Carnall 3

Electronic Energy Levels in the Trivalent Lanthanide Aquo Ions. I.
Pr3+,
Nd3+, Pm3+, Sm3+, Dy3+, Ho3+, Er3+, and Tm3+
W. T. Carnall, P. R. Fields, and K. Rajnak
Citation: J. Chem. Phys. 49, 4424 (1968); doi: 10.1063/1.1669893

View online: http://dx.doi.org/10.1063/1.1669893
View Table of Contents: http://jcp.aip.org/resource/1/JCPSA6/v49/i10
Published by the American Institute of Physics.
Additional information on J. Chem. Phys.

Journal Homepage: http://jcp.aip.org/
Journal Information: http://jcp.aip.org/about/about_the_journal
Top downloads: http://jcp.aip.org/features/most_downloaded
Information for Authors: http://jcp.aip.org/authors
Downloaded 28 Mar 2013 to 129.174.21.5. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://jcp.aip.org/about/rights_and_permissions
T'HE JOURNAL OF CHEMICAL PHYSICS VOLUME 49, NUMBER 10 15 NOVEMBER 1968
Electronic Energy Levels in the Trivalent Lanthanide Aquo Ions.

.
I PrH, N d H, PmH, Sm H , DyH, HOH, Er3+, and Tm3 +*
W. T. CARNALL, P. R. FIELDS, AND K. RAJNAKt
Chemistry Division, Argonne National Laboratory, Argonne, Ininois
(Received 2 April 1968)
The free-ion energy-level schemes of the PtJ+, Nd3+, Pm3+, Sm3+, Dya+, Hoa+, EtJ+, and Tm3+ aquo ions
have been determined from their absorption spectra in dilute acid solution at 25°. Energy-level assignments
were made by comparison with crystal spectra, and on the basis of correlations between calculated and
observed band intensities. For most of the ions, it was possible to identify several transitions giving rise
to bands at energies as high as 45 000-50 000 cm-1 • Sufficient numbers of assignments were made to justify
inclusion of the effects of configuration interaction in the calculation of the energy-level parameters. Varia-
tion of the electrostatic, spin-orbit coupling, and configuration-interaction parameters across the lanthanide
series is examined.
In 1907, BecquereP first showed that the absorption trum of Nd3+ in Nd(Br03)s·9H20 by Satten in 19538
bands of crystalline lanthanide compounds could be was apparently the first attempt to analyze the states
resolved into sharp lines at low temperatures. He fur- resulting from more than two equivalent f electrons
ther reported that many of these lines exhibit a Zeeman assuming Russell-Saunders coupling. Subsequent stud-
effect. By 1929, Bethe2 had developed the theory for ies (for example see Refs. 9 and 10) have yielded an
term splitting which results from the influence of a intermediate coupling treatment of several of the low-
crystalline electric field of a given symmetry on an lying levels in the crystal spectra of practically all the
atom, but the gross term structure and the nature of lanthanides. The problem of extending these analyses
the sharp lanthanide absorption lines were not under- to higher energies is formidable because of the increase
stood. Thus, while there was considerable interest mani- in the density of levels to be characterized, and because
fest in the spectroscopic properties of the lanthanides the parameters used to predict the location of terms
in the early thirties,8 the interpretation was very lim- are normally not well enough determined to provide a
ited. reliable guide to making assignments.
Experimentally, it was extremely difficult to prepare Many of the transitions characterized in crystals
high-purity lanthanide compounds prior to the develop- containing lanthanide ions can be readily identified in
ment of the ion-exchange processes used today. Prandt14 solution spectra. However, until recently there was no
referred to his work with the rare earths over a period independent method of establishing the SLJ assign-
of 20 years, which finally led to the preparation of rela- ment of a band in solution. The success of the JuddlL
tively pure compounds of most of the members of the Ofeltl2 theory in deducing intensity relationships in
4fN series. These were used in 1934 to obtain the first lanthanide spectra13 ,14 suggested its application to this
complete (except for Pm8+) set of lanthani~e-solution problem. In the present communication we will discuss
absorption spectra in the range 2200-7000 A.4 Prandtl the use of calculated intensities in extending the analysis
expressed the hope that the results in solution would of lanthanide spectra both to higher energies and to
stimulate further theoretical interpretation. more complex systems than has heretofor been at-
A comparison between theory and experiment was tempted.
given for the spectrum of Tm2(S04)3·8H20 in 1937,6 On the basis of the numerous new assignments that
using the methods of Condon and Shortley.6 However have been made, it was possible for the first time to
from the standpoint of computation, the complete consider the second-order effects of configuration inter-
treatment of more complex configurations was not action in defining the energy-level schemes of the lan-
practical until the development of the tensor operator thanides.
methods of Racah. 7 Thus, the investigation of the spec- ENERGY-LEVEL CALCULATIONS
* Work performed under the auspices of the U.S. Atomic Since the theoretical treatment of lanthanide absorp-
Energy Commission. tion spectra has already been discussed in detail (for
t Present address: Kalamazoo College, Kalamozoo, Mich.
1 J. Becquerel, Radium 4, 328 (1907); 5, 5 (1908). J. Becquerel 8 R. A. Satten, J. Chero. Phys. 21, 637 (1953).
and H. K. Onnes, ibid. 5,227 (1908). »G. H. Dieke, Advances in Quantum Electronics, J. R. Singer,
2 H. Bethe, Ann. Physik 3, 133 (1929). Ed. (Columbia University Press, New York, 1961).
3 S. Freed, Rev. Mod. Phys. 14, 105 (1942). 10 G. H. Dieke and H. M. Crosswhite, Appl. Opt. 2,675 (1963).
'W. Prandtl and K. Scheiner, Z. Anorg. Algero. Chem. 220, 11 B. R. Judd, Phys. Rev. 127, 750 (1962).
12 G. S. Ofelt, J. Chem. Phys. 37,511 (1962).
107 (1934).
6 H A Bethe and F. H. Spedding, Phys. Rev. 12,454 (1937). 13 W. T. Carnall, P. R. Fields, and B. G. Wybourne, J. Chem.
8 E: U: Condon and G. H. ShortleYl Theory of Atomic Spectra Phys. 42, 3797 (1965).
(Cambridge University Press, Cambndge, 1957). If W. T. Carnall and P. R. Fields, Advan. Chem. Ser. 71, 86
7 G. Racah, Phys. Rev. 62, 438 (1942). (1967) .
4424
ENERGY LEVELS OF TRIVALENT LANTHANIDE AQUO IONS. I 4425
example see Refs. 15 and 16), only the general outline As Wybourne16 has pointed out, this approximation
of the calculations need be included here. happens to yield ratios of Slater integrals that are use-
The observed spectra of the trivalent lanthanides in ful in the calculations and does not imply anything
solution or in crystal lattices arise from "forbidden" about the actual shape of the eigenfunctions. It should
transitions within the ground 4jN configuration. I6 To a be noted that the splitting of the ground multiplet is
first approximation, the energy-level structure of the rather insensitive to the values of Fk • Initial estimates
4fN configuration arises from electrostatic and magnetic of Fk and t4! for all of the lanthanides have been tabu-
interactions between the 41 electrons. These interactions lated. l4 •16
may be represented by the Hamiltonian When the methods of Racah7 are used to calculate
the matrix elements of electrostatic interaction, opera-
X=X.+X.o, (1)
tors are constructed such that they have simple trans-
where X. involves the Coulomb interaction between formation properties with respect to the symmetry
pairs of 1 electrons and Xso takes into account the cou- groups used to classify the states. In this system the
pling of spin and orbital angular momenta which is the electrostatic energy becomes
most important magnetic interaction for 1 electrons.
Expansion of the electrostatic interactions in terms (6)
of Legendre polynomials allows separation of variables
and leads to the result
where the ek's are the angular parts of the new opera-
6
E.= 2:.PFk (k even). (2) tors and the Ek's are linear combinations of the Fk'S.I5
k=O It has been shown I7- 20 that the effects of configura-
In Eq. (2), the electrostatic energy (Ee) is expressed tion interaction must frequently be considered in order
in terms of products of Slater radial integrals, F k , and to obtain an adequate interpretation of the energy-level
angular coefficients, p. Similarly, the energy arising schemes of lanthanide ions. This can be accomplished
from spin-orbit interaction (E.o) may be written . in second-order perturbation theory by adding to the
Hamiltonian certain model interactions acting wholly
(3) within thefN configuration. 21 Such a procedure accounts
in which A.o represents the angular part of the spin- for that part of the interaction which does not have
orbit interaction and t4{ is a radial integral. The total the same angular form as the coefficients, ek, of the
energy in this approximation is, then, E.+ Eso. electrostatic-interaction parameters. The part that does
The angular terms are evaluated using Racah's ten- have the same form as ek is automatically absorbed in
sor operator formalism,1 and an intermediate coupling the Ek parameters.
scheme with Russell-Saunders (SIJ) basis states. In ForjN configurations, the additional interactions may
practice the radial integrals are treated as parameters be written as
to be determined from the experimental data. Neglect-
X' =aL(L+ 1) +/1G(G.!) +yG(R7 )
ing the effect of fOFo, which is only a scaling factor,
there remain three parameters associated with electro- +three-particle terms, (7)
static interactions, F2, F4, F 6 , and one parameter, t4{'
associated with spin-orbit interactions. There exist in where a, /1, and 'Yare linear combinations of radial
the literature very few cases where all four parameters integrals and are treated as adjustable parameters.
G(G2 ) and G(R 7 ) are eigenvalues of Casimir's opera-
were derived by a least-squares fit to energy levels for
which J assignments have been made. Further approxi- tor for the groups G2 and R 7 , respectively, and have
mations are usually necessary, since to adequately been tabulated by Wybourne. 16 The effects of the three-
determine all the parameters simultaneously requires particle terms are smaller than those of a, /1, and 'Y,
assignments to levels within a number of different and the accuracy of the present data does not warrant
multiplets. In many instances only levels in the ground their consideration here. I8 For some ions, the param-
multiplet and the first two or three excited multiplets eters a and /1 have been determined previously (the 'Y
have been clearly identified. One frequently used ap- used by Trees22 is 1/12 of the /1 used here) but crystal
proximation assumes that the 41-radial wavefunction studies frequently do not allow assignment of a suffi-
is hydrogenic. This makes it possible to calculate ratios cient number of levels for a least-squares fit of six
of the Slater integrals16 ; parameters (E1, E2, £3, tv, a, and j3). 'YG(R7) is re-
F 4/F 2 =O.138, (4) 17 K. Rajnak and B. G. Wybourne, J. Chem. Phys. 41, 565
(1964) .
F 6 /F 2 =O.0151. (5) 18 K. Rajnak, J. Chem. Phys. 43, 847 (1965).
111 J. Sugar. Phys. Rev. 14, 731 (1965); H. M. Crosswhite
16 J. P. Elliott, B. R. Judd, and W. A. Runciman, Proc. Roy. G. H. Dieke, and W. J. Carter, ]. Chem. Phys. 43, 2047 (1965):
Soc. (London) A240, 509 (1957). 20 K. Ra~nak and W. F, Krupke, J. Chem. Phys. 46, 3532 (1967).
16 B. G. Wyboume, Spectroscopic Properties of Rare Earths 21 K. RaJnak and B. G. Wyboume, Phys. Rev, 132, 280 (1963).
(John Wiley & Sons, New York, 1965). 22 R. E. Trees, J. Opt. Soc. Am, 54, 651 (1964).
4426 CARNALL, FIELDS, AND RAJNAK
lated to the correction (3Q used by Racah23 and Trees. 22 to small variations in Fk and t4j. Thus, it appeared
'Y is expected to be important only in cases where large reasonable to assume that changes in these param-
components in the eigenvectors of the observed levels eters which were occasioned by progressively improv-
have a seniority number different from that of the ing the energy-level fit would produce only minor
ground state. An attempt to determine 'Y has previ- variations in the magnitude of the matrix elements
ously been possible only for the free-ion spectrum of of U(lI). A single set of three parameters, :lA, is required
PrH.24 by Judd's theory to obtain all the calculated intensities
It should be noted that 'Y and El are not independ- for a given lanthanide ion in a given medium. 13 The
ent; they are related by a common term S(S+1) in intensity of a particular transition is proportional to
