7
50
O. Palumbo et al / Journal of Alloys and Compounds 404–406 (2005) 748–751
To check that the peak is due to a mechanism originat-
ing from the sodium aluminium hydride, rather than the
compactant with which it is mixed, we prepared pure KBr
prismatic bars and submitted them to several hydrogen
charging treatments, in H2 and in plasma atmospheres. In all
cases the dissipation assumed values comparable with the
background.
The analysis of the peak in terms of a Debye process indi-
cated that the experimental curve is remarkably broader than
a single Debye process. This means either that the relaxing
complexes are strongly interacting, or that they perform dif-
ferent types of jumps having close relaxation rates; this fact
implies a distribution of the activation energy, E, and of the
pre-exponential factor, τ0, of the Arrhenius relaxation time.
The best fit to the data, performed with Gaussian distributions
of E and τ0, gave the following mean values for the relaxation
parameters and the respective widths:
Fig. 2. Low temperature dependence of the elastic energy loss function of:
NaAlH4-Ti2%, sample Ti2-1 as prepared (•). NaAlH4-Ti2%, sample Ti2-2
after cycling up to 436 K: (ꢀ) 1.1 kHz; (ꢁ) 4.8 kHz. NaAlH4-Ti2%, sample
Ti2-2 after subsequent ageing of 1 h at 415 K: (ꢂ) 1.1 kHz; (♦) 4.8 kHz. KBr
as prepared (◦).
E = 0.126 eV
σ(E) = 0.022 eV
τ0 = 7 × 10−
14
s
σ(τ ) = 3 × 10
−15
s
modulus and its softening, after cycling up to 343 and 368 K
to the evolution of reaction (1).
0
We also observed that, after heating to a temperature not
exceeding that of the previous cycle, the modulus and dissi-
pation heating curves are retraced on cooling; this constitutes
strong indication that the chemical reaction does not proceed
appreciably in this case.
Important information is obtained from the prefactor; in fact,
its value is typical of point defect relaxation, and thus may
provide the key to the interpretation of the nature of the re-
laxing entity. Preliminarly, we suggest that, in view of the
fast dynamics of the mobile species causing the peak, as well
In the 7th cycle (Fig. 1) the sample was heated to 436 K,
and it is known [10] that at this temperature the 316 prod-
uct from reaction (1) is transformed to NaH, according to
equation:
as of the observation of H outgassing during the reactions,
hydrogen is very likely involved in the point-defect complex
causing the detected relaxation.
2
To ascribe the peak to a possible physical mechanism, we
consider that the peak cannot be ascribed to the products of
reaction (1), like Na AlH or aggregated Al, as it manifests
3
2
Na3AlH → 3NaH + Al + H2(↑)
(2)
6
3
6
itself only after the onset of reaction (2). Rather, the peak
mechanism should involve one of the possible point-defects
or point defect complexes produced by the decomposition
reaction (2), and they may be: (i) a stoichiometry defect
Therefore, the marked instability and the modulus harden-
ing measured during cycling to 436 K is clearly linked to the
evolution of decomposition (2). We point out that the phe-
nomena observed in sample Ti2-2 are strictly due to intrinsic
properties of the alanates, and are not due to KBr, whose
modulus and dissipation curves are reproducible on heating
and cooling. Fig. 2 displays the coefficient of elastic energy
dissipation, from room temperature to 4 K of the Ti-doped
sample Ti2-1 after preparation, which shows a monotonic de-
crease with decreasing temperature without any visible pro-
cess. The dissipation of a pure KBr sample, to be considered
as the background dissipation, is also drawn in Fig. 2. Af-
ter cycling to 436 K, the low temperature measurement was
repeated at two different frequencies, and the result is also
shown in Fig. 2. Surprisingly, a well developed peak appeared
at about 70 K which shifts to higher temperature at the higher
frequency, indicating that the peak is caused by a thermally
activated relaxation process. The peak is caused by a species
which was absent before the T.T. at 436 K; this entity is very
of Na3AlH . Indeed, if one or more of the six H atoms are
6
missing, jumping is possible for the remaining H atoms,
and in other words H vacancy dynamics may take place;
(ii) the relaxation of H in the NaH compound produced
by the decomposition reaction (2); (iii) the reorientation
of H around Ti, or a Ti–Al complex, or a vacancy which
acts as an attracting cluster for H after the transformation
◦
at 150 C.
The present results clearly indicate that new models in-
volving the H mobility should be considered to understand
the catalytically enhanced kinetics.
4. Conclusions
The complex elastic modulus in alanates have been
measured for the first time. The dynamic Young modulus
clearly reflects the evolution of the decomposition reactions
as a function of temperature and time. Moreover, the elastic
3
mobile, as it performs about 5 × 10 jumps/s at 70 K, cor-
responding to a relaxation rate of about 1011
temperature.
s
−1
at room
Palumbo
