Professional Documents
Culture Documents
Crystal-Field6-15-07 Hundus Rule PDF
Crystal-Field6-15-07 Hundus Rule PDF
1
Contents
1. Introduction: a brief history on the modern magnetism
2. Fundamentals
2.1. Angular momentum and Magnetic momentum of one electron
2.2 Periodic table of iron group elements
2.3 Magnetic moment of atom
2.4 Spin-orbit interaction in an electron around the nucleus
3. Hund’s rule
3.1 Electron states in the atom
3.2 The electron configuration (3d)n (n = 1 - 10)
3.3 Spin orbit interaction of isolated atom
4. Crystal field
4.1 Overview
4.2 Series expansion of the crystal field
4.3 Mathematica program
4.4 Summary
5. p-electrons: quenching of the orbital angular momentum;
5.1. 2p-electron wave functions
5.2 2p electron in the octahedral field
5.3. Quenching of orbital angular momentum
A. Orbital angular momentum (quantum mechanics)
B. Calculation of orbital angular momentum by Mathematica
C. Theorem
5.4. Zeeman splitting of the orbital energy levels
5.5 Mathematica program: eigenvalue problem
6. 3d-electrons in the crystal field
6.1. 3d-electron wave functions
6.2. Orhorhombic crystal field
6.3. Wave function of (3d)1 electron in the orthorhombic field
6.4. Wave function of (3d)1 electron in the tetragonal field
A. Eigenvalue problem
B. Eigenvalue problem
C. Mathematica program: Eigenvalue problem
D. Quenching of orbital angular momentum
E. Mathematica program
7. The energy diagram of Cu2+ with (3d)9 electron configuration: Rule-1
8. Energy diagram of (3d)n electron configuration
8.1. Rule-2
8.2. Ground state for (3d)n electron configuration (n>1)
8.3. Excited states for (3d)2 electron configuration
8.4. Excited states for (3d)3 electron configuration
8.5. Excited states of (3d)1 electron configuration
8.6. Excited states of (3d)4 electron configuration
8.7. Energy diagram: application of rule-1 and rule-2
9. Jahn-Teller effect
2
10. Low spin and high spin states
11. Spin Hamiltonian of spin systems with the orbital singlet as a ground state
11.1. Pryce spin Hamiltonian
11.2 g-factors and the single ion anisotropy D
A. Definition
B. g-factors in Cu2+ ion in the tetragonal field (c>a = b)
11.3 Ni2+
11.4 Cr3+
12. Spin Hamiltonian for S = 1/2, 1, 3/2, 2 and 5/2
12.1 S = 1/2
12.2. S =1
A. Eigenvalue problem for S = 1
B. Magnetic susceptibility with the quenching of the spin angular momentum
C. Mathematica program: energy diagram of the spin Hamiltonian with S = 1
in the presence of magnetic field (the general case)
12.3 S = 3/2
A. Simple case with ε = 0 and δ>0
B. Kramers doublet
C. Mathematica program: energy diagram of the spin Hamiltonian with S =
3/2 in the presence of magnetic field
12.4 Energy diagram for S = 2
12.5 S = 5/2
A. Simple case for S = 5/2
B. Mathematica program for S = 5/2 (general case)
13. Spin Hamiltonian of Fe2+ and Co2+ in the trigonal crystal field
13.1 Fe2+
A. Energy diagram
B. Spin Hamiltonian with fictitious spin s = 1.
C. Mathematica program: energy diagram of Fe2+ in the trigonal field
13.2 Co2+
A. Energy diagram of Co2+
B. Spin Hamiltonian with fictitious spin s = 1/2
C. Mathematica program: energy diagram of Co2+ in the trigonal field
14. Conclusion
References
Appendix
3
developments. The first outstanding early attempt to understand magnetism at the atomic
level was provided by the semi-empirical theories of Langevin and Weiss. These theories
were able to explain experimental results on the magnetic properties of materials.
Langevin assumed that an atomic or molecular magnet carries a permanent magnetic
moment. He was quantizing the system without realizing it. If one applies classical
dynamics and statistical mechanics consistently, one finds that the diamagnetic and
paramagnetic contributions to the magnetic susceptibility exactly cancel. Thus there
should be no magnetism. The break-through in understanding of magnetic phenomena at
the atomic level occurred in 1913, when Niels Bohr introduced the significant concept of
the quantization of the orbital angular momentum, as a part of his remarkable theory of
the hydrogen spectrum. The quantization of electron orbits implied the existence of an
elementary magnetic moment, the Bohr magneton. In 1922, Stern and Gerlach
experimentally verified the quantized orbital angular momentum and hence the orbital
magnetic moment.
The advent of quantum mechanics in 1926 furnished at last the key to the quantitative
understanding of magnetism, (i) the discovery of the matrix form of quantum mechanics
by Heisenberg and Born, (ii) the alternative but equivalent wave mechanical form by de
Broglie and Schrödinger, and (iii) the introduction of electron spin by Uhlenbeck and
Goulsmit. A quantum mechanics without spin and the Pauli’s exclusion principle would
not have been able to understand even the structure of the periodic table or most magnetic
phenomena. Originally spin was a sort of the appendage to the mathematical framework,
but in 1928, Dirac synthesized everything in his remarkable four first order simultaneous
equations which is relativistically invariant under the Lorentz transformation. The
electron spin and the factor of two came naturally out of the calculation. In 1928,
Heisenberg has shown how the previously obscure Weiss molecular field could be
attributed to a quantum mechanical exchange effect, arising from the Pauli’s exclusion
principle that no two electrons occupy the same state. The forces of interaction between
neighboring atoms give rise to a exchange coupling between unpaired spinning electron.
This leads to a scalar isotropic interaction of two spins with an exchange interaction
constant (see the book written by Hoddeson et al.7, “Out of the Crystal Maze” for more
detail of the above review).
In the early 1930’s there appeared two major textbooks devoted to the topics of
magnetism, Van Vleck’s Theory of Electric and Magnetic Susceptibilities (1932)8 and
Stoner’s Magnetism and Matter (1934).9 These are considered to be the best classic texts
in modern magnetism.
John H. Van Vleck (born March 13, 1899, Middletown, Connecticut, died Oct. 27, 1980,
Cambridge, Massachusetts.) American physicist and mathematician who shared the
Nobel Prize for Physics in 1977 with Philip W. Anderson and Nevill F. Mott. The prize
honored Van Vleck's contributions to the understanding of the behavior of electrons in
magnetic, non-crystalline solid materials. Educated at the University of Wisconsin,
Madison, and at Harvard University, where he received his Ph.D. in 1922, He developed
during the early 1930s the first fully articulated quantum mechanical theory of magnetism.
Later he was a chief architect of the ligand field theory of molecular bonding.
((Note))
4
In this note we use the following notations for the 3d orbitals.
dε = t2g = d(xy), d(yz), and d(zx).
dγ = eg = d(x2 – y2) and d(3z2-r2).
2. Fundamentals
A. Angular momentum and magnetic momentum of one electron
Fig.1 Orbital (circular) motion of electron with mass m and a charge –e. The
direction of orbital angular momentum L is perpendicular to the plane of the motion
(x-y plane).
5
Fig.2 Acceptable wave on the ring (circular orbit). The circumference should be
equal to the integer n (=1, 2, 3,…) times the de Broglie wavelength λ. The picture of
fitting the de Broglie waves onto a circle makes clear the reason why the orbital
angular momentum is quantized. Only integral numbers of wavelengths can be fitted.
Otherwise, there would be destructive interference between waves on successive
cycles of the ring.
6
2μ B ˆ
μˆ = − S, (2.8)
h
h
using operators (Dirac). When Sˆ = σˆ , we have μˆ = − μ Bσˆ . The spin angular
2
momentum is described by the Pauli matrices (operators)
h h h
Sˆ x = σˆ x , Sˆ y = σˆ y Sˆ z = σ̂ z . (2.9)
2 2 2
Using the basis,
⎛1⎞ ⎛ 0⎞
+ = ⎜⎜ ⎟⎟ , − = ⎜⎜ ⎟⎟ . (2.10)
⎝ 0⎠ ⎝1⎠
we have
⎛0 1⎞ ⎛0 − i⎞ ⎛1 0 ⎞
σˆ x = ⎜⎜ ⎟⎟ , σˆ y = ⎜⎜ ⎟⎟ , σˆ z = ⎜⎜ ⎟⎟ . (2.11)
⎝1 0⎠ ⎝i 0 ⎠ ⎝ 0 − 1⎠
The commutation relations are valid;
[σˆ x , σˆ y ] = 2iσˆ z , [σˆ y , σˆ z ] = 2iσˆ x , [σˆ z , σˆ x ] = 2iσˆ y . (2.12)
The resultant magnetic moment of an electron is given by
μB
μ=− (L + 2S) . (2.13)
h
l=0
n2
There are two values for ms (= ±1/2).
Thus, corresponding to any value of n, there are 2n2 states.
K shell
n l m s ms
1 0 0 1/2 ±1/2 (1s)2
L shell
n l m s ms
2 0 0 1/2 ±1/2 (2s)2
2 1 1,0,-1 1/2 ±1/2 (2p)6
M shell
n l m s ms
3 0 0 1/2 ±1/2 (3s)2
3 1 1,0,-1 1/2 ±1/2 (3p)6
3 2 2,1,0,-1,-2 1/2 ±1/2 (3d)10
N shell
7
n l m s ms
4 0 0 1/2 ±1/2 (4s)2
4 1 1,0,-1 1/2 ±1/2 (4p)6
4 2 2,1,0,-1,-2 1/2 ±1/2 (4d)10
4 3 3,2,1,0,-1,-2,-3 1/2 ±1/2 (4f)14
2 2 6 2 6 10 2 6 10 14 2 6 10
(1s) |(2s) (2p) |(3s) (3p) (3d) |(4s) (4p) (4d) (4f) |(5s) (5p) ((5d) ….
Iron-group elements:
Ti3+, V4+ (1s)2|(2s)2(2p)6|(3s)2(3p)6|(3d)1
3+
V (1s)2|(2s)2(2p)6|(3s)2(3p)6|(3d)2
Cr3+, V2+ (1s)2|(2s)2(2p)6|(3s)2(3p)6|(3d)3
2+ 3+
Cr , Mn (1s)2|(2s)2(2p)6|(3s)2(3p)6|(3d)4
Mn2+, Fe3+ (1s)2|(2s)2(2p)6|(3s)2(3p)6|(3d)5
2+
Fe (1s)2|(2s)2(2p)6|(3s)2(3p)6|(3d)6
Co2+ (1s)2|(2s)2(2p)6|(3s)2(3p)6|(3d)7
2+
Ni (1s)2|(2s)2(2p)6|(3s)2(3p)6|(3d)8
Cu2+ (1s)2|(2s)2(2p)6|(3s)2(3p)6|(3d)9
Atoms with filled n shells have a total angular momentum and a total spin of zero.
Electrons exterior these closed shells are called valence electrons.
8
Fig.3 Basic classical vector model of orbital angular momentum (L), spin angular
momentum (S), orbital magnetic moment (μL), and spin magnetic moment (μS). J (=
L + S) is the total angular momentum. μJ is the component of the total magnetic
moment (μL + μS) along the direction (-J).
Suppose that
L = aJ + L ⊥ and S = bJ + S ⊥ , (2.18)
where a and b are constants, and the vectors S ⊥ and L ⊥ are perpendicular to J.
Here we have the relation a + b = 1 , and L ⊥ + S ⊥ = 0 . The values of a and b are
determined as follows.
J ⋅L J ⋅S
a= 2 , b= 2 . (2.19)
J J
Here we note that
J 2 − L2 − S 2 J 2 − L2 + S 2
J ⋅ S = (L + S ) ⋅ S = S 2 + L ⋅ S = S 2 + = , (2.20)
2 2
or
J 2 − L2 + S 2 h 2
J ⋅S = = [ J ( J + 1) − L( L + 1) + S ( S + 1)] , (2.21)
2 2
using the average in quantum mechanics. The total magnetic moment μ is
μB μB
μ=− (L + 2S) = − [(a + 2b)J + ( L⊥ + 2S ⊥ )] . (2.22)
h h
Thus we have
μB μB g J μB
μJ = − (a + 2b)J = − (1 + b)J = − J, (2.23)
h h h
9
with
J ⋅ S 3 S ( S + 1) − L( L + 1)
gJ = 1 + b = 1 + = + . (2.24)
J2 2 2 J ( J + 1)
((Note))
The spin component is given by
S = bJ + S ⊥ = ( g J − 1)J + S ⊥ , (2.25)
with b = g J − 1 ,
The de Gennes factor is defined by
( g J − 1) 2 J 2
2
= ( g J − 1) 2 J ( J + 1) . (2.26)
h
In ions with strong spin-orbit coupling the spin is not a good quantum number, but rather
the total angular momentum , J = L+S. The spin operator is described by
S = ( g J − 1)J . (2.27)
Fig.4 Simple model for the spin-orbit interaction. The orbital current due to the
circular motion of the nucleus (with velocity vN and charge Ze) produces an magnetic
field at the center where the electron is located.
The current I due to the movement of nucleus (charge Ze, e>0) is given by
Idl = Zev N , (2.28)
dl
where v N is the velocity of the nucleus and = v N . Note that
dt
Δq dl
Idl = dl = Δq = Zev N . (2.29)
Δt dt
The effective magnetic field at the electron at the origin is
I dl × r1
B eff = , v N = veθ , (2.30)
c r13
where v is the velocity of the electron. Then we have
10
Ze v N × r1 Ze veθ × r1
B eff = = . (2.31)
c r13 c r1
3
or
Ze veθ × r Zev eθ × r Zev e z Zev e z Zemv
B eff = − = = = = ez . (2.33)
c r3 c r2 c r2 c r 2 mcr 2
The Coulomb potential energy is given by
Ze 2 dVc (r ) Ze 2
Vc ( r) = − , = 2 . (2.34)
r dr r
Thus we have
Zemrv Ze 2 mrv Ze 2 m0 rv
B eff = e z = e z = ez
mcr 3 mcer 3 m cerr 2
. (2.35)
1 Ze 2 1 dVc (r )
= m0 rve z = Lz e z
mcer r 2 mcer dr
or
1 1 dVc (r )
B eff = Lz e z , (2.36)
mce r dr
where Lz is the z-component of the orbital angular momentum, Lz = mvr .
The spin magnetic moment is given by
2μ
μs = − B S . (2.37)
h
The Zeeman energy is given by
1 1 ⎛ 2μ ⎞ ⎛ 1 1 dVc (r ) ⎞
H LS = − μ s ⋅ B eff = − ⎜ − B S ⎟ ⋅ ⎜ L⎟
2 2 ⎝ h ⎠ ⎝ mce r dr ⎠
1 1 dVc (r )
= S ⋅ L = ξ (S ⋅ L) . (2.38)
2m 2c 2 r dr
((Thomas correction))
Thomas factor 1/2, which represents an additional relativistic effect due to the
acceleration of the electron. The electron spin, magnetic moment, and spin-orbit
interaction can be derived directly from the Dirac relativistic electron theory. The
Thomas factor is built in the expression.
H LS = ξS ⋅ L , (2.39)
with
2
1 1 dVc (r ) 1 ⎛ e ⎞ 1
ξ= 2 2
= Z⎜ ⎟ . (2.40)
2m c r dr 2 ⎝ mc ⎠ r3 av
When we use the formula
11
Z3
r −3 = 3
, (2.41)
n 4 a0 l (l + 1 / 2)(l + 1)
the spin-orbit interaction constant ξ is described by
e2Z 4 me8 Z 4
ξ= = , (2.42)
2m 2c 2 n 4 a0 l (l + 1 / 2)(l + 1) 2c 2 n 4 h6l (l + 1 / 2)(l + 1)
3
where
h2
a0 = = 0.52917720859 Å. (2.43)
me 2
(Bohr radius) (from NIST physics constants)
The energy level (negative) is given by
Z 2 me 4 Z 2e 2
| En |= 2 = . (2.44)
n 2h 2 2n 2 a0
The ratio h 2ξ / En is
h 2ξ e4 Z 2 (αZ ) 2 1
= 2 2 2 = , (2.45)
En c n h l (l + 1 / 2)(l + 1) n l (l + 1 / 2)(l + 1)
2
with
e2 1
α= = . (2.46)
hc 137.037
((Note)) For l = 0 the spin-orbit interaction vanishes and therefore ξ = 0 in this case.
3. Hund’s rule
3.1. Electron states in the atom
For a given l, the number m takes 2l +1 values. The number s is restricted to only two
values ±1/2. Hence there are altogether 2(2l+1) different states with the same n and l.
There states are said to be equivalent.
According to Pauli’s principle, there can be only one electron in each such state. Thus
at most 2(2l+1) electrons in an atom can simultaneously have the same n and l.
Hund’s rule is known concerning the relative position of levels with the same
configuration but different L and S.
Hund’first law
(1) The maximum values of the total spin S allowed by the exclusion principle.
Hund’s second law
(2) The maximum values of the total orbital angular momentum L consistent with this
value of S.
