I am working on the formation of amino acids in the interstellar medium. most of the time I use radicals like CH, Cooh, CH2, NH etc in different steps of my reaction like CH+COOH----->HCCooh etc. now I have following questions
1) In these reactions, I seldom found any pre-reaction complex. each time when I try to form the pre-reaction complex, reactants joins to form product even if I reduce the stepsize of optimization by 1/3 (in gaussian 09 default stepsize is 30 and I have use 10). how to form reaction complex in these systems? Is it ok to compare the product and TS energy with the sum of reactant energy.
2) most of the time I did not found any transition state for these system using QST2 and QST3. I have performed the scan along the assumed reaction coordinate (C--C distance in above reaction) however, most of the time scan fails with "error in internal coordinates" or "convergence failure". what else I can do to look for transition state and if it does not exists, how to verify that?
3) I am using wb97xd/6-311+(d,p) model chemistry for my work. Is it enough for these open shell systems? Do I need to go for mp2 or large basis functions?
4) Sometimes I have molecules with different spin muliplicity. Is it ok to use molecules with different spin muliplicity as reactants? what will be the effect of there spin on the reaction?
The idea of doing a scan of the reaction coordinate is a sound one; if there is no TS, the energy plotted along the reaction coordinate should be monotonic (either increasing or decreasing).
I don't recall how to fix "error in internal coordinates" except to make sure you have the atoms mapped from reactant(s) to product(s) correctly. Even then it might fail.
There are techniques to prevent "convergence failure". The simplest is (in Gaussian) scf(qc,maxcyc=999). The qc means to use a different method of achieving convergence of the wave function (called quadratic convergence). The maxcyc specifies how many cycles one is willing to wait for convergence, in this case 999. There are other techniques if these don't work; look up the scf keyword in the Gaussian manual. If you are using a different program, you will need different keywords of course.
As for spin, in general the total spin of the reactants should equal the total spin of the products. (There is some possibility to violate this thanks to relativity, but 99+% of reactions obey the above rule. Moreover, quantum chemical programs have much more difficulty in modeling the >1% of reactions that don't.). For example, two doublets (spin +/- 0.5) can react to form a singlet (spin 0) only if the spins are opposite: +0.5 + -0.5 = 0.
In the case of CH + COOH --> HCCOOH, there are several possibilities for the spins. CH can be a doublet (spin +/- 0.5) or a quartet (spin +/- 1.5). COOH would I presume be a doublet. HCCOOH is a carbene, which can exist in singlet and triplet (spin +/-1) states. The reaction surface is different for each spin case; for example, quartet CH would have a somewhat different C-H bond length than doublet CH. So you actually have several potential energy surfaces just for this one reaction.
- Badges
- Science topic
- Similar topics
- Chemistry
- Theoretical Chemistry