PHYSICAL REVIEW C 74, 027601 (2006)
Spin-excitation mechanisms in Skyrme-force time-dependent Hartree-Fock calculations
J. A. Maruhn,1 P.-G. Reinhard,2 P. D. Stevenson,3 and M. R. Strayer4
1Institut fu¨r Theoretische Physik, Universita¨t Frankfurt, Max-von-Laue-Str. 1, D-60438 Frankfurt am Main, Germany
2Institut fu¨r Theoretische Physik II, Universita¨t Erlangen-Nu¨rnberg, Staudtstrasse 7, D-91058 Erlangen, Germany
3Department of Physics, University of Surrey, Guildford, Surrey GU2 7XH, United Kingdom
4Physics Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831-6373, USA
(Received 26 April 2006; published 11 August 2006)
We investigate the role of odd-odd (with respect to time inversion) couplings in the Skyrme force on collisions
of light nuclei, employing a fully three-dimensional numerical treatment without any symmetry restrictions and
with modern Skyrme functionals. We demonstrate the necessity of these couplings to suppress spurious spin
excitations owing to the spin-orbit force in free translational motion of a nucleus but show that in a collision
situation there is a strong spin excitation even in spin-saturated systems which persists in the departing fragments.
The energy loss is considerably increased by the odd-odd terms.
DOI: 10.1103/PhysRevC.74.027601 PACS number(s): 24.10.−i, 25.70.Lm, 21.60.Jz
Time-dependent Hartree-Fock (TDHF) enjoyed a period
of large attention in nuclear physics about 30 years ago;
for reviews see, e.g., Refs. [1–3]. These early calculations
delivered a great number of useful insights into the basic
mechanisms of heavy-ion collisions, even with the large prac-
tical restrictions of that time concerning the model, degrees of
freedom, and symmetries. However, it was soon recognized not
to be as comprehensive a description as originally expected.
For example, widths in the distributions of fragments and
kinetic energies are systematically underestimated, a fact
which had been traced back in parts to missing correlations
[4,5]. More puzzling was that average quantities, such as fusion
cross sections, did not come out all that well although they
should be predictable by mean field dynamics. Already at that
time there were indications that the many restrictions in the
calculations spoil their predictive value and that, for example,
simply the proper handling of the spin-orbit (l∗s) force can
improve the results considerably towards the experimentally
observed dissipation [6,7]. Computer limitations halted these
developments for a while. The subsequent dramatic advance
in computational power now allows three-dimensional TDHF
calculations with a full-fledged Skyrme force, without any
symmetry restrictions, and for any nuclear size. Accordingly,
there is a renewed interest in TDHF studies as seen from recent
publications on resonance dynamics [8–10] and heavy-ion
collisions [11]. The present manuscript also deals with recent
3D TDHF calculations and aims to investigate the importance
of a full treatment of the l∗s force and related dissipation
mechanisms.
TDHF in a nuclear context means a time-dependent mean-
field theory derived from an effective energy functional. The
most widely used is the Skyrme functional which was proposed
long ago as a quantitative self-consistent model for the nuclear
ground state [12] and dynamics [13]. The Skyrme energy-
density functional consists of free kinetic energy, Coulomb
energy with exchange in the Slater approximation, and an
effective-interaction part depending on density ρ, kinetic
density τ , l∗s density �J , current � , and spin density �σ , for
a detailed explicit expression see, e.g., Ref. [14]. Pairing is
not considered in the present case where we deal mostly with
closed shell nuclei. For the purpose of later discussions, we
display here the l∗s part of the functional
Eevenls = −
∫
d3r
(
b4ρ∇ · �J + b′4
∑
q
ρq(∇ · �Jq)
)
, (1)
E
(odd)
ls = −
∫
d3r
(
b4 �σ · (∇ × � ) + b′4
∑
q
�σq · (∇ × �q)
)
.
(2)
The index q ∈ {p, n} labels protons and neutrons. The spec-
ification of the energy-density functional fixes all that is
needed for TDHF and the stationary initial states. The TDHF
equations are derived by time-dependent variation with respect
to the single-nucleon wavefunctions ϕ+α and the corresponding
stationary HF equation by analogous stationary variation.
