P. Wodniecki et al. / Journal of Alloys and Compounds 312 (2000) 17–24
19
ture of its high-temperature phase is of D022 type. This
high-temperature phase is reported to be stable between
700 and 9508C [7].
The octahedral unit cell of Hf2Al3 (space group Fdd2 )
contains eight molecules. All atoms occupy lattice sites of
low symmetry: the Hf atoms are located on 16(b) sites,
while the Al atoms populate the two non-equivalent 8(a)
and 16(b) sites in the atomic ratio 1:2 [2]. According to the
phase diagram, Hf2Al3 is stable up to 16408C [1].
PAC measurements with the 111In (EC)111Cd probes; as
usual, the 171–245-keV cascade involving the 245-keV
isomeric state in 111Cd served to measure the hyperfine
perturbation [11]. The PAC experiments covered the
temperature range between 30 and 1113 K. The measure-
ments below room temperature were carried out in a
closed-cycle helium cryostat, those above 300 K using
small resistive ovens of low g-ray absorption.
The analysis of all experimental perturbation spectra
R(t) was performed by adopting the expression of the
perturbation factor G2(t) [11,12] valid for a number (k) of
static electric quadrupole interactions:
3. Experimental details
k
3
All samples of about 300 mg weight were produced by
multiple arc melting, under argon atmosphere, the proper
amounts of high-purity components of the alloys, followed
by several days of annealing at 8008C in evacuated and
sealed quartz tubes and quenching in water. The powder
X-ray diffraction pattern established the D023 low-tem-
perature structure of the investigated HfAl3 samples and
also confirmed the Fdd2 crystallographic structure of
Hf2Al3. After this preparatory stage, the PAC samples of
some 3–5 mg were cut and doped with the hyperfine
probes (see below). In order to reach the high-temperature
phase of HfAl3, the sample temperature was increased
from 700 to 10008C during the PAC cycles. During this
annealing procedure, we noticed that some Al had covered
the quartz tube from inside, indicating the dissociation of
HfAl3 into HfAl2 1Al. The phase analysis via X-ray
diffraction gave 96% of HfAl3 (low-temperature phase)
and 4% HfAl2,, which after an annealing for 28 h at
10008C turned into about 16% HfAl2. Balducci et al. [8]
have recently studied the disintegration of HfAl3 and
established the reaction rate of the process
HfAl3(solid)→HfAl2(solid)1Al(gas). In order to calibrate
the PAC spectra after the high-temperature annealing, we
also took some PAC spectra in HfAl2, which was prepared
from stoichiometric amounts of Hf and Al as described
above. These results will be communicated separately [3].
Doping of the samples with the 181Hf activity was done
by neutron irradiation in the pile of the MARIA reactor at
Swierk, at a flux of about 1014 neutrons/cm2 per s. In
order to remove all irradiation defects, the samples were
then annealed for 1 day at 6008C prior to the 181Hf (b2)
181Ta PAC experiments. Doping with the 111In activity was
achieved by 111In ion implantation at 400 keV by means of
G2(t) 5
O
fi
O
s2n(hi ) cos(gn(hi )nQit) exp(2gn(hi )dit) (1)
i51 n50
The least-square fits of this expression to the experimental
perturbation factors yielded the fractions fi of probes
exposed to the different EFGs characterized by the quad-
rupole frequencies nQi and asymmetry parameters hi. A
broadening of the EFGi was described by a Lorentzian
nQi-distribution having the width di. The various observed
fractions fi and their EFG parameters indicate the popula-
tion of non-equivalent probe sites in the sample. Since the
samples showed evidence of a non-random orientations of
the crystallites (texture), the PAC data had to be fitted with
free s2n parameters.
From the measured quadrupole frequency nQ, the princi-
pal EFG component V was calculated for the 111Cd or
zz
181Ta probes according to
V 5 hnQ /eQ
(2)
zz
In these calculations the known quadrupole moment Q5
0.83(13) b for 111Cd [13] and 2.36(5) b for 181Ta [14]
were adopted, but the errors of the Q values were not taken
into consideration, when comparing the various values of
V . Hence, the quoted errors are only instrumental and
zz
statistical ones. The measured temperature variation of the
EFGs was fitted using either a linear or a T 3 / 2 dependence
[15]:
V (T) 5V (0)[1 2 aT], or
zz
zz
V (T) 5V (0)[1 2 bT 3 / 2
]
(3)
zz
zz
4. Results
4.1. The low-temperature HfAl3 (D023 ) phase
¨
the Gottingen heavy ion implanter IONAS [9,10]. Since
1-day annealings of the samples after implantation at
5008C turned out not to be enough for removal of all
radiation damage, an additional annealing at 6008C was
applied.
The room temperature PAC spectra taken for the D023-
type HfAl3 phase, as prepared by a 24 h annealing at
6008C, are shown in Fig. 2, for 181Ta and 111Cd probes,
respectively. For both probes, the PAC spectra reveal the
existence of one frequency triplet related to a single probe
site with well-defined, axially symmetric environment, i.e.
with asymmetry parameter h50. The corresponding quad-
The PAC measurements with 181Hf (b2) 181Ta were
performed using a standard four BaF2 detector set-up
having a time resolution of 0.8 ns for the g-rays of 134 and
482 keV in the cascade [11]. A four NaI(Tl) detector
apparatus having a time resolution of 3 ns was used for the