Course of lectures on

magnetism of lanthanide ions

under varying ligand and magnetic fields
RWTH Aachen University, February, 18-21, 2008
Heiko Lueken
Institute of Inorganic Chemistry

Contents
1 Magnetic properties of free lanthanide ions: an overview
1.1 Electronic structure and Hunds rules . . . . . . . . . . . . . . . . . . .
1.2 Energetic effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3 Magnetic behaviour of Ln ions . . . . . . . . . . . . . . . . . . . . . . .
1.3.1 Ground multiplet thermally isolated (E(J, J 1) kB T ) . .
1.3.2 Ground multiplet thermally not isolated (E(J, J 1) kB T ) .
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.
.
.
.
.
.

.
.
.
.
.
.

3
3
3
4
4
5
6

2 Theory of free lanthanide ions

2.1 Functions . . . . . . . . . . . . . . . . .
2.2 Angular momenta . . . . . . . . . . . . .
2.2.1 Orbital angular momentum . . .
2.2.2 Spin . . . . . . . . . . . . . . . .
Problems . . . . . . . . . . . . . . . . . .
2.3 Quantum-mechanical procedures . . . . .
2.3.1 Operators . . . . . . . . . . . . .
2.3.2 Perturbation theory . . . . . . . .
2.4 Applications . . . . . . . . . . . . . . . .
2.4.1 Spin-orbit coupling (p1 , f 1 , f 13 ) .
Problems . . . . . . . . . . . . . . . . . .
2.4.2 Magnetic susceptibility, Van Vleck

.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.

7
7
10
10
12
12
12
12
13
17
17
23
23

.
.
.
.
.
.
.

26
26
27
27
29
30
30
30

. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
equation .

.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.

3 Lanthanide ions in cubic and non-cubic ligand fields

3.1 Ligand-field (LF) operators . . . . . . . . . . . . . . .
3.2 Introduction: p1 system . . . . . . . . . . . . . . . . .
cyl
3.2.1 Cylindrical LF: HLF
+ HSO . . . . . . . . . . .
or
3.2.2 Orthorhombic LF: HLF
+ HSO . . . . . . . . . .
3.3 Cubic LF . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.1 Group theoretical tools I . . . . . . . . . . . . .
cub
3.3.2 4f 1 system (Ce3+ , 2F : HSO + HLF
) . . . . . . .
1

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.

3.3.3 4f N systems . . . . . . . . . . . . . . . . . . .
3.3.4 Magnetic behaviour of Cs2 NaLnCl6 (overview)
cub
3.3.5 4f 2 system (Pr3+ , 3H4 : HLF
) . . . . . . . . .
2
3+ 3
cub
3.3.6 4f system (Pr , H4 : HLF + Hmag ) . . . . .
3.3.7 Prediction of LF splittings . . . . . . . . . . .
Problems . . . . . . . . . . . . . . . . . . . . . . . . .
Non-cubic LF . . . . . . . . . . . . . . . . . . . . . .
3.4.1 Ligand field operators . . . . . . . . . . . . .
cyl
3.4.2 4f 1 system (Ce3+ , 2F5/2 : HLF
) . . . . . . . . .
cyl
1
3+ 2
3.4.3 4f system (Ce , F5/2 : HLF + Hmag ) . . . .
Problems . . . . . . . . . . . . . . . . . . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.

35
37
39
41
42
52
53
53
53
54
56

4 Isotropic spin-spin exchange couplings in dinuclear compounds

4.1 Parametrization of exchange interactions . . . . . . . . . . . . . . .
4.2 Heisenberg operator . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3 Exchange-coupled species in a magnetic field . . . . . . . . . . . . .
4.4 Setup of computational procedures . . . . . . . . . . . . . . . . . .
4.5 Homodinuclear lanthanide complexes . . . . . . . . . . . . . . . . .
4.5.1 Gd3+ Gd3+ . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.5.2 Group theoretical tools II . . . . . . . . . . . . . . . . . . .
4.5.3 Ce3+ Ce3+ . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.5.4 Nd3+ Nd3+ (cub) . . . . . . . . . . . . . . . . . . . . . . . .
4.5.5 Sm3+ Sm3+ (cub) . . . . . . . . . . . . . . . . . . . . . . . .
4.5.6 Tb3+ Tb3+ (cub) . . . . . . . . . . . . . . . . . . . . . . . .
4.5.7 Dy3+ Dy3+ (cub) . . . . . . . . . . . . . . . . . . . . . . . .
4.5.8 Dy3+ Dy3+ (cyl) . . . . . . . . . . . . . . . . . . . . . . . .
4.5.9 LnLn pair: Temporary resume of eff for Jex > 0 . . . . . .
4.6 Heterodinuclear transition metallanthanide systems . . . . . . . .
4.6.1 p(2 P1/2 )s(2 S1/2 ) molecule . . . . . . . . . . . . . . . . . . .
4.6.2 Gd3+ Cu2+ . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.3 Ce3+ (cub)Cu2+ . . . . . . . . . . . . . . . . . . . . . . . .
4.6.4 Nd3+ (cub)Cu2+ ) and Nd3+ (cub)Cr3+ . . . . . . . . . . . .
4.6.5 Dy3+ (cub)Cu2+ , Dy3+ (cyl)Cr3+ , Tb3+ (cyl)Cr3+ . . . . .
4.6.6 Gd3+ V3+ . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

56
56
59
59
63
64
64
65
67
69
70
70
71
72
72
73
73
75
76
78
79
80
82

3.4

.
.
.
.
.
.
.
.
.
.
.

5 Solutions

.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.

83

6 Appendix 1
87
6.1 Magnetic quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
6.2 Energy equivalents and conversion factors . . . . . . . . . . . . . . . . . . 88
7 Appendix 2: Matrix elements of operator equivalents

89

References

91

Magnetic properties of free lanthanide ions:

an overview

1.1

Electronic structure and Hunds rules

The 4f electrons of free Ln ions are influenced by nuclear attraction, interelectronic repulsion (Hee ) and spin-orbit coupling (HSO ), where Hee > HSO . To determine the free ion
ground multiplet 2S+1 LJ the following scheme is helpful (N: no. of 4f electrons):
Ln3+

Ce

Pr

Nd Pm
3

Sm

Eu

Gd

Tb

Dy Ho

Er

Tm

Yb

Lu

11

12

13

14

12 21

12

12 12

2 3

10

ms

+ 12 + 12 + 12

+ 12

+ 12

+ 12

+ 12

21

21

ml

+3 +2 +1

2 3

+3

+2 +1

1. The term with maximum S lies lowest in energy (

ms,i = MS S).

P
2. For a given spin multiplicity, the term with highest L lies lowest in energy ( i ml,i =
ML L).
3. For less than half-filled subshells, the level with the lowest value of J lies lowest
(J = |L S|), while the highest J lies lowest when a subshell is more than half full
(J = L + S).
Examples:

1.2

Pr3+
Dy3+

[4f 2 ]:
[4f 9 ]:

S = 1,
S = 25 ,

L = 5,
L = 5,

J = 4;
J = 15
;
2

(3H4 )
(6H15/2 )

Energetic effects

The energy splittings of the electronic states of a magnetic system 1) are produced by Hee ,
HSO , the ligand field (HLF ), interatomic exchange interactions (Hex ), and the applied
magnetic field (Hmag ). The following scheme shows the order of Hee , HSO , and HLF .

3dN

4f N

Hee

>

HLF

>

HSO

weak field

HLF

>

Hee

>

HSO

strong field

HLF

Hee

>

HSO

intermediate field

Hee

>

HSO

>

HLF

(a)

Hee

>

HSO

HLF

(b)

(a) strong field lanthanide system ; (b) weak field lanthanide system

Table 1 gives approximate sizes of the corresponding energy splittings, including Hex
and Hmag .
1)

Localised electrons are presupposed.

Table 1 Effects acting on d and f electrons.

Effect
Electron-electron interaction

Ligand-field potential

System
3d, 4d, 5d

Hee

4f, 5f

HLF

4f

3d, 4d, 5d

HSO

4f, 5f
nd

Exchange interaction

Hex

4f
nd4f

Magnetic field
a)
b)

wavenumber/cm1 a)
3d > 4d > 5d 104
4f > 5f 104

b)

b)

3d, 4d, 5d 3d < 4d < 5d 2 104

5f

Spin-orbit coupling

Energy equivalent

Hmag

b)

102
103

3d < 4d < 5d 103

4f < 5f 103

b)

b)

102
<1

< 10
0.5 (1 T)

Conversion to other units, see Table 34 (Appendix).

Approximate value according to ref. [1].

Inspecting Table 1, two points are important for 4f ions: (i) Hex is at least two orders of
magnitude smaller than HLF (and HSO ). (ii) Hmag has to be chosen carefully on account
of comparable magnitude of Hex and Hmag . Therefore, in order to detect Hex , HLF should
be modelled most exactly and Hmag should be very weak.

1.3

Magnetic behaviour of Ln ions

The variation of m , (permanent magnetic moment corresponding with the Curie constant C), eff (effective Bohr magneton number2) ), m T , Mm as a function of T and B0
serves to characterize the low-lying electronic states of the magnetic systems.
1.3.1

Ground multiplet thermally isolated (E(J, J 1) kB T )

Neglecting HLF , Hunds formula eq. (1)

gJ

J(J + 1) B

(except 4f 4 , 4f 5 , 4f 6

systems)

(1)

applies, where J is the total angular momentum of the ground multiplet and gJ the Lande
factor
gJ = 1 +

J(J + 1) + S(S + 1) L(L + 1)

.
2J(J + 1)

(2)

S, L, and J correspond to the total spin angular momentum, the total orbital angular
momentum and the total angular momentum, respectively, of the ground multiplet.
2)

Note that the empirical number eff has no connexion with the permanent moment except when
Curies law is obeyed [2].

A graph showing the variation eff vs. T is recommended for 4f/5f systems 3 ) . According to Kramers rule [4], ions with an odd number of 4f electrons (half integral values
of J) have always a degenerate ground state obeying the Curie law at low temperature
(regardless the symmetry of HLF , but presupposed that (i) solely the ligand field (LF)
ground state is occupied and (ii) cooperative magnetic effects are absent). For an even
number of f electrons, degenerate as well as singlet ground states can be found depending on the symmetry of HLF . A singlet ground state produces temperature independent
paramagnetism (TIP) at low temperature.
Going to 5f electron systems, HSO and HLF increase according to the larger effective
nuclear charge and according to the fact that 5f electrons are more accessible for ligands.
No simple approximation can be made. So, the only possibility to predict the magnetic
behaviour for 5f ions are computational methods.
1.3.2

Ground multiplet thermally not isolated (E(J, J 1) kB T )

General equation: m

where J


LS J(J + 1)
(2J + 1)J exp
2
kB T
NA 2B J=LS


= 0
(3)
L+S
P
3kB T
LS J(J + 1)
(2J + 1) exp
2
kB T
J=LS

kB T
and LS =
= gJ2 J(J + 1) + 2(gJ 1)(gJ 2)
LS
2S
L+S
P

The gJ values are given by eq. (2) (exception: gJ = L + 2 for J = 0). The term spin-orbit
coupling parameter LS is positive for ions with less than half-full subshell and negative
for ions with more than half-full subshell.
Example 1.1 4f 6 (Sm2+ , Eu3+ )
NA 2B 2

where
3kB T eff





kB T

27
kB T
2
1
144
eff = Z
exp
+
9

6k T


  B





kB T

kB T
135
exp
+ 189 21
exp
15
+
2

2kB T

k T



 

 B

kB T
5
5
1485
kB T
+ 405 27
exp
+
exp
33

3kB T
2

2kB T




2457
7
kB T
+
exp
and
39
2

2kB T








+ 5 exp
+ 7 exp
Z =
1 + 3 exp
6kB T
2kB T
kB T






5
5
7
+9 exp
+ 11 exp
+ 13 exp
3kB T
2kBT
2kB T
m = 0

3)

(4)

We recommend this plot or alternatively 2eff vs. T instead of the m T vs. T plot for two reasons:
(i) The number eff has the same numerical value in the SI and CGS-emu
system, (ii) eff is easily
p
compared with the lanthanides corresponding free-ion value eff = gJ J(J + 1) [5].

Eq. (4) is calculated with gJ = 3/2 for the multiplets J = 1, 2, 3, 4, 5, 6, gJ = 5 for J = 0,

and LS = /6.
Table 2: Lanthanide ions: term symbol (ground state), one-electron spin-orbit
coupling parameter 4f [cm1 ], gJ , gJ J, gJ [J(J + 1)]1/2 and exp
eff (295 K) [7]

Ln3+

4f N

2S+1

LJ

4f a)

b)
gJ J gJ [J(J + 1)]1/2 exp
eff

gJ

La3+c) 4f 0

S0

Ce3+

4f 1

F5/2

625

6/7

15/7

2.535

2.32.5

Pr3+

4f 2

H4

758

4/5

16/5

3.578

3.43.6

Nd3+

4f 3

I9/2

884 8/11 36/11

3.618

3.43.5

Pm3+

4f 4

I4

1000

3/5

12/5

2.683

2.9d)

Sm3+

4f 5

H5/2

1157

2/7

5/7

0.845

1.6

Eu3+

4f 6

F0

1326

3.5

Gd3+

4f 7

S7/2

1450

7.937

7.87.9

Tb3+

4f 8

F6

1709

3/2

9.721

9.79.8

Dy3+

4f 9

H15/2 1932

4/3

10

10.646

10.210.6

Ho3+

4f 10

I8

2141

5/4

10

10.607

10.310.5

Er3+

4f 11

I15/2

2369

6/5

9.581

9.49.5

Tm3+

4f 12

H6

2628

7/6

7.561

7.5

Yb3+

4f 13

F7/2

2870

8/7

4.536

4.5

Lu3+b) 4f 14

S0

a)
The relation between 4f and LS of the Russell-Saunders ground term is given by LS =
(4f /2S), where (+) and () sign correspond to N 2l + 1 and N 2l + 1, respectively.
b)
Definition: m = 0 NA 2B 2eff /(3 kB T ).
c)
diamagnetic
d)
observed for Nd2+ compounds.

Problems
1. Verify the Russell-Saunders ground multiplets for Nd2+ and Er3+ given in Table 2.
2. Evaluate the susceptibility equation for a free p1 system perturbed by spin-orbit
coupling. (Hint: Apply eq. (3). Note that for the single-electron system LS .
Solutions are given in section 5.)
6

Theory of free lanthanide ions

2.1

Functions

Schrodinger equation (spin ignored):

h

i
h
2 2
eV (r) (r) = E(r)
2me

(5)

For convenience the eigenfunctions (atomic orbitals) are given in spherical polar coordinates:
z

x = r sin cos
y = r sin sin
z = r cos

P(x,y,z)
r. cos q
q
r. sin q. cos f

.
f r sin q

r. sin q. sin f

(6)

r 2 = x2 + y 2 + z 2
cos = z/r
tan = y/x

(7)

Fig. 1:
Relation between cartesian coordinates and spherical polar coordinates

Yml l (, )

(r) = n,l,ml (r, , ) = Rn,l (r)

=
| {z } | {z }
radial f. angular f.

Rn,l (r) lml()

1 iml
e
2

The functions Yml l (, ) are spherical harmonics specified by the quantum numbers l
and ml (Table 3). Real functions (Table 4) are gained by linear combinations
1
2
1

i 2

[n,l,ml + n,l,ml ] =
[n,l,ml n,l,ml ] =

1
2
1

i 2

[n,l,ml + n,l,ml ] =
[n,l,ml n,l,ml ] =

1 Rn,l (r)l
ml ()

1
Rn,l (r)l
ml ()

1 Rn,l (r)l ()
ml

1 Rn,l (r)l ()
ml

cos ml

ml odd

sin ml

ml odd
(8)

cos ml

ml even

sin ml

ml even

For a complete description of the wave function, the spin has to be taken into consideration. Ignoring spin-orbit coupling, the total function (spin orbital) reads
(r, , ; ) = (r, , ) () where = 21 .

(9)

Table 3: Spherical harmonics for l = 0, 1, 2, 3

l ml
0

0
0

1
1
0
2 1
2
0
1

2
3
0
1
4 2
3
4
a)

Yml l (, ) a)

Yml l (x, y, z)

 1 1/2
4
 3 1/2
cos
4
 3 1/2
sin ei

8
 5 1/2
(3 cos2 1)
16
 15 1/2
cos sin ei

8
 15 1/2
sin2 ei2
32
 7 1/2
(5 cos3 3 cos )
16
 21 1/2
sin (5 cos2 1)ei

64
 105 1/2
cos sin2 ei2
32
 35 1/2

sin3 ei3
64
 9 1/2
(35 cos4 30 cos2 + 3)
256
 45 1/2

sin (7 cos3 3 cos ) ei

64
 45 1/2
sin2 (7 cos2 1) ei2
128
 315 1/2
sin3 cos ei3

64
 315 1/2
sin4 ei4
512

 1 1/2
4
 3 1/2 z
4
r
 3 1/2 x iy

8
r
 5 1/2 3z 2 r2
16
r2
 15 1/2 z(x iy)

8
r2
 15 1/2 (x iy)2
32
r2
 7 1/2 z(5z 2 3r2 )
16
r3
 21 1/2
(5z 2 r2)
(x iy)

64
r3
 105 1/2 z(x iy)2
32
r3
 35 1/2 (x iy)3

64
r3
 9 1/2 (35z 4 30z 2r2 + 3r4)
256
r4
 45 1/2
(7z 3 3zr2 )

(x iy)
64
r4
 45 1/2
2
2
2 (7z r )
(x iy)
128
r4
 315 1/2 z(x iy)3

64
r4
 315 1/2 (x iy)4
512
r4

Phase factors correspond to the Condon-Shortley convention: 1 for odd positive ml and +1 otherwise.

Table 4: Real orthonormal linear combinations of the spherical harmonics Yml l (, ) for
l = 1, 2, 3.

function

 3 1/2
cos
4
 3 1/2
1
sin cos
4
 3 1/2
sin sin
4
 5 1/2
(3 cos2 1)
16
 15 1/2
cos sin cos
4
 15 1/2
2
cos sin sin
4
 15 1/2
sin2 cos 2
16
 15 1/2
sin2 sin 2
16
 7 1/2
(5 cos3 3 cos )
16
 21 1/2
sin (5 cos2 1) cos
32
 21 1/2
sin (5 cos2 1) sin
32
 105 1/2
3
cos sin2 sin 2
16
 105 1/2
cos sin2 cos 2
16
 35 1/2
sin3 cos 3
32
 35 1/2
sin3 sin 3
32

designation
 3 1/2 z
=
4
r
 3 1/2 x
=
4
r
 3 1/2 y
=
4
r
 5 1/2 3z 2 r2
=
16
r2
 15 1/2 xz
=
4
r2
 15 1/2 yz
=
4
r2
 15 1/2 x2 y 2
=
16
r2
 15 1/2 xy
=
4
r2
 7 1/2 z(5z 2 3r2)
=
16
r3
 21 1/2 x(5z 2 r2 )
=
32
r3
 21 1/2 y(5z 2 r2 )
=
32
r3
 105 1/2 xyz
=
4
r3
 105 1/2 z(x2 y 2 )
=
16
r3
 35 1/2 x(x2 3y 2)
=
32
r3
 35 1/2 y(3x2 y 2 )
=
32
r3
9

pz
px
py
dz 2
dxz
dyz
dx2 y2
dxy
fz 3
fxz 2
fyz 2
fxyz
fz(x2 y2)
fx(x2 3y2)
fy(3x2 y2 )

2.2
2.2.1

Angular momenta
Orbital angular momentum

Classical definition of the angular momentum l (Fig. 2):

l = r p.

(10)

r
p
Fig. 2: Definition of the angular momentum
l = lx i + ly j + lz k
= (ypz zpy )i + (zpx xpz )j + (xpy ypx )k

(11)

Length of the angular momentum vector:

|l|2 = lx2 + ly2 + lz2 .

(12)

Quantum mechanical operators lx , ly , lz are derived by substituting the position operator

and the linear momentum operator for the corresponding classical quantity, i. e.
q q = q

pq pq =

h

i q

(q = x, y, z;

i=

1)

lx = ypz zpy ;
ly = zpx xpz ;
lz = xpy ypx







lx = h

y z
; ly =
z x
; lz =
x
y
i
z
y
i
x
z
i
y
x

(13)
(14)

Commutation relations
[lx , ly ] = i
hlz ;

[ly , lz ] = i
hlx ;

[lz , lx ] = i
hly ,

(15)

Operator lz in spherical polar coordinates:

lz = h
i

(16)

lz acts on the depending part of the atomic orbitals (Table 3):







lz  l ml
= lz  l ml = ml h
 l ml
Dirac notation
 



l ml lz  l ml
= ml h
l ml  l ml
{z
}
|
1

10

 = E , ( normalised eigenfunction of H)

generally H
Z
Z
d = E d = E
H
| {z }
1
Z
d h|H|i

H
matrix element (Dirac notation)
Application of lz :

 
lz  2 2

=
=
=
=



l2  l ml =

h
Y22 (, )
i

#
"
1/2
h

15
Rn,2 (r)
sin2 ei2
i
32

1/2
h
15
ei2
Rn,2 (r)
sin2
i 32



1/2
 
h
15
Rn,2 (r)i2
sin2 ei2 = 2
h22
i 32


l(l + 1)
h2  l ml
Rn,2 (r)

(17)

Shift operators:

l+ = lx + ily ;

l = lx ily .

(18)

Reverse operations:
ly = 1 (l+ l ).
lx = 1 (l+ + l );
2
2i




lz  l ml
= ml h
 l ml




l2  l ml
= l(l + 1) h
2  l ml


p


l  l ml
l(l + 1) ml (ml 1) h
 l ml 1 .
=

(19)

(20)

ml =+2
+1

Fig. 3: Specified orientation of l (l = 2)

with regard to the component lz while lx and
ly are unspecified

1
2

11

2.2.2

Spin

Spin orbital of a one-electron system:

(r) () = n,l,ml (r, , ) () = Rn,l (r) Yml l (, ) ()
|
{z
}
atomic orbital
(

ms = 12 :
Spin function: () | s ms i
ms = 12 :




s2  s ms = s(s + 1) h
2  s ms with s = 21




 s ms with ms = 12
sz  s ms = ms h


p


s(s + 1) ms (ms 1) h
 s ms 1
s  s ms =

(21)

where s2 = s2x + s2y + s2z ; s+ = sx + i

sy ;
s = sx i
sy
1
1
s+ + s ); sy = 2i (
s+ s )
sx = 2 (
Problems
3. Use the spherical harmonics in Table 3 and construct the real functions dx2 y2 and
dxy with the help of eqs. (8). Verify the results with the data given in Table 4.

2.3

Quantum-mechanical procedures

2.3.1

Operators

Postulate 1. The state of a system is fully described by the wavefunction (r1 , r2 , . . . , t).
Postulate 2. Observables are represented by operators chosen to satisfy the commutation
relation

(22)
qpq pq q = [
q , pq ] = i
h
(q = x, y, z; i = 1)
Example 2.1 Verification of the commutation relation
x = x

and px =

h
d
i dx

h
d
i dx


h
d(x)
d
h

px x =
+x
=
i dx
i
dx
h

(
xpx px x) = = i
h
i
xpx = x

12

Tab. 5: Classical and quantum-mechanical forms of Ekin and Epot

quantity Da)
Ekin

1
3

Epot c)

a)
b)
c)

classical
p2x
me vx2
=
2
2me
2
p
2me

quantum-mechanical

2
h
d
h
2 d2
1
p2x
=
=
2m
2me i dx
2me dx2
e

2
2
2
h
2 2
h
2
+
+

2me x2 y 2 z 2
2me

eV (x)

eV (x) = eV (x)

eV (r)

eV (r) = eV (r)

b)

Dimension.
is the Nabla operator.
Valid for one electron with charge e in the potential V .

2.3.2

Perturbation theory

1. Non-degenerate states
(0) (0) = E (0) (0)
H
n
n
n

unperturbed system

(23)

Hamilton operator of the perturbed system:

=H
(0) + H
(1)
H
Schrodinger equation of the perturbed system:
n = En n ;
H

find En , n

(24)

Series expansion of n und En :

(1)
2 (2)
n = (0)
n + n + n + . . .
En = En(0) + En(1) + 2 En(2) + . . . .

(25)
(26)

Insert the series into eq. (24):

(0) + H
(1) )((0) + (1) + 2 (2) + . . .) =
(H
n
n
n
(0)
(1)
2 (2)
(1)
2 (2)
(En + En + En + . . .)((0)
n + n + n + . . .),
Ordering of the terms with regard to powers of :




(0) (2) + . . . =
(1) (1) + H
(0) (0) + H
(1) (0) + H
(0) (1) + 2 H
H
n
n
n
n
n


(0) (1)
(1) (0)
+
En(0) (0)
n + En n + En n


(0) (2)
(1) (1)
+ ....
2 En(2) (0)
n + En n + En n
13

2
..
.

(0) (0) = E (0) (0)

H
n
n
n
(0) E (0) )(1) = (E (1) H
(1) )(0)
(H
n
n
n
n
(0)
(0)
(2)
(2) (0)
(1)

(1) )(1)
(H En )n = En n + (En H
n
..
.

(27)
(28)
(29)

(1)

The first-order correction to the energy En

(0)
(premultiply both sides of eq. (28) with n and integrate)
Z
Z
Z
(0) (1)
(0) (0) (1)
(0)
(1)
(1) (0)
n n d = En (0)
n H n d En
n H
n d
|
{z
}
0


(1) |n
En(1) = n|H

(30)

The first-order correction to the wavefunction:


(1) |n
X m|H
(1)
(0)
n =
m
(0)
(0)
m6=n Em En

(31)

The second-order correction to the energy:




(1) |n 2
X  m|H
En(2) =
(see Fig. 4)
(0)
(0)
m6=n Em En

(32)

E
Y2(0)
E 2(0)

Y2(0)

H12

(0)

E 1(0)

E 2(0)

E 2(0)

H22

Y1

E 2(0) H22

H21 H12
E 1(0)

E 2(0)

Y1(0)
E 1(0)
Y1(0)

H21

(0)

E 2(0)

E 1(0)

E 1(0)

H 11

Y2

E 1(0) H11

(a)

H12 H21
E 2(0)

E 1(0)

(c)

(b)

Fig. 4: Illustration of the possible effects of a perturbation on two nondegenerate levels; (a) 0th ,(b) 1st ,(c) 2nd order
14

2. Degenerate states
Eq. (30) (32) apply also in this case; additionally: determination of the correct zerothorder wavefunctions (see Fig. 5)
Example: doubly degenerate pair of states
(0) (0) = E (0) (0)
H
n
n,i
n,i
(0)

(i = 1, 2)

(33)

(0)

(0)
n = u1 n,1 + u2 n,2

(34)

The correct linear combinations are those which correspond to the perturbed functions
for 0.
Determination of u1 and u2 :

Substituting eq. (34) in eq. (25); eq. (28) now reads:

(0)

(0) E (0) (1) = E (1) H
(1) u1 (0)
H
n,1 + u2 n,2
n
n
n
(0)

(35)

(0)

Multiply with n,1 and n,2 , respectively and integrate:

u1 H11 En(1) + u2 H12 = 0


u1 H21 + u2 H22 En(1) = 0

where Hij =


(0) (1) (0)
(1) |j
n,i H
n,j d = i|H

To find the non-trivial solutions of this pair of equations, the determinant of the coefficients
of the constants u1 and u2 must disappear:



 H E (1)
H12
H11 , H22 : diagonal elements
n

 11
(36)

(1)  = 0;

H21
H22 En 
H12 , H21 : off-diagonal elements

p


(1)
En(1,2) = H11 + H22 /2 (H11 H22 )2 /4 + |H12 |2
(1)

u1(1,2) (H11 En(1,2) ) + u2(1,2) H12 = 0;

x(1,2) =

u1(1,2)
H12
=
(1)
u2(1,2)
H11 E

n(1,2)

Normalisation:

x2(1,2) u22(1,2) + u22(1,2) = 1

q
u
=

2(1,2)

x2(1,2) + 1

x(1,2)

u1(1,2) = x(1,2) u2(1,2) = q

x2(1,2) + 1
(1)

Correct zeroth-order wavefunctions for the energy En(1,2) :

(0)

(0)

(0)

n(1,2) = u1(1,2) n,1 + u2(1,2) n,2

15

E
Y2(0)

Y2(0)

E 2(0)

E 2(0)

(0)

(0)
u 2(2) Y1,2

(0)

(0)
u 2(1) Y1,2

(0)
Y1(2) = u 1(2) Y1,1
(0)
(0)
Y1,1
, Y1,2

E 1(0)

E 2(0) H22

(0)
Y1(1) = u 1(1) Y1,1

E 2(0)

H22

E 1(0)

E 1(2)

H21(1) H1(1)2 H21(2) H1(2)2

E 1(0)

(1)

H21(2) H1(2)2

(1)

E 1(0) E 1(2)

E 1(0)

E 1(0)

E 2(0)

(1)

H21(1) H1(1)2

(1)

(a)

E 1(0)

E 1(1)

E 1(0) E 1(1)

(a)

E 2(0)

(b)

E 2(0)

E 1(0)

(c)

Fig. 5: Illustration of the possible effects of a perturbation on a doubly degenerate ground state and a non-degenerate excited state; (a) 0th , (a) correct
0th , (b) 1st ,(c) 2nd order

Alternative procedure to solve 2 2 determinants (cf. [11], p. 119):

General
solution for 2 2 determinants:

 H E
H12 
 11
tan 2 = 2H12 /(H11 H22 )

 = 0.
 H12
H22 E 
(
H11 H12 cot ; 1 = sin 1 cos 2
(1) |j i
E=
; Hij = hi |H
H22 + H12 cot ; 2 = cos 1 + sin 2
(37)

16

2.4

Applications
Spin-orbit coupling (p1 , f 1 , f 13 )

2.4.1

Example 2.2 p1 system

SO = (r)
H
l
s

where

(r) =

p1 system: l = 1, s = 12 , j1 = l s =

1
2

e 1 V (r)
.
2m2e c2 r r
and j2 = l + s =

(0)

3
2

(see Figs. 6 and 7, Table 6)

n/cm-1

spin-orbit coupling
(4)

(38)

16973

1z
2

P3/2

17cm

16956

P1/2

(6)
1

(2)
off

S 1/ 2

-z

on

l = 589.8nm

Fig. 6: Splitting of the p1 levels

by spin-orbit interaction (: oneelectron spin-orbit coupling constant)

589.2nm

Fig. 7:
Term scheme of the
sodium atom

Unperturbed states sixfold degenerate

(0) (0) = E (0) (0)
H
i
i

(i = 1, 2, . . . , 6).

Eq. (35) reads in this case:

(0) E (0) ) (1) = (E (1) H
(1) ) u1 (0) + . . . + u6 (0) ).
(H
1
6

(39)

(0)

Premultiplication with 1 and integration result in:

Z
Z

(1)




(0) (0)
(0)
(0)
(1) u1 (0) + . . . + u6 (0) d
1 H E d = 1 E (1) H
1
6
|
{z
}
0
Z
Z
(0) (0)
(0) (0)
(1)
(1)
1 6 d
1 1 d + . . . + u6 E
0 = u1 E
Z
Z
(0) (1) (0)
(0) (1) (0)
u1 1 H 1 d . . . u6 1 H
6 d.

17

(40)

Using the abbreviation

(0)

(1) (0) d Hij we obtain a system of six equations:

H
j

0 = u1 H11 E (1) + u2 H12 + + u6 H16


0 = u1 H21 + u2 H22 E (1) + + u6 H26
..
..
.
.


0 = u1 H61 + u2 H62 + + u6 H66 E (1)

(41)

Non-trivial solutions for the coefficients of u1 , u2, . . . , u6 :

Calculation of the integrals Hij
Z


SO n,l,m (r, , ) m () r 2 dr sin d d d.
n,l,ml (r, , ) ms () H
s
l
Z

Rn,l (r) (r) Rn,l(r) r 2 dr

0
Z Z 2 Z 1/2




m
s Yl l (, ) ms () sin d d d.
Ylml (, ) ms () l
0

(42)

(43)

1/2

hc n,l = h

Rn,l (r) (r) Rn,l(r) r 2 dr.

(44)

n,l : one-electron spin-orbit

constant
 coupling


basis in Dirac notation: ml ms
 1  1 
 
 

1
0
 1 1
1 1
0 1
2
2
2
2
2
Integral eq. (43) has the short form



 1 1 .
2

  
hcn,l


m
m
l
s ml ms .
l
s
h
2

(45)

(46)

To determine the 36 matrix elements of the spin-orbit coupling operator

l
s = lx sx +
ly sy + lz sz , replace the x and y components by the step operators (see eq. (19,21)):

l
s = lz sz + 21 (l+ + l ) 12 (
s+ + s ) + 2i1 (l+ l ) 2i1 (
s+ s )
1
= lz sz + 4 (l+ s+ + l s+ + l+ s + l s l+ s+ + l s+ + l+ s l s )
= lz sz + 1 (l+ s + l s+ )
2

The general matrix element (46) is



  


l
s ml ms = ml ms lz sz + 21 (l+ s + l s+ ) ml ms
ml ms 




= ml ms lz sz  ml ms




+ 21 ml ms l+ s  ml ms




+ 12 ml ms l s+  ml ms .

 (1)  
n :
General hints to the evaluation of matrix elements mH
18

(47)

(48)

 
(1) n . This will result in a constant a multiplied by a wavefunction which
(i) Evaluate H
 
 
(1) 

may or may not be the same as the original. For
 present
 us

the

let
 assume H n = a n .

(ii) The result of (i) is then premultiplied
man .

  by
m  giving



(iii) Since a
is a constant
we have
  = a m n and we are
thus left
  m an
 with the task of





evaluating m n . Provided m and n are orthonormalised, m n = 1 when m = n
but is zero otherwise.
On account of orthonormalised states



(49)
ml ms  ml ms = ml ,ml ms ,ms ,

the integral is not zero when ml = ml and ms = ms . The wavefunctions are eigenfunctions
of lz und sz , so that the application of the operator products in eq. (48) on the wavefunction
on its right-hand side yields:




lz sz  ml ms = ml ms h
2  ml ms


l+ s  ml ms =
p
p


l(l + 1) ml (ml + 1) s(s + 1) ms (ms 1) h
2  ml + 1 ms 1


l s+  ml ms =
p
p


l(l + 1) ml (ml 1) s(s + 1) ms (ms + 1) h
2  ml 1 ms + 1

where s = 12 . For diagonal elements only lz sz may contribute, whereas for off-diagonal
elements only the step operators may account:







ml ms l s+  ml + 1 ms 1
ml ms l+ s  ml 1 ms + 1
Matrix elements (46) which may contribute are restricted to the condition
ml + ms = ml + ms

(50)

The non-zero matrix elements are:







1 12 l+ s  0 21
0 12 l+ s  1 12





1
1 21 l s+  0 12 .
0 2 l s+  1 12
ml ms

1
1 2


1 12 

1
0 2


0 12 


1 21 


1 12 

 1 
 
1 1
0
1
2
2
1
2

21

1
2

1
2

1
2

 
 
 

0 1
 1 1
 1 1
2
2
2

0
0
q

1
2

1
2

21
1
2

19

(51)

H11 =
H22 =
H33 =
H44 =
H55 =
H66 =
H23 =
H45 =



1   1 
z 1 2 = 1 12 1 21  1 21 = 21
2 1 2 lz s
h

1 
 1 1 = 1 ( 1 ) = 1
l
s

z
z
2
2
2
2
h
2




0 12 lz sz  0 12 = 0 12 = 0
h
2




1 
z  0 12 = 0 ( 12 ) = 0
2 0 2 lz s
h

1 
 1 1 = (1) 1 = 1
1
l
s

z
z
2
2
2
2
h
2



1 12 lz sz  1 12 = (1) ( 12 ) = 12
h
2


 1 1
 q1

1 1
1
1 

1

l
s

0
=

= 2 = H32
2

1
1

2 2 +
2
2
2
2
h
2
q


 1

1
1 1

1 2 = 2 2 1 = 12 = H54
2 0 2 2 l+ s
h

Diagonalisation of the 2 2 blocks of the H matrix:



q
 1

1

 2 E (1)

2

 = ( 1 E (1) )(E (1) ) 1 2 = 0
q

2
2
1
(1) 


E 
2
(1)

E(1) =

1
2

(1)

E(2) = .
(1)

Evaluation of the zeroth-order functions for E(1) = 21 :

q


1
1
0 = 2 2 u2(1) + 12 u3(1)
q
q
u2(1) q 1
x(1) =
= 2;
u2(1) = 13 ;
u3(1) = 23
u3(1)
q 
 q2  1 
1
1 
1 2 + 3 0 2 .
2 =
3

(52)

(1)

For E(2) = , the result is:

q


0 = 21 + u2(2) + 12 u3(2)
q

u2(2)
= 2;
u2(2) = 23 ;
x(2) =
u3(2)
q 
 q1  1 
1
2 
3 = 3 1 2 + 3  0 2 .

u3(2) =

1
3

Evaluating the second 2 2 block the resulting states are

q 
 q2 

(1)
1
1 
1
E(1) = 21 :
4 =
+ 3  0 12
3
2
q 
 q1 

(1)
1
2 
0 1 .
E(2) = :
5 =
1

3
2
3
2
The results are given in Table 6.

20

(53)

(54)
(55)

The functions are not only eigenfunctions of the operators

l
s and s
l but also of

l2 +
l
s + s
l + s2 =
l + s 2 = 2 .
(56)
If 2 acts on a quartet state function Q (1 , 2 , 4 , 6 ) and on a doublet state function
D (3 , 5 ), respectively, the result is


2 Q =
l2 + 2
l
s + s2 Q


2 ( 15
) Q = h
2 ( 32 )( 52 ) Q
= h
2 2 + 2 12 + 34 Q = h
4
2 D

= h
2 j2 (j2 + 1) Q
where j2 = 23


=
l2 + 2
l
s + s2 D


= h
2 2 2 1 + 34 D = h
2 ( 34 ) D = h
2 ( 21 )( 32 ) D
= h
2 j1 (j1 + 1) D

(57)

where j1 = 21 .

(58)

Table 6: Functions and energies of the spin-orbit coupled p1 system



 ml ms


1  1 12
q 
 q2  1 
1
1

2
3 1 2 +
3 0 2
q 
 q2 

1
1
0 1
4
+
1
3
2
3
2


6  1 21
q 
 q1  1 
2
1
3 3 1 2 + 3  0 2
q 
 q1 

2
1
0 1

5
1
3
2
3
2



 j mj

mj = ml + ms j E (1)

3 3

2 2
3 1

2 2
3 1

2
2
3 3

2
2
1 1

2 2
1 1

2
2

3
2
1
2

21

3
2

1
2

1
2

23
1
2

21

Example 2.3 Spin-orbit coupling of the f 1 and the f 13 system

Energy eigenvalues and eigenfunctions of 4f 1 (Ce3+ ) and 4f 13 (Yb3+ )
On account of j = l s = 3 12 (see Example 2.2) for one-electron and one-hole f systems
we have
energies
N

Ln3+ [4f ]

ground multiplet

Ce3+ [4f 1 ]

Yb3+ [4f 13 ]

F5/2 (j = 5/2)
F7/2 (j = 7/2)

E
2Ce

32 Yb

excited multiplet

+ 23 Ce

+2Yb

F7/2 (j = 7/2)
F5/2 (j = 5/2)

Eigenfunctions are obtained with the help of vector coupling coefficients (Clebsch-Gordan
coefficients) (see Table 7).
21

Table 7: Vector coupling coefficients for

systems with j2 = 1/2
j=

m2 =

j1 +

1
2

j1

1
2

1
2

j1 + m + 21
2j + 1
s 1
j1 m + 12

2j1 + 1

m2 = 21
s
j1 m +
2j1 + 1
s
j1 + m +
2j1 + 1

1
2

1
2

Table 8: Spin-orbit coupled eigenfunctions of Ce3+ and Yb3+ free ions



 j mj

a)



1 ; 6  52 52


2 ; 5  52 32


3 ; 4  52 12


1 ; 8  72 72


2 ; 7  72 52


3 ; 6  72 32


4 ; 5  72 12



(Ce3+ )
(Yb3+ )
 ml ms
EJ
EJ
q 
 q1 

1
6 
7 3 2 7  2 12
q 
 q2 

1
5 
2
2
7 2 2 7  1 12
q 
q 


47  1 12 37 0 12


 3 1
q  2
 q6 

1
1 
3

+ 7  2 12
7
2
3
q 

32
 q5 

2
1
1
2 

2 2 + 7 1 2
q7 
 q4 

3 
1
1

+ 7 0 12
7
2

=
=
=
=
=
=
=

a)

Short form of the functions: the first symbol refers to the upper sign, the second
to the lower one.



The calculation of the coefficients is demonstrated for  52 25 (first line in Table 8).
Assignments: j = 5/2, m = mj = 5/2, j1 = l = 3, and m2 = ms = 1/2 (j2 = s = 1/2)
The roots of the lower row of Table 7 become
s
s
q
1
j1 + m + 2
3 + 52 21
m2 = 12 :
=
= 67
2j1 + 1
7
s
s
q
j1 m + 12
3 52 + 21
1
m2 = 2 :
=
= 17
2j1 + 1
7
Since the Condon-Shortley standard assignment is j1 s and j2 l the sign of the coefficients has to be changed according to the phase relation |jb ja jmi = (1)ja +jb j |ja jb jmi.
Finally, weq
obtain
5 5

 q 


= 6 3 1 + 1 2 + 1 .
2 2

22

Problems
4. Calculate the matrix elements hl, ml |lq |l, ml i (where q stands for z, +, ):
2
(a) h0, 0|lz |0, 0i, (b) h2, 2|l+ |2, 1i, (c) h2, 2|l+
|2, 0i, (d) h2, 0|l+l |2, 0i.
5. The 14 microstates |ml ms i of a f 1 system (l = 3, s = 12 ) yield under the influence
of spin-orbit coupling 14 eigenstates |jmj i which, apart from the states | 27 72 i,
are linear combinations of two microstates each. Use Table 7 to evaluate the vector
coupling coefficients for the coupled states | 52 12 i, | 25 12 i, | 27 32 i, and | 27 32 i. Control
your results with the entries of Table 8.
6. What levels (multiplets J) may arise from the terms (a) 1 S, (b) 2 P , (c) 3 P , (d) 3 D,
(e) 4 D? How many states (distinguished by the quantum number MJ ) belong to
each level?
2.4.2

Magnetic susceptibility, Van Vleck equation

Fundamental magnetisation equation

P
P

n exp(En /kB T )
(En /B) exp(En /kB T )
n
n
P
= NA P
Mm = NA
exp(En /kB T )
exp(En /kB T )
n

(59)

Van Vleck equation

=H
(0) + B H
(1)
operator: H
z
(1)

En = Wn(0) + Bz Wn(1) + Bz2 Wn(2) + . . . .

(2)

Wn , Wn : First- and second-order Zeeman coefficient

n = En /B = Wn(1) 2BWn(2) . . .
P
(1)
(2)
(0)
[(Wn )2 /kB T 2Wn ] exp(Wn /kB T )
m = 0 NA n
P
(0)
exp(Wn /kB T )

(60)

Eq. (60) is valid for applied magnetic fields B 0.

Zeeman operator:




one-electron system Hmag = e l + 2

s B where e = e/(2me )


mag,z = lz + 2
H
s
z Bz
e



single term Hmag = e L + 2S B



mag,z = L
z + 2Sz Bz
H
e
mag = g J B
single multiplet H
e J
mag,z = g Jz Bz
H
e J

23

(61)
(62)
(63)
(64)
(65)
(66)

Example 2.4 Magnetic susceptibility of the 4f 1 /4f 13 free ion systems

The 14 spin-coupled eigenfunctions of the 4f 1 /4f13 systems are given in Table 8. Applying Zeeman operator (63) the assigned matrix elements in scheme (67) are to be evaluated:
1 2 3 4 5 6 1 2 3 4 5 6 7 8
1

 (1)  
z  1
Calculation of the off-diagonal element 2 H

5 5

 (1)  


 1 = 7 5  e (lz + 2
s
)
2 H
z
z
2 2
2 2
q
q

Dq

 q1
E
 6 
1
1
1
1 
1
6

l
+
2
s
+
3

2
+
= e
3

2
+

z
7
2
7
2  z
7
2
7
2



e h

sz 3 12 3 12 lz + 2
sz 2 + 12
6 3 12 lz + 2
7
{z
}
{z
}
|
|
0
(3 1)
h

i
6
1 
1
1
1
+ 6 2 + 2 lz + 2
B
sz 3 2 6 2 + 2 lz + 2
sz  2 + 2 =
7
|
|
{z
}
{z
}
0
(2 + 1)
h

(2)

We obtain with eq. (32) for W| 5

5
i
2 2

(2)
W| 5 5 i
2 2

2
 75
h
| e (lz + 2
sz )| 25 52 i
2 2
(0)

(0)

W7/2 W5/2

(6/49) 2B
12 2B
=
.
(7/2)
343

24

(67)



(0)
 J MJ
Wn
5 5

2
2
5 3

2
2
2
5 1

2
2
7 7

2
2
7 5

2
2
3
7 3 2

2
2
7 1

2
2

(0)
Tab. 9: Energies Wn and Zeeman co(1)
(2)
efficients Wn , Wn of a free f 1 system
(gJ1 = 6/7 for 2F5/2 , gJ2 = 8/7 for 2F7/2 )

(0)

(1)

(1)

Wn /B

(2)

Wn /2B

(5/2) gJ1 [12/(343 )]

(3/2) gJ1 [20/(343 )]
(1/2) gJ1 [24/(343 )]
(7/2) gJ2
(5/2) gJ2 +[12/(343 )]
(3/2) gJ2 +[20/(343 )]
(1/2) gJ2 +[24/(343 )]

(2)

In Table 9 the quantities Wn , Wn and Wn are collected. The final susceptibility

equation for the f 1 /f 13 system is obtained by summing up the contributions of the 14
levels:
NA 2B 2
m = 0

where 2eff =
3kB T eff





 
144 12kB T
7
45 16kB T
+4
exp
+

3
7
49
7
49
2kB T



.
(68)
7
3 + 4 exp
2kB T
For Ce3+ the parameter in eq. (68) refers to the single-electron spin-orbit coupling constant while in the case of Yb3+ the term spin-orbit coupling constant LS = has to be
replaced for . (Notice that LS = /(2S).)
The absolute energy separation between 2F7/2 and 2F5/2 is 27 Yb = 72 2870 cm1
10 000 cm1 for the Yb3+ ion and 27 Ce = 27 625 cm1 2 200 cm1 for the Ce3+ -Ion.
In magnetochemical practice the higher lying multiplet is often ignored in eq. (68). There
are two possibilities for simplification:
1.) (4f 1 ) and LS (4f 13 ) respectively:
m (4f 1 ) = 0

2eff (4f 1 )

NA 2B 45
3kB T |{z}
7
2eff

45
=
=
7

m (4f 13 ) = 0

 2    
5
7
6
7
2
2

NA 2B 144
3kB T |{z}
7
2eff

2eff (4f 13 )

144
=
=
7

(69)

 2    
7
9
8
.
7
2
2

In this approximation 2eff is temperature independent.

2.) In a less drastic approximation the contribution of the excited multiplet is again
neglected, but not W (2) of the ground multiplet, so that in the case of the 4f 1 system the
susceptibility equation reads


NA 2B 45 16kB T
NA gJ2 J(J + 1)2B
1
m (4f ) = 0
,
in general m = 0
+
+ 0 (70)
3kB T 7
49
3kB T
0 is positive, leading to a weak increase of eff with increasing temperature.
25

Lanthanide ions in cubic and non-cubic ligand fields

The evaluation of LF effects on f systems requires more effort than on d systems:

On account of the larger angular momentum quantum number l, the number of
ligand field parameters is larger, for example two instead of one for cubic symmetry;
The number of microstates used in complete magnetochemical analyses is larger for
f systems (for example 91 microstates for f 2 compared to 45 for d2 ).
Ground multiplets of the f 1 to f 13 ions:

f N f 14N

ground
multiplet

3.1


f 1 f 13

2
F 5 2F 7
2


f 4 f 10

5
I4 5I8


f 3 f 11

4
I 9 4I 15


f 2 f 12

3
H4 3H6


f5 f9

6
H 5 6H 15
2


f6 f8

7
F0 7F6

f7

(71)

S7

Ligand-field (LF) operators [7]

General form of the LF operator for one-electron systems:

=
H
LF

X
+k
X

Aqk r k Cqk (, ),

where

(72)

k=0 q=k

(R) k
= e(1)
C (, ) dR
Rk+1 q
{z
}
|
geometrical coordination factor

Aqk

and

Cqk
|

LF operator for cubic ligand fields:

1/2
4
=
Yqk (, )
2k + 1
{z
}
Racah tensor

for f 1 system

cub
H
LF

z
}|
{
h
h
p
p

i

i
6
0 6
4
0 4
4
4
6
6
= A4 r C0 + 5/14 C4 + C4 +A6 r C0 7/2 C4 + C4 , where
{z
}
|

(73)

for d1 system

C04 =

4
C4
=
C06 =
6
C4

1
(35 cos4 30 cos2 + 3)
8

1
70 sin4 ei4
16
1
(231 cos6 315 cos4 + 105 cos2
16

3
14 sin4 (11 cos2 1) ei4
32

5)

General LF matrix element:

Z

LF n,l,m (r, , )d
n,l,m
(r, , ) H
l
l

(74)

Specific LF matrix element with k = 4 and q = 0:

LF parameter B04
z Z
}|
{ Z Z

p
4
Ymll (, ) [C04 + 5/14(C44 + C4
A04
Rn,l (r) r 4 Rn,l (r) r 2 dr
)] Yml (, ) sin d d
l
{z
} 0 0
|0
radial integral < r 4 >
26

h
h
i
p
p

i
4
cub
4
6
4
4
6
6
6

HLF = B0 C0 + 5/14 (C4 + C4 ) + B0 C0 7/2 C4 + C4

(75)

With the help of integrals of the type hYml l |Cqk |Yml i [3], listed in Table 10 (p electrons)
l
and Table 13 (f electrons), LF effects can be evaluated. The k values in eq. (72) comprise
terms with k 2l and k even. The term with k = 0 has spherical symmetry, produces
no splitting and is therefore omitted. The relevant k values for LF terms are k = 2 (p
electrons), k = 2 and 4 (d electrons), k = 2, 4, and 6 (f electrons). Notice that the
operators Cqk act only on the orbital part of a wave function.

Introduction: p1 system

3.2
3.2.1

cyl
Cylindrical LF: HLF
+ HSO

Example 3.1 The p1 ion in a cylindrical LF

To introduce the use of LF operators and electrostatic matrix elements, the p1 system
in a cylindrical LF (Dh) is discussed4 ) . The LF operator consists of a single term with
k = 2 and q = 0:
cyl = B 2 C 2 ,
H
0 0
LF

where B02 = A02 < r 2 >

and C02 = 21 (3 cos2 1)

(76)

With the help of the matrix components of the electrostatic interactions, shown in
Table 10, the 3 3 matrix (77) is obtained. Notice that only the red numbers corresponding to c2 are relevant. Since the operator C02 does not change ml , matrix elements
< ml | C02 | ml > with ml 6= ml are automatically zero.
Table 10: Matrix components of electrostatic interactions, ck (lml , l ml ) for l = l =

1; ck (l ml , lml ) = (1)ml ml ck (lml , l ml ).

Red numbers are relevant for Examples 3.1
and 3.3, the blue number for Example 3.3
only.

ml

ml

c0

5 c2 a)

1 1 +1 1

1
0 0 + 3
0
0 +1 +2

1 1 0 6
a)

The numerical factor

5 is the denominator for
all ck values of the column.

ml
3 3 matrix of operator (76):

|1i

h 1 | 51 B02
h 1 |
h0|

| 1 i | 0 i
15 B02

(77)
2 2
B
5 0

Results: (i) All off-diagonal matrix elements are zero; (ii) LF operator eq. (76) splits
the threefold degenerate p1 system (in the absence of spin-orbit coupling) into an orbital
doublet | 1i with energy ELF = B02 /5 and a singlet | 0 i with energy ELF = 2B02 /5.
4)

Electronic p states are not split by cubic ligand fields.

27

Example 3.2 The p1 system under the action of spin-orbit coupling and a cylindrical
ligand field (cf. spin-orbit matrix (51))
+H
cyl = (r)
H
l
s + B02 C02
SO
LF

ml ms


1 21 


0 21 


1 12 

12

1
2

1
2



1 1
2

B02 /5

(78)
 1 

0
0 1
2
2

q
1
21 B02 /5

2
q
1

2B02 /5
2


1 12 

12 B02 /5

1
2

B02 /5

1/5

= 2 0 12 lz sz  0 12 + B02 0C02  0 21  21 = 2B02 /5

| {z } | {z }
{z
}
h
|
1

2/5

  
1  1 


= 2 0 12 lz sz  0 12 + B02 0C02  0
= 2B02 /5
{z
}
| {z } | 2 {z 2 }
h
|
0

H55 =

H66 =

1 21 lz sz  1 21 + B02 1C02  1 12  21 = 21 B02 /5

2
|
{z
}
{z
} | {z }
h
|
1

1/5

 



1  1  1

1 
 1 1 + B02 1C02  1
2  2 = 2 B02 /5
1

l
s

z
z
2
2
2
|
{z
}
{z
}|
{z
}
h
|
2 /2
h

H23 =

2/5

h2 /2

1 1
0
=
l
s

2 2
2
{z
}
h
2 |

2/2

H45

1/5

H44

1
2

1
2

  
1  1 


= 2 1 12 lz sz  1 12 + B02 1C02  1
= 12 B02 /5
| {z } | 2 {z 2 }
{z
}
h
|

h2 /2

H33



 1 1

 2 
1  1  1

1   1 
2
 

+
B
= B02 /5
l
s

1
1
z
z
0 1 C0 1
2
}
| {z } | 2 {z 2 } 2
h
2 | 2 {z
2 /2
h

H22

2B02 /5


1 21 

H11 =



 1 1

1/5

1
2

= H32



 q1

1
1 1

= 2 0 2 2 l+ s 1 2 = 2 = H54
{z
}
h
|

2/2

Application of scheme eq. (37) on the 2 2 matrices yields

28

H22 = /2 B02 /5,

H33 = 2B02 /5,

2
2 H23
=
=
tan 2 =
H22 H33
/2 + 3B02 /5

1
2

H23 = H32 = / 2
(79)

3 2
B /
5 0

cyl
Result: The combined effect of HSO + HLF
splits the sixfold degenerate spin orbital
states into three doublets. The composition of the wave functions depends on the ratio
B02 /. No simple closed expressions can be given for the functions.

3.2.2

or
Orthorhombic LF: HLF
+ HSO

Example 3.3 The p1 system in orthorhombic ligand fields

q


or
2 2
2
2
2
2

HLF = B0 C0 + B2 C2 + C2
where C2 = 38 sin2 exp(2i)

Matrix elements of operator eq. (80):

ml

|1i

h1|

15 B02

h 1 |

6
5

| 1 i

6
5

(80)

|0i

B22

15 B02

B22

2 2
B
5 0

h0|

Determination of the correct zeroth order functions by application of scheme eq. (37):

H11 = H22 = B02 /5,

H12 = H21 = 6 B22 /5
2 H12
tan 2 =
=
H11 H22

2 6 B22 /5
B02 /5 + B02 /5

6 B22 /5
0

+ for B22 < 0

for B22 > 0

2 = 90 = 45 ; cot 45 = 1

E1,2 = B02 /5 6 B22 /5,

1,2 = ( 2)1 (| 1 i | 1 i)
E3 = 2 B02 /5,
3 = | 0 i

Result: Neglecting spin, the orthorhombic ligand field splits the threefold degenerate
orbital states into three singlets.
Example 3.4 The p1 system under the action of spin-orbit coupling and an orthorhombic
ligand field
2
+H
or = (r)
H
l
s + B02 C02 + B22 (C22 + C2
)
SO
LF

(81)

The matrix elements of this operator are stated in scheme (82). The E (1) s in the
diagonal elements are omitted and also the multiplicative factor hc for the spin-orbit
coupling matrix elements which are usually given in the energy equivalent cm1 . The
empty off-diagonal elements are automatically zero. The diagonalisation of the 3 3

29

blocks, the energies and the composition of the wave functions must be calculated with
the aid of a computer.

ml ms

1 21 


0 12 

12


1 21 

1
2


1
2

B02 /5
0

6B22 /5



0 1
2
0

2B02 /5
q
1



 1 1
2

6B22 /5
q
1

3.3.1

12 B02 /5
12


0 21 

1
2

B02 /5

6B22 /5


1 12 

3.3



1 1

 1
0



 1 1

6B22 /5

1
2

2B02 /5
0

(82)

0
1
2

B02 /5

Cubic LF
Group theoretical tools I [9, 8]

Character tables of the point groups serve Tab. 11: Character table of the symmetry
to assign the LF states with respect to symgroups O and Td
metry and degeration. For the cubic sysO E 8 C3 3 C2 6 C4 6 C2
tems O and Td Table 11 is relevant. It is
applicable to Ln ions with integer J. The Td E 8 C3 3 C2 6 S4 6 d
degree of degeneration of the LF states is re- A1 a) 1
1
1
1
1 1 b)
sponsible for the orbital contribution of the A
1
1
1 1 1 2
2
lanthanides magnetic moment: The singlet
E
2 1
2
0
0 3
states A1 (1 ) and A2 (2 ) as well as the dou0 1
1 1 4
blet state E(3 ) are non-magnetic, while the T1 3
0 1 1
1 5
triplet states T1 (4 ) and T2 (5 ) are mag- T2 3
netic. For half-integer J all states are on
a)
Mulliken nomenclature.
principle magnetic; for cubic LF character
b)
Bethe nomenclature.

tables of the respective double groups O and

Td are relevant [8, 10] (cf. Table 12). Usually Bethes nomenclature is used for f systems,
see , e. g., the article of Lea, Leask, Wolf (LLW[16]).
3.3.2

cub
4f 1 system (Ce3+ , 2F : HSO + HLF
)

Ce3+ is a half-integer system. So, O , Td are relevant. Under the action of the LF operator
q
q
n
o
o
n
6
4
4
6
6
6
4
cub
1
4
5
7

+
B0 C0 2 [C4 + C4 ] ,
HLF (4f ) = B0 C0 + 14 [C4 + C4 ]
|
{z
}
irrelevant for restricted basis 2F5/2
the free ion multiplets split according to 2F5/2 G (8 ) + E (7 ) and 2F7/2 G (8 )
+ E (7 ) + E (6 ). If HSO is distinctly stronger than HLF it is convenient to consider
30

Tab. 12: Character table of the double groups O and Td

4 C3

4 C23

3 C2

3 C34

3 C4

6 C2

4 C23 R 4 C3 R 3 C2 R 3 C34 R 3 C4 R 6 C2 R
4 C3

4 C23

3 C2

3 S34

3 S4

6 d

Td

A1 (1 )a)

A2 (2 )

E(3 )

T1 (4 )

T2 (5 )

E (6 )b)

2 2

4 4

E (7 )

2 2

G (8 )
a)
b)

4 C23 R

4 C3 R 3 C2 R

1
1

3 S34 R

3 S4 R 6 d R

1
0
0

1 5 : integer J; 1 3 : non-magnetic.
6 8 : half-integer J; magnetic.

HLF acting only on the ground multiplet 2F5/2 [ G (8 ) + E (7 )]. This approximation
cub can be restricted to the fourth degree term (k = 4), since
has the advantage that H
LF
the corresponding LF matrix elements of the sixth degree term are zero. Using Table 8
cub in scheme (83) are obtained.
and Table 13 the matrix elements of H
LF
Normally, the spin-orbit coupled states |J MJ i are used as a basis for the multiplet.
In the following examples, however, we write them as |ML MS i basis with the advantage
cub on the orbital function |M i becomes obvious.
that the action of H
LF
L
Table 13: Matrix components of electrostatic interactions, ck (lml , l ml ) for l =

l = 3; ck (l ml , lml ) = (1)ml ml ck (lml , l ml ) [3]

ml

ml

c0

3 3 +1
3 2

3 1

2 2 +1
2 1
2

0
0

1 1 +1

15 c2 a) 33 c4 a)
5

+5

10
0

+ 15

20
+3

+3

30

+ 54

63

429
5

c6 a)

ml

ml

c0

+ 7

28

+ 84

7
+6

+ 32 105

3 + 224
+1

15

0 +1

15 c2

+ 2

33 c4

+ 15

+4

+6
0

3 3

3 1

2 1

3 2
2 2

1 1

0
+ 42

0
+ 70

0
14

24 40

429
5

c6

+ 350
+20

924

+ 462

210

+ 504

378

420

The numerical factor is the denominator for all ck values of a column. (Notice that only
the red terms are relevant in Example 3.5.)
a)

31

cub 5
LF
| 2 i of matrix (83)
Example 3.5 Evaluation of the diagonal element H11 = h 25 |H

H11 =

1
7

 cub 


 6 | 3, 1 i + | 2, 1 i
6 h 3, 12 | + h 2, 12 | H
LF
2
2

(i) Integration with respect to the spin, (ii) insert of ck (lml , l ml ) values 5) :


cub | 3 i + h 2 |H
cub | 2 i =
H11 = 17 6 h 3 |H
LF
LF
h



i
4
6
6
4
1
6
4
+
h
2
|C
|
2
i
+
B
6
h
3
|C
|
3
i
+
h
2
|C
|
2
i
=
B
6
h
3
|C
|
3
i
0
0
0
0
7
| {z0 }
|
{z
}
| {z0 }
c4 (33,33)=3/33

HLF matrix of 2F5/2

1
21

B04 .

c4 (22,22)=7/33

   3  5 3 1  1





4f 1 (cub)  52
2
2
2
2
2

5
1
5

2
21
21

3  5
3
2  21
21

5
1
5
2 
21
21

3
3
5

21
2
21

1
2

2
21

1
2
2 
21

(83)

cub | 3 i
Example 3.6 Evaluation of the off-diagonal element h 25 |H
LF
2
H12 =

1
7

H12 =

1
7

 cub 


 2 | 1, 1 i + 5 | 2, 1 i
6 h3, 12 | + h2, 21 | H
LF
2
2

cub | 1 i + 5 h 2 |H
cub | 2 i =
12 h 3 |H
LF
LF
h q 
i

4
4
1
5
4
B
12
h
3
|C
|
1
i
+
5
h
2
|C
|
2
i
0
4
4
7
14
| {z
| {z
}
}
42/33

70/33

q
i
h

=
72 B06
12h 3 |C46 | 1i + 5 h 2 |C46 | 2 i
{z
}
|

5
21

B04

The HLF matrix (83) has two identical 1 1 and two identical 2 2 blocks. The former
two correspond to the |MJ i states | 21 i and | 12 i yielding the first-order correction to the
energy E (1) = ELF = (2/21) B04. The energy of the other states is obtained by applying
eqn. (36):


 1 4


5
 B0 E
E1 = 2 B04

B04
21
21
 21

=
0
=





3
5
E2 = 4 B 4 .

B04
21
B04 E 
0
21
21

cub , applied to 2F5/2 [nf 1 ], has no influence, that is,

Note that the sixth degree term (k = 6) of H
LF
the corresponding matrix elements vanish.
5)

32

The sixfold degenerate multiplet 2 F5/2 is split by a cubic LF in a doublet (E ; 7 ) and a

quartet (G ; 8 ) (see Table 14). In octahedral surrounding with negatively charged ligands
E is the ground state (A40 , B04 > 0), whereas in tetrahedral symmetry the splitting pattern
is reversed. The splitting E(G , E ) = (6/21) B04 is in the range of a few 100 cm1 , that
is, several orders of magnitude smaller than for d ions, caused by smaller radial integrals
and larger metalligand distances for the 4f ions.
Table 14: Wavefunctions and energies of the
2
F5/2 multiplet perturbed by a cubic LF
i a)b) |MJ i
|M
q
q
|G , i 56 | 52 i 16 | 23 i
| 1 i
q 2
q
5
1

|E i
| 2 i 56 | 32 i
6

ELF
2
B4
21 0

|G , i

c)

4
21
B04

G 8 ; E 7 .
For the states in short cf. ref. [6].
c)
Phase
P for Kramers doublets:
|i = J,M CJ,M |JM i
P

|i = J,M CJ,M
(1)JM |J M i.
a)
b)

Magnetic susceptibility
Zeeman operator (eq. (66)):
mag,z = g J B
H
e J z z

where e = e/(2me ),

gJ = 6/7

Application of the z component of the Zeeman operator onto the six symmetry-adapted
linear combinations of the ground multiplet 2F5/2 (Table 14) yields the matrix elements
in Table 15 (in units of gJ B Bz and = ELF (G ) ELF (E ) = (6/21) B04).
(1)

(0)

Wn
(0)

(1)

(2)

Table 15: Wn , Wn and Wn of the 4f 1 system

cub
perturbed by HLF
( (6/21)B04)

|G, i
|G , i
i
|E M

Diagonal element:
mag,z |G i
H11 = hG |H
 5 5

= ( 6 )( 2 ) + ( 16 )( 23 ) gJ B Bz = ( 11
)gJ B Bz
6

Off-diagonal element:
mag |E i
H15 = hG |H
q q
i
hq q z

1
5 5
1
5
3
2 5
(
)

(
)
g

B
=
(
) gJ B Bz .
=
J B z
6
6 2
6
6
2
3
33

Wn
gJ B

(2)

Wn
gJ2 2B

11/6 +

20
9

1/2

5/6

20
9

(0)

(1)

(2)

Substitution of Wn , Wn , and Wn into the Van Vleck eq. (60) yields the magnetic
susceptibility of the 4f 1 system in cubic ligand fields:
NA 2B 2

where
3kB T eff





40
130
40

2 25
gJ
+
kB T +

kB T exp
12 3
12
3
kB T



=
.

1 + 2 exp
kB T

m = 0

2eff

(84)

1
Fig. 8 shows the eff T and the 1
(LF ground state
m T behaviour for = +605 cm

4
1
1

4
E (7 ), B0 = 2 119 cm , oct.) and = 605 cm (G (8 ), B0 = 2 119 cm1 , tetrah.,
cube).

a
c

2.0

500
-3
5
c-1
mol /10 mol m

2.4

meff

1.6

1.2
0.8
0.4
0.0

400
f

300
200

b
100
0

100 200 300 400

T/K

100 200 300 400

T/K

1
Fig. 8: Ce3+ in cubic LF; eff T (a,c,e) and 1
oct.,
m T diagrams (b,d,f); = 605 cm
1
(e,f), = 605 cm tetrah., cube, (c,d); straight lines (a,b) correspond to the free ion.

Discussion
1. The ligand field effect produces a distinct deviation from free ion magnetic behaviour, except Ln ions with 4f 7 configuration (Eu2+ , Gd3+ ).
2. At very low temperature the 1
m vs. T curves for systems with an odd number of
4f electrons (half-integral J) become nearly straight lines running into the origin
(Curie-like behaviour, explicable with eq. (84)), provided that cooperative magnetic
effects are absent.
3. In the case of integral J the low-temperatur magnetic behaviour is more complicated
(see Section 3.3.4).
4. Warning: The upper region of the 1
m versus T curves does not follow the Curie
7
or Curie-Weiss law (except 4f ions). To compare the magnetic behaviour of Ln
ions in a crystal with that of free ions, the eff versus T plot is suited.
34

5. Cooperative effects between magnetically active ions can be described by a so-called

molecular-field parameter MF:
1
1
m = m (LF) MF .

(85)

6. Susceptibility equations of the type (84) are not only suited for molecular and ionic
systems, but also for intermetallics with stable Ln valence. The 6s and 5d valence
electrons are delocalised to a high degree and yield normally only a small TIP.
7. Ce, Yb, Eu are candidates for mixed-valence behaviour (Ce3+/4+ , Yb2+/3+ , Eu2+/3+ ).
Paramagnetism of Ce3+ , Yb3+ , U5+ ions in an octahedral LF
Figs. 9 and 10 exhibit the typical paramagnetic behaviour of octahedrally coordinated
Ce(III), Yb(III) and U(V) compounds as 1
m T and eff T plots calculated with program
CONDON [12], using the parameter values (in cm1 ):
Ce(III): B04 = 2 119, B06 = 261, = 623, see Tab. 16 [13];
Yb(III): B04 = 1 471, B06 = 0, = 2 903 [13];
U(V): B04 = 23 100, B06 = 3 750, = 2 200 [14].
The 4f 13 ion with ground multiplet 2 F7/2 is the strongest paramagnet, the 5f 1 ion (2 F5/2 )
the weakest one. The lower eff value of the latter compared to eff of Ce(III) is the
consequence of the much stronger LF effect.

160

5+

3+

Yb

120

meff

-3
6
c-1
mol / ( 10 mol m )

200

80
Ce3+

40

Ce 3+

3+

Yb
0

100

200
T/K

300

0
0

400

Fig. 9: Typical 1
m T diagrams of compounds with Ce(III), Yb(III) and U(V) in an
octahedral ligand field.

3.3.3

100

200
T/K

300

5+

400

Fig. 10: Typical eff T diagrams of compounds with Ce(III), Yb(III) and U(V) in an
octahedral ligand field.

4f N systems

cub reads
For many-electron lanthanide systems H
LF
cub
LF
H
(4f N ) = B04

N n
X
i=1

C04 (i)

5
14

[C44 (i)

4
C4
(i)]

35

B06

N n
q
o
X
6
C06 (i) 72 [C46 (i) + C4
(i)] (86)
i=1

Coupling schemes: The 4f states are controlled by Hee , HSO (see Fig. 11) and the
ligand field 6) (HLF ), where the order of energetic effects is Hee > HSO > HLF (except for
the 4f 4 , 4f 5 , 4f 6 systems where HSO HLF )7) . To describe the electronic situation, three
coupling schemes are distinguished, depending on the relative strength of HLF and on the
desired accuracy in LF parameter determination [15].

