., 2015.
Original Russian Text L.N. Savushkin, 2015, published in Fizika Elementarnykh Chastits i Atomnogo Yadra, 2015, Vol. 46, No. 6.
1. BASIC CONCEPTS
Relativistic nuclear theory is an elegant and power
ful tool for describing nuclear properties, and in this
approach it is defined as a relativistic and essentially
nonlinear Fermi system with a pronounced isovector
structure [15].
A relativistic description of nuclei was developed in
line with successes of meson theory, in particular the
meson theory of the NN interaction.
Now lowenergy nuclear physics is completely rel
ativized, and the relativistic formalism reproduces a
wide range of physical phenomena (see below).
According to Yukawas idea, the atomic nucleus
should be treated as a set of nucleons affected by
meson fields of different nature, i.e., meson fields with
different spacetime transformation properties [14].
The fields to be considered are scalar S, vector V, pseu
doscalar P, and other meson fields including both isos
calar and isovector fields (meaning the properties of
fields under Lorentz transformations). For this reason
1
ponents, namely,
, , ,
= i [ , ], 5, 5 ,.
2
(1)
859
860
SAVUSHKIN
JG, JP
Meson field
designation
pseudoscalarisovector
139
1, 0+
scalarisoscalar
4001200
0+, 0+
Nucleon
coupling constant
g ps*
f pv
g**
vectorisoscalar
782
vectorisovecto
770
a0 scalarisovector
980
pseudoscalarisoscalar
548
a1 axialisovector
0, 1
1+, 1
1260
g***
= 0, 1, 2, 3
g***
= 0, 1, 2, 3
1, 0+
g a0
0+, 0
ga
,1
* Interaction of the pion and nucleon fields can be presented either in the pseudoscalar form (with the coupling constant g) or in the
pseudovector form (with the coupling constant f), see below.
** OBEP models involve a phenomenological element. The problem in this approach is the scalarisoscalar boson ( or ), an impor
tant component of the models. Empirical evidence for this meson with the required parameter is contradictory so far.
*** The vector fields and interact with the nucleon field either via the direct coupling (with the coupling constants
g and g, respectively) or via the tensor coupling (with the coupling constants f and f, respectively).
Yc(r ) =
g
2
2
4 m
2
(2)
e mr e r 2 m2
r ,
1 +
r
2
r
i = p + M + S(r ) 3 N Z S (r )
A
t
+ V (r ) 3 N Z V (r )
f
+ i 1 3 N Z V (r ) ,
g 2M
A
(3)
Vol. 46
No. 6
2015
861
Table 2. Mesons and fields involved in Eqs. (3) and (4) and their parameters: coupling constants (their numerical values are
given below) and masses (in MeV)
g
(4001200)
g a0
g
g
(782)
V(r) 0+, 1isovectorisoscalar potential (repulsive)
V (r) 1 , 1 vectorisovector potential (contributes to isovector structure of nuclear potential) (770)
M *(r ) = M + S(r ) 3 N Z S (r ),
A
(4)
2(r ) = 2M + S 3 N Z S (r )
A
1 + 3
N
Z
V + 3
V (r )
C + ,
2
A
(5)
Vol. 46
(6)
H D = , = ,
where HD is given by Eq. (3). The notion of (r) is dis
cussed at length in the Section 3.
The order of magnitude of M*(r) can be estimated
as follows:
Write Eq. (3) for N = Z, which results in an equa
tion involving only S and V.
Take into consideration that the combination
S + V determines the depth of the nuclear potential and
S V determines the spinorbit potential; both are
known from experiment.
On this basis, it is possible to obtain the fields
S ~ 420 MeV and V ~ +330 MeV of the effective
mass 0.6M. The calculation is detailed below.
Here we only note that due to large values of the fields
S and V in comparison with M, the relativistic formu
lation of nuclear theory has to be used. The correct
magnitude and sign of the singleparticle spinorbit
potential in the nucleus was the first and most impor
tant result of this approach obtained in [911], which
resolved the problem that had remained unsolved
since the development of the shell model until 1973.
2. LINEAR VERSION
OF RELATIVISTIC NUCLEAR THEORY
Let us introduce the notations
g (r ) = S(r ), g a0 a0(r ) = S (r ),
(7)
= ( , ) = 0,
i
No. 6
2015
862
SAVUSHKIN
= 0(, S, (V ), , A )
+ int (linear in boson fields).
(8)
int = g
+ g a0
a0
scalar isovector
density
+ g
+ g
vector
density
scalar
density
vector isovector
density
(9)
2M tensor isovector
density
or i
g
.
5
5
m pseudovector
pseudoscalar isovector
isovector
density
density
(10)
It contains the static part Vc(r) (see (2) and Appendix B),
and its last five terms are relativistic corrections of
order v2/c2 with respect to Vc(r) (see Appendix B). In
Eq. (10), only Vc(r) is parameterized, and all relativis
tic corrections are a uniquely defined structure of
those relativistic equations which were used to obtain
those corrections. The relativistic corrections do not
involve any additional adjustable parameters. Only
Vc(r) is parameterized, and the adjustable parameters
are the coupling constants (g2, f) of some mesons with
nucleons, masses (m) of some mesons, and regulariza
tion parameters (). It is for this reason that the num
ber of adjustable parameters used in the OBEP scheme
is as few as 5 to 10. Note that the twoparticle spin
orbit LS and tensor forces are the relativistic correc
tions (together with V, V, V) of order v2/c2 (where
v is the velocity of nucleons in a nucleus, and c is the
speed of light) in the OBEP scheme and also the com
ponents of theory that do not contain additional
adjustable parameters.