The full Skyrme functional and the subsequent TDHF
equations meet all symmetries of space-time, in particular
invariance under Galilei transformations, a condition which
must be fulfilled for a meaningful theory of heavy-ion reactions
[13]. Galilean invariance imposes restrictions on the form of
the odd-odd terms, i.e., those terms containing the time-odd
pieces, current � and spin density �σ . This means that the
kinetic term always appears in the boost-invariant combination
ρτ − � 2. Of particular importance here is the correct interplay
between the even-even and odd-odd parts of the l∗s term,
Eqs. (1) with (2). We shall now discuss their effect on
heavy-ion collisions and, in particular, show that their omission
leads to inconsistencies.
The Skyrme functional allows a very precise description
of nuclear ground state properties and excitations [14]. There
exists, in fact, a great variety of parametrizations of the Skyrme
functional in the literature which differ in quality and bias
of fitted data. In order to distinguish generic effects from
particularities of a certain parametrization, we considered
several different Skyrme forces and show results for two:
SkM∗ [15] and SLy6 [16]. Calculations with other forces did
not differ significantly for the purposes of this work.
0556-2813/2006/74(2)/027601(4) 027601-1 ©2006 The American Physical Society
BRIEF REPORTS PHYSICAL REVIEW C 74, 027601 (2006)
The practical solution of the TDHF equations employs a
representation of wavefunctions, potentials, and densities on
a three-dimensional Cartesian coordinate-space grid. Deriva-
tives are evaluated in Fourier-transformed space using the
fast Fourier transformation (FFT) [17]. We work with a grid
spacing of 1 fm. The accelerated gradient iteration is employed
to find the stationary ground state solution [17,18]. The
Coulomb field is calculated by solving the Poisson equation
on a grid which is twice as large as the physical grid and
with periodic boundary conditions with the method of Ref.
[19]. Note that the reflection of emitted nucleons from the
boundaries of the numerical box lead to an uncertainty of
2–3 MeV in the final relative-motion energy, an effect which
is also seen in giant resonance calculations [20]. A Taylor
series expansion up to sixth order of the unitary mean-field
propagator is used for the dynamical time stepping [21]. The
conservation of particle number and total energy provides a
rather stringent check of numerical accuracy. In practice, we
tune our numerical parameters such that we observe over all
time a change in the particle number of less than 0.01, and
a drift in the total energy of less than 0.1 MeV. A time step
of �t = 0.2 fm/c was found adequate in all cases considered,
independent of the bombarding energy and also of whether the
odd-odd l∗s couplings were included or not.
The standard test case throughout this paper is an 16O+16O
collision atEc.m. =75–150 MeV. This system is one of the most
frequently studied with TDHF and has also been the focus of
investigations of the effect of the l∗s force on dissipation [7].
The fragment wave functions are placed symmetrically on
the grid to an initial center of mass distance of 16 fm and
then boosted to the desired relative center-of-mass (c.m.)
energy. This prepares the initial state from which the TDHF
propagation is calculated. We also compare results for systems
including 12C, 13C, and 48Ca in order to get a first impression
of the systematic variations.
The most interesting observable in heavy-ion collisions is
the kinetic energy of relative motion of the two fragments.
This quantity is deduced using a two-body analysis of the
time-dependent density distribution. For this purpose, we
calculate the principal axes of the mass quadrupole tensor, then
examine the density along the axis of minimum quadrupole
moment to find whether it shows the characteristics of two
maxima separated by a low-density region. The point of
lowest density along this line then defines a dividing plane
perpendicular to this axis, and two fragments are assumed to
exist on both sides of this plane. Calculating the centers of
mass of each fragment yields a new straight line connecting
them, which is used to repeat the process. This is iterated
until the definitions of the fragment centers of mass and the
dividing plane have stabilized. The principal result of this
analysis are the fragment masses and charges Mi,Qi, i = 1, 2,
the separation distance R of the fragments, the relative velocity
˙R as well as the angular velocity of rotation in the scattering
plane ˙θ calculated from the positions in two successive time
steps. The energy of relative motion can then be calculated
from the radial kinetic and rotational energy of the fragment
minus the remaining Coulomb interaction. The Coulomb en-
ergy is approximated by the expression for two point charges,
which should be good for larger distances. By comparing this
0 10 20 30 40 50 60 70 80
t [fm/c]
26
27
28
29
T[
Me
V]
Case
C12
C13
O16
FIG. 1. The c.m. energy of a single free nucleus moving through
the grid for three different test cases as indicated. The nearly constant
lines correspond to the full Skyrme treatment, while the other results
were calculated omitting the odd-odd l∗s coupling.
