My friends Ken and Armel have reopened a discussion that some of us have
gone through more times than we need, but I know that many people read
the Rietveld list as a way to learn about crystallography from powder
diffraction, so I would like to separate truth from hype.
What Armel says about resolution is more or less correct:
> >Though neutron data are underprivileged by lowest resolution
> >than X-ray data...
> Let us examine the lowest FWHM measured on powder patterns
> at constant wavelength :
> neutron : minimal FWHM ~0.10� (2-theta), pushing the machine
> conventional X-ray : minimal FWHM ~0.04�
> synchrotron radiation : minimal FWHM < 0.01�
and he also correctly makes the point that far more complex structures
have been solved using x-ray data than neutron.
Lest there be confusion, I would like to differentiate *solving*
structures from *determining* them. Solving comes when you have a
material with an unknown structure, for example a zeolite with an
unknown framework and you want to figure out approximately where the
atoms are. Determining a structure comes when you refine a model to best
fit the data and from this learn bond distances, bond angles, vacancy
sites...
For solving structures you need very accurate peak positions for
indexing and then minimal reflection overlap to get the most information
into the solution pipeline. Good resolution at lower 2theta values is
what you need and this is the forte of x-ray diffraction. (The best way
to discuss resolution is Q vs FWHM in Q, but for this e-mail I will
pretend that all instruments use a wavelength of ~1.54). Note that
resolution is a function of 2theta and typically degrades rapidly for
x-ray instruments as 2theta increases.
To refine complex structures without using overly simplified models
(rigid bodies, soft constraints...), you need many observations. Several
reflections that are separated by much less than a FWHM would be a
single observation. The number of observations is dictated by two
factors, FWHM, particularly at high angles and signal-to-noise. Both are
typically the forte of neutron diffraction and the literature backs this
up for publication of complex crystallographic models.
The BT-1 neutron diffractometer at NIST offers resolution of 0.12-0.2
degrees FWHM, not so good compared to the minima that Armel cites, but
we can obtain that resolution from 4 to 130 degrees! We still see peaks
well above background at 150-160 degrees. This is rare with x-rays, even
with synchrotron instruments.
The HRPD neutron diffractometer at ISIS is an exception to Armel's
comments (and an exceptional instrument). By my estimates it offers a
resolution better than "0.15 degrees FWHM" out to "~150 degrees" and is
competitive or exceeds synchrotron resolution for almost all "2theta".
Alas, in the USA at present we do not have any TOF instruments that
offer resolution equivalent to 0.2 degrees FWHM out past say "75
degrees". This is why we in the USA need to support the construction of
the new spallation source (the SNS).
Brian
********************************************************************
Brian H. Toby, Ph.D. Leader, Crystallography Team
[EMAIL PROTECTED] NIST Center for Neutron Research, Stop 8562
voice: 301-975-4297 National Institute of Standards & Technology
FAX: 301-921-9847 Gaithersburg, MD 20899-8562
********************************************************************
