|
|
|
|
|
|
|
|
|
|
|
|
Calibration
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Various
computational models (method/basis) were investigated for calculation
of
the efg's. Each was calibrated by linear regression analysis
of
the
calculated efg's versus the experimental nqcc's for a chosen set of
molecules.
Calculations of the efg's were made on the experimental structures of
these
molecules. Although not independent, all three diagonal
components
of the efg tensor were plotted against the corresponding components of
the nqcc tensor. This assures, because the tensors are
traceless,
that the regression line pass through the origin, as required by
equation
(1). The coefficient eQ/h in equation (1) is the slope of
this
line.
Having thus determined eQ/h for a given model, the model may then be
used
for calculation of nqcc's in molecules in addition to those chosen for
calibration. The premise that underlies this procedure is
that
errors
inherent in the computational model are systematic and can be corrected
- partially, at least - by the best-fit coefficient eQ/h.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The
goal is to
reproduce accurately, and efficiently, the experimental nqcc's.
The best models, therefore, are those which show the best
linear
relationship
between the calculated efg's and the experimental nqcc's - that is, the
least
residual standard deviation (RSD). It is sufficient that Qeff
- derived from the model-dependent, best-fit value of eQ/h -
approximate
Q to within a few percent.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The
effective
moment is given by
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Qeff
=
(eQ/h)/234.9647,
|
|
(2)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
where
Qeff
is in barns (b) when eQ/h is in MHz/a.u. (Physical constants
and
unit conversion factors are all contained in the numerical constant.)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Previous page
|
|
|
|
Next page
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Table
of Contents
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|