Mechanistic understanding, gamma phase formation and morphological behaviour of the Reduction?Diffusion processed U-Mo alloy

Article abstract of DOI:10.1016/j.jallcom.2019.03.204

Reduction?Diffusion (R-D) process has been attempted to synthesize U-Mo alloy powder for the first time, and the mechanism of alloy formation has been investigated. Calciothermic reduction of uranium dioxide in the presence of the alloying element Mo, followed by high temperature soaking; subsequent leaching to remove the calcia slag yields U-Mo alloy in powder form. Experiments were carried out to synthesize U-10 wt%Mo alloy powder. The alloy powder products were characterized by Scanning Electron Microscopy and X-ray Diffraction. Complete stabilization of the metastable isotropic gamma phase has been achieved in U-10 wt%Mo powder synthesized by R-D process, where the final soaking was carried out at 1300 °C and 4 h. Attempts were made to understand the mechanism of formation of U-Mo alloy by R-D process on two fronts, i.e. alloying within a single particle, and particle-to-particle homogenization. The intended application of the alloy powders is in high uranium density dispersion fuels for high flux nuclear research and test reactors.

Full text of DOI:10.1016/j.jallcom.2019.03.204

Journal of Alloys and Compounds 791 (2019) 109e120
Contents lists available at ScienceDirect
Journal of Alloys and Compounds
journal homepage: http://www.elsevier.com/locate/jalcom
Mechanistic understanding, gamma phase formation and
morphological behaviour of the ReductionꢀDiffusion processed U-Mo
alloy
,
Sonal Gupta a *, Raj Kumar a, Sanjib Majumdar b, Jitendra Sonber b, S.K. Satpati a,
M.L. Sahu a
a Uranium Extraction Division, Bhabha Atomic Research Centre, Mumbai 400085, India
b Materials Processing & Corrosion Engineering Division, Bhabha Atomic Research Centre, Mumbai 400085, India
a r t i c l e i n f o
a b s t r a c t
Article history:
ReductionꢀDiffusion (R-D) process has been attempted to synthesize U-Mo alloy powder for the first
time, and the mechanism of alloy formation has been investigated. Calciothermic reduction of uranium
dioxide in the presence of the alloying element Mo, followed by high temperature soaking; subsequent
leaching to remove the calcia slag yields U-Mo alloy in powder form. Experiments were carried out to
synthesize U-10 wt%Mo alloy powder. The alloy powder products were characterized by Scanning
Electron Microscopy and X-ray Diffraction. Complete stabilization of the metastable isotropic gamma
phase has been achieved in U-10 wt%Mo powder synthesized by R-D process, where the final soaking
was carried out at 1300 ꢁC and 4 h. Attempts were made to understand the mechanism of formation of U-
Mo alloy by R-D process on two fronts, i.e. alloying within a single particle, and particle-to-particle
homogenization. The intended application of the alloy powders is in high uranium density dispersion
fuels for high flux nuclear research and test reactors.
Received 3 December 2018
Received in revised form
15 February 2019
Accepted 13 March 2019
Available online 16 March 2019
Keywords:
Actinide alloys and compounds
Nuclear reactor materials
Powder metallurgy
Diffusion
Scanning electron microscopy
SEM
© 2019 Elsevier B.V. All rights reserved.
X-ray diffraction
1. Introduction
behave in a stable manner under low-burnup, high-temperature
conditions, on account of swelling through fission gas nucleation
High uranium density alloys and compounds are required for
use in dispersion fuels, for high flux research and test reactors
[1e5]. UAl3, U3Si, U3Si2, U6Me (Me ¼ Fe, Ni, Mn, Co or Ge), Uꢀ10 wt
%Mo are some of the candidate alloys/compounds which have been
studied in this regard [6e10]. Among them U-Mo alloys were found
to be the most promising [3e5,11,12]. Unalloyed uranium has
orthorhombic crystal structure from room temperature up to
and growth [4,5], unlike
under irradiation, and thermal and irradiation growth under power
ramping [3,4]. In addition, -phase is desirable because of its
a-U which undergoes cavitational swelling
g
resistance to corrosion in water and water ꢀ steam environments
[11]. Hence, uranium needs to be alloyed with elements like Nb, Hf,
Zr, Mo, Ti, Pu, Ru, Rh, Pd, Re which can stabilize the g-phase for an
indefinitely long period at room temperature, and for a long time
period at elevated temperatures [3e5]. Among these, U-Mo alloys
663 ꢁC (i.e.
a-uranium) and hence is anisotropic and lacks dimen-
sional stability and irradiation stability [3e5,13,14]. Alpha uranium
transforms to beta uranium at 663 ꢁC, which has tetragonal struc-
ture; beta uranium transforms to gamma uranium at 770 ꢁC which
is body e centred cubic [13]. The high temperature phase g-ura-
nium has body centred cubic crystal structure and hence it is
isotropic and stable under irradiation environments, but not ther-
form g-phase at relatively higher range of solute concentration,
which makes them more favourable [15]. Alloying with Mo im-
proves thermal cycling stability of uranium and also imparts su-
perior fabrication characteristics [16]. Unalloyed alpha uranium is
very brittle, whereas unalloyed gamma uranium is too soft and
weak. The uranium density in the fuel cladded in Al matrix which is
achieved with U-10 wt% Mo alloy is 8e9 g cmꢀ3, as compared to
U3Si2 and UAl3, where the densities are about 4.8 g cmꢀ3 and
modynamically stable at room temperature. The g-U is known to
1.7e2.3 g cmꢀ3 respectively [17,18]. In U-Mo alloys, the cubic
g-
phase can be retained at room temperature by adding minimum
8 wt% Mo under furnace cooling condition, and by addition of
* Corresponding author.
E-mail address: sonalg@barc.gov.in (S. Gupta).
https://doi.org/10.1016/j.jallcom.2019.03.204
0925-8388/© 2019 Elsevier B.V. All rights reserved.

