1: \documentclass[aps,pra,floatfix,twocolumn,showpacs]{revtex4}
2: \topmargin 0.0in
3: \usepackage{epsfig}
4: \usepackage{graphicx}
5: \usepackage{dcolumn}
6: %documentclass[aps,pra,showpacs,twocolumn,nofootinbib]{revtex4}
7: %documentclass[aps,pra,showpacs,twocolumn]{revtex4}
8: %usepackage{graphicx}
9: %usepackage{dcolumn}
10: \begin{document}
11: \title{Fine Structure Anomalies and Search for Variation of the Fine
12: Structure Constant in Laboratory Experiments}
13: \author{V. A. Dzuba$^1$}
14: \email{V.Dzuba@unsw.edu.au}
15: \author{V. V. Flambaum$^1$$,$$^{2}$}
16: \email{V.Flambaum@unsw.edu.au}
17: \affiliation{$^1$School of Physics, University of New South Wales, Sydney 2052,
18: Australia\\
19: $^2$ Physics Division, Argonne National Laboratory, Argonne, Illinois 60439-4843, USA }
20:
21: \date{\today}
22:
23:
24: \begin{abstract}
25:
26: Configuration interaction in many-electron atoms may cause anomalies in
27: the fine structure which make the intervals
28: small and very sensitive
29: to variation of the fine structure constant. Repeated precision
30: measurements of these intervals over long period of time can put strong
31: constrain on possible time variation of the fine structure constant.
32: We consider the $5p^4 \ ^3$P$_{2,1,0}$ fine structure multiplet in the
33: ground state of neutral tellurium as an example.
34: Here the effect of change of the fine structure constant is
35: enhanced about one hundred times in the relative change of the small
36: energy interval between the $^3$P$_1$ and $^3$P$_0$ states.
37:
38: \end{abstract}
39:
40: \pacs{PACS: 31.25.-v, 31.25.Eb, 31.25.Jf}
41:
42: \maketitle
43:
44: \section{introduction}
45:
46: The possibility of fundamental constants to vary is predicted by theories
47: unifying gravity with other interactions
48: (see, e.g. review by Uzan~\cite{uzan}). There is an evidence found in
49: quasar absorption spectra that the fine structure constant $\alpha$
50: ($\alpha = e^2/\hbar c$) might be smaller in early cosmological
51: epoch~\cite{1,2,3,4}. However, similar analysis applied by other authors to
52: different sets of data does not support this claim \cite{Srianand,Quast}.
53: Recent progress in improving atomic clocks and developing optical frequency
54: standards (see, e.g. \cite{clocks}) made it possible to put strong
55: constrains on possible time variation of fundamental constants in laboratory
56: measurements. In particular, constrains obtained for the fine structure
57: constant $\alpha$ need only about one order of magnitude improvement
58: to see whether they are consistent with the quasar absorption spectra
59: data if the same rate of change is assumed over cosmological time.
60: These constrains were obtained with the full use of the technique
61: developed for highly accurate optical standards. However, the
62: choice of atomic transitions was not optimal for searching of variation
63: of the fine structure constant. Frequencies used so far change at about
64: the same rate as $\alpha$ if $\alpha$ changes, i.e. there is no enhancement.
65:
66: An alternative approach was suggested in Refs.\cite{flambaum1,flambaum2}.
67: It was proposed to measure small frequency between two close states with
68: different dependence on $\alpha$.
69: Here small change of $\alpha$ may lead to
70: orders of magnitude larger
71: %///
72: relative
73: change in frequency. The enhancement
74: factor $k$
75: can be expressed as
76: %///
77: $k = 2 q_{12}/\omega$, where $ q_{12}$ is the
78: difference in relativistic energy shifts of two levels and $\omega$ is
79: transition frequency. A good candidate for such experiments is dysprosium
80: atom~\cite{flambaum1,flambaum2,budker}. It has two almost degenerate states
81: of opposite parity and corresponding enhancement factor is of order
82: of $10^{8}$~\cite{flambaum2}. The measurements for dysprosium are currently
83: underway at Berkeley~\cite{budker}.
84:
85: While it is
86: hard to find anything better than dysprosium in terms of
87: enhancement factor, it has its disadvantages too. The levels involved
88: are not so narrow. While one of the states lives long enough
89: to cause no problem the other level does not.