Note the contributions from different mesons to
the particular components of the nucleonnucleon
OBEP (10) (see Appendix C). It is also worth noting
that (i) the scalarisoscalar meson provides (strong)
Scalar mesons ()
Vector mesons (, )
Vc V LS V V ,
Vc V LS V V ,
V = 0 VT = 0
V VT
Pseudoscalar mesons ()
(11)
Vc = 0 V LS = 0 V = 0 V = 0.
V
VT
In the OBEP scheme, the contribution to the two
particle spinorbit forces comes from the , , and
mesons while the tensor forces are determined by the
Vol. 46
No. 6
2015
V LS , VT .
(12)
H
U S 0 = 1 2 1 d [V (r ) S(r )] + 3 N Z
A
4M r dr
(14)
f
S (r ) 1 + 2 V (r ) .
g
d
(15)
U SH0 = C H 1 ,
r dr
where CH is the constant calculated within this proce
dure from the Hartree approximation OBEPs, is the
nuclear density, is the angular momentum operator,
and is the spin operator; this constant depends only
on the parameters (g2, f, ) of the OBEP used for the
calculation [9, 10].
H
(S )
E LS (H ) = E LS
(H )
(16)
(V )
S
(V )
(H ) + E LS
(H ) + E LS
(H ).
+ E LS
In [10], the effect of Fock (exchange) matrix ele
ments on the spinorbit splitting was taken into
account. In this section we consider the results of
those calculations without applying selfconsistency
(the complete selfconsistent problem is discussed in
the next sections). In [10], important relations were
established for doubly magic (and doubly magic a
nucleon) nuclei (in the approximation of shortrange
twoparticle spinorbit forces)
3 PVS
863
No. 6
E S (HF ) = 1 2 + N
2
A
E S (HF ) = 1 2 + Z
2
A
E V (HF ) = 1 + 3 N
2A
E V (HF ) = 1 + 3 Z
2A
2015
864
SAVUSHKIN
Table 3. Spinorbit splittings obtained within the HartreeFock (HF) approximation for different OBEPs. Spinorbit
splittings in the Hartree approximation for the same OBEPs are given in parentheses [10]. All values are in MeV; results a
d correspond to different OBEP models. Note that H and HF results in each column of the table are compared for the same
set of OBEP parameters (used in the corresponding column)
a [16]
b [16, 17]
Model II
c [16, 17]
Model III
7.65
9.45
13.8
d [16, 17]
10.1
E1sof ( 41Ca)
6.50
(9.32)
3.30
(4.25)
1.82
(5.56)
2.25
(5.98)
2.37
E 2sop ( 41Ca)
2.00
(2.22)
3.53
(1.01)
2.62
(1.32)
2.88
(1.41)
2.94
2.47
(2.22)
11.2
(1.16)
8.33
(1.43)
9.22
(1.52)
9.37
(7.08)
2.03
(3.70)
1.49
(4.59)
1.66
(4.82)
1.69
(1.28)
(0.667)
(0.830)
(0.874)
E1(iso)(209 Pb)
(so) 209
E3d (
4.57
0.98
Pb)
) (
) (
E nV (H )
(26)
(27)
E V (HF ) =
Exp
E Vp (H )
(9 + 12 f g )
(2 + 4 f g )
(28)
Vol. 46
No. 6
2015
865
Table 4. Contributions from different OBEP components to the spinorbit splittings (in MeV) for the 208Pb nucleus,
Model II [1, 18] and Model III [1, 18]
Splittings (in MeV)
E (1i)
Forces
Model II
E (2 g )
Model III
Model II
E (3d )
Model III
Model II
Model III
3.02
2.54
1.23
1.02
0.548
0.456
2
2
1 2 ( p Yc + Yc p )
2M
0.546
0.571
0.226
0.232
0.100
0.103
1.18
1.17
0.501
0.493
0.22
0.217
1.42
1.45
0.568
0.576
0.253
0.258
3.55
3.62
1.42
1.44
0.633
0.646
Spinorbit forces
8.33
9.22
2.62
2.88
1.49
1.66
Total splitting
3.81
5.11
0.813
1.26
0.682
0.93
2
1 2 ( Y c )
4M
1 ( 2Y )
c 1
2
12M 2
Tensor forces
Experiment
4.57
2.47
ing into the operator that has the following form (this
result is obtained in the shortrange tensor force
approximation):
(29)
U soT ,n = 1 ( J n + J p ) ,
r
(30)
U soT , p = 1 ( J p + J n ) ,
r
where the superscripts/subscripts n and p correspond
to neutrons and protons respectively, while the spin
orbit densities are defined as
J (r ) = 1 3
4r
(2 j
+ 1)
(31)
2
3
j( j + 1) ( + 1) R(r ),
4
where summation is taken over the occupied proton
states and (or) occupied neutron states; note that for
spinsaturated nuclei J(r) = 0 (on the assumption that
the nucleus features spin symmetry, i.e., that the wave
functions R for two states of the spinorbit doublet
are identical), and in (29) and (30) are constants
that are calculated via parameters (of OBEP tensor
forces, for example).
