View source for Rep:Jb406p3proj

=Module 3 - Physical, part3=

==Optimising the Chair and Boat conformer transition states of the cope reaction==

In this part of the exercise I familiarised myself with the different ways to gain a transition state optimisation by investigating the transition state of the interconversion of hexadiene. This had a clear analogy to Diels-Alder reactions which I would be studying later. It also offererd other advantages: this simple molecule would take less time to calculate, and so the mistakes I would make here would be simple to identify and correct, whilst with more complicated molecules I could be running a non productive calculation for many hours before I realised I had made an error.


First of all I formed the fragment molecules (allyl fragment - CH2CHCH2) and optimised them via a Hartree-Fock optimisation using a 3-21g basis set, to use as the basis of my transition states.

Chair:<jmol>
<jmolApplet>
<title>BCl3</title><color>white</color><size>150</size>
<script>zoom 200; cpk -20;;</script>
<uploadedFileContents>Jb406 Chair ts guess.mol</uploadedFileContents>
</jmolApplet>
</jmol>

Boat:<jmol>
<jmolApplet>
<title>BCl3</title><color>white</color><size>150</size>
<script>zoom 200; cpk -20;;</script>
<uploadedFileContents>Jb406 TS GUESS (PART C).mol</uploadedFileContents>
</jmolApplet>
</jmol>

I did this by first constructing the allyl fragments then copy and pasted it so I has 2  in the same file. I then orientated them so that the distance between the carbons forming the bonds was around 2.2Å, and arranged them so that they resembled the transition states.
 
===Part 1: Optimising the transition state via simple gaussian optimisisation===

This is more a intermidiatary step, but it gives a good idea to what the transition state will look like. I ran a gaussian optimisation, with calculation of frequency, on my rough transition states, setting the input to ''optimise to a TS (Berny)'' and adding in the additional keywords ''Opt=NoEigen''.

The gaussian input file looked like: <pre># opt=(calcfc,ts,noeigen) freq hf/3-21g geom=connectivity

Title Card Required

0 1
 C
 H                  1              B1
etc.
</pre>


[[image:Jb406 vibration.jpg|thumbnail|left|Vibration of the TS]][[image:jb406_Chair_frequencies.jpg|thumbnail|The frequnecies seen]] To determine whether the optimisation had occured as planned there should have been a imaginary frequency at -818cm<sup>-1</sup>. This was not seen after the first optimisation; it required further optimisation of this failed optimisations checkpoint file before I opened the frequency file and saw the imaginary frequency. The vibration of the TS was indeed analogus to the bond breaking and the bond formation seen in the cope rearrangement.












===Part 2:Use of the frozen coordinate method===

In this part I ensured that the carbons of the seperate allyl fragments that were forming the new bonds were a set distance apart - 2.2Å. This makes the TS more accurate since the bonds are indeed the right distance apart. I did this by going to the  Redundant coord editor and selecting the respective atoms - the pairs that were forming and breaking the bonds respectively - and setting the connection to ''unidentified'', and making it so that a fixed distance would always lay between them, the distance being 2.2Å.

The input:<pre># opt=modredundant rhf/3-21g guess=read geom=connectivity

opt to a ts guess force constant (part B)

0 1
 C
 H                  1              B1 
etc.</pre>

The optimised structure:

<jmol>
<jmolApplet>
<title>BCl3</title><color>white</color><size>150</size>
<script>zoom 200; cpk -20;;</script>
<uploadedFileContents>jb406_Part_c.mol</uploadedFileContents>
</jmolApplet>
</jmol>

This optimised structure is not very different from the original optimisation, but the fragment seperation is more correct.

===Part 3:Optimisation by derivision of the bond length===

Here I took the checkpoint out file and I changed the nature of the interaction between the bond breaking/forming carbons in the redundant coord editor, to bond, and the changed ''add'' to ''derivative''. I then set up a gaussian optimisation, leaving the choice for calculating the force constant as ''never''. I ran the calculation.

Input file: <pre># opt=(ts,modredundant,noeigen) rhf/3-21g geom=connectivity

opt to a ts guess force constant (part d)

0 1
 C
 H                  1              B1

etc.
</pre>

Result:<jmol>
<jmolApplet>
<title>BCl3</title><color>white</color><size>150</size>
<script>zoom 200; cpk -20;;</script>
<uploadedFileContents>jb406_Part_d.mol</uploadedFileContents>
</jmolApplet>
</jmol>

===Comparison of the seperate parts===

When I compared the 3 different optimisations, there did not seem to be any great difference between them just from examination of the molecules within gaussview. however when I used the inquirer tool I found several differences