<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=ISO-8859-1">
</head>
<body bgcolor="#ffffff" text="#000000">
I know that many of you are starting to work on NICS
calculations. Some of you have also started to realize that there
is a myriad of output from Gaussian that you have to sort through.
Here are some pieces of advice.<br>
<br>
<ol>
<li>Just as in any experiment work, you must demonstrate first
that your theoretical method produces values in accordance with
what has previously been done. This means that you need to find
a journal article (preferably the seminal article on the
subject) and try to reproduce their results. <br>
</li>
<li>In the case of NICS you must pay very close attention to the
GEOMETRY, the THEORETICAL METHOD, and the BASIS SET. It has
been surmised that the actual NICS calculated value is less
dependent on the basis set than the other factors, but that
still does not change the fact that everything has to be the
same to get the same result. Remember what I said in CHM 331 --
a calculation done at the same geometry, basis set, and level of
theory MUST produce the same value. <br>
</li>
<li>Look very carefully at how the calculations were done. In
many of the papers I have read the geometry was optimized using
one method and basis set and then the NICS calculation was done
using another method and basis set using the previously
optimized geometry. That being said, if you are going to try
and reproduce the work of one of these papers then you will have
to follow the same procedure. My recommendation is to optimize
first and then to the optimized structure add the ghost atom for
the NICS calculation.</li>
<li>Make sure you specify your ghost atom with the symbol Bq. The
symbol X can be used to define dummy atoms for building
Z-matrices, but Gaussian will throw away those coordinates (and
no longer use that atom position) once the Z-matrix is
complete. By using Bq you are telling Gaussian that you want
information about the electronic wavefunction as that particular
point in the molecule.</li>
<li>A Gaussian NMR calculation by default uses the
Gauge-Independent Atomic Orbital (GIAO) method. This is the NMR
method used in many of the papers so you should not have to
specify anything else. What you do have to search for in the
Gaussian Output file is the magnetic shilding values. Do not
concern yourself with the actual values found in the shielding
tensor (XX, xY, etc,) but do look at the isotropic and
anisotropic magnetic shieldings for your ghost atom. You
should only need one of these -- but I will let you figure that
out based on the literature.</li>
</ol>
<br>
<pre class="moz-signature" cols="72">--
Andrew J. Pounds, Ph.D. (<a class="moz-txt-link-abbreviated" href="mailto:pounds@theochem.mercer.edu">pounds@theochem.mercer.edu</a>)
Associate Professor of Chemistry and Computer Science
Mercer University, Macon, GA 31207 (478) 301-5627
</pre>
</body>
</html>