41 |
|
|
42 |
|
\begin{document} |
43 |
|
|
44 |
< |
\preprint{AIP/123-QED} |
44 |
> |
%\preprint{AIP/123-QED} |
45 |
|
|
46 |
< |
\title[Taylor-shifted and Gradient-shifted electrostatics for multipoles] |
47 |
< |
{Real space alternatives to the Ewald |
46 |
> |
\title{Real space alternatives to the Ewald |
47 |
|
Sum. I. Taylor-shifted and Gradient-shifted electrostatics for multipoles} |
48 |
|
|
49 |
|
\author{Madan Lamichhane} |
76 |
|
results for ordered arrays of multipoles. |
77 |
|
\end{abstract} |
78 |
|
|
79 |
< |
\pacs{Valid PACS appear here}% PACS, the Physics and Astronomy |
79 |
> |
%\pacs{Valid PACS appear here}% PACS, the Physics and Astronomy |
80 |
|
% Classification Scheme. |
81 |
< |
\keywords{Suggested keywords}%Use showkeys class option if keyword |
81 |
> |
%\keywords{Suggested keywords}%Use showkeys class option if keyword |
82 |
|
%display desired |
83 |
|
\maketitle |
84 |
|
|
1419 |
|
the number of dipoles per unit volume, and $\mu$ is the strength of |
1420 |
|
the dipole. |
1421 |
|
|
1422 |
< |
To test the new electrostatic methods, we have constructed very large |
1423 |
< |
($N = 8,000, 16,000, 32,000$) arrays of dipoles in the orientations |
1424 |
< |
described in \ref{tab:LT}. For the purposes of this paper, the primary |
1425 |
< |
quantity of interest is the analytic energy constant for the perfect |
1426 |
< |
arrays. Convergence to these constants are shown as a function of both |
1427 |
< |
the cutoff radius, $r_c$, and the damping parameter, $\alpha$ in Fig. |
1428 |
< |
XXX. We have simultaneously tested a hard cutoff (where the kernel is |
1429 |
< |
simply truncated at the cutoff radius), as well as a shifted potential |
1430 |
< |
(SP) form which includes a potential-shifting and self-interaction |
1431 |
< |
term, but does not shift the forces and torques smoothly at the cutoff |
1422 |
> |
To test the new electrostatic methods, we have constructed very large, |
1423 |
> |
$N$ = 8,000~(sc), 16,000~(bcc), or 32,000~(fcc) arrays of dipoles in |
1424 |
> |
the orientations described in table \ref{tab:LT}. For the purposes of |
1425 |
> |
testing the energy expressions and the self-neutralization schemes, |
1426 |
> |
the primary quantity of interest is the analytic energy constant for |
1427 |
> |
the perfect arrays. Convergence to these constants are shown as a |
1428 |
> |
function of both the cutoff radius, $r_c$, and the damping parameter, |
1429 |
> |
$\alpha$ in Figs. \ref{fig:energyConstVsCutoff} and XXX. We have |
1430 |
> |
simultaneously tested a hard cutoff (where the kernel is simply |
1431 |
> |
truncated at the cutoff radius), as well as a shifted potential (SP) |
1432 |
> |
form which includes a potential-shifting and self-interaction term, |
1433 |
> |
but does not shift the forces and torques smoothly at the cutoff |
1434 |
|
radius. |
1435 |
|
|
1436 |
+ |
\begin{figure} |
1437 |
+ |
\includegraphics[width=4.5in]{energyConstVsCutoff} |
1438 |
+ |
\caption{Convergence to the analytic energy constants as a function of |
1439 |
+ |
cutoff radius (normalized by the lattice constant) for the different |
1440 |
+ |
real-space methods. The two crystals shown here are the ``B'' array |
1441 |
+ |
for bcc crystals with the dipoles along the 001 direction (upper), |
1442 |
+ |
as well as the minimum energy bcc lattice (lower). The analytic |
1443 |
+ |
energy constants are shown as a grey dashed line. The left panel |
1444 |
+ |
shows results for the undamped kernel ($1/r$), while the damped |
1445 |
+ |
error function kernel, $B_0(r)$ was used in the right panel. } |
1446 |
+ |
\label{fig:energyConstVsCutoff} |
1447 |
+ |
\end{figure} |
1448 |
+ |
|
1449 |
+ |
The Hard cutoff exhibits oscillations around the analytic energy |
1450 |
+ |
constants, and converges to incorrect energies when the complementary |
1451 |
+ |
error function damping kernel is used. The shifted potential (SP) and |
1452 |
+ |
gradient-shifted force (GSF) approximations converge to the correct |
1453 |
+ |
energy smoothly by $r_c / 6 a$ even for the undamped case. This |
1454 |
+ |
indicates that the correction provided by the self term is required |
1455 |
+ |
for obtaining accurate energies. The Taylor-shifted force (TSF) |
1456 |
+ |
approximation appears to perturb the potential too much inside the |
1457 |
+ |
cutoff region to provide accurate measures of the energy constants. |
1458 |
+ |
|
1459 |
+ |
|
1460 |
|
{\it Quadrupolar} analogues to the Madelung constants were first |
1461 |
|
worked out by Nagai and Nakamura who computed the energies of selected |
1462 |
|
quadrupole arrays based on extensions to the Luttinger and Tisza |