0
1
2

15/2
5

13/2

11/2

E / 10 cm

-1

3
2

7/2
0

J=4

3+

Pr

J= 5/2

9/2

3+

Sm

J=0

Eu

3+

J=6

Tb 3+

Fig. 11: Energies of ground state multiplets of selected Ln3+ ions (kB T =
208.6 cm1 at 300 K)
Russell-Saunders coupling scheme (LS coupling scheme)
Including spin-orbit coupling in 1st order, S, L, and J are good quantum numbers
in this coupling scheme, i. e., this model requires multiplets 2S+1LJ that are (i)
well separated in energy and (ii) only slightly influenced by HLF . To estimate LF
parameters it is sufficient to study the action of HLF solely on the ground multiplet
2S+1
LJ [16] (except the 4f 4 , 4f 5 , 4f 6 systems).
Intermediate coupling scheme
The intermediate coupling scheme is applicable, if S and L are no longer good
quantum numbers on account of the competing perturbations by Hee and HSO while
J remains a good quantum number. This model is adequate, if the ligand field
splitting is not too strong, i. e., in the case of the weak-field lanthanide system 8) .
Examples: lanthanide chlorides and bromides (Cl , Br ).
J-mixing scheme
The J-mixing coupling scheme is favourable for strong-field lanthanide systems. If
J-mixing, caused by HLF , is considered, a 2S+1LJ multiplet may have contributions
from terms with other J values, that is, J is no longer a good quantum number.
Examples: lanthanide oxides and fluorides (O2 , F ).
6)

For lanthanides the term crystal field is often used instead of ligand field.
For actinides in high oxidation states there is no energetic order, that is, Hee HSO HLF .
8)
The terms weak field and strong field here have another meaning than for transition metal
compounds. For a strong-field transition metal system, HLF is a stronger interaction than Hee whereas
for lanthanides HLF is always weaker than Hee .
7)

36

3.3.4

Magnetic behaviour of the series Cs2 NaLnCl6 (overview)

On the basis of spectroscopic data of the series Cs2 NaLnCl6 9) [13] (see Tab. 16) the magnetic behaviour is calculated with program CONDON 10) , developed by Schilder and
Lueken [12]. The eff T data for the representative metal ions Pr3+ [4f 2 ], Nd3+ [4f 3 ],
Sm3+ [4f 5 ] and Eu3+ [4f 6 ] are compared with those of the free ions.

Cl

Cs

Ln
Na
Fig. 12: Cs2 NaLnCl6 : unit cell of the elpasolite type
Table 16: LF parameters B04 and B06 of octahedrally coordinated Ln3+ ions in Cs2 NaLnCl6
Ln3+
B04 a)
B06 a)
Ce3+
2119
261
Pr3+
1938
290
3+
Nd
1966
258
3+
b)
Sm
(1671)
(228) b)
Eu3+
2055
308
3+
Gd
1776
136
a)
b)

Ln3+
Tb3+
Dy3+
Ho3+
Er3+
Tm3+
Yb3+

B04
1624
1614
1593
1492
1498
1471

B06
150
148
171
163
159
[0]

Values in cm1 .
Values for Cs2 NaYCl6 : Sm3+ .

The chemical environment of the Ln ions, compared to the one of d ions, has only a
minor effect on the 4f electrons. The LF effect produces splittings of HLF 102 cm1
9)

Elpasolite type structure, Ln3+ point symmetry Oh .

CONDON considers ligand field effects on the J-mixing level and, beyond this model, the applied
field dependence of the magnetic susceptibility as well as intermolecular spin-spin exchange couplings.
10)

37

leading generally to paramagnetism, described by the temperature dependence of the

effective Bohr magneton number eff :

1/2
3kB T m
NA 2eff 2B
where eff =
= 797.7(T m)1/2
(87)
m = 0
3kB T
0 NA 2B
As a rule, eff approaches the free Ln ion value for T > 200 K (see Table 2, Figs. 8, 10,
13, 14).
Pr3+ : Applying B04 = 1938 cm1 and B06 = 290 cm1 the 1
m T plot in Fig. 13 (left)
is obtained. The octahedral LF effect produces temperature independent paramagnetism
below 100 K in agreement with the predicted non-magnetic ground state A1 (see Table 17).
Nd3+ : The magnetic behaviour of Cs2 NaNdCl6 exhibits Curie paramagnetism in the
temperature range under investigation (see Fig. 13, right). This behaviour is expected
for Ln ions with an odd number of 4f electrons in magnetically diluted systems without
cooperative interactions.
Sm3+ : On account of low lying multiplets, eff of the free Sm3+ ion is already temperature dependent (see Fig. 14 (left), curve b, and Fig. 11). After switching on the cubic
ligand field (curve a), a drastic modification is observed at T < 200 K.

4.0

240

200
3.0
2.0

120
80

150
c-1mol /105mol m-3

160

meff

5
-3
c -1
mol /10 mol m

200

1.0

40
0

0.0
100 200 300 400
T/K

100
50

0
0

3.6
3.2
2.8
2.4
2.0
1.6
1.2
0.8
0.4
0.0
100 200 300 400

meff

Eu3+ : In the complete temperature range the magnetic properties of Eu3+ compounds
are determined by the TIP on account of the non-magnetic ground multiplet with J = 0
(see Fig. 14, right). The effect of a cubic ligand field is negligible.

T/K

3+
Fig. 13: 1
(left) and Nd3+ (right), calculated with the
m T and eff T diagrams for Pr
spectroscopically determined data of Cs2 NaPrCl6 and Cs2 NaNdCl6 , respectively (solid
lines: full basis; dottet lines: ground multiplet only; dashed lines: free ions)

38

4.0

8
d

meff

2.0
b
a

1.0

5
4

300

200
T/K

100

3
2

c
d

0.0
0

6
-7
-3
c-1
mol /10 mol m

3.0

400

200
T/K

100

300

400

3+
Fig. 14: eff T (left) and 1
and Eu3+ , calculated with
m T diagrams (right) for Sm
3+
the spectroscopically determined data of Cs2 NaYCl6 :Sm (a) and Cs2 NaEuCl6 (d), respectively; free Sm3+ ion (b), free Eu3+ ion (c).
cub
4f 2 system (Pr3+, 3H4 : HLF
)

3.3.5

cub reads:
For a 4f 2 system H
LF
cub (4f 2 ) = B 4
H
LF
0

2 n
X

C04 (i) +

i=1

(88)
2 n
q
o
o
X
4
4
6
6
6
6
5
7
[C
(i)
+
C
(i)]
+
B
C
(i)

[C
(i)
+
C
(i)]
4
4
0
0
4
4
14
2
i=1

Compared to the two multiplets of the 4f 1 (Ce3+ ) free ion, the 4f 2 (Pr3+ ) free ion has
13 multiplets and, in consequence, exhibits a more complicated LF splitting pattern (see
Table 17 and ref. [13]).
Table 17: Spectroscopically determined levels of the Pr3+ ion (Oh , octahedron)
in Cs2 NaPrCl6 . The energy increases from top to bottom and from left to right.
3

H5

A1 [1 ](0 a) ) T1

(1) b)

T1 [4 ](236)

T2

T2

E[3 ](422)

A1

T2 [5 ](701)

T1

H4

(2)

F2

E c)

T1

(1)

T2

H6

G4

D2

P0

P1

P2

A1

T2

A1

T1

A1

T2

A1

T2

T1

T1

A2

A1

T1

T2

T2

T2

A2

F3

A2 [2 ]

F4

T2

(2)

T2

Experimental energy data in cm1 ;

(1)

(2)

T1

a)

I6

E
b)

2300 cm1;

39

c)

4392 cm1 [13].

S0

According to the procedure outlined in 3.3.2, only the ground multiplet 3H4 will be
considered for the 4f 2 ion, that is, one starts with the |JMJ i basis consisting of the
2J + 1 = 9 states |4 4i, |4 3i, . . . , |4 4i. The 4f 2 situation is much more complicated
than the 4f 1 one for the following reasons: In order to apply the entries of Table 13,
the |JMJ i functions have to be decomposed into the microstates of the two electrons
(cf. ref. [7], page 120). Subsequently, the operator eq. (88) must be applied. Then the
respective LF matrix elements have to be summed up and collected in a 9 9 matrix.
After diagonalisation of this matrix, the symmetry adapted functions and their energies
are obtained as a function of the LF parameters B04 and B06 .
These calculations would be time-consuming and must be worked out by a computer,
using, for example, the program CONDON [12] which is at our disposal. In general, this
program uses the complete set of basis functions, but can be restricted to the ground
multiplet11) . The results, obtained by CONDON with the restricted basis, are presented
in Table 18. A splitting in a singlet (A1 ), a doublet (E) and two triplets (T1 , T2 ) is
obtained and the entries of Table 17 are confirmed.
Tab. 18: Functions and energies after perturbation of the 3H4 (4f 2 )
ground multiplet by a cubic ligand field

functions |MJ i (J = 4)
q

1
14
|
0
i
+
5
|
4
i
+
5
|
4
i
24
q

1

10
|
0
i
+
7
|
4
i
+
7
|
4
i
24
q
1
(| 2 i + | 2 i)
2
q
q
18 | 3 i 78 | 1 i
q
1
(| 4 i | 4 i)
2
q
q
1
7
|
3
i
+
| 1 i
8
8
q
q
7
| 3 i 18 | 1 i
8
q
1
(| 2 i | 2 i)
2
q
q
78 | 3 i + 18 | 1 i
a)
b)

A1 1 , E 3 , T1 4 , T2 5 [6].
b4 = (15/2)J B04 , b6 = (315/4)J B06 ;

11)

ia) )
|M

ELF

b)

|A1 a1 i

28 b4 80 b6

|Ei

4 b4 + 64 b6

|Ei
|T1 1i
|T1 0i

14 b4 + 4 b6

|T1 1i
|T2 1i
|T2 0i

26 b4 20 b6

|T2 1i
4
,
J = 5 445

J =

272
4 459 455 .

As an alternative, a rather simple method is available to determine LF effects for cubic lanthanide
systems, if the description of the LF effect within the ground multiplet is sufficient. In this method an
operator is applied that consists of so-called operator-equivalents instead of the Racah tensors. We will
come back to its practice in 3.3.7 and in Appendix 2.

40

3.3.6

cub
4f 2 system (Pr3+, 3H4 : HLF
+ Hmag )

Applying the Zeeman operator eq. (66) on the eigenfunctions, the matrix elements (89)
are obtained (in units of gJ B Bz ; gJ = 4/5).
(89)

A1

A1

a1

a1

1
T1

0
1

T1

20
3

28
3

1
2
28
3

7
2

0
21

20

T2

q3

0
q

0
q

T2

7
2

7
2
5
2

7
2

(0)

52
(1)

(2)

In Table 19 the energies Wn and the Zeeman coefficients Wn , Wn are listed.

(0)

(1)

(2)

Tab. 19: Energies Wn and Zeeman coefficients Wn , Wn of

the 4f 2 system (Pr3+ ) after perturbation of the 3H4 multiplet by
a cubic ligand field.
(1)

(0)
Wn a)

|A1 a1 i

28 b4 80 b6

|Ei

4 b4 + 64 b6

|Ei
|T1 1i

14 b4 + 4 b6

|T1 0i
|T2 1i
|T2 0i
a)

26 b4 20 b6

(2)

Wn b)
gJ2 2B
(20/3)
0

(T1 ; A1 )
(28/3)
0

(T1 ; E)
4
0

(T2 ; E)
(7/4)
1/2

(T2 ; T1 )
(20/3)
(28/3)
0
+
+
(T1 ; A1 ) (T1 ; E)
(7/4)
5/2
+
(T2 ; T1 )
4
0
+
(T2 ; E)

Wn
gJ B

(0)

Further contributions to Wn beside
ELF are equal in magnitude

4
for all levels and are reduced; b4 = 15
b6 = 315
J B06 .
2 J B0 ,
4
(0)
(0)
b)
(m ; n ) = Wm Wn .

41

(0)

(1)

(2)

Setting Wn , Wn and Wn into the Van Vleck eq. (60), the susceptibility equation
(0)
for the Pr3+ ion in cubic ligand fields is developed (where X Wn ()/kB T ):
NA 2B 2
m (Pr3+ ) = 0

where 2eff = gJ2

3kB T eff




56kB T
24kB T
40kB T
exp(XA1 ) +
exp(XE )
+
(T1 ; A1 )
(T1 ; E) (T2 ; E)


3
21kB T
40kB T
56kB T
+
exp(XT1 )
+

2 (T2 ; T1 ) (T1 ; A1 ) (T1 ; E)




75
21kB T
24kB T
+
exp(XT2 )

2
(T2 ; T1 ) (T2 ; E)

1
exp(XA1 ) + 2 exp(XE ) + 3 exp(XT1 ) + 3 exp(XT2 )

(90)

3+
Fig. 13 (left) exhibits calculated 1
ion in an
m T and eff T diagrams for the Pr
4
1
6
1
octahedral ligand field where B0 = 1938 cm and B0 = 290 cm . Dotted lines refer to
calculations with the reduced basis 3H4 (eq. (90)), whereas the solid lines correspond to
calculations with the complete basis (91 microstates). Differences in m values between
the two models amount to 3 %. In addition, the expected magnetic behaviour of the free
ion is shown for comparison (dashed lines).

3.3.7

Prediction of LF splittings by means of the operator-equivalent method

The magnetic behaviour of Ln ions depends on position and charge of the ligands, both
controlling sign and amount of the LF parameters Bqk . If one accepts the qualitative
validity of the point charge electrostatic model (PCEM) for the geometrical coordination
factor Aqk (defined in eq. (72)), the sign of the corresponding Bqk = Aqk < r k >, that is,
the sign of B04 and B06 in the case of cubic coordination polyhedra (octahedron, cube,
tetrahedron), can be predicted reliably12) . We consider the point charges Qe 13) at the
vertices of regular polyhedra (octahedron, tetrahedron, cube) with the distance R to the
centre of the polyhedron, and calculate A04 and A06 .
Geometrical coordination factors
A04 For the octahedron the integration in eq. (72) reduces to a summation of the six
ligands with (j , j ) coordinates (0, 0), (, 0), (/2, 0), (/2, /2), (/2, ), (/2, 3/2):
A04 (oct.)

6
6
Qe2 X 4
Qe2 X 1
7 Qe2
4
2
= 5
C0 (j , j ) = 5
,
(35
cos

30
cos

+
3)
=
j
j
R j=1
R j=1 8
2 R5

(91)

For negatively charged ligands (Q > 0) which coordinate the Ln ion octahedrally, A04 (oct.)
is positive and consequently also A04 < r 4 > = B04 , whereas the A04 s for tetrahedron and
cube are negative (angle coordinates for tetrahedron: (, /4), (, 5/4), ( , 3/4),
12)

Notice, that radial integrals < rk > are positive quantities.

The elementary charge e is positive; the charge of the electron is e. Ligands with the charge Qe
are negatively charged for Q > 0.
13)

42


( , 7/4), where cos = 1/ 3; additionally for the cube (, 3/4), (, 7/4), (
, /4), ( , 5/4)):
A04 (tetr.) =

14 Qe2
;
9 R5

A04 (cube) =

28 Qe2
9 R5

(92)

A06
A06 (oct.)

6
Qe2 X 6
C (j , j ),
=
R7 j=1 0

where

C06 (j , j )

4
13

1/2
Y06 (j , j )

6
Qe2 X 1
3 Qe2
6
4
2
=
(231
cos

315
cos

+
105
cos

5)
=
j
j
j
R7 j=1 16
4 R7

A06 (tetr.) =

8 Qe2
;
9 R7

A06 (cube) =

16 Qe2
9 R7

(93)

(94)

For negatively charged ligands (Q > 0), A06 is positive, regardless the polyhedron.
Having fixed the sign of the LF parameters B04 and B06 , the possible LF ground states
of the Ln ion under investigation can be determined. If the program CONDON is at
our disposal, we need only the values of B04 and B06 as an input to obtain energy and
composition of the LF states. Otherwise, we can benifit from tables of eigenvectors and
diagrams of energy eigenvalues, expressed as a function of the ratio between the fourth
cub , that are published for all J-manifolds of the lanthanides
and sixth degree terms of H
LF
by Lea, Leask, Wolf [16]. The immediate application of these diagrams and tables is
the prediction of spectroscopic energy level schemes and possible g-factors for lanthanide
ions in cubic coordination. We start with introductory remarks concerning the operator
equivalent technique.
Operator equivalents: To describe the fundamental magnetic behaviour of Ln ions
in a solid, a model is sufficient that, according to the Russell-Saunders coupling scheme,
LF into consideratakes solely the ground multiplet 2S+1LJ and its perturbation by H
tion (exceptions: Sm3+ [4f 5 ], Eu3+ [4f 6 ]). The restriction of the basis functions allows
|J M i by application of the soa simple calculation of the matrix elements hJ MJ |H
LF
J
called operator-equivalent technique. Using this technique, the calculation of the integrals
hli ml1 |Cqk (i)|li ml1 i entailing a troublesome decoding of the |J MJ i basis with regard to
the microstates is unnecessary.
The operator equivalent technique is founded on the fact that the operators Cqk =
p
4/(2k + 1) Yqk , written in cartesian coordinates, can be replaced by operators Jx , Jy ,
qk are
Jz of the total angular momentum [17] 14) . By this means operator equivalents O
designed whose matrix elements are proportional to the corresponding matrix elements of
LF . The restriction to the ground multiplet has the advantage that the matrix elements
H
cub |JM i are easily evaluated: Given the wave functions in the |JM i basis, the
hJMJ |H
LF
J
J
cub on the orbital part of the functions is taken into consideration by a simple
action of H
LF
cub (eq. (72)) are
factor (Stevens factor [18]). If x, y, z in the Racah tensors Cqk (x, y, z) of H
LF

replaced by the total angular momentum operators Jx , Jy , Jz [17], operator equivalents

14)

Originally, the operator equivalents are related to the spherical harmonics Yqk [18, 21]. Conversion
factors are published in ref. [22]. Details are given in the Final remarks, p. 49.

43

 k are obtained, whose matrix elements are proportional to the matrix elements of H
cub .
O
q
LF
Simple examples of operator equivalents are:
x2 y 2 (Jx2 Jy2 ),

3z 2 r 2 [3Jz2 J(J + 1)],

xy ( 21 )(Jx Jy + Jy Jx ).

The multiplicative (Stevens) factor depends on k:

J (k = 2)

J (k = 4)

J (k = 6)

(see Table 23 and example 3.8)

For cubic symmetry only terms with k = 4 and 6 are relevant, so that the multiplet under
consideration is generally characterised by the two figures J and J 15) . The two following
equations show the operator and its operator equivalent:
cub (4f N ) = B 4
H
LF
0

N n
N n
q
q
o
o
X
X
4
4
4
6
6
6
6
5
7
C0 (i) + 14 [C4 (i) + C4 (i)] + B0
C0 (i) 2 [C4 (i) + C4 (i)]
i=1

i=1

q 
q 
i
h
i
h
6
5
4 + O
4
6 7 O
6 + O
6
cub = B 4 O
4 +
O
+
B

O
H
4
4
0 J
0
4
4
LF
0 J
0
14
2

(95)

q4 as well as Cq6 and O

q6 are related (see ref. [17]) by
The operators Cq4 and O
N
X

C04 (i)

4 = J
= J O
0

1
8

i=1
N
X

4
4 = J
C4
(i) = J O
4

i=1
N
X

6 =
C06 (i) = J O
0
J

i=1

o

n
4
2
2
2

35J z [30J(J + 1) 25]J z + 3J (J + 1) 6J(J + 1)

1
8

1
16

n

70

1
2

J4

(96)

231 J6z [315 J(J + 1) 735]J4z + [105 J 2 (J + 1)2

525 J(J + 1) + 294]J2z 5 J 3 (J + 1)3 +40 J 2 (J + 1)2 60 J(J + 1)
N
q
o
X

n
6
6
63
1

2z J(J + 1) 38]J4 + J4 [. . .]
C4 (i) = J O4 = J 128
[11
J
(97)
4
i=1

k |JM i that are necessary to set up HLF matrices are availThe matrix elements hJMJ |O
q
J
able and collected in the Tables 35 38 on pages 89 90. In the following three examples
the handling of operator equivalents is shown.
Example 3.7 Application of the operator-equivalent method to 2F5/2 of Ce3+ [4f 1 ]
The HLF operator equivalent for the cubic 4f 1 system reads
q 
i
h
5
4 + O
4
cub = B 4 O
4 +
O
H
4
4
LF
0 J
0
14

(98)

Inspecting the Tables 35 and 36 the relevant data for the basis | 52 MJ i are:
4| 1 i =
h 12 |O
15
0
2
3
3 4
h 2 |O0 | 2 i = 45/2

4| 5 i =
h 52 |O
15/2
0
2

3
5 4
h 2 |O4| 2 i = 15 14/2.

15)

(99)

Exceptions are the multiplets 2F5/2 (Ce3+ [4f 1 ]) and 6H5/2 (Sm3+ [4f 5 ]) with J = 5/2, where only
is relevant.
the fourth degree term (k = 4) in H
LF

44

04 | 5 i = 15 , given in eq. (99)

Verification of h 25 |O
2
2
n














04 | 5 i = 1 35 5 4 30 5 7 25 5 2 + 3 5 2 7 2 6
h 25 |O
2
8
2
2
2
2
2
2


1
960
15
= 128 (21 875 23 750 + 3 675 840) = 128 = 2

5
2

7
2

o

Notice that, applying operator equivalents, the factor h

is always omitted.

To complete the matrix elements they must be multiplied by J =

Example 3.8 shows the evaluation of the Stevens factor J .

2
315

(see Table 23).

Example 3.8 Evaluation of J for 2F5/2 [4f 1 ]

4 | 5 i (Example 3.7):
J is derived by comparing h 52 |C04 | 52 i (Example 3.5) with h 25 |O
0 2
2
2
4| 5 i J =
=
h 52 |C04 | 25 i = J h 25 |O
0 2
2
3 57
315
| {z }
| {z }
1/21

15/2

2
On the basis of the operator-equivalent matrix elements (99) and J = 315
(Example 3.8)
cub
2

the HLF matrix within the basis F5/2 can be set up. It is identical to matrix (83).

