5
522
M.P. Pitt et al. / Journal of Alloys and Compounds 509 (2011) 5515–5524
and Zr Fe alloy phases can be described as:
mal rise, consistent with typical thermal expansion in the absence
of D.
3
◦
(
480–530 C) Zr FeD
(I) + 0.17Zr FeD
2
0.62
2
1.78
+
(P) 4.29ZrD1.66(c) → 0.32Zr FeD0.62(I) + 4.02ZrD1.45
2
3
.4. Amorphous Zr1 Fex phases during thermal
−x
+
(c) 0.46ZrFe + 0.42Zr Fe
(6)
2
3
disproportionation of Zr FeD
5−x
2
Reactant terms have been removed to show only the phase for-
For reactions (2)–(8) described above, we have consistently
observed a discrepancy in the balance of Zr and Fe atoms. For reac-
tions (2), (3) and (5) covering 280–430 C, we observe a loss of Zr
mation contributing to the growth of ZrFe2 and Zr Fe, the majority
3
of which occurs by depletion of the I4/mcm Zr FeD5−x phase, with
2
◦
a minor amount of Zr being sourced from cubic ZrD2 . The atom
−x
and Fe atoms from the observed crystalline phases in the sample.
balance indicates a ca. 6% increase in Zr and a 47% increase in Fe,
again, consistent with the ongoing crystallisation of an amorphous
◦
For reactions (4) and (6)–(8) covering 430–680 C, we observe a gain
in Zr and Fe atoms. Fig. 10 shows the total atom count of Zr, Fe and D
across the entire temperature range of Zr2FeD5 disproportionation.
Table 1 also summarises the change in composition of all crystalline
phases, the temperature range in which they are stable, and what
their unit cell dimensions range over. Inspection of Fig. 10 shows
the temperature range over which Fe recovery occurs is very large,
Zr1 Fex phase. The significant proportion of ca. 10 mol.% Zr Fe
suggests it has nucleated well before 530 C.
−x
3
◦
◦
.3. Zr Fe alloy growth period from 530 to 680 C
3
3
This high temperature period is well characterised by the dom-
inant growth of the alloy phase Zr Fe (see Fig. 4), a moderate rise
◦
◦
covering 200 C from 430 to 630 C. The large temperature range of
3
recovery suggests that several amorphous (a-) Zr1 Fex composi-
−x
in ZrFe2 proportion, and strong depletion of cubic Fm3m ZrD
.
2−x
tions may be present in the sample, and indeed, when summing the
Inspection of Fig. 5 shows that the Zr Fe is weakly occupied by
3
◦
discrepancy from the total (fully crystallised) atom count at 680 C,
D, which can only be sourced from the cubic Fm3m ZrD2 phase,
−x
we can infer that a-Zr71Fe29 must be present in the originally arc
suggesting a morphology of intimate mechanical contact between
the phases, allowing D diffusion that is not blocked by the pres-
ence of ZrFe . There are four D sites available in the Zr FeD
melted sample, and that a-Zr56Fe44 formation occurs from 280 to
◦
4
30 C. It is not unreasonable to expect an amorphous composi-
2
3
6.7
tion in an inhomogenous arc melt close to the Zr Fe composition.
2
structure [30], D1 (4c Zr Fe ), D2 (8f Zr Fe), D3 (8f Zr ), and D4
3
2
3
4
Nor is it surprising to expect the formation of a-Zr1 Fex phases,
−x
(
5
8f Zr ). The D3 and D4 sites show weak occupancy <12% from
4
particularly when the crystalline phases in the phase transition all
have such dissimilar lattice parameters. The most plausible expla-
◦ ◦
80 to 630 C and <5% occupancy of the D2 site at 630 C. This is
consistent with [30], where typically only the Zr4 tetrahedra are
occupied at weak D concentrations. The trigonal bypyramid Zr Fe
nation is that the domination of the cubic Fm3m ZrD2 phase in the
−x
3
2
◦
multi-phase region from 330 to 530 C is the origin of considerable
◦
is not occupied. The Zr Fe structure does not hold any D at 680 C,
3
lattice parameter misfit with other phases, resulting in the spread
and by this temperature, the cubic Fm3m ZrD2
reduced to ca. 2.5 mol.%. From 530 to 580 C, ZrFe2 and Zr FeD
growth occurs mostly from the cubic Fm3m ZrD2
according to:
proportion is
−x
of dislocations and interfacial growth of a-Zr Fex phases. There is
◦
1−x
3
6.7−x
phase
a significant resolvable strain component in the cubic Fm3m ZrD2
−x
−x
lineshape when it is initially forming, and dislocation loop punch-
ing is likely, as is observed for needle like ZrHx precipitates [31].