Huns’s third law
(i) J = L − S for less than half full (spin-orbit interaction, the discussion will be
made later)
(ii) J = L + S for more than half full (spin-orbit interaction).
12
A d shell corresponds to l = 2, with five values of ml. Multiplying this by 2 for the
spin states gives a total of 10. Then the configuration (3d)10 represents a full shell. It is
non-degenerate, and the state is 1S0. This is a general rule for a full shell. It follows
because each of electrons must have a different pair of ml and ms values.
(3d)2: V3+
3
F2
L =3, S = 1, J = 2,
2
Fig.5(b) Hund’s law for the (3d) electron configuration.
L = 2, S = 2, J = 0
4
Fig.5(d) Hund’s law for the (3d) electron configuration.
13
L = 0, S = 5/2, J = 5/2
Fig.5(e) Hund’s law for the (3d)5 electron configuration.
(3d)6: Fe2+
5
D4
L = 2, S = 2, J = 4
Fig.5(f) Hund’s law for the (3d)6 electron configuration.
(3d)7: Co2+
4
F9/2
L = 3, S = 3/2, J = 9/2
Fig.5(g) Hund’s law for the (3d)7 electron configuration.
(3d)8: Ni2+
3
F4
L = 3, S = 1, J = 4
8
Fig.5(h) Hund’s law for the (3d) electron configuration.
(3d)9: Cu2+
2
D5/2
L = 2, S = 1/2, J = 5/2
Fig.5(i) Hund’s law for the (3d)9 electron configuration.
14
(3d)9
This configuration represents a set of electrons one short of a full shell. Since a full shell
has zero angular momentum (both orbital and spin), it follows that if one electron is
removed from a full shell, the spin angular momentum of the remainder are minus those
of the one that was removed. So the L, S, and J values of remainder are the same as if
there were only one electron in the shell.
(3d)10
A d shell corresponds to l = 2, with five values of ml. Multiplying this by two for the spin
states gives 10. Thus the configuration (3d)10 represents a full shell. L = 0. S = 0. J = 0.
L = 3, S = 3/2 L = 3, S = 3/2
15
1 1 n
where S = (2l + 1) − [n − (2l + 1] = (2l + 1) − . This condition favors J = L+S.
2 2 2
The third Hund’ rule is a consequence of the sign of the spin-orbit interaction.
4. Crystal field
4.1 Overview
Rare-earth: The 4f shell responsible for paramagnetism lies deep inside the ions,
within the 5s and 5p shells. J is a good quantum number. (L-S coupling>> crystal field).
Iron group: The 3d shell responsible for paramagnetism is the outermost shell. The 3d
shell experiences the intense inhomogeneous electric field produced by neighboring ions.
The inhomogeneous electric field is called the crystal field. (Crystal field>>L-S coupling).
Two major effect of the crystal field
(i) The coupling of L and S is largely broken up. So that the states are no longer specified
by their J values.
(2L+1)(2S+1) degeneracy → orbital splitting due to the crystal field (degeneracy
2S+1)→L-S coupling
(ii) 2L+1 sublevels belonging to a given L, which are degenerate in the free ion may now
be split by the crystal field. The quenching of the orbital angular momentum.
(2J+1) degeneracy (Hund’s rules)→L-S coupling → crystal field.
We note that there are many excellent textbooks,10,14 review articles,15 and original
papers16,17 on the ligand field theory, including Griffiths (The Theory of Transition-Metal
Ions),10 Abraham and Bleaney (Electron Paramagnetic Resonance of Transition Ions),11
Sugano et al. (Multiplets of Transition-Metal Ions in Crystals),12 Inui et al. (Group
Theory and Its Applications in Physics),13 and Ballhausen (Introduction to Ligand Field
Theory).14
16
Fig.7 crystal structure of octahedral and tetragonal structures. A magnetic ion (red
solid circle) with one electron (a charge –e) is located at the origin. Six ions (open
circles, each ion has a –Ze charges) are located on the x-, y-, and z-axes at the
coordinates (± a, 0, 0), (0, ± b, 0), and (0, 0, ± c), where a, b, and c are distances. a =
b = c for the octahedral structure. a = b ≠ c for the tetragonal structure.
We consider an electrostatic potential energy of an electron (a charge –e, typically 2p, 3d,
or 4f electrons) of the M ion, due to the Coulomb field (ligand), defined by
⎡ 1 1 1
V (r ) = Ze 2 ⎢ + +
⎢⎣ ( x − a) + y + z
2 2 2
( x + a) + y + z
2 2 2
x + ( y − b) 2 + z 2
2
1 1 1 ⎤
+ + + ⎥
x 2 + ( y + b) 2 + z 2 x 2 + y 2 + ( z − c) 2 x 2 + y 2 + ( z + c) 2 ⎥⎦
(4.1)
where a = b = c for the orthorhombic (cubic) field and a = b (≠ c) for the tetragonal field.
We use the Mathematica to expand V(r) around the origin in terms of the powers xpyqzr
(the maximum of p + q + r = n; n = 2 for the 2p electrons, n = 4 for the 3d electrons, and
n = 6 for the 4f electrons).
è!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
! è!!!!!!!!!!!!!!!!!!!!!!!!!!!!! è!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!
1 1 1
H− a + xL2 + y2 + z2 Ha + xL2 + y2 + z2 x2 + H− b + yL2 + z2
+ + +
h[α_,β_]:=If[0≤α≤β,1,0];
General case (the highest order=4)
eq1 =
Series@ Vc, 8x, 0, 4<, 8y, 0, 4<, 8z, 0, 4<D êê
FullSimplify@#, 8a > 0, b > 0, c > 0<D & êê Normal êê Expand;
eq12 = SumAxp yq zr CoefficientAx y z eq1, xp+1 yq+1 zr+1 E h@ p + q + r, 4D ,
8p, 0, 4<, 8q, 0, 4<, 8r, 0, 4<E
J N x2 + J Nx +
2 2 2 1 4 2 2 1 48 18 18 4
+ + + − − + +
a b c 2 a 3 b 3 c 3 24 a5 b5 c5
J− N y2 + J− N x2 y2 +
1 2 4 2 1 24 24 6
+ − − +
2 a 3 b3 c 3 4 a 5 b5 c5
J N y4 + J− N z2 + J− N x2 z2 +
1 18 48 18 1 2 2 4 1 24 6 24
+ + − + + −
24 a 5 b 5 c5 2 a3 b 3 c3 4 a5 b5 c5
J Ny z + J Nz
1 6 24 24 2 2 1 18 18 48 4
− − + +
4 a 5 b 5 c5 24 a5 b5 c5
The crystal field of the d electrons
17
The case of a = b = c (orthorhombic (cubic))
eq2=eq12/.{b→a,c→a}//Expand
6 7 x4 21 x2 y2 7 y4 21 x2 z2 21 y2 z2 7 z4
+ − + − − +
a 2 a5 2 a5 2 a5 2 a5 2 a5 2 a5
This can be written as
Jx4 + y4 + z4 − Ix + y + z M N êê Expand
6 35 3 2 2 2 2
eq3 = +
5
a 4a 5
6 7 x4 21 x2 y2 7 y4 21 x2 z2 21 y2 z2 7 z4
+ − + − − +
a 2 a5 2 a5 2 a5 2 a5 2 a5 2 a5
eq2-eq3//Simplify
0
The case of a = b but not equal to c (tetragonal)
eq4=eq12/.{b→a}//Expand
4 2 x2 x2 11 x4 3 x4 y2 y2 12 x2 y2 3 x2 y2 11 y4
+ + − + + + − − + + +
a c a3 c3 4 a5 4 c5 a3 c3 a5 2 c5 4 a5
3 y4 2 z2 2 z2 9 x2 z2 6 x2 z2 9 y2 z2 6 y2 z2 3 z4 2 z4
− + − − − − + +
4 c5 a3 c3 2 a5 c5 2 a5 c5 2 a5 c5
This can be written as
+ +J N Ix2 + y2 − 2 z2M + J N Ix4 + y4 + z4M −
4 2 1 1 20 15
eq5 = − +
a c a3 c3 4 a5 4 c5
J N Iz4 + 6 x2 y2M − J N Ix2 + y2 + z2M êê Expand
5 5 9 12 2
− +
4 a5 4 c5 4 a5 4 c5
4 2 x2 x2 11 x4 3 x4 y2 y2 12 x2 y2 3 x2 y2 11 y4
+ + − + + + − − + + +
a c a3 c3 4 a5 4 c5 a3 c3 a5 2 c5 4 a5
3 y4 2 z2 2 z2 9 x2 z2 6 x2 z2 9 y2 z2 6 y2 z2 3 z4 2 z4
− + − − − − + +
4 c5 a3 c3 2 a5 c5 2 a5 c5 2 a5 c5
eq4-eq5//Simplify
0
Crystal field for the p electrobns (the highest order = 2)
a, b and a are different.
eq13 = SumAxp yq zr CoefficientAx y z eq1, xp+1 yq+1 zr+1 E h@ p + q + r, 2D ,
8 p, 0, 4<, 8q, 0, 4<, 8r, 0, 4<E
J N x2 +
2 2 2 1 4 2 2
+ + + − −
a b c 2 a 3 b3 c3
J− N y2 + J− N z2
1 2 4 2 1 2 2 4
+ − − +
C1 = CoefficientAeq13, z2E;
A1+B1+C1//Simplify
0
Crystal field for the f electrobns (the highest order = 6)
eq6 =
Series@ Vc, 8x, 0, 6<, 8y, 0, 6<, 8z, 0, 6<D êê
FullSimplify@#, 8a > 0, b > 0, c > 0<D & êê Normal êê Expand;
eq61 = SumAxp yq zr CoefficientAx y z eq6, xp+1 yq+1 zr+1 E h@ p + q + r, 6D ,
8 p, 0, 6<, 8q, 0, 6<, 8r, 0, 6<E
18
J N x2 +
2 2 2 1 4 2 2
+ + + − −
a b c 2 a3 b3 c3
J N x4 + J Nx +
1 48 18 18 1 1440 450 450 6
+ + − −
24 a5 b5 c5 720 a7 b7 c7
J− N y2 + J− N x2 y2 +
1 2 4 2 1 24 24 6
+ − − +
2 a 3 b 3 c 3 4 a5 b 5 c5
J− Nx y + J Ny +
1 720 540 90 4 2 1 18 48 18 4
+ − + +
48 a 7 b 7 c 7 24 a5 b5 c5
J N x2 y4 + J− Ny +
1 540 720 90 1 450 1440 450 6
− − + −
48 a 7 b 7 c 7 720 a7 b7 c7
J− N z2 + J− N x2 z2 +
1 2 2 4 1 24 6 24
− + + −
2 a 3 b 3 c 3 4 a5 b 5 c5
J− Nx z + J N y2 z2 +
1 720 90 540 4 2 1 6 24 24
− + − −
48 a 7 b 7 c 7 4 a 5 b5 c5
J N x2 y2 z2 + J− N y4 z2 +
1 180 180 180 1 90 720 540
+ + − +
8 a7 b 7 c 7 48 a7 b7 c7
J N z4 + J Nx z +
1 18 18 48 1 540 90 720 2 4
+ + − −
24 a 5 b 5 c5 48 a7 b7 c7
J− Ny z + J− N z6
1 90 540 720 2 4 1 450 450 1440
+ − − +
48 a 7 b 7 c 7 720 a7 b7 c7
Octahedral case (a=b=c)
eq7=eq61/.{b→a,c→a}
6 7 x4 3 x6 21 x2 y2 45 x4 y2 7 y4 45 x2 y4 3 y6 21 x2 z2
+ + − − + − + − −
a 2 a5 4 a7 2 a5 8 a7 2 a5 8 a7 4 a7 2 a5
45 x4 z2 21 y2 z2 135 x2 y2 z2 45 y4 z2 7 z4 45 x2 z4 45 y2 z4 3 z6
− + − + − − +
8 a7 2 a5 2 a7 8 a7 2 a5 8 a7 8 a7 4 a7
4.4 Summary
The summary of the result (Mathematica) is as follows for the 3d electrons (n = 4).
(i) Potential in the case when a, b and c are different (general case)
2 2 2
V (r ) = Ze 2 [ + +
a b c
1 4 2 2 1 2 4 2 1 2 2 4
+ ( 3 − 3 − 3 ) x 2 + (− 3 + 3 − 3 ) y 2 + (− 3 − 3 + 3 ) z 2
2 a b c 2 a b c 2 a b c
1 8 3 3 4 1 3 8 3 4 1 3 3 8 4
+ ( 5 + 5 + 5 )x + ( 5 + 5 + 5 ) y + ( 5 + 5 + 5 )z
4 a b c 4 a b c 4 a b c
3 4 4 1 3 1 4 4 3 4 1 4
+ (− 5 − 5 + 5 ) x 2 y 2 + ( 5 − 5 − 5 ) y 2 z 2 + + (− 5 + 5 − 5 ) z 2 x 2 ]
2 a b c 2 a b c 2 a b c
(4.2)
(ii) Potential in the case of a = b, but c being different from a and b (tetragonal field)
4 2 1 1 5 4 3
V (r ) = Ze 2 [ + + ( 3 − 3 )( x 2 + y 2 − 2 z 2 ) + ( 5 + 5 )( x 4 + y 4 + z 4 )
a c a c 4 a c
5 1 1 3 3 4
− ( 5 − 5 )( z 4 + 6 x 2 y 2 ) − ( 5 + 5 )r 4 ] . (4.3)
4 a c 4 a c
where r 2 = x 2 + y 2 + z 2 .
(iii) Potential in the case of a = b = c (orthorhombic field)
19
6 7 21
V (r ) = Ze 2 [ + 5 ( x 4 + y 4 + z 4 ) − 5 ( x 2 y 2 + y 2 z 2 + z 2 x 2 )] , (4.4)
a 2a 2a
or
6 35 3
V (r ) = Ze 2 [ + 5 ( x 4 + y 4 + z 4 − r 4 )] . (4.5)
a 4a 5
83<
-0.05
82<
0.05 00.05
0
-0.05
5
0.2
81<
0.1 0.2
0
-0.1
-0.2
2 0
0.05 0.05
0 0
-0.05
- -0.05 -0.2
-0.2
-0.1 0.05
0 0 -0.05
0.05
0.1 -0.05 0
0.05
0 05
0.2
0 2
Fig.8 Angular parts of the wavefunctions for (1) 2px, (2) 2py, and (3) 2pz
n = 2, l = 1 (2p electron)
r
r exp(− )
2a0
R21 (r ) = 5/2
, (5.3)
2 6a0
20
1 3
Y11 (θ ,φ ) = − eiφ sin θ
2 2π
1 3
Y10 (θ ,φ ) = cosθ , (5.4)
2 π
1 − iφ 3
Y1−1 (θ ,φ ) = e sin θ
2 2π
1 1 3 1 3 R21 (r )
r px = [−Y11 (θ , φ ) + Y1−1 (θ , φ )]R21 (r ) = sin θ cos φR21 (r ) = x
2 2 π 2 π r
1 1 1 3 1 3 R21 (r )
r py = i [Y1 (θ , φ ) + Y1−1 (θ , φ )]R21 (r ) = sin θ sin φR21 (r ) = y ,
2 2 π 2 π r
1 3 1 3 R21 (r )
r p z = Y10 (θ , φ ) R21 (r ) = cos θ = z
2 π 2 π r
(5.5)
with the radial wave function given by
r
exp(− )
1 3 R21 (r ) 1 2a0
f (r ) = = . (5.6)
2 π r 4 2π a0
5/ 2
The energy is split because of the crystal field due to negative ions. W3 > W2 > W1 ,
because of Coulomb repulsion. Since c<a<b, one can find the longest Coulomb repulsion
between negative ions and electron, for the z-axis, for y-axis, and for the x axis, in order.
W2 = p y Vˆ p y = −24a0 πB
2
. (5.9)
W3 = pz Vˆ pz = 24a0 π ( A + B)
2
21
Note
p x Vˆ p y = 0, p x Vˆ p z = 0, p y Vˆ p z = 0, ... , (5.10)
The perturbation theory (quantum mechanics) can be applied to the ground states which
are orbital-triplet states. As a result of the perturbation, these states are split into non-
degenerate orbital-singlet states.
We assume that the new state is given by ψ with the energy eigenvalue E.
Eigenvalue problem
⎛ px Vˆ p x − E 0 0 ⎞⎛ p x ψ ⎞ ⎛ 0 ⎞
⎜ ⎟⎜ ⎟ ⎜ ⎟
⎜ 0 p y Vˆ p y − E 0 ⎟⎜ p y ψ ⎟ = ⎜ 0 ⎟ , (5.11)
⎜⎜ ⎟⎜ ⎟
⎝ 0 0 p z Vˆ p z − E ⎟⎠⎝ p z ψ ⎠ ⎜⎝ 0 ⎟⎠
The value of E is obtained from
W1 − E 0 0
0 W2 − E 0 = 0. (5.12)
0 0 W3 − E
As we expect, the degenerate ground state is separated into three orbital-singlet states;
eigenstate px with the energy W1, eigenstate p y with the energy W2, and eigenstate
pz with the energy W3.
