carbonyl charges

User 870ab5b546

29-04-2009 18:33:15

Consider the charges on the carbonyl C atoms in these two compounds:


[bob@epoch WEB-INF]$ evaluate -e "charge()" "CC(=O)OC(=O)OC" -p 4
0.0226;0.1824;-0.2456;-0.0663;0.1674;-0.2028;-0.1672;0.0477
[bob@epoch WEB-INF]$ evaluate -e "charge()" "CC(=O)OC=O" -p 4
0.0225;0.1704;-0.2481;-0.0966;0.1651;-0.2532

The first result makes sense.  The carbonyl C attached to O should be less positive than the carbonyl C attached to C.  The second result makes less sense.  The carbonyl C attached to C should be less positive than the carbonyl C attached to H.  Should I be using a different function, or is there something wrong with the calculation, or what?

User 851ac690a0

16-06-2009 14:50:55

Hi,


 


This is a "bug" of the calculation.  The "H" atom does  not have suitable parameter in the charge calculation. 


 


 


I propose the calculation of  the "bond charge":


Add the "C" charge to the avarage charge of  the "H" atoms  which are attached to the  "C". 


This biased "C" charge may be show more reliable picture from the "C" atoms relatively to ecah others.


For example:


bond charge of "C 5 " =  0.3042;      0.1651+0.1391=0.3042


bond charge of " C 1" = 0.0561;       ( 0.0225+ (0.1008/3)  )=0.0561


bond charge of " C 2" = 0.1704;


 


Jozsi

User 870ab5b546

16-06-2009 15:58:04










Jozsi wrote:

Hi,


 


This is a "bug" of the calculation.  The "H" atom does  not have suitable parameter in the charge calculation. 


 


 


I propose the calculation of  the "bond charge":


Add the "C" charge to the avarage charge of  the "H" atoms  which are attached to the  "C". 


This biased "C" charge may be show more reliable picture from the "C" atoms relatively to ecah others.


For example:


bond charge of "C 5 " =  0.3042;      0.1651+0.1391=0.3042


bond charge of " C 1" = 0.0561;       ( 0.0225+ (0.1008/3)  )=0.0561


bond charge of " C 2" = 0.1704;


 


Jozsi



 


I guess that could be a fix in the very short term, but are you planning to fix the problem of the H not having a suitable parameter, instead of asking us to rely on this arbitrary-looking fudge?  If so, when can we expect the new parameter?


For that matter, when can we expect the improved parameters that Hans suggested for energyE() calculations?

User 851ac690a0

16-06-2009 16:10:04

Hi,


 


My suggestion for your problem is to use the "bond charge".  May be it will help you. I don't think that I will improve the "H" parameter of the charge calculation in the near future.


As far as Hans's suggestion is concerned I don't know when I will be able to deal with that proposal again. I started but I had to postponed.


 


 


Thanks.


Jozsi