110
S. Gupta et al. / Journal of Alloys and Compounds 791 (2019) 109e120
4.2 wt% under critical cooling conditions [4,19]. However, for
desired fuel behaviour and performance in the reactor, the reten-
inter-particle homogenization mechanism, in R-D process in the U-
Mo system.
tion of this metastable
560 ꢁC during fuel fabrication and irradiation is desirable, as the
equilibrium phases in this temperature regime are orthorhombic
g-phase in alloys at temperatures below
2. Materials and methods
a
and tetragonal g0 (U2Mo) [3,5,20]. Literature studies report that
with increased addition of Mo in U, the transformation becomes
more and more “sluggish” [5,20]. Thus, U-10 wt% Mo is inherently
stable against decomposition; hence the alloy of this composition
has been the subject of interest of so many researchers, particularly
after 1970s [1e5].
The R-D experiments were carried out in batches on a 100 g UO2
scale. U-10 wt%Mo was the target alloy composition. Uranium di-
oxide of nuclear purity (>99.6% U) was prepared by the reduction of
UO3 with ammonia. UO3 was obtained by calcining ammonium di-
uranate at 475 ꢁC. Molybdenum metal powder (produced by
hydrogen reduction of MoO3) of 99.9% purity was used. Molybde-
num powder was taken in 10% excess over the stoichiometric
requirement for U-10 wt%Mo alloy, to ensure complete alloying and
to account for volatilization of Mo as MoO3. The raw materials were
mixed in a double cone blender. The blended mixture was charged
into a yttria-coated graphite crucible, and the remaining space of
the crucible was capped with pre-calcined CaO. The crucible was
then put inside a recrystallized alumina retort which had four
connections; for thermocouple (B-type, Pt-Rh-30%/Pt-Rh-6%),
argon inlet, argon outlet and vacuum. The retort was then loaded
into a resistive pit-type furnace. The chamber was evacuated and
back flushed with argon three times and thereafter argon was
purged continuously to maintain an inert atmosphere. The heating
schedule for the experiment was decided based on the observa-
tions of the DTA experiment (discussed in Results). DTA was carried
out using Setsys Evolution TGA-DTA/DSC (Make: Setaram) to deter-
mine the onset of the reduction temperature. The final soaking
temperatures and duration are indicated in Table 2. Soaking was
done for alloying and homogenization. The reduced mass obtained
after CTR was subjected to leaching with acetic acid in an agitated
glass beaker. This was followed by washing with acetone and vac-
uum drying, whereby it finally turned into a silvery grey U alloy
powder, followed by packing and sealing. Drying, packing and
sealing were carried out under inert (argon) atmosphere inside
glove box as U (metal/alloy) powder is pyrophoric. The presence of
even a very small amount of air/oxygen can cause violent reaction
of U and oxygen causing the powder to catch fire.
For use in the dispersion type fuels, U-Mo alloy is required in
powder form. U-Mo alloy powder was prepared in the past by
centrifugal atomization, rotating electrode process, hydriding-
dehydriding (HDH) and mechanical grinding [12,18,21e26]. All
these processes mentioned above require the starting material of
U-Mo alloy ingot to be prepared by melting-casting and afterwards
be subjected to the respective powder preparation technique.
Reduction-Diffusion (R-D) process is a novel method, which has
been used predominantly for the synthesis of rare earth magnetic
alloy powders directly from their oxides (or chlorides) [27,28]. The
rare earth oxide (or chloride) is reduced using a suitable reductant,
mostly calcium in the presence of the alloying element [27,28]. The
reduced metal then forms alloy with the other metal in the charge
by mutual diffusion, and the heat balance of the reaction is such
that the alloy is directly obtained in powder form [27e29]. The alloy
powder particles remain separated by solid calcium oxide slag
network; hence on solidification and leaching to remove the slag,
the final alloy is obtained in powder form [29]. R-D process has
been a great success for the synthesis of Sm-Co, Nd-Fe-B, La-Ni, Sm-
Fe, La-Fe-Si, Al-Ti, V-Ti, Y-Ni alloy powders [27,28,30e39]. However,
no such work on the synthesis of U-Mo powder by R-D has been
reported till date. The process is simpler and eliminates the need of
melting the metals to form alloy ingot, and hence can be more cost-
efficient process. In the present study, this method has been
employed for the first time to prepare U-Mo alloy powder. A sys-
tematic comparison of the pros and cons of the various methods
mentioned above along with the R-D method, is presented in
Table 1.
The alloy powder was characterized by X-Ray Diffraction and
Scanning Electron Microscopy techniques. X-Ray diffraction was
carried out using INEL Equinox 3000 X-Ray Diffractometer with
The present study involves the preparation of U-Mo alloy
powder by the calciothermic reduction (CTR) of uranium dioxide in
the presence of Mo powder in the required stoichiometric amount
at different times and temperatures, and characterization of the
alloy powders using XRD and SEM. The residual calcium content
was also analyzed for one of the representative powder samples
using volumetric analysis methods. Attempts were made to un-
derstand the mechanism of alloying within a single particle and
curved Position Sensitive Detector with 2
copper target with K- of wavelength 1.54056 Å. The XRD acquisi-
q
range from 0 to 120ꢁ and
a
tion was carried out for about 1 h (generator: 40 kV, 30 mA) to
ensure that peaks corresponding to different phases develop with
intensities good enough for analysis and get well resolved from the
background (i.e. peak to background ratio should be high). PDF-2 is
the reference database used for analysis using Match! Software®.
Table 1
Comparison of the different methods of alloy powder synthesis.
Atomization/Rotating Electrode Process
Hydriding- Dehyriding (HDH) process
Reduction - Diffusion process
Pros Tighter control over particle size distribution Does not require re-melting of ingot as in
ꢂ Simpler process, eliminates the need of melting the metals to
form alloy ingot, process steps are reduced.
ꢂ Raw material is oxide, does not require high purity metal as
starting material
[40]
atomization, the solid ingot can be reacted with
hydrogen directly and then de-hydrided.
Cons ꢂ Process steps are more
ꢂ Alloy ingot prepared by melting-casting which ꢂ Interaction between material of construction and alloy.
ꢂ Alloy needs to be synthesized by melting- is then hydrided.
However, in the present study this problem was overcome
casting and is then re-melted for carrying ꢂ Gamma U-Mo hydride is not very brittle; hence by using graphite crucible with yttria coating on it.
out atomization (energy-intensive).
milling operation needs to be carried out before
dehydriding [26].
ꢂ HDH might be easy for U powder (brittle
a-U
hydride); but for alloy powder, process steps
are more.
ꢂ Very fine powders produced, which are difficult
to handle and often pyrophoric

