[Schrödinger]

Hey guys!

I'm working on finding and optimizing transition states for organic reactions. I wanted to try out with a simple Diels-Alder reaction between butadiene and acetylene. So while trying to draw a reasonable input structure for the TS, I wanted to draw a double bond with a partial pi bond on acetylene as is the case (Bond order = 2.5). Apparently, this isn't possible. (gview on Gaussian 09). I went ahead anyway and left the triple bond unchanged for the rest of the TS optimization. But I found that the TS gave a structure with C-H bonds being partial in nature and one or two important bonds missing, although bond lengths dont reflect that. It's just that the TS not having solid lines for C-H bonds troubles me. So, is there a way to get around this?

[curiouscat]

Hey guys!

I'm working on finding and optimizing transition states for organic reactions. I wanted to try out with a simple Diels-Alder reaction between butadiene and acetylene. So while trying to draw a reasonable input structure for the TS, I wanted to draw a double bond with a partial pi bond on acetylene as is the case (Bond order = 2.5). Apparently, this isn't possible. (gview on Gaussian 09). I went ahead anyway and left the triple bond unchanged for the rest of the TS optimization. But I found that the TS gave a structure with C-H bonds being partial in nature and one or two important bonds missing, although bond lengths dont reflect that. It's just that the TS not having solid lines for C-H bonds troubles me. So, is there a way to get around this?

Gaussian uses DFT; and it doesn't really care about bonds at all. Single, double bonds etc. are merely abstractions.

DFT code doesn't fundamentally use them.

[Schrödinger]

So, all I have to do is draw some random bond, and just specify the bond distance and DFT will do its work? Neat! B-) But then those dashed bonds in Gaussian.. What do they mean then? There are 2 types of dashed bonds : Single bond + dashed - Bond order = 1.5; and just a dashed bond - Bond order = 1. Do they really matter while doing a caulculation?

[curiouscat]

So, all I have to do is draw some random bond, and just specify the bond distance and DFT will do its work? Neat! B-) But then those dashed bonds in Gaussian.. What do they mean then? There are 2 types of dashed bonds : Single bond + dashed - Bond order = 1.5; and just a dashed bond - Bond order = 1. Do they really matter while doing a caulculation?

Gaussian probably guesses those bonds based on proximity (atom-atom distances) and valency. I guess the code also could use electron density and distribution info. to decide the bonding but I don't think it does.

To my knowledge those shown bonds are irrelevant in any strict DFT-only calculation. All that matters is atom positions.

[vex]

Gaussian probably guesses those bonds based on proximity (atom-atom distances) and valency. I guess the code also could use electron density and distribution info. to decide the bonding but I don't think it does.

I am 99.5% sure that it's based on proximity alone.

There is a way to calculate bond order using Huckel theory without imposing any kind of valence-bond stuff on your calculation. From your Gaussian output you could conceivably get the relative atomic orbital contribution to each MO in terms of a coefficient, and then plug that into the Huckel bond order formula. However, I've only ever seen this done for pi systems and never with the output of a Gaussian calculation, so if that's the route you need to take, may The Force be with you.