Isomers of Mo(CO)4L2

Start by optimising the ground state structures of cis and trans Mo(CO)4(PCl3)2 To get a good result you need to do this in three steps:

  • Use the B3LYP method and start with a low level basis set and pseudo-potential like LANL2MB to get the rough geometry right, you will also need to set loose convergence criteria you do this by adding "opt=loose" in the "Additional keywords" box.
  • Convergence here relates to the first derivative of the energy, if you use the normal convergence critera the limits set are more accurate than the method employed and your calculation may not converge because this limit will never be reached
  • Wowha! Has your job come back and the P-Cl bonds are gone!
  • Gaussview is not very smart, it has an internal list of bond distances (setup mainly for organic compounds) and if the computed bonds are outside of this Gaussview doesn't put them in, typically inorganic bonds can be longer than those found in organic complexes. This does not mean that the bonds don't exist, after all a bond is not just a line on a structure. A good question to discuss amoungst yourseves or with the demonstraitor is: What is a bond? How do we define bonds? However, for the moment ignore Gaussview, you know better! The bonds are there Gaussview just hasn't drawn them in.
  • Optimising at this low level will in general supply good bond lengths and angles. However, from experience the dihedral angles will not be so well described, particularly the rotation of the PCl3 groups. It is easy, if you don't start in exactly the right orientation, to find a minima that is not the lowest energy one.

    I have already searched over the rotational profile for you and worked out a good place to start that will get you to the right minima ... So take your optimised LANL2MB structure and alter the torsion angle of the PCl3 groups in the following way:
    • for the cis conformer ensure that one Cl points up parallel to the axial bond, and that one Cl of the other group points down. Make sure you rotate the whole group and not just a single Cl atom
    • for the trans conformer ensure that both PCl3 groups are eclipsed and that one Cl of each group lies parallel to one Mo-C bond.
  • Now starting from the new geometries, optimise using the B3LYP method with the LANL2DZ pseudo-potential and basis sets using the normal optimisation criteria, ie make sure that you remove "opt=loose" (but make sure you are still doing an optimisation!), we will also increase the electronic convergence by adding "int=ultrafine scf=conver=9" to the "additional keywords".
  • LANL2-DZ, has DZ and not MB as in LANL2-MB, DZ stands for double zeta and is a much better basis set and pseudo-potential than the MB or minimal basis option. With the much better pseudo-potential and basis set we need tighter convergence criteria, hence the extra options. If you want to know more about the details of these please ask a demonstrator!
  • to give you an idea of how long these jobs take, my LANL2MB job took about 13min, and my LANL2MB job took about 30min once they had got to the front of the queue on the hpc. Make sure you have submitted this second job by the end of Day 2
  • deposit your completed optimisation in the chemical database "D-space", and record the unique identifier on your wiki
  • see below for an optional advanced section or go on to do the frequency calculation in the next section

Advanced (and Optional)

  • if you feel well in control of what you are doing try this extension. Even the LANL2DZ pseudo-potential and associated basis set are not quite good enough, the phosphorous atom still has only a minimal basis (that is valence s and p orbital functions) however, we know that P likes to be hypervalent, and to use it's low lying dAOs.
  • If you have been looking at the output *.log (here the * means "some general name") files with a text editor you may have seen the rather alarming message: 
    Warning!  P  atom   10 may be hypervalent but has no d functions.
Warning!  P  atom   11 may be hypervalent but has no d functions.
    Your job will still run without these functions, but we really should include d atomic orbital functions in our description if we want more accurate results.
  • We can add these dAO functions, but gaussview is not capable of doing it, so we must edit the gaussian input file directly.
    • open the final *.log file from your previous runs and then save a new copy eg newjob_cis.com or newjob_cis.gjf (windows)
    • open the new job *.com file in a text editor
    • add the keyword "extrabasis" to the line beginning with a #, for example 
      # opt b3lyp/lanl2dz geom=connectivity 
  int=ultrafine scf=conver=9 extrabasis
    • then go right to the end of the file, add a blank line as indicated below and then the text shown in between (you can copy and paste this bit) then add another blank line. The blank lines are important, the job will not run properly without them. 
      (blank line)
P 0
D  1  1.0
0.55  0.100D+01
****
(blank line)
      This text says add to the P atom (line 1) a D function (line 2) of a specific width and decay rate (line 3) then the **** tells the program that this is the end of the extra basis input.
    • now submit this adulterated *.com file directly via the scan/server web-interface (do NOT open it in gaussview)
  • compleate the geometry optimisation and frequency analysis with these options
Top