their coefficients. Thus, changes in El are expected L:A :lAu(1/;J II V().) 111/;'1')2. Thus, large matrix elements
when 'Y is added to the calculation. of U(A) for a transition imply a large calculated in-
tensity.
ENERGY-LEVEL FITTING PROCEDURE It was found that throughout the visible and ultra-
violet regions of the spectrum (to 50000 em-I), there
In the trivalent lanthanides, the free ion is only were numerous instances where single transitions with
slightly perturbed by the electric field of the crystal large matrix elements of U(4) and/or U(6) occurred near
or solution matrix in which it is observed. Each free- the same energy as a number of transitions with very
ion level, characterized by the total angular momentum small matrix elements of U(A). In many cases it was
quantum number, J, is split by this electric field into clear that the theory could account for the experimen-
a manifold of not more than 2J 1 components. The + tal data only if the transition with large matrix ele-
splitting does not usually encompass more than ,...,200- ments were assigned to a relatively intense absorption
300 em-I. band. Thus, the intensity calculations, based upon ap-
In crystal spectra, Zeeman and polarization studies proximate values of Fk and S4f> were used to make
have made it possible to characterize the J manifolds additional assignments to the observed spectrum. In-
to which many of the lower-lying crystal levels belong. tensity relationships then became a tool for improving
At higher energies this becomes more difficult. There the energy-level parameters.
is an overlapping of components of many different J The fitting procedure adopted here consisted of first
levels, and unambiguous assignments are not possible. making energy-level assignments to the centers of those
In solution spectra the crystal-field components are absorption bands, which were clearly characterized by
usually not resolved, so the observed absorption bands comparison with data obtained in the analysis of crys-
can only be characterized in terms of the free-ion levels tal spectra. The parameters Ek and f4! were determined
involved. At lower energies the different J manifolds by a least-squares process. 13 The resulting eigenvectors
are frequently isolated and the energy of the transition were then characteristic of the levels observed in solu-
may be assigned to the center of the observed band. tion. These eigenvectors were used to calculate the
The assignments at this point are based on compari- matrix elements of U().) for all transitions to 50000
son with crystal spectra. These led to an initial estimate em-I. Several additional assignments were obvious at
of the values of the parameters Fk and S4! for all of the this point for each lanthanide considered. The fitting
lanthanide aquo ions. 14 However, the limited usefulness process employing all assigned levels was then repeated
of these first parameters should be emphasized. In the to determine new values of Ek and f4!.
lanthanides, the lower levels are usually rather pure In most cases, new assignments were made at only
L-S states, whose composition does not change appre- slightly higher energies than those involved in previous
ciably with small changes in Fk and S4!. It might be assignments, and the calculations were repeated a num-
anticipated that estimates of these parameters based ber of times. Thus, there were no abrupt changes in
only on relatively pure low-lying levels would yield Ek and f4!. Actually, as would be predicted, the value
poor predictions for the energies of higher states which of S4! was relatively well defined early in the fitting
are not as pure. procedure.
Based upon approximate Fk and S4f parameters we For Pr3+, Nd3+, Pm3+, Ho3+, Er3+, and Tm3+, the
showed13 •14 that the Judd-Ofelt theory was able to cor- number of assignments which could be readily made
relate the intensities of the solution absorption bands was sufficient to justify inclusion in the fit of the
observed in the near-infrared and a limited portion of configuration-interaction parameters, a and {3, as soon
the visible region of the spectrum, for all of the lan- as the first two or three calculations were completed.
thanides from Pr3-I- to Yb3+. We noted that the values Some higher energy levels were assigned subsequently
of the calculated matrix elements of U(lI), as discussed such that 'Y could be defined. In the spectra of Sm3+
in the accompanying paper/Ii were rather insensitive and Dy3+, some of the bands in the visible range were
found to be strongly dependent upon 'Y. Failure to
G. Racah, Phys. Rev. 85, 381 (1952).
23 recognize this fact, by including assignments for these
K. Rajnak, J. Opt. Soc. Am. 55, 126 (1965).
24
levels in a six-parameter fit, led to poor agreement
W. T. Carnall, P. R. Fields, and K. Rajnak, J. Chern. Phys.
26
49,4412 (1968). between observed and calculated energies over the
ENE R G Y LEV E L S 0 F T R I V ALE N T LAN T HAN IDE A QUO ION S. I 4427
whole spectrum. Exclusion of these strongly /'-depend- TABLE~. Comparison .between the centers of gravity of levels
for Nd3+ m LaC]' and m LaF. crystals observed at liquid-He
ent levels in a six-parameter fit resulted in a good fit temperature and at room temperature.
for all other levels, but a difference of as much as
300-400 cm-I between calculated and observed ener- Nd3+ in LaC], Nd3+in LaFa
gies of /'-dependent levels. ,,(ern-1) ,,(ern-1)
S'L'J' (liquid He)a ,,(ern-1) b (liquid He) 0 ,,(ern-1) b
ln1iuence of Stark Components in the Ground
State on Bandwidth 41111. 4010 3 937 4089 4000
'116/'
At room temperature, it can usually be assumed 'F'/2 11 439 11 441 11 613 11600
'Fi/i 12 466 12 422 12 610}
that a small percentage of the transitions may arise 12 610
even from the highest Stark component of the ground 'H 9!1 12 618 12 707
level. One would, therefore, predict that a band ob- 'Fm 13 437}
13 386 13646 13 586
served at room temperature would be appreciably 'SIll 13 524
broadened by the presence of additional components 'Fm 14 722 14577 14 849 14 792
on the low-energy side compared to the band observed 'Huts 15 935 15 847 16 050 16025
'Gi/2 17 135 17094 17 328d }
at low temperatures. Thus, by assigning to the transi- 17 391
tion in question an energy corresponding to the center !GT/I 17 551 d
'KlIII
of such a broadened band, an error, always in the 'GT/. 19 020 18 939 19 239 19 230
?irection of too Iowan energy, would result. Actually, 'Glli 19434 19 379 19 700 b 19 607
It can be demonstrated that the indicated effect is IKl 6/, 20 876
IGB/. 21 056 21 141
small. (tD,IF).!1 21 129}
Since the energies of the crystal-field components of 21 459
many of the levels for Nds+ in LaCIa 26 and in LaFs 27 'Gtus 21 426
sPl/1 23 214 23239 23468 23474
have already been recorded at low temperature, we IDi/1 23 780 23 781
examined the spectra of similar crystals at room tem-
a Reference 26.
perature. The spectra were recorded with the same b Present measurement for center of band observed at room temperature.
instrument that was used in the other work reported o Reference 27.
here. 26 Assignments of energies were made on the same d Approximate center of gravity calculated from incomplete data.
basis as was used for the solution spectra. The results
obtained are given in Table I. It was found that in each case, the literature was searched for the results
general the satellites to a band arising from transitions of investigations in crystal media, and a single com-
parison standard for the solution spectrum was chosen.
which originate in one of the higher Stark components
of the ground level are relatively weak and do not The most common crystal host employed by other
appreciably affect the center of the band observed at investigators f~r such studies has been LaCIa. It happens
room temperature as compared to that established at that the free-IOn energy levels of the lanthanides in
this medium generally correspond quite closely to the
liquid-He temperature. In some cases, the satellites
are well resolved from the principal band and can be centers of bands observed in solution.
clearly identified. For example, the incompletely re- The energy of the ground state in solution was
solved group of bands centered at 11 441 cm-I for Nd3+ equated to the average energy of the ground-state
in LaCIa could be assigned to transitions arising from Stark components obtained from crystal spectra. Simi-
the lowest Stark component in the ground level and larly, where transitions to higher levels of the ground-
terminating in 4FS/2 level. Two isolated satellites, one term multiplet have been observed in a reference crys-
centered at 11 322 cm-I and a second weaker band at tal but not in solution, the assumption was made that
11198 em-I, were assigned to transitions from known they would be observed in solution at essentially the
Stark components of the ground Ievel26 at 115 and same energy.
244 em-I, respectively. We conclude that except in The experimental conditions and techniques of our
cases of resolved satellites, the process of assigning own work, which comprises all of the solution results
energies based on centers of bands observed at room reported here, were identical to those described in Refs.
t~mperature does not appear to be subject to any sig-
13, 14, and 25. Thus, the deuterated and normal aque-
mficant error. ous solutions employed made possible measurements
in the range ",6000-50000 em-I. References are cited
DATA ANALYSIS AND RESULTS for a few additional bands in the range ",4000-6000
cm-1 which we have observed for lanthanide nitrates
In this section we will discuss some of the individual dissolved in molten LiN03-KNOs eutectic. 21HO We have
assignments made to the lanthanides considered. In
28 W. T. Carnall, D. M. Gruen, and R. L. McBeth J. Phys.
26 E. H. Carlson and G. H. Dieke, J. Chern. Phys. 34, 1602 Chern. 66, 2159 (1962). '
(1?r61); F. Varsanyi and G. H. Dieke, ibid. 33, 1616 (1960). : w. T. Carnail, J. Phys. Chern. 67, 1206 (1963).
H. H. Caspers, H. E. Rast, and R. A. Buchanan, J. Chern. W. T. Carnall, P. R. Fields, and G. E. Toogood J. Phys.
Phys. 42, 3214 (1965). Chern. 68, 2351 (1964). '
~
N
00
TABLE II. Energy-level parameter values calculated for the trivalent lanthanides in various media (in cm-1).
No of
Matrix E1 E2 E3 t4/ a (3 'Y levels fit
Pr3+
r=ion.
LaClsb
LaFsa
4864.6±5.5
4713.8±5.0
4548.1±7.2
23. 138±O .15
21.890±O.1O
21. 659±O.16
488 . l1±O. 97
464. 23±O. 73
470.02±O.88
758.82±6.9
742. 68±4,4
743.24±6.0
23.684±1.2
22.899±O.87
18.642±O.88
-585.41±72
-676.97±46
-754.20±55
727. 78±45
599.51±42
1396.2±60
13
13
13
(J
aq 4548.2±7.5 21. 937±O.15 466. 73±O.96 740. 75±8. 7 21.255±1.2 -799.94±72 1342.9±62 13
;..-
879.42±5.7 -0.8174±1.2
t
aclsd 4974.6±13 23. 734±O.22 478.03±1.1 -163.94±53 22 :;d
NdH Z
aq 4739.3±16 23.999±O.16 485.96±1.0 884.58±6.3 O.5611±1.4 -117. 15±56 1321.3±84 32 ;..-
Pm8+ aq 4921.6±62 24.522±O.74 525.53±1.6 1000.8±7.4 10.991±2.5 -244.88±70 789. 74±370 22 r
r
5594.9±28 27.365±0.57 545.50±3.4 1162.5±15 16.192±2.8 -54O.36±176
t
aclso 20
SmH
aq 5496.9±42 25.809±0.14 556. 4O± 1. 7 1157.3±5.8 22.250±1.3 -742.55±53 796. 64±216 48 ...."'i
trJ
aClaf 6411.2±19 28.544±0.25 603.84±2.0 1923,4±8.0 38.661±2.3 -1184.7±89 16
t
r
DyH t::1
aq 6119.6±58 3O.O12±0.20 610. 14±2.2 1932.0±7.8 37.062±l.9 -1139.1±74 2395.3±300 34 en
6520.2±104 31.438±0.52 620.76±2.1 2138.3±9.2 23.490±1.6 -803.30±63 887.56±516 23
t
aClall
HOH ;..-
aq 6440.6±35 30.220±0.18 624.39±O.95 2141.3±4.9 23.635±O.88 -807.20±36 1278.4±190 41 Z
t::1
6855.3±17 32 . 126±O .22 645.57±2.4 2369,4±11 20.385±2.6 -666.60±107 27
Er3+ r=w.·
LaFa l
LaClai
aq
6884.4±25
6885.3±22
6769.9±130
32.586±O.21
32. 272±0. 24
32.388±0.14
649.64±2.0
64l.14±2.6
646.62±1.2
2380.0±9.7
2379.5±11
2380.7±8.3
17.044±2.4
17.431±2.2
18.347±l.2
-527.03±83
-655. 28±86
-509. 28±58 649. 71±640
21
20
27
:;d
;..-
......
Z
;..-
{EthYl sulfate k 7150.8±92 33. 886±O. 62 675.73±5.2 2628.1±28 14.688±6.4 -74l.59±333 12
Tm8+ I"'l
aq 7142.4±60 33. 795±0.41 674.27±3.4 2628.7±18 14.677±4.2 -631. 79±216 12
& Reference 19. IIReference 20; G. H. Dieke and B. Pandy. J. Chern. Phy•• 41, 1952 (1964).
b Reference 32. b Reference 36.
• Reference 33. I Reference 37.
d Reference 26. i Reference. 17 and 38.
• Reference 35. k References 39--42.
tH. M. Cl'OISWhite and G. H. Dieke. J. Chern. Phys. 35,1535 (1961); J. D. A1ce and G. H. Dieke,
iIM. 37, 2364 (1962).
ENERGY LEVELS OF TRIVALENT LANTHANIDE AQUO ION S. I 4429
TABLE III. Energy-level assignments for Pt'+.
Pt'+ Free ion a Pt'+ in LaClab Pt'+in LaF.·