In the HF method, J(r) for spinsaturated nuclei is
zero (accurate to the equality of the radial wave func
tions for the spinorbit doublet Rj = 1/2 Rj = + 1/2
(spin symmetry); it is known that this symmetry works
well for finite nuclei). In [1, 18], contributions to the
spinorbit splittings (of singleparticle states) from
twoparticle forces (associated with the OBEP) of any
6 In
d
(32)
1 + 1 J (r) ,
r dr r
where (r) is the nuclear density, and J(r) is the spin
orbit density defined by (31). It should be stressed that
in the mean field (Hartree) approximation the compo
nent J(r) does not appear, at least in the v2/c2 limit.
Note that the role of the Fock (exchange) matrix ele
ments in calculations of spinorbit splittings is very
important (and different for spinsaturated and spin
unsaturated nuclei). This procedure allowed us to use
the potential radius technique for calculating matrix
elements of various OBEP components. In [1, 18],
considering matrix elements of different OBEP com
ponents, we used the following shortrange approxi
mation: matrix elements of different OBEP compo
U
5 Here
0.98
No. 6
LS
866
SAVUSHKIN
27
Mg
NNforce
Si
Total
11.4
7.6
10.7
14.5
12.3
7.7
11.1
15.7
10.9
2.4
3.8
12.3
11.8
2.4
4.0
13.4
V i (r, p; 1, 2; 1, 2 ); (c i = LS;T , ),
the spinorbit splitting of the odd nucleon state with
the quantum numbers
(33)
i
i
= E (nn = 1) E (nn = + 1),
where Ei is the contribution from the potential V i to
the singleparticle energy of the nucleon. The calcula
tions were performed analytically with the wave func
tions of proper symmetry.
The role of the tensor forces is effectively demon
strated in Tables 4 and 5. In particular, Table 5 presents
the spinorbit splitting of the neutron level |Nn =
|202 in the 25Mg and 27Si nuclei (in the spherical
basis the spinorbit splitting of this level can be com
pared with the splitting of the 1d1/21d3/2 neutron state
in the WoodsSaxon potential with A ~ 27). As is evi
dent from Table 5, the tensor force contributions are
appreciably higher for 25Mg (compared to 27Si)
because this nucleus is spin unsaturated in both neu
trons and protons while 27Si is spin unsaturated only in
neutrons. The tensor force contribution to the spin
orbit splitting emerges through the Fock matrix ele
ments and spinunsaturated shells, considerably
depending on the number of these shells. Twoparticle
spinorbit forces are a specialized component of the
OBEP (and any other nucleonnucleon potentials):
in spinsaturated (SS) nuclei the total spinorbit split
ting is completely determined by the twoparticle
spinorbit force (this statement strictly holds under
spin symmetry, when radial wave functions of two
states of any spinorbit doublet are identical, which is
valid to a high degree of accuracy for real nuclei).
Note that two spinorbit coupling operators (see
(31)) depend on A (in spherical basis) as follows:
(1 r )(d dr ) A 2 / 3,
spinorbit splitting
1 r J (r ) depends on the number of spinunsaturated nuclear shells.
(34)
Vol. 46
No. 6
2015
867
3. WALECKA MODEL,
RELATIVISTIC SATURATION MECHANISM,
AND PARAMETERS OF THE SKYRME
HARTREEFOCK METHOD
Saturation [22, 23] is an important feature of mod
ern relativistic approaches resulting from consider
ation of the small component of the nucleon wave
function (which leads to a decrease in the mathemati
cal expectation of the singleparticle operator for
kinetic energy).
In 1974, Walecka [24, 25] proposed an elegant
method for constructing a theory using the Lagrangian
W that involved nucleons and two meson fields, sca
lar and vector
(44)
= ,
where is the large component of the bispinor, is
its small component, and comprises a complete set
of the quantum numbers of the state under consider
ation; S and V can thus be presented as
W = (i M g g )
2 2
+ 1 1 m 1 + 1 m ,
2
2
4
2
where for we have
= .
(35)
(36)
(37)
( + m2 ) = g ,
(38)
(39)
( + m2 ) = g .
In the static case, terms with a time derivative disap
pear; if there is invariance under time reversal, the spa
tial components of the meson field vanish, = 0. In
this case we have
( m2 ) = g ,
(
m2 )0
(40)
= g ,
0
(41)
0
S =
V =
V =
(43)
(45)
+ .
(46)
(49)
while bearing in mind that M ( is a nonrelativistic
eigenvalue. Thus, from (48) we have
1 p .
2(r )
(50)
(42)
No. 6
1 p = p 1 p + 1 d 1 , (51)
2(r )
2(r )
r dr 2(r )
2015
868
SAVUSHKIN
= p 1 p + V (r )
2(r )
(52)
1
1
d
+ S(r ) +
.
r dr 2(r )
d 1 = d V ,
LS
dr 2(r )
dr
(54)
and
1 = 1 + (r ).