S. Gupta et al. / Journal of Alloys and Compounds 791 (2019) 109e120
111
Table 2
Details of the process parameters of the different R-D experiments.
Exp. Code
Process Parameters
Sample Preparation for SEM
RD-1
RD-2
RD-3
RD-4
RD-5
1150 ꢁC, 4 h
1300 ꢁC, 1 h
1300 ꢁC, 4 h
1300 ꢁC, 5 min
970 ꢁC, 4 h
Powder sprinkled on graphite tape
Powder sprinkled on graphite tape, powder also mounted in cold setting resin and polished to reveal the cross-section
Powder sprinkled on graphite tape
Powder sprinkled on graphite tape, powder also mounted in cold setting resin and polished to reveal the cross-section
Samples of reduced mass mounted in cold setting resin and polished to view the slag morphology
SEM and EDS were carried out using CamScan Scanning Electron
Microscope (Model: CamScan MV 100) with tungsten filament and
Energy Dispersive Spectrometer attachment of Oxford Instruments
(X-max 80). The accelerating voltage in the SEM was 20 kV.
detector. To determine oxygen content, the sample was put in
nickel flask and fused in graphite crucible in purified helium at-
mosphere at 2700 ꢁC. Oxygen is released as CO and traces of CO2. All
CO is oxidised to CO2 by CuO. Oxygen is determined by quantifi-
cation of CO2 by IR detector.
The detailed chemical analysis of alloy powder of a represen-
tative batch was carried out to determine U, Mo, residual Ca, C, N
and O concentrations. 1 g of the powder sample was dissolved in
2 N HNO3 and titrated by potassium dichromate to get uranium
concentration. To analyze Mo concentration, about 0.5 g of the
powder sample was dissolved in 4 N HNO3 and diluted to a make-
up volume 25 mL. U was extracted from the solution by 30% TBP
(Tri-Butyl Phosphate), it forms a complex with TBP. The solution
free from U was analyzed by Inductively Coupled Plasma Atomic
Emission Spectroscopy for Mo concentration. The same procedure
was followed for residual Ca analysis. U has to be first removed
from the solution because it exhibits various oxidation states, and
hence has characteristic peaks in the entire region of the emission
spectrum; which might interfere with the peaks of Ca and Mo.
To analyze nitrogen content, the sample was put inside a nickel
flask and fused in a graphite crucible in purified helium atmosphere
at 2700 ꢁC. The released nitrogen as N2 was passed through
chemical traps and quantified from the thermal conductivity dif-
ference of reference gas flow (He) and sample gas flow (N2 þ He).
The combustion of the alloy sample in purified oxygen atmosphere
at ~1100 ꢁC inside an induction furnace was carried out to detect
carbon content; whereby carbon in the sample converted to CO2.
CO2 was passed through chemical traps and amount of C is quan-
tified by measuring the transmittance against the characteristic IR
absorption peak of CO2 at wave number 2350 cmꢀ1 by employing IR
3. Results
3.1. Thermal analysis
Differential thermal analysis was carried out to determine the
temperature at which reduction of UO2 by Ca takes place. The DTA
plot is shown in Fig. 1. An endothermic peak at around 640 ꢁC is
observed, where weight loss is seen in the TG curve, which in-
dicates volatilization of Mo as MoO3 [41]. Two more endothermic
peaks are observed, one between 830 and 840 ꢁC and the other at
~1130 ꢁC; which correspond to the melting of calcium and uranium,
respectively. The melting points of Ca and U are 842 ꢁC and 1132 ꢁC
respectively. An exothermic peak appears at around 850e900 ꢁC,
next to the melting of Ca which marks the onset of reduction of UO2
to U metal. Melting of Ca triggers the reduction reaction by forming
a uniform layer over the uranium oxide particles facilitating the
reduction. There is a slight drop in the weight of the charge in the
reduction region, which occurs due to local volatilization of Ca
because of the heat released during reduction of UO2. The initial
drop in weight in the TG curve is due to the removal of moisture
and volatile impurities during heating.
The heat of reduction reaction is sufficient to melt the uranium,
but not sufficient to maintain U in molten state for prolonged time,
Fig. 1. The DTA-TG plot for calciothermic reduction of UO2 in the presence of Mo.

112
S. Gupta et al. / Journal of Alloys and Compounds 791 (2019) 109e120
Fig. 2. The XRD plot of U-Mo alloy powders prepared by R-D process.
Fig. 3. SEM images of the raw materials used for the experiments, (a) UO2 powder (b) Mo powder.
hence external heating is continued up to the final soaking tem-
perature. To ensure efficient and complete reduction of oxide to U
metal before melting of U occurs, it was decided to give a soaking of
about 10 min at 900 ꢁC in the main experimental trials.
The entries in the PDF-2 database used to identify the phases in
Match! Software® are 00-024-0748 (
a-uranium), 00-024-0749 (g-
uranium) and 00-012-0296 (g0 or U2Mo).
3.3. Morphological and chemical analyses
3.3.1. Morphology
3.2. Phase analysis
The XRD pattern shown in Fig. 2 is characteristic of the U-Mo
alloy powders prepared by R-D process. The XRD plot of U-10 wt%
Mo alloy which was soaked at 1300 ꢁC for 4 h contains peaks cor-
SEM images of the raw materials are shown in Fig. 3. It is clear
that UO2 particle (Fig. 3(a)) is porous and irregular in nature, and
particles have size in the range of 20e30
exhibit two morphologies. Mo particles (Fig. 3(b)) of around
~10e15 m are faceted. Except for these particles, most of the Mo
powder particles are in the form of agglomerates of small faceted
particles (individual particles of ~1 m).
mm. Mo powder particles
responding to only
material) are not seen in the XRD plot. Both
absent, which indicates complete stabilization of
g
-phase. Peaks corresponding to UO2 (starting
and g0 peaks are
-phase under the
a
m
g
mentioned soaking conditions. The XRD plot for one such experi-
ment carried out with 4 wt% Mo under the same soaking condi-
m
The effect of soaking temperature and soaking time on the final
morphology and distribution of Mo in an alloy particle is illustrated
in Fig. 4. In case of RD-3 (1300 ꢁC, 4 h), the alloy particles exhibit
two morphologies, one of rounded type and the other of skeleton-
like structure (see Fig. 4(c)). The composition at the particle surface
is close to 10 wt% Mo (as is evident from EDS point scan data,
Fig. 4(h)) and the average composition obtained through the EDS
tions, showed the presence of all the three (
-phase was only partially stabilized. g0 peaks occur at 2
which are very close to those corresponding to peaks; it is
because
a
,
g0 and
g) phases, i.e.
g
q
values
g
g
’ (tetragonal) is an ordered structure of the metastable
g
(cubic) phase. Hence, it is difficult to distinguish between
g
and g0
peaks.