with the full numeric calculation of the Coulomb interaction
energy, we could establish that for R� 12 fm it is accurate
within about 0.02 MeV.
As a first critical test, we consider the free translational
motion of a nucleus. The results for three different nuclei
shown in Fig. 1 strikingly illustrate, on the one hand, the
accuracy of the code, and on the other, the need for odd-odd
couplings. If the full Skyrme force is used, the kinetic energy
of the c.m. just shows small oscillations by |�T | < 0.2 MeV
in the period covered, while with the odd-odd couplings
omitted there is an immediate deceleration followed by large
oscillations. At this point we can already conclude that
the omission of these terms leads to unacceptable physical
behavior.
The spurious dissipation is caused by the intrinsic excitation
of a spin-twist mode. For a nucleus moving with constant
velocity �v, the coupling term contains
∇× � = (∇ρ)×�v = dρ
dr
�r�v
r
, (3)
where spherical symmetry was assumed for simplicity. This is
an azimuthal vector field which thus couples to a spin field of
the same character. Omitting the odd-odd coupling thus leads
to a spurious excitation of a ring-like spin density, which with
the full Skyrme force is suppressed by the odd-odd terms.
An examination of the spin density in the cases without the
odd-odd l∗s terms shows that the actual excitation of this mode
accounts for about 95% of the energy loss. The rest is due to
additional excitations caused by the deceleration (note that the
total energy is conserved in any case).
Figure 2 shows the relative c.m. energy for a collision.
The initial phase (up to 40 fm/c) is free c.m. motion of the
two nuclei, and we see again the spurious dissipation of
c.m. energy as soon as the odd-odd l∗s term is omitted. In
contrast, the full Skyrme interaction preserves the relative
motion energy very well until contact. The bottom part of
the figure shows the energy contained in the odd-odd l∗s term
(2). The “no-odd” case shows a substantial increase in that
energy which is obviously properly compensated in the full
027601-2
BRIEF REPORTS PHYSICAL REVIEW C 74, 027601 (2006)
0 50 100 150 200 250
t [fm/c]
-10
10
30
50
70
90
110
130
E[
Me
V]
Odd
No-oddEcm
Eodd-odd
FIG. 2. Relative c.m. energy (top) and odd-odd l∗s energy
(bottom) in a head-on 16O+16O collision at 125 MeV. The energies
values loose their meaning in the contact regime, which is between
about 50 and 120 fm/c. “odd” and “no-odd” refer to calculation with
and without the odd-odd l∗s couplings.
treatment. The energetic relations are reversed in the exit
channel. The unphysical case without odd-odd l∗s departs
with more residual c.m. energy while the full interaction
produces more true dissipation. This is caused again by
the mechanism sketched in the spin-coupling term (3), but
now is not spurious: it occurs because the l∗s terms in the
single-particle Hamiltonian and the counterbalancing odd-odd
l∗s coupling come out of synchronization due to the physical
change in current pattern during the collision, so that a net
spin-twist excitation remains. Figure 3 visualizes the pattern
of the spin-twist excitation in a stage where the two fragments
start interacting. This spin density distribution pattern was
observed in the numerical results. The excitation persists after
full separation. The extra amount of energy stored in this mode
explains the enhanced dissipation observed in the exit stages of
Fig. 2. The spin-twist mode and the unambigous detection of
its deexcitation following a collision certainly deserves further
detailed investigations.
