NNR IN RAPIDLY ROTATED METALS By - Nottingham eTheses ...

- 109 -
in the same way to the precise determination of the 51V Knight
shift. The detected FIDs (and hence the absorption spectra)
of the vanadium sample slugs were found to be largely unaffected
by magic angle rotation at speeds up to 4.5 kHz. However an
estimate of the Knight shift was made by exchanging the metal and
reference sample in the regular Bruker probe. In the frequency
domain the resonance frequency is defined quite accurately by the
Lorentzian type centre of the absorption lineshape. This accur-
acy was similarly apparent in the time domain where the large
signal intensity and short T1 enabled the null position to be set
to ±60 Hz.
There is no obvious reference compound that should be used in
the measurement of the 51V Knight shift. In the past, values have
been quoted with respect to Nb(VO 3)2
and NaVO3 (123)
(120),
KVO 3
(121),
and vanadium dissolved in nitric acid ý124).
VOC1 3
was ruled out as a possible reference compound, because chemical
shift measurements have shown that the 51V resonance in VOC13 is
shifted paramagnetically compared to most other observable
0
(122)
VOC13
vanadium
(125)
compounds. For this work sodium metavanadate was chosen in
preference to potassium metavanadate and lead metavanadate because
of its greater solubility in water. Thus the Knight shift of the
powdered metal was measured against a saturated solution of NaVO3
in distilled water. The value obtained at a room temperature of
298 K was 5755 ±5 ppm, where the error refers to the uncertainty
in setting the null position of the metal and reference solution.
This result is to be compared with the value of 5805 ± 10 ppm
recorded by Drain (123) at a similar temperature. Vanadium has a
large magnetic susceptibility so the discrepancy between these two