Fig.9 Schematic diagram of the energy splitting of 2p orbital under the crystal field.
Lˆ x px = 0 , Lˆ x p y = ih p z , , Lˆ x pz = −ih p y , . (5.14)
Lˆ y px = −ih pz , Lˆ y p y = 0, Lˆ y pz = ih px
Then we have
22
p x Lˆ z p x = p y Lˆ z p y = p z Lˆ z p z = 0
p x Lˆ x p x = p y Lˆ x p y = p z Lˆ x p z = 0 . (5.15)
p x Lˆ y p x = p y Lˆ y p y = pz Lˆ y p z = 0
p y Lˆ z p x = ih, p x Lˆ z p y = −ih,
p z Lˆ x p y = ih, p y Lˆ x p z = −ih, . (5.16)
p x Lˆ y p z = ih, p z Lˆ y p x = −ih.
Note that
( Lˆ x + Lˆ y + Lˆ z ) px = −ihLˆ y pz + ihLˆ z p y = 2h 2 px
2 2 2
( Lˆ x + Lˆ y + Lˆ z ) p y = 2h 2 p y
2 2 2
. (5.17)
( Lˆ x + Lˆ y + Lˆ z ) pz = 2h 2 pz
2 2 2
The orbital angular momentum is zero (quenching of the orbital angular momentum).
"#################### "####################
{Lz[ψx],Lz[ψy],Lz[ψz]}//Simplify
: y — fB x2 + y2 + z2 F, − x — fB x2 + y2 + z2 F, 0>
"###################
# "####################
{Lx[ψx],Lx[ψy],Lx[ψz]}//Simplify
:0, z — fB x2 + y2 + z2 F, − y — fB x2 + y2 + z2 F>
"#################### "###################
#
{Ly[ψx],Ly[ψy],Ly[ψz]}//Simplify
:− z — fB x2 + y2 + z2 F, 0, x — fB x2 + y2 + z2 F>
C. Theorem
If the state ψ is not degenerate, then the wave function r φ should be real.
Suppose that
r ψ = ψ (r ) = f (r ) + ig (r ) , (5.18)
where f and g are real.
Hψ (r ) = H [ f (r ) + ig (r )] = E[ f (r ) + ig (r )] , (5.19)
or
Hf (r ) = Ef (r )
. (5.20)
Hg (r ) = Eg (r )]
23
In other words, f(r) and g(r) are the eigenfunctions. This is inconsistent with the above
assumption. Then ψ (r ) is real.
ψ * (r ) = ψ (r ) . (5.21)
We consider the expectation
ψ Lˆ ψ = ∫ dr r ψ L r ψ = ∫ drψ * (r )Lψ (r ) .
*
(5.22)
h
L = r×p = r×∇
i
. (5.23)
h
L = − r×∇ =
*
i
ψ L ψ = ∫ drψ (r )L*ψ (r ) = − ∫ dr[ψ (r )Lψ (r )] = − ψ Lˆ ψ .
ˆ *
(5.24)
Since L is a physical quantity (observable), we have
ψ Lˆ ψ = ψ Lˆ ψ .
*
(5.25)
Then we have
ψ Lˆ ψ = 0 (the quenching of L). (5.26)
For the orbital singlet
ψ Lˆ ψ = 0 . (5.27)
μB H ˆ . (5.30)
= − Axˆ 2 − Byˆ 2 + ( A + B) zˆ 2 + Lz
h
We assume that the new state is given by ψ with the energy eigenvalue E.
Eigenvalue problem
⎛ μB H ⎞
⎜ p x Vˆ p x − E p x Lˆ z p y 0 ⎟
⎜ h ⎟⎛⎜ px ψ ⎞⎟ ⎛⎜ 0 ⎞⎟
⎜ μ B H p Lˆ p p y Vˆ p y − E 0 ⎟⎜ p ψ ⎟ = ⎜ 0 ⎟ ,
⎜ h y z x
⎟⎜ y ⎟ ⎜ ⎟
⎜ 0 0 p z Vˆ p z − E ⎟⎝ pz ψ ⎠ ⎝ 0 ⎠
⎜ ⎟
⎝ ⎠
(5.31)
or
24
⎛W1 − E − iμ B H 0 ⎞⎛ px ψ ⎞ ⎛ 0 ⎞
⎜ ⎟⎜ ⎟ ⎜ ⎟
⎜ iμ B H W2 − E 0 ⎟⎜ p y ψ ⎟ = ⎜ 0 ⎟ . (5.32)
⎜ 0
⎝ 0 W3 − E ⎟⎠⎜⎝ pz ψ ⎟⎠ ⎜⎝ 0 ⎟⎠
Effect of the crystal field on the energy level of the 2p electrons (n = 2, l = 1) wave
functions
rwave@n_, __, r_D :=
è!!!!!!!!!!!!!!!
1
Hn + _L !
i
j è!!!!!!!!!!!!!!!!!!!!!
j21+_ a0 2 Hn − _ − 1L !
−_− 3 − r
k
a0 n n−_−2 r_
2r y
LaguerreLA−1 + n − _, 1 + 2 _, Ez
z;
a0 n {
rwave@2, 1, rD
− r
2 è!!!
6! a05ê2
2a0 r
ψpx =
rwave@2, 1, rD
i
j
j
j $%%%%%
1%
SphericalHarmonicY@1, 1, θ, φD +
j−
k 2
y
z
$%%%%%% SphericalHarmonicY@1, −1, θ, φDz
z êê
z
1
2 {
FullSimplify@#, φεRealsD &
r Cos@φD Sin@θD
− r
4 a05ê2 è!!!!!!
2 π!
2a0
ψpy =
rwave@2, 1, rD
i
j
j
j $%%%%%% HSphericalHarmonicY@1, 1, θ, φD +
j
1
k 2
y
z
SphericalHarmonicY@1, −1, θ, φDLz
z
z êê FullSimplify@#, φεRealsD &
{
General ::spell1 : Possible spelling error : new
4 a05ê2 è!!!!!!
2 π!
ψpz= rwave[2,1,r] SphericalHarmonicY[1,0,θ,φ]//Simplify
name "ψpz " is similar to existing symbols 8ψpx , ψpy <. More…
General ::spell : Possible spelling error : new symbol
25
r Cos@θD
− r
4 a05ê2 è!!!!!!!
2a0
2π
rule1={x→r Sin[θ] Cos[φ],y→r Sin[θ] Sin[φ],z→r Cos[θ]}
V = −A x2 − B y2 + HA + BL z2 ê. rule1 êê Simplify
{x→r Cos[φ] Sin[θ],y→r Sin[θ] Sin[φ],z→r Cos[θ]}
H12//TableForm
−24 A a02 π 0 0
24 a02 HA + BL π
0 −24 a02 B π 0
0 0
24 a02 HA + BL π
E1=H13[[1,2]];E2=H13[[1,1]];E3=H13[[1,3]]
Zeeman effect
We consider the eigen value problem
i
1 j "################################################ y
jW1 + W2 − W12 − 2 W1 W2 + W22 + 4 H2 μB2 z
Eigensystem[HH]//Simplify
::W3, z,
2 k {
1 i
j "################################################ y
jW1 + W2 + W12 − 2 W1 W2 + W22 + 4 H2 μB2 z z>,
2 k {
è!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
I−W1 + W2 + W12 − 2 W1 W2 + W22 + 4 H2 μB2 M
:80, 0, 1<, : , 1, 0>,
è!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2 H μB
IW1 − W2 + W12 − 2 W1 W2 + W22 + 4 H2 μB2 M
:− , 1, 0>>>
2 H μB
eq1=Eigenvalues[HH]
26
i
1 j "################################################ y
:W3, jW1 + W2 − W12 − 2 W1 W2 + W22 + 4 H2 μB2 z z,
2 k {
1 i
j "################################################ y
jW1 + W2 + W12 − 2 W1 W2 + W22 + 4 H2 μB2 z z>
2 k {
rule3={μB→1,A→2,B→1 ,a0→1}
{μB→1,A→2,B→1,a0→1}
i "###########################
#y
0.5 j 5684.89 + 4. H2 z
W11=eq1[[2]]/.rule2/.rule3//N
j−226.195 − 1. z
k {
i "###########################
#y
0.5 j 5684.89 + 4. H2 z
W22=eq1[[3]]/.rule2/.rule3//N
j−226.195 + z
k {
W33=eq1[[1]]/.rule2/.rule3//N
226.195
Plot[Evaluate[{W11,W22}/.a0→1],{H,0,100},PlotStyle→{Hue[0],
Hue[0.7]},Prolog→AbsoluteThickness[1.5],PlotPoints→100,Back
ground→GrayLevel[0.7],AxesLabel→{"H","Energy"}]
Energy
H
20 40 60 80 100
-50
-100
-150
-200
Graphics
Fig.10 Zeeman splitting of the energy level W1 and W2. We assume that a0 = 1,
μB = 1, A = 2 and B = 1.
27
1
ψ xy = −i [Y22 (θ ,φ ) − Y2− 2 (θ ,φ )]
2
1 1
ψ yz = i [Y2 (θ ,φ ) + Y2−1 (θ ,φ )]
2
1
ψ zx = − [Y21 (θ ,φ ) − Y2−1 (θ ,φ )] , (6.1)
2
1
ψ x2 − y2 = [Y22 (θ ,φ ) + Y2− 2 (θ ,φ )]
2
ψ 3 z 2 − r 2 = Y20 (θ ,φ )
or
1
ψ 1 = ψ xy = −i [ l = 2, m = 2 − l = 2, m = −2 ]
2
1
ψ2 = ψ yz = i [ 2,1 + 2,−1 ]
2
1
ψ3 = ψ zx = − [ 2,1 − 2,−1 ] . (6.2)
2
1
ψ4 = ψ x2 − y2 = [ 2,2 + 2,−2 ]
2
ψ 5 = ψ 3z 2
−r 2
= 2,0
Note that the notation of the spherical harmonics used here is the same as that used in the
Mathematica.
Table[{2,m,SphericalHarmonicY[2,m,θ, φ]},{m,-
2,2,1}]//TableForm
2 −2 1 −2 φ $%%%%%%%
15 % Sin@θD2
4 2π
2 −1 1 − φ $%%%%%%%
15 % Cos@θD Sin@θD
2 2π
2 0 1 $%%%%%
5 % H− 1 + 3 Cos@θD2L
4 π
2 1 −1 φ $%%%%%%%
15 % Cos@θD Sin@θD
2 2π
2 2 1 2 φ $%%%%%%%
15 % Sin@θD2
4 2π
28
2 1 r
R3, 2 (r ) = 2 7/2
r 2 exp(− ). (6.3)
15 81a0 3a0
There are two types of orbital states: the dε orbits dxy, dyz, dzx, the dγ orbits: dx2-y2 and
d 3 z 2 −r 2 . The complete wavefunctions are
given as follows.
dε (t2 g )
1 r 2 xy 2 1 r
r d xy = exp(− ) = exp(− ) 3 xy
81a0
7/2
3a0 2π 6π 81a0
7/2
3a0
1 r 2 yz 2 1 r
r d yz = exp(− ) = exp(− ) 3 yz
81a0
7/2
3a0 2π 6π 81a0
7/2
3a0
1 r 2 zx 2 1 r
r d zx = exp(− ) = exp(− ) 3 zx ,
81a0
7/2
3a0 2π 6π 81a0
7/2
3a0
dγ (eg )
1 r x2 − y2 2 1 r 3( x2 − y 2 )
r dx2 − y2 = exp(− ) = exp(− )
2π 6π 81a0
7/2 7/2
81a0 3a0 3a0 2
1 r 3z 2 − r 2 2 1 r 3z 2 − r 2
r d3 z 2 − r 2 =
exp( − ) = exp( − )
6π 6π 81a0
7/2 7/2
81a0 3a0 3a0 2
(6.4)
For convenience, we use the notations: r d xy = ψ 1 , r d yz = ψ 2 , r d zx = ψ 3 ,
r d x 2 − y 2 = ψ 4 , r d3 z 2 − r 2 = ψ 5 . The complete wavefunctions are given by
82<
83<
0.2
81<
0.1 0.05-0.2 -0.1
0
0 -0.05
0.05 0
0.1
-0.1 0.2
-0.2
2
0.2
0.2
0.1
0
0.1
0.05
0.2 0
0 0
-0.05 0.1
-0.1
-0
-0.1
-0.2
0.2 0
-0.1 -0.2
-0.1 -0.2
0
0.1 -0.05
-0.2 0
0.2
0 2 0.05
0 05
29
85<0
84<
-0.1
0.1-0.05
0.05
05 0.05
0.1
-0.05 0
-0.1
0.4
0.2
0.05 0
0 0.2
-0.05
0 -0.2
-0.2
0 -0.4
-0.2
0.2
Fig.11 Angular parts of the wavefunctions for (1) dxy, (2) dyz, (3) dzx, (4) d x 2 − y 2 , and
(5) d 3 z 2 −r 2 .
0.2
0.1
x
-0.2 -0.1 0.1 0.2
-0.1
-0.2
30
Fig.12(a) Representation of dxy orbit in the x-y plane.
The Coulomb repulsion energy is small for the dxy because of the existence of the
neighboring negative charges located at (± a, 0, 0) and (0, ± a, 0).
0.2
0.1
z
-0.2 -0.1 0.1 0.2
-0.1
-0.2
The Coulomb repulsion energy is small for the dyz because of the existence of the
neighboring negative charges located at (0,± a, 0) and (0, 0, ± a).
0.2
0.1
z
-0.2 -0.1 0.1 0.2
-0.1
-0.2
The Coulomb repulsion energy is small for the dzx because of the existence of the
neighboring negative charges located at (0, 0, ± a, 0) and (± a, 0,0).
31
84<y
0.3
0.2
0.1
x
-0.3 -0.2 -0.1 0.1 0.2 0.3
-0.1
-0.2
-0.3
The Coulomb repulsion energy is large for the d x 2 − y 2 because of the existence of the
neighboring negative charges located at (± a, 0, 0) and (0, ± a, 0).
85<x
0.1
0.05
z
-0.4 -0.2 0.2 0.4
-0.05
-0.1
The Coulomb repulsion energy is large for the d 3 z 2 −r 2 , because of the existence of the
neighboring negative charges located at (0,0,± a).
32
Fig.13 Splitting of the energy level (2D) of ground state of free (3d)1 under the
octahedral field.
((Note))
We have some comment on the independence of the dγ-orbital wave function (there are
only two independent states).
1 r x2 − y2
r d x2 − y 2 = exp( − ) (6.12)
2π
7/2
81a0 3a0
1 r 3z 2 − r 2 1 r 3z 2 − ( x 2 + y 2 + z 2 )
r d3 z 2 − r 2 = exp( − ) = exp( − )
6π 6π
7/2 7/2
81a0 3a0 81a0 3a0
(6.13)
or
1 r 2z 2 − (x 2 + y 2 )
r d 3 z 2 −r 2 = exp(− )
6π
7/2
81a 0 3a 0
. (6.14)
1 r (z 2 − x 2 ) + (z 2 − y 2 )
= exp( − )
6π
7/2
81a 0 3a 0
There are two independent states since
r d x2 − y2 + r d y2 −z2 + r d z2 −x2 = 0, (6.15)
r dz2 − x2 − r d y2 − z2
r d3 z 2 − r 2 = . (6.16)
3
33
4 2
V (r ) = Ze 2 [ + + A( x 2 + y 2 − 2 z 2 ) + D( x 4 + y 4 + z 4 )
a c
3 3 4
+ Q( z 4 + 6 x 2 y 2 ) − ( 5 + 5 )r 4 ].
4 a c
(6.17)
where
1 1 5 4 3
A=( 3
− 3 ), D = ( 5 + 5 ) > 0,
a c 4 a c
. (6.18)
5 1 1 3 3 4
Q = − ( 5 − 5 ), ε = ( 5 + 5 ) > 0.
4 a c 4 a c
The sign of A and Q changes depending on the ratio of c/a. On the other hand, D and ε
are always positive. Here the matrix element is defined by
⎛W11 0 0 0 0 ⎞
⎜ ⎟
⎜ 0 W22 0 0 0 ⎟
i Vˆ j = ∫ r d (i ) V (r ) r d ( j ) = ⎜ 0 0 ⎟ . (6.19)
*
0 W33 0
⎜ ⎟
⎜ 0 0 0 W44 0 ⎟
⎜ ⎟
⎝ 0 0 0 0 W55 ⎠
Ze 2
[4c + a(2 + 72ca0 A + 1215ca0 (11D + 19Q − 21ε ))]
2 4
W11 =
ca
Ze 2
[4c + a(2 + 9ca0 (−4 A + 135a0 (11( D + Q) − 21ε ))]
2 2
W22 =
ca
Ze 2
[4c + a(2 + 9ca0 (−4 A + 135a0 (11( D + Q) − 21ε ))]
2 2
W33 =
ca
Ze 2
[4c + a (2 + 72ca0 A + 1215ca0 (15D + 7Q − 21ε ))]
2 4
W44 =
ca
Ze 2
[4c + a(2 + 9ca0 (−8 A + 405a0 (5( D + Q) − 7ε ))]
2 2
W55 =
ca
Here we use the perturbation theory (degenerate case) (see the Mathematica program
below).