We have investigated a great variety of central and non-
central collisions at various collisional energies. They all show
similar effects. As one example, Fig. 4 shows the loss of
c.m. energy versus impact parameter, mainly for SkM∗ at
FIG. 3. Artist’s view of the spin excitation generated in a central
collision of two 16O nuclei. Closely based on the numerical spin-
density vectors produced in the calculations.
FIG. 4. The final c.m. kinetic energies in noncentral collisions of
16O on 16O at 100 MeV at the impact parameters given, and with or
without the inclusion of the odd-odd l∗s terms. The Skyrme force
used was SkM∗. For comparison, results from Ref. [7] are shown,
with and (completely) without l∗s-force.
various levels of approximation, one result from SkI3 for
comparison, and results from older TDHF studies in axial
approximation [6,7]. The old calculations were done with a
variant of SkM∗ replacing the gradient terms by finite folding.
At that time, it was great success to include the even-even
l∗s term. This brought a substantial jump in dissipation and
resolved the puzzle of too much transparency in the then
older TDHF calculations. The present calculations without
odd-odd terms still differ from the previous ones in that they
are now fully triaxial. This makes no effect at low impact
parameter (the minor difference is probably due to the folding
approximation) but a visibly enhanced dissipation for grazing
collisions which is reasonable because non-central collisions
break axial symmetry and call for a triaxial treatment. The most
interesting effect here is the additional dissipation caused by
the step to the full Skyrme functional (compare up-triangles
with full dots). It remains very similar up to an impact
parameter of to about 5 fm which is, not suprisingly, close
to the nuclear radius. For larger b, it rapidly vanishes as we
get to peripheral collisions. The spin-twist mode thus leads
to excitations in the final fragments that remain relevant for a
large range of impact parameters. The most striking effect is
the qualitative difference that thew new calculations without
symmetry restrictions and with all odd-odd terms included
predict a large regime of fusion whereas all the other do not.
We can estimate a fusion cross section for the system 16O+16O
at collisional energy of 100 MeV to be somewhat larger than
500 mb. This compares very well with 550 mb deduced from
the systematics of Ref. [22] while all approximate calculations
fail in that respect. Fig. 4 also shows the result for full SkI3,
which is very similar to SkM∗. The same similarity is seen
for other forces investigated. Calculations with older Skyrme
forces showed a much more dramatic and systematic force
dependence [6,7,23,24].
The dependence of the additional dissipation (as compared
to restricted calculations) on the mass of the colliding nuclei
was tested by a few calculations for other collision partners.
027601-3
dell
线条
dell
线条
dell
线条
dell
线条
BRIEF REPORTS PHYSICAL REVIEW C 74, 027601 (2006)
For 12C on 12C the additional energy loss was similar and
possibly even larger reaching 30 MeV, and similarly for 12C
on 16O. On the other hand, for 48Ca on 48Ca it was almost
negligible with about 1 MeV. The reason for this is not yet
clear and needs to be investigated.
In this work we have investigated the energy loss in
heavy-ion collisions as described by TDHF handled in full
3D and including consistently all terms of the given Skyrme
functional. Particular attention was paid to the time reversal
odd-odd spin-orbit (l∗s) term which is often neglected in TDHF
calcuations. The main findings can be concluded in brief:
The odd-odd l∗s terms establish full Galilean invariance of
the functional and they are crucial to provide properly free
translation of a nucleus over the grid. The odd-odd l∗s terms
add substantially to the dissipation observed in heavy-ion
collisions. That effect persists up to impact parameters of
order of the nuclear radius. It is large for small nuclei and
seems to decrease for heavier ones. The enhanced dissipation
is associated with the strong excitation of a pronounced
spin-twist mode which is present even in the collision of
spin-saturated nuclei and persists after separation in both
fragments. The two main tasks for future research are:
first, large scale investigations of dissipation under varying
scattering conditions, and second, a closer inspection of that
most interesting spin-twist mode, working out directions for
an experimental assessment.
This work was supported by BMBF under contracts no. 06 F
131 and 06 ER 808 and the UK EPSRC grant GR/S96425/01.
We gratefully acknowledge support by the Frankfurt Center
for Scientific Computing.
[1] J. P. Svenne, Adv. Nucl. Phys. 11, 179 (1979).
[2] J. W. Negele, Rev. Mod. Phys. 54, 913 (1982).
[3] K. T. R. Davies, K. R. S. Devi, S. E. Koonin, and M. R. Strayer,
in Treatise on Heavy-Ion Physics, Vol. 3 Compound System
Phenomena, edited by D. A. Bromley (Plenum Press, New York,
1985), p. 3.
[4] P.-G. Reinhard, K. Goeke, and R. Y. Cusson, Nucl. Phys. A398,
141 (1983).
[5] J. B. Marston and S. E. Koonin, Phys. Rev. Lett. 54, 1139 (1985).
[6] A. S. Umar, M. R. Strayer, and P.-G. Reinhard, Phys. Rev. Lett.
56, 2793 (1986).
[7] P.-G. Reinhard, A. S. Umar, K. T. R. Davies, M. R. Strayer, and
S.-J. Lee, Phys. Rev. C 37, 1026 (1988).
[8] T. Nakatsukasa and K. Yabana, Phys. Rev. C 71, 024301 (2005).
[9] J. A. Maruhn, P.-G. Reinhard, P. D. Stevenson, J. R. Stone, and
M. R. Strayer, Phys. Rev. C 71, 064328 (2005).
[10] C. Simenel and P. Chomaz, Phys. Rev. C 68, 024302 (2003).
[11] A. S. Umar and V. E. Oberacker, Phys. Rev. C 73, 054607 (2006).
[12] D. Vautherin and D. M. Brink, Phys. Rev. C 5, 626 (1972).
[13] Y. M. Engel, D. M. Brink, K. Goeke, S. J. Krieger, and
D. Vautherin, Nucl. Phys. A249, 215 (1975).
[14] M. Bender, P.-H. Heenen, and P.-G. Reinhard, Rev. Mod. Phys.
75, 121 (2003).
[15] J. Bartel, P. Quentin, M. Brack, C. Guet, and H.-B. Ha˚kansson,
Nucl. Phys. A386, 79 (1982).
[16] E. Chabanat, P. Bonche, P. Haensel, J. Meyer, and R. Schaeffer,
Nucl. Phys. A627, 710 (1997).
[17] V. Blum, G. Lauritsch, J. A. Maruhn, and P.-G. Reinhard,
J. Comput. Phys. 100, 364 (1992).
[18] P.-G. Reinhard and R. Y. Cusson, Nucl. Phys. A378, 418 (1982).
[19] J. W. Eastwood and D. R. K. Brownrigg, J. Comput. Phys. 32,
24 (1979).
[20] P.-G. Reinhard, P. D. Stevenson, D. Almehed, J. A. Maruhn, and
M. R. Strayer, Phys. Rev. E 73, 036709 (2006).
[21] H. Flocard, S. E. Koonin, and M. S. Weiss, Phys. Rev. C 17,
1682 (1978).
[22] J. R. Birkelund, L. E. Tubbs, J. R. Huizenga, J. N. De, and
D. Sperber, Phys. Rep. 56, 107 (1979).
[23] J. A. Maruhn, K. T. R. Davies, and
本文档为【Spin-excitation mechanisms in Skyrme-force time-dependent Hartree-Fock calculations】,请使用软件OFFICE或WPS软件打开。作品中的文字与图均可以修改和编辑,
图片更改请在作品中右键图片并更换,文字修改请直接点击文字进行修改,也可以新增和删除文档中的内容。
该文档来自用户分享,如有侵权行为请发邮件ishare@vip.sina.com联系网站客服,我们会及时删除。
[版权声明] 本站所有资料为用户分享产生,若发现您的权利被侵害,请联系客服邮件isharekefu@iask.cn,我们尽快处理。
本作品所展示的图片、画像、字体、音乐的版权可能需版权方额外授权,请谨慎使用。
网站提供的党政主题相关内容(国旗、国徽、党徽..)目的在于配合国家政策宣传,仅限个人学习分享使用,禁止用于任何广告和商用目的。