S. Gupta et al. / Journal of Alloys and Compounds 791 (2019) 109e120
113
Fig. 4. SEM ꢀ BSE images of U-Mo alloy powders prepared by R-D process under different soaking conditions: (a) 1300 ꢁC, 5 min (b) 1300 ꢁC, 1 h (c) 1300 ꢁC, 4 h (d) 1150 ꢁC, 4 h; the
corresponding EDS data for the points marked 1 through 7 are shown in the EDS spectrums Fig. 4(e)e4(h).
data is 10.58 wt% Mo. The Mo composition observed in the volu-
metric (bulk) analysis is also 9.7 wt% (Table 3). Thus, the Mo
composition throughout the particle is about 10 wt%, which is also
the Mo composition in every alloy particle.
some particles did not contain Mo at all. Hence, the particle was
mounted and polished to reveal a particle cross-section, and the
average Mo concentration analyzed over different points at the
particle cross-section was found to be around 14 wt% (refer Fig. 4(b)
and (g) for the SEM and corresponding EDS data respectively),
which was higher than the Mo concentration at the surface, indi-
cating that a Mo concentration gradient existed from particle core
When the soaking time was limited to 1 h at 1300 ꢁC (RD-2), it
was found that there was a large variation in the surface compo-
sition of the alloy particles (in EDS point scan), and apparently,
Table 3
Chemical Analysis of the U-Mo alloy produced by the optimized R-D process.
Exp. Code
U wt%
Mo wt%
9.7 0.2
Ca wt%
O wt%
C wt%
N wt%
RD-3, U-10 wt% Mo
90.6 0.3
0.0445 0.0022
0.0595 0.0060
0.0195 0.0020
0.0350 0.0030

114
S. Gupta et al. / Journal of Alloys and Compounds 791 (2019) 109e120
to surface, which if given sufficient soaking time, would have
vanished due to homogenization. This was an important inference
from RD-2, which provided an impetus for investigation of the
alloying mechanism within a single particle. An experiment (RD-4)
was planned to capture and analyze particles where incomplete
alloying could be seen by soaking for just 5 min. As expected, un-
alloyed Mo was present in the core of the particle as seen in the
SEM image in the backscattered mode (Fig. 4(a)). Hence, Mo con-
centration in the centre of the particle was close to 100%. The EDS
results (refer Fig. 4(f)) quote this concentration as 85 wt% Mo
(marked 3 in Fig. 4(a)), because part of the intensity (~9e10 wt%) is
also attributed to gold, which was present because the mounted
sample was coated with gold to make it conducting. The actual Mo
concentration hence will be much greater than 85 wt% at the
centre. The elemental maps for U and Mo distribution were ob-
tained using EDS for the same alloy particle, as shown in Fig. 5. The
system indicates a shrinking core type mechanism of alloy forma-
tion [42], which has been discussed in detail in Section 4.
ceramic CaO slag.
3.3.2. Chemical composition
The sample which, through XRD and SEM/EDS analysis
confirmed to be homogeneous U - 10 wt% Mo (RD-3) alloy powder
was also analyzed for U wt%, Mo wt%, residual Ca and other im-
purities by the methods described in Section 2. The chemical
composition of the respective sample is reported in Table 3.
4. Discussion
4.1. Mechanism and evolution of morphology
The reaction UO2 þ 2Ca / U þ 2CaO is exothermic in nature
(D
Hꢁ
¼ ꢀ44.2 kcal/mol,
D
Gꢁ
¼ ꢀ42.15 kcal/mol, taken from
298
298
JANAF Tables). As Mo dissolves in liquid uranium, the liquidus
temperature of the alloy increases [43]. The liquidus of the con-
cerned alloy U-10 wt%Mo is 1252 ꢁC (see U-Mo phase diagram,
Fig. 11, [43]). To understand the effect of temperature on the pro-
cess, two temperatures were chosen, one above the alloy liquidus,
i.e. 1300 ꢁC and one between the melting point of U and the alloy,
i.e. 1150 ꢁC. Soaking above the alloy liquidus provides superheat to
ensure that the alloy is molten and has sufficient flowability so that
it can move through the pores of the CaO slag. The porous nature of
CaO slag has been discussed in Section 3.3.1. The average compo-
sition of the alloy powder obtained under different soaking con-
ditions and the standard deviation values calculated from the EDS
data are reported in Table 4, which clearly indicate that the spread
in the Mo concentration, i.e. the inhomogeneity decreases as the
soaking time at higher temperature (1300 ꢁC) increases. The spread
The SEM image of RD-1 alloy which was soaked at 1150 ꢁC
(Fig. 4(d)) shows faceted morphology and also agglomerates of
faceted particles, which resemble the morphology of the Mo
powder used for the experiments. These particles of RD-1 powder
confirm the presence of both U and Mo. Hence, it seems that the
liquid U envelops the Mo particle, and dissolution of Mo into the U
envelope increases the melting point of the U in the envelope,
which therefore might have solidified under the temperature
conditions of the system; hence the morphology of Mo powder is
retained in the alloy powder particles.
The SEM images in Fig. 4 also depict that the particle size of the
alloy powder is a function of soaking temperature and soaking
time. However, to quantitatively correlate the particle size with the
process parameters, further studies need to be done.
The reduced hard mass obtained after calciothermic reduction
of UO2 in RD-5 (970 ꢁC, 4 h) was mounted and polished, and viewed
under SEM to examine the morphology of the CaO slag which
formed during the reduction reaction. The SEM images are shown
in Fig. 6.
The temperature was chosen just above the temperature where
reduction occurs. In general, every SEM image had three colours,
white for the U alloy region, greyish to dark greyish for the CaO slag,
and black for the empty pockets based upon Z contrast in back
scattered electron imaging in SEM; and was confirmed by the EDS
which was done for the same images (refer Fig. 6(e) and (f)). The U
particles are in general spherical or rounded and are dispersed in
the calcium oxide slag as is evident from Fig. 6(a) and (b). The CaO
slag contains empty pockets or pores (see Fig. 6(b)), of irregular
geometry, which are interconnected. Some interconnected parti-
cles which show signs of coalescence were also observed, as in
Fig. 6(c) and (d). Presence of Au in the spectrum is due to the gold
coating done to make the sample conducting, since it contains
in the Mo concentration is maximum (
s
¼ 25.57) even with a
soaking of 4 h when the soaking temperature is 1150 ꢁC. This in-
homogeneity is lower even for process parameters (1300 ꢁC, 5 min)
and (1300 ꢁC, 1 h). Thus, it can be inferred that the soaking tem-
perature must be well above the alloy liquidus; and at such a
temperature, increasing the soaking time increases both the intra-
particle homogeneity (i.e. absence of unalloyed Mo core inside the
particle) and inter-particle homogeneity of the alloy powder.
4.1.1. Alloying within a single particle: the single-particle system
In experiment RD-4 (1300 ꢁC, 5 min), large inhomogeneity is
observed in the final alloy powder due to insufficient soaking time
available. The powder was mounted and polished to reveal a cross-
section of the particle, and then viewed under SEM in backscattered
mode. The SEM (BSE) image shown in Fig. 4(a), shows dark grey
patches on the white surface of the particle cross-section, which
confirm the presence of Mo (low Z element, Z ¼ 42) in EDS. The
white region confirms the presence of U (high Z, Z ¼ 92) with much
smaller concentration of Mo than that in the core of the particle.
The elemental map for U (in red colour) generated using EDS
Fig. 5. (a) SEM ꢀ BSE image of the cross-section of an alloy particle which was soaked for 5 min at 1300 ꢁC, line scan for U and Mo distribution through the cross-section is also
shown (b) Elemental map for U distribution over the cross-section (c) Elemental map for Mo distribution.