E.xptl Eeal.d AE E..ptl Ecalod AE Eexp\l E •• I.d AE
S'L']' (em-I) (em-I) (em-I) (em-I) (em-I) (em-I) (em-I) (em-I) (em-I)
3H. 0 20 -20 99.3 118 . -9 200 228 -28

3H, 2 152.2 2 147 5 2209 2200 9 2363 2 309 54
3H, 4389.1 4 375 14 4390 4380 10 4464 4490 -26
8F, 4996.7 4 989 8 4 906 4911 -5 5 215 5 182 33
BF. 6 415.4 6406 9 6 322 6 305 17 6 555 6 581 -26
IF. 6 854.9 6888 -33 6766 6790 -24 7 031 7044 -13
IG. 9 921.4 9903 18 9 766 9 753 13 9 997 9 989 8
IDs 17 334.5 17 329 5 16 714 16709 5 16 846 16 850 -4
aP D 21 390.1 21 362 28 20 475 20 450 25 20 925 20 899 26
aPI 22007.6 21 984 24 21 081 21 057 24 (21 473)· 21 528 -55
11& 22 211.6 22 211 1 21 410 21 410 0 21 519 21 519 0
ap, 23 160.9 23 214 -53 22 227 22 279 -52 22 754 22 722 32
ISo 50090.3 50 090 0 48800 48 800 0 46 900 46900 0
rms Deviation 36 33 47
a Reference 19. d The parameters used to generate this set of energies are given in
b Reference 32. Table II.
• Reference 33. • Based on the lowest three Stark levels in the 'P,-II6 manifold.
found that where comparison is possible lanthanide rise to the band. 13 •14 Such a correlation provides addi-
absorption bands in the latter medium occur at essen- tional criteria for making assignments.
tially the same energies and with nearly the same The intensity calculation has been discussed else-
intensities as those observed in aqueous solution. The where. 13 •26 The tables included here"contain the ma-
only exception involves those bands which show a trix elements of U().) , since thesetmatrix elements
hypersensitivity to medium effects. 13 •31 We have al- form the basis for making most of thelassignments.
ready pointed out that there is a one-to-one corre- We adopted the following convention for U('A) =
spondence between the hypersensitivity of a lanthanide (l/IJ II U().) II l/I'J') 2: where (l/IJ II U().) IIl/I'J') <10-7 ,
absorption band, and very large calculated values of U(A) =0; where 0.00005< U(A) > IX 10-14, U(A)~O.
the matrix element of U(2) for the transition giving The parameters, Ek, !;4J, 01, (3, and /" are summarized
TABLE IV. Energy-level assignments and matrix elements of U().) for Pr8+ (aq).
E ••ptl E....lo· AE
S'L']' (em-I) (em-I) (em-I) U(2)b U(4)b U(6)b
3H. 200° 245 -45

3H6 2 360- 2 322 38 0.1095 0.2017 0.6109
aHs 4 500d 4 496 ·4 0.0001 0.0330 0.1395
aF2 5 200d 5 149 51 0.5089 0.4032 0.1177
3Fa 6500 6540 -40 0.0654 0.3469 0.6983
'F. 6 950 6 973 -23 0.0187 0.0500 0.4849
IG. 9 900 9 885 15 0.0012 0.0072 0.0266
IDs 16840 16840 0 0.0026 0.0170 0.0520
ap. 20750 20706 44 0 0.1728 0
3PI 21 300 21 330 -30 0 0.1707 0
lIs 21 500 21500 0 0.0093 0.0517 0.0239
ap, 22 520 22 535 -15 ",0 0.0362 0.1355
ISO 46 900° 46 900 0 0 0.0070 0
rms Deviation 48
• The parameters used to generate t "i. set of energies are given In ° Assumed energy based on data in Ref. 33.
Table II. d Based on data in Ref. 28.
bU(h) -(,yJ II U().) ",y'J')I.
•1 C. K. JfIlrgensen and B. R. Judd, Mol. Phys. 8,281 (1964); B. R. Judd, J. Chern. Ph:ts. 44,839 (1966) •
4430 CARNALL, FIELDS, AND RAJN AK
TABLE V. Energy-level assignments and matrix elements of U(X) for NdH (aq).
N dH in LaClaa NdH (aq)
Euptl Eoa1.b I1E E ••ptl E.a1.b I1E

S'L']' (em-I) (em-I) (em-I) (em-I) (em-I) (em-I) U(2)c U(4)c U(6)c
41911 146 178 -32 146 d 130 16