(55)
LS V
2(r ) 2
Integrating (54), we used the boundary conditions
(r) = M and (r) = 0 for r , which are quite obvi
ous: the nuclear density becomes zero outside the
nucleus while the effective mass becomes equal to the
free nucleon mass. The HFS also leads to (55), but it is
an exact equation within the HFS, while in the relativ
istic case a more accurate result can be obtained (see
[4] and references therein). From Eqs. (54) and (55)
we can obtain
= 2M + S V = 0. 6.
(56)
V S = 0 . 8M = 750 MeV,
(57)
M
2M
Using this result, we obtain
V + S = M U sh 90 MeV.
(58)
(59)
= i (1 + 4 . 123(r )) p .
2M
(60)
g 2
,
2 S
m
(61)
g 2
V ,
m2
(62)
V = g 0 = +
that is, the fact that meson fields are directly expressed
in terms of the corresponding densities. The density V
is expressed in terms of the Fermi momentum in a
conventional way
V =
2 pF
.
3 2
Vol. 46
(63)
No. 6
2015
/M V, fm3
1.0
0.5
0.8
0.4
869
U, UII, MeV
196
U
/M
98
0.6
0.3
0.4
0.2
0.2
0.1
UII
0.1
0.1
0.2
0.3
0.4
S, fm3
0.2
0.3
0.4
0.5
V, fm3
98
S =
p F
4M * d 3p
(2)
p +M
2
= M2*
p + pF2 + M *2
pF pF2 + M *2 M *2 ln F
, (64)
M*
2
g
M * = M + S = M 2 S .
m
Considering (63), Eq. (64) is a transcendental equa
tion that determines dependence of S on V and can
be solved analytically for V 0 and V . We
obtain
V 0,
V ,
S =
2
2
(Mm ) (g ), V .
2
(65)
g 2
g 2
2 V
2 S
m
m
g2 g2
(67)
U I = 2 2 ,
m m
i.e., it has no minimum (as a function of density) at all
and infinitely increases, which gives rise to a collapse
in accordance with Wigners wellknown result [29]
(see also references therein). Thus, in the relativistic
nuclear shell model the saturation property has a pure
relativistic nature. This property occurs only because
in the nucleus there are two fields (S and V) with the
transformation properties of the relativistic scalar and
the relativistic vector and also because the nucleon
wave function has a small component. In the Har
treeFock nonrelativistic theory with Skyrme forces
(HFS), the saturation property is ensured by the para
bolic dependence of the nuclear central potential on
the density of the following form (it is also presented in
Fig. 2):
(68)
U II = a + b 2.
Requiring that the function UII have the minimum at
the same point as the relativistic nuclear central
potential (66) calculated with the parameters of the
Walecka model
(66)
Vol. 46
No. 6
2
2
CV2 = M2 g 2 = 195 . 7, C S2 = M2 g S2 = 266 . 9,
m
mS
2015
(69)
870
SAVUSHKIN
Table 6. Comparison of the relativistic Hartree approximation [26] parameters with the Skyrme II set of parameters [28]
(MeV fm5)
a (MeV fm3)
b (MeV fm6)
c (MeV fm5)
e (MeV fm5)
85.5
101
816
783
1889
1750
200
169
125
101
we obtain
(70)
V (r ) =
g 2
(r ) + 12 V (r ),
2 V
m
m
(71)
g2
S(r ) = 2 S (r ) + 12 S(r ),
m
m
(72)
1U = c(r ),
c=
g 2
4
m
g 2
4
m
2U = e(r ), (r ) =
(73)
2
g 2 g 2
5
1
e=
+ = 125 MeV fm ,
2 m2 m2
where (r) is the kinetic energy density. Thus we have
obtained HFS results both for the effective nucleon
mass (formula (55)) and for the nuclear potential
( )
LS d
, (74)
r dr
and calculated all HFS parameters for N = Z nuclei,
beginning with the parameters of the relativistic shell
model. In Table 6 these parameters are compared with
the Skyrme II set of parameters [28], one of the most
often used in HFS calculations. From the table and
Eq. (74) it is evident that the HFS method is actually
an imitation of the nuclear relativistic method in non
relativistic terms.
Note that the relativistic mean field approximation
RMFA and the HFS method have much in common
but are different in some detail.
As to the general differences of the two approaches,
it should be borne in mind that, for example, the spin
orbit interaction operator arises quite naturally in the
U HFS = a + b2 c + e(r ) +
4. ROLE OF NONLINEARITY
IN THE RELATIVISTIC MEAN
FIELD APPROXIMATION
In its initial form, the relativistic mean field approx
imation involves only isoscalar mesons ( and ) and
one vectorisovector meson (the latter is introduced
to describe properties of isotopes and asymmetric
nuclear matter). However, this model only qualita
tively describes ground states of atomic nuclei and
nuclear matter. The compression modulus of nuclear
matter and properties of nuclear surface cannot be
reproduced within this simple version ([15] and ref
erences therein). In [38, 39], the authors investigated
the role of the selfinteraction of the scalar meson field
in a nuclear medium like (3 + 4). Selfinteraction of
meson fields is associated with multiparticle forces.
The compression modulus and surface properties of
nuclei are well reproduced in the model with the scalar
field selfinteraction, but binding energies and radii of
nuclei cannot be simultaneously well reproduced
without these components (see [5] and references
therein). Inclusion of the 4 selfinteraction was the
next step made by Boguta and Bodmer [38, 40] (cubic
selfinteraction in the Lagrangian ~3 cannot be used
for the reason of parity).
