Ceilidh ... Re^4: Particle energy is less than sum of components energies

Firefly and PC GAMESS-related discussion club

Re^4: Particle energy is less than sum of components energies

Hi,
thanks for your useful comments, Andrei.
I agree with you, that my complexes are not capable to keep so much cations, how many I would like :-). My purpose is methodological calculations. I've not deal with crystal lattice, my objects it is ionic melts.
I have small experience of quantum-chemical calculations and for this reason I've asked, whether there is any "magic" option in FF which would reject unstable structures. From your answer it follows, apparently, that such option is not present.
I do not think that such option should be turned on always as sometimes it is useful to look at unstable structures, but it would be quite good to have it as a warning signal.
There's nothing to be done, let's live without this option. How it is spoken? What cannot be cured must be endured?..
Thanks once more!
Vyacheslav

--------------------------------------------------------------------
On Fri Apr 23 '10 0:09am, Andrei V Scherbinin wrote
---------------------------------------------------
>Hello,

>Now it seems more clear for me.

>1. Your negatively charged complex is probably not able to bind as many M+ cations as you wish
>for simple electrostatic reasons. For example, in ionic crystalline solid, the coordination sphere
>may include a lot of counter-ions just because of presence of the Madelung potential of the surrounding lattice. So if your aim is to simulate the condensed phase, you should provide adequate environoment, otherwise it makes little sense.

>2. You use local optimization routines which do not guarantee that you reach the global minimum.
>Moreover, there are no numerical methods allowing to do this with a guarantee. So if you still
>suspect that the problem is in the algorithms you use, you could examine a kind of multi-start
>methodology, i.e., perform a number of local optimization runs starting with different initial
>structures generated manually or randomly.

>However, I think that the real problem is 1. rather than 2.

>Best wishes,

>Andrei
>>

Fri Apr 23 '10 4:09pm

This message read 791 times