S. Gupta et al. / Journal of Alloys and Compounds 791 (2019) 109e120
115
Fig. 6. SEM ꢀ BSE images of the reduced hard mass obtained after calciothermic reduction of UO2 in the presence of Mo (Fig. 6(a) through 6(d)); Fig. 6(e) and (f) represent the
typical EDS spectrums for the white and grey regions in the SEM images of the reduced hard mass.
Table 4
Effect of the process parameters on the characteristics of the final alloy powder in the R-D experiments.
Exp. Code Process Parameters Avg. Composition of the R-D powder (wt% Mo, from EDS data) Standard Deviation (s or SD) Phases present in the alloy powder
RD-1
RD-2
RD-3
RD-4
RD-5
1150 ꢁC, 4 h
1300 ꢁC, 1 h
1300 ꢁC, 4 h
1300 ꢁC, 5 min
970 ꢁC, 4 h
23.21
2.59
10.58
14.12
25.57
3.08
1.59
a
a
g
a
,
,
g
g
,
,
g0
g0
10.44
,
g,
g0
The reduced mass was not leached to obtain powder in this case.
(Fig. 5(b)) confirms the presence of U predominantly at the pe-
riphery of the particle, whereas there is no uranium at all in the
core of the particle. The elemental map for Mo (Fig. 5(c)) shows an
intensified green colour in the core of the particle, and a green
colour of lower intensity elsewhere. Thus, it is clear that the particle
contains an unalloyed Mo core, surrounded by region containing
both U and Mo. The concentration of U outside the core, as seen in
the line scan, is nearly constant over the region between the core
and the periphery. This is due to the fact that there is a continuous
supply of liquid U from the surrounding pores or pockets in the slag
to the region of lower U (higher Mo) concentration, as a result of
chemical driving force arising due to difference in Mo (or U)
concentration.
capture and analyze particles where incomplete alloying could be
seen (Fig. 5 and Fig. 4(a)) and hence unalloyed Mo was present in
the core surrounded by U-Mo alloy region. The alloying occurs via a
shrinking core mechanism [42], which has also been used to
explain the alloy formation in case of Sm2Fe17 and Y-Ni alloys by R-
D process [34,39].
As the alloying progresses, the Mo core will shrink in size due to
dissolution of Mo in uranium. Hence, the topology of Mo particles is
retained initially in the alloy particles as seen in Fig. 4(d), which
vanishes with time, as the molten alloy when it is kept above the
liquidus temperature (1252 ꢁC) of the Ue10 wt%Mo alloy for pro-
longed time, will grow into spherical particles to minimise the total
surface energy. When it tries to grow against the constraints put by
the slag matrix, depressions arise over the surface of the alloy
particle, as seen in Fig. 4(c).
These observations indicate that U-Mo alloy formation occurs in
steps, (i) enveloping of a Mo particle by liquid U, and (ii) outward
dissolution or diffusion of Mo (solute) into U (solvent). Due to very
less time available for homogenization, i.e. 5 min, it was possible to
Based upon the inferences drawn so far, the single particle
alloying mechanism was identified to be the outward diffusion or

116
S. Gupta et al. / Journal of Alloys and Compounds 791 (2019) 109e120
Fig. 7. Schematic showing the mechanism of alloying within a single particle.
1) If a smaller pocket has lower Mo wt%, and a nearby bigger
pocket has higher Mo wt%, the liquid U will move from smaller
pocket to bigger pocket to equalize the concentration of the
solute (Mo, to 10 wt%) / coarsening at such locations will occur,
bigger particle will grow consuming the smaller one; thus giving
rise to rounded interconnected particles with protrusions on the
surface.
dissolution of Mo into liquid U envelope surrounding the Mo par-
ticle. This mechanism is illustrated schematically in Fig. 7.
4.1.2. Particle-to-particle homogenization: the multi-particle
system
To achieve particle-to-particle homogeneity, there should be
exchange of one element or the other among the alloy particles,
over time to attain equilibrium composition (close to 10 wt% Mo).
The U particles are in general spherical or rounded and are
dispersed in the calcium oxide slag as can be seen in Fig. 6(a) and
(b). The CaO slag contains empty pockets or pores, of irregular
geometry, which are interconnected. Hence, the liquid alloy can
move from one pocket to other through these interconnected
paths. Some particles which show signs of coalescence were also
observed, as in Fig. 6(c) and (d). Such a morphology is possibly the
result of liquid U alloy (containing Mo) flowing from one pocket to
the other under the effect of chemical driving force, due to differ-
ence in Mo concentration existing between the slag pockets. As the
soaking temperature was 970 ꢁC which is much below the melting
point of U, the local rise in temperature due to exothermic heat of
the reduction reaction might have raised the temperature locally to
keep the alloy (containing U and Mo) liquid in the regions seeing
the high temperature, but the heat of the reaction would not be
sufficient to maintain the charge at higher temperature for pro-
longed time. Hence, the liquid that had flown between the pockets
solidified in situ and retained the shape of the pockets, as soon as its
temperature fell below its melting point; thus giving rise to
morphology as is seen in Fig. 6(d).
2) If a bigger pocket has lower Mo wt%, and a nearby smaller
pocket has higher Mo wt%, the liquid U will move from bigger
pocket to smaller one, thus shrinking in size and increasing the
size of the smaller one; to equalize the concentration of the
solute (Mo, to 10 wt%) / no coarsening will occur at such lo-
cations. Instead, the morphology will resemble that of two
spheres connected by a pipe, which when repeated in three
dimensions will give rise to skeleton-like structure even at
higher soaking temperature. This proposition is supported by
the fact that the skeleton-like structure observed in SEM of RD-1
(1150 ꢁC, 4 h) shown in Fig. 8(a) is preserved even when soaking
is done for 4 h at 1300 ꢁC (Fig. 8(b)). However, at higher tem-
perature (above the liquidus of the alloy), the individual parti-
cles in the skeleton are more connected and the skeleton has
more developed arms as compared to that at lower temperature,
because liquid has sufficient flowability so as to move from one
pocket to another. Moreover, the skeleton-like structure seen in
case of soaking for 4 h at 1150 ꢁC is devoid of Mo, because that
particular region did not see higher temperature for longer time
to maintain U in liquid state, so the U could not move to enve-
lope an adjacent Mo particle and solidified in-situ. In contrast,
the skeleton-like structure at 1300 ꢁC has a composition close to
10 wt% Mo, which explains the need to hold the charge at a
temperature above the liquidus of the alloy for homogenization
to occur.
The movement of liquid alloy from one slag pocket to the other
occurs under the effect of chemical driving force due to Mo con-
centration gradient existing in a direction opposite to movement of
the liquid. This movement has two different aspects:

S. Gupta et al. / Journal of Alloys and Compounds 791 (2019) 109e120
117
Fig. 8. SEM ꢀ BSE images of the skeleton-like structures observed under different soaking conditions (a) 1150 ꢁC, 4 h (b) 1300 ꢁC, 4 h; the corresponding EDS spectrums in Fig. 8(c)
and (d) show the EDS data of the regions marked with boxes (yellow) in Fig. 8(a) and (b) respectively. (For interpretation of the references to colour in this figure legend, the reader
is referred to the Web version of this article.)
The average concentration of Mo in the system will eventually
approach the equilibrium one, i.e. 10 wt%, when sufficient soaking
time is available above the liquidus temperature of the alloy, as is
achieved with a soaking of 4 h at 1300 ꢁC. The movement of liquid
U, the final particle size and morphology and particle-to-particle
homogeneity is decided by the following factors: (i) size of the
slag pockets (ii) the concentration of Mo (solute) in the liquid (U)
within each pocket (iii) soaking temperature and (iv) soaking time
at higher soaking temperature. The mechanism governing the ho-
mogenization of Mo concentration throughout the system is
Fig. 9. Schematic showing the mechanism of particle-to-particle homogenization.

118
S. Gupta et al. / Journal of Alloys and Compounds 791 (2019) 109e120
Fig. 10. Different Stages in the reduction e diffusion process: (a) The mixture at room temperature before blending (b) The mixture after blending; particles come closer together so
that the only empty spaces are the voids between any two particles (c) Calcium melts at 842 ꢁC filling the voids between the particles and covers the particles uniformly, facilitating
reduction (d) After reduction and melting of U, molten metal droplets and solid Mo particles are entrapped in porous CaO slag which is solid; Mo dissolves in liquid uranium (e) U-
Mo alloy liquid in the pores of the solid CaO slag (f) The liquid U at higher temperature moves from its higher concentration (lower Mo wt%) to its lower concentration (higher Mo
wt%) under a chemical driving force, thus homogenization occurs with time such that every particle is U e 10 wt% Mo alloy.
temperature, the average concentration of Mo in the system will
eventually approach the equilibrium one, i.e. 10 wt% (Fig. 10(f))
illustrated by a schematic diagram in Fig. 9.
The different changes occurring in the state of the system as a
function of time (t) and temperature (T) as depicted in Fig. 10 are
1) At T ¼ Room Temperature, all the charge is solid. Blending is
done to bring particles closer together, so that the only empty
spaces left would be the voids between any two particles (t ¼ 0,
T ¼ RT: Mechanical mixture before blending; t ¼ t1, T ¼ RT:
Blended mixture, Fig. 10(a)and (b))
4.2. Gamma phase stability
The intent of this research work is to achieve fully stabilized
gamma phase in the U-Mo alloy, which can be achieved under
furnace cooling condition with a minimum of 8 wt% Mo, as
explained earlier [4,19]. Hence, this was true for U-10 wt%Mo alloy
prepared by R-D under the suitable soaking conditions. The XRD
plot shown in Fig. 2 corroborates this fact.
The present work however involves in situ formation of U-Mo
alloy powder after reduction of UO2, where metal is melted locally
(dispersed in non-melting CaO slag) and alloying is achieved by
local diffusion of Mo into U. Therefore the local cooling rates are
expected to be higher than that of conventional alloy making
processes, even when the average cooling conditions are similar to
furnace cooling condition. So, it was expected that even with 4 wt%
Mo, gamma phase could be stabilized completely. However, the
XRD plot for U-4wt%Mo alloy in Fig. 2 shows the presence of peaks
2) At t ¼ t2 and T ¼ 842 ꢁC, calcium melts and surrounds/covers the
particles intimately. The blended mixture of UO2 and Mo is
submerged in molten calcium. Reduction of UO2 by Ca occurs
and heat of the reduction reaction is sufficient to melt uranium;
but may not be enough to sustain melting and also heat losses
will be more for smaller size of the system; hence heating is
continued above the melting point of U and soaking at higher
temperature is given (Fig. 10(c))
3) At t ¼ t3, the system takes the form of a porous solid (high
melting 2572 ꢁC) CaO slag network with molten uranium pre-
sent in the slag pockets or pores; Mo is distributed randomly
among these slag pockets, submerged in the molten uranium.
The liquid U envelops the Mo particle and alloying occurs due to
dissolution of Mo into the liquid U (Fig. 10(d)and 10(e))
4) The liquid U moves from one pocket where concentration of
liquid is more (solute Mo is less) to other pocket in the vicinity
where concentration of liquid is less (solute Mo is more) under
chemical driving force, so as to equalize U wt% and thus Mo wt%
in each particle. Given soaking for sufficient time at higher
corresponding to all the three phases, a, g and g'; which are the
equilibrium phases at 4 wt% Mo as is evident from the U-Mo phase
diagram shown in Fig. 11 [43]. This clearly explains that the cooling
rates in case of R-D process are in no way greater than the cooling
rates in furnace cooling conditions, which necessitates a Mo con-
centration greater than 8 wt% for complete stabilization of the
isotropic
g phase.