Example 3.9 Application of the operator-equivalent method to Pr3+ [4f 2 ], 3H4 , J = 4

For the 4f 2 system both the fourth and the sixth degree terms of the LF operator equivalents eq. (95) have to be applied. According to the 1st column in Table 17, the 9-fold
degenerate ground multiplet 3H4 decomposes under the action of a cubic LF into the
terms A1 , E, T1 , and T2 . In order to verify the splitting by use of the operator-equivalent
k
0k |MJ i and hMJ |O
4
method, we need the matrix elements hMJ |O
|MJ 4i for k = 4 and
k = 6 (Table 20 and Tables 35 38 in Appendix 2) as well as J = (4/5 445) and
J = (272/4 459 455), given in Table 23. The 9 9 matrix is divided into one 3 3 block
and three 2 2 blocks:

14 b4 + 4 b6


0
70 (b4 6 b6 )

where b4 =

0


15

J B04 ,


 3
2

70 (b4 6 b6 )

18 b4 20 b6



 4
0

70 (b4 6 b6 )
and b6 =


 1

70 (b4 6 b6 )
14 b4 + 4 b6

315
4

J B06
 
2

(100)


 2

3  21 b4 17 b6
7 (5 b4 + 3 b6 )


1 
7 (5 b4 + 3 b6 )
9 b4 + b6


2
11 b4 + 22 b6 15 b4 + 42 b6


2 
15 b4 + 42 b6 11 b4 + 22 b6

cub matrix yields the energies and eigenfunctions given in Table 21.
Evaluation of the H
LF
The splitting in a singlet (A1 ), a doublet (E) and two triplets (T1 , T2 ) is confirmed.
45

qk in the basis
Tab. 20: Matrix elements of the operator equivalents O
2
|JMJ i (J = 4) for 4f and cubic symmetry (see Appendix 2)

qk |M i
hMJ |O
J
0k | 0 i
h 0 |O

k=6

qk |M i
hMJ |O
J

135 1575

0k | 4 i
h4 |O

k=4

k| 1 i
h1 |O
0

135
2

315
4

k| 2 i
h2 |O
0

165
2

3465
2

0k | 3 i
h3 |O

315
2

5355
4

k=4

k=6

105
315

k | 2 i 45 70 945 14
h2 |O
4
2
2

k | 1 i 105 10 945 2
h3 |O
4
2
4

k
4
h4 |O
|0i
105
945 5

Tab. 21: Functions and energies after perturbation of the 3H4 (4f 2 )
multiplet by a cubic ligand field

functions |MJ i (J = 4)
q

1
14
|
0
i
+
5
|
4
i
+
5
|
4
i
q 24

1
10 | 0 i + 7 | 4 i + 7 | 4 i
24
q
1
(| 2 i + | 2 i)
2
q
q
1
8 | 3 i 78 | 1 i
q
1
(| 4 i | 4 i)
q2
q
1
7
|
3
i
+
| 1 i
q8
q8
7
| 3 i 18 | 1 i
q8
1
(| 2 i | 2 i)
2
q
q
7
8 | 3 i + 18 | 1 i
a)
b)

ia) )
|M

ELF

b)

|A1 a1 i

28 b4 80 b6

|Ei

4 b4 + 64 b6

|Ei
|T1 1i
|T1 0i

14 b4 + 4 b6

|T1 1i
|T2 1i
|T2 0i

26 b4 20 b6

|T2 1i

Assignment of the states see ref. [6]; A1 1 , E 3 , T1 4 , T2 5 .

b4 = (15/2)J B04 ; b6 = (315/4)J B06 .

Level tuning [16]: The ground state of an Ln ion in a cubic LF depends on J B04 and
J B06 . We want to obtain information about the succession of ELF of the LF states as a
function of the ratio J B04 /J B06 . Using the operator equivalent matrix elements in Tables
35 38, the (2J + 1) (2J + 1) matrix can in principle be drawn up. The tables contain
factors common to all the matrix elements, F (4) and F (6). These factors are separated
out in order to keep the energy values in the same numerical range for all ratios of the
fourth and sixth degree terms. We start out from eq. (95)
q 
q 
i
i
h
h
6 6
cub
7
4
6
6
44 + O
4

04 + 5 O
LF
O
+
O
O

+
B
H
= J B04 O
4
4
0
J
0
14
2
|
|
{z
}
{z
}
4
6
b
b
O
O
46

b 4 + B6 O
b6
= J B04 O
J 0
b6
b4
O
O
+ J B06 F (6)
.
= J B04 F (4)
F (4)
F (6)

(101)

In order to cover all possible values of the ratio between the fourth and sixth degree terms
a mixing parameter x has been introduced which is governed by the relative magnitude
of B04 and B06 . Furthermore, a factor W has been defined that scales ELF :
J B04 F (4) = W x,
| {z }

J B06 F (6) = W (1 |x|)

| {z }

b4

where

1 < x < +1

(102)

b6

Solving both eqs. (102) for W , an equation is obtained that covers all possible values of
the ratio between B04 and B06 . It follows that
J B04
x
F (6)
,
=
6
J B0
1 |x| F (4)

(103)

so that (J B04 /J B06 ) = 0 for x = 0, while (J B04 /J B06 ) = for x = 1. In consequence, eq. (101) can be rewritten as
"
!
!#
b4
b6
O
O
cub = W x
H
+ (1 |x|)
(104)
LF
F (4)
F (6)
For each J-manifold x is allowed to take the values in the range 0 1. The energy
expressions for the LF states as a function of B04 F (4) = b4 and B06 F (4) = b6 are
replaced by W x and W (1 |x|), respecticely. Table 22 exhibits the results for the 4f 2
system and Fig. 15 displays the corresponding diagram ELF /W versus x.
Tab. 22: Energies ELF of the 4f 2 system in cubic ligand fields as
a function of W and x as well as b4 and b6 , see eqs. (102)

ELF /W

ELF

x = 0 x = 1

A1 28 b4 80 b6

W [28x 80(1 |x|)]

80

28

4 b4 + 64 b6

W [4x + 64(1 |x|)]

64

T1 14 b4 + 4 b6

W [14x + 4(1 |x|)]

14

20

26

T2 26 b4 20 b6 W [26x 20(1 |x|)]

Diagrams like Figs. 15 and 16, displaying LF energy eigenvalues ELF /W as a function of
the LF parameters, are available for all J-manifolds of the lanthanides ground multiplets
[16]. In examples 3.10 and 3.11 the diagrams are used to determine the LF ground state for
Pr3+ [4f 2 ] and Nd3+ [4f 3 ] in cubic LF. In example 3.12 the construction of the x ELF /W
diagram for Pr3+ is demonstrated.
47

60

E (G3)
40

40

20

T1 (G4)

20

(2)

(2)

( G8 )

E (G6)

x
T2(G5)

-40

A1(G1)
-1

X
-20
-40

-60

Fig. 15: x(ELF /W ) plot of

4) in cubic ligand fields [16]

-1

Pr3+ [4f 2 ]

(1)

(1)

( G8 )
1

Fig. 16: x(ELF /W ) plot of Nd3+ [4f 3 ] (J =

9/2) in cubic ligand fields [16]

(J =

Example 3.10 Ground state of Pr3+ in cubic LF

Pr3+ : The signs of W and x for an octahedral LF:
(i) The PCEM predicts that for an octahedral LF both geometrical parameters, A04 and
A06 , are positive (see eqs. (91) and (93)) and, consequently, also B04 and B06 16) .
(ii) Since both F (4) and F (6) are positive (for all Js), it is obvious from the right eq. (102)
that the sign of W is determined by the sign of J B06 , since (1 |x|) is always positive
for 1 < x < +1.
(iii) From eq. (103) we see that the sign of x is determined by the sign of (J B04 /J B06 ).
For the Pr3+ system under investigation, J , B04 , and B06 are positive, while J is negative
(see Table 23). So, the sign of x is negative.
The x(ELF /W ) diagram in Fig. 15 shows that for the situation x < 0 and W > 0 the
singlet A1 is the ground state. Thus, an octahedrally coordinated Pr3+ ion should exhibit
temperature-independent paramagnetism at low temperature. Fig. 13 (left) confirms this
magnetic behaviour.
Pr3+ : The signs of W and x for a tetrahedral/cubical LF:
Since A04 (B04 ) is negative, x is positive; the sign of W doesnt change compared to the
octahedral case. Consequently, both W and x are positive and the LF ground state may
be a singlet (A1 ) or a triplet (T2 ), depending on the magnitude of x (see the region x > 0
in Fig. 15).
Example 3.11 Ground state of Nd3+ [4f 3 ] in cubic LF
Nd3+ : The signs of W and x for an octahedral LF:
(i) Applying the PCEM, the situation is identical to the Pr3+ case: B04 , B06 > 0 17) .
(ii) Since J is negative, it is obvious from the right eq. (102) that W < 0.
(iii) From eq. (103) we see that the sign of x is determined by the sign of (J B04 /J B06 ).
For the Nd3+ system under investigation both J and J are negative (see Table 23). So,
the sign of x is positive.
16)

B04 =
17)

The result of the PCEM calculation is supported by spectroscopic investigations, showing that
1 938 cm1 and B06 = 290 cm1 (see Table 16).
B04 = 1 966 cm1, B06 = 258 cm1 (see Table 16).

48

The x(ELF /W ) diagram in Fig. 16 shows that for the situation x > 0 and W < 018)
either the quartet G or the doublet E is the ground state. Thus, ignoring interionic spinspin couplings, an octahedrally coordinated Nd3+ ion should exhibit Curie paramagnetism
regardless T . Fig. 13 (right) exhibits the magnetic behaviour.
Note, that the diagrams in Figs. 15 and 16 differ distinctly: While for the Pr3+ ion
exclusively straight lines result, one straight line and two curves are observed for the
Nd3+ ion (see Remarks on page 49 for further details).
Nd3+ : The signs of W and x for a tetrahedral/cubical LF:
Since both W and x are negative (according to J , J , B04 < 0 and B06 > 0, eq. (103)) the
ground state is the quartet G .
Final remarks
According to todays standard practice, we use the operator equivalents that are
related to the Racah tensors Cqk [17] while Stevens definition relates to the spherical harmonics Yqk [18, 19]. Since the article of Lea, Leask, Wolf [16] is based on
Stevens definition Yqk , some hints to the conversion factors between both derivations
is informative with respect to the application of the LLW diagrams:
(i) The LLW diagrams can be applied directly, since the conversion factor for both
W and x is 1. In other words, the values O4 /F (4) (based on Yqk ) in equation
b4 /F (4) (based on C k ) in eq. (104) are identical and the
(5) of ref. [16] and O
q
same is true for the sixth degree terms (k = 6).

(ii) Conversion
(see eq.
(96) and (97)):
p

factors 4for operator

equivalents
4
=
6 = 63/128 O 6
4 = 1 O 4, O
6 = 1 O 6, O
O
70/8
O
;
O
0
4
0
4
0
4
0
4
8
16
Consequence of the conversion factors:
Conversion
of F (4) and F (6) [21]
into F (4) and F (6) (Table 23), respectively:
1
F (4) = 81 F (4),
F (6) = 16
F (6)

(iii) The LF parameters, used by LLW, B04 (LLW) and B06 (LLW) [16], include the
Stevens factors J and J , while in todays standard practice, used by us,
both are separated:
B04 (LLW) = J B04 ,
B06 (LLW) = J B06
Within the x(ELF /W ) diagrams there are two types of i curves to be considered:
(i) curves that are straight lines and (ii) curves that deviate more or less from
straight lines.
(i) A state i whose energy eigenvalue is a linear function of x, occurs once only in
the decomposition of the J-manifold. For example, in the cubic Pr3+ systems
with ground multiplet J = 4 each split term A1 (1 ), E(3 ), T1 (4 ), T2 (5 )
occurs once only. So, the diagram exhibits only straight lines. The state
functions have fixed compositions (see Table 21) independent of B04 and B06 .
Other examples are ions in cubic LFs with J = 52 (E (7 ), G (8 )) and 27 (E
(6 ), E (7 ), G (8 )).

18)

According to the negative sign of W , one has to mirror the x(ELF /W ) diagram at the x axis with
the consequence that the ELF /W axis shows downwards for W < 0.

49

Tab. 23: Ground states of Ln3+ ions in cubic ligand fields [16]

Ln3+ [4f N ]
2S+1

LJ

Ce3+ [4f 1 ]
2

F5/2

Pr3+ [4f 2 ]

Nd3+ [4f 3 ]

H4

I9/2

Sm3+ [4f 5 ]
6

H5/2

Tb3+ [4f 8 ]
7

F6

2
35

52
2 475

7
1 089

13
315

1
99

2
315

4
5 445

136
467 181

26
10 395

2
16 335

272
4 459 455

1 615
42 513 471

1
891 891

F (4)a)

15
2

15
2

21
2

15
2

15
2

F (6)

315 b)
4

315

945
2

[O]c)

+ d)
E [ +1
]

A1 [ +
]

G , E [
]
+

+
E [ +1
]

A2 , A1 [
]

[T ]

G [ +1
]

+
A1 , T2 [ +
]

G [
]

G [ +1
]

A2 , E [
]
+

Ln3+ [4f N ]
2S+1

LJ

Dy3+ [4f 9 ] Ho3+ [4f 10 ] Er3+ [4f 11 ] Tm3+ [4f 12 ] Yb3+ [4f 13 ]
6

H15/2

I8

I15/2

H6

F7/2

2
315

1
450

4
1 575

1
99

2
63

8
135 135

1
30 030

2
45 045

8
49 005

2
1 155

4
3 864 861

5
3 864 861

8
3 864 861

5
891 891

4
27 027

F (4)

15
2

105
2

15
2

15
2

15
2

F (6)

3 465
4

3 465
4

3 465
4

945
2

315 b)
4

[O]

]
E , E [ +

]
E, A1 [ +

]
E , G [ +
+

]
A2 , A1 [

+
]
E [

[T ]

E , G [ +
]
+

E, T2 [
]

E , E [ +
]

A2 , E [
]
+

E , E [ +
]
+

F (4) and F (6) are equivalent to F (4) and F (6) [16] and consider the different
qk and Oqk .
definitions of the operator equivalents O
b)
Note that there is a misprint in the corresponding Table 4.39 of ref. [7].
c)
Indices O and T identify the expected LF ground state for octahedral and
tetrahedral (cubal) symmetry and negatively charged ligands (A1 1 ; A2 2 ; E
3 ; T1 4 ; T2 5 ; E 6 ; E 7 ; G 8 ).
 
d)
Combination of signs for the W and x parameter W
x .
a)

50

(ii) Systems with J > 4 decompose as well into states which occur more than once
and deviate from straight lines. Taking J = 29 (Nd3+ ) as an example, decom(1)
(2)
cub
position by HLF
takes place into the states E (6 ), G (8 ), G (8 ). Here
(1)
(2)
the energy of the states G (8 ) and G (8 ) depends in a more complicated
way on the ratio B04 /B06 and consequently on x (see Fig. 16)19) . This is the
consequence of the fact that the functions of both states transform according
to the same irreducible representation of the cubic group. The state E (6 ),
however, occurs once only. So, its energy is a linear function of x.
The multiplets of Ln3+ ions with the same J, e. g., the two pairs Tb3+ , Tm3+ (J = 6)
and Dy3+ , Er3+ (J = 15/2), respectively, are each reduced in cubic ligand fields into
the same irreducible representations.
Exact values for LF parameters cannot be calculated reliably. However, the sign
of Bqk values can often be predicted. Consequently, the signs of x and W can be
deduced, leading to the determination of the ground state.
Systems with non-integer J have exclusively magnetic states (showing Curie paramagnetism 20) ), while in the systems with integer J ligand field ground states may be
produced which are nonmagnetic, for example A1 (1 ), A2 (2 ) and E(3 ), yielding
temperature independent paramagnetism (TIP).
The x(ELF /W ) diagrams are helpful not only for cubic systems but also for coordination polyhedra that dont deviate greatly from cubic symmetry, that is, have
cubic pseudo symmetry.
Example 3.12 : Energy of the LF states of Pr3+ in Cs2 NaPrCl6
To verify Fig. 15 we use Pr3+ in Cs2 NaPrCl6 as an example with B04 = 1 938 cm1 and
B06 = 290 cm1 (Table 16). With the help of the entries in Table 23 as well as eqs. (102)
and (103), the following data result:

 
4
15
4

b4 = J F (4)B0 = W x =
1938 cm1 = 10.68 cm1
5 445
2



315
272
6

290 cm1 = 1.393 cm1

b6 = J F (6)B0 = W (1 |x|) =
4 459 455
4
x
b4
b4
=
= 7.665
=
x = 0.8845, W =
= 12.07
b6
1 |x|
x
Table 24 exhibits the LF energies corresponding with x = 0.8845 and W = 12.07.
Note that the calculated energy values, appearing in the last column of Table 24, are not
identical with the observed ones (see Table 17). One reason for the discrepancy is caused
by the fact that in our model only the ground multiplet 3H4 has been considered instead
of the full basis.
The energy difference between the LF term of highest energy (T2 ) and lowest energy
(A1 ) of the ground multiplet, usually called crystal field overall splitting (CFOS for short),
amounts to 660 cm1 (calc.) and 701 cm1 (obs.).
19)
20)

The states are listed as a function of x in ref. [16].

Curie paramagnetism means that m increases with decreasing T .

51

Tab. 24: Calculated energies ELF of the LF states of Pr3+ in Cs2 NaPrCl6 on
the basis of spectroscopically determined values B04 = 1938 and B06 = 290 cm1
[13].

ELF,calc. /cm1

a)

ELF,obs. b)

A1 (1 )

28 b4 80 b6 W [28x 80(1 |x|)]

410

T1 (4 )

14 b4 + 4 b6 W [14x + 4(1 |x|)]

144

266

236

E(3 )

4 b4 + 64 b6 W [4x + 64(1 |x|)]

46

456

422

T2 (5 )

26 b4 20 b6 W [26x 20(1 |x|)]

250 660c)

701

a)
b)
c)

W = 12.07; x = 0.8845; see Example 3.12.

Ref. [13].
Crystal field overall splitting (CFOS).

Problems
7. Determine the crystal field overall splitting (CFOS) for the Dy3+ ion in Cs2 NaDyCl6
on the basis of Fig. 35. Apply eqs. (102) and (103) as well as the entries in Table 23
and take the B04 and B06 values for Dy from Table 16. (Solutions are specified in
section 5.)

52

3.4
3.4.1

Non-cubic LF
Ligand field operators (single electron systems)

Lowering the symmetry results in LF operators with an increasing number of terms21) :

cylindrical (Dh, Cv 22) )
cyl = B 2 C 2 + B 4 C 4 + B 6 C 6
H
0 0
0 0
0 0
LF

(105)

hexagonal (D6h , D3h , D6 , C6v )

hex = B 2 C 2 + B 4 C 4 + B 6 C 6 + B 6 C 6 + C 6
H
LF
0 0
0 0
0 0
6
6
6
tetragonal (D4h , D4 , C4v , D2d )

tet = B 2 C 2 + B 4 C 4 + B 4 C 4 + C 4 + B 6 C 6 + B 6 C 6 + C 6
H
LF
0 0
0 0
4
4
4
0 0
4
4
4

(106)

(107)

trigonal (D3d , D3 , C3v )

tri
4
6
6
LF
H
= B02 C02 +B04 C04 +B34 C3
C34 +B06 C06 +B36 C3
C36 +B66 C6
+ C66 (108)

For d electron systems only terms with k 4 are relevant. The number of terms for f(d)
systems with cylindrical, hexagonal, tetragonal, trigonal symmetry is 3(2), 4(2), 5(3),
operators (including the symmetry Ih ) is given
6(3), respectively. A complete list of H
LF
in ref. [15].
3.4.2

cyl
4f 1 system (Ce3+ , 2F5/2 : HLF
)

In the following calculations the basis functions are restricted to the multiplet 2F5/2 23) .
cyl (2F5/2 ) = B 2 C 2 + B 4 C 4
Ligand field operator: H
0 0
0 0
LF

(109)

Operator (109), acting on 2F5/2 , produces solely diagonal elements (q = 0). The matrix
elements hMJ |C0k |MJ i for k = 4 are known from matrix (83). The data for k = 2 are
obtained with the help of Table 8 and Table 13:
h21 |C02 | 12 i =

8
;
35

h32 |C02 | 32 i =

2
;
35

h52 |C02 | 52 i = 10
.
35

In Table 25 the results are given.

4
= 41 35 sin3 cos ei3 ;
C02 = 12 3 cos2 1 ; C3

1
1
6
6
105 sin3 cos 11 cos2 3 ei3 ; C6
= 32
231 sin6 ei6 .
C3
= 16
cyl
22)
HLF applies also for D4d .
23)
cyl and H
hex are identical.
Note that under this restriction the effects of H
LF
LF
21)

53

|MJ i

cyl
Table 25: Splitting of the 2 F5/2 [4f 1 ] multiplet by H
LF

(0)

En,LF Wn

| 21 ia)

8
B2
35 0

2
B4
21 0

| 23 i

2
B2
35 0

3
B4
21 0

| 25 i

B2 +
10
35 0

1
B4
21 0

a)

Wavefunctions adapted to
the magnetic field in x-direction:


| i = 12 | 21 i + | 12 i .

In a cylindrical /hexagonal LF 2F5/2 is split in three doublets whose energy separations

are given by B02 and B04 .
3.4.3

cyl
4f 1 system (Ce3+ , 2F5/2 : HLF
+ Hmag )

In a cylindrical/hexagonal LF the multiplet 2F5/2 is split in three doublets depending on

B02 and B04 (see section 3.4.2). To gain the susceptibility equation the influence of the
magnetic field must be considered parallel to the unique axis z and in a second direction
perpendicular to z, that is, x.
mag,z = eg Jz Bz (g = 6 )
Magnetic field parallel to the unique axis:
H
J
J
7
The results are available in the third column of Table 26. Applying Van Vlecks equation
(60) the susceptibility equation for mk is
NA 2B 2

where
3kB T effk





1
9 25
1
2
3
+ +
exp
exp
4
kB T
4
4
kB T
2




= gJ
,
1
2
exp
+ 1 + exp
kB T
kB T

(110)

mk = 0

2effk
and

(0)

(0)

(0)

(0)

1 = W1/2 W3/2 =

6
B2
35 0

gJ =

6
7

5
B4
21 0

B2 +
2 = W5/2 W3/2 = 12
35 0

4
B4.
21 0

cyl , H
magk (z) and H
mag (x);
Table 26: Splitting of the 2 F5/2 [4f 1 ] multiplet by H
LF
(0)
(1)
(1)
(2)
energies ELF Wn , Zeeman coefficients Wn,z and Wn,x , Wn,x
|MJ i

(0)

En,LF Wn

(1)

(1)

(2)

Wn,z /gJ B

Wn,x /gJ B

Wn,x /gJ2 2B a)

| 12 ib)

8
B2
35 0

2
B4
21 0

21

32

2/1

| 32 i

2
B2
35 0

3
B4
21 0

23

5/(42 ) 2/1

| 52 i

10
B2 +
35 0

1
B4
21 0

25

5/(42 )

a)
b)

(0)

(0)

(0)

(0)

5
6
2
B02 + 21
B04 ;
2 = W5/2 W3/2 = 12
1 = W1/2 W3/2 = 35
35 B0 +


mag : | i = 1 | 1 i + | 1 i .
Wave functions adapted to H
2
2
2

54

4
4
21 B0

Magnetic field perpendicular to the

unique axis:


1
mag,x = egJ JxBx = egJ
H
J+ + J Bx
2
mag | 5 i
h 25 |H
x
2
mag | 5 i and h 3 |H
mag | 3 i are zero. On account of the
The matrix elements h 25 |H
x
x
2
2
2
1
3
1
off-diagonal element h 2 |Hmagx | 2 i = 2 gJ B Bx , the correct zeroth-order functions have
to be determined:


 W (1) 3 g 


n

(1)
1
1
3
2 J B 
1
|
g

;
|
i
=
i
+
|

i
.
=
0;
W
=

n(1,2)
2 J B
2
2
2
 2 gJ B Wn(1) 
The H matrix reads (in units of gJ B Bx ):

5

2

5
2 

3

2

3
2 


5  5 3  3 
 





 +

2
2
2
2

5
2

5
2

5
2

3
2

5
2

23
(0)

(1)

(2)

In Table 26, column 4 and 5 the results are collected. Substituting Wn , Wn,x and Wn,x
in eqn. (60) yields the expression for m :

NA 2B 2

where
(111)
3kB T eff



 



9 4kB T
1
5kB T
4
5
2
2
= gJ 3
exp
+
kB T

+
exp
4
1
kB T
1 22
22
kB T


1


2
1
+ 1 + exp
exp
kB T
kB T

m = 0
2eff

and

(0)

(0)

(0)

(0)

1 = W1/2 W3/2 =

6
B2
35 0

5
B4
21 0

2 = W5/2 W3/2 = 12
B2 +
35 0

4
B 4.
21 0

For a polycrystalline sample the average susceptibility is

m = 31 mk + 2 m .