B. Eigenvalue problem
The eigenvalues
E1 = W11 for dxy, E2 = W22 for dyz, E3 = W33 for dzx, E4 = W44 for d x 2 − y 2 ,
and
E5 = W55 for d 3 z 2 −r 2 .
We note that
E4 = E5 = E1 + 10 Dq , E1 = E2 = E3
in the limit of A→0, Q→0, and ε→0. In the general case,
E4 − E1 30Q E5 − E1 8A 10Q
= 10 − = 10 − 2
−
Dq D Dq 27 a0 D D
34
E3 − E1 2A 20Q E3 − E2
=− 2
− = 0 (E2 and E3 are degenerate).
Dq 9a0 D D Dq
E5 − E4 8A 20Q
=− 2
+
Dq 27a0 D D
We now make a plot of the energy levels (normalized by Dq) as a function of c/a0, where
a/a0 is fixed as a parameter and a0 is the Bohr radius. For convenience we choose a/a0 =
10 (a = 10 a0 = 5.3 Å).
Energy êDq
22.5
20
17.5
15
12.5
10
cêa0
7.5 9.5 10 10.5 11
Fig.14 Energy levels for the tetragonal case, as a function of c/a0. a/a0 = 10. [dxy
(green), dyz (red), dzx (red), d x 2 − y 2 (blue), and d 3 z 2 −r 2 (purple). The energy level of dyz
and dzx (degenerate) is higher than that of dxy for c<a, while the energy level of dyz
and dzx (degenerate) is lower than that of dxy for c>a. The energy level of d 3 z 2 −r 2 is
higher than that of d x 2 − y 2 for c<a, while the energy level of d 3 z 2 −r 2 is lower than that
of d x 2 − y 2 for c>a.
Fig.15 Energy diagram for (3d)1 ion in the octahedral and tetragonal cases for (i) c >
a = b. (ii) c < a = b.
35
C. Mathematica program: Eigenvalue problem
Using the Mathematica program we solve the eigenvalue problems. The result
obtained is shown in Fig.14.
rule1 = 9A1 → J N, D1 → J N , Q1 → −J N,
1 1 20 15 5 5
− + −
a3 c3 4 a5 4 c5 4 a5 4 c5
ε1 → J N=
9 12
+
4 a5 4 c5
:A1 → >
1 1 5 15 5 5 9 3
− , D1 → + , Q1 → − + , ε1 → +
a3 c3 a5 4 c5 4 a5 4 c5 4 a5 c5
rule2={x→r Sin[θ] Cos[φ],y→r Sin[θ] Sin[φ],z→r Cos[θ]}
{x→r Cos[φ] Sin[θ],y→r Sin[θ] Sin[φ],z→r Cos[θ]}
V1=V/.rule2//TrigFactor
1
−
r2 H1 + 3 Cos@2 θDL
ψ@5D =
− r
6π
r2 H1 + 3 Cos@2 θDL
− r
81 a07ê2
36
2 $%%%%%%
2% − r
3a0 r2
15
81 a07ê2
Average of r4
r4av = IntegrateA r4 r2 rwave@3, 2, rD2, 8r, 0, ∞<E êê
FullSimplify@#, Re@a0D > 0D &
25515 a04
q1= < >
r4av êê Simplify
2
q4 =
105
486 a04
The definition of D0: D0=D1 Z ,
D0 = D1 Z e02
D1 e02 Z
Average of r2
r2av = IntegrateA r2 r2 rwave@3, 2, rD2, 8r, 0, ∞<E êê
FullSimplify@#, Re@a0D > 0D &
126 a02
Matrix element calculations
J[p_,q_]:=ψ[p] V1 ψ[q]
J[1,1]//Simplify
1
i − 3a0
839808 a a07 c π
j
j e02 r4 Z H128 a + 256 c − 32 a A1 c r2 + 42 a c D1 r4 + 42 a c Q1 r4 −
2r
k
64 a c r4 ε1 + 8 a c r2 H− 12 A1 + HD1 + Q1L r2L Cos@2 θD +
14 a c HD1 + Q1L r4 Cos@4 θD − 4 a c D1 r4 Cos@2 Hθ − 2 φLD +
12 a c Q1 r4 Cos@2 Hθ − 2 φLD + a c D1 r4 Cos@4 Hθ − φLD −
3 a c Q1 r4 Cos@4 Hθ − φLD + 6 a c D1 r4 Cos@4 φD − 18 a c Q1 r4 Cos@4 φD +
a c D1 r4 Cos@4 Hθ + φLD − 3 a c Q1 r4 Cos@4 Hθ + φLD − 4 a c D1 r4
y
Cos@2 Hθ + 2 φLD + 12 a c Q1 r4 Cos@2 Hθ + 2 φLDL Sin@θD4 Sin@2 φD2z
z
{
ψ[q]
H1@p_, q_D :=
Norm1[p_,q_]:=ψ[p]
{{1,0,0,0,0},{0,1,0,0,0},{0,0,1,0,0},{0,0,0,1,0},{0,0,0,0,1}
}
Eigenvalue problem:
eq3=Table[H1[p,q],{p,1,5},{q,1,5}]//Simplify;eq4=Eigensystem
[eq3]//Simplify
37
e02 Z H4 c + a H2 + 72 a02 A1 c + 1215 a04 c H15 D1 + 7 Q1 − 21 ε1LLL
:: ,
>,
E4=eq4[[1,1]]//Simplify;E3=eq4[[1,2]]//Simplify;E2=eq4[[1,3]
e02 Z H4 c + a H2 + 72 a02 A1 c + 1215 a04 c H11 D1 + 19 Q1 − 21 ε1LLL
]//Simplify;E5=eq4[[1,4]]//Simplify;E1=eq4[[1,5]]//Simplify
1215 H3 α5 + 4 β5L
ê. rule1 ê. rule3 êê Simplify
E4
E40 =
2 H23085 β5 + 72 α2 β5 + 4 α4 β5 + 2 α5 H1215 − 36 β2 + β4LL
D0 q4
1215 H3 α5 + 4 β5L
ê. rule1 ê. rule3 êê Simplify
E3
E30 =
4 H− 3645 β5 − 18 α2 β5 + 2 α4 β5 + α5 H−4860 + 18 β2 + β4LL
D0 q4
1215 H3 α5 + 4 β5L
ê. rule1 ê. rule3 êê Simplify
E2
E20 =
4 H− 3645 β5 − 18 α2 β5 + 2 α4 β5 + α5 H−4860 + 18 β2 + β4LL
D0 q4
1215 H3 α5 + 4 β5L
ê. rule1 ê. rule3 êê Simplify
E1
E10 =
4 H− 9720 β5 + 36 α2 β5 + 2 α4 β5 + α5 H1215 − 36 β2 + β4LL
D0 q4
1215 H3 α5 + 4 β5L
Energy={E10,E20,E30,E40,E50}
38
8H4 H−9720 β5 + 36 α2 β5 + 2 α4 β5 + α5 H1215 − 36 β2 + β4LLL ë H1215 H3 α5 + 4 β5LL,
H4 H−3645 β5 − 18 α2 β5 + 2 α4 β5 + α5 H− 4860 + 18 β2 + β4LLL ë H1215 H3 α5 + 4 β5LL,
H4 H−3645 β5 − 18 α2 β5 + 2 α4 β5 + α5 H− 4860 + 18 β2 + β4LLL ë H1215 H3 α5 + 4 β5LL,
H2 H23085 β5 + 72 α2 β5 + 4 α4 β5 + 2 α5 H1215 − 36 β2 + β4LLL ë H1215 H3 α5 + 4 β5LL,
H2 H10935 β5 − 72 α2 β5 + 4 α4 β5 + 2 α5 H7290 + 36 β2 + β4LLL ë H1215 H3 α5 + 4 β5LL<
Energy diagram
Plot[Evaluate[Energy/.α→10],{β,9,11},PlotStyle→Table[Hue[0.
2 i],{i,0,5}],
Prolog→AbsoluteThickness[1.5],Background→GrayLevel[0.7],
AxesLabel→{"c/a0","Energy/Dq"}]
Graphics
Energy/.α→β//Simplify
:−4 + >
4 β4 4 β4 4 β4 4 β4 4 β4
, −4 + , −4 + , 6+ , 6+
2835 2835 2835 2835 2835
ψ i Lˆz ψ i = 0
ψ i Lˆx ψ i = 0 , ( Lˆ x + Lˆ y + Lˆ z ) ψ i = 6h 2 ψ i , (i = 1 – 5).
2 2 2
(6.22)
ψ i Lˆ y ψ i = 0
E. Mathematica program
Using the Mathematica we show the formula of the orbital angular momentum in
quantum mechanics.
((Mathematica-1))
Orbital angular momentum of 3d electron
39
Hx D@#, yD − y D@#, xDL &; Lx := Hy D@#, zD − z D@#, yDL &;
— —
Lz :=
è!!!!!!!!!!!!!!!!!!!!
2
I3 z2 − Ix2 + y2 + z2MM fA x2 + y2 + z2 E;
1
ψ5 =
2
è!!!! "###################
# è!!!! "####################
{Lz[ψ1],Lz[ψ2],Lz[ψ3],Lz[ψ4],Lz[ψ5]}//Simplify
:− 3 Hx2 − y2L — fB x2 + y2 + z2 F, − 3 x z — fB x2 + y2 + z2 F,
è!!!! "###################
# è!!!! "###################
#
3 y z — fB x2 + y2 + z2 F, 2 3 x y — fB x2 + y2 + z2 F, 0>
è!!!! "###################
#
{Lx[ψ1],Lx[ψ2],Lx[ψ3],Lx[ψ4],Lx[ψ5]}//Simplify
: 3 x z — fB x2 + y2 + z2 F,
è!!!! "###################
# è!!!! "####################
− 3 Hy2 − z2L — fB x2 + y2 + z2 F, − 3 x y — fB x2 + y2 + z2 F,
è!!!! "###################
# "###################
#
− 3 y z — fB x2 + y2 + z2 F, − 3 y z — fB x2 + y2 + z2 F>
è!!!! "###################
#
{Ly[ψ1],Ly[ψ2],Ly[ψ3],Ly[ψ4],Ly[ψ5]}//Simplify
:− 3 y z — fB x2 + y2 + z2 F,
è!!!! "###################
# è!!!! "###################
#
3 x y — fB x2 + y2 + z2 F, 3 Hx2 − z2L — fB x2 + y2 + z2 F,
è!!!! "###################
# "###################
#
− 3 x z — fB x2 + y2 + z2 F, 3 x z — fB x2 + y2 + z2 F>
((Mathematica-2))
Lx=Table[Jx[2,n,m],{n,2,-2,-1},{m,2,-2,-
1}];Ly=Table[Jy[2,n,m],{n,2,-2,-1},{m,2,-2,-
1}];Lz=Table[Jz[2,n,m],{n,2,-2,-1},{m,2,-2,-1}];Lp=Lx+
Ly;Lm=Lx-
Ly;I1={{1,0,0,0,0},{0,1,0,0,0},{0,0,1,0,0},{0,0,0,1,0},{0,0,
0,0,1}}
Column matrix of y1, y2, y3, y4, y5
40
ψ@1D = − 81, 0, 0, 0, −1<; ψ@2D = 80, 1, 0, 1, 0<;
è!!!! è!!!!
2 2
ψ@3D = − 80, 1, 0, −1, 0<; ψ@4D = 81, 0, 0, 0, 1<;
è!!!! è!!!!
1 1
2 2
ψ@5D = 80, 0, 1, 0, 0<
Table@ψ@iD∗.Lx.ψ@1D, 8i, 1, 5<D
{0,0,0,0,0}
Calculation of L+.y[i]
Table@ψ@iD∗.Lp.ψ@1D, 8i, 1, 5<D
9− , 0, 0, −1, 3 =
Table@ψ@iD∗.Lp.ψ@4D, 8i, 1, 5<D
41
è!!!! è!!!!
90, − 3 , − 3 , 0, 0=
Calculation of L-.y[i]
Table@ψ@iD∗.Lm.ψ@1D, 8i, 1, 5<D
9− , 0, 0, 1, − 3 =
Table@ψ@iD∗.Lm.ψ@4D, 8i, 1, 5<D
Fig.16 Energy diagram for (3d)9 in the octahedral and tetragonal cases for (i) c > a =
b. (ii) c < a = b.
In the (3d)9 electron configuration, we have (3d)9: 2S+1 = 2, 2L+1 = 5 (L = 2). The
ground state is a Kramers doublet. Since S = 1/2, there is no single ion anisotropy (see
12.1 for detail). A full shell of electrons has a charge cloud which is spherically
symmetric so that removal of one electron leaves a distributed charge deficiency
coincident with the space charge of the electron moved. This gap or “hole” behaves like a
single positively charged electron.
For the (3d)9 ion in an octahedral crystal field, the argument therefore proceeds as for
3d1 except the de and dγ levels will now be reversed in the energy level diagram because
the ion has a lesser energy when the lobes of the positive hole’s wavefunction are
directed towards neighboring negative ions rather than between the ions. In some cases
the derivation from regularity might arise spontaneously from the crystal-field effect
themselves because, if the ground-state in a regular field is orbitally degenerate, slight
distortion of the environment can lower the ground-state energy and at the same time
remove some degeneracy, pushing some levels up and other down. Both the doubly
degenerate ground-state and the triply-degenerate level are split into two. If the distortion
in the z direction, the d(yz) and d(zx) states of the triplet remain symmetrically disposed
42
and, though displaced upwards in energy, they remain degenerate. The d(xy) state, on the
other hand, is displaced downward, in energy, and the center of gravity of the triplet
remains unchanged to the first order. The d(3z2-r2) and d(x2 – y2) states of the unperturbed
ground-state are not symmetrical with respect to the z-axis and are separated by the
distortion, d(3z2-r2) becoming an orbitally non-degenerate ground state.
In conclusion, The energy diagram of (3d)9 and (3d)1 configuration are inverted in
relation to each other because (3d)9 may be treated as a (3d)1 positive hole in a filled
(3d)10 shell.
Fig.17(a) Ground state of (3d)1: orbital triplet, where the red circle denote electron
up-spin state. dε1
(ii)
43
Fig.17(b) Ground state of (3d)2: orbital triplet. dε2
(iii)
Fig.17(c) Ground state of (3d)3: orbital singlet. dε3. The quenching of the orbital
angular momentum
(iv)
Fig.17(d) Ground state of (3d)4: orbital doublet. The possibility of Jahn-Teller effect.
dε3dγ1
(v)
Fig.17(e) Ground state of (3d)5: orbital singlet. dε3dγ2. The quenching of the orbital
angular momentum
(vi)
44
Fig.17(f) Ground state of (3d)6: orbital triplet. dε4dγ2. The red circle and blue circle
denote electron up-state and down-state.
(vii)
(viii)
Fig.17(h) Ground state of (3d)8: orbital singlet. dε6dγ2. The quenching of the orbital
angular momentum
(ix)
45
Fig.17(i) Ground state of (3d)9: orbital doublet. dε6dγ3. The possibility of the Jahn-
Teller effect
In Summary we have the following energy diagram for (3d)2. Note that when the
Hund’s second law is also taken into account, the six degeneracy of the first excited state
is split into the two triplet states.
46
Fig.18(d). Energy diagram of (3d)2 electron configuration (two-electrons system). Γ4:
the ground state (orbital triplet), Γ5: the first excited state (orbital triplet), Γ2: the
second excited state (orbital singlet).
Fig.19 (a) Ground state, the first excited state, and the second excited state of (3d)3
electron configuration (one-electron model).
47
8.5. Excited states of (3d)1 electron configuration
Ground state Orbital triplet (3C1 = 3)
Excited state Orbital doublet (2C1 = 2).
Fig.20(a) Ground state and excited states of (3d)1 electron configfuration (one
electron model)
In summary, we have
Fig.20(b) Ground state and excited states of (3d)1 electron configuration. The energy
levels are the same as shown in Fig.20(a).
48
Fig.21(a) Ground state and excited states of (3d)4 electron configuration (one
electron model)
Fig.21(b) Ground state and excited states of (3d)4 electron configuration (four
electrons system).
49
Fig.22 Energy diagram of (3d)n (n = 1, 2, 3, 4) with the rule-1 and rule-2. Γ4 and Γ5
triplets can be represented by a fictitious orbital angular momentum.