S. Gupta et al. / Journal of Alloys and Compounds 791 (2019) 109e120
119
and all other staff in our lab for their extended support while car-
rying out the calciothermic reduction and leaching experiments.
The authors also thank Dr. Bhaskar Paul of MP&CED for his support
in carrying out the DTA experiment. The authors also thank Shri
K.N. Hareendhran and Shri S.V. Kadam of Quality Control Section,
UED in carrying out the chemical analysis of the alloy powders.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.jallcom.2019.03.204.
References
[1] A. Travelli, The U.S. Reduced enrichment research and test reactor (RERTR)
program, in: Proceedings of the 1978 International Meeting on Reduced
Enrichment for Research and Test Reactors, ANL/RERTR/TMe1, Argonne, Illi-
nois, November 9-10, 1978, pp. 3e18.
[2] S. Van den Berghe, P. Lemoine, Review of 15 years of high-density low-
enriched UMo dispersion fuel development for research reactors in Europe,
Nucl. Eng. Technol. 46 (2) (2014) 125e146. https://doi.org/10.5516/NET.07.
2014.703.
[3] V.P. Sinha, P.V. Hegde, G.J. Prasad, G.K. Dey, H.S. Kamath, Phase transformation
of metastable cubic gamma-phase in U-Mo alloys, J. Alloys Compd. 506 (1)
(2010) 253e262. https://doi.org/10.1016/j.jallcom.2010.06.187.
[4] V.P. Sinha, P.V. Hegde, G.J. Prasad, G.K. Dey, H.S. Kamath, Effect of molybde-
num addition on metastability of cubic g-uranium, J. Alloys Compd. 491 (1)
(2010) 753e760. https://doi.org/10.1016/j.jallcom.2009.11.060.
[5] S. Neogy, M.T. Saify, S.K. Jha, D. Srivastava, M.M. Hussain, G.K. Dey, R.P. Singh,
Microstructural study of gamma phase stability in U-9wt.%Mo alloy, J. Nucl.
Mater. 422 (1e3) (2012) 77e85. https://doi.org/10.1016/j.jnucmat.2011.12.
005.
[6] J.L. Snelgrove, R.F. Domagala, G.L. Hofman, T.C. Wiencek, G.L. Copeland,
R.W. Hobbs, R.L. Senn, The Use of U3Si2 Dispersed in Aluminum in Plate-type
Fuel Elements for Research and Test Reactors, Argonne National Laboratory,
1987. ANL/RERTR/TMe11.
[7] V.P. Sinha, G.P. Mishra, S. Pal, K.B. Khan, P.V. Hegde, G.J. Prasad, Development
of powder metallurgy technique for synthesis of U3Si2 dispersoid, J. Nucl.
Mater. 383 (1e2) (2008) 196e200. https://doi.org/10.1016/j.jnucmat.2008.08.
046.
[8] M. Ugajin, A. Itoh, M. Akabori, N. Ooka, Y. Nakakura, Irradiation behavior of
high uranium-density alloys in the plate fuels, J. Nucl. Mater. 254 (1) (1998)
78e83. https://doi.org/10.1016/S0022-3115(97)00347-4.
[9] T.C. Wiencek, S.L. Hayes, G.L. Hofman, M.K. Meyer, Irradiation behavior of
U6Mn-Al dispersion fuel elements, J. Nucl. Mater. 278 (2e3) (2000) 358e363.
https://doi.org/10.1016/S0022-3115(99)00280-9.
[10] R.F. Domagala, G.L. Copel, G.L. Hofman, Irradiation behavior of low-enriched
U6Fe-Al dispersion fuel elements, J. Nucl. Mater. 150 (2) (1987) 238e243.
[11] Development Status of Metallic, Dispersion and Non-oxide Advanced and
Alternative Fuels for Power and Research Reactors, IAEA-TECDOC-1374,
Vienna, 2003, p. 49.
[12] K.H. Kim, D.B. Lee, C.K. Kim, G.E. Hofman, K.W. Paik, Characterization of U-2
wt% Mo and U-10 wt% Mo alloy powders prepared by centrifugal atomization,
J. Nucl. Mater. 245 (2e3) (1997) 179e184. https://doi.org/10.1016/S0022-
3115(97)00011-1.
[13] A.N. Holden, Physical Metallurgy of Uranium, Addison-Wesley Publishing Co.,
Inc., USA, 1958.
[14] D.D. Keiser Jr., A.B. Robinson, J.-F. Jue, P. Medvedev, D.M. Wachs, M.R. Finlay,
Microstructural development in irradiated U-7Mo/6061 Al alloy matrix
dispersion fuel, J. Nucl. Mater. 393 (2) (2009) 311e320. https://doi.org/10.
1016/j.jnucmat.2009.06.018.
[15] V.P. Sinha, G.J. Prasad, P.V. Hegde, R. Keswani, C.B. Basak, S. Pal, G.P. Mishra,
Development, preparation and characterization of uranium molybdenum al-
loys for dispersion fuel application, J. Alloys Compd. 473 (1e2) (2009)
238e244. https://doi.org/10.1016/j.jallcom.2008.05.061.
[16] W.D. Wilkinson, in: Uranium Metallurgy, vol. 2, Interscience Publishers, 1962,
pp. 855e1230. Uranium Corrosion and Alloys.
[17] A. Travelli, Status and Progress of the RERTR Program in the Year 2004, IAEA-
CNe140, International Atomic Energy Agency (IAEA), 2004. https://inis.iaea.
org/collection/NCLCollectionStore/_Public/36/069/36069517.pdf.
[18] K.H. Kim, D.B. Lee, C.K. Kim, I.H. Kuk, K.W. Baek, Development of Uranium-
Silicide and U-Mo Alloy Fuels by Centrifugal Atomization, 1997. IAEA-
TECDOCe970, https://inis.iaea.org/collection/NCLCollectionStore/_Public/29/
000/29000279.pdf.
Fig. 11. The partial U-Mo phase diagram in the compositional range of interest [43].
5. Conclusions
U-10 wt%Mo alloy has been successfully prepared by R-D pro-
cess and complete stabilization of desired isotropic
g-phase has
been achieved at optimum process parameters of 1300 ꢁC and 4 h.
Understanding of the mechanism of the R-D process for the U-Mo
system has also been achieved and the mechanism has been pro-
posed based on experimental evidence. It has been identified that
atomistic diffusion of U or Mo, through the porous slag, between
two slag pockets, followed by inter-diffusion of U and Mo within a
single particle, are two dominant mechanisms adopted by the
system, for homogenization and reaching the equilibrium compo-
sition. The above two mechanisms are sped up if the temperature of
the charge (i.e. reduced metal þ Mo þ slag) is kept sufficiently
higher than the liquidus of the alloy of interest. The flow of liquid
between two slag pockets and the inter-diffusivities of each species
increase dramatically in the liquid state, due to its open structure.
At soaking temperatures below the liquidus of the particular alloy,
only solid state diffusion of Mo in U will be prevalent, which is itself
sluggish in nature. This necessitates longer soaking periods to
achieve homogeneous and equilibrium composition. In addition,
the driving force for preferential solute movement ceases to exist
due to absence of liquid U. Therefore, it was preferred to choose a
soaking temperature above the liquidus temperature of the alloy, in
order to achieve homogeneity and equilibrium in practical time
spans.
Acknowledgements
[19] K. Tangri, G.I. Williams, Metastable phases in the uranium molybdenum
system and their origin, J. Nucl. Mater. 4 (2) (1961) 226e233. https://doi.org/
10.106/0022-3115(61)90129-5.
[20] S. Neogy, M.T. Saify, S.K. Jha, D. Srivastava, G.K. Dey, Ageing characteristics of
the metastable gamma phase in Ue9 wt.% Mo alloy: experimental observa-
tions and thermodynamic validation, Phil. Mag. 95 (26) (2015) 2866e2884.
The authors are grateful to Dr. Madangopal Krishnan, Associate
Director, Materials Group for providing the opportunity to embark
upon this research work and showing his keen interest in the work.
The authors are thankful to Shri Pramendra Singh, Shri U.B. Naik