As such, we would expect considerable microstructure commensu-
◦
(
530–580 C) ZrD1.44(c) + 0.08Zr FeD0.62(I) → 0.86ZrD1.45
2
rate with ZrD2 nucleation, and the formation of a-Zr
Fex phases
−x
1−x
+
(c) 0.10ZrFe + 0.12Zr FeD
+ 0.38D (g)
(7)
2
3
0.34
2
along the grain boundaries. Such features obviously suggest the use
of transmission electron microscopy (TEM) to be vindicated.
The total atom count of D also suggests the presence of D in
This equation has reactant and product terms removed to show
only those phases contributing to ZrFe2 and Zr FeD growth
3
6.7−x
non-crystalline phases. Fig. 3 shows the release of D from crys-
(
the D content in the equation has been balanced to emphasise D
◦
talline Zr FexDy phases is minimal in the 380–480 C temperature
release from the sample). On balance, there is a 9% and 21% increase
1−x
◦
range. The thermal desorption data in [30] clearly show deuterium
in Zr and Fe respectively. This indicates that even at 580 C, there
◦
evolution from Zr FeD5 disproportionation in this temperature
is still amorphous Zr1 Fex crystallising. After 580 C, there is little
2
−x
range. An explanation can be found in [20], where it has been
change in the ZrFe2 proportion, and the growth of Zr FeD
can
3
6.7−x
observed that 0.2–0.7 wt.%H can be solved into a-Zr Fe33. The
be directly attributed to cubic Fm3m ZrD2 depletion as:
67
−x
◦
presence of H in a-Zr67Fe33Hx increases Tcryst by ca. 40 C. Accord-
◦
(
580–680 C) ZrD1.44(c) → 0.05ZrD1.34(c) + 0.97Zr3
ing to the concentration dependent crystallisation temperatures
◦
reported in [10], a-Zr71Fe29 will begin crystallising at ca. 365 C, or
+
Fe 0.69D (g)
2
(8)
◦
ca. 405 C for a-Zr71Fe29Dx, assuming no significant isotope effect.
◦
Reactant and product terms are removed to show the phases
These crystallisation temperatures fall in the 380–480 C range, and
contributing to Zr Fe growth. Here we have the strongest indica-
the obvious inference is that D released from the sample in this tem-
perature range can come from a-Zr71Fe29Dx present in the initially
deuterated arc melt. Crystallisation of a-Zr71Fe29Dx beginning at
3
tion yet of non-crystalline phases in the sample crystallising and
releasing atoms to contribute to the growth of existing crystalline
phases, with a 44% and 41% increase in the Zr and Fe atom count
◦
ca. 405 C is also consistent with the Fe recovery observed from ca.
◦
◦
respectively. At 680 C, we do not yet observe the formation of
430 C onward. The crystallisation of a-Zr71Fe29 and a-Zr71Fe29Dx
Zr Fe, which is below the temperature at which it is observed in
the binary phase diagram, 780 C [14]. Fig. 9 shows the variation
is shown by the dashed lines in Fig. 10. As such, the complex
2
◦
◦
crystalline multi-phase period from 330 to 530 C is further com-
◦
in unit cell parameters of ZrFe2 and Zr FeD
temperature from 530 to 680 C. The absolute change in magni-
tude of these lattice parameters is minor for example compared to
as a function of
plicated by the formation of a-Zr56Fe44 from 280 to 430 C, and the
3
6.7−x
◦
◦
crystallisation of a-Zr71Fe Dx and release of D from ca. 405 C. Zr
29
◦
and Fe recovery after 530 C is also consistent with the crystallisa-
tion of a-Zr56Fe44 starting at ca. 535 C according to [10]. It is also
◦
the variation of P4/ncc Zr FeD5−x lattice parameters in Fig. 6. How-
2
ever, all metrics of the Zr FeD
and fall, commensurate with the weak rise and fall in D occupancy
phase can be observed to rise
clear that the boundaries for crystallisation of a-Zr71Fe29 and a-
3
6.7−x
◦
Zr56Fe44 are very close in temperature to the beginning (330 C) and
◦
across the D2–D4 sites. In contrast, the ZrFe unit cell displays a nor-
end (530 C) respectively of the complex multi-phase period, and
2