9. Jahn-Teller effect
We now consider the energy diagram of Cu2+ for the (3d)9 electron configuration. The
energy level of Cu2+ consists of ground state (orbital doublet Γ3, dγ) and the excited state
(orbital triplet Γ5, dε) in the octahedral field (c = a = b). The ground state is split into the
two orbital singlets [d(x2 – y2) and d(3z2 – r2)] in the tetragonal field, as a result of the
lattice distortion (c>a = b). Here the energy level of d(x2 – y2) is lower than that of d(3z2 –
r2). In this case, there are two contributions to the total energy. One is the energy loss due
to the lattice distortion (the increase of the distance c). The other is the energy gain due to
the lowering of the energy level of the ground state. If the energy gain is larger than the
energy loss, a spontaneous displacement occurs (Jahn-Teller effect).
50
What happens when the crystal field is strong (Vcr>VH). In this case all 6 states in the
de are occupied by 3 up-spin states and 3-spin down states since Vcr>VH. Then we have
the S = 0 (we call this low spin state).
11. Spin Hamiltonian of spin systems with the orbital singlet as a ground state
11.1. Pryce spin Hamiltonian
Now we apply the perturbation theory (degenerate case) where the orbital ground
state is singlet.16 The Hamiltonian is described by
Hˆ = Hˆ c + Hˆ ′ , (11.1)
where Ĥ c is the unperturbed Hamiltonian of the system under the octahedral crystal field
The perturbation Ĥ ′ is given by
Hˆ ′ = λLˆ ⋅ Sˆ + μ B (Lˆ + 2Sˆ ) ⋅ H , (11.2)
where the first term is the spin-orbit interaction and the second term is the Zeeman
energy. L̂ and Ŝ are the operators of the orbital and spin angular momentum in the
quantum mechanics. We now transform Ĥ ′ into the so-called spin Hamiltonian by a
method proposed by Pryce, where the orbital dependence is projected out.
Let us evaluate the expectation value of Ĥ ′ for a non-degenerate ground state (orbital
singlet Ψ0 .3 To the second order perturbation for the non-degenerate case (see the
Appendix for the perturbation theory)
Here
Ψ0 Hˆ ' Ψ0 = Ψ0 λLˆ ⋅ Sˆ + μ B (L ˆ + 2Sˆ ) ⋅ H Ψ = Ψ μ (2Sˆ ) ⋅ H Ψ = 2μ Sˆ ⋅ H
0 0 B 0 B
ˆ ˆ ˆ ˆ ˆ
Γ' , γ ' H ' Ψ = Γ' , γ ' λL ⋅ S + μ (L + 2S) ⋅ B Ψ
0 B 0
(11.5)
or
H (Sˆ ) = ∑ ( μ B g μν H μ Sˆν − λ2 Λ μν Sˆμ Sˆν − μ B Λ μν H μ Hν ) ,
2
(11.6)
μ ,ν
where E (Γ' , γ ' ) − E0 is the energy gap between the ground state and the excited states.
g μν = 2(δ μν − λΛ μν ) . (11.7)
51
What is the physical meaning of the above spin Hamiltonian?
(i) The g tensor which is different from 2, because of spin-orbit interaction.
(ii) The second term represents the single-ion anisotropy. Note that Λ μν reflects the
symmetry of the system.
We neglect the last term which is related to the Van Vleck susceptibility. We also assume
that
⎛ Λ xx 0 0 ⎞
⎜ ⎟
Λ = ⎜ 0 Λ yy 0 ⎟ . (11.8)
⎜ 0 0 Λ zz ⎟⎠
⎝
⎧ g xx = 2(1 − λΛ xx )
⎪
⎨ g yy = 2(1 − λΛ yy ) . (11.9)
⎪ g = 2(1 − λΛ )
⎩ zz zz
The spin Hamiltonian should reflect the symmetry of the crystal field. Then the spin
Hamiltonian can be written as
Hˆ (Sˆ ) = −λ2 (Λ zz Sˆ z2 + Λ xx Sˆ x2 + Λ yy Sˆ y2 ) + μ B ( g xx Sˆ x H x + g yy Sˆ y H y + g zz Sˆ z H z )
1 1 1
= −λ2 [Λ zz − (Λ xx + Λ yy )][Sˆ z2 − S ( S + 1)] − λ2 (Λ xx − Λ yy )(Sˆ x2 − Sˆ y2 ) .
2 3 2
+ μ B ( g xx Sˆ x H x + g yy Sˆ y H y + g zz Sˆ z H z )
(11.10)
Note that for simplicity we use the unit of h = 1 . We put
1
δ = −λ2 [Λ z − (Λ x + Λ y )] , (11.11)
2
ε = −λ2 (Λ x − Λ y ) / 2 . (11.12)
Then the effective spin Hamiltonian Hˆ (Sˆ ) is described as
1
Hˆ (Sˆ ) = δ [ Sˆ z2 − S ( S + 1)] + ε ( Sˆ x2 − Sˆ y2 ) + μ B ( g xx Sˆ x H x + g yy Sˆ y H y + g zz Sˆ z H z ) .
3
(11.13)
Note that an important experimental fact is that the crystal-field parameters (δ, ε, and g)
do not change appreciably for the same types of the systems.
52
1
D = −λ2 [Λ z − (Λ x + Λ y )]
2
2 ⎡
λ ⎛ g ⎞ 1⎛ g g y ⎞⎤
= − ⎢⎜1 − z ⎟ − ⎜⎜ 2 − x − ⎟⎥
λ ⎢⎣⎝ 2 ⎠ 2⎝ 2 2 ⎟⎠⎥⎦
. (11.15)
⎡ gz 1 ⎤
= −λ ⎢ − + (g x + g y )⎥
⎣ 2 4 ⎦
λ⎡
g z − (g x + g y )⎥
1 ⎤
= ⎢
2⎣ 2 ⎦
We assume that gz = gc, gx = gy = ga for simplicity. The constant D is rewritten as
λ
D= ( gc − ga ) . (11.16)
2
Note that the sign of the spin-orbit interaction constant λ is as follows,
⎧λ > 0 (less than half; (3d) n n < 5))
⎨
⎩λ < 0 (more than half; (3d) n > 5) )
n
Then we have
If λ > 0, then gc > ga ⇒ D > 0.
If λ < 0, then gc > ga ⇒ D < 0.
Note that the g-value of the free electron is not precisely 2. There are quantum
electrodynamics corrections which leads to the value g = 2.0023193043622.
53
4λ
g z = 2(1 − )
E1 − E4
λ
g x = 2(1 − ), (11.19)
E2 − E4
λ
g y = 2(1 − )
E3 − E4
where λ (≈ -850 cm-1) and E2 = E3>E1>E5>E4. We find that g x = g y < g z .
11.3. Ni2+
As a typical example, we consider the case of Ni2+ [((3d)8; S = 1, L = 3] under the
octahedral crystal field. The energy diagram consists of a orbital singlet state (Γ2) as a
ground state and an orbital triplet (Γ5). The energy difference between the ground state
and the excited state is 1 eV and is much larger than the magnitude of the spin-orbit
coupling constant λ (= -335 cm-1 = -0.0415 eV), Note that 1 meV = 8.0655 cm-1.
(i) Since the ground state is the orbital singlet, the quenching of the orbital angular
momentum occurs.
(ii) The spin degeneracy is (2S + 1) = 3. S, ms with S = 1, and ms = 1, 0, and -1.
Fig.23 Crystal-field splitting of Ni2+ ion ground state in the ochtahedral and
tetragonal fields.
In the ground-orbital singlet, the mean value of orbital angular momentum is zero and the
value of spin S = 1. The spin anisotropy arises only as a 2nd-order effect through the
combined effect of spin-orbit coupling and trigonal field within the upper states which
are linked with the ground state in the second order perturbation. For a trigonal distortion,
we expect the anisotropy to be adequately represented by a term DS z2 .
11.4 Cr3+
Cr3+ (3d3) has a ground state 4F. In an octahedral field, the sevenfold orbital
degeneracy (2L+1 = 7) is removed, and the levels are split into a lower orbital singlet (Γ2)
and two higher lying orbital triplets (Γ5, Γ4). The low-lying singlet has a fourfold spin
degeneracy (2S+1 = 4, S = 3/2). It is not removed even by the combination action of the
54
octahedral and spin-orbit coupling, forming two Kramers doublets (see the detail in
Sec.12). It can be removed only by the external magnetic field.
12.2. S =1
A. Eigenvalue problem for S = 1
For S = 1 in the case of (3d)8, Ni2+ ion), the spin operators are described by (3x3) spin
matrices,
⎛ 1 ⎞ ⎛ −i ⎞
⎜ 0 0 ⎟ ⎜ 0 0 ⎟
⎜ 2 ⎟ ⎜ 2 ⎟ ⎛1 0 0 ⎞
⎜ 1 1 ⎟ ⎜ i −i ⎟ ⎜ ⎟
ˆ
Sx = ⎜ 0 , ˆ
S = 0 , Sˆ = ⎜ 0 0 0 ⎟.
2⎟ ⎜ 2 2⎟
y z
2 ⎜ 0 0 − 1⎟
⎜ 1 ⎟ ⎜ i ⎟ ⎝ ⎠
⎜⎜ 0 0 ⎟⎟ ⎜⎜ 0 0 ⎟⎟
⎝ 2 ⎠ ⎝ 2 ⎠
(12.6)
Then
55
⎛1 1⎞ ⎛ 1 1⎞
⎜ 0 ⎟ ⎜ 0 − ⎟ ⎛1 0 0⎞
⎜2 2⎟ ⎜ 2 2⎟ ⎜ ⎟
Sˆ x = ⎜ 0
2
1 0⎟, Sˆ y2 = ⎜ 0 1 0 ⎟ , Sˆ = ⎜ 0 0 0 ⎟ .
2
z
⎜1 0
1⎟ ⎜− 1 0 1 ⎟ ⎜0 0 1⎟
⎝ ⎠
⎜2 2 ⎟⎠ ⎜ 2 2 ⎟⎠
⎝ ⎝
(12.7)
Here we consider the eigenvalue problem for the effective spin Hamiltonian given by
1
Hˆ = δ [ Sˆ z2 − S ( S + 1)] + ε ( Sˆ x2 − Sˆ y2 ) . (12.8)
3
We have the determinant given by
δ
−W 0 ε
3
2δ
Δ = det( Hˆ − WIˆ) = 0 − −W 0 = 0, (12.9)
3
δ
ε 0 −W
3
or
2 δ δ
W = − δ , −ε , +ε
3 3 3
where W is the energy eigenvalue. For Ni2+ it is known that δ = ε = 1 cm-1.
The ground state with spin triplet (orbital singlet) is split into three spin singlets due to
the spin-orbit interaction.
What is the wave function of the ground state?
2
For W = W1 = − δ ,
3
1 0 −1
⎛δ ⎞
⎜ − W1 0 ε ⎟
1 ⎜3 ⎟⎛⎜ c1 ⎞⎟ ⎛⎜ 0 ⎞⎟
⎜ 0 2δ , (12.10)
0 − − W1 0 ⎟⎜ c2 ⎟ = ⎜ 0 ⎟
⎜ 3 ⎟⎜ ⎟ ⎜ ⎟
−1 ⎜ δ ⎟⎝ c3 ⎠ ⎝ 0 ⎠
⎜ ε 0 − W 1⎟
⎝ 3 ⎠
or
56
⎛ δ 0 ε ⎞⎛ c1 ⎞ ⎛ 0 ⎞
⎜ ⎟⎜ ⎟ ⎜ ⎟
⎜ 0 0 0 ⎟⎜ c2 ⎟ = ⎜ 0 ⎟ , (12.11)
⎜ ε 0 δ ⎟⎜ c ⎟ ⎜ 0 ⎟
⎝ ⎠⎝ 3 ⎠ ⎝ ⎠
or
δc1 + εc3 = 0
εc1 + δc3 = 0
c1 = c3 = 0
ψ1 = 0
δ
For W = W2 = −ε
3
⎛δ ⎞
⎜ − W2 0 ε ⎟
⎜3 ⎟⎛⎜ c1 ⎞⎟ ⎛⎜ 0 ⎞⎟
⎜ 0 2δ
− − W2 0 ⎟⎜ c2 ⎟ = ⎜ 0 ⎟
⎜ 3 ⎟⎜ ⎟ ⎜ ⎟
⎜ δ ⎟⎝ c3 ⎠ ⎝ 0 ⎠
⎜ ε 0 − W2 ⎟
⎝ 3 ⎠
or
⎛ε 0 ε ⎞⎛ c1 ⎞ ⎛ 0 ⎞
⎜ ⎟⎜ ⎟ ⎜ ⎟
⎜ 0 − δ + ε 0 ⎟⎜ c2 ⎟ = ⎜ 0 ⎟ . (12.12)
⎜ε 0 ε ⎟⎠⎜⎝ c3 ⎟⎠ ⎜⎝ 0 ⎟⎠
⎝
ε (c1 + c3 ) = 0 and c 2 = 0 ∴ c3 = −c1
1
ψ2 = ( 1 − −1 )
2
δ
For W = W3 = +ε ,
3
⎛δ ⎞
⎜ − W3 0 ε ⎟
⎜3 ⎟⎛⎜ c1 ⎞⎟ ⎛⎜ 0 ⎞⎟
⎜ 0 2δ
− − W3 0 ⎟⎜ c2 ⎟ = ⎜ 0 ⎟
⎜ 3 ⎟⎜ ⎟ ⎜ ⎟
⎜ δ ⎟⎝ c3 ⎠ ⎝ 0 ⎠
⎜ ε 0 − W3 ⎟
⎝ 3 ⎠
or
⎛−ε 0 ε ⎞⎛ c1 ⎞ ⎛ 0 ⎞
⎜ ⎟⎜ ⎟ ⎜ ⎟
⎜ 0 − δ − ε 0 ⎟⎜ c2 ⎟ = ⎜ 0 ⎟ , (12.13)
⎜ ε 0 − ε ⎟⎠⎜⎝ c3 ⎟⎠ ⎜⎝ 0 ⎟⎠
⎝
− εc1 + εc3 = 0 and c 2 = 0 ∴ c1 = c3
1
ψ3 = ( 1 + −1 )
2
57
Since the ground state is a spin singlet
1
⎯⎯⎯⎯⎯ ψ3 = ( 1 + −1 )
2
1
⎯⎯⎯⎯⎯ ψ2 = ( 1 − −1 )
2
⎯⎯⎯⎯⎯ ψ1 = 0
we find that the quenching of the spin angular momentum occurs in the systrem
ψ1 Sz ψ1 = 0 Sz 0 = 0 . (12.14)
number and kB is the Boltzmann constant. When we assume that g = 2, we have the T
dependence of the susceptibility shown here, where T0 is changed as a parameter. The
susceptibility has a maximum at T / T0 = 0.6835 .
58
χ Hemu êmole L
0.3
0.25
0.2
0.15
0.1
0.05
T HKL
2 4 6 8 10 12 14
At low temperature, χ becomes zero, because of the factor eδ / k B T . The energy of ground
state does not change with the magnetic field. (the system is in the ground state with
singlet.)
1 è!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
Jx@__, n_, m_D := H_ − mL H_ + m + 1L KroneckerDelta@n, m + 1D +
1 è!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2
H_ + mL H_ − m + 1L KroneckerDelta@n, m − 1D;
1 è!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2
Jy@__, n_, m_D := − H_ − mL H_ + m + 1L KroneckerDelta@n, m + 1D +
1 è!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2
H_ + mL H_ − m + 1L KroneckerDelta@n, m − 1D;
59
g HHx − HyL μB
:: , ε>,
è!!!
2!
δ
+ g Hz μB,
: >,
è!!!
2! è!!!
2!
,− ,
g HHx + HyL μB
3
:ε, Hδ − 3 g Hz μBL>>
è!!!
2!
1
,
3
Eigenvalue problems: eigenvalues and eigenfunctions; Hx
1 i
j "############################################### y
j−δ + 3 ε − 3 δ2 + 2 δ ε + ε2 + 4 g2 Hx2 μB2 z
eqx1=Eigensystem[Hamil[Hx,0,0]]//Simplify
:: Hδ − 3 εL, z,
6 k {
1
i
j i "############################################### yy
j−δ + 3 j jε + δ2 + 2 δ ε + ε2 + 4 g2 Hx2 μB2 z zz
3
z>,
k k {{
1
i i "############################################### yz
y
:8−1, 0, 1<, :1, − j jg Hx μB j jδ − 3 ε + δ2 + 2 δ ε + ε2 + 4 g2 Hx2 μB2 z
6
zz ì
k k {{
i
jè!!!! i
jδ ε + ε2 + g2 Hx2 μB2 − ε "############################################## #y
zyz, 1>,
j 2j δ2 + 2 δ ε + ε2 + 4 g2 Hx2 μB2 z z
k k {{
i i "############################################### yy
:1, j
jg Hx μB jj−δ + 3 ε + δ2 + 2 δ ε + ε2 + 4 g2 Hx2 μB2 z zz
zì
k k {{
i
jè!!!
