Sciencemadness Discussion Board

Rotational barrier calculations

Nicodem - 2-10-2010 at 09:28

Is there anybody able to guide me step by step on how to use Jaguar to calculate the energy vs. torsion angle graph of a bond such as the C-C single bond? The hybridization and the substituents of these carbons can be anything. You can use any structure as a pedagogic model, butane, propanol, biphenyls, whatever. At the moment I would only like to learn how this is done rather than apply it to something specific.
I could not find a tutorial on this in Jaguar's quick start guide, but perhaps someone knows a complete beginner step-by-step example?

I never used the Schroedinger package for anything except once Glide for docking and even that I forgot nearly all of it. For the Maestro interface I think I could find my way trough intuitively, as it is almost human friendly, but in general my experience with Jaguar is close to nonexisting - I only had a short one hour introducing course and that was some time ago. So consider that you need to instruct a completely clueless beginner. There is currently no need for exceptionally high accuracy of the result, so the level of theory used can be any semiempirical method that still finishes the calculation in no more than a couple of hours or a day utmost on a normal medium speed PC.

Thanks in advance!

Polverone - 2-10-2010 at 15:50

The bad news is that I don't have access to this software and can't guide you through step by step, but the good news is that this kind of calculation is possible through Maestro. You want to look at section 4.4 in the full Jaguar user manual.

You can pick the angle or dihedral you want to alter, starting value, ending value, and total number of intermediate steps to calculate. If you use a relaxed scan all the geometric coordinates except the variable of interest are allowed to optimize at each step (potentially time-consuming), otherwise the structure stays rigid apart from the change you are forcing.

As an example that is easily checked against published results, I suggest the rotation of ethane. I haven't used Maestro, but if it is like other software you can get ethane from a structure library or create a carbon in the building interface, attach another carbon, and then ask Maestro to add hydrogens. Optimize the initial ethane structure at whatever level of theory you are using, then set up a relaxed scan for the rotation.

For such a small molecule you can use large basis sets and advanced methods if you please, because the calculation should still complete quickly. Semi-empirical theory is probably quite accurate in the case of ethane, but it might be interesting to compare with Hartree-Fock, DFT, and MP2 levels of theory. Use at least basis set 6-31G* or similar or higher quality.

It may also be interesting to compare the relaxed scan with the rigid. Ethane has only a few degrees of freedom, so the difference should not be as dramatic as in some other molecules, but small geometric differences can produce large energy differences.

Sandmeyer - 2-10-2010 at 16:12

The other, non-English forum has some really knowledgeable people, test your luck there as well. There is a skilled member with nick longimanus around here also, but he is unfortunately not so often around, so a PM might be a good idea... Good luck...

Nicodem - 3-10-2010 at 13:23

Thanks Polverone. I don't know how was I able to miss the full manual, especially since I already visited that page. I will try out using that as a reference, but my major problem is that the computer with the licensed package installed, is neither at my home nor my lab, but in some other lab in some other institution, so that I can't devote much time to self teaching. In the worst case I can always ask someone to show me how it is done, but everyone skilled enough is always busy and prefers to solve a specific problem, I come to them with, by themselves rather than teach me how it is done in general. Also, nobody is especially good in using Jaguar. Seems like these computational guys are either totally obsessed with Gaussian (at least those that deal with the theory) or just waste theirs and computer time running docking jobs on proteins (which is in my opinion just a great way to waste tax payer's money!). Yet, if you teach a men to catch a fish...

Polverone - 3-10-2010 at 14:09

This may be a long shot, but is there any sort of remote desktop connection available to the site where Maestro/Jaguar are available? That way you could use it from your usual work place or even your home if you have a fast enough internet connection.

Gaussian appears to be the most popular program, and although it is not the fastest or least expensive it does offer a wide variety of methods. Apart from the variables usually disclosed about computational results like basis set size, input geometries, and electron correlation methods, there are a number of often-unmentioned tunable parameters whose defaults vary across software. Examples include geometry optimization algorithms, numerical tolerances, grid sizes, and spherical vs. cartesian basis sets. It is possible to treat a problem with two different software packages and get identical results, but it can require considerable knowledge of both implementations. That means that if you are trying to follow up on someone's published results, it is easiest to just use the same program as in the prior work. Hence lots of Gaussian results in the past mean more Gaussian results published in the future. I think that any published computational work should require full software input files and output log files as part of the supplementary information but this does not appear to be standard practice.

RiP057 - 26-10-2010 at 20:23

why use jaguar and not use gaussian???

if you were using gaussian I could help you all day... its called a rigid scan

in gaussian you take your z matrix and for the command line
its
%mem
%chk
%nprocshared
b3lyp/cbsb7 scan

title

molecule specs

in the zmatrix there should be a variable for the dihedral it should be d# with # being some number
in the definition of variable section
d#=180.0 36 10.0

this will make it do a rigid scan rotating the dihedral by 10 degrees each calculation for 36 iterations starting with a dihedral of 180.0 degrees.... gives some pretty graphs...

ditch jaguar get gaussian

ps unless you have a really good computer system dont use the complete basis set and use 6-31G(d,p) or 6-31G.... I just prefer cbsb7

[Edited on 27-10-2010 by RiP057]

cattoo - 1-8-2012 at 12:15

Hi all. I need to calculate rotational barrier for some 1,3-butadienes by gaussian09. I used rigid scan method. But i need more accurate barrier values.I think i need to calculate rotational transition state energies. Do you have any idea about how to do? I am new at computational chem. I appreciate your help. thanks

arsphenamine - 11-8-2012 at 07:01

Quote: Originally posted by RiP057  
why use jaguar and not use gaussian???
[Edited on 27-10-2010 by RiP057]
Perhaps cost is an issue.

Eliteforum - 11-1-2013 at 06:11

Is Gaussian not available to pirate?