(112)

This model, extended by a molecular field parameter MF and a temperature-independent

parameter 0
(m 0 )1 = m1 MF
has been applied to the intermetallic CePt5 (CaCu5 -type: Ce3+ [4f 1], D6h ). A good adaptation to the experimental data was obtained with the parameter values B02 = 1108 cm1 ,
B04 = 104 cm1 , MF = 0.43 105 mol m3 , and 0 = 63 1011 m3 mol1 [24, 25].
55

Problems
8. In which respect does the magnetic susceptibility of a mononuclear lanthanide systems with Ln point symmetry (i) Oh , (ii) Dh, (iii) D2h differ?I

Isotropic spin-spin exchange couplings in dinuclear

compounds

4.1

Parametrization of exchange interactions

Heitler-London model of H2

r1 2
1

ra 2

rb 1

rb2

ra1
a

ra b

Fig. 17: H2 model; a and b assign the nuclei, 1 and 2 the

electrons

Valence bond ansatz: construction of products with orbital configuration a b using the
four spin orbitals
a

Product states in consideration of the Pauli principle:

D1
D2
D3
D4

=
=
=
=

a (1)(1) b(2)(2) a (2)(2) b(1)(1)

a (1)(1) b(2)(2) a (2)(2) b(1)(1)
a (1)(1) b(2)(2) a (2)(2) b(1)(1)
a (1)(1) b(2)(2) a (2)(2) b(1)(1)

(113)
(114)
(115)
(116)

Construction of eigenfunctions of the total spin

= S
1 + S
2
S

(S1 = S2 = 1/2)

2 = (S
1 + S
2 )2 and Sz = Sz1 + Sz2 with S = 1
D3 and D4 are eigenfunctions | S MS i of S
and MS = 1 and 1 (Spin triplet functions | 1 1 i and | 1 1 i), respectively, while linear
combinations of D1 and D2 yield functions with MS = 0:
D1 + D2 = | 1 0 i (spin triplet function);
D1 D2 = | 0 0 i (spin singlet function)
1 = D1 D2
|00i

2 = D3
|11i

3 = D1 + D2
|10i
56

4 = D4
| 1 1 i

a and b are normalised:

Z
Z

a (1) a (1)d1 = b (2)b (2)d2 = 1,

but not orthogonal:
overlap integral Sab =

a (1) b (1)d1 =

a (2) b (2)d2 6= 0

Normalised functions of the dinuclear unit:

q
1 = Ng [a (1)b(2) + a (2)b (1)] 12 [(1)(2) (2)(1)]
|
{z
}|
{z
}
sym
anti

(1)(2)
q
1
= Nu [a (1)b(2) a (2)b (1)]
[(1)(2) + (2)(1)]
3
2
|
{z
}

4
anti
(1)(2)
|
{z
}
sym

2
where Ng = (2 + 2Sab
)

1/2

2
and Nu = (2 2Sab
)

1/2

Symmetry of the functions with regard to exchange of electrons

total function:
singlet function:
triplet functions:

anti
i (1, 2) = i (2, 1)
orbital sym (g), spin function anti
orbital anti (u), spin function sym

Symmetry of the orbital forces a distinct multiplicity of the spin function on account
of the Pauli principle

Evaluation of the energy E(S) and E(T ) of the singlet and triplet states
h
2 2
e2
e2
e2
e2
e2
h
2 2

+
H=
(1)

(2)

2me
r
rb1
2me
r
rb2
r12
{z a1
}|
{z a2
}
|

h(1)
h(2)
1 g i = 2(h + hab Sab ) + Jab + Kab
E(S) = h1 g1 |H|
1
2
1 + Sab
3 u i = 2(h hab Sab ) + Jab Kab
E(T ) = h3 u2 |H|
2
2
1 Sab
where
57

(117)
(118)

h =
=
hab =
Jab =
Kab =



E
 
(one-centre
a (i) h(i) a (i)


D
E
 
b (i) h(i)
one-electron integral)
 b (i)


E
D


(transfer or hopping integral)
a (i) h(i) b (i)
 2



a (1)b(2) e /r12  a (1)b (2)
(Coulomb integral)

 2


(Exchange integral).
a (1)b(2) e /r12  a (2)b (1)

Example: Evaluation of E(S) =

1. Integration over the spin:
1
2

g1

 
E
 1 g
H  1

h(1)(2) (1)(2)|(1)(2) (1)(2)i =


1
h(1)(2)|(1)(2)i + h(2)(1)|(2)(1)i
2 |
{z
} |
{z
}
1
1

h(2)(1)|(1)(2)i h(1)(2)|(2)(1)i = 1.
{z
} |
{z
}
|
0

2. Integration over the space:

E(S) =


D


2
2

Ng a (1)b(2) + b (1)a (2) h(1) + h(2) + e /r12  a (1)b (2)+

E
b (1)a (2) = 2Ng2 [2(h + hab Sab ) + Jab + Kab ] .

(119)

Singlet-triplet splitting:
E(T, S) = E(T ) E(S)
2
2Kab 4hab Sab + 2Sab
(2h + Jab )
Application of the Heitler-London model to dinuclear complexes having S1 = S2 =
centres

1
2

Example: Ln Cua L Cub Ln

As distinguished from the strong covalent bond in H2 the interactions between both
magnetically active electrons is weak.
small E(T, S).
The highest singly occupied antibonding orbitals a and b of the fragments Ln Cua L and
LCub Ln , respectively take over the role of the 1s orbitals of the H atoms. a and b have
mainly d character. They are centered at the metal ions and partially delocalised in the
direction of the ligands.

58

4.2

Heisenberg operator

The interaction between the unpaired electrons of centre 1 with the ones of centre 2 are
phenomenologically described by an apparent spin-spin coupling, whose magnitude and
sign are given by the spin-spin coupling parameter (exchange parameter) Jex :
ex = 2Jex S
1 S
2
H

where

2Jex = E(T, S)

(120)

is an effective operator, describing but not explaining the spin-spin coupling.

H
ex
to 1 g (S = 0) and 3 u (S = 1, i = 2, 3, 4):
Application of H
ex
1
i




g
g
1
1
3
1 S
2 3 ui = 1 Jex 3 ui
1 S
2 1 =

,
2J
S
2Jex S
J
ex
ex
1
2
}
| 2{z }
| {z
E(S)
E(T )

E(T, S) = E(T ) E(S) = 2Jex

(121)

Jex < 0: singlet ground state (intramolecular antiferromagnetic interaction)

Jex > 0: triplet ground state (intramolecular ferromagnetic interaction)

Hints to the evaluation of E(T ) and E(S):


2
2 = S
1 + S
2 = S
21 + S
22 + 2S
1 S
2
S


1 S
2 = S
2 S
21 S
22 = h
2S
2 S (S + 1) S1 (S1 + 1) S2 (S2 + 1)
| {z } | {z } | {z }
3
3
0 or 2
4

Heisenberg operator for more than two centres:

X
ex = 2
i S
j
H
Jex,ij S
(122)
i<j

4.3

ex = 2Jex
H

i S
j
S

(123)

i<j

Exchange-coupled species in a magnetic field



ex = 2Jex S
1 S
2 = 2Jex Sz1 Sz2 + Sx1 Sx2 + Sy1 Sy2
H
h

i
= 2Jex Sz1 Sz2 + 1 S+1 S2 + S1 S+2
2

(124)

Basis: spin functions in the form | MS MS i where the first MS refers to electron 1 and
the second to electron 2
H-Matrix:
1


1
2 2

MS1 MS2

1 1
 Jex /2
2 2

1 1
2 2 

1 1
2 
2

1 1
2 2 

1
2 2

 1 1  1 1


2
2
2
2

Jex /2

Jex

Jex

Jex /2
Jex /2
59

(125)

Evaluation of the diagonal element H11 :





2Jex 21 21  Sz1 Sz2  21 12 = 2Jex 21 21 = Jex /2

Evaluation of the off-diagonal element H23 :





2Jex 12 21  12 S1 S+2  12 12 = Jex (1)(1) = Jex
Results:

Tab. 27: Spin functions and exchange energies of the S1 = S2 = 21 system

Spin function
MS S
E
 1 1   1 1 

3
 
1
0 0
J
2
2
2 2
2 ex
2
1 1

1
2 2





 1 1 +  1 1
1
0 1 21 Jex
2
2
2 2
2
 1 1

1
2
2

2 S +1

E
30 J

20 J

12 J

6J
2J
0

11

S
5

Fig. 18: Relative energies and multiplicities of the spin states of a dinuclear Fe3+ complex
(S = 52 ); for Cu2+ (S = 21 ) only the two lowest levels are relevant, while for Gd3+ (S = 72 ) the
two levels with S = 6 (E = |42Jex |) and S = 7 (E = |56Jex |) have to be added.

60

 Magnetic susceptibility of a spin-spin-coupled system with S1 = S2 =

10
MS

/ 10 m mol

-7

S=1 4

-1

E(T,S)

-1

S= 0 0

E / cm

-1

10

1
2

-2

3
2
B0 / T

Fig. 19: Correlation diagram of a S1 =

S2 = 21 system under the influence of
isotropic intramolecular spin-spin coupling (Jex = 2 cm1 ) and applied field

15
10
T /K

20

25

Fig. 20: m versus T diagram of a

S1 = S2 = 12 exchange-coupled system
with Jex = 2 cm1 at applied fields of
B0 = 0.01 T (a), 3.5 T (b), and 5 T (c)

Application of the Van Vleck equation (60) to a dinuclear system with S1 = S2 =

Zeeman operator:
Mz = e g(Sz1 + Sz2 ) Bz = e g Sz Bz
H
{z
}
|
(1)

1
2

(126)


 

 
 
0 0
1 0
 1 1
S MS  1 1


1 1  gB Bz


0
1 0


1 1 
gB Bz


0 0
0

Matrix elements:

(1)

|11i = g Bz W
h11|H
B
Mz
|11i = gB
(1)

|1 1i = g Bz W
h1 1|H
B
Mz
|11i = gB
(1)

(1)

(0)

The remaining matrix elements (Zeeman coefficients) are zero. W|11i , W|11i , WS
(0)
WT

E(S), and
= E(T ) are substituted into the Van Vleck equation. After dividing by
2, the Bleaney-Bowers expression (m per centre) is obtained, here extended by 0 .

1

1
2Jex
NA 2B g 2
1 + exp
+ 0 , only applicable to a (Fig. 20)
(127)
m = 0
3kB T
3
kB T
61

e
c

0.8

-3
7
-1
cmol
/ ( 10 mol m )

cmol / ( 10-7 m3 mol -1 )

1.0

d
0.6
b

0.4
0.2

0.0
200

100

300

100

200

300

400

T/K

Model calculations concerning

the system S1 = S2 = 12 with
positive and negative Jex ;

2.5
e

2.0

Jex values [cm1]:

curve a: 50
curve b: 25
curve c: 0
curve d: +25
curve e: +50

d
c

1.5

meff

a
b
c
d
e

400

T/K

10
9
8
7
6
5
4
3
2
1
0

1.0
0.5

0.0
0

100

200

300

Fig. 21 a: m T diagram
Fig. 21 b: 1
m T diagram
Fig. 21 c: eff T diagram

400

T/K

Polynuclear unit of n equivalent centres:

m =

0 NA 2B g 2
n

3kB T



)
S (S + 1)(2S + 1)(S ) exp E(S
kB T


P

E(S )

S (2S + 1)(S ) exp kB T

Evaluation of S , E(S ) and (S ):

S of the coupled states:

S = nS, nS 1, . . . , 0 (nS integer) or

1
2

(nS half integer)

Relative energies E(S ):

zJex
E(S ) =
[S (S + 1) nS(S + 1)]
n1
z: number of nearest neighbours of a centre
n: number of interacting centres
62

(128)

Frequency (S ) of the states S : (S ) = (S ) (S + 1)

(S ) is the coefficient of X S in the expansion (X S + X S1 + . . . + X S )n

Scope of validity of the Heisenberg model
1. The magnetic moments are localised (no band magnetism)
2. S of the metal ions is a good quantum number
3. the ground term of the metal ion has no first order orbital moment
3d5 high spin
Examples:

3d (Oh ), 3d (Td )
4f 7

4.4

3d3 (Oh ), 3d7 (Td )

A1

A2

S7/2

3d (Oh )

A2

Setup of computational procedures

Application of the computer program CONDON to

4fN -4fN (homodinuclear complexes)
4fN -S

where S =

1
2

[Cu2+ ],

1,

3
2

[Cr3+ ],

7
2

[Gd3+ ] (heterodinuclear compl.)

 X
N
N
2
e2 X
h

i + V (ri ) +
(ri)
li si

+
H =
2m
r
e
ij
i>j
i=1
i=1
{z
} | {z } |
{z
}
|
(0)

SO
ee
H
H
) N
(H
k
N

X
XX
X


k
Bqk Cq
(i) + (1)q Cqk (i)
B0k C0k (i) +
B (
li + 2
si ) B0
+
+
N 
X

q=1

i=1 k=0

ex = 2Jex
H

N X
N
X

{z
LF
H

sk
sl

|i=1

{z

Hmag

B0 = 0.001 T (in general)

k=1 l=1

With regard to the LF symmetry of the lanthanide our choice is cubic and cylindrical for
two reasons: (i) the former is isotropic and the latter anisotropic, (ii) the number of LF
parameters is low [15] (cf. scheme (129)) 24 ) .
cubic
point group
no. of Bqk s
Bqk
24)

Oh
Td
2
B04 (+B44 )
B06 (+B46 )

cylindrical hexagonal tetragonal

Dh
Cv
3
B02 , B04
B06

D6h , D3h
D6 , C6v
4
B02 , B04
B06 , B66

trigonal

D4h , D4
D3d , D3
C4v , D2d
C3v
5
6
B02 , B04 , B44 B02 , B04 , B34
B06 , B46
B06 , B36 , B66

(129)

For the calculations we use only B04 as LF parameter, except for Ce3+ (cyl) where B02 is considered.

63

Cubic (Oh , Td )
N
X

cub =
H
LF

i=1

n
n
p
p
 4
o
 6
o
B04 C04 (i)+ 5/14 C4
(i) + C44 (i) + B06 C06 (i) 7/2 C4
(i) + C46 (i)
B44 =

fixed relations:

p
B46 = 7/2B06

p
5/14 B04 ,

Cylindrical (Dh )
cyl =
H
LF

N
X

B02 C02 (i) + B04 C04 (i) + B06 C06 (i)

i=1

Bqk and Jex data (cm1 )

Symmetry options
symmetry

cub

cyl

Ln

Bqk

Bqk

B04 (+B44 )

B02 (Ce3+ ) or B04

Ce (cub)

1 100

Ce (cyl)

coordination polyhedron
B0k > 0

octahedron

linear

Nd,Sm,Tb,Dy

B0k < 0

cube, tetrahedron

square antiprism

(cub, cyl)

4.5

Homodinuclear lanthanide complexes

4.5.1

Gd3+ Gd3+

2 000

S1 , S2 = 27 , n = 2:

0 NA 2B g 2 e2x + 5e6x + 14e12x + 30e20x + 55e30x + 91e42x + 140e56x

kB T
1 + 3e2x + 5e6x + 7e12x + 9e20x + 11e30x + 13e42x + 15e56x
where x = Jex /(kB T )

(130)

12
-1

Jex = 0.5 cm
Jex = 0

10

-1

Jex = -0.5 cm

eff

8
6

Fig. 22: Homodinuclear complex of

Gd3+ : eff vs. T plot, Jex = 0, 0.5
cm1 (B0 = 0.001 T).

4
2
0
0

50

100

150

200

250

300

T/K

64

5
1

a)
Occasionally, the selected |Jex |
values are larger than realistic values in
order to inspect more distinctly the Hex
effect (cf. Tab. 1).

Application of eq. (128) with S = |S1 S2 |, |S1 S2 + 1|, . . . S1 + S2 ;

m =

600

Jex a)

4.5.2

Group theoretical tools II

We saw in 3.3.1 that on the basis of group theory it is possible to assign the LF split
terms of Ln ions. Symmetry aspects are also useful to assign the splitting produced by
cub (eq. (86)) + H
iso (eq. (120)) on the product functions of the
the combined effect of H
LF
ex
two interacting metal ions. The following examination is restricted to the simple case of
homodinuclear Ln systems equipped with orbital contributions of the magnetic moments.
iso in singlets, doublets and triplets.
The product states split under the action of H
ex
The split pattern is obtained in two steps:
application of ab (R) = a (R)b (R) , where i is the character and R is a symmetry operation, given in the headlines of Table 28.
Tab. 28: Character table of the symmetry groups O and Td

O
Td
1 (A1 )a)
2 (A2 )
3 (E)
4 (T1 )
5 (T2 )
1 1 b)
2 2
3 3
4 4
5 5
1 2
a)
b)

E 8 C3 3 C2 6 C4 6 C2
E 8 C3 3 C2 6 S4 6 d
1
1
1
1
1
1
1
1 1 1
2 1
2
0
0
3
0 1
1 1
3
0 1 1
1
1
1
1
1
1
1
1
1
1
1
4
1
4
0
0
9
0
1
1
1
9
0
1
1
1
1
1
1 1 1

irred. reps. of the

coupled states

1
1
1 2 3
1 3 4 5
1 3 4 5
2

Bethe (Mulliken) nomenclature.

Selection of direct products for dinuclear systems with integer J.

a (R) and b (R) are the characters of the irreducible representations a and b ,
while ab (R) is the character of the (in general reducible) product state.
Example: The character system of the product functions ab
3 resulting from the
single-ion states a3 and b3 , 3 3 , is given in line 8 of Table 28.
P
application of ai = h1 R i (R) (R) to determine the irreducible parts of ab
3 ;
ai : frequency of i in ; h: order of the group;
i (R): character of the ith irreducible representation
i for the group element R;
P
(R): character of of the group element R;
R : sum over all group elements.
1
24 (1
1
24 (1
1
24 (1

a1

a2

a3

a4

= a5 = 0;

1 4 + 8 1 1 + 3 1 4 + 6 1 0 + 6 1 0) = 1

1 4 + 8 1 1 + 3 1 4 6 1 0 6 1 0) = 1

2 4 + 8 (1) 1 + 3 2 4 + 6 0 0 + 6 0 0) = 1
=

3 3 1 2 3

Result: 3 3 is reducible to two singlets (1 , 2 ) and a doublet (3 ). The energies

of the states depend on sign and size of Jex .
65

For metal ions with half-integer J three types of cubic states exists: the doublets 6 (E ),
7 (E ), and the quartet 8 (G ) (cf. Tab. 29).
Tab. 29: Character table of the double groups O and Td

Td

1 (A1 )a)
2 (A2 )
3 (E)
4 (T1 )
5 (T2 )
6 (E )b)
7 (E )
8 (G)

1
1
2
3
3
2
2
4

4 C3
4 C23
3 C2
3 C4
3 C34
6 C2
4 C23 R 4 C3 R 3 C2 R 3 C34 R 3 C4 R 6 C2 R
4 C3
4 C23
3 C2
3 S4
3 S34
6 d
R
2
3
4 C3 R 4 C3 R 3 C2 R 3 S4 R 3 S4 R 6 d R
R

1
1
2
3
3
2
2
4

1
1
1
0
0
1
1
1

1
1
1
0
0
1
1
1

1
1
2
1
1
0
0
0

1
1
0
1
1

2
0

1
1
0
1
1

2
0

1
1
0
1
1
0
0
0

0
0
0
0
0
0

0
0
1
1
0
0
1
1
1
1
0
0

0
0
1
1
0
0
1
1
1
1
0
0

2
2
0
2
0
0

2
0

2
0
0
0
0
0

2
2
0
2
0
0

2
0

2
0
0
0
0
0

0
0
0
0
0
0

4 6 d) 6 6
5 6
6 6
2 6
2 2
3 6
4 4
4 7
6 6
5 7
6 6
2 7
2 2
3 7
4 4
2 8
4 4
3 8
8 8
4 8
12 12
5 8
12 12

1
1
1
1
1
1

6 6 c) 4
7 7
4
8 8
16
6 7
4
6 8
8
7 8
8

a)
b)
c)
d)

4
4
16
4
8
8

1
1
1
1
1
1

0
0
0
0
0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0
0
0
0
0

irreducible representations
of the coupled states

1 4
1 4
1 2 3 24 25
2 5
3 4 5
3 4 5
6 8
7 8
7
8
7 8
6 8
6
8
8
6 7 8
6 7 2 8
6 7 2 8

1 5 : integer J; 1 3 : non-magnetic.
6 8 : half-integer J; magnetic.
Direct products where both metal ions have half-integer J or S.
Direct products where one metal has integer J and the other has half-integer J or S.

66

4.5.3

Ce3+ Ce3+ (cub)

Fig. 23 exhibits the 4f 1 ions cubic

LF splitting of ground and excited
multiplet 2F5/2 and 2F7/2 , respectively
(2F5/2 7 + 8 ; 2F7/2 7 + 8 +
6 ). The LF ground state in octahedral coordination is the doublet 7
and in tetrahedral/cubical coordination the quartet 8 .
iso
Switching on Hex
with Jex = 5 cm1
between the two Ce3+ ions each
in octahedral LF the four product states 7 7 split into 1 4
(see Fig. 24) with ground state 1
for Jex < 0 and 4 for Jex > 0.
The splitting of the excited 16 prodiso
uct functions 8 8 by Hex
is more
complicated (see caption of Fig. 24).
Figs. 25 and 26 display the variations
Fig. 23: Ce3+ [4f 1 ]: Cubic LF splitting of the
eff vs. T for octahedral and tetrahemultiplets 2F5/2 and 2F7/2 ; left: octahedral LF;
dral/cubical LF and positive and negright: tetrahedral/cubical LF.
ative Jex . Note that eff for the 8 LF
ground state is distinctly larger than for the doublet 7 (maximum at 2.2 and 1.3, respectively, at T = 1 K).

Fig. 24: Splitting of the states of an exchange-coupled CeCe unit in an octahedral LF

(B04 = 1100 cm1 ); 7 7 1 4 : Jex > 0, triplet ground state (4 ); Jex < 0, singlet
ground state (1 ); 8 8 1 2 3 2 4 2 5 .

67

Fig. 25: eff vs. T plot of an exchange-coupled

Ce-Ce unit in an oct. LF (B04 = 1 100 cm1 );
Jex > 0, triplet ground state (4 ); Jex < 0,
singlet ground state (1 ).

Fig. 26: eff vs. T plot of an exchange-coupled

Ce-Ce unit in a cubical/tet. LF (B04 = 1 100
cm1 ); Jex > 0, triplet ground state; Jex < 0,
singlet ground state.

Remark to Fig. 26: The product states 8 8 are split by Hex into 1 (singlet (1))
2 (singlet (1)) 3 (doublet (2)) 2 4 (triplets (3)) 2 5 (triplets (3)). For Jex > 0, the
sequence of split terms is given with increasing energy by (3) (3) (3) (1) (3) (2)
(1); for Jex < 0, this order is reversed. (Remember that only the triplets are magnetic.)
Ce3+ Ce3+ (cyl)
310

(2)

| 1/2 i
(2)

| 3/2 i
ELF /cm1
0

(2)

| 5/2 i

Scheme: Cylindrical LF splitting of the

multiplet 2F5/2 (B02 = 600 cm1 ; ground
state: | 25 52 i). Figure 27 exhibits that for
k

Jex > 0 and T 0 eff = 5.2 while

eff 0
(not shown in the figure), that is, strong
magnetic anisotropy is observed;
eff
and
eff represent the temperature
dependence of the average eff values for
positive and negative Jex , respectively.