Vol. 46
No. 6
2015
1 g 4 1 g 5 2
2
2
(which are substantiated by chiral symmetry). In
[4042] this type of Lagrangian was used together with
the (3 + 4) terms for investigating properties of finite
nuclei and nuclear matter. The investigation showed
that the model with the interaction described
binding energies and charge radii of nuclei in a wide
range of A. In addition to K = 265 MeV (where K is the
compression modulus of nuclear matter), the model
successfully reproduces resonance energies of the
breathing mode of the monopole isoscalar giant reso
nance. The behavior of the equation of state of nuclear
and neutron matter at high densities is appreciably
softer than in the case of only the scalar nonlinear
model (with nonlinearities of only the scalar field).
Similar calculations were performed for deformed
nuclei.
At present, nonlinear terms are phenomenologi
cally introduced in the mean field approximation
scheme, and their role is similar to that of density
dependent forces in the nonrelativistic formalism.
Properties of nuclei were extensively investigated
within this approach with different types of nonlinear
ity, and it was demonstrated that an atomic nucleus
was a pure nonlinear relativistic system [15].
An alternative possibility of introducing nonlinear
ity in relativistic theory of the nuclear structure is
through the density dependence of mesonnucleon
coupling constants. This was first done in [43] and
later in [4446]. This method is now referred to as the
relativistic densitydependent Hartree method
(RDDH), later extended to be used in the framework
of the relativistic HartreeFock method (see below).
The relativistic Hartree method involves the fol
lowing equations that are solved using the selfconsis
tent procedure:
1. The Dirac equation for nucleon wave functions
(r).
2. The KleinGordon equation for meson fields:
scalarisoscalar field S(r),
vectorisoscalar field V(r),
vectorisovector field V (r).
3. The KleinGordon equation for the Coulomb
field A0(r).
The sources of the meson fields are
(a) Scalar density S(r).
(b) Vector (baryon) isoscalar density V(r).
(c) Vectorisovector density 3(r).
PHYSICS OF PARTICLES AND NUCLEI
Vol. 46
871
v ,
(75a)
v ,
(75b)
S =
V =
3 =
v
2
0 3 ,
(75c)
p =
v
2
1 + 3
.
2
(75d)
E E F
(76)
v 2 = 1 1
,
2
(E E F )2 + 2
for the quantum state with singleparticle energy E.
2
(77)
,
2
2
E
E
(
)
F
where N is the number of protons or neutrons. BCS
calculations allow considering singleparticle states up
to one more shell above the Fermi energy [49]. For
most nuclei not very close to the drip line this method
presents no problems. However, when approaching
the drip line the singleparticle states above the Fermi
level become unbound. In this case, for mere practical
reasons, contributions only from bound states are
taken into account in (76). Also, the pairing energy
=N =1
2
E E F
E pair =
u v
(78)
No. 6
2015
872
SAVUSHKIN
E ZPE =
2
F Ptotal
F
2M total
(79)
is used, where
2
M total = AM , Ptotal
=
p .
2
i
(80)
E ZPE =
2
F Ptotal
F
(81)
, MeV.
2M total
This simple form is used in some sets of parameters
for heavy nuclei because the nuclear binding energy is
high while the correction for the centerofmass
motion is small. For light nuclei, the corrections
become appreciably large, and the more appropriate
choice is |F = |FH (relativistic Hartree method). Pair
ing correlations play a very important role in nuclear
physics, in particular in relativistic nuclear theory.
This topic is discussed at greater length within the rel
ativistic approach in [4, 5, 5052].
5. SPIN AND PSEUDOSPIN SYMMETRY
IN FINITE NUCLEI
The condition S + V S V is one of the main fea
tures of all reasonable modern relativistic models that
ensure, among other things, the correct value and sign
of the spinorbit force in the nucleus. At the same
time, inadequate use of this very condition in some of
the recent publications has led to inadequate treat
ment of pseudospin symmetry (PSS) in finite nuclei
an interesting phenomenon in nuclear physics closely
related to spin symmetry. The notion of PSS was intro
duced in nuclear physics about 50 years ago [5361],
and it is still one of the central topics in the nuclear
structure research. A lot of interesting results have
been obtained in the past 20 years [6280].
In the central potential the singlenucleon wave
functions have the form [4, 5]
iG(r) jem
(82)
= 1
= ,
r F(r)( n) jem
where G(r) and F(r) are the large and small radial
components of the wave function, jem is the spin
r
angular function, n = ,
and is a quantum number
r
1
of relativistic theory defined as = j + for j =
2
1/2. Apart from the eigenvalue (see (47), (48)),
(83)
Vol. 46
No. 6
2015
= ',
(84)
(86)
Considering (85) and (86), the Dirac equation in
spherical coordinates can be written as
d G(r ) = G(r ) + BF (r ),
(87)
dr
r
d F (r ) = AG(r ) + F (r ),
(88)
dr
r
where G(r) and F(r) are the large and small compo
nents of the Dirac spinor, and = E M is the single
particle energy of the nucleon with mass M and rela
tivistic energy E. From Eq. (88) we easily obtain
(89)
G(r ) = 1 d F (r ) + F (r ) .