4Iu/'I. 2 028 2043 -15 2 028 d 2007 21 0.0194 0.1073 1.1652
411311 4 010 4 029 -19 40500 4005 45 0.0001 0.0136 0.4557
'11512 6 058 6092 -34 6 050 6 080 -30 0 0.0001 0.0452
4Fm 11 438 11 475 -37 11 460 11 527 -67 0 0.2293 0.0549
'Fb12 12 466 12 517 -51 12 480 12 573 -93 0.0010 0.2371 0.3970
2H9/2 12 618 12 710 -92 12 590 12 738 -148 0.0092 0.0080 0.1154
'F112 13 437 13 504 -67 13 500 13 460 40 0 0.0027 0.2352
'Sm 13 524 13 435 89 13 500 13 565 -65 0.0010 0.0422 0.4245
'F912 14 722 14 793 -71 14 700 14 854 -154 0.0009 0.0092 0.0417
2H11I2 15 935 16001 -66 15 870 16 026 -156 0.0001 0.0027 0.0104
'G. '2 17 135 16 999 136 17 300 17 167 133 0.8979 0.4093 0.0359
2G1/2 17 194 17 460 17 333 127 0.0757 0.1848 0.0314
IKta/2 18 874 19 018 0.0068 0.0002 0.0312
'G7/t 19 020 18 965 55 19 160 19 103 57 0.0550 0.1570 0.0553
'G9IJ 19 434 19 369 64 19 550 19 544 6 0.0046 0.0608 0.0406
2KlS/2 20 858 21 000 21 016 -16 0 0.0052 0.0143
2G912 21 056 21060 -4 21 300 21 171 129 0.0010 0.0148 0.0139
(2D,2P)3/2 21 129 21 191 -62 21 300 21 266 34 0 0.0188 0.0002
'GUll 21 426 21 357 69 21 650 21 563 87 ",0 0.0053 0.0080
2Pl/I 23 214 23 050 164 23 250 23 140 110 0 0.0367 0
2D5/1 23 780 23 750 30 23900 23 865 35 ",0 0.0002 0.0021
(IP,2D)3/2 26160 26 158 2 26300 26 260 40 0 0.0014 0.0008
'Dm 27 981 27 990 -9 28300 28 312 -12 0 0.1960 0.0170
'D.,2 28 187 28500 28477 23 0.0001 0.0567 0.0275
2IU/2 28 388 28 624 0.0049 0.0146 0.0034
'D1I2 28 505 28 563 -58 28 850 28 894 -44 0 0.2584 0
2LI.,2 29260 0 0.0248 0.0097
2111/2 29 966 0.0001 0.0013 0.0017
'Dl/1 30500 30 554 -54 ",0 0.0037 0.0080
ILll/! 30 747 0 0.0010 0.0012
2H9/t 32 567 0.0001 0.0085 ",0
IDa/l 33400 33 481 -81 0 0.0112 0.0012
tHu/'/. 33 913 0.0001 0.0001 0.0002
JDm 34 450 34 474 -24 0.0007 0.0006 0.0034
IF./2 38500 38504 -4 0.0021 0.0033 ",0
2F1I2 39 950 39 926 24 ",0 0.0004 0.0007
2G912 47 700 47 696 4 ",0 0.0015 0.0001
2Gl/2 48 600 48 586 14 0.0004 0.0024 0.0002
rms Deviation 84 rms Deviation 88
a Reference 26. • U (1\) = (1/11 II u(X) 11.p'J') t.

b The parameters used to generate this set of energies are given in d Assumed energy based on data in Ref. 26.
Table II. e Based on data in Ref. 28.
in Table II for all the systems examined. The values particularly relevant. The assignments, together with
given are all based upon fitting procedures carried out the levels calculated based upon seven-parameter fits
during the present investigation. to the crystal and free-ion data, carried out during the
present investigation, are shown in Table III. The
PrH (p)
experimental results in solution are given in Table IV.
There are numerous references to the energy-level It was concluded that in this case the energies assigned
structure of PrH, but for purposes of comparison with to solution bands corresponded most closely with those
the data obtained in solution, the free-ion levels19 and obtained for PrH in LaF3 •
those reported for p r3-t in LaCla 32 and in LaF3 33 are Most of the assignments shown in Table IV were
12 R. Sarup and M. H. Crozier, J. Chern. Phys. 42,371 (1965); to resolved absorption bands. Curv~ fitting25 to the 3p
J. Markovsky, W. Low, and S. Yatsiv, Phys. Letters 2, 186 group resolved a band centered at 21500 cm- I which
(1962) . was attributed to lIe. Since in this case it was necessary
NIH. H. Caspers, H. E. Rast, and R. A. Buchanan, J. Chem.
Phys.43, 2124 (1965). to assign an energy to the ISO transition in order to
determine the value of the configuration-interaction only three had large matrix elements of U(A): 5Do, 5DI ,
parameter "I, it was assumed that the energy of the and 5D2• The observed relative intensities of the bands
transition in solution corresponded to that assigned in were in keeping with the predicted intensity pattern.
LaFa. In solution, a relatively intense absorption bandl4
is observed at >43500 cm- I due to transitions 4JN---'lo
4JN-I 5d. Calculation indicates that the ISO band should
be weak, and indeed no transition is observed super- The energies of the 6FI / 2 and 6Fs/2 levels in Sm3+ were
imposed on the intense j---'lod absorption. assigned after comparing results in aqueous solution
The parameters of a six-parameter fit excluding the with those in a fused nitrate salt. 28 Preliminary calcula-
assignment to ISO, except for E, are identical to those tions showed that the intensity of the 6F3/ 2 transition
which include "I and the ISO energy. was in good agreement with that of the band observed
in aqueous solution at 6630 cm-I . Since a large matrix
Nd 3+ (P) element of U(2) was calculated for the 6FI / 2 level (Table
VII), the corresponding absorption band was identi-
A comparison of the energies for Nd3+ transitions in fied with a hypersensitive transition. In aqueous solu-
LaCIa 26 and LaFa 27 with the centers of band systems tion only a tailing to the infrared of the band associated
in aqueous solution revealed a close correlation with with the 6Fs/2 level could be observed, but in the nitrate
the data for N d3+ in LaCIa (Table V). melt an intense absorption was found centered at 6400
The 4(;5/2 transition is hypersensitive l3 ,31 and thus cm-I • The latter energy was, therefore, also assigned
was clearly identified with a band at 17 500 cm-t, to the transition to the 6F1/ 2 level in aqueous solution.
whereas a resolved component at 17460 cm-1 was as- Although a very large number of transitions occur in
signed to the weaker 2G7 / 2 level. Of the transitions to the range 29 000-50 000 cm-I, only a very few had
2K13/2 and 4G 7/ 2 which calculated near to the same en-
calculated matrix elements of U(A) that could account
ergy, the latter was assigned to the center of a band for the magnitude of the observed absorption bands.
at 19 160 cm-I since the former could not account for The levels whose eigenvectors contain appreciable
the observed intensity. components of states with seniority number different
Isolated bands are observed in Nd3+ even at high than that of the ground state are usually found at
energies. The intensity calculation based on parameters higher energies. However, in Sm3+ the eigenvectors of
fit to the assignments made at <35000 cm-I predicted the 4G6/ 2 and 4Fs/2 levels, which are identified with iso-
only four relatively weak transitions in the 35000- lated absorption bands near 18000 cm-t, contain appre-
50000 cm-I range. Since the eigenvectors of 2F6/2 and ciable contributions from states of seniority number 3.
2F7/2 revealed that both states involve large percent-
Thus, to obtain a good fit to these two levels where
ages of components differing in seniority number from the assignments are unequivocal, the parameter "I
that of the ground state, it was possible to determine should be included. Unfortunately, there have been
the parameter "I after assigning energies for these two relatively few assignments made to the spectrum of
levels. Sm3+ in LaCIa.a5 Most of the assigned levels above
17000 cm-I (Table VII) are expected to be dependent
on "I, so it is not possible to obtain a six-parameter
A center for the Stark components of the ground fit while excluding assignments dependent on "I. The
term 514 was assumed to be ,.......150 cm-I . No spectro- parameters given for Sm3+ in LaCIa (Table II) must
scopic investigations of Pm3+ in a crystal lattice have therefore be considered only approximate. A sufficient
been reported. The energies of the 517 and 518 levels number of assignments were made to determine all
are unknown, but the other levels in the ground-term seven parameters from the solution data.
multiplet have been observed (Table VI).
Preliminary intensity calculations showed that it was
reasonable to assign the five transitions 5F1 ,2,3,4, 5S2 to
the centers of resolved bands in the 12 000-15 000 cm- I The assignments for Dy3+(aq) were derived primar-
range, and to place both sF5 and 3K1 at 15 900 cm- I , ily from intensity correlations since investigations in
where the absorption was well isolated. Both the transi- crystals have been very limited. A known hypersensitive
tions to 5G2 and 5G3 were identified with known hyper- band29 at 7700 cm- I confirmed the assignment of the
sensitive bands at 17700 and 18300 cm-I •30 transition to the 6Fu /2 level at this energy (Table VIII).
Gruber and Conway34 reported finding three rela- Recently, we attempted to define approximate electro-
tively intense, well-resolved bands near 30000 cm- I . static and spin-orbit parameters for Dy3+(aq) by as-
Using the assignments at <30000 cm-I as the basis, signments to several well-defined, lower-energy absorp-
a preliminary calculation showed that among the nu- tion bands. 14 A poor correspondence was found between
merous transitions predicted to occur near 30 000 cm-t, calculated levels and those observed levels which were
M J. B. Gruber and J. G. Conway, J. Inorg. Nuc!. Chern. 14, 36 M. S. Magno and G. H. Dieke, J. Chern. Phys. 37, 2354
303 (1960). (1962) .
4432 CARNALL, FIELDS, AND RAJN AK
TABLE VI. Energy-level assignments and matrix elements of U(X) for Pm3+ (aq).
E ••v'] Eoaloa llE