Fig. 27: eff vs. T plot of an exchange-coupled

CeCe unit in a cylindrical LF (B02 = 600 cm1 ).

68

4.5.4

Nd3+ Nd3+ (cub)

For B04 > 0 (oct. LF; W < 0, x > 0; cf. Ex. 3.11)
the LF ground state is 6 while for B04 < 0 (cub/
(2)
tet) 8 is lowest in energy (see Fig. 28).

B04 < 0

B04 > 0

Fig. 28: Splitting of the Nd ground multiplet

(1)
(2)
by cubic LFs (4I9/2 6 8 8 ) [16].

Fig. 29: Splitting of the 8 8 product states

of NdNd by Hex ( 1 2 3 2 4 2 5 ).
4.0

4.0

-1

-1

Jex = 1 cm

Jex = 1 cm

Jex = 0

Jex = 0

-1

Jex = -1 cm

-1

Jex = -1 cm

3.5

3.5

3.0

eff

eff

3.0

2.5

2.5

2.0

2.0

1.5

1.5
0

10

20

30

40

50

10

20

30

40

50

T/K

T/K

Fig. 30: eff vs. T plot of an exchange-coupled

NdNd unit in an octahedral LF (LF ground
state 6 ). For Jex > 0, the ground state is
magnetic (triplet 4 ), while for Jex < 0, the
ground state is nonmagnetic (singlet 1 ), that
is, eff 0 at T 0.

Fig. 31: eff vs. T plot of an exchange-coupled

NdNd unit in a cubical/tetrahedral LF (LF
ground state 8 ). For both situations, Jex < 0
and Jex > 0, the ground state is a singlet, that
is, eff 0 as well as eff 0 at T 0 (cf.
Fig. 29).

Results: Fig. 30: Only a little increase of eff is observed after switching on Hex ;
Fig. 31: For the ferromagnetically coupled Nd-Nd the peak of eff at 1 K is remarkable.
69

4.5.5

Sm3+ Sm3+ (cub)

0.8

0.40

-1

Jex = 1 cm

-1

Jex = 1 cm

Jex = 0

0.7

Jex = 0

0.35

-1

Jex = -1 cm

-1

0.30

0.6

0.25

0.5

0.20

0.4

eff

eff

Jex = -1 cm

0.15

0.3

0.10

0.2

0.05

0.1
0.0

0.00
0

10

12

14

T/K

10

12

14

T/K

7 7 1 4

8 8 1 2 3 24 25
Fig. 33: eff vs. T plot of an exchange-cou-

Fig. 32: eff vs. T plot of an exchange-coupled SmSm unit in oct. LF (LF ground state
7 ). For Jex > 0 the eff values are only little
enhanced; eff 0 for Jex < 0 .

4.5.6

pled SmSm unit in cub/tet. LF (LF ground state

8 ). Although Jex > 0 the eff values fall below
the curve of the uncoupled system. (Obviously the
magnetic contribution of the uncoupled Sm3+ ion with LF
ground state 8 is larger than the contribution of the Hex
triplet ground state of the Sm-Sm pair.)

Tb3+ Tb3+ (cub)

Taking into consideration that in the
case of Tb3+ [4f 8 ] the parameter W is
negative for both octahedral and cubical/tetrahedral coordination polyhedra (see Table 23) a) , the inspection of Fig. 34 shows that apart
from the point of intersection of 2 ,
(2)
3 and 5 at E 85, x = 0.8,
exlusively non-magnetic LF states
are ground states. In consequence,
the Tb3+ ion in cubic LF is not a
good candidate for the investigation
of spin-spin couplings between each
other.
a)

Both Examples 3.10 and 3.11 show

how to read the xELF /W diagrams.

B04 > 0

TIP at T 10 K

B04 < 0

Fig. 34: Splitting of the Tb ground multiplet by a cubic

(1)
(2)
LF (7F6 1 2 3 4 5 5 ) [16].

70

Dy3+ Dy3+ (cub)

4.5.7

B04 > 0

B04 < 0
(1)

Fig. 35: Splitting of the Dy3+ ground multiplet by a cubic LF (6H15/2 6 7 8

(2)

(3)

8 8 ) [16]. The sequence of LF states for octahedral coordination is given at x = 1

(B04 = 2000 cm1 , LF ground state 6 ) and for cubical/tetrahedral coordination at x = 1
(1)
(B04 = 2000 cm1 , LF ground state 8 ).
12

12

10

10
-1

-1

Jex = 1 cm

Jex = 1 cm

Jex = 0

Jex = 0

8
-1

-1

Jex = -1 cm

eff

eff

Jex = -1 cm

0
0

20

40

60

80

100

120

140

20

40

60

80

100

120

140

T/K

T/K

Fig. 36: eff versus T plot of an exchangecoupled DyDy unit in an octahedral LF (B04 =
2 000 cm1 ; LF ground state 6 ).

Fig. 37: Exchange-coupled dinuclear Dy unit

in a cubical/tetrahedral LF (B04 = 2 000 cm1 ;
(1)
LF ground state 8 ).

Results: For LF ground state 6 (Fig. 36) Jex > 0 enhances eff in the whole temperature
range while for Jex < 0 eff diminishes. For LF ground state 8 (Fig. 37) it is remarkable
that for Jex > 0 eff drops steeply from 11.3 to 2.2 far below eff = 8.9 of the uncoupled
Dy-Dy pair, caused by a very weak magnet contribution of the ground triplet.
71

Dy3+ Dy3+ (cyl)

4.5.8

(2)

420 | 11/2 i
(2)

404 | 9/2 i
..
..
..
.
.
.
(2)

315 | 7/2 i
(2)

305 | 13/2 i
..
..
..
.
.
.
(2)

206 | 5/2 i
..
..
..
.
.
.
(2)

114 | 3/2 i

ELF /cm1
(2)

62

| 1/2 i

(2)

| 15/2 i

Fig. 38: Exchange-coupled dinuclear Dy unit in cylindrical LF

(B04 = 2 000 cm1 ; LF ground state |J MJ i: |15/2 15/2i).
DyDy(cyl)
Dy(cyl)Cr
Tb(cyl)Cr

B04 = 2000 cm1

4.5.9

eff

eff

see 4.6.5
see 4.6.5

LnLn pair: Temporary r

esum
e of eff for Jex > 0 (rating +/)
eff a)

3+

Ln

free ion
cub
cyl
a)

Jex /cm1
+1
+5
+5

Ce3+

Nd3+

Sm3+

Tb3+

Dy3+

2.54

3.62

0.84

9.72

10.64

+/

2.3

3.8

0.48

TIP

11.2

++

b)

b)

++

12.7

24.0

5.2

At T = 2 K.

b)

Examination is still pending.

72

1. Cubic Ln systems are no good

candidates for large eff ;
2. Particularly Sm and Tb are
unsuitable in cubic LF;
3. Dy in cylindrical LF is an excellent candidate.

4.6

Heterodinuclear transition metallanthanide systems

4.6.1

p(2 P1/2 )s(2S1/2 ) molecule

As an introduction, the spin-spin-coupling Hex within a p1 s1 unit is considered, where

the six states of p1 are split by HSO into the quartet 2 P3/2 and the doublet 2 P1/2 . We
assume HSO Hex and restrict the calculations to 2 P1/2 2 S1/2 . The corresponding basis
functions |JMJ i = | 21 12 i of the multiplet 2 P1/2 are given in Table 3025) .
Table 30:

P1/2 functions |JMJ i = | 12 12 i of the spin-orbit coupled p1 system



 ML MS



 J MJ

q 
 q 
 1

3 23  1 12 + 13  0 12
2
5

1
2

q 
 q1 
 1 1
1
2 

1
3  0 12
3
2
2
2

MJ = ML + MS J
1
2

1
2

P1/2

21

ex can be
Since the calculations are restricted to the ground multiplet J = 1/2, H
applied in the form


ex = 2Jex (g 1) J

H
S
=
2(g

1)J
J
S
+
J
S
+
J
S
ex
z z
x x
y y
J
| J {z }
De Gennes

i
h
1

(131)
= 2(gJ 1)Jex Jz Sz + 2 J+ S + J S+
with the advantage, that for the p system the basis functions |J MJ i can be used.

Basis functions of ps: | MJ MS i, MJ referring to the p electron and MS to the s electron

According to g1/2 = 23 for 2 P1/2 (see eq. (2)), the De Gennes factor is g1/2 1 = 13 , and
(gJ 1)J is the component SJ of S in direction of J .
H-Matrix of operator (131):

  1 1 1 1  1 1



MJ MS  21 12
2 2
2
2
2
2

1 1 1

J
2 2
6 ex

1 1
1
2 2 
J
16 Jex
3 ex

1 1
1
2 
J
16 Jex
2
3 ex

1 1
1
J
2 2 
6 ex

(132)

Evaluation of the diagonal element H11 :





2( 31 )Jex 12 21  Jz Sz  12 21 = ( 23 )Jex 12
25)

1
2

= 16 Jex

Tab. 31: Functions and exchange energies of the J = 21 ,

S = 12 system
Function
E
 1 1   1 1 

 
1
12 Jex
2
2
2 2
2
1 1

2 2
 1 1   1 1 

1
 +
1
J
2
2
2 2
6 ex
2
 1 1

2
2

The complete set of basis functions of 2 P1/2 and 2 P3/2 is given Table 6.

73

Evaluation of the off-diagonal element H23 :





( 23 )Jex 12 21  ( 21 )J S+  12 12 = ( 32 )Jex ( 12 )(1)(1) = 31 Jex

Result: Compared to the singlet-triplet splitting of an exchange-coupled s1 s1 molecule,

the situation is inverted for a p1 s1 molecule if strong spin-orbit coupling within the p1
system is taken into consideration: In the case of antiferromagnetic exchange coupling
(Jex < 0), the triplet state is lowest in energy.

74

4.6.2

Gd3+ Cu2+ (S = 1/2)

Fig. 39: eff vs. T plot of a ferromagnetically coupled and an uncoupled Gd3+ Cu2+ unit as a
function of the applied field: Jex = 5 cm1 (solid lines), Jex = 0 (dotted lines); gGd = gCu = 2;
B0 = 0.01 (a), 0.1 (b), 1.0 T (c).

The solid lines in Fig. 39 refer to a heterodinuclear Gd3+ Cu2+ complex (SGd = 7/2;
SCu = 1/2) of ferromagnetically coupled spins with the exchange parameter Jex = 5 cm1 .
Solely for B0 = 0.01 T (a) the expected plateau at eff = 8.94 is obtained at T
2 K except for T 0.2 K. In order to adjust Jex by fitting procedures to corresponding
experimental data, it is justified to apply a susceptibility equation valid for B0 0 (Van
Vleck approach). For measurements with the stronger field 0.1 T (b) and all the more 1 T
(c) deviations are obvious, which are not adequately described by the simple model. They
could be misinterpreted, however, as an additional antiferromagnetic intermolecular spinspin coupling. A similar problem arises if a corresponding uncoupled system is considered
(Fig. 39, dotted lines). A straight line parallel to the T axis is obtained at eff = 8.13 for
B0 0 (curve a) while increasing deviations from this reference line to lower values are
again obvious with increasing field (b, c), falsely interpreted as an intramolecular spinspin coupling of antiferromagnetic nature. Hence, either the measurements are carried
out at very weak applied fields or the extended susceptibility formula is used as in the
given model calculations.
75

4.6.3

Ce3+ (cub)Cu2+ (S = 1/2)

After discussion of the p1 s1 case under the influence of HSO + Hex , we pass to the f 1 s1
system perturbed by HSO + HLF + Hex . To clarify the computational procedures, the lowtemperature magnetic behaviour of a system consisting of a 4f 1 ion, exchange-coupled
to a pure spin system with S = 12 , is described with restricted basis functions. After
oct
switching on HSO and HLF
acting on the Ce3+ ion, only the ground doublet 7 (E ) is
considered for spin-spin coupling with the 2 S1/2 system (represented here by Cu2+ [3d9 ]).
The procedure is outlined by three steps:

4f 1

SO

F5/2

LF

7 + 8

2 5

7 6

H oct
Hex

F5/2 + 2 F7/2

The wave functions of the Ce3+ doublet 7 are given in Tab. 32. The state is assumed to
be thermally isolated (similar to 2 P1/2 of the ps molecule (see section 4.6.1) on account
of HSO > HLF (oct) Hex ) and exchange-coupled to the S = 12 pure spin system.
Table 32: Doublet wavefunctions of 2 F5/2
perturbed by a cubic ligand field
i
|M
|MJ i

q  E q 
E
E

1  5
5 
3
=

E


6
2
6
2
q 

E q  E
E


1 
52 56  23
=
E
6 

ex (eq. (131)) with g = 6 produces the 4 4 matrix given in scheme

Application of H
J
7
(133) which is similar to scheme (132), apart from the signs of the matrix elements:

  1   1  

E ,
E ,
 E , 1
MJ MS  E , 21
2
2
2

1 
5
E , 2
Jex
42

1 
5
5
E , 2
J
21
Jex
42 ex

1 
5
5
21
E , 2 
Jex
J
42 ex

1 
5
E , 2 
42
Jex

(133)

Evaluation of the diagonal element H11 (De Gennes factor: (gJ 1) = 76 1 = 71 ):







5
2( 71 )Jex E , 21  Jz Sz  E , 12 = ( 72 ) ( 61 )( 52 )( 12 ) + ( 56 )( 23 )( 12 ) Jex = 42
Jex
Results:
15
Esinglet = 42
Jex

5
Etriplet = 42
Jex

10
E = Etriplet Esinglet = 21
Jex

Simulation calculations: For Jex > 0, the triplet state 5 is the ground state while for
Jex < 0 the singlet 2 has the lowest energy (as is predicted for the restricted basis).
76

Figs. 40 and 41 exhibit the corresponding eff vs. T plots and Figs. 42 and 43 the 1
m vs.
26)
T plots . Inspecting Fig. 40 it is remarkable that at T > 5 K the antiferromagnetic spinspin coupling yields higher eff values than the ferromagnetic spin-spin coupling. This
finding is the consequence of the fact that corresponding to Hunds third rule J is L S
and that for a spin-spin coupling of antiferromagnetic nature the spin parts of the JS
dinuclear unit cancel to a certain degree, so that the orbital part is enhanced. As opposed
to that, the spin-spin coupling of ferromagnetic nature leads to a further weakening of
the orbital part of the 4f 1 configuration. The fact that the eff values of the uncoupled
CeCu unit at T < 2 K are larger than the eff values of both coupled systems is the
consequence of the reduction of orbital parts according to the exchange splitting.
3.0

1.0
-1

Jex = 10 cm
Jex = 0

-1

2.5

Jex = -10 cm

0.8

2.0

-1

Jex = 10 cm

eff

eff

0.6

-1

Jex = -10 cm

1.5

0.4

1.0

0.2

0.0
0.0

0.5
0

10

20

30

40

50

0.5

1.5

2.0

T/K

T/K

Fig. 40: eff vs. T plot of an exchange-coupled

Ce3+ Cu2+ (S = 12 ) unit; Ce in octahedral LF
(B04 = 1 100 cm1 ).
2.5

Fig. 41: eff vs. T plot of an exchange-coupled

Ce3+ Cu2+ (S = 12 ) unit at T <1.5 K; Ce in
octahedral LF (B04 = 1 100 cm1 ).
Fig. 42: 1
m vs. T plot of the exchange3+
coupled Ce (oct)Cu2+ (S = 21 ) unit.

Jex = 10 cm-1
Jex = 0
Jex = -10 cm-1

-3

2.0

1.0

m / 10 mol m

Notice that, apart from the lowest temperatures, the spin-spin coupling produces
small parallel shifts of the 1
m T curve of
the uncoupled system (Fig. 42, black line).
The shifts are described by


1
= 1
m
m MF

1.5

-1

1.0

0.5

where (1
m ) corresponds with the shifted
curve and MF is the molecular field parameter.

0.0
0

50

100

150

200

250

300

T/K
26)

The data presented in Figs. 40 43 have been calculated with the full basis set of 14 2 = 28
functions (using the computer program CONDON).

77

Fig. 43 displays that at T < 5 K the blue curve (Jex < 0) tends to temperature
independent paramagnetism (TIP), whereas the red curve (Jex > 0) exhibits Curie paramagnetism.

1.4
-1

Jex = 10 cm
Jex = 0

1.2

-1

Jex = -10 cm

0.8

m / 10 mol m

-3

1.0

-1

0.6
0.4
0.2
0.0
0

T/K
..
.

(2)

316
ELF /cm1
7
0

..
.

3
..
.

(3)

321
ELF /cm1

(1)

(3)

Jex > 0

4/5
..
.
(3)

(1)

Jex < 0

5
2

1 vs. T plot of an exchange-coupled Ce3+ (oct)Cu2+ (S = 1 ) unit, showing the
Fig. 43: m
2
crossing of the red and blue curve around 0.2 K and the energy splitting of the two lowest states
5 and 2 due to Hex .

In Fig. 43 the black reference curve of the uncoupled system follows Curies law in
a good approximation on account of the thermally isolated states 7 (Ce) and 6 (Cu).
nd
order Zeeman
Both states show linear 1
m T behaviour (apart from a very small 2
contribution of 7 ).
4.6.4

Nd3+ (cub)Cu2+ (S = 1/2) and Nd3+ (cub)Cr3+ (S = 3/2)

Figs. 44 and 45 exhibit the eff vs. T plots of the heterodinuclear species Nd(S = 21 )
(Cu2+ ) and Nd(S = 23 ) (Cr3+ ), respectively, in cubic LF. Inspecting both Figures it
is remarkable that, similar to the Ce-Cu system (Figs. 40 43) the antiferromagnetic
spin-spin coupling leads to higher eff values than the ferromagnetic coupling on account
78

of the fact that the residual component of the spin-spin coupling strengthens the orbital
part of the lanthanides moment.
4.5
4.0
3.5

-1

Jex = 5 cm
Jex = 0

3.0

-1

Jex = -5 cm

eff

2.5
2.0
1.5
1.0
0.5
0.0
0

50

100

150

200

250

300

T/K

Fig. 44: eff vs. T diagram of an exchangecoupled Nd3+ Cu2+ unit with Nd in a cubic
LF (B04 = 2 000 cm1 ) and Cu regarded as a
pure spin system S = 21 .

Fig. 45: eff vs. T diagram of an exchangecoupled Nd3+ Cr3+ unit with Nd in a cubic
LF (B04 = 2 000 cm1 ) and Cr regarded as a
pure spin system S = 32 .

Dy3+ (cub)Cu2+ (S = 21 ),Dy3+ (cyl)Cr3+ (S = 32 ),Tb3+ (cyl)Cr3+ (S = 32 )

4.6.5
12

-1

Jex = 5 cm
Jex = 0

-1

Jex = -5 cm

eff

10

4
0

10

20

30

40

50

T/K

Fig. 46: eff vs. T diagram of an exchange- Fig. 47: DyDy(cyl):

coupled Dy3+ Cu2+ unit where Dy is in a cubic
Dy(cyl)Cr:
LF (B04 = 2 000 cm1 ) whereas Cu is regarded
Tb(cyl)Cr:
as a pure spin system S = 21 .

79

eff
k
eff
k
eff

eff

eff

eff

Gd3+ V3+

4.6.6

The 4f ion (Gd3+ , 8 S7/2 ) is regarded as a pure spin system, unaffected by ligands, zerofield splitting ignored. The electronic situation of the transition metal ion (for example
V3+ ) depends on HLF and HSO , demonstrated in Figure 48.
3

4
2

eff

mol

-1

/ 10

mol m

-3

0
0

100

200

300

T/K
Fig. 48: eff versus T and 1
m versus T diagrams showing the predicted mag3+
2
netic behaviour of V [3d ] ion in octahedral (solid lines) and orthorhombic surrounding (dashed lines); parameter values [6]: Racah parameters B = 861 cm1 ,
C = 4 165 cm1 , = 209 cm1 , B04 = 25 000 cm1 for the octahedral coordination,
B02 = 34 000, B22 = 16 000, B04 = 3 500, B24 = 1 800, B44 = 2 300 cm1 for
orthorhombic symmetry.

While the octahedral system shows between 100 K and 300 K nearly temperature independent paramagnetism, the orthorhombic one approximates to a Curie paramagnet with
only slightly varying eff . The two situations differ distinctly in the orbital contributions
to the magnetic moment. The expected magnetic properties of the V3+ Gd3+ unit is
given in Figures 49 and 50).
If V3+ is octahedrally coordinated, eff in the uncoupled situation (Jex = 0) is temperature dependent owing to the loss of magnetic activity of vanadium with decreasing
temperature (see the solid reference line in Figure 49, eff = 8.37 (300 K) 8.05 (50 K)).
When T approaches zero solely the magnetism of Gd3+ contributes. The uncoupled system
with vanadium in orthorhombic symmetry, however, exhibits a temperature independent
eff , except for T 25 K. This is the consequence of the nearly total quenching of orbital
contributions of vanadiums valence electrons owing to the low-symmetric ligand field. In
consequence, the transition metal ion behaves in a wide temperature range as a pure spin
system S = 1.
80

10

eff

7
J

ex

-1

[cm ]

10
5

[T]
0

2
0

-2
-5

-10

100

50

100

200

150

300

T / K

Fig. 49: eff versus T diagram (B0 = 0) showing the predicted magnetic behaviour of an exchange-coupled 3d2 (oct)4f 7 species where Jex =
0, 2, 5, 10 cm1 ; insert: model calculations for Jex = 2 cm1 and B0 = 0,
1, and 5 T.
10

eff

10

Jex [cm

-1

10

B0 [T]

-2

-5

-10

100

50

200

100

150

300

T/K
Fig. 50: eff versus T diagram (B0 = 0) showing the predicted magnetic behaviour of an exchange-coupled 3d2 (rhomb)4f 7 species where Jex =
0, 2, 5, 10 cm1 ; insert: model calculations for Jex = 2 cm1 and B0 = 0,
1, and 5 T.

81

Switching on the intramolecular spin-spin coupling Jex we see a distinct dependence

of the magnetic properties on the sign of Jex and the transition metals symmetry: Regarding first the cubic case with a ferromagnetic coupling (Jex > 0), eff increases, for
antiferromagnetic coupling (Jex < 0) eff decreases compared to the reference line. However, in contrast to pure spin systems, where we expect a plateau at low temperature,
the course is more complicated reflecting the reduction of vanadiums paramagnetism.
If the transition metal site has only orthorhombic symmetry significant alterations are
ascertained in the course of eff versus T : At low temperature plateaus become apparent
at eff 9.9 and 5.9 for positive and negative Jex corresponding to systems with total
spin S = 9/2 and 5/2, respectively. The decrease of eff at the lowest temperatures is
again due to small orbital contributions of vanadiums electronic system.

Problems
9. The Bleaney-Bowers formula, eq. (127), approaches for high temperature the CurieWeiss law m = C/(T ). Determine the relation between
Jex and . Is the result
P
in agreement with the relation = (2S(S + 1)/3kB ) i zi Jex ?

