<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=GB2312" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Dear Lu,<br>
if you want to test the first way (the simplest way to do it), you
should first increase the Al RMT as much as you can in order to put the
2p Al states into the core (you can do it only for the first one). Then
you should check how much core charge you have out of the sphere. If it
is reasonable, you can expect to describe correctly to hole into those
states as it will contract the p states.<br>
So, you can:<br>
- change the in1 file for the first Al1 by first removing the 2p
states for the first Al and put them into the inc file (with a hole on
these states).<br>
- reduce the number of electrons in the in2 file<br>
- allow charge compensation in the inm file<br>
This will normally shift down the energy of the Al 2p states. I then
suggest to modify back the in1 file in order to still use a extended
basis with the 2p and 3p states for the Al. But in order not put back
the Al2p states into the clmval states, you will have to increase the
emin above the Al1 2p* (those with the hole).<br>
Normally these procedure is rather stable and it is recommended to use
it if it does not give too much charge leakage (grep :NEC in the scf
file).<br>
You can have a look in the recent paper we publish for information
<p class="Titrepublication"
style="margin-left: 28.2pt; text-indent: -28.2pt; line-height: normal;"><a
name="OLE_LINK1"><!--[if !supportLists]--><span
style="font-size: 10pt;" lang="EN-US"><span style=""><span
style="font-family: "Times New Roman"; font-style: normal; font-variant: normal; font-weight: normal; font-size: 7pt; line-height: normal; font-size-adjust: none; font-stretch: normal;"></span></span></span><!--[endif]--><i><span
style="font-size: 10pt;" lang="EN-GB">Ab initio </span></i></a><span
style=""><span style="font-size: 10pt;" lang="EN-GB">simulation of the
Electron Energy-Loss Near-Edge Structures at the Li
K-edge in Li, Li<sub>2</sub>O and LiMn<sub>2</sub>O<sub>4</sub>.</span></span><span
style=""></span><span style="font-size: 10pt;" lang="EN-US"><o:p></o:p></span></p>
<p class="Auteurs"><span style="font-size: 10pt;">V. Mauchamp, F.
Boucher, G.
Ouvrard, and P. Moreau<o:p></o:p></span></p>
<p class="Auteurs"><i><span style="font-size: 10pt;" lang="EN-GB">Phys.
Rev. </span></i><span style="font-size: 10pt;" lang="EN-GB">B <b
style="">74</b>, 115106 (2006).<o:p></o:p></span></p>
I have adapted all the para script if you are interested to run the
second option.<br>
Concerning the convergence with respect to the k point sampling. If,
for a given number of atoms the convergence is reach with 500kpt, you
can roughly divide by two the number of kpoints if you increase by two
the number of atoms into the cell.<br>
Regards<br>
Florent<br>
<br>
lu wei a ¨¦crit :
<blockquote type="cite">
<div>Dear All and Prof. Peter;</div>
<div> Thank you very much for your kindly and rapidly help!</div>
<div> I want to simulate a-Al2O3 Al-L2,3 edge, The supercell is
Al11O18Al*. I use the first way you mentioned to solve this problem, I
set Al Rmt=2.1, O Rmt=1.39 (R0=0.00005 for all of them). The
case.outputst does not give any warnning. Is it ok?</div>
<div> As to the second way, I don't understand how to operate it
since I am still not very familar with this code. In the following I
have attached the structer, in1c,in2c files. Could you show me a demo
on the operation? And where you have reduced an electron in this second
way? How to merge the 'two windown' calculation results into together
at last?</div>
<div> Another question: I test a-Al2O3 with kpoint=500 can
convergence, how much k-point I should use in the supercell
(Al11O18Al*) to keep convergence?</div>
<div> Sorry for so many questions.</div>
<div>Yours;</div>
<div> </div>
<div> </div>
<div><em>2 ways to solve this:<br>
<br>
Use as large as possible spheres for Al and choose a separation energy<br>
such that Al-2p is in the core. Check the core leakage, which should not<br>
be too large (but maybe a small miss is tolerable)<br>
<br>
Perform a "2-window" calculation. This is a bit "tricky" but should be<br>
the optimal solution.<br>
After the "normal init_lapw" do:<br>
cp case.in1 case.in1s (or in1c -> in1cs)<br>
cp case.in2 case.in2s<br>
<br>
Edit these files and put proper EMIN values in case.in1 (so that the Al<br>
2p states are NOT included; this should lead to a message like X<br>
EIGENVALUES below yy) , proper NE values in in2/in2s (The in2<br>
calculation should calculate all valence states except the Al 2p; the<br>
in2s should only calculate the Al 2p states).<br>
<br>
I haven't done this for a long time, but it should still work.<br>
<br>
For the XSPEC/TELNES calculations you would have to put the 2p state<br>
into case.inc anyway; or you could also use "optic" to calculate this.<br>
<br>
> Dear All;<br>
> I want to simulate Al atom L2,3 edge ELNES. Since L2,3 edge <br>
> corresponde to 2p->3d transition. To consider core hole effect,
I have <br>
> to reduce one electron from Al 2p state (one Al in supercell). Am
I right?<br>
> However, when we use the recommed energy -6 Ry to seperate
core <br>
> state, 2p state does not include in the core state ( I did not
find 2p <br>
> state in case.inc ). So how can I resolve this problem?<br>
> Thank you in advance !<br>
<br>
<br>
</em></div>
<pre wrap=""> </pre>
</blockquote>
<br>
<br>
<pre class="moz-signature" cols="80">--
-------------------------------------------------------------------------
| Florent BOUCHER | |
| Institut des Mat¨¦riaux Jean Rouxel | <a class="moz-txt-link-freetext"
href="Mailto:Florent.Boucher@cnrs-imn.fr">Mailto:Florent.Boucher@cnrs-imn.fr</a> |
| 2, rue de la Houssini¨¨re | Phone: (33) 2 40 37 39 24 |
| BP 32229 | Fax: (33) 2 40 37 39 95 |
| 44322 NANTES CEDEX 3 (FRANCE) | <a class="moz-txt-link-freetext"
href="http://www.cnrs-imn.fr">http://www.cnrs-imn.fr</a> |
-------------------------------------------------------------------------</pre>
</body>
</html>