90: One has to work inside the line width
91: to have the desirable accuracy. This
92: %///
93: puts
94: certain limitations on achievable
95: constrains on $\alpha$-variation. Therefore, it would be important to find
96: something which combines the advantages of having metastable states and
97: strong enhancement. A number of such transitions were considered in
98: recent work~\cite{saveliy}. It has been suggested to look at close
99: metastable states of different configurations. Energies of different
100: configurations change at different rates when $\alpha$ is changing.
101: This, together with small energy intervals between the states ensures
102: strong enhancement~\cite{saveliy}.
103:
104: In present work we consider slightly different possibility. We suggest
105: to study anomalously small
106: fine structure intervals in ground configurations of many-electron atoms.
107: The use of the ground state fine structure multiplet ensures that states
108: are metastable. They can only decay via $M1$-transitions which are very
109: much suppressed due to small value of transition frequency. On the other
110: hand, configuration
111: interaction in many-electron atoms can
112: reduce fine structure intervals and lead to strong sensitivity to
113: the change in the fine structure constant. In next sections we consider
114: in detail fine structure of the ground state of tellurium and discuss
115: some other possibilities.
116:
117: \section{Theory and results for tellurium}
118:
119: To study sensitivity of atomic frequencies to variation of the fine
120: structure constant $\alpha$ it is convenient to present them in the
121: vicinity of the physical value of $\alpha$ ($\alpha = \alpha_0 = 1/137.036$)
122: in the form
123: \begin{equation}
124: \epsilon = \epsilon_0 + qx,
125: \label{omegaq}
126: \end{equation}
127: where $x=(\alpha/\alpha_0)^2-1$ and coefficient $q$ defines the sensitivity
128: of the frequency to variation of $\alpha$. In general, its value can be
129: found from atomic calculations. If $\alpha$ changes
130: %///
131: the relative
132: frequency changes at
133: the rate
134: \begin{equation}
135: \frac{\Delta \omega}{\omega} = \frac{2q_{12}}{\omega_0}
136: \frac{\Delta \alpha}{\alpha_0} \equiv k \frac{\Delta \alpha}{\alpha_0},
137: \label{kfact}
138: \end{equation}
139: where $k=2q_{12}/\omega_0$ is enhancement factor. To search for variation of
140: the fine structure constant one needs to compare atomic frequencies
141: with different enhancement factors over long period of time. The larger
142: this difference the more sensitive the experiment to variation of $\alpha$.
143: Note that $k$ can have negative value which means that changes of $\alpha$
144: and frequency go in opposite directions: frequency decreases when $\alpha$
145: increases and vise versa.
146:
147: For ``normal'' fine structure intervals which are proportional to
148: $(Z\alpha)^2$ formula (\ref{omegaq}) is valid for all values of $\alpha$,
149: $0 < \alpha < \alpha_0$. Therefore,
150: $q_{12}=\omega_0$ and $k=2$.
151: In other words, the factor $k$
152: is the same for all ``normal'' fine structure intervals
153: and comparison between them cannot reveal any drift of $\alpha$.
154: These fine structure intervals can still be used in search for variation
155: of
156: the fundamental constants
157: if they are compared to hyperfine structure or to frequencies
158: of suitable optical transitions.
159:
160: Situation changes dramatically if fine structure multiplet is strongly
161: perturbed by configuration interaction with neighboring states.
162: We consider neutral tellurium atom in its ground state as an example.
163:
164: Experimental and theoretical energies of the ground state $5p^4$
165: configuration of tellurium are presented in Table~\ref{tab1}.
166: There are strong anomalies in the fine structure multiplet $^3$P$_{2,1,0}$.
167: The $^3$P$_2$ - $^3$P$_1$ and $^3$P$_1$ - $^3$P$_0$ intervals have
168: opposite signs and differ in value more than hundred times. This is
169: because of configuration interaction between the $^3$P$_2$ and $^1$D$_2$
170: states and between the $^3$P$_0$ and $^1$S$_0$ states while the $^3$P$_1$
171: state has no close neighbors to mix with. Configuration interaction leads
172: to accidental almost exact cancellation between spin-orbit and Coulomb
173: terms in the energy interval between the $^3$P$_1$ and $^3$P$_0$ states.
174: Since spin-orbit
175: interaction
176: is much more sensitive to the change of $\alpha$ than
177: the Coulomb term, it is natural to expect that the $^3$P$_1$ - $^3$P$_0$
178: energy interval is very sensitive to the variation of $\alpha$.
