Chemistry Reference
In-Depth Information
and reconstructed charge-density distributions. The charge-density community
would take advantage of such indicators as a whole, because everybody would
learn from everybody else, if those indicators existed and if the charge-density
community as a whole was open enough to discuss such “critical” things.
1.5
1001 Hidden Sources of Error
Some sources of error were already given above. What else may cause a mismatch
between predicted and measured entities? This list might get quite long. The reason
is that the more accurate the experiment gets the finer the effects are that disturb or
even destroy the results - sometimes without a warning and with no obvious sign.
Whereas standard structure determination is rather robust and may forgive small
systematic errors, the situation differs in charge-density studies. Two examples are
given in the following.
Suppose there are Renninger [
5
] reflections in the measured data. In standard
structure determination, these play a role mainly by preventing the determination of
the correct space group, as systematic absences are obstructed by multiple dif-
fracted intensities. Once the correct space group has been established, the problem
of Renninger reflections can in most cases safely be ignored, although these diffract
not only in otherwise systematic absent reflections but in others, too. Also, the
missing intensity from the strong reflection that is multiply diffracted does not play
a crucial role. In charge-density studies, however, all of these errors will affect the
values of model parameters, e.g., the monopole values.
The other example is thermal diffuse scattering (TDS). It is well known that TDS
may account for more than 25% of the intensities of the weak reflections [
6
].
In standard structure determination, this will not affect bond distances, but thermal
motion parameters. These are, however, affected only little. In charge-density studies,
these artificially large intensities may be described by artificially induced changes in
the mono- and multipoles and will certainly affect the anharmonic motion parameters
(if refined), which values are determined mainly from the high order data.
Both topics (and many more), Renninger reflections and TDS, will become more
and more important in future charge-density studies as for higher and higher accuracy
these effects will be more and more disturbing. By now, there may be hundreds of
experimental data sets waiting for new methods to correct the data a posteriori or the
theory for these effects. More, possibly hidden, sources of error are:
l
Errors on operating the diffractometer, the data processing software or the
refinement software
l
Accidental disturbance of the measurement by temporary power supply break-
down, temperature fluctuations at the detector, the crystal or at the X-ray source,
cosmic rays, natural radioactivity
l
Errors in absorption and extinction corrections, Renninger reflections