S'L'J' (em-1) (em-1) (em-1) U(2)b U(4)b U(6)b
51, 150- 99 51
515 1 577 0.0246 0.1172 0.9694
516 3 186 0.0018 0.0301 0.6893
517 4 85()d 4 876 -26 0 0.0025 0.1581
5Is 6 600d 6611 -11 0 ~ 0.0103
oF1 12 400 12 398 2 0 0.1404 0
oFs 12 820 12 811 9 0.0026 0.1992 0.1264
sFs 13 600 13 651 -51 ",0 0.1041 0.4253
5S, 14 300 14 337 -37 ~ 0.0011 0.2295
5F, 14 650 14 562 88 0.0005 0.0291 0.2403
5F. 15 900 15 863 37 ",0 0.0021 0.0346
aKa 15 900 15 875 25 0.0024 0.0025 0.0104
8K7 17 163 0 0.0020 0.0200
8H, 17 300 17 327 -27 0.0064 0.0210 0.0240
5c, 17 700 17 857 -157 0.7215 0.2433 0.0041
DCa 18300 18 256 44 0.1444 0.2655 0.0454
3K a 18 719 0 0.0003 0.0088
3H. 19 617 0.0001 0.0057 0.0062
DG, 20 250 20 181 69 0.0093 0.0957 0.0787
3Ga 21 102 0.0228 0.0652 0.0075
5G. 21 900 21 998 -98 0.0003 0.0103 0.0365
3D2 22 178 0.0072 0.0025 0.0028
5Ge 22300 22 262 38 ~ 0.0002 0.0008
3L7 22300 22 372 -72 0 0.0016 0.0099
3D1 23 321 0 0.0001 0
3L8 23500 23444 56 0 0.0019 0.0133
3G. 23 897 0.0013 0.0112 0.0044
8H e 23995 ~ 0.0002 0.0041
3L9 24 412 0 0 0.0019
(3M,1Lh 24 462 0 0.0045 0.0013
3Da 24800 24800 0 0.0007 0.0032 0.0095
3P, 25066 0 ~ 0
(1D,SPh 25 538 ~ 0.0002 0.0003
BGo 26 235 0.0002 0.0006 0.0007
IF, 26643 ",0 0.0005 0.0008
sPI 27 051 0 0.0003 0
3Mg 27 804 0 0 0.0007
8Ft 27 894 0.0091 0.0062 ",0
3Is 27916 0.0042 0.0004 0.0006
3F, 28 193 0.0048 0.0097 0.0025
3M1o 28 395 0 0 ",0
3Fa 28 810 0.0002 0.0031 0.0013
3le 29 078 0.0003 0.0006 0.0011
1L8 29 189 0 0.0015 0.0008
8f? 29 587 0 0.0001 0.0002
5Do 30040 29 979 61 0 0.1520 0
5D1 30 480 30 471 9 0 0.2361 0
sFa 30 816 0.0012 0.0096 0.0046
5D2 31 250 31 266 -16 0 0.1013 0.0214
rms Deviation 78
• The parameters used to generate this set of energy levels are given in • Assumed center of Stark components of the ground level.
Table II. d Based on data in Ref. 30.
II II
b U (h) = <1/IJ U(~) 1/1' JI) ••
included in the fit. As we extended the earlier study energies, the error was 200-300 cm-1 in a number of
by including the parameters IX and fJ, the fit to observed cases. We could not make any satisfactory progress in
energies remained poor. Although calculated matrix the fitting procedure until we recognized the crucial
elements of U<X) clearly indicated that certain transi- importance of 'Y in the fitting process. At the outset,
tions should be observed within rather narrowly defined because of the paucity of assignments, the use of a
TABLE VII. Energy-level assignments and matrix elements of U().) for SmH (aq).
SmH in LaClaa SmH (aq)
Eexpd Eoalob t.E E ••pi! Eoal.b t.E

S'L'J' (em-I) (em-I) (em-I) (em-I) (em-I) (em-I) U(2)o U(4)o U(6)c
6H5/1 36 12 24 36d 46 -10

6H7I2 1 080 1 067 13 1 080 d 1 080 0 0.2062 0.1962 0.0952
6H 9/J 2 286 2 279 7 2 286d 2 290 -4 0.0256 0.1395 0.3267
6HU/2 3608 3 618 -10 3 608 d 3 624 -16 0 0.0240 0.2649
6HI3 12 5 014 5044 -30 5 000- 5042 -42 0 0.0007 0.0659
6F1I2 6 299 6400- 6397 3 0.1939 0 0
6HUI ! 6 523 6508 0 0 0.0043
sFal! 6544 6630 6641 -11 0.1444 0.1364 0
6F6/! 7 033 7 041 -8 7 100 7 131 -31 0.0332 0.2840 0
6F m 7900 7898 2 8000 7977 23 0.0020 0.1429 0.4301
6F 91! 9 075 9 069 6 9 200 9 136 64 ...v() 0.0206 0.3413
6Fu/s 10460 10 458 2 10500 10 517 -17 0 0.0006 0.0515
'G6/! 17 860 18 053 -193 17 900 17 924 -24 0.0002 0.0007 0
'Fa/! 18860 18 632 228 18900 18 832 68 0.0003 ...v() 0
'Gm 20 010 20100 -90 20 050 20014 36 0.0004 0.0018 0.0025
'[912 20 600 20556 44 20526 0.0022 0.0005 0.0014
4M15I2 20279 20 800 20627 173 0 0 0.0307
'[UI2 20977 20996 -19 21 100 21 096 4 0 ",0 0.0108
'1 13/2 21 562 21 497 65 21 600 21 650 -50 0 0.0030 0.0228
'F6/! 22 129 22 172 -43 22200 22 098 102 0.0004 0.0002 0
'Ml7/2 22 006 22 370 0 0 0.0053
'G9/2 22 850 22 749 101 22 700 22 706 -6 0.0001 0.0010 0.0028
'II6/Z 22 797 22 966 0 0 0.0002
'MI9 /! 23 543 23 902 0 0 0
(6P, 'Ph12 23 784 23 776 8 24 050 24 101 -51 .....,0 0.0263 0
'L 13 /2 24 560 24 424 136 24 570 24 562 8 0 0.0081 0.0096
'F7I! 24783 24 775 0.0002 0.0012 0.0003
eP812 24 631 24 950 24 999 -49 .....,0 0.1684 0
'Ku/! 24 532 24 752 -220 25 177 0 0.0004 0.0027
'M!lIs 25 224 0 0 0
'L15/2 25 650 25638 12 0 0 0.0060
'GU/2 25 718 0 0.0001 0.0010
'DU2 26 573 0.0001 0 0
eP712 26750 26 660 90 .....,0 0.0016 0.0751
'47/2 26 749 0 0 0.0002
'Kla/2 26 967 0 0.0005 0.0011
'F9/2 27 207 .....,0 0.0004 ...v()
'D3/t 27 700 27 714 -14 0.0001 0.0251 0
('D,6Ph/! 27700 27 722 -22 .....,0 0.0170 0
'Hm 28 250 28 396 -146 0.0013 0.0006 .....,0
'KIS/! 28 732 0 0 .....,0
'H 9/2 29 100 29 012 88 0.0001 0.0002 0.0006
'D 7/2 29 100 29 107 -7 ...v() 0.0005 0.0375
('K, 'L) 1711 29 178 0 0 .....,0
'49/s 29 322 0 0 0
'Hul! 29 381 0 0.0001 0.0005
'HIB !2 29 827 0 ...v() 0.0002
'G71! 29 980 .....,0 0.0006 0.0004
'G 9/! 30 200 30 099 101 0.0001 0.0009 0.0013
'G6/2 30232 0.0002 0.0002 0
'Pill 31 093 ...v() 0 0
'Gu 12 31 349 0 .......,0 0.0003
2£'6/2 31408 0 0 0.0002
'Pa/2 31 550 31 508 42 .....,0 0.0136 0
'P 6/! 32 800 32 706 94 .....,0 0.0021 0
2F5/S 33800 33 767 33 ...v() 0.0005 0
TABLE VII ( Continued)
SmH in LaCl s" SmH (aq)
Enptl Eea",b dE E•• vtl E •• lob dE

S'L'J' (em-I) (em-I) (em-I) (em-I) (em-I) (em-I) U(2)o U(4)o U(6)·
2K18/2 33800 33 825 -25 0 0.0002 0.0001

'F9/2 34 061 ~ .....,0 .....,0
2LI7/2 34 357 0 0 .....,0
('I, 'Fh/2 34600 34 591 9 .....,0 0.0004 0.0003
2NU/2 35 385 0 0 0
'PII2 35 718 ",,0 0 0
4F7/2 35 785 ,....,,0 0 ,....,,0
'1 u /2 36000 36 053 -53 0 0.0002 0.0003
2NU/! 36 238 0 0 0
'Fo/2 36 520 ",,0 rvO 0
'110/2 36 586 0 0 ",0
'Fa/2 36 700 36 586 114 ~ 0.0002 0
'[13/1 36 700 36 757 -57 0 ,....,,0 0.0001
2Ml us 36 982 0 0 rvO
(IH, '[)m 37 600 37 487 113 ,....,,0 0.0005 0.0001
('D, 'F)3/2 38300 38 270 30 ~ 0.0003 0
'Ful 38 803 0 .....,0 .....,Q
IKIS/t 38 949 0 0 .....,Q
'GU2 39 057 0.0001 ",0 ,....,,0
IHu/! 39 250 39 188 62 0 0.0002 .....,Q
'MUll 40 417 0 0 0
'Dol! 40664 .....,0 ",,0 0
'[U/2 40 990 0 0.0001 .....,0
'KII" 41 269 0 0.0001 .....,0
'Dill 41 369 ",,0 ",,0 0
'G9/2 42000 41 941 49 '"'-0 0.0002 0.0001
'028/1 42 022 0 0 0
'021/1 42 406 0 0 0
'Go/s 42 714 0.0016 0.0010 0
'G7/1 42900 42 965 -65 0.0013 0.0025 0.0006
'Kw2 43 028 0 0 .....,0
('H, 'G)g/l 43 100 43 250 -150 0.0003 0.0008 0.0015
(1[, 'Hhlll 43 414 0 0.0006 0.0006
(II, 'Hh./I 43504 0 0.0002 .....,Q
('G, IH, 'H) nil 43800 43 845 -45 0 0.0008 0.0006
('G, 'H) 7/1 44 237 0.0002 ",0 ~
'G9/2 44 832 .....,0 ",0 0.0002

('G, 'H, 'G) 7/1 45 269 ",,0 0.0006 .....,0
'GeJl 45 615 ",0 0.0006 0
I[a/I 45 801 0 .....,0 .....,0
'Gtlll 46 123 0 0.0001 0.0002
1L.7/1 46 370 0 0 .....,0
'HIllS 46500 46 500 0 0 0.0004 0.0002
'Hu/! 46500 46 554 -54 0 0.0005 .....,0
('P,'Dh/l 46 500 46 669 -169 0.0002 0.0010 0
'Dill 47 127 .....,0 ",,0 0
tHIlil 47400 47 307 93 0 0.0010 0.0001
'L.6/1 47 843 0 0 .....,Q
(IF, 'P, 'Dh/I 47940 0 0.0003 0
'Hg/I 48 288 ",,0 ,....,,0 .....,Q
rms Deviation 119 80
a Reference 35. d Assumed based on data for Sm'+ in LaCIa (Ref. 35).
b The parameters used to generate this set of levels are given in Table II. e Based on data in Ref. 28.
II II
• U (A) - (tfJ U(A) >It'fl'.
TABLE VIII. Energy-level assignments and matrix elements of U(~) for Dyl+ (aq).
Dy8+ in LaC)." Dya+ (aq)

