Full text of DOI:10.1557/JMR.2000.0014

Journal of
MATERIALS RESEARCH
Welcome
Comments
Help
Micropyretic synthesis of NiAl containing Ti and B
G.K. Dey and A. Arya
Materials Science Division, Bhabha Atomic Research Centre, Trombay, Mumbai 400 085, India
J.A. Sekhar
International Center for Micropyretics, University of Cincinnati, P.O. Box 210012,
Cincinnati, Ohio 45221
(Received 12 June 1998; accepted 30 September 1999)
The effect of alloying additions of Ti and B on the process of micropyretic synthesis
on NiAl and on the microstructure of the synthesized alloy was examined. It was
observed that the combustibility of the quaternary alloy is good despite the presence of
the alloying elements because of an additional combustion reaction between Ti and B.
The microstructure of the quaternary alloy was found to consist primarily of the NiAl
and Ti boride phases. The effect of preheating of the specimen prior to synthesis on
the process of synthesis was also examined. It was observed that preheating not only
can change the morphology of the phases but also influence the nature of the phases
present in the alloy. The mechanism of the formation of the two phase microstructure
during the synthesis from the elemental powders was established by stopping the
combustion front and by carrying out a detailed microstructural characterization of
regions around the stopped combustion front.
I. INTRODUCTION
ence the combustibility of alloys during micropyretic
synthesis in various ways depending on the nature of the
diluents and their amounts.9 Most of the earlier studies
have dealt with the effect of diluents which do not par-
ticipate in the combustion reaction.9,10 The present study
differed from the earlier ones in that it examined the
effect of additions of two alloying elements which par-
ticipated in the combustion reaction. The process param-
eters that were monitored were the temperature of the
combustion reaction and the velocity of the combustion
front. Despite the fact that quite a large amount of Ti was
used as a diluent, it was observed that the combustibility
of the quaternary alloy was quite good, primarily because
of an additional combustion reaction leading to the for-
mation of titanium boride phases.
Earlier studies have shown that preheating the speci-
men prior to synthesis not only can influence the com-
bustion temperature and the combustion velocity but can
also have a significant influence on the microstructure of
the alloy.11–14 Besides, the influence of the diluents, the
effects of preheating on the combustion process and the
microstructure were examined in this study. It was ob-
served that preheating could not only change the mor-
phology of the phases but also affect the nature of
the phases present in the alloy and the distribution of
porosity.
NiAl has been the focus of several investigations be-
cause of its attractive properties.1 The lack of ductility
has been one of the major drawbacks of this alloy. In a
recent study, it has been demonstrated that this interme-
tallic compound can show significant ductility under cer-
tain conditions of impurity and imperfection content,
heat treatment, and surface perfection.2 Observations of
this nature are likely to enhance the interest in this alloy.
Several alