysis of Mo V [11].As described there "Interpretation of the spectrum was guided by calculations of the level structures and transition probabilities with the Hartree-Fock code of Cowan [12].Further guidance was provided by construction of two-dimensional transition arrays with the computer spreadsheet method described by Reader [13]." The odd parity energy levels are given in Table 2, the even levels in Table 3.In addition to the usual spectroscopic designations in either LS or J 1 l (pair) coupling, the levels are given shorthand designations that are used in the classification of the spectral lines.The shorthand designations are explained in the footnotes to Tables 2 and 3.As described in [11] "The values of the energy levels were optimized with the computer program ELCALC, an iterative procedure in which the observed wave numbers are weighted according to the inverse square of their uncertainties.The uncertainties of the level values given by this procedure are also listed."(The program ELCALC was written by L. J. Radziemski of the Research Corporation, Tucson, Arizona 85712.The procedure and definition of level value uncertainties have been described by Radziemski and Kaufman [14].)For the level optimization only the most reliably classified lines were used.That is, lines that were very weak or that appeared with suspiciously high intensities were excluded.
Figure 1 shows a schematic overview of the positions of the 4s 2 4p 5 , 4s4p 6 , 4s 2 4p 4 4d, 5s, 5p, 5d, and 6s, configurations.It also shows the calculated positions of the 4s 2 4p 4 4f and 4s4p 5 4d configurations.a Designations are given with a short form of the configuration (two places) followed by the ordinal number of the calculated J-value for the configuration (one place) and the J value (one place).For example, 5p73 indicates the seventh level with J = 3 for the 4p 4 5p configuration.p5 3 and p5 1 indicate the J = 3/2 and 1/2 levels of the 4p 5 configuration, respectively.
The 4p 4 4d ( 3 P) 4 F 9/2 (4d19) and ( 1 D) 2 G 9/2 (4d29) levels are necessarily based on only a single transition.However, the lines assigned to these transitions are both very strong and place the J = 9/2 levels close to their predicted positions.There is no doubt as to their identifications.
The structure of the 4p 4 4d configuration is shown in Figure 2.This is similar to Figure 1 of [4], except that we show here the observed positions of levels that were previously unknown.
The 4p 4 4d ( 3 P) 4 F9/2 (4d19) and ( 1 D) 2 G9/2 (4d29) levels are necessarily based on only a single transition.However, the lines assigned to these transitions are both very strong and place the J = 9/2 levels close to their predicted positions.There is no doubt as to their identifications.
The structure of the 4p 4 4d configuration is shown in Figure 2.This is similar to Figure 1 of [4], except that we show here the observed positions of levels that were previously unknown.

4s 2 4p 4 5s Levels
The levels of the 4s 2 4p 4 5s configuration, which were complete in [4], have improved values as a result of their combinations with 4p 4 5p.In Figure 3 we give the structure of the 4p 4 5s configuration.This is the same as Figure 2 of [4], except that here we designate the levels in J 1 l-coupling, rather than J 1 j-coupling.

4s 2 4p 4 5s Levels
The levels of the 4s 2 4p 4 5s configuration, which were complete in [4], have improved values as a result of their combinations with 4p 4 5p.In Figure 3 we give the structure of the 4p 4 5s configuration.This is the same as Figure 2 of [4], except that here we designate the levels in J1l-coupling, rather than J1j-coupling.

4s 2 4p 4 5p Levels
All levels of this configuration have been located.Of the 21 levels of this configuration given in [6], only three could be confirmed (345444, 360851, and 374278 cm −1 ).The levels at 342193 and 355076 cm −1 were confirmed, but were found to have incorrect J-values.The structure of the 4p 4 5p levels is

4s 2 4p 4 5p Levels
All levels of this configuration have been located.Of the 21 levels of this configuration given in [6], only three could be confirmed (345444, 360851, and 374278 cm −1 ).The levels at 342193 and 355076 cm −1 were confirmed, but were found to have incorrect J-values.The structure of the 4p 4 5p levels is shown in Figure 4.The levels are designated in J 1 l-coupling.

4s 2 4p 4 5p Levels
All levels of this configuration have been located.Of the 21 levels of this configuration given in [6], only three could be confirmed (345444, 360851, and 374278 cm −1 ).The levels at 342193 and 355076 cm −1 were confirmed, but were found to have incorrect J-values.The structure of the 4p 4 5p levels is shown in Figure 4.The levels are designated in J1l-coupling.The 4p 4 5d and 6s configurations lie very close in energy and are treated together.The levels are shown in Figure 5; they are designated in J 1 l-coupling.As with 4p 4 4d, the J = 9/2 levels could be established by only a single line.However, there is little doubt as to the identifications.A few of the levels of these configurations given in [5] could be confirmed, although some of the J-values and configuration assignments had to be revised.All of the 4p 4 5d levels of [6] were found to be spurious.The 4p 4 5d and 6s configurations lie very close in energy and are treated together.The levels are shown in Figure 5; they are designated in J1l-coupling.As with 4p 4 4d, the J = 9/2 levels could be established by only a single line.However, there is little doubt as to the identifications.A few of the levels of these configurations given in [5] could be confirmed, although some of the J-values and configuration assignments had to be revised.All of the 4p 4 5d levels of [6] were found to be spurious.3.5.4s 2 4p 4 6d and 4s 2 4p 4 7s Levels Based on our calculations, we were able to assign a number of low wavelength lines with clear Y V character as transitions to the ground term from levels of 4p 4 6d and 7s.For pairs of lines with wave number differences that closely match the 4p 5 2 P interval, the implied levels are relatively certain.However, the designations are considered to be tentative.Where the levels are based on Atoms 2016, 4, 31 26 of 36 3.5.4s 2 4p 4 6d and 4s 2 4p 4 7s Levels Based on our calculations, we were able to assign a number of low wavelength lines with clear Y V character as transitions to the ground term from levels of 4p 4 6d and 7s.For pairs of lines with wave number differences that closely match the 4p 5 2 P interval, the implied levels are relatively certain.However, the designations are considered to be tentative.Where the levels are based on single transitions, the line and level identifications are even less certain.None of these levels were included in the least-squares-fits, described below.

4s 2 4p 4 4f and 4s4p 5 4d Configurations
Extensive efforts to find levels of these configurations were not successful.Levels of 4p 4 4f were given in [6], but it is almost certain that all of them are spurious.

Odd Parity Configurations
As in [11] "The observed configurations were interpreted theoretically by making least-squares fits of the energy parameters to the observed levels with the Cowan suite of codes, RCN (Hartree-Fock), RCG (energy matrix diagonalization), and RCE (least-squares parameter fitting) [12].The Hartree-Fock code was run in a relativistic mode (HFR) with a correlation term in the potential.Breit energies were not included.For the initial calculations the HFR values were scaled by factors of 0.85 for the direct electrostatic parameters F k , the exchange electrostatic parameters G k , and the configuration interaction parameters R k ."The odd configurations 4s 2 4p 5 , 4s 2 4p 4 5p, 4s 2 4p 4 4f, and 4s4p 5 4d were treated as a single group.
The Hartree-Fock and least-squares fitted parameters for the odd configurations are given in Table 4.For these calculations, the 4p 4 5p exchange electrostatic parameters, G 0 (4p5p) and G 2 (4p5p), were linked at their HFR ratio.The LSF/HFR ratio of 0.836 is satisfactory.The configuration interaction (CI) parameters for the 4s 2 4p 5 -4s 2 4p 4 5p interaction were held fixed at their scaled HFR values.All other CI parameters and parameters for 4s 2 4p 4 4f and 4s4p 5 4d were fixed at their scaled HFR values.The value of the effective interaction parameter α(4p4p) for the 4p 4 5p configuration was fixed at the value observed for the 4p 4 core of Y VI [7].In Table 4, only values for the observed configurations 4s 2 4p 5 and 4s 2 4p 4 5p are given.The calculated level values and eigenvector compositions for the odd configurations are given in Table 5.This table gives the percentage compositions for the three leading eigenvector states in LS-coupling and the percentage for the leading eigenvector state in J 1 l-coupling.As can be seen there is not much mixing between the 4s 2 4p 5 and the 4s 2 4p 4 5p configurations, and 4s 2 4p 4 5p has essentially no mixture of either 4s 2 4p 4 4f or 4s4p 5 4d.

Even Parity Configurations
The parameters for the even configurations are given in Table 6.Here, the 4s4p 6 , 4p 4 4d, 5s, 5d, 6s, 6d, and 7s configurations were treated as single group.For the initial calculations the HFR values were scaled by factors of 0.85 for the direct electrostatic parameters F k , the exchange electrostatic parameters G k , and the configuration interaction parameters R k .All the parameters that were allowed to vary were well defined in the fit and have reasonable ratios to the HFR values.The exchange parameters G 1 (4p5d) and G 3 (4p5d) were linked at their HFR ratio.The CI parameters for the 4s4p 6 -4s 2 4p 4 4d and 4s4p 6 -4s 2 4p 4 5d interactions were also linked at their HFR ratio.The fitted values are reasonable.The other CI parameters and all of the parameters for 4p 4 6d and 4p 4 7s were held fixed at their scaled HFR values.As described in [4] the interaction of 4s4p 6 2 S 1/2 with the 4s 2 4p 4 4d ( 1 D) 2 S level is great, with a mutual repulsion of ~31,000 cm −1 .On the other hand, interaction between 4s4p 6 and 4s 2 4p 4 5d is negligible.The value of the effective interaction parameter α(4p4p) for the 4p 4 4d, 5s, 5d, and 6s configurations was again fixed at the value observed for the 4p 4 core of Y VI [7].The calculated level values and eigenvector compositions for the even levels are given in Table 7.This table gives the percentage compositions for the three leading eigenvector states in LS-coupling and the percentage for the leading eigenvector state in J 1 l-coupling, where appropriate.As can be seen, the purity of the states of the 4p 4 4d configuration in LS-coupling is low, leading to low leading percentages for many of the levels.Even though the 4p 4 5d and 4p 4 6s configurations are practically coincident, there is not much mixing of states.

4s4p 6 -4s 2 4p 4 5p Transitions
Transitions between the 4s4p 6 and 4s 2 4p 4 5p configurations are normally forbidden as two electron jumps.However, because of configuration interaction between 4s4p 6 and 4s 2 4p 4 4d, they can in fact take place.We observe six of them in Y V.The wavelengths for these transitions are long relative to the resonance lines and serve to improve the accuracy of the excited levels.

Ritz Wavelengths
We determined Ritz wavelengths for all of the lines by differencing the energy level values in Tables 2 and 3.The Ritz wavelengths are given in Table 1.The uncertainties of the calculated wavelengths correspond to the square root of the sum of the squares of the uncertainties of the combining levels.The Ritz values have uncertainties that are as low as ±0.0004Å.Those lines with uncertainties in the Ritz wavelengths of ±0.0020 Å or less should serve well as wavelength standards in the deep VUV.

Oscillator Strengths
Table 1 lists the transition probabilities g U A and log g L f for each observed line as calculated with wavefunctions obtained from the fitted energy parameters.Here, f is the oscillator strength, g U is the statistical weight of the upper level 2J U +1 and g L is the statistical weight of the lower level 2J L +1.The A-values are compared with recently published ab initio values in Section 9 below.
Since there are no experimental values for the transition probabilities of Y V, it is difficult to estimate the uncertainty of the calculated values.One guide is the cancellation factor.This is the ratio of the calculated transition probability to a value calculated with all parts of the wave function taken as positive [12].Low cancellation factors generally indicate a larger uncertainty in the calculated values.Indeed, many of the values in Table 1 have low cancellation factors.The present calculated transition probabilities can be considered as qualitative estimates of the relative intensities of the lines.Based on general experience, we estimate the uncertainties to be about ±50%.

Ionization Energy
An ionization energy of 605,000 ± 4000 cm −1 was obtained in [4] by estimating a value for n*(4p 4 5s) of 2.98 ± 0.02.On the basis of their observed 4s 2 4p 4 ns(n = 5-7) and nd(n = 4-6) series, Zahid-Ali, Chaghtai, and Singh [5] revised this downward slightly to 604,700 ± 2500 cm −1 .Since many of the levels used in their determination are now known to be spurious, this value must be re-determined.
For our new determination, we use the centers-of-gravity of the 4p 4 5s and 4p 4 6s configurations together with an estimated value for the change in effective quantum number ∆n*(4p 4 6s-4p 4 5s) = n*(4p 4 6s)-n*(4p 4 5s).This allows us to find the limit of the 4p 4 ns series, which is the center-of-gravity of the 4p 4 configuration of Y VI.

Comparison with ab Initio Calculations
Recently, two sets of ab initio calculations for the levels and oscillator strengths of Y V have appeared.Singh et al. [17] used a multiconfiguration Dirac-Fock (MCDF) approach to make calculations for transitions within the n = 4 complex; 4s 2 4p 5 , 4s4p 6 , 4s 2 4p 4 4d.Aggarwal and Keenan [18] used the General-purpose Relativistic Atomic Structure Package (GRASP) for calculations within the same complex of n = 4 configurations.Both calculations are based on new versions of the Grant atomic structure code.Froese Fischer [19] has discussed the accuracy that might be expected from calculations for complex atoms with GRASP, in particular as applied to the Br-like ion W 39+ .
Comparisons of our present results with those of the ab initio calculations of [17,18] are given in Tables 8-10.The index numbers for the levels in these tables are those used in [17,18].The wavelengths for Aggarwal and Keenan [18] in Table 8 are differences of the GRASP3 energies in their Table 3.It should be noted that the level with index 25 in [17] is misprinted 4s 2 4p 4 ( 1 D)4d 2 P 3/2 ; it should be 4s 2 4p 4 ( 1 S)4d 2 D 3/2 , as given in [18].
The main difference between the results of [17,18] and our present results is that the energies of the levels designated 4s 2 4p 4 ( 3 P)4d 2 P 1/2 (index 28) and 4s 2 4p 4 ( 1 D)4d 2 S 1/2 (index 30) are reversed in order of energy.That is, the level with index 28 corresponds to our level 4d51, and the level with index 30 corresponds to our level 4d41.
That our present order is correct can be seen from the fact that ( 3 P) 2 P has little interaction with 4s4p 6 2 S, and its position is largely fixed by the internal parameters of 4p 4 4d.If omitted from the LSF calculation, the calculated energy is very close to the observed value.So, there is no doubt about this assignment.This leaves the level at 294,965 cm −1 as the only possibility for ( 1 D) 2 S. The position of ( 1 D) 2 S is harder to pin down, because it is affected not only by the internal parameters of 4p 4 4d, but also by the amount of its upward displacement due to interaction with 4s4p 6 2 S. In our present calculations this uncertainty is removed, because when the level is included in the LSF, the CI parameter R 1 (4p4p,4s4d) takes a fitted value that has a reasonable ratio to HFR.This conclusion is supported by the observed line intensities, which follow the predicted pattern for these two levels.See for example the lines at 339.023, 353.976, 330.398, and 344.583Å in Table 8.It is clear that in the MCDF calculations the upward displacement of 4s 2 4p 4 ( 1 D)4d 2 S 1/2 due to interaction with 4s4p 6 2 S 1/2 is a little too large.The LSF/HFR scale factor of 0.769 for this interaction in Table 6 also reflects this circumstance.
In Table 8 we compare the wavelengths and transition probabilities A (s −1 ) found from GRASP with our present results.The values of A(present) in this table are those given in Table 1 divided by the statistical weight of the upper level 2J u +1.A notable disagreement for the transition probabilities for the 4s 2 4p 5 2 P 3/2 -4s 2 4p 4 4d ( 3 P) 4 F 3/2 transition (indices 1-12), observed at 419.792 Å.Both Singh et al. [17] and Aggarwal and Keenan [18] find an extremely low transition probability for this transition.However, we obtain a somewhat higher A-value, and it is indeed observed as a reasonably strong line.This transition is nominally forbidden as an inter-combination line in LS-coupling because of the change of spin.However, although the 4p 4 4d level (238,215 cm −1 observed value) has a leading percentage composition in LS coupling of 63% 4p 4 4d ( 3 P) 4 F 3/2 , the full percentage compositions show that it actually has a total doublet character of about 31%.This accounts for our calculated transition probability and observed line strength.Singh et al. [17] report a composition of 88% 4p 4 4d ( 3 P) 4 F 3/2 for this level, with no secondary percentage mentioned.Percentage compositions were not reported by Aggarwal and Keenan [18].The present percentage compositions for Y V are practically the same as were given in [4].This paper was not cited in either [17] or [18].
Other striking differences can be seen in Table 8.The values found by all three calculations for the 4s 2 4p 5 2 P 3/2 -4s 2 4p 4 4d( 1 S) 4 D 5/2 transition (indices 1-26) are extremely discrepant.The value of Aggarwal [18] is a little closer to our present value.The values for the 4s 2 4p 5 2 P 1/2 -4s 2 4p 4 4d ( 3 P) 4 D 3/2 transition (indices 2-6) also disagree by a large amount.Still, they all predict that this will be a very weak line, and in fact it has not been observed.Table 8.Comparison of wavelengths λ (Å) and transition probabilities A (s −1 ) for Y V calculated with the MCDF2 method of Singh et al. [17] and the GRASP3 method of Aggarwal and Keenan [18] with present values.Index numbers are those used in [17,18].Blank spaces indicate that line was not observed.Designations are for the upper levels in the transition.Both Singh et al. [17] and Aggarwal and Keenan [18] compare their calculated level values with the observed values given in the NIST Atomic Spectra Database [20].Since we have made a number of revisions to the 4p 4d levels, a new comparison is called for.This is given in Table 9.

Lower
Table 9.Comparison of level energies E (cm −1 ) for Y V calculated with the MCDF2 method of Singh et al. [17] and the GRASP3 method of Aggarwal and Keenan [18] with present experimental energies.Index numbers are those used in [17,18].

aa
Designations are explained in Table2; 4p61 indicates the J = 1/2 level of 4s4p 6 ; b Tentative designation; not included in LSF; c Tentative level with tentative designation; not included in LSF.Designations are explained in Table2; 4p61 indicates the J = 1/2 level of 4s4p 6 ; b Tentative designation; not included in LSF; c Tentative level with tentative designation; not included in LSF.

Figure 1 .
Figure 1.Schematic overview of the observed configurations of Y V.The calculated positions of the 4s 2 4p 4 4f and 4s4p 6 4d configurations are also shown.

Figure 1 .
Figure 1.Schematic overview of the observed configurations of Y V.The calculated positions of the 4s 2 4p 4 4f and 4s4p 6 4d configurations are also shown.

Figure 1 .
Figure 1.Schematic overview of the observed configurations of Y V.The calculated positions of the 4s 2 4p 4 4f and 4s4p 6 4d configurations are also shown.

Table 1 .
Observed spectral lines of Y V. Wavelengths and wave numbers are in vacuum.Wavelength values in parentheses are Ritz values.General uncertainty of the observed wavelengths is ±0.007Å. Uncertainties for less certain wavelengths are given in Section 2 of the text.|CF| is the cancellation factor (see text).Unc (Å) is the uncertainty of the Ritz wavelength.
a Symbols: dc, doubly classified; p, perturbed; u, unresolved from close line; s, shaded to shorter wavelength; l, shaded to longer wavelength; x, not included in level optimization; d, double line; c, complex.A, blended or obscured by Y VI; B, perturbed by O II; C, perturbed by Si IV; D, intensity much higher than expected; E, perturbed by second order line; F, perturbed ghost of Si IV line; G, perturbed by Si IV; H, uncertain stage of ionization; J, perturbed by Y III; K, perturbed by Si II; L, perturbed by C I; M, perturbed by unknown impurity; b Level codes are explained in Table2.

Table 3 .
Even parity energy levels (cm −1 ) of Y V.

Table 4 .
[7]tree-Fock and least-squares fitted parameters for the odd configurations of Y V. Mean error of fit 179 cm −1 .Fixed at value from 4p 4 of Y VI[7]; b Linked in LSF fit; c Fixed at scaled HFR value. a

Table 5 .
Calculated energy levels (cm −1 ) and percentage compositions for the odd levels of Y V.

Table 6 .
[7]tree-Fock and least-squares fitted parameters for the even configurations of Y V. Mean error of fit 273 cm −1 .Fixed at value from 4p 4 of Y VI[7]; b,c Linked in groups in LSF fit; d Fixed at scaled HFR value. a