179:
180: To check how strong is sensitivity of the fine structure intervals
181: of Te~I we perform model configuration interaction (CI) calculations
182: for the $5p^4$ configuration of the atom. First, we perform Hartree-Fock
183: calculations
184: for open shells to find the $5p_{1/2}$ and $5p_{3/2}$
185: states of neutral tellurium. Then we use the CI technique to construct
186: four-electron states of the $5p^4$ configuration and to calculate their
187: energies
188: %///
189: (in fact, in this approximation CI technique is reduced
190: to diagonalization of the interaction Hamiltonian describing
191: direct mixing of different $5p^4$ states).
192: %---------------------------------------
193: It turns out that some extra fitting is needed to have good
194: agreement with experiment. Namely, we reduce the value of the
195: $F_2(5p_{3/2},5p_{3/2})$ Coulomb integral by 25\%. This reduction
196: simulates the effect of screening of Coulomb interaction between
197: valence electrons by core electrons. The results for energies are
198: presented in Table~\ref{tab1}. One can see that in spite of simple
199: approximation used in calculations the agreement with experiment
200: is very good.
201:
202: Coefficients $q$ (see Eq.(\ref{omegaq})) are found by varying
203: $\alpha$ in computer codes:
204: \[ q=\frac{\epsilon(x=+0.1)-\epsilon(x=-0.1)}{0.2}. \]
205: The results for $q$ are also presented in Table~\ref{tab1}. Since
206: we have good agreement with experiment for the energies it is natural
207: to assume that the accuracy for the $q$-coefficients is also good.
208:
209: \begin{table}
210: \caption{\label{tab1}Experimental and theoretical energies and
211: $q$-coefficients (cm$^{-1}$) for the $5p^4$ ground-state configuration
212: of Te~I.}
213: \begin{ruledtabular}
214: \begin{tabular}{ccrrr}
215: State & $J$ & \multicolumn{1}{c}{$E$(exp)} & \multicolumn{1}{c}{$E$(theor)}
216: & \multicolumn{1}{c}{$q$} \\
217: \hline
218: $5p^4 \ \ ^3P$ & 2 & 0 & 0 & 0 \\
219: & 1 & 4751 & 4770 & 5927 \\
220: & 0 & 4707 & 4713 & 3594 \\
221: $5p^4 \ \ ^1D$ & 2 & 10559 & 10745 & 5207 \\
222: $5p^4 \ \ ^1S$ & 0 & 23199 & 22845 & 8571 \\
223: \end{tabular}
224: \end{ruledtabular}
225: \end{table}
226:
227: Table~\ref{tab2} presents frequencies and enhancement factors for transitions
228: between all states of the $^3$P$_{0,1,2}$ fine structure multiplet. The
229: enhancement for the $^3$P$_1 \ - \ ^3$P$_0$ is more than one hundred due
230: to anomalously small frequency of the transition.
231: Ratio of
232: this small frequency to almost any other atomic frequency
233: is extremely sensitive to variation of $\alpha$. However, other transitions
234: from Table~\ref{tab2} can also be used.
235:
236: \begin{table}
237: \caption{\label{tab2}Frequencies (in cm$^{-1}$) and enhancement factors
238: ($k=2\Delta q/\omega$)
239: for transitions within the $^3$P$_{0,1,2}$ fine structure multiplet of Te~I}.
240: \begin{ruledtabular}
241: \begin{tabular}{ccrr}
242: Transition & Type & \multicolumn{1}{c}{$\omega$} & \multicolumn{1}{c}{$k$} \\
243: \hline
244: $^3$P$_1 \ - \ ^3$P$_2$ & M1 & 4751 & 2.2 \\
245: $^3$P$_0 \ - \ ^3$P$_2$ & E2 & 4707 & 1.5 \\
246: $^3$P$_1 \ - \ ^3$P$_0$ & M1 & 44 & 106 \\
247: \end{tabular}
248: \end{ruledtabular}
249: \end{table}
250:
251: Because of very exotic behavior of the fine structure intervals of Te~I
252: it is very interesting to see what happens to them when $\alpha$ varies
253: from zero (non-relativistic limit) to its physical value $\alpha_0$.
254: We have performed such calculations and results for five low states of the
255: $5p^4$ ground-state configuration of Te~I are presented on
256: Fig.~\ref{fig}. Valence energies (energy to remove
257: all four valence electrons from atom) are shown as functions of
258: $(\alpha/\alpha_0)^2$. All three energies of the $^3$P$_{0,1,2}$
259: multiplet have the same value at $\alpha=0$ and fine structure intervals
260: are proportional to $\alpha^2$ at small $\alpha$. At larger values
261: of $\alpha$ interaction
262: with the $^1$D$_2$ and $^1$S$_0$ leads to significant perturbation of the
263: $^3$P multiplet. In particular, repulsion between the $^1$S$_0$ and
264: $^3$P$_0$ levels
265: %///
266: causes
267: the latter
268: to cross with the $^3$P$_1$ level
269: in the vicinity of the physical value of $\alpha$. This explains
270: anomalously small energy interval between the two states.
271:
272: \begin{figure}
273: \centering
274: \epsfig{figure=te1.eps,scale=0.45}
275: \caption{Energy levels of lower states of Te~I as functions of the fine
276: structure constant.}
277: \label{fig}
278: \end{figure}
279:
280: Let us mention in the end of this section that both upper states of the
281: $^3$P multiplet are metastable. The $^3$P$_1$ state decays mostly by M1
282: transition to the ground state. Its lifetime is about 1 sec.
283: The $^3$P$_0$ state decays via E2 transition to the ground state and
284: corresponding lifetime (if no other factors are involved) is about
285: $4 \times 10^3$ sec.
286:
287: \section{Polonium and cerium}
288:
289: There are many other examples of the anomalous fine structure in the
290: ground and low excited states of many-electron atoms which involve
291: metastable states and large enhancement and therefore suitable for
292: search of variation of the fine structure constant. The actual choice
293: between them would involve consideration of many other factors which
294: are important for experimentalists but not discussed in present paper.
295: Below we discuss two more examples just to have broader picture.
296:
297: Brief look at the spectra of elements presented in Moore's tables~\cite{moore}
298: reveals that practically all elements with the $np^4$ configuration in the
299: ground state have anomalies in the fine structure (though none of them has
300: so small energy interval as Te~I). The most interesting case after Te~I
301: is probably polonium (Po~I). It has the largest nuclear charge $Z$ ($Z=84$)
302: among the $np^4$ elements which mean strong relativistic effects and
303: large $q$ and $k$ coefficients (see formulas (\ref{omegaq}) and (\ref{kfact})).
304: Experimental and theoretical data similar to those considered for Te~I
305: are
306: presented in Tables \ref{tabpo} and \ref{kpo}. We see that enhancement
307: factors $k$ are large and different for different transitions. This is
308: exactly what is needed for the search of variation of the fine structure
309: constant.
310:
311: \begin{table}
312: \caption{\label{tabpo}Experimental and theoretical energies and
313: $q$-coefficients (cm$^{-1}$) for the $6p^4$ ground-state configuration
314: of Po~I.}
315: \begin{ruledtabular}
316: \begin{tabular}{ccrrr}
317: State & $J$ & \multicolumn{1}{c}{$E$(exp)} & \multicolumn{1}{c}{$E$(theor)}
318: & \multicolumn{1}{c}{$q$} \\
319: \hline
320: $6p^4 \ \ ^3P$ & 2 & 0 & 0 & 0 \\
321: & 1 & 16831 & 17248 & 26720 \\
322: & 0 & 7514 & 7397 & 1910 \\
323: $6p^4 \ \ ^1D$ & 2 & 21679 & 22189 & 25850 \\
324: $6p^4 \ \ ^1S$ & 0 & 42718 & 45101 & 54130 \\
325: \end{tabular}
326: \end{ruledtabular}
327: \end{table}
328:
329: \begin{table}
330: \caption{\label{kpo}Frequencies (in cm$^{-1}$) and enhancement factors
331: ($k=2\Delta q/\omega$)
332: for transitions within the $^3$P$_{0,1,2}$ fine structure multiplet of Po~I.}
333: \begin{ruledtabular}
334: \begin{tabular}{ccrr}
335: Transition & Type & \multicolumn{1}{c}{$\omega$} & \multicolumn{1}{c}{$k$} \\
336: \hline
337: $^3$P$_1 \ - \ ^3$P$_2$ & M1 & 16831 & 3.2 \\
338: $^3$P$_0 \ - \ ^3$P$_2$ & E2 & 7514 & 0.5 \\
339: $^3$P$_1 \ - \ ^3$P$_0$ & M1 & 9317 & 5.3 \\
340: \end{tabular}
341: \end{ruledtabular}
342: \end{table}
343:
344: Another interesting example is the positive ion of cerium (Ce~II).
345: An extract from the tables~\cite{martin} presenting experimental
346: energies and $g$-factors Lande of lower states of Ce~II are presented in
347: Table~\ref{tabce} together with non-relativistic (NR) values of the
348: $g$-factors. Non-relativistic $g$-factors were calculated according
349: to the formula
350: \begin{equation}
351: g = 1 + \frac{J(J+1)-L(L+1)+S(S+1)}{2J(J+1)}.
352: \label{gfactor}
353: \end{equation}
354:
355: The data presented in Table~\ref{tabce} reveals an interesting picture.
356: Fine structure multiplets of Ce~II intersect. There are many states with
357: the same total momentum $J$ within an energy interval spanned by singe
358: fine structure multiplet. For example, there are four(!) states of $J=4.5$
359: within energy interval of the lowest fine structure multiplet $^4$H$^o$.
360: It is clear that they must be strongly mixed. Fine structure intervals
361: do not obey
362: Lande's rule and experimental $g$-factors deviate significantly
363: from the non-relativistic values. All these suggest strong configuration
364: mixing and sensitivity of the intervals to variation of $\alpha$.
365: An interval of particular interest is the $^4$H$^o_{11/2} \ - \ ^4$H$^o_{9/2}$
366: one. It is small, only 299~cm$^{-1}$ and states involved are strongly
367: mixed with other close states. This ensures strong enhancement of the
368: change of $\alpha$ in the
369: %///
370: relative
371: change of frequency.
372:
373: \begin{table}
374: \caption{\label{tabce}Energies (cm$^{-1}$) and $g$-factors
375: of lower states of Ce~II.}
376: \begin{ruledtabular}
377: \begin{tabular}{cccrrr}
378: Config. & Term & $J$ & \multicolumn{1}{c}{$E$(exp)} &
379: \multicolumn{1}{c}{$g$(exp)} & \multicolumn{1}{c}{$g$(NR)} \\
380: \hline
381: $4f5d^2$ & $^4$H$^o$ & 3.5 & 0.000 & 0.794 & 0.667 \\
382: & & 4.5 & 2581.257 & 1.023 & 0.970 \\
383: & & 5.5 & 2879.695 & 1.123 & 1.133 \\
384: & & 6.5 & 4203.934 & 1.189 & 1.231 \\
385: & & & & & \\
386: $4f5d^2$ & & 4.5 & 987.611 & 0.948 & 1.000 \\
387: & & & & & \\
388: $4f5d^2$ & $^4$I$^o$ & 4.5 & 1410.304 & 0.856 & 0.727 \\
389: & & 5.5 & 2563.233 & 0.968 & 0.965 \\
390: & & 6.5 & 3793.634 & 1.128 & 1.107 \\
391: & & 7.5 & 5455.845 & 1.196 & 1.200 \\
392: & & & & & \\
393: $4f5d^2$ & & 3.5 & 1873.934 & 0.806 & 1.000 \\
394: & & & & & \\
395: $4f5d^2$ & & 0.5 & 2140.492 & 0.985 & 1.000 \\
396: & & & & & \\
397: $4f5d6s$ & & 4.5 & 2382.246 & 1.039 & 1.000 \\
398: \end{tabular}
399: \end{ruledtabular}
400: \end{table}
401:
402: Calculations for Ce~II are more difficult than for Te~I and Po~I because
403: of valence states of high angular momentum ($5d$ and $4f$).
404: Therefore we believe that it is premature to attempt them now.
405: The presence of enhancement is obvious but its actual value would
406: become important only on the stage of planning the measurements.
407: We are ready to perform the calculations if there is any interest
408: from experimentalists.
409:
410: Cases of fine structure anomalies similar to what is presented here for
411: Ce~II can be easily fond in spectra of many other rare-earth elements.
412: Which of them are suitable for the search of $\alpha$ variation is the
413: question which needs further consideration.
414:
415: \section{Conclusion}
416:
417: We present an alternative way to search for variation of the fine structure
418: constant in laboratory measurements. We propose to use fine structure
419: intervals in the ground or low excited states of many-electron atoms which
420: are strongly perturbed by configuration interaction with neighboring states.
421: This method has double advantages. On one
422: hand,
423: the use of low states
424: ensures that they are metastable. This is important for very accurate
425: frequency measurements. On the other
426: hand, strong perturbation may lead to anomalously small fine structure
427: interval and
428: strong enhancement of the relative
429: sensitivity of the frequencies to the change
430: of the fine structure constant.
431: Because of the high relative sensitivity one does need extremely accurate
432: absolute measurements of the frequencies (this is the difference
433: with conventional atomic clock measurements).
434: Large value of the effect/frequency ratio may also help to
435: reduce importance of some systematic effects (e.g. the Doppler shift
436: and broadening).
437: Note, however, that we do not consider in this paper any practical
438: measurement scheme.
439:
440: Enhanced and highly non-linear dependence of the small fine structure intervals
441: on the magnitude of the relativistic corrections
442: also presents certain theoretical interest.
443:
444: \section{Acknowledgments}
445:
446: This work was supported by the Australian Research Council and Department
447: of Energy, Office of Nuclear Physics, Contract No. W-31-109-ENG-38.
448:
449:
450: \begin{thebibliography}{20}
451:
452: \bibitem{uzan}
453: J-P. Uzan, Rev. Mod. Phys. {\bf 75}, 403 (2003).
454:
455: \bibitem{1} J. K. Webb , V. V. Flambaum, C. W. Churchill, M. J. Drinkwater,
456: and J. D. Barrow, Phys. Rev. Lett. {\bf 82}, 884 (1999).
457:
458: \bibitem{2} J. K. Webb,
459: M. T. Murphy, V. V. Flambaum, V. A. Dzuba, J. D. Barrow, C. W. Churchill,
460: J. X. Prochaska, and A. M. Wolfe, Phys. Rev. Lett.
461: {\bf 87}, 091301 (2001).
462:
463: \bibitem{3}
464: M. T. Murphy, J. K. Webb, V. V. Flambaum, V. A. Dzuba,C. W. Churchill,
465: J. X. Prochaska, J. D. Barrow, A. M. Wolfe,
466: Mon. Not. R. Astron. Soc. {\bf 327}, 1208 (2001).
467:
468: \bibitem{4}
469: M. T. Murphy, J. K. Webb, V. V. Flambaum,
470: Mon. Not. R. Astron. Soc. {\bf 345}, 609 (2003).
471:
472: \bibitem{Srianand}
473: R. Srianand, H. Chand, P. Petitjean, and B. Aracil, Phys. Rev. Lett.
474: {\bf 92}, 121302 (2004).
475:
476: \bibitem{Quast}
477: R. Quast, D. Reimers, and S. A. Levshakov,
478: Astron. Astrophysics {\bf 415}, L7 (2004).
479:
480: \bibitem{clocks} S. G. Karshenboim and E. Peik (Eds.).
481: {\em Astrophysics, Clocks and Fundamental Constants\/},
482: (Springer, Berlin, Heidelberg, 2004), Vol. 648.
483:
484: \bibitem{flambaum1} V. A. Dzuba, V. V. Flambaum, and J. K. Webb,
485: Phys. Rev. A {\bf 59}, 230 (1999).
486:
487: \bibitem{flambaum2} V. A. Dzuba, V. V. Flambaum, and M.V. Marchenko,
488: Phys. Rev. A {\bf 68}, 022506 (2003).
489:
490: \bibitem{budker} A. T. Nguyen, D. Budker, S. K. Lamoreaux, J. R. Torgerson,
491: Phys. Rev. A {\bf 69}, 022105 (2004).
492:
493: \bibitem{saveliy} S. G. Karshenboim, A. Nevsky, E. Angstmann, V. A. Dzuba,
494: V. V. Flambaum, to be published.
495:
496: \bibitem{moore} C. E. Moore, {\em Atomic Energy Levels}, Natl. Bur.
497: Stand. (U.S.), Circ. No. 467 (U.S. GPO, Washington, D. C., 1958),
498: Vols. 1-3.
499:
500: \bibitem{martin} W. C. Martin, R. Zalubas, and L. Hagan,
501: {\em Atomic Energy Levels, The Rare-Earth Elements},
502: Natl. Bur. Stand. (U.S.), (U.S. GPO, Washington, D. C., 1978).
503:
504: \end{thebibliography}
505:
506: \end{document}
507: