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

- 77 -
in each case so as to give the best fit to the experimental data.
The same values of 6D were used for both 63 Cu and 65 Cu in each
halide and also for
81 Br in the samples of cuprous bromide.
From Figures 6.1 and 6.2 it can be seen that although at lower
temperatures Van Kranendonk's function fits the data well, at
higher temperatures the T1 values from all three halides fall progres-
sively below the theoretical curves. This probably arises from
a contribution to the overall spin-lattice relaxation time from
increasingly rapid ionic diffusion, as has been noted elsewhere.
Nevertheless the general tendency for Tl to decrease asymptotically
to a T-2 behaviour at higher temperatures, but deviate from such
a temperature dependence at lower temperatures is still apparent.
(78'79)
The values of T1 from the two specimens of cuprous bromide are
in very close agreement for
63Cu, 65Cu and 81Br over the entire
temperature range. The same is also true of the copper isotopes
in the two cuprous iodide specimens of differing purity, showing
that over the range of measurements the recorded T1 values were
largely insensitive to sample purity or origin.
The Debye temperatures found for each of the three cuprous
halides are shown in Table 6.1. The choice of value of 8D needed
to secure agreement between theory and experiment is not critical
so there is an uncertainty of approximately ±25 K in the values
quoted. Previous estimates of 8D as found from T1 measurements on
pure single crystals of CuCl and CuBr(80) and a powdered sample of
CuI(81) are included in the same table. The Debye temperatures
given by Domngang and Wucher for CuCl and CuBr are in reasonable
agreement with those quoted here, but there is some discrepancy in
the two values for CuI. This may simply reflect the insensitivity