! i
jδ ε + ε2 + g2 Hx2 μB2 + ε "############################################## #y
zyz, 1>>>
j 2j δ2 + 2 δ ε + ε2 + 4 g2 Hx2 μB2 z z
k k {{
Eigenvalue problems: eugenvalues and eigenfunctions; Hy
1 i
j "############################################### y
j−δ − 3 ε − 3 δ2 − 2 δ ε + ε2 + 4 g2 Hy2 μB2 z
eqy1=Eigensystem[Hamil[0,Hy,0]]
:: Hδ + 3 εL, z,
6 k {
1
i
j "############################################### y
j−δ − 3 ε + 3 δ2 − 2 δ ε + ε2 + 4 g2 Hy2 μB2 z
3
z>,
k {
1
i i è!!! ! è!!! !
:81, 0, 1<, :− 1, j j j
6
j 2 g Hy δ μB + 3 2 g Hy ε μB +
k k
è!!!! "############################################### y
zyzì
2 g Hy μB δ2 − 2 δ ε + ε2 + 4 g2 Hy2 μB2 z z
{{
j2 i
i j−δ ε + ε2 + g2 Hy2 μB2 + ε "##############################################
#yzy
z, 1>,
j j δ2 − 2 δ ε + ε2 + 4 g2 Hy2 μB2 z z
k k {{
i i è!!!! è!!!!
:−1, − j
j j j 2 g Hy δ μB + 3 2 g Hy ε μB −
k k
è!!!! "############################################### yz
y
2 g Hy μB δ2 − 2 δ ε + ε2 + 4 g2 Hy2 μB2 z zz ì
{{
i
j2 i
jδ ε − ε2 − g2 Hy2 μB2 + ε "##############################################
#yzy
z, 1>>>
j j δ2 − 2 δ ε + ε2 + 4 g2 Hy2 μB2 z z
k k {{
Eigenvalue problems: eigenvalues and eigenfunctions; Hz
2δ 1 i jδ − 3 "######################### y 1 i
jδ + 3 "######################### y
ε2 + g2 Hz2 μB2 z ε2 + g2 Hz2 μB2 z
eqz1=Eigensystem[Hamil[0,0,Hz]]
::− j z, j z>,
3 k { 3 k {
,
è!!!!!!!!!!!!!!!!!!!!!!!!!
3
è!!!!!!!!!!!!!!!!!!!!!!!!!
ε
:− , 0, 1>>>
− g Hz μB − ε + g Hz μB2
2 2 2
ε
rule1={g→1,μB→1}
{g→1,μB→1}
Magnetic field // z, E/d vs
60
energyHz=eqz1[[1]]/δ/.{Hz→δ y,ε→δ
2 1 "############
# 1 "############
#
x}/.rule1//Simplify[#,{x>0,δ>0}]&
:− , − x2 + y2 , + x2 + y2 >
GraphicsArray
Fig.26 Zeeman splitting of the energy levels for S = 1 as a function of gμ B H z / δ .
H//z. The index shows the value of ε (= 0 – 1.0).
i
j "############################# y
j−1 + 3 x + 3 1 + 2 x + x2 + 4 y2 z
3
z>
k {
1
6
61
px@ξ_D := PlotA Evaluate@energyHx ê. x → ξD, 8y, 0, 2<,
PlotStyle → Table@ Hue@0.15 iD, 8i, 0, 4<D,
Prolog → AbsoluteThickness@3 D, PlotLabel → 8ξ<,
AxesLabel → 9" ", "Eêδ"=, Background → GrayLevel@0.7D,
gμBHx
δ
DisplayFunction → IdentityE;
ptx = Evaluate@Table@ px@ξD, 8ξ, 0, 1, 0.2<DD;
Show@GraphicsArray@Partition@ptx, 2DD,
DisplayFunction → $DisplayFunctionD
GraphicsArray
Fig.27 Zeeman splitting of the energy levels for S = 1 as a function of gμ B H x / δ .
H//x. The index shows the value of ε (= 0 – 1.0).
Magnetic field êê y, ε ê δ vs
gμBHy
δ
energyHy=eqy1[[1]]/δ/.{Hy→δ y,ε→δ
1 i
j "############################# y
j− 1 − 3 x − 3 1 − 2 x + x2 + 4 y2 z
x}/.rule1//Simplify[#,{x>0,δ>0}]&
: z,
6 k {
1
+ x,
i
j "############################# y
j−1 − 3 x + 3 1 − 2 x + x2 + 4 y2 z
3
z>
k {
1
6
62
py@ξ_D := PlotA Evaluate@energyHy ê. x → ξD, 8y, 0, 2<,
PlotStyle → Table@ Hue@0.15 iD, 8i, 0, 4<D,
Prolog → AbsoluteThickness@3 D, PlotLabel → 8ξ<,
AxesLabel → 9" ", "Eêδ"=, Background → GrayLevel@0.7D,
gμBHy
δ
DisplayFunction → IdentityE;
pty = Evaluate@Table@ py@ξD, 8ξ, 0, 1, 0.2<DD;
Show@GraphicsArray@Partition@pty, 2DD,
DisplayFunction → $DisplayFunctionD
1 1
gμBHy gμBHy
0.5 1 1.5 2 δ 0.5 1 1.5 2 δ
-1 -1
-2 -2
1 1
gμBHy gμBHy
0.5 1 1.5 2 δ 0.5 1 1.5 2 δ
-1 -1
-2 -2
GraphicsArray
Fig.28 Zeeman splitting of the energy levels for S = 1 as a function of gμ B H x / δ .
H//x. The index shows the value of ε (= 0 – 1.0).
12.3 S = 3/2
A. Simple case with ε = 0 and δ>0
For simplicity we consider the spin Hamiltonian with ε = 0 and δ > 0.
1
Hˆ = δ [ Sˆ z2 − S ( S + 1)] . (12.18)
3
Here note that
Hˆ mz = ε (mz ) mz , (12.19)
with
5
ε (mz ) = mz Hˆ mz = δ (mz2 − ) , (12.20)
4
63
3 1
where mz is the eigenket of Ŝ z with the eigenvalue m z = ± , ± ,
2 2
ˆ
S z mz = m z mz , (12.21)
For this we get
⎧ ⎛ 3⎞
⎪⎪ ε ⎜ mz = ± 2 ⎟ = δ
⎝ ⎠ (12.22)
⎨
⎛ 1 ⎞
⎪ε ⎜ mz = ± ⎟ = −δ
⎪⎩ ⎝ 2⎠
Both states are spin doublets (Kramers doublet). This is true for any half-integer spin (S =
1/2, 3/2, 5/2,…)
Fig.29 Schematic diagram of the energy levels due to the crystal-field for S = 3/2.
There are two Kramers doublets.
B. Kramers doublet
((Time reversal operator))
Most operators of interest are either even or odd under the time reversal.
ˆ
ΘAΘˆ ˆ −1 = ± Aˆ (+: even, -: odd).
(1) ˆ iΘ
Θ ˆ −1 = −i1̂ (i is a pure imaginary, 1̂ is the identity operator).
(2) ˆ pˆ Θ
Θ ˆ −1 = −pˆ
(3) Θˆ rˆ Θˆ −1
= rˆ : ( Θ̂ r = r ).
(4) ˆ Sˆ Θ
Θ ˆ −1 = −Sˆ ( Ŝ is the spin angular momentum).
2
(5) Θ ˆ −1 = Hˆ , when Hˆ = pˆ + V ( xˆ ) and V (xˆ ) is a potential energy. The relation
ˆ Hˆ Θ
2m
is independent of the form of V (xˆ ) .
See the Appendix 2 for the detail of the time reversal operator for spin 1/2
((Kramers theorem))
We introduce the time reversal operator Θ̂ . Suppose that Ĥ is invariant under time
reversal,
[ Hˆ , Θ
ˆ ] = 0̂ . (12.23)
Let φn and Θ̂ φn be the energy eigenket and its time-reversed states, respectively.
Hˆ Θ
ˆ φ =Θ
n
ˆ Hˆ φ = ΘˆE φ = E Θ
n n
ˆφ .
n n n (12.24)
Θ̂ φn and φn belong to the same energy eigenvalue.
When Θ ˆ 2 = −1̂ (half-integer), Θ̂ φ and φ are orthogonal.
n n
64
This means that Θ̂ φn and φn (having the same energy) must correspond to distinct
states (Kramers doublet).
S = 1 (even number of electrons) singlet
S = 3/2 (odd number of electrons) doublet (Kramers doublet)
Fig.30 Schematic diagram of the energy levels for the odd number of electrons
under the crystal field and spin-orbit interaction. The ground state is a Kramers
doublet.
For the odd number of electrons in the incomplete shell, there remains the levels with
double degeneracy in the absence of B in spite of any crystal field (Kramers doublet).
((Note))
Co2+: (3d)7 n = 7 odd number ⇒ Kramers doublet
2+ 8
Ni : (3d) n = 8 even number ⇒ No Kramers doublet
S → 3 ê 2;
3
65
Matrix element in the general case
3 g Hz μB 1 è!!!! è!!!!
Hamil[Hx,Hy,Hz]//Simplify
::δ + , 3 g HHx − HyL μB, 3 ε, 0>,
1 è!!!! è!!!!
2 2
: 3 g HHx + HyL μB, −δ + , g HHx − HyL μB, 3 ε>,
g Hz μB
è!!!! 1 è!!!!
2 2
: 3 ε, g HHx + HyL μB, H−2 δ − g Hz μBL, 3 g HHx − HyL μB>,
1
è!!!! 1 è!!!!
2 2
:0, 3 ε, 3 g HHx + HyL μB, δ − >>
3 g Hz μB
2 2
Eigenvalue problems: eugenvalues and eigenfunctions for H = Hx, Hy, and Hz
eqx1=Eigensystem[Hamil[Hx,0,0]]//Simplify;eqy1=Eigensystem[H
amil[0,Hy,0]]//Simplify;eqz1=Eigensystem[Hamil[0,0,Hz]]//Sim
"########################################################
plify
::−
1
g Hz μB − δ2 + 3 ε2 − 2 g Hz δ μB + g2 Hz2 μB2 ,
"########################################################
2
1
− g Hz μB + δ2 + 3 ε2 − 2 g Hz δ μB + g2 Hz2 μB2 ,
g Hz μB "########################################################
2
− δ2 + 3 ε2 + 2 g Hz δ μB + g2 Hz2 μB2 ,
g Hz μB "########################################################
2
+ δ2 + 3 ε2 + 2 g Hz δ μB + g2 Hz2 μB2 >,
è!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2
::0, − , 0, 1>,
è!!!
3! ε
δ − g Hz μB + δ2 + 3 ε2 − 2 g Hz δ μB + g2 Hz2 μB2
è!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
:0, , 0, 1>,
è!!!3! ε
−δ + g Hz μB + δ2 + 3 ε2 − 2 g Hz δ μB + g2 Hz2 μB2
è!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
: , 0, 1, 0>,
è!!!
3! ε
δ + g Hz μB − δ2 + 3 ε2 + 2 g Hz δ μB + g2 Hz2 μB2
è!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
: , 0, 1, 0>>>
è!!!
3! ε
δ + g Hz μB + δ2 + 3 ε2 + 2 g Hz δ μB + g2 Hz2 μB2
rule1={g→1,μB→1}
{g→1,μB→1}
Magnetic field // z, e/d vs
energyHz=eqz1[[1]]/δ/.{Hz→δ y,ε→δ
i
j "########################### y "###########################
j−2 3 x2 + H−1 + yL2 − yz
x}/.rule1//Simplify[#,{x>0,δ>0}]&
: z, 3 x2 + H− 1 + yL2 − ,
k {
1 y
i "########################
jy − 2 3 x2 + H1 + yL2 z # y
z, "#########################
2 2
j + 3 x2 + H1 + yL2 >
k { 2
1 y
2
66
pz@ξ_D := PlotAEvaluate@energyHz ê. x → ξD, 8y, 0, 2<,
PlotStyle → Table@ Hue@0.15 iD, 8i, 0, 4<D,
Prolog → AbsoluteThickness@2.5 D, PlotLabel → 8ξ<, PlotPoints → 100,
AxesLabel → 9" ", "Eêδ"=, Background → GrayLevel@0.7D,
gμBHz
δ
DisplayFunction → IdentityE;
ptz = Evaluate@Table@ pz@ξD, 8ξ, 0, 1, 0.2<DD;
Show@GraphicsArray@Partition@ptz, 2DD,
DisplayFunction → $DisplayFunctionD
GraphicsArray
Fig.31 Zeeman splitting of the energy levels for S = 3/2 as a function of gμ B H z / δ .
H//z. The index shows the value of ε (= 0 – 1.0).
1 i "######################################
jy − 2 1 + 3 x2 − y + 3 x y + y2 z y
z, y + "#####################################
#
2 2
j 1 + 3 x2 − y + 3 x y + y2 >
2 k { 2
67
px@ξ_D := PlotAEvaluate@energyHx ê. x → ξD, 8y, 0, 2<,
PlotStyle → Table@ Hue@0.15 iD, 8i, 0, 4<D,
Prolog → AbsoluteThickness@2.5 D, PlotLabel → 8ξ<, PlotPoints → 100,
AxesLabel → 9" ", "Eêδ"=, Background → GrayLevel@0.7D,
gμBHx
δ
DisplayFunction → IdentityE;
ptx = Evaluate@Table@ px@ξD, 8ξ, 0, 1, 0.2<DD;
Show@GraphicsArray@Partition@ptx, 2DD,
DisplayFunction → $DisplayFunctionD
GraphicsArray
Fig.32 Zeeman splitting of the energy levels for S = 3/2 as a function of gμ B H x / δ .
H//x. The index shows the value of ε (= 0 – 1.0).
Magnetic field êê y, E ê δ vs
gμBHy
δ
energyHy=eqy1[[1]]/δ/.{Hy→δ y,ε→δ
i
j "###################################### y y "######################################
jy − 2 1 + 3 x2 − y − 3 x y + y2 z
x}/.rule1//Simplify[#,{x>0,δ>0}]&
: z,
k { 2
1
+ 1 + 3 x2 − y − 3 x y + y2 ,
y "#####################################
# y "######################################
2
− − 1 + 3 x + y + 3 x y + y , − + 1 + 3 x2 + y + 3 x y + y2 >
2 2
2 2
68
py@ξ_D := PlotAEvaluate@energyHy ê. x → ξD, 8y, 0, 2<,
PlotStyle → Table@ Hue@0.15 iD, 8i, 0, 4<D,
Prolog → AbsoluteThickness@2.5 D, PlotLabel → 8ξ<, PlotPoints → 100,
AxesLabel → 9" ", "Eêδ"=, Background → GrayLevel@0.7D,
gμBHy
δ
DisplayFunction → IdentityE;
pty = Evaluate@Table@ py@ξD, 8ξ, 0, 1, 0.2<DD;
Show@GraphicsArray@Partition@pty, 2DD,
DisplayFunction → $DisplayFunctionD
GraphicsArray
Fig.33 Zeeman splitting of the energy levels for S = 3/2 as a function of gμ B H y / δ .
H//y. The index shows the value of ε (= 0 – 1.0).
69
Sx = Table@Jx@2, n, mD, 8n, 2, −2, −1<, 8 m, 2, −2, −1<D;
Sy = Table@Jy@2, n, mD, 8n, 2, −2, −1<, 8 m, 2, −2, −1<D;
Sz = Table@Jz@2, n, mD, 8n, 2, −2, −1<, 8 m, 2, −2, −1<D;
I1 = 881, 0, 0, 0, 0<, 80, 1, 0, 0, 0<, 80, 0, 1, 0, 0<, 80, 0, 0, 1, 0<,
80, 0, 0, 0, 1<<;
Hamil@ Hx_, Hy_, Hz_D :=
J g μB H Hx Sx + Hy Sy + Hz SzL + δ JSz.Sz − S HS + 1L I1N +
1
3
ε HSx.Sx − Sy.SyLN ê. S → 2;
Matrix element for H=0
è!!!!
992 δ, 0, 6 ε, 0, 0=, 80, −δ, 0, 3 ε, 0<,
Hamil[0,0,0]//Simplify
è!!!!!!!!!!!!!!! è!!!!!!!!!!!!!!!
$%%%%%
2 Iδ + δ2 + 3 ε2 M $%%%%%
2 I−δ + δ2 + 3 ε2 M
, 0, 1>, :1, 0, , 0, 1>>>
3 3
−
ε ε
"#############
# "##############
energy=eq1[[1]]/δ/.{ε→δ x}//Simplify[#,{x>0,δ>0}]&
:2, −1 − 3 x, − 1 + 3 x, −2 1 + 3 x2 , 2 1 + 3 x2 >
PlotAEvaluate@energyD, 8x, −2, 2<,
PlotStyle → Table@ Hue@0.15 iD, 8i, 0, 4<D,
Prolog → AbsoluteThickness@2.5 D, AxesLabel → 9" ", "Eêδ"=,
ε
δ
Background → GrayLevel@0.7DE
Eêδ
6
4
2
ε
-2 -1 1 2 δ
-2
-4
-6
Graphics
Fig.34 Energy diagram for S = 2. Ε/δ vs ε/δ. The ground state is a spin singlet.