120
S. Gupta et al. / Journal of Alloys and Compounds 791 (2019) 109e120
https://doi.org/10.1080/14786435.2015.1081300.
531.181.
[21] J.M. Park, H.J. Ryu, K.H. Kim, D.B. Lee, Y.S. Lee, J.S. Lee, B.S. Seong, C.K. Kim,
M. Cornen, Neutron diffraction analyses of Ue(6e10 wt.%) Mo alloy powders
fabricated by centrifugal atomization, J. Nucl. Mater. 397 (1e3) (2010) 27e30.
https://doi.org/10.1016/j.jnucmat.2009.11.026.
[22] J. Lisboa, J. Marin, M. Barrera, A. Navarrete, L. Olivares, Atomization of U-Mo
Particles under nitrogen atmosphere, World J. Nucl. Sci. Technol. 6 (2016)
43e52. https://doi.org/10.4236/wjnst.2016.61004.
[23] C.R. Clark, B.R. Muntifering, J.-F. Jue, Production and characterization of
atomized U-Mo powder by the rotating electrode process, in: 2007 Interna-
tional Meeting on Reduced Enrichment for Research and Test Reactors
(RERTR), INL/CON-07-13156, Idaho National Laboratory (INL), 2007. https://
inldigitallibrary.inl.gov/sites/sti/sti/3874558.pdf.
[24] M.K. Meyer, C.R. Clark, Fuel powder production from ductile uranium alloys,
in: 1998 International Meeting on Reduced Enrichment for Research and Test
Reactors (RERTR), INIS-BRe3884, Sao Paulo, Brazil, October 18 - 23, 1998.
https://www.rertr.anl.gov/Fuels98/CClark.pdf.
[33] D.L. Martin, Nickel-lanthanum alloy produced by a reduction-diffusion pro-
cess, U.S. Patent 3,883,346, issued May 13, 1975r.
[34] G. Deng, Q. Jing, X. Wang, G. He, X. Ye, Synthesis mechanism of Sm2Fe17 alloy
produced in reduction-diffusion process, J. Rare Earths 28 (2010) 420e424.
https://doi.org/10.1016/S1002-0721(10)60357-2.
[35] S. Okada, K. Suzuki, E. Node, K. Takagi, K. Ozaki, Y. Enokido, Preparation of
submicron-sized Sm2Fe17N3 fine powder with high coercivity by reduction-
diffusion process, J. Alloys Compd. 695 (2017) 1617e1623. https://doi.org/
10.1016/j.jallcom.2016.10.306.
[36] C.S. Teixeira, L. Caron, A. Anastasopol, S.W.H. Eijt, J.A. Lozano, E. Brück,
P.A.P. Wendhausen, A new feature of the reductionediffusion process applied
for the synthesis of magnetocaloric LaFe13ꢀxSix compounds, J. Alloys Compd.
541 (2012) 84e87. https://doi.org/10.1016/j.jallcom.2012.05.112.
[37] R.O. Suzuki, T. Ueki, M. Ikezawa, T.H. Okabe, T. Oishi, K. Ono, A fundamental
study on preparation of Al3Ti powders by calciothermic reduction of oxides,
Mater. Trans., JIM 32 (3) (1991) 272e277. https://doi.org/10.2320/
matertrans1989.32.272.
[38] R.O. Suzuki, K. Tatemoto, H. Kitagawa, Direct synthesis of the hydrogen
storage VeTi alloy powder from the oxides by calcium co-reduction, J. Alloys
Compd. 385 (1e2) (2004) 173e180. https://doi.org/10.1016/j.jallcom.2004.04.
137.
[39] M. Ilayaraja, L.J. Berchmans, S.R. Sankaranarayanan, Synthesis of Y-Ni alloy by
calciothermic reduction diffusion process, Metall. Mater. Eng. 21 (2) (2015)
65e72. https://doi.org/10.30544/96.
[25] J.P. Blackledge, G.G. Libowitz, in: W.M. Mueller (Ed.), Metal Hydrides, 1968.
[26] M. Durazzo, C.J. Rocha, F. Mestnik, M. Ricardo, N. Leal, Powder fabrication of U-
Mo alloys for nuclear dispersion fuels, in: International Nuclear Atlantic
Conference, INIS-BRe10680 Brazil, 2011 (2011).
[27] R. Cech, Rare earth intermetallic compounds by a calcium hydride reduction
diffusion process, U.S. Patent 3,748,193, issued July 24, 1973.
[28] T.S. Krishnan, Preparation of rare earth-cobalt magnet alloy by reduction-
diffusion process, Bull. Mater. Sci. 2 (3) (1980) 161e165.
[29] C.K. Gupta, Chemical Metallurgy, Wiley-VCH Verlag GmbH
&
Co. KGaA,
[40] A.S.M. Handbook, Powder metal technologies and applications, ASM Int. 7
(1998) 89e91.
Weinheim, 2003, pp. 383e384.
[30] G. Qi, M. Hino, A. Yazawa, Experimental study on the reduction-diffusion
process to produce FeeNd, FeeSm, CoeNd and CoeSm alloys, Mater. Trans.,
JIM 31 (6) (1990) 463e470. https://doi.org/10.2320/matertrans1989.31.463.
[31] J. Lin, S. Liu, X. Qian, J. Bayi, M. Su, Synthesis of Nd2Fe17 by reduction-diffusion
process, J. Alloys Compd. 238 (1e2) (1996) 113e115. https://doi.org/10.1016/
0925-8388(95)02130-2.
[41] J.-M. Oh, W. Kim, B.-K. Lee, C.-Y. Suh, H.-S. Kim, J.-W. Lim, New method for
preparation of low oxygen Mo powder by reduction of MoO3 using Ca,
Powder Metall. 57 (4) (2014) 291e294. https://doi.org/10.1179/
1743290113Y.0000000083.
[42] Octave Levenspiel, Chemical Reaction Engineering, John Wiley & Sons, Inc.,
1999, pp. 569e580.
ꢀ
[32] Eneida da G. Guilherme, H.R. Hechenberg, Jose Octavio A. Pascoal, Reduction-
[43] M.A. Steiner, E. Garlea, S.R. Agnew, Modeling solute segregation during the
diffusion preparation of Nd15Fe77B8, NdFe11Ti, NdFe10.5Mo1.5 and
NdFe10.75Mo1.25 alloys for magnets, Mater. Sci. Forum 530 (2006) 181e186.
Trans Tech Publications, https://doi.org/10.4028/www.scientific.net/MSF.530-
solidification of
https://doi.org/10.1016/j.jnucmat.2016.02.035.
g-phase U-Mo alloys, J. Nucl. Mater. 474 (2016) 105e112.