A(r ) dr
r
The solutions G(r) and F(r) can be found as solutions
of the equations
( + 1) + AB G = 0,
(90)
G" + B' G' + +
B G r
r2
'(' + 1) + AB F = 0,
(91)
F " + A' F ' +
A F r
r2
where A(r) and B(r) are represented by Eqs. (85) and
(86) while the quantity
( )
( )
AB = 2MA + 2 V + (S 2 V 2 ) 2,
(92)
Vol. 46
873
1
d(S V ) .
(94)
+ 2M + S V r dr
The term in (91) (which will be referred to as the
F term) is determined by the factor A'/A, which
will be referred to as the pseudospinorbit potential
(PSOP)
d(S + V ) ;
1
(95)
S V r
dr
the latter operator is often treated as the spinorbit
potential of the small component.
The solution of Eq. (90) with the same number of
nodes nr of the large component G(r) and the same
value of forms a spin doublet (SD) while the solution
of Eq. (91) with the same number of nodes n r of the
No. 6
2015
874
SAVUSHKIN
( + 1)
A'
(ii) The F terms is small
2
Ar
r
[67, 68].
(iii) Various contributions to energy (91) partially
compensate for one another [69, 70].
PSS and S + V. As is shown in [64], in the limit
S + V = 0 two pseudospin partners a and b have iden
tical energies, and the functions Fa and Fb are identical
up to a phase
(96)
a = b and Fa = Fb.
Let us designate this particular type of pseudospin
symmetry as PSS* (actually, it is this particular type of
pseudospin symmetry that is considered in the above
mentioned works of Jinocchio). Note that what we
designate as PSS is the case where a b (without the
requirement that Fa Fb). The condition S + V = 0
allows no bound states (except for models with too
unrealistic S V values). In real nuclei, (S + V) is
small (+50 MeV) (but not zero). On this basis, it is
stated in [62, 63] that we can expect approximate
PSS* (i.e., a b and Fa Fb). In [74] it is shown that
neither PSS nor PSS* necessarily improve as |S + V|
decreases. On the other hand, all pseudospin doublets
that become degenerate (a = b) for a given S + V
value are split if S + V varies, in particular if it
decreases. Also, Fa becomes appreciably different
from Fb as S + V decreases, when a or b become close
to the continuum. These arguments allow stating that
neither PSS nor PSS* can be based on the assumption
of smallness of S + V.
SS and S V. In the limit S V = 0, two states of
the spin doublet (SD) have identical energy (i.e., there
is exact spin symmetry); moreover, the corresponding
two functions G are identical (we designate this special
type of SS as SS*). In real nuclei |S V| is large, and
therefore neither SS nor SS* could be expected.
Indeed, nuclei have large spinorbit splittings, but the
G functions of the spin partners are very similar (actu
ally, much more similar than the small components F
for the pseudospin partners) [7376]. Furthermore, if
Vol. 46
No. 6
2015
S(r ) = g (r),
1 + 3
(98)
A0(r),
2
where the contributions from the vectorisovector
3
0 ( r ) and Coulomb A0 components are taken into
account. In this section we do not consider the contri
bution from the tensor coupling of the meson field
with the nucleon field (this coupling was introduced
by Pauli in electrodynamics to reproduce the anoma
lous part of the electron magnetic moment). The cor
responding interaction appears in Eq. (3) (term
~f/g). Its role in the kink effect will be considered
below.
Note also that the pion field and the spatial compo
nents of the vector meson fields ( and ) become zero
in the problem under consideration.
3
(r) =
B
12
12
(99)
= 1 + (r),
2M
(97)
V (r ) = g 0(r) + 3 g 30(r) + e
9 A kink in the behavior of charge radii was also observed in Kr, Rb,
875
No. 6
2015
(100)
876
SAVUSHKIN
rc, fm
5.53
NL3
NLSH
L
EXP
5.52
5.51
5.50
5.49
208
206
210
Vcent = S + V + S V + V + Vcent ,
M
2M
(101)
2
V cent = 1 1 W + 1 W + 1 W ' ,
r
2M 4
2
(102)
S' V '
(103)
,
2M + + S V
(104)
VS 0(r) = 1 2W S.
2M r
The potential Vcent(r) depends on energy, and we con
sequently can introduce effective mass M defined by
W =
10
the equation
M = 1 dVcent 1 V .
(105)
M
d
M
Writing the Dirac equation in this form (100), we can
separately investigate the influence of different com
ponents in this equation on the kink effect (e.g., we
can investigate whether particular components of this
form are of relativistic origin or not).
In [88] the kink effect calculations were performed
for the Pb isotope chain within the standard relativistic
Hartree method using a linear model (L) [94] and two
nonlinear models with the scalar field selfinteraction
[95, 96].
10
Vol. 46
No. 6
2015
Vol. 46
877
rc, fm
NLSH
5.52
5.51
5.50
c
a
5.49
5.48
d
b
206
208
210
No. 6
2015
878
SAVUSHKIN
2
0(, , , , , A )
= (i M ) 1 m22 + 1 ( )
2
2
(107)
2
1
1
1
+ m F F + m2
2
4
2
2 2
1
1
G G + ( m ) 1 H H ,
4
2
4
where
F ,
(108)
G , H A A.
(109)
f
+ x ig 5
2M
f
+ (1 x ) 5 e 1 (1 + 3 ) A.
m
2
int (mesonnucleon) = g g
g
2M
other hand
11
= 0(, , , , , A )
+ int (mesonnucleon) U NL(mesonmeson).
11 Consideration
(106)
Vol. 46
No. 6
2015
where
2
m*2 = m2 + bg 2 M ( g ) + cg 2 ( g )
U NL = 1 bM (g )3 + 1 c (g )4
3
4
+ dMg g 2 1 e (g )2 g 2
4
2
2
1
f (g ) + 1 m2() 2,
4
2
(110)
i + M + g + g + g
(111)
x ig 5 (1 x ) 5 + = 0 .
m
U NL
*2 = m2 +
m
.
(114)
Vol. 46
( g 0 ) 2 1 2
2
eg ( g 0 ) .
g
(115)
(116)
m*2 = m2 (1 + ) .
(117)
(113)
+ dg M
2
*2 ) = g ,
( + m
879
p2 ( g ) mc
(118)
No. 6
2015
880
SAVUSHKIN
Table 7. Adjustable parameters (m, g , g , g, b , c ), used in our models. For the pion, the value used in the calculation was
2
2
f 4 0 . 076, which corresponds to the N pseudoscalar coupling constant g 4 = 14. The PVa model fully takes into
account the tensor force contribution. In the PVe and PVe' models the tensor force contribution is entirely eliminated using the
method developed in [113, 123, 124]. The PBc' model retains about 1/3 of the pion tensor force (PTF) [113, 123, 124]
Model
PVa
PVe
PVe'
PVc'
m, MeV
g 2 4
g 2 4
g 2 4
b 10 3
c 10 3
443.28
441.57
441.72
441.85
5.3215
4.9822
5.0154
5.0898
10.393
9.3441
9.5096
9.7595
0.72
0.62
0.67
0.69
4.361
5.497
5.260
4.970
7.257
9.308
8.948
8.366
Vol. 46
No. 6
2015
881
(b)
56
0 1g9/2
2p1/2
2p3/2
1f5/2
1f7/2
[28]
7.78
Exp.
[20]
2s1/2
E/A, MeV
Ei , MeV
20
2p1/2
1f5/2
2p3/2
1f7/2
1d3/2
1d5/2
40
PVe
PVe'
PVc'
PVa
7.82
[8]
7.86
1p1/2
1p3/2
Pb
60 1s1/2
7.90
PVa PVe PVe' PVc'
192
Exp.
198
204
210
Fig. 5. (a) Spectrum of Ei levels for the 56Ni nucleus; (b) binding energy per particle E/A of the 208Pb isotope chain.
TRENDS, CONCLUSIONS
Vol. 46
No. 6
2015
882
SAVUSHKIN
A = ( A1, A 2, A 3) = { A i }, i = 1,2,3.
(A.1)
A = g A ,
(A.2)
A B = A0 B0 A B,
(A.3)
x 2 = x x = t 2 x 2.
(A.4)
Vol. 46
No. 6
2015
,
x
.
x x
5 + 5 = 0, ( 5)2 = 1,
(A.18)
(A.5)
5 5 = 0, 5 + 5 = 0,
(A.19)
(A.6)
( 5)+ = 5, [ 5, ] = 0.
(A.20)
0
{ i } = =
, i = 1,2,3,
0
I 0
= 0 = 0 =
,
0 I
883
(A.7)
I ; 5, , 5 , .
(A.21)
(A.8)
16
X =
A,
(A.22)
0
(A.9)
= 0 =
.
0
Since the Dirac Hamiltonian must be Hermitian, we
have + = , + = , and
where
0 1
(A.10)
5 = = i = .
1 0
It should be stressed that some authors use (A.10) mul
tiplied by i as the matrix 5. Next,
x A = 1 Tr(X A ),
4
0 1 2 3
= i [ , ], = , , = 0,1,2,3, (A.11)
2
where 1 is the 2 2 unit matrix, and is the Pauli
2 2 spin matrix, so that
0 1
x = 1 = ,
1 0
(A.12)
0 i
1 0
2
3
y = =
, z = = 0 1 .
i 0
(1)2 = ( 2 )2 = ( 3)2 = 1, ( 0 )2 = 1.
(A.13)
(A.14)
(A.15)
i + i = 2i,
(A.16)
+ = 0.
They all are Hermitian matrices.
(A.17)
Vol. 46
A =1
P = 0PNR,
(A.23)
(A.24)
y 0
T = i
K,
0 y
(A.25)
No. 6
2015
884
SAVUSHKIN
Scalar mesons:
a ( Y c (r ))
,
4
V(r ) = 0, VT (r ) = 0,
2
V c (r ) = Y c (r ) +
(B.1)
(B.2)
s
2 1 dY c
V LS (r ) = 1 a
,
2 r dr
(B.3)
V(r ) = a 2Ycs (r ),
(B.4)
V(r ) = a 2 1 d Ycs (r ).
r dr
Vector mesons:
(B.5)
(B.6)
f
2 V
Yc (r ) ,
g
(B.7)
f dY (r )
V LS (r ) = 3 a 2 1 + 4 1 c
,
2 3 g r dr
(B.8)
f
Vc (r ) = YcV (r ) 1 a 2 2YcV (r ) ,
2 g
2
V(r ) = 1 a 2 1 +
6
dY V (r )
f
VT (r ) = 1 a 2 1 + r d 1 c ,
2
g dr r dr
V(r ) = a
V (r ) = a
YcV (r ),
V
2 1 dY c (r )
dr
(B.9)
(B.12)
(B.13)
(B.11)
Pseudoscalar mesons:
V(r ) = 1 a 2 2YcPS (r ) ,
12
1 dY cPS (r )
2 2
VT (r ) = 1 a r 1 d
12
r dr r
0
mesons Vtens(r) = 0, for mesons V T = 0 and V T 0,
(B.10)
.