12.5. S = 5/2
A. Simple case for S = 5/2
70
Fig.35 Schematic diagram of the energy levels for S = 5/2 under the crystal field.
There are three Kramers doublets.
(2S + 1) = 6 degeneracy
1
Hˆ = D[ Sˆ z2 − S ( S + 1)] . (12.25)
3
Here note that
Hˆ mz = ε (mz ) mz . (12.26)
1 5 7
ε (mz ) = mz Hˆ mz = D(mz2 − ⋅ ⋅ ) (12.27)
3 2 2
where mz = 5/2, 3/2, -1/2, -3/2,-5/2.
From this we get
⎛ 5⎞ ⎛ 25 35 ⎞ 10
ε ⎜ mz = ± ⎟ = δ ⎜ − ⎟ = δ
⎝ 2⎠ ⎝ 4 12 ⎠ 3
⎛ 3⎞ ⎛ 9 35 ⎞ 2
ε ⎜ mz = ± ⎟ = δ ⎜ − ⎟ = − δ
⎝ 2⎠ ⎝ 4 12 ⎠ 3
⎛ 1⎞ ⎛ 1 35 ⎞ 8
ε ⎜ mz = ± ⎟ = δ ⎜ − ⎟ = − δ
⎝ 2⎠ ⎝ 4 12 ⎠ 3
71
Sx = Table@Jx@5 ê 2, n, mD, 8n, 5 ê 2, −5 ê 2, −1<, 8m, 5 ê 2, −5 ê 2, −1<D;
Sy = Table@Jy@5 ê 2, n, mD, 8n, 5 ê 2, −5 ê 2, −1<, 8m, 5 ê 2, −5 ê 2, −1<D;
Sz = Table@Jz@5 ê 2, n, mD, 8n, 5 ê 2, −5 ê 2, −1<, 8m, 5 ê 2, −5 ê 2, −1<D;
I1 = 881, 0, 0, 0, 0, 0<, 80, 1, 0, 0, 0, 0<, 80, 0, 1, 0, 0, 0<,
80, 0, 0, 1, 0, 0<, 80, 0, 0, 0, 1, 0<, 80, 0, 0, 0, 0, 1<<;
Hamil@ Hx_, Hy_, Hz_D :=
J g μB H Hx Sx + Hy Sy + Hz SzL + δ JSz.Sz − S HS + 1L I1N +
1
3
ε HSx.Sx − Sy.SyLN ê. S → 5 ê 2;
Hamil@0, 0, 0D
Matrix element for H=0
eq1 = ê. 8ε → δ x< êê Simplify@#, 8x > 0, δ > 0<D &
è!!!!!! è!!!!
δ
:: , 0, 10 x, 0, 0, 0>, :0, − , 0, 3 2 x, 0, 0>,
10 2
è!!!! è!!!!!!
3 3
:0, 0, 3 2 x, 0, − , 0>, :0, 0, 0, 10 x, 0, >>
2 10
3 3
Eigenvalue problems: eugenvalues and eigenfunctions for H = 0
eq1 = Eigenvalues@eq1D êê Simplify;
PlotAEvaluate@eq1D, 8x, −2, 2<,
PlotStyle → Table@ Hue@0.15 iD, 8i, 0, 4<D,
Prolog → AbsoluteThickness@2.5 D, PlotPoints → 100,
AxesLabel → 9" ", "Eêδ"=, Background → GrayLevel@0.7DE
ε
Eêδ
δ
10
ε
-2 -1 1 2 δ
-5
-10
Graphics
Fig.36 Energy diagram for S = 5/2. Ε/δ vs ε/δ. All the three states are a Kramers
doublet.
13. Spin Hamiltonian of Fe2+ and Co2+ in the trigonal crystal field
13.1. Fe2+
A. Energy diagram
For Fe2+ [(3d)6], the ground state is the orbital triplet. Thus we cannot use the method
proposed by Pryce. In this case, we need to determine the energy diagram by taking both
the crystal field and the spin orbit interaction into account.18
72
Fig.37 Crystal field splitting of Fe2+ ion ground states.
The free-ion (3d)6, 5D state (L = 2, S = 2) of the Fe2+ is split by the cubic crystal field
into the orbital doublet (E) and orbital triplet (T2), the latter being the lowest one. The
orbital wavefunction of the ground orbital triplet are represented by
2 1
− 1 = ψ d1 = φ− 2 + φ1
3 3
0 = ψ s = φ0 (13.1)
2 1
1 = ψ d2 = φ2 − φ−1
3 3
where we use the notation, l = 3, m = φm and we have the relation Lˆ z φm = m φm (m =
-3, -2, -1, 0, 1, 2, 3). Hereafter we do not use the operators in quantum mechanics for
convenience. We consider the splitting of the orbital triplet by the perturbing Hamiltonian
given by
r v 2
H 0 = − kλl ⋅ S − δ (lz − ) .
2
(13.2)
3
where k (≈ 1) is a constant, λ (= -100 cm-1 for Fe2+) is a spin-orbit coupling constant, and
S is the spin angular momentum of the magnitude 2. The second term is the tetragonal
field. A fictitious angular momentum l of the magnitude 1 represents the triplet state (l is
antiparallel to the real orbital angular momentum L (= -kl).
Since lz + Sz is a constant of the motion, its eigenvalue m can be used to classify the
various states, where m = lz’+ Sz’, lz|lz’>= lz’|lz’> (lz’ = 1, 0, -1), and Sz|Sz’>= Sz’|Sz’>
(Sz’ =2, 1,0,-1,-2). There are 15 states. A notation l z ' , S z ' is used to specify these states.
m = lz’+ Sz’, Eigenfunctions
3 lz ' = 1, S z ' = 2
2 1,1 , 0,2
1 1,0 , 0,1 , − 1,2
73
0 1,−1 , 0,0 , − 1,1
-1 1,−2 , 0,−1 , − 1,0
-2 0,−2 , − 1,−1
-3 − 1,−2
We have the splitting of the ground orbital triplet by the spin-orbit coupling λ’ and
the trigonal field δ (>0). The energy levels are denoted by E3 (m = ± 3), E2(±) (m = ± 2),
δ
E1(i) (i = 1, 2, 3) (m = ±1), E0(0) and E0(±) (m = 0). Here we use x = .
λ'
(i) m = ±3 (E3)
E3 x
=− −2
λ' 3
(ii) m = ±2 [E2]
E2 x 1 1
= − ± 9 + 2x + x2
λ' 6 2 2
(iii) m = ±1 [E1(1), E1(2), E1(3)]
(i )
E1 x
= − + 1 + ξi (i = 1, 2, 3).
λ' 3
where
3 2
x = ξi + 1 − −
ξi + 1 ξi − 1
(iv) m = 0 [E0 , E0(±)]
(0)
The ground level is either E0(+) or E1(1), depending on the sign of x. All the energy states
except for E1(1) and E0(+) might be neglected in the first approximation, since these
lowest levels lie 100 cm-1 below the others.
Thus we may use a fictitious spin s = 1 for the lowest three states denoted by the
eigenkets |ψ0> for the singlet and |ψ±1> for the doublet:
ψ ±1 = c1 ± 1,0 + c2 0,±1 + c3 m 1,±2 , for E = E1(1), (13.3)
ψ 0 = a1 1,−1 + a2 0,0 + a3 − 1,1 for E = E0(+). (13.4)
The parameters c1, c2, c3, a1, a2, and a3 are defined by
2 2 2 2 2 2
c1 + c2 + c3 = 1 , a1 + a2 + a3 = 1 , (13.5)
3α 2α
c1 = − , c2 = α , c3 = , (13.5)
1 + ξ1 1 − ξ1
3 ξ0
a1 = a3 = − , a2 = , (13.6)
ξ0 2 + 6 ξ0 2 + 6
where
74
3 2
α2 =[ +1+ ]−1 , (13.7)
(1 + ξ1 ) 2
(1 − ξ1 ) 2
and the parameters ξ0 and ξ1 are related to the energy E1(1) and E0(+) through
(1) (+)
E1 x E0 x
= − + 1 + ξ1 , = − + 1 + ξ0 . (13.8)
λ' 3 λ' 3
75
2
H = − D ∑ ( siz − ) − 2 J ∑ s i ⋅ s j − 2 J A ∑ siz s jz ,
2
(13.21)
i 3 <i, j > <i, j >
where the fictitious spin s = 1. J = q2Jij D ≈ δ/10 (>0) is the single ion anisotropy, and JA
(=J (p2-q2)/q2) is the anisotropic exchange interaction. The second term is the isotropic
exchange interaction, and third term is the anisotropic exchange interaction. The spin
anisotropy parameter Deff is defined as Deff (= D (s -1/2) + 2zsJA. The XY (Ising)
symmetry appears when Deff<0 (Deff>0).
x 1 1 è!!!!!!!!!!!!!!!!!!!!!
3 6 2 2
E2n = −J − − x2 + 2 x + 9 N;
6 2 2
E13@x1_D := ModuleA8t, y<, t = x1; eq11 = t == Hy + 1L −
3 2
− ;
y+ 1 y−1
s1 = NSolve@eq11, yD êê Flatten; E13@tD = −J− + Hy + 1LN ê. s1@@1DDE;
t
3
E12@x2_D := ModuleA8t, y<, t = x2; eq22 = t == Hy + 1L −
3 2
− ;
y+ 1 y−1
s2 = NSolve@eq22, yD êê Flatten; E12@tD = −J− + Hy + 1LN ê. s2@@2DDE;
t
3
E11@x3_D := ModuleA8t, y<, t = x3; eq33 = t == Hy + 1L −
3 2
− ;
y+ 1 y−1
s3 = NSolve@eq33, yD êê Flatten; E11@tD = −J− + Hy + 1L ê. s3@@3DDNE;
t
3
E0 = −J− + 1N;
x
3
E0p@x4_D := ModuleA8t, y<, t = x4; eq44 = Jy + − 1N Jy − N − 6 0;
t 2t
76
c1@x1_, k1_D := ModuleA8t, k, y, s1, y1<, t = x1; k = k1;
eq11 = t == Hy + 1L − ; s1 = NSolve@eq11, yD êê Flatten;
3 2
−
y+ 1 y− 1
i −è!!!
j
! y
3 z
y1 = y ê. s1@@3DD; c1@t, kD = j
j zE;
j 1 + y1 z
z
1
$%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% k {
H1+ y1L H1− y1L
3 2
2 + 1+ 2
y1 = y ê. s1@@3DD; c2@t, kD = E;
1
$%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%
H1+ y1L H1− y1L
3 2
2 + 1+ 2
$%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% k {
H1+ y1L H1− y1L
3 + 1 + 2
2 2
è!!!!
3 3
y0 = Jz + − 1N ê. s2@@2DD; a1@t, kD = −
è!!!!!!!!!!!!!
E;
t 3
3 6 + y02
a2@x1_, k1_D := ModuleA8t, k, z, s2, y0<, t = x1; k = k1;
eq22 = Jz + − 1N Jz − N − 6 0; s2 = NSolve@eq22, zD êê Flatten;
t 2t
3 3
y0 = Jz + − 1N ê. s2@@2DD; a2@t, kD = è!!!!!!!!!!!!! E;
t y0
3 6 + y02
a3@x1_, k1_D := ModuleA8t, k, z, s2, y0<, t = x1; k = k1;
eq22 = Jz + − 1N Jz − N − 6 0; s2 = NSolve@eq22, zD êê Flatten;
t 2t
è!!!!
3 3
y0 = Jz + − 1N ê. s2@@2DD; a3@t, kD = −
è!!!!!!!!!!!!!
E
t 3
è!!!!
3 6 + y02
q@x_, k_D := 3 Hc1@x, kD a1@x, kD + c2@x, kD a2@x, kDL +
è!!!!
2 c3 @x, kD a3@x, kD; p@x_, k_D := c2@x, kD2 + 2 c3@x, kD2;
gc@x_, k_D := −k c1@x, kD2 + 2 c2@x, kD2 + Hk + 4L c3@x, kD2;
ga@x_, k_D := −k Hc1@x, kD a2@x, kD + c2@x, kD a3@x, kDL +
è!!!! è!!!!
2 3 Hc1@x, kD a1@x, kD + c2@x, kD a2@x, kDL + 2 2 c3@x, kD a3@x, kD
Overview of the Energy levels, E3, E2+, E2-, E13[x], E12[x], E11[x], E0, E0+[x],E0-[x]
as a function of x
Plot[Evaluate[{E3,E2p,E2n,E13[x],E12[x],E11[x],E0,E0p[x],E0n
[x]}],{x,-8,8}, PlotStyle→Table[Hue[0.1
77
i],{i,0,8}],Prolog→AbsoluteThickness[3],
AxesLabel→{"x=δ/λ'","E/|λ'|"}, Background→GrayLevel[0.7]]
Eê»λ'»
x=δêλ'
-7.5 -5 -2.5 2.5 5 7.5
-2
-4
-6
Graphics
Fig.38 The splitting of the ground orbital triplet by the spin-orbit coupling and the
trigonal field for Fe2+ ion in the trigonal field. E3 (red), E2(-) (yellow), E1(3) (green),
E0(-) (purple), E0(0) (blue), E1(2) (green), E2(+) (orange), E1(1) (ligh blue), and E0(+) (blue).
-3.5
-4
-4.5
x=δêλ'
-6 -4 -2 2 4 6
Graphics
Fig.39 The detail of the two energy levels E1(1) (red) and E0(+) (light blue) as a
ground state for Fe2+ ion in the trigonal field.
Plot of the g-factors gc[x,k] and ga[x,k] as a function of x with k = 0.9, 0.95, and1.0
Plot[Evaluate[Table[{gc[x,k],ga[x,k]},{k,0.9,1,0.05}]],{x,-
10,10}, PlotStyle→Table[Hue[0.3
i],{i,0,10}],Prolog→AbsoluteThickness[3],
AxesLabel→{"x=δ/λ'","gc"},Background→GrayLevel[0.7]]
78
gc
4.5
3.5
2.5
x=δêλ'
-10 -5 5 10
Graphics
Fig.40 Plot of gc (red) and ga (blue) as a function of x for Fe2+ ion in the trigonal
field. k = 0.9, 0.95, and 1.
1.8
1.6
1.4
1.2
x=δêλ'
-7.5 -5 -2.5 2.5 5 7.5
Graphics
Fig.41 Plot of the parameters q (red) and p (green) for Fe2+ ion in the trigonal field. k
= 0.9, 0.95, and 1.
p2 − q2
Plot of as a function of x . k = 0.8, 0.9, 1, 1.1
p@x, kD2 − q@x, kD2
q2
PlotAEvaluateATableA , 8k, 0.8, 1.1, 0.1<EE,
q@x, kD2
8x, −8, 8<, PlotStyle → Table@ Hue@0.5 iD, 8i, 0, 8<D,
Prolog → AbsoluteThickness@3D, AxesLabel → 8"x=δê λ'", "D"<,
Background → GrayLevel@0.7DE
79
D
1.5
0.5
x=δêλ'
-7.5 -5 -2.5 2.5 5 7.5
-0.5
Graphics
Fig.42 Plot of the spin anisotropy parameter p2/q2 – 1 as a function of x for Fe2+ ion
in the trigonal field. k = 0.8, 0.9, 1, and 1.1.
3.5
2.5
gc
2.5 3 3.5 4 4.5 5
1.5
Graphics
Fig.43 Calculated relation between gc and ga-values of Fe2+ ion in the trigonal field.
k = 0.9 (red), 0.95 (yellow), and 1.0 (green).
80
0.8
0.6
0.4
0.2
-10 -5 5 10
Graphics
Fig.44 Plot of the probability amplitudes c12 (red), c22 (blue), and c32 (green), as a
function of x for Fe2+ ion in the trigonal field. k = 0.9
0.8
0.6
0.4
0.2
-10 -5 5 10
Graphics
Fig.45 Plot of the probability amplitude an2 (n = 1 – 3) for Fe2+ ion in the trigonal
field. k = 0.9.
13.2. Co2+
A. Energy diagram of Co2+
81
Fig.46 Crystal-field splittings of Co2+ ion ground states.
For Co2+ [(3d)7], the ground state is the orbital triplet. Thus we cannot use the method
proposed by Pryce. In this case, we need to determine the energy diagram by taking both
the crystal field and the spin orbit interaction into account.19,20
In a cubic crystal field the free-ion 3d7 (L = 3, S = 3/2), 4F state is split into two
orbital triplets and one orbital singlet with a triplet the lowest,
5 1
− 1 = ψ d1 = φ2 + φ−1
6 6
2 1 5
0 = ψs = φ0 − [ φ3 − φ− 3 ] , (13.22)
3 3 2
5 1
1 = ψ d2 = φ− 2 − φ1
6 6
where we use the notation, l = 2, m = φm and we have the relation Lˆ z φm = m φm (m =
-2, -1, 0, 1, 2). We can verify that
3
± 1 Lˆ z ± 1 = m , 0 Lˆ z 0 = 0 , ± 1 Lˆ z 0 = 0 Lˆ z ± 1 = ± 1 Lˆz m 1 = 0 .