b tt.E E •• ptl Eo.lob tt.E
E •• ptl E ... lo
S'L'J' (em-I) (em-I) (em-I) (em-I) (em-I) (em-I) U(2)- U(4)- U(6)"
IHI5/2 52 35 17 52d 40 12
IH 1312 3 517 3480 37 3 517d 3 506 11 0.2457 0.4139 0.6824
IHU/2 5 835 5804 31 5 850 5 833 17 0.0923 0.0366 0.6410
IH9I2 7 667 7 637 30 7 700 7 692 8 0 0.0176 0.1985
I FU/2 7667 7663 4 7 700 7 730 -30 0.9387 0.8292 0.2048
IFS/! 9 014 9011 3 9 100 9087 13 0 0.5736 0.7213
6H 7I2 9 014 9099 -85 9 100 9 115 -15 0 0.0007 0.0392
6H5/1 10 140 10146 -6 10200 10 169 31 0 0 0.0026
IF7I1 10 926 10 945 -19 11 000 11 025 -25 0 0.1360 0.7146
IFsl2 12 323 12 341 -18 12400 12 432 -32 0 0 0.3452
IFI/2 13 116 13 109 7 13 250 13 212 38 0 0 0.0610
IFlit 13 650 13 760 0 0 0
'F 9/1 20 963 21056 -93 21 100 21 144 -44 0 0.0047 0.0295
'116/2 21 954 21 956 -2 22 100 22 293 -193 0.0073 0.0003 0.0654
'GUll 23 303 23 288 15 23400 23 321 79 0.0004 0.0145 0.0003
'Fm 25 640 25800 25 754 46 0 0.0768 0.0263
'111/2 25 539 25800 25 919 -119 0.0041 0.0013 0.0248
'MU/2 25 562 26 341 0 0.0102 0.0822
'Kim 25 627 26400 26 365 35 0.0109 0.0048 0.0935
'MIDIS 26 374 27400 27 219 181 0.0004 0.0166 0.1020
(4P, 'D)a/2 26 832 27400 27 254 146 0 0 0.0448
'Pm 27 038 27400 27 503 -103 0 0 0.0697
'11111 27 728 28 152 0.0001 ",0 0.0074
'P7/1 28 326 28346 -20 28 550 28 551 -1 0 0.5222 0.0125
('M, '1)15/1 28 575 29244 0.0023 0.0005 0.0009
('F, 'Dh/l 29480 29 383 97 29600 29 593 7 0 0 0.0249
'I 91t 29 885 0 0.0003 0.0003
'Gm 30200 0 0.0014 0.0005
'PIII 30 800 30 803 -3 0 0 0.1095
4M17/1 30 892 0.0032 ",0 0.0012
('G,2F)7I1 31 560 0 0.0066 0.0002
'KlD/1 31 795 0.0028 0.0001 0.0093
'Dl/ 2 31 842 0 0 0
('D, 'Ghll 31 857 0 0 ",0
'M,D/S 32 187 0.0004 0.0203 0.0056
'HulS 33500 33471 29 0.0099 0.0100 0.0016
'FilS 33642 0 0 0.0045
('K, 'Lhllt 33 776 0.0060 0.0024 "-'0
'Dm 33900 33 834 66 0 0.0661 0.0016
,H11I1 34 307 0.0029 0.0018 0.0022
'G9/2 34311 0 0.0007 0.0041
'Fill 34 398 0 0 0.0004
'GU /2 34900 34954 -54 0.0016 0.0054 0.0075
, M,712 34900 35047 -147 0.0003 0.0088 0.0018
('G, 'Hh/2 35900 35891 9 0 0.0042 0.0001
'Kult 36432 0.0001 ",0 0.0001
(4G, 'PhIl 36 550 36 524 26 0 0 0.0040
4GB" 36 794 0 0.0013 0.0014
(4L, 'K)I812 37 103 0.0003 0.0001 0.0006
4M,W2 37 230 0.0001 0.0016 0.0002
'Pllt 37 435 0 0 0
4H7/1 37 807 0 rvO 0.0002
'Fm 37900 38 031 -131 0 0 0.0043
(2K,ILh5/2 38811 0.0003 0.0001 0.0007
('G, 'PhiZ 38900 38860 40 0 0 0.0061
('P, 'F)." 39 100 39 127 -27 0 0 0.0091
TABLE VIII (Continued)
DyH in LaCla" Dy3+ (aq)
EeIJltl Eoalob t:.E

S'L'J' (em-I) (em-I) (em-I) U(2)o U(4)o U(6)·
40 922 0 0 ~O
41 050 41 035 15 0 0.0015 0.0002

fJI6/' 41 700 41 596 104 0.0030 0.0043 0.0001
'4712 41 848 0.0001 ~ 0.0002
4F7/! 42 127 0 0.0021 0.0002
2Pm 42 807 0 0 0.0005
rms Deviation 56 86
• See Table II. Footnote f. • u 0.) = (Il-'J II Uo.) IIljIJ')t.

b The 'parameters used to generate this set of energies are given in d Assumed energy based upon data in Table II. Footnote f.
Table II.
seven-parameter fit required estimates for some of these it was necessary to employ the data published by sev-
parameters. However, this procedure did permit an eral different authors for Tm(C2H 5S04 )a-9H20,3lH2 as
immediate, striking improvement of the fit, and pre- shown in Table XII. Since the value of 'Y in this case
dictions of transition energies and intensities which depends wholly on the energy assigned to the transition
were readily verified by reference to the observed spec- to ISO, which has not been observed, it was necessary
trum.I4 to limit the fit to six parameters.
Ho3+ (flO)
DISCUSSION
In this and in subsequent heavy lanthanides, all
excited levels in the ground-term multiplet were ob- Intensity calculations have proven to be a useful
served directly either in aqueous solution or in a molten tool in the assignment of energy levels in lanthanide
nitrate salt medium. Bands at 22 100 (5G 6) and 27 700 solution spectra. The energy parameters summarized
cm-I (3G 6 ) (Table IX) are known to be hypersensitive. 25 in Table II for the aquo ions should also provide a
The absence of any large background corrections in valuable guide to further assignments in crystals. Since
the ultraviolet range made it possible to assign several all three parameters of configuration interaction have
transitions to relatively weak bands between 40000 been determined here for the first time for trivalent
and 50000 cm-I. lanthanides, it now becomes possible to examine their
behavior across the series. We will also be interested
Er+ (f11) in the variation of the values of Flo and tv determined
A considerable amount of experimental work on the here as a function of the atomic number Z. That part
energy level scheme of Er+ has been reported includ- of the configuration interaction which has the same
ing a study of the free-ion spectrum.86 The published angular form as the coefficients, e", of the E" param-
assignments for Er8+ (free-ion), Er+ in LaF3,37 and eters [Eq. (6) ] provides a negative screening of the
Er+ in LaCla 38 are given in Table X, together with E"'s which is not removed by the addition of a, (:1, and
the results of a six-parameter fit to the data in each case. 'Y. This screening results in the empirical parameters
The experimental and calculated results for Er8+ (aq) being smaller than those calculated by Hartree-Fock
are shown in Table XI. Practically all of the states (H-F) methods. 43
which would be strongly dependent on 'Y lie at > 50 000
cm-l . Only in solution was it possible to obtain an Slater Integrals (F,,)
estimate of the value of this parameter.
A plot of F2 vs Z for the lanthanides considered here
Tm3+ (f12) is linear to a good approximation as shown in Fig. 1.
In order to obtain experimental results in a crystal 811 J. B. Gruber and J. G. Conway, J. Chern. Phys. 32, 1178
medium that were comparable to those for Tm3+ (aq) , (1960); 32, 1531 (1960).
40 E. Y. Wong and I. Richman, J. Chern. Phys. 34, 1182 (1961).
as w. J. Carter and H. M. Crosswhite, J. Chern. Phys. (to be 41 B. C. Gerstein, L. D. Jennings, and F. H. Spedding, J. Chern.
published) . Phys. 37, 1496 (1962).
1fT W. F. Krupke and J. B. Gruber, J. Chern. Phys. 41, 1225 42 W. F. Krupke and J. B. Gruber, Phys. Rev. 139, A2008
(1964) . (1965) .
18 F. Varsanyi and G. H. Dieke, J. Chern. Phys.36, 2951 (1962); 48 A. J. Freeman and R. E. Watson, Phys. Rev. 127, 2058
G. H. Dieke and S. Singh, ibid. 35, 555 (1961). (1962) .
TABLE IX. Energy-level assignments and matrix elements of U(~) for Ho3+ (aq).
Hos+ in LaCl3" H 0 3+ (aq)
E•• ptl E ... lcb AE E.xp'l Ealcb tJ.E

S'L'J' (em-I) (em-I) (em-I) (em-I) (em-I) (em-I) U(2)c U(4)c U(6)c
51a 107.7 140 -32 108d 80 28

617 5 154.8 5 169 -13 5 130· 5 116 14 0.0250 0.1344 1.5216
61d 8 656.9 8 668 -11 8 580 8 614 -34 0.0084 0.0386 0.6921
610 11 218.9 11 216 3 11 120 11 165 -45 0 0.0100 0.0936
61. 13 284.1 13 272 12 13 300 13 219 81 0 ......,0 0.0077
6F. 15 482.0 15508 -26 15500 15 519 -19 0 0.4250 0.5687
6S2 18 430.3 18 284 146 18 500 18 354 146 0 0 0.2268
6F, 18 553.9 18 619 -65 18500 18 612 -112 0 0.2392 0.7071
6Fa 20 614.0 20644 -30 20 600 20 673 -73 0 0 0.3460
6F2 21 069.2 21 058 11 21 100 21 130 -30 0 0 0.1921
3Ka 21 350.7 21 310 41 21 370 21 308 62 0.0208 0.0334 0.1578
6GB 22 100.6 22 118 -17 22 100 22 094 6 1.5201 0.8410 0.1411
6FI 22 304 22 375 0 0 0
(SG,3G). 23 936.4 23 782 154 23 950 23 887 63 0 0.5338 0.0002
DG. 25 820.4 25 798 22 25800 25 826 -26 0 0.0315 0.0359
3K7 26 139.7 26 114 26 26200 26 117 83 0.0058 0.0046 0.0338
(sG,3H). 27 648.8 27640 9 27700 27 653 47 0 0.0790 0.1610
sHa 27 703 27 700 27 675 25 0.2155 0.1179 0.0028
(6F, SF, 6Gh 27968.8 28 136 -167 28 250 28 301 -51 0 0 0.0041
iGS 28 810.6 28 786 24 28 800 28 816 -16 0 0 0.0133
sLg 28 911.4 28 988 -77 29000 29 020 -20 0.0185 0.0052 0.1536
(SF, 3H, 3G). 29 915 30 000 30 017 -17 0 0.1260 0.0047
sKs 29 994 30000 30 036 -36 0.0026 0.0002 0.0026
6G2 30 687 30900 30 813 87 0 0 0.0010
3Da 33 066.5 33 130 -64 33 200 33 339 -139 0 0 0.0030
3PI 33 152 33 398 0 0 0
aM10 34375 34200 34264 -64 0.0003 0.0696 0.0808
aLa 34 273 34200 34 306 -106 0.0017 0.0005 0.0108
(6G, 6D, 3G), 34 728.3 34 761 -33 34800 34794 6 0 0.3040 0.0492
(3F,3G)a 34977 35200 35 224 -24 0 0 0.0036
3PO 35 852 36050 0 0 0
(3H, 6D, IG). 35 980 36000 36046 -46 0 0.2635 0.0041
3F2 36 006 36 364 0 0 0.0035
ILa 36488 36500 36 516 -16 0.0002 0.0056 0.0016
(3H,3G). 36 722.2 36684 38 36 800 36773 27 0 0.0024 0.0032
(3P,ID)z 37 686 37 845 0 0 ......,0
3~ 37 924 38000 38 022 -22 0.0020 ......,0 0.0036
sh 38 314.2 38 268 46 38500 38 470 30 0.0157 0.0003 0.0080
3F. 38 500 38 509 -9 0 0.0084 0.0023
31s 39271 0 0.0008 0.0008
3Mp 39500 39 435 65 0.0005 0.0057 0.0029
31s 39 830 0.0065 ......,0 0.0043
(3D, 6D, sPh 39 982 0 0 0
iD3 40000 39 992 8 0 0 0.0293
(3F,6D). 41 550 41 532 18 0 0.2577 0.0144
6D2 41900 41 922 -22 0 0 0.0128
iDO 42 582 0 0 0
(SF,6D)a 42 850 42 811 39 0 0 0.0079
6DI 43 066 0 0 0
(lD,3Dh 45 286 0 0 0.0004
3Ma 45 691 ......,0 ......,0 0.0001
3H. 45 705 0 0.0068 0.0003
(3H, II). 45 700 45 724 -24 0.0001 0.0118 0.0005
3F3 47448 0 0 0.0008
(3H, 3G, 31). 48200 48 102 98 0 0.0034 ~
(ID,3Fh 48 736 0 0 ~
IH" 49 335 0 0.0016 ~
rms Deviation ~2 62
• See Table II. Footnote g. d Assumed energy based on data in G. H. Dieke and B. Pandy, J. Chern .
b The parameters used to generate this set of energies are given in Phys. 41. 1952 (1964).
Table II. • Based on data in Ref. 29.
II
• Uo..) = (!/tJII U IA ) ,yJ/) ••
TABLE X. Energy-level assignments for E1"+.

---------- ====--===================
Er3+ (free ion)' Er3+ in LaCl{
S'L'J'
o 96 -96 217 232 -15 108 130 -22

6 485.9 6 567 -81 6697 6 732 -35 6 589 6 630 -41
10 123.6 10 140 -16 10340 10 335 5 10219 10226 -7
12 345.5 12 294 61 12 567 12 504 63 12 438 12 377 61
15 182.9 15 202 -19 15 452 15 432 20 15 283 15 260 23
18 299.6 18 445 -145 18 570 18 696 -126 18 398 1~ 536 -138
19 010.8 19 096 -85 19 334 19 398 -64 19 144 19 181 -37
20 494.1 20 391 103 20 709 20 652 57 20 515 20 449 66
22 181.8 22 014 168 22 378 22 305 73 22 174 22 097 77
22 453.9 22 368 86 22 711 22645 66 22 516 22 449 67
24 475.2 24 367 108 24744 24644 100 24 565 24 471 94
26 376.9 26 348 29 26 585 26 680 -95 26 367 26 397 -30
27 319.2 27 333 -14 27 629 27 693 -64 27 327 27 358 -31
27 584.9 27 597 -12 27 857 27 605 27 643 -38
27 825.0 27 843 -18 28 298 28 176 122 27 987 27909 78
31 414.4 31 496 -82 31 718 31 844 -126 31 497 31 625 -128
32 972.2 32 862 110 33 139- 33 145 -6 32 963 32 927 36
33 088.4 33 266 -178 33666 33 284
33 240.2 33 218 12 33 629 33 418
33 849.1 33 840 9 34211 34 209 2 33 944 33 921 23
34 733.3 34 534 199 35 055 35 010 45 34 755 34 743 12
36 479.5 36 326 144 36641 36 676 -35 36440 36501 -61
38 406.1 38 375 31 38 827 1 38 816 11
39 268.5 39 026 242 39 530 39 519 11
2111/2 40 808
41 438
41 854.7 42 028 -173
42 426.2 42 710 -284
43 507
46 829
47 420.0 47 532 -112
rms Deviation 141 84 84
• Reference 36. • This energy was assigned to the transition to <G./. in Ref. 37. but in-
b Reference 37. tensity considerations suggest the indicated assignment.
• References 17 and 38. f Not all of the Stark components of this level were observed.
d The parameters used to generate this set of energies are given in
Table II.
ENERGY LEVELS OF TRIVALENT LANTHANIDE AQUa ION S. I 4439
TABLE XI. Energy-level assignments and matrix elements of U(~) for Er3+ (aq).
E..ptl Ecalea tiE

S'[J]' (em-I) (em-I) (em-I) U(2)b U(4)b U(6)b
'11012 108e 109 -'-1

'IIaII 6600 6 610 -10 0.0195 0.1173 1.4316
'IUI2 10 250 10 219 31 0.0282 0.0003 0.3953
'1 912 12 400 12 378 22 0 0.1733 0.0099
'F.12 15 250 15 245 5 0 0.5354 0.4618
'Sill 18 350 18 462 -112 0 0 0.2211
2Hl112 19 150 19 256 -106 0.7125 0.4125 0.0925
4F712 20 450 20 422 28 0 0.1469 0.6266
'Fbll 22 100 22 074 26 0 0 0.2232
'Fall 22500 22 422 78 0 0 0.1272
(2(;, 'F, IH) 912 24 550 24 505 45 0 0.0189 0.2256
4{;1112 26400 26 496 -96 0.9183 0.5262 0.1172
'G 912 27400 27 478 -78 0 0.2416 0.1235
2Klol 2 27801 0.0219 0.0041 0.0758
2(;711 28 000 27 979 21 0 0.0174 0.1163
(2P, 2D, 'F)all 31 600 31 653 -53 0 0 0.0172
'Kla12 33200 33 085 115 0.0032 0.0029 0.0152
'Go12 33400 33 389 11 0 0 0.0026
'Plil 33 453 0 0 0
'G 712 34 050 34 022 28 0 0.0334 0.0029
2Dal2 34850 34800 50 0 0 0.0228
(2H,2(;)912 36 550 36566 -16 0 0.0501 0.0001
2Dal! 38600 38 576 24 0 0 0.0267
'Dl/2 39200 39 158 42 0 0.8921 0.0291
211112 41 150 41 009 141 0.0002 0.0284 0.0034
2L1712 41 650 41 686 -36 0.0047 0.0664 0.0327
'D812 42300 42 257 43 0 0 0.0126
(2D,'PhI2 42966 0 0 0.0002
'Ita/! 43 550 43 717 167 0.0050 0.0170 0.0050
'DIll 47040 0 0 0
2H912 47 822 0 0.0038 0.0001
'LW2 47 916 0.0002 0.0026 0.0021
(2D, 'Dhl! 49000 49033 -33 0 0 0.0096
2Hl11' 51 000 0.0001 0.0082 ~O
rms Deviation 81
a The parameters used to generate this set of energies are !liven b U (X) = (1/;J II U (~) II 1/;' f)'.
in Table U. o Assumed energy based on data in Ref. 38.
TABLE XII. Energy-level assignments and 'matrix elements of U(}.) for Tm8+ (aq).
Tm (C2H~SO.) s· 9H2O" TmH (aq)

E.xptl Eoalob t..E E oxptl Eoal." t..E
S'L'J' (em-I) (em-I) (em-I) (em-I) (em-I) (em-I) U(2)o U(4)o U(6)'
BH6 170 141 29 170d 202 -32

sF. 5846 5 787 59 5 900 5 811 89 0.5375 0.7261 0.2382
aH. 8 250 8 326 -76 8400 8 390 10 0.1074 0.2314 0.6383
aH. 12 665 12 673 -8 12 700 12 720 -20 0.2373 0.1090 0.5947
BFa 14 464 14 514 -50 14500 14510 -10 0 0.3164 0.8411
BF2 15 192 15 112 80 15 100 15 116 -16 0 ~O 0.2581
IG. 21 300 21 306 -6 21 350 21 374 -24 0.0483 0.0748 0.0125
ID2 27 916 28 014 -98 28000 28 032 -32 0 0.3156 0.0928
lIs 34 870 34 851 19 34900 34 886 14 0.0106 0.0388 0.0134
BPO 35444 35 641 -197 35 500 35 637 -137 0 0 0.0756
aPI 36 461 36 295 166 36400 36 298 102 0 0 0.1239
aP2 38 257 38 175 82 38 250 38 193 57 0 0.2645 0.0223
ISo 79 720 79 592 0 0 0.0002
rms Deviation 143 94
• References 39-42. o U (X) = (!f;J II Ul~) II !f;'JI), •