(B.14)
.
(B.15)
VS = YcS (r ), VV = YcV (r ), VPS = 0,
and the operator Vtot (see (10)) is presented as a sum of
the static part of (B.15) (independent of velocity) and
the relativistic corrections of order v2/c2. Operator (10)
has a structure such that for each type of meson and for
the given g, f, m, and all relativistic corrections (in
particular, spinorbit and tensor forces) are uniquely
determined by the static part of the corresponding
(r)
(C.2)
S12.
r2
The central part of the NN coupling via the meson is
always attractive. As a result, it leads to simple mono
Vtens = [vT0 + vT 1 2]
Vol. 46
No. 6
2015
885
W (35)
3
4
U = U + U + U () = 1 g 2 + 1 g 3
3
4
2
2
1
1
c3( ) d3( ) ,
4
4
(E.1)
(E.2)
2
2
2
m* (r ) = m + c30,
(E.3)
(E.4)
g
g
= 1 g 2 2 ()3 + 1 g 3 2 ()4 . (D.1)
3 m
4 m
The exchange (Fock) matrix elements were taken
into account for this operator in an exact form using
the Fierz transformation in [110].
Consideration of nonlinearities in the RH and RHF
theories appears to be a very important element of the
selfconsistent description. It is for this reason that the
PHYSICS OF PARTICLES AND NUCLEI
Vol. 46
No. 6
886
SAVUSHKIN
(F.1)
3
4
To this Lagrangian there corresponds the effective
mass of the scalar meson
2
2
2
(F.2)
m* = m + g 2 + g 3 ,
where m is the mass of the free scalar meson. In this
case the scalarisoscalar field is found as a solution
of the equation
2
2
( m* ) = g S
with S denoting the scalar density and
(F.3)
I = g 2 1 3 g 3 1 4 .
(F.4)
6
4
Thus, the equation for the scalar meson field is linear
ized with respect to the field operator , which can be
presented as
= S (x, y)(y)(y)d 4 y,
S (r, r ') =
(F.7)
(i)(r< )(i)(r> )
,
W
(F.8)
(i )
E =1
2
(T + E
i
M ) + 1()d 3r,
(F.12)
= lin + NL,
(F.9)
(G.1)
where
W = ()'(r)(i)(r) ()(r)(i)'(r).
=0 m=
S (r, r ') =
(F.11)
APPENDIX G.
ZERORANGE SELFINTERACTION METHOD
2 1
(r )
(r )
, K
r mW
r mW
(F.6)
S(q) = (m* + q ) ,
where q is the fourmomentum of the meson pre
sented in the same form as in the case of nonlinearities
with m replaced by m * , and m * do not depend on
the momentum. However, they are calculated in a self
consistent way.
For finite nuclei, S(r, r') is presented as
I (m r )
r2
that are regular at the beginning and in the infinity
respectively. Replacement of the Yukawa operator for
the meson with the function S(r, r') leads to the
replacements
(F.5)
()
(i)
g
,
m2
(G.2)
g
g
()3 + 1 g 3 2 ()4, (G.3)
2
4 m
m
1
ZRL
NL = g 2
3
Vol. 46
No. 6
2015
ACKNOWLEDGMENTS
The author is grateful to V. N. Fomenko, B.L. Birbrair,
and S. Marcos for the collaboration and to the referee
for very helpful comments.
REFERENCES
1. L. N. Savushkin and V. N. Fomenko, Sov. J. Part.
Nucl. 8, 371 (1977).
2. L. N. Savushkin and V. N. Fomenko, Lectures for
young scientists: introduction in the meson theory of
nuclear interactions and nuclear systems, Preprint
OIYaI R483369 (Joint Institute for Nuclear
Research, Dubna, 1983) [in Russian].
3. N. V. Giai and L. N. Savushkin, Sov. J. Part. Nucl. 23,
373 (1992).
4. L. N. Savushkin and H. Toki, The Atomic Nucleus as a
Relativistic System (Springer Verlag, 2004).
5. L. N. Savushkin, Relativistic Nuclear Shell Model
(St. Petersburg Nuclear Physics Institute, 2011).
6. R. J. Furnstahl, New generation relativistic models,
Lect. Notes Phys. 641, 129 (2004); B. D. Serot,
Covariant effective field theory for nuclear structure
and nuclear currents, Lect. Notes Phys. 641, 3163
(2004); D. Vretenar and W. Weise, Exploring the
nucleus in the context of low energy QCD, Lect.
PHYSICS OF PARTICLES AND NUCLEI
Vol. 46
No. 6
887
888
SAVUSHKIN
Vol. 46
No. 6
2015
Vol. 46
889
No. 6
2015
890
SAVUSHKIN
Translated by M. Potapov
Vol. 46
No. 6
2015