[Pw_forum] second energy parameter setting in RRKJ USPP

Stefano Baroni baroni at sissa.it
Mon Aug 11 21:42:18 CEST 2008


Dear Wenmei:

besides the very sound suggestion by Hande, I would also recommend  
that you take a glance at any (possibly, old) textbook in elementary  
quantum mechanics. Why old? Because modern texts often given  
elementary stuff for granted, which usually it is not.

Coming to the specific issue that you raise. The number of nodes in  
the solution of a 1D Schroedinger equation depends on the energy  
parameter. The lower the energy, the fewer the nodes.  Of course, for  
an arbitrary energy parameter the resulting wavefunction is not  
normalizable, that is why it has to be discarded, but for very  
special, discrete, values, and energy is quantized. It is interesting  
to understand "intuitively" how quantization of energy comes about. As  
you rise the energy from any given value, the outer node moves out  
furthhr and futher from the origin, but the number of nodes stays  
constant if you do not change the energy too much. If the energy is  
raised too much, a further node will appear. For some critical value  
of the energy, the "extra" node will be at infinity: that's when an  
energy eigenvalue is crossed.

I do not expect that, if you have never heard about similar  
arguments,  you will understand mine, which have been stated rather  
sloppily. Actually, if you decide to study this matter in a serious  
book, your ability to understand my "sloppy arguments" will be a check  
of your actual understanding of the serious arguments expounded in the  
book.

Hope this helps, a bit
Stefano

On Aug 11, 2008, at 6:36 PM, Hande Ustunel wrote:

> Dear Wenmei,
>
> I usually look at the output of the all-electron calculation and  
> choose an
> energy close to the ae orbital energy. Once I manage to generate the
> pseudopotential I go as far away as possible from that reference  
> without
> upsetting the code.
>
> Hope this helps...
> Hande
>
> On Tue, 12 Aug 2008, [gb2312] Ã÷ÎÄÃÀ wrote:
>
>>
>> Dear all:
>>
>>      I want to generate my own RRKJ type ultrasoft pseudopotential  
>> with atomic code in PWSCF,
>> I encountered a very confusing situation:
>>
>>  when I chose -0.20Ry as the second energy used to pseudize the 5s  
>> orbit, the ld1.x always complains :
>> =================================================
>>    from gener_pseudo : error #         1     too many nodes
>> =================================================
>>   however if I change the energy to be -0.50Ry, the corresponding  
>> complain disappeared.
>>
>> How should I choose these energy parameters to generate a good  
>> pseudopotential and how should I have a
>> correct understanding of these parameters?
>>
>> Thanks very much .
>>
>> Wenmei
>>
>> My &inputp file is also listed as :
>>
>> 64D  3  2   10   0.00  1.00   2.004D  3  2   0    -0.20 1.00    
>> 2.005S  1  0   2    0.00  2.00   2.505S  1  0   0    -0.50  2.00    
>> 2.505P  2  1   3    0.00  2.66   2.985P  2  1   0    -0.20  2.66    
>> 2.98
>>
>>
>>
>>
>>
>> _________________________________________________________________
>> ¿ìÀ´¿´¿´ÕâЩèßäÓжය£¬±¬Ð¦£¡
>> http://cnweb.search.live.com/video/results.aspx?q=%E5%8F%AF%E7%88%B1%E7%8C%AB%E5%92%AA&Form=MEVHAA
>
> -- 
> Hande Toffoli
> Department of Physics
> Office 439
> Middle East Technical University
> Ankara 06531, Turkey
> Tel : +90 312 210 3264
> http://www.physics.metu.edu.tr/~hande
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum



More information about the Pw_forum mailing list