10. What magnetic behaviour is obtained, if in the Bleaney-Bowers formula, eq. (127),
Jex is set to 0?
11. What are the electronic (4f N ) and the LF symmetry relevant preconditions for building up a homodinuclear Ln molecule with large eff and high magnetic anisotropy?
12. Magnetochemical results are often presented as m T , 1
m T or eff T diagrams.
What type of diagram is suited for (a) Curie paramagnetism, (b) intramolecular
ferromagnetic interactions, (c) intramolekular antiferromagnetic interactions, (d)
diamagnetic behaviour, (e) TUP behaviour?
13. To reliably characterise magnetic properties measurements at different field strength
are essential. What is the reason?
14. A frequent mistake in magnetochemical investigations is the application of too strong
magnetic fields. Why may this be unfavourable?
15. To evaluate the paramagnetic part of the susceptibility of a compound with macrocyclic ligands, the problem may occur that the incremental method for the diamagnetic correction is not as precise as necessary. What is to be done?
16. For a polynuclear complex one observes at high temperature Curie-Weiss behaviour
with > 0 and in the low-temperature region field-dependent susceptibilities. What
magnetic collective effect can be expected?
17. With decreasing temperature the eff data of a homodinuclear compound increase
weakly and then, after passing a maximum at low temperature, steeply drop. (i)
What is the reason for this behaviour? (ii) What model (susceptibility expression)
should be tried to simulate the behaviour?
18. You notice that the paramagnetic properties of a homodinuclear molecular compound is not satisfactorily described with the corresponding eq. (128). What simple
extensions of the model are in principle possible?
82

Solutions
1. Verify the Russell-Saunders ground multiplets for Nd2+ and Er3+ given in Table 2.
Nd2+ [4f 4 ]: L = 6, S = 2, J = 4 = 5 I4 .
Er3+ [4f 11 ]: L = 6, S = 3/2, J = 15/2 = 5 I15/2 .
2. Evaluate the susceptibility equation for a free p1 system perturbed by spin-orbit
coupling. (Hint: Apply eq. (3); note that for the single-electron system LS .)

3




1
+ 21 32 2
2
2

3

= 23
g1/2 = 1 +
1
2

5
2 1
2 3

3
+
2
= 43
g3/2 = 1 + 2 2 3
2 5
2
2 2 2







2

k T
2
1
3
2
B
+
2

2
= 31 + 8 k9BT
1/2 = 32
2
2
3
3

2 3
5





3/2 = 43
+ 2 34 1 43 2 kBT = 20
4 k9BT
2
2
3
Z
NA 2B
, where
m = 0
3k T
N
 B




8 kB T
1
Z = 2 3 + 9 exp 8 k3BT + 4 20

3





N = 2 exp 8 k3BT + 4 exp 815

kB T
m = 0

NA 2B
3kB T

4 kB T
9




exp 815
kB T





1 8 kB T
3
40 8 kB T
exp
+
+

3
9
3
9
2 kB T



3
1 + 2 exp
2kB T

(134)

To verify eq. (134), it is recommended to develop the susceptibility equation for the
borderline case 0 (use
p the approximation exp x 1 x). For an uncoupled
1
system the relation = L(L + 1) + 4S(S + 1) B holds,
so that for the p system
(L = 1, S = 1/2) the permanent magnetic moment = 5 B and the susceptibility
equation m = 0 5NA 2B /(3kB T ) must be the results.
3. Use the spherical harmonics in Table 3 and construct the real functions dx2 y2 and
dxy with the help of eqs. (8). Verify the results with the data given in Table 4.
1
2

[n,l,ml + n,l,ml ] =

[n,l,ml

hq

15
32

sin (cos 2 + i sin 2 + cos 2 i sin 2)

sin2 cos 2 = dx2 y2

i
hq
2
15
sin

(cos
2
+
i
sin
2

cos
2
+
i
sin
2)
n,l,ml ] = i1 2
32
q
15
=
sin2 sin 2 = dxy
16
=

1
2

1
2

15
16

83

4. Calculate the matrix elements hl, ml |lq |l, ml i (where q stands for z, +, ):
2
(a) h0, 0|lz |0, 0i, (b) h2, 2|l+ |2, 1i, (c) h2, 2|l+
|2, 0i, (d) h2, 0|l+l |2, 0i.
(Hint: Apply eqs. (20))

(a) h0, 0|lz |0, 0i = 0

(b) h2, 2|l+ |2, 1i = 2 3 1 2 h
h2, 2|2, 2i = 2 h

| {z }
1
2
2

2
h2, 2|l+ |2, 1i = 2 6 h
h2, 2|2, 2i = 2 6 h

(c) h2, 2|l+

|2, 0i = 6 h
| {z }
1
p

h2, 0|l+|2, 1i = 6 + (1 + 1) h
6h
h2, 0|2, 0i = 6 h
(d) h2, 0|l+ l |2, 0i = 6 h
2
| {z }
1

5. The 14 microstates |ml ms i of an f 1 system (l = 3, s = 21 ) yield under the influence

of spin-orbit coupling 14 eigenstates |jmj i which, apart from the states | 72 27 i are
linear combinations of two microstates each. Use Table 7 to evaluate the vector
coupling coefficients for the coupled states | 52 12 i, | 25 12 i, | 27 32 i, and | 27 32 i. Control
your results with the entries of Table 8.


5 1


1 1 + b 0 1

=
a
2 2
2
2
Assignments: j = 25 , m = mj , j1 = l = 3, m2 = ms = 12 (j2 = s = 21 )
q 1 1
q 1 1
q
q
q
q
j1 +m+ 12
3+ 2 + 2
j1 m+ 12
3 2 + 2
4
a=
=
= 7
b=
=
= 37 27)
2j1 +1
7
 5 1  2j1 +1  1  7 


= a 1 2 + b  0 12
2
2q
q 1 1
q 1 1
q
q
q
j1 m+ 12
3+ 2 + 2
j1 +m+ 12
3 2 + 2
4
a=
=
= 7
b=
=
= 37
2j1 +1
7
2j1 +1
7


7 3



= a  2 12 + b  1 21
2 2
Assignments: j = 27 , m = mj , j1 = l = 3, m2 = ms = 12 (j2 = s = 21 )
q 3 1
q 3 1
q
q
q
q
j1 m+ 12
3 2 + 2
j1 +m+ 12
3+ 2 + 2
2
a=
=
= 7
b=
=
= 57
2j1 +1
7
2j1 +1
7


7 3



= a 2 21 + b 1 12
2 q
2
q 3 1
q 3 1
q
q
q
j1 +m+ 12
3 2 + 2
j1 m+ 12
3+ 2 + 2
2
a=
=
= 7
b=
=
= 57
2j1 +1
7
2j1 +1
7

6. What levels (multiplets J) may arise from the terms (a) 1 S, (b) 2 P , (c) 3 P , (d) 3 D,
(e) 4 D? How many states (distinguished by the quantum number MJ ) belong to
each level?
(a) 1 S: J = 0 (1 state)
(b) 2 P : J = 12 (2), J = 32 (4)
(c) 3 P : J = 0 (1), J = 1 (3),
(d) 3 D: J = 1 (3), J = 2 (5),
(e) 4 D: J = 21 (2), J = 32 (4),

J = 2 (5)
J = 3 (7)
J = 52 (6),

J=

7
2

(8)

7. Determine the crystal field overall splitting (CFOS) for the Dy3+ ion in Cs2 NaDyCl6
on the basis of Fig. 35. Apply eqs. (102) and (103) as well as the entries in Table 23
and take the B04 and B06 values for Dy from Table 16.
J F (4)B04
27)

8
= Wx =
135 135



15
2

1 614 cm1 = 0.7166 cm1

According to the phase convention one has to change the sign of a and b (cf. Example 2.3).

84

J F (6)B06

= W (1 |x|) =

x
= 5.401
1 |x|

4
3 864 861



3465
4

x = 0.8438,

148 cm1 = 0.1327 cm1

W = 0.8493

On account of W > 0 the ELF /W axis points upwards. Consequently, the LF

ground state at x = 0.8438 is 6 . The corresponding LF state of highest energy
(3)
(3)
is 8 . The energy separation ELF /W between 8 and 6 is 490 cm1 , so that
CFOS = ELF = 490 0.8493 cm1 = 416 cm1 , in agreement with the data given
in ref. [26], p. 2591.
8. In which respect does the magnetic susceptibility of a mononuclear lanthanide systems with Ln point symmetry (i) Oh , (ii) Dh, (iii) D2h differ?
(i) Oh : isotropic (cubic) system (x = y = z )
(ii) Dh : anisotropic (cylindrical) system (x = y 6= z )
(iii) D2h : anisotropic (orthorhombic) system (x 6= y 6= z )
9. The Bleaney-Bowers formula, eq. (127), approaches at high temperature the CurieWeiss law m = C/(T ). Determine the relation between
Jex and . Is the result
P
in agreement with the relation = [2S(S + 1)/3kB ] i zi Jex,i ?
NA 2B g 2

= 0
3kB T

1


(exp x 1 x for x 0)
1
2Jex
1 + exp
3
kB T
 
2 2
0 NA 2B g 2
0 NA 2B g 2
3
1
NA B g
= 
=



0
2Jex
4
2Jex
Jex
3kB
4

T
T 4kB
3kB T
3 3kB T
T
2kB
X
Jex
Jex
;
[2S(S + 1)/3kB ]
zi Jex,i =
for S = 12 , i = 1, zi = 1
=
2kB
2k
B
i

10. What magnetic behaviour is obtained, if in the Bleaney-Bowers formula, eq. (127),
Jex is set to 0?
NA 2B g 2

m = 0
3kB T

1
NA 2B g 2
ex =0
 J

m =
2Jex
1
3kB T
1 + exp
3
kB T

 
3
4

Curie law

11. What are the electronic (4f N ) and LF symmetrical preconditions for building up a
homodinuclear Ln molecule with large eff and high magnetic anisotropy?
(i) Ln from the second half of the Ln series (on account of free ion ground states
with J = L + S);
(ii) Ln with half-integer J (nonmagnetic LF ground states of the Ln ion are ruled
out);
(iii) Ln with large S and large L;
(iv) anisotropic LF, for example unique axis anisotropy (cylindrical, hexagonal,
tetragonal)
85

12. Magnetochemical results are often presented as m T , 1

m T or eff T plots. What
type of diagram is suited for molecular systems showing (a) Curie paramagnetism,
(b) intramolecular ferromagnetic interactions, (c) intramolekular antiferromagnetic
interactions, (d) diamagnetic behaviour, (e) TUP behaviour?
(a) 1
m T

(b) eff T

(c), (d), (e) m T

13. To reliably characterise magnetic properties measurements at different field strength

are essential. What is the reason?
To furnish proof of (i) ferromagnetic impurities, (ii) magnetic ordering effects
14. A frequent mistake in magnetochemical investigations is the application of too strong
magnetic fields. Why may this be unfavourable?
Weak magnetic ordering effects may be overlooked.
15. To evaluate the paramagnetic part of the susceptibility of a compound with macrocyclic ligands, the problem may occur that the incremental method for the diamagnetic correction is not as precise as necessary. What is to be done?
Measurement of the magnetic susceptibility of the ligand.
16. For a polynuclear complex one observes at high temperature Curie-Weiss behaviour
with > 0 and in the low-temperature region field-dependent susceptibilities. What
magnetic collective effect can be expected?
Ferromagnetism
17. With decreasing temperature the eff data of a homodinuclear compound increase
weakly and then, after passing a maximum at low temperature, steeply drop. (i)
What is the reason for this behaviour? (ii) What model (susceptibility expression)
should be tried to simulate the behaviour?
(i) Intramolecular spin-spin coupling of ferromagnetic nature, intermolecular spinspin coupling of antiferromagnetic nature.
1
(ii) Consideration of a molecular field parameter MF : 1
MF
m = (m )
18. You notice that the paramagnetic properties of a homodinuclear molecular compound is not satisfactorily described with the corresponding eq. (128). What simple
extensions of the model are in principle possible?
(i) Addition of a TIP term 0 ;
(ii) Consideration of intermolecular exchange interactions by a molecular field parameter MF .

86

Appendix 1

6.1

Magnetic quantities

Table 33: Definitions, units and conversion factors [5]

0
B
H

Quantity

SI

CGS-emu

Permeability
of vacuum
Magnetic
induction
Magnetic field
strength

0 =
4 107 H/m b
B = 0 (H + M )
T = Vs/m2
H
A/m
B0 = 0 H
T = Vs/m2
M
A/m
m = MV
Am2 = J/T
1
B = e
h/(2me )
Am2
= M /c
Am2 /kg
Mm = M M/e
Am2 /mol
a = Mm /(NA B )
Am2
1
M = H
1
g = /
m3 /kg
m = M/
m3 /mol
[3kB /0 NA 2B ]1/2 [m T ]1/2
1

B0

Magnetic field

Magnetization

Magnetic
dipole moment

m/B
B

Mm
a

Bohr magneton
specific magnetic
dipole momentd
Molar
magnetizationf
Atomic magnetic
dipole momentg

a /B

g
m
eff

Magnetic volume
susceptibility
Magnetic mass
susceptibility
Molar magnetic
susceptibility
Effective Bohr
magneton number [2]

B = H (ir) + 4M
G
Oe

87

104 T/G
103
(A m1 )/Oe
4

104 T/G

103 (A/m)/G

m = MV
G cm3
1
B = e
h/(2me )
Gcm3
= M /
Gcm3 /g
Mm = M M/
Gcm3 /mol
a = Mm /(NA B )
Gcm3
1
M = (ir) H (ir)
1
(ir)
g = (ir) /
cm3 /g
(ir)
m = (ir) M/
cm3 /mol
[3kB /0 NA 2B ]1/2 [m T ]1/2
1

103 Am2 /Gcm3

Multiplicative factor applied to the value in CGS-emu units to obtain the value in SI units.
H = Henry; H/m = Vs/(Am).
c
= mass density.
d
Specific magnetization; s = specific saturation magnetization.
e
M = molar mass.
f
s
Mm
= molar saturation magnetization.
g s
a = atomic magnetic saturation moment.
b

Factora

10

1
Am2 /Gcm3
Am2 /kg
Gcm3 /g

103 Am2 /Gcm3

1
4
4 m3 /kg
103 cm3 /g
4/106 m3 /cm3
1

 The legal SI units are not generally accepted. The CGS-emu system is still widely
used in magnetochemistry. Therefore, use magnetic quantities which are independent of the two systems, e. g., eff or 2eff instead of m T .
Do not mix the systems, e. g., use G (Gauss) instead of T (Tesla) in the CGS-emu
system.
Use B0 = 0 H in graphical representations (conversion factor 104 T/Oe).

6.2

Energy equivalents and conversion factors

Tab. 34: Energy equivalents and conversion factors

1J
1 eV
1 s1
1 cm1
1K
1T
1 kJ/mol

s1

cm1

eV

6.24151 1.50919 5.03411 7.24292 1.07828 6.02214

1.0218
10

1018
1

19

1033

2.1799

10

14

6.62607 4.13567 1
1034

1022

8.6554

10

1015

10

10

1011

10

1022

1.6045

10

1011

1023

1020

104

101

1011

1013

1.72760 9.64853

1.43877 2.14195 1.19626

10

102

1.38066 8.61739 2.08367 6.95039 1

1023

105

1010

101

24

10

1.48874 8.31451

9.27402 5.78839 1.39963 4.66864 6.71710 1

10

10

10

kJ/mol

3.33564 4.79922 7.14477 3.99031

1.98645 1.23984 2.99792 1

23

10

10

103

5.58494

103

1.66054 1.03642 2.50607 8.35933 1.20272 1.79053 1

1021

102

1012

88

101

102

102

Appendix 2: Matrix elements of operator equivalents

4 [20]
Matrix elements of the operator equivalents O
0

Tab. 35:

F (4)a)

5
2
7
2
9
2
11
2
13
2
15
2

15/2

12

32

72

92

11
2

13
2

15
2

3 13

15/2

21/2

18

15

28

15/2

108

15/2

189 129
0

12 13

63 13
2

18

105/2

15/2

84

64

3/2

105/2

5
2
7
2
9
2
11
2
13
2
15
2

31

22

18

33

27

92 132

11

33
77

143

91

14

21

273
7 8

54

96

66

99

24

39

39

251 249 704 869 429 1001

17

13

52

04 etc. are listed.

In ref. [20] the respective products F (4) O

Tab. 36:

9 11

756 621

23 101 201 221

15/2

a)

3 17

36

hMJ ||MJ i

25

4 [20]
Matrix elements of the operator equivalents O
4

F (44)

h 25 || 32 i h 72 || 21 i

3 35/2
5

3 35/2
5 3
35
p

21 5/2
5 3
5 2

3 35
35
3 105
p

3 35/2
42 5
35 6
p

3 35/2
42 15 10 231

hMJ ||MJ 4i
h 92 || 12 i

p
p
p
p
p

35/2
35/2
35/2
35/2
35/2

3 2

5 21

15 22

15 77

h 11
|| 3 i
2 2

h 13
|| 5 i h 15
|| 7 i
2 2
2 2

165

15 11

5 429

715

5005

1365

h2|| 2i h3|| 1i
h4||0i
h5||1i
h6||2i h7||3i h8||4i

15
5 7
70

35
5 42
3 70
210

70
21 10 15 14
5 66
3 55

126
70 3 5 462 15 33 5 143
1001

15 015 5 273 2 455

210 6 1155 15 154 5 1001

89

 6 [20]
Matrix elements of the operator equivalents O
0

Tab. 37:

F (6)

5
2
7
2
9
2
11
2
13
2
15
2

23

52

72

29

11
2

13
2

15
2

20

315/4
315

945/2
135
3465/4

200

11 31

11

227 11 319

143

87

59

12

36

29 48

15

200 125

20

945/4

3465/4 120

40
40

22 17

20

22

85

39 117

43

50

197

176

93

128

55

22

55 286
65

65

143

78 169 104

6 [20]
Matrix elements of the operator equivalents O
4

Tab. 38:

25 185

945/2

25

10 11

0
315/2

45

25

315/4

5
2
7
2
9
2
11
2
13
2
15
2

75

hMJ ||MJ i

12

F (64)

h 25 || 32 i

0
0

45 14/4
7 3

315 2/4
16 3

135 7/2
63

135 14/2 56 5

495 14/4 42 15

135 14/4

45 14/4

135 14/4

45 14/4

495 14/4

h2|| 2i
14

168
168

1260
252

h 72 || 21 i

3 35

6 2

105

21 6

6 231

hMJ ||MJ 4i
h 29 || 12 i

h 11
|| 3 i
2 2

h 13
|| 5 i
2 2

h 15
|| 7 i
2 2

30 2

13 21

13 22

3 77

7 165

46 11

7 429

6 715

3 5005

5 1365

h3|| 1i
h4||0i
h5||1i
h6||2i
h7||3i
h8||4i

7
2 70

13 42
12 70 15 210

35 10
8 14 21 66
28 55

546 3 6 462 147 33 126 143 45 1001

6 1155 6 154 3 1001 2 15 015 19 273 12 455

90

References
[1] A. F. Williams, A Theoretical Approach to Inorganic Chemistry, Springer, Berlin
(1979).
[2] J. H. Van Vleck, Electric and Magnetic Susceptibilities, Oxford University Press,
Oxford 1932.
[3] E. U. Condon, G. H. Shortley, The Theory of Atomic Spectra, University Press,
Cambridge 1970; p. 179.
[4] A. Abragam, B. Bleaney, Electron Paramagnetic Resonance of Transition Ions,
Clarendon Press, Oxford 1970; p. 289.
[5] (a) S. Hatscher, H. Schilder, H. Lueken, W. Urland: Practical Guide to Measurements and Interpretation of Magnetic Properties (IUPAC Technical Report), Pure
Appl. Chem. 2005, 77, 497 - 511; (b) German version: Angew. Chem. 2006, 118,
8283 - 8240.
[6] J. S. Griffith, The Theory of Transition-Metal Ions, Cambridge University Press,
Cambridge, 1971.
[7] H. Lueken, Magnetochemie, Teubner, Stuttgart, Leipzig 1999.
[8] H. L. Schlafer, G. Gliemann, Einf
uhrung in die Ligandenfeldtheorie, Akad. Verlagsges., Wiesbaden 1980; engl. edition: Basic Principles of Ligand Field Theory,
Wiley-Interscience, New York 1969.
[9] P. W. Atkins, R. S. Friedman, Molecular Quantum Mechanics, 4th ed., Oxford
University Press, Oxford 2005.
[10] H. Bethe, Termaufspaltung in Kristallen, Ann. Phys. 1929, 5, 133 - 208; further
publication of this article in: Symmetry in chemical theory, Benchmark papers in
inorganic chemistry, vol. 4, ed. J. P. Fackler, Jr., Dowden, Hutchinson, Ross, Inc.
1973 (ISBN: 0-87933-018-X).
[11] C. J. Ballhausen, Ligand Field Theory, McGraw-Hill, New York 1962.
[12] H. Schilder, H. Lueken, Computerized magnetic studies on d, f, dd, ff, and dS,
fS systems under varying ligand and magnetic fields J. Magn. Magn. Mater. 2004,
281, 17- 26.
[13] P. A. Tanner, V. V. Ravi Kanth Kumar, C. K. Jayasankar, M. F. Reid, Analysis
of spectral data and comparative energy level parametrizations for Ln3+ in cubic
elpasolite crystals, J. Alloys Comp. 1994, 215, 349 - 370.
[14] M. J. Reisfeld, G. A. Crosby, Analysis of the absorption spectrum of cesium uranium(V) hexafluoride, Inorg. Chem. 1965, 4, 65 - 70.
[15] C. Gorller-Walrand, K. Binnemans, Rationalization of Crystal-Field Parametrization, in K. A. Gschneidner, Jr., L. Eyring, Handbook on the Physics and Chemistry
of Rare Earths, Elsevier, Amsterdam 1996, vol. 23, ch. 155, p. 121 - 283.
91

[16] K. R. Lea, M. J. M. Leask, W. P. Wolf: The Raising of Angular Momentum Degeneracy of f-Electronic Terms by Cubic Crystal Fields, J. Phys. Chem. Solids, 1962,
23, 1381 - 1405.
[17] D. Smith, J. H. M. Thornley, The use of operator equivalents, Proc. Phys. Soc. 1966,
89, 779.
[18] K. W. H. Stevens, Matrix elements and operator equivalents connected with the
magnetic properties of rare earth ions, Proc. Phys. Soc. A [London] 1952, 65, 209.
[19] R. J. Elliott, K. W. H. Stevens, Proc. Roy. Soc. A [London] 1953, 218, 553.
[20] R. J. Birgeneau, Tables of matrix elements of Racah tensor operator equivalents,
Can. J. Phys. 1967, 45, 3761 - 3771.
[21] M. Hutchings, Point-charge calculations of energy levels of magnetic ions in crystalline electric fields, Solid State Phys. 1964, 16, 227.
[22] A. J. Kassman, Relationships between the coefficients of the tensor operator and
operator equivalent methods, J. Chem. Phys. 1970, 53, 4118 - 4119.
[23] R. J. Elliott, K. W. H. Stevens, The Theory of magnetic resonance experiments on
salts of the rare earths Proc. Phys. Soc. 1953, A 218, 553.
[24] H. Lueken, M. Meier, G. Klessen, W. Bronger, J. Fleischhauer, J. Less-Common
Met. 1979, 63, P35 - P44.
[25] A. Schroder, R. van den Berg, H. von Lohneysen, H. Lueken, Solid state commun.
1988, 65, 99 - 101.
[26] C. A. Morrison, R. P. Leavitt, and D. E. Wortman, J. Chem. Phys. 1980, 73, 2580
- 2598.

92