2
(13.23)
We can show that all the matrix element of L̂ within the states 0 , ± 1 , are exactly the
same as the matrix element of -3l/2, where l is a fictitious angular momentum of the
magnitude 1. Hereafter we do not use the operators in quantum mechanics for
convenience.
Now we consider the splitting of the ground orbital triplet by the tetragonal crystal
field together with the spin-orbit coupling,
82
3 r v 2
H 0 = − kλl ⋅ S − δ (l z − ) ,
2
(13.24)
2 3
where λ’= kλ, λ is the spin -orbit coupling constant and may be different from its free-ion
value of -180 cm-1, and k is the orbital reduction factor due to admixture of 4P into 4T1
and is less than but of order unity, δ is the trigonal field strength, and S is the spin angular
momentum of the magnitude 3/2. A fictitious angular momentum l of the magnitude 1
represents the triplet state (l is antiparallel to the real orbital angular momentum L = -
3kl/2).
Since lz + Sz is a constant of the motion, its eigenvalue m can be used to classify the
various states, where m = lz’+ Sz’, (lz’ = 1, 0, -1 and Sz’ = 3/2, 1/2,-1/2,-3/2). There are
12 eigenfunctions.
m = lz’+ Sz’, Eigenfunctions
5 3
lz ' = 1, S z ' =
2 2
3 1 3
1, , 0,
2 2 2
1 1 1 3
1,− , 0, , − 1,
2 2 2 2
1 1 1 3
− − 1, , 0,− , 1,−
2 2 2 2
3 1 3
− − 1,− , 0,−
2 2 2
5 3
− − 1,−
2 2
The secular equation reduces to three separate equations for energy. The energy
eigenvalues are as follows.
Es (m = ±5/2),
Eq(±) (m = ±3/2),
Ec(0), Ec(1), and Ec(2) (m = ±1/2).
For all values of x, Ec(0) is the lowest energy. The wave functions |ψ±1>, |ψ±3>, and
|ψ±4> for m = ±1/2 are given by
3 1 1
ψ ±1 = c1 m 1,± + c2 0,± + c3 ± 1,m for E±1 = Ec(0), (13.25)
2 2 2
3 1 1
ψ ± 3 = c4 m 1,± + c5 0,± + c6 ± 1,m for E±3 = Ec(1), (13.26)
2 2 2
3 1 1
ψ ±4 = c7 m 1,± + c8 0,± + c9 ± 1,m for E±4 = Ec(2)). (13.27)
2 2 2
where the parameters ci (i = 1 - 9) are defined by
83
6 8
c1 = β 0 , c 2 = − β 0 , c3 = β0 ,
ς0 ς0 + 2
6 8
c4 = β 1 , c5 = − β 1 , c6 = β1 , , (13.28)
ς1 ς0 + 2
6 8
c7 = β 2 , c8 = − β 2 , c 9 = β2 ,
ς2 ς0 + 2
6 8
with β j = [ 2 + 1 + ]−1 / 2 (j = 0,1,2). The parameter ζj (j = 0, 1, 2) is related to
ςj (ς j + 2) 2
Ec(j)/λ’ as
Ec ( j ) x 3
= − + (ς j + 3) , (13.29)
λ' 3 4
and
δ 3 9 6
x = = (ς j + 3) − − . (13.30)
λ' 4 2ς j ς j + 2
The wave functions |ψ±2> and |ψ±5> for m = ±3/2 are given by
3 1
ψ ±2 = d1 0,± + d 2 ± 1,± for E±2 = Eq(+), (13.31)
2 2
3 1
ψ ± 5 = d3 0,± + d 4 ± 1,± for E±5 = Eq(-), (13.32)
2 2
where the parameter di (i = 1- 4) is defined by
(+)
9 2 E
d1 = γ (+)
, d2 = γ (+)
( x + q ), (13.33)
2 6 3 λ'
(+)
9 2 E
d3 = γ (−)
, d4 = γ (+)
( x + q ), (13.34)
2 6 3 λ'
with
(±)
9 2 E
γ (±)
= [( ) + ( x + q ) 2 ]−1 / 2 .
2
(13.35)
2 6 3 λ'
84
with Vz = -(3k/2)lz + 2Sz and Vx = -(3k/2)lx + 2Sx. If we take the z axis parallel to the c
axis, and x, y axes perpendicular to it, we may finally replace the true spin S = 3/2 within
the ground doublet, by the s= 1/2 fictitious spin, Sx = qsx, Sy = qsy, and Sz = psz;
p = 2 ψ ±1 ± S z ψ ±1 = 3c1 + c2 − 3c3 ,
2 2 2
(13.38)
q = 2 ψ ±1 S x ± iS y ψ m1 = 2 3c1c2 + 2c2 .
2
(13.39)
Let us now consider the Heisenberg exchange interaction given by
H = −2 J 0S i ⋅ S j = −2 J 0 q 2si ⋅ s j − 2 J 0 ( p 2 − q 2 ) siz ⋅ s jz , (13.40)
or
H = −2 J ∑s
<i, j >
i ⋅ s j − 2 J A ∑ siz s jz .
<i, j >
(13.41)
where J = q2J0 and JA = (p2/q2- 1)J. The first term is a Heisenberg-type (isotropic)
exchange interaction and the second term is an anisotropic exchange interaction: JA = J0
(p2 – q2). Since s = 1/2, there is no single ion anisotropy. The ratio JA/J (= (p2/q2-1)
provides a measure for the spin symmetry of the system. The spin dimensionality of the
system is Ising-like for JA>0, XY-like for JA<0, and Heisenberg-like for JA = 0.
85
c1@x4_, k1_D := ModuleA8t, y, k<, t = x4; k = k1;
eq44 = t == Hy + 3L − ; s4 = NSolve@eq44, yD êê Flatten;
3 9 6
−
4 2 y y+ 2
è!!!!
c1@t, kD = ê. s4@@3DDE;
6
y
$%%%%%%%%%%%%%%%%%%%%%%%%%%%
H y+2L2
6 + 1+ 8
y2
c2@t, kD = ê. s4@@3DDE;
−1
$%%%%%%%%%%%%%%%%%%%%%%%%%%%
H y+2L
6 + 1+ 8
2y 2
c3@t, kD = ê. s4@@3DDE;
8
y+2
$%%%%%%%%%%%%%%%%%%%%%%%%%%%
H y+2L2
6 + 1+ 8
y2
c4@t, kD = ê. s4@@1DDE;
6
y
$%%%%%%%%%%%%%%%%%%%%%%%%%%%
H y+2L
6 + 1+ 8
2y 2
c5@t, kD = ê. s4@@1DDE;
−1
$%%%%%%%%%%%%%%%%%%%%%%%%%%%
H y+2L
6 + 1+ 8
2y 2
c6@t, kD = ê. s4@@1DDE;
8
y+2
$%%%%%%%%%%%%%%%%%%%%%%%%%%%
H y+2L
6 + 1+ 8
2y 2
c7@t, kD = ê. s4@@2DDE;
6
y
$%%%%%%%%%%%%%%%%%%%%%%%%%%%
H y+2L2
6 + 1+ 8
y2
x=δêλ'
-10 -5 5 10
-2
-4
-6
Graphics
Fig.47 Energy levels of six Kramers doublets for Co2+ ion in the trigonal field. Ec(0)
(light green), Eq- (green) Ec(1) (blue), Ec(2) (purple), Eq+ (yellow), and Es (red)
0.8
0.6
0.4
0.2
x
-10 -5 5 10
Graphics
Fig.48 Plot of the probability amplitudes cn2 (n = 1 – 9) for Co2+ ion in the trigonal
field. k = 0.9.
87
0.8
0.6
0.4
0.2
-10 -5 5 10
Graphics
Fig.49 Plot of the probability amplitudes dn2 (n = 1 – 4) for Co2+ ion in the trigonal
field. k = 0.9.
Plot of gc vs ga
k = 0.9, 0.95, 1
p1=ParametricPlot[Evaluate[Table[{gc[x,k],ga[x,k]},{k,0.9,1.
0,0.05}]],{x,-
200,100},Prolog→AbsoluteThickness[2],AxesLabel→{"gc","ga"},
PlotStyle→Table[Hue[0.3
i],{i,0,5}],PlotRange→{{0,10},{0,6}},
Background→GrayLevel[0.7]]
88
ga
6
gc
2 4 6 8 10
Fig.51 Calculated gc and ga-values of Co2+ ion in the trigonal field. k = 0.9 (red),
0.95 (green), and 1.0 (blue).
Graphics
Parameters p and q as a function of x
k = 0.8, 0.9, 1.0, 1.1
Plot[Evaluate[Table[{p[x,k],q[x,k]},{k,0.8.1.1,0.1}]],{x,-
5,5}, PlotStyle→Table[Hue[0.15
i],{i,0,6}],Prolog→AbsoluteThickness[3],AxesLabel→{"x","p,q
"}, Background→GrayLevel[0.7]]
p,q
2.4
2.2
x
-4 -2 2 4
1.8
1.6
1.4
1.2
Graphics
Fig.52 Plot of the parameters p (red) and q (yellow) as a function of x for Co2+ ion in
the trigonal field. k = 0.8, 0.9, 1.0, and 1.0.
p2
Parameter − 1 as a function of x
q2
PlotA EvaluateATableA p@x, kD2 ê q@x, kD2 − 1, 8k, 0.8, 1.1, 0.1<EE,
k = 0.8, 0.9, 1.0, 1.1
89
Hp2 êq2 L−1
4
x
-4 -2 2 4
Graphics
Fig.53 Plot of p2/q2 – 1 as a function of x for Co2+ ion in the trigonal field. k =
0.8, 0.9, 1.0, and 1.0.
14. CONCLUSION
We have shown that the spin Hamiltonian well accurately describes the magnetic
properties of magnetic ions in the crystal field. When these magnetic ions are
magnetically coupled through exchange interactions, they may undergo a second-order
phase transition at a critical temperature [Curie temperature (ferromagnet) or a Néel
temperature (antiferromagnet)]. Spins are ordered below the critical temperature. The
static critical exponents depend on the dimensionality (d = 3, 2,) of the system and the
symmetry dimensionality of the order parameter [Ising (n = 1), XY (n = 2), and
Heisenberg (n = 3)]. The spin dimensionality can be determined from the spin
Hamiltonian.21 In two dimensions, there are phase transitions for the Ising and XY spin
symmetry. For example, K2CoF4 and Rb2CoF4 are two-dimensional (2D) Ising-like
antiferromagnet (fictitious spin 1/2), and K2MnF4 is a 2D Heisenberg antiferromagnet. In
contrast, K2CuF4 (spin 1/2) is a 2D XY-like ferromagnet.22 For the 2D Ising model, there
is an exact Onsager solution,23 while for the 2D XYmodel, the system is predicted to
show a Kosterlitz-Thouless (KT) transition.24 The spins form vortices at low temperatures.
and below the KT transition temperature the vortices become bound. The critical
exponents of critical behaviors of the spin systems with short range interactions depends
only on the dimensionality d and the spin dimensionality n (the universality).
Since early 1980’s, we have been studying the magnetic phase transitions of quasi 2D
spin systems such as stage-2 CoCl2-, NiCl2-, CuCl2-, FeCl3-, MnCl2-, and CrCl3- graphite
intercalation compounds.25-31 We find that the magnetic properties of these compounds
are well described by the appropriate spin Hamiltonians of Co2+, Ni2+, Cu2+, Fe3+, Mn2+,
and Cr3+ in the trigonal crystal field.
REFERENCES
1. C. Kittel, Introduction to Solid State Physics, seventh edition (John Wiley & Sons,
Inc., New York, 1996).
2. J.H. Van Vleck, Nobel Lecture (December 8, 1977), Quantum Mechanics The
Key to Understanding Magnetism.
3. R.M. White, Quantum Theory of Magnetism (Springer-Verlag, New York,2007).
4. K. Yosida, Theory of Magnetism (Springer-Verlag, Berlin, 1996).
5. J. Kanamori, Magnetism (Baifukan, Tokyo, 1969, in Japanese).
6. M. Date, Electron Spin Resonance (Baifukan, Tokyo, 1978, in Japanese).
90
7. L. Hoddeson, E. Braun, J. Teichmann, and S. Weart, Out of the Crystal Maze
(Oxford University Press, New York, 1992).
8. J.H. Van Vleck, Theory of Electric and Magnetic Susceptibilities (Oxford
University Press, Oxford,1932).
9. E.C. Stoner, Magnetism and Matter (Methuen, London, 1934).
10. J.S. Griffith, The Theory of Transition-Metal Ions (Cambridge University Press,
Cambridge, 1961).
11. A. Abraham and B. Bleaney, Electron Paramagnetic Resonance of Transition
Ions (Claendon Press, Oxford, 1970).
12. S. Sugano, T. Tanabe, and H. Kamimura, Multiplets of Transition-Metal Ions in
Crystals (Academic Press, New York, 1993).
13. T. Inui, Y. Tanabe, and Y. Onodera, Group Theory and Its Applications in
Physics (Springer-Verlag, Berlin, 1990).
14. C.J. Ballhausen, Introduction to Ligand Field Theory (McGraw-Hill Book
Company, Inc., New York, 1962).
15. K.W.H. Stevens, in Magnetism Vol.1 p1-22, edited by G.T. Rado and H. Suhl
(Academic Press, New York 1963).
16. A. Abraham and M. H. L. Pryce, Proc. Roy. Soc. (London) A205, 135 (1951).
17. B. Bleaney and K.W.H. Stevens, Reports on Progress in Physics 16, 108 (1953).
18. K. Inomata and T. Oguchi, J. Phys. Soc. Jpn. 23, 765 (1967).
19. M.E. Lines, Phys. Rev. 131, 546 (1963).
20. T. Oguchi, J. Phys. Soc. Jpn. 20, 2236 (1965).
21. M.F. Collins, Magnetic Critical Scattering (Oxford University Press, New York,
1989).
22. L.J. de Jongh and A.R. Miedema, Advances in Physics 50, 947 (2001).
23. L. Onsager, Phys. Rev. 65, 117 (1944).
24. J.M. Kosterlitz and D.J. Thouless, J. Phys. C 6, 1181 (1973).
25. T. Enoki, M. Suzuki, and M. Endo, Graphite Intercalation Compounds and
Applications (Oxford University Press, Oxford, 2003). Chapter 7. (review).
26. M. Suzuki and I.S. Suzuki, Phys. Rev. B 57, 10674 (1998). (stage-2 CrCl3 GIC).
27. D.G. Wiesler, M. Suzuki, I.S. Suzuki, and N. Rosov, Phys. Rev. B. 55, 6382
(1997). (stage-2 MnCl2 GIC).
28. M. Suzuki and I.S. Suzuki, Phys. Rev. B 58, 371 (1998). (stage-2 FeCl3 GIC).
29. M. Suzuki and I.S. Suzuki, Phys. Rev. B 58, 840 (1998). (stage-2 CoCl2 GIC).
30. I.S. Suzuki and M. Suzuki, J. Phys. Cond. Matter 10, 5399 (1998). (stage-2 NiCl2
GIC).
31. M. Suzuki, I.S. Suzuki, C.R. Burr, D.G. Wiesler, N. Rosov, and K. Koga, Phys.
Rev. B 50, 9188 (1994). (stage-2 CuCl2 GIC).
Appendix
91
where Hˆ 0 is an unprtturbed Hamiltonian and Hˆ 1 is the perturbation. ψ n
( 0)
is the
eigenket of Hˆ 0 with an eigenvalue En (non-degenerate case).
(0)
Hˆ ψ =E ψ
(0) (0) (0)
0 n n . n (A.2)
Then the perturbation energy and the new eigenket are given by
ψ k ( 0 ) Hˆ 1 ψ n ( 0 ) ψ n ( 0) Hˆ 1 ψ k ( 0 )
En = En + λ ψ n H 1 ψ n + ∑
( 0) ( 0) ˆ ( 0)
(0) ( 0)
+ ....
k ≠n ( En − Ek )
(A.3)
(0) ˆ
ψ k H1 ψ n (0)
+ ∑∑ ψ k
(0) k 1 l l 1 n
( 0) ( 0) ( 0) (0)
+ ...
l ≠n k ≠n ( E n − E k )( E n − E l )
ˆ σˆ Θ
Θ ˆ −1 = −σˆ . (A.10)
z z
92
(ii) ψ and ψ~ = Θ ˆ ψ are orthogonal, since ψ~ ψ = (− η *C η *C )⎛⎜ C+ ⎞⎟ = 0
− + ⎜ ⎟
⎝ C− ⎠
We now consider the system with N electrons. N is an even or an odd integer number.
For convenience we use η = 1. In general, the N-spins state Φ is described by
Φ = u1 + 1 + 2
+ 3 + 4 .... + N −1
+ N
+ u2 + 1 + 2
+ 3 + 4 ..... + N −1
− N
.(A.11)
+ u3 + 1 + 2
+ 3 + 4 .... − N −1
− N
+ ......... + u2N − 1 − 2
− 3 − 4 .... − N −1
− N
93