b The parameters used to generate this set of energies are given d Assumed energy based upon data in Refs. 40 and 41.
in Table U.
Using an equation of the same form as that given by The approximate values of F2 obtained from four-
Satten,8 we obtain parameter fits to the data for Eu3+, Gd3+, and Tb3+ 14,15
lie well above the line determined by Eq. (8). Since
F2 = 12.820(Z-34) , (8) several'observed levels for each of these ions are strongly
where a screening constant CT=34 is used in the expres- dependen t on 'Y,'" the effect is probably a reflection of
sion for the effective nuclear charge (Z-CT). the insensitivity of the low-lying levels to the values
of F 2• The H-F calculations of Freeman and Watson43
yield values of F2 that are, except for Pr3+ and Nd3+,
AT- -----® 0.017
o r--'-'---;;r-'-'--'--'~::-'--'-'-
®-- __ -®......... / 41'hydrooenic rotio
~2 0,015
'''®''
0.013
F2
(eM") o 0.011
400
o
0.16
~
-',-e---- __ .-® .----_---3
350 F2 • 12.820 (Z-34)
0.14 =--._._. ___ .______®'.______ -:-
F4 4f·hydrOQenic rotlo
o 0.12 F;
0.10r-
I
300~--~--~--L-~--~--~--~~--~--~ Pr 3 + Nd'+ Pm'· 5m 3+ Eu'''' Gd 3+ Tb 3 + 0)'3+ Ha'· Er 3 + Tm'34-
Pr 3 + Nd 3+ Pm 3+- 5m 3'" Eu 3+ Gd 3+ Tb 3+ 3 + Ho 3+ Er 3+ Tm 3+
OV (59) (61) (63) (65) (67) (69)
(59) 16il (63) (65) (67) (59) ATOMIC NO, Z
ATOMIC NO. Z
FIG. 2. Variation aeross the lanthanide series of the ratios
FIG. 1. Variation across the lanthanide series of the parameter F4/F2 and Fs/F4 calculated from a fit of lanthanide aquo-ion
F2 obtained from a fit to lanthanide aquo-ion data. data.
ENE R G Y LEV E L S 0 F T R I V ALE NT LAN T HAN IDE A QUO ION S. I 4441
3000~--~---r--~--r--.--~---r--'---r--.---.--~,
2800
2600
2400
2200
FIG. 3. Variation across the lanthanide 1800

series of the spin-orbit coupling constant,
r". calculated from a fit to lanthanide
1600
aqua-ion data.
1400
1200
1000
800
ATOMIC NO. Z
a linear function of Z with the same slope as that of experimental data are limited to components of only
the present experimental results shown in Fig. 1. We two or three different multiplets.
find the relationship F2 (H-F)::1.4 F2 [Ln3+ (aquo)], Attempts were made to express !41 both as a linear
where Ln3+ represents those lanthanides heavier than and as a quadratic function of Z. At the outset it was
Nd3+. clear that the average deviation from the experimen-
Values of F4/F2 as a function of Z are quite regular. tal value was appreciably increased when the data for
As shown in Fig. 2, the average is near 0.15 which is Ce3+ and Yb3+ were included, regardless of whether
somewhat larger than the ratios calculated assuming the expression was quadratic or linear. Results exclud-
that 4/ wavefunctions are hydrogenic [Eq. (4)] or by ing Ce3+ and Yb3+ showed that the average deviation
using H-F methods (::0.130).43 was appreciably less when two linear functions with a
There is an apparent change in the ratio F s/F 2 from common value at Gd3+ were employed as compared to
somewhat less than the hydrogenic ratio [Eq. (5) ] for a single quadratic function. The linear functions were
the light lanthanides to greater than this ratio for the
heavy members of the series (Fig. 2). Hartree-Fock !4/=142Z-7648 (Pr3+ through Gd3+) , average devi-
calculations give Fs/F2::0.0139.43 There is no obvious a tion from exptl = 11.3,
explanation for this nonlinear variation in the experi- !4/=231.142Z-13330.3 (Gd3+ through Tm3+) , aver-
mental results, but the parameters are well-enough age deviation from exptl=9.5.
determined to suggest that the effect is real.
Spin-orbit coupling constants for most of the tri-
Spin-Orbit Coupling Constant C!;4/) valent lanthanides have been calculated by Blume,
Freeman, and Watson, using analytic nonrelativistic
The variation of !4/ across the lanthanide series is H-F functions. 46 These authors have discussed at length
shown in Fig. 3, where the values for Ce3+ 44 and Yb3+ 13 the fact that the calculated coupling constants are
have also been included. A good approximation to the larger than those obtained experimentally. We found
values of this parameter14 can be obtained for Eu3+, that the calculated values only differ from those given
Gd3+, and Tb3+ even though the assignments to the in Fig. 3 by a scaling factor. Two linear functions of Z,
"R. A. Buchanan, H. E. Rast, and H. H. Caspers, J. Chem. "M. Blume, A. J. Freeman, and R. E. Watson, Phys. Rev.
Phys. 44,4063 (1966). 134, A32 (1964).
40 in these parameters as a function of Z in the 4fN series

,0 is much greater than that observed in the dN series.
20
/0
0
-\7 We note that in general, the magnitude of {J/l2 is
approximately twice that of a.
Where it was possible to define a value for 'Y, we
found it to be positive, and the average was near
1100-1200 em-l. The only previous determination of
/20 'Y was from the free-ion spectrum of Pr2+, and in this
-fJ/12
-\:=t_~I-
I 80 case a negative value for the parameter was obtained. 24
::;; l
<.)
40 The latter result depended upon only one somewhat
doubtful level.
0
In view of the present results, a negative value for
'Y in Pr2+ seems very unlikely. It is highly probable
6000
y
that the 2F6f2 level from which the value of 'Y was
4000 determined is spurious. This view is further supported
by the calculation of Trees22 which indicated that 'Y
2000
~!--I should be positive.
0 Recent H-F calculations for Pr2+ show that the major
Pr Nd Pm Sm Eu Gd Tb Oy Ho Er Tm
(59) (61) (63) (65) (67) (69) contribution to 'Y comes from interaction with 4d84f6.48
ATOMIC NO. Z
Although there are negative terms arising from inter-
FlG. 4. Variation across the lanthanide series of the configura- action with configurations such as 4f6p and 4p64f4,
tion-interaction parameters a, {3, and l' calculated from a fit to these are not sufficient to overcome the large positive
lanthanide aquo-ion data. contributions from 4d84f5.
CONCLUDING REMARKS
having the same slope as those of Fig. 3, but with a At present it is possible to compare the solution pa-
common value at Eu3+ instead of Gd3+, give an excellent rameters determined here with those of the free-ion
fit to the H-F data. only for Pr+ and Er+. l9 ,36 One would expect covalent
bonding effects in solution to reduce the parameters
Configuration-Interaction Parameters (a, {J, 'Y) slightly from those obtained for the free ion. The reduc-
tion is about 5% for the Ek and t41 values of Pr+.
The systematic behavior of the configuration inter- The configuration-interaction parameters are somewhat
action parameters has been studied previously only in more sensitive. The Ek and t4/ values for Er+ are
the 3dN and 4dN configurations. By assuming that each slightly higher for the solution than for the free ion,
of the parameters follows an equation of the form while a and {J are lower. However, the rms errors on
the parameters are nearly the same as the differences,
so the effect may not be significant. We would expect
the present results to provide a good starting point for
where x=N-6 and y=x2-l0, Shadmi46 ,47 was able further analyses of both the crystal and free-ion spec-
to fit simultaneously the energy levels of an entire tra of 4JN configurations.
series by the determination of only the "general pa-
rameters," Ai. In doing this he assumed that a and {J ACKNOWLEDGMENTS
vary linearly with N, i.e., A 2 =O. However, if one com-
pares the values of ex obtained on the basis of this We gratefully acknowledge the large contribution of
assumption with those obtained as a result of individual B. G. Wybourne, University of Canterbury, Christ-
fits to the levels of a single ion, there are frequently church, New Zealand, which made much of this work
large differences well outside the rms deviation of the possible. We would also like to thank F. L. Clark, C.
parameters. 47 Thus, a linear variation of a with N is Chamot, and S. Katilavas for their assistance in pro-
not apparent in the case of the single-ion fits. gramming and carrying out the computer calculations.
In the case of the lanthanides, the absolute magni- We are indebted to J. G. Conway and T. Clements,
tudes of a and {J appear to increase in the first half of Lawrence Radiation Laboratory, University of Cali-
the series after Pr3+, and decrease in the second half fornia, for providing us with a number of computer
(Fig. 4). It is clear that the magnitude of the variation programs and tapes used in this work.
46 Y. Shadmi, Phys. Rev. 139, A43 (1965). 48 M. Wilson and K. Rajnak (pri vate communication, November
47 y. Shadmi, J. Res. Nat!. Bur. Std. A70, 435 (1966). 1967) .

Carnall 3

Uploaded by

Document Informationclick to expand document informationElectronic Energy Levels in the Trivalent Lanthanide Aquo Ions. I. Pr3+, Nd3+, Pm3+, Sm3+, Dy3+, Ho3+, Er3+, and Tm3+

Document Informationclick to expand document information

Copyright:

Available Formats

Carnall 3

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Carnall 3

Uploaded by

Copyright:

Available Formats

Electronic Energy Levels in the Trivalent Lanthanide Aquo Ions. I.

Citation: J. Chem. Phys. 49, 4424 (1968); doi: 10.1063/1.1669893

Additional information on J. Chem. Phys.

Electronic Energy Levels in the Trivalent Lanthanide Aquo Ions.

TABLE III. Energy-level assignments for Pt'+.

Pt'+ Free ion a Pt'+ in LaClab Pt'+in LaF.·

3H. 0 20 -20 99.3 118 . -9 200 228 -28

3H. 200° 245 -45

N dH in LaClaa NdH (aq)

Euptl Eoa1.b I1E E ••ptl E.a1.b I1E

41911 146 178 -32 146 d 130 16

a Reference 26. • U (1\) = (1/11 II u(X) 11.p'J') t.

E ••v'] Eoaloa llE

SmH in LaClaa SmH (aq)

Eexpd Eoalob t.E E ••pi! Eoal.b t.E

6H5/1 36 12 24 36d 46 -10

TABLE VII ( Continued)

SmH in LaCl s" SmH (aq)

Enptl Eea",b dE E•• vtl E •• lob dE

2K18/2 33800 33 825 -25 0 0.0002 0.0001

'G9/2 44 832 .....,0 ",0 0.0002

Dy8+ in LaC)." Dya+ (aq)

TABLE VIII (Continued)

DyH in LaCla" Dy3+ (aq)

EeIJltl Eoalob t:.E

41 050 41 035 15 0 0.0015 0.0002

• See Table II. Footnote f. • u 0.) = (Il-'J II Uo.) IIljIJ')t.

Hos+ in LaCl3" H 0 3+ (aq)

E•• ptl E ... lcb AE E.xp'l Ealcb tJ.E

51a 107.7 140 -32 108d 80 28

TABLE X. Energy-level assignments for E1"+.

o 96 -96 217 232 -15 108 130 -22

E..ptl Ecalea tiE

'11012 108e 109 -'-1

Tm (C2H~SO.) s· 9H2O" TmH (aq)

BH6 170 141 29 170d 202 -32

• References 39-42. o U (X) = (!f;J II Ul~) II !f;'JI), •

AT- -----® 0.017

FIG. 3. Variation across the lanthanide 1800

40 in these parameters as a function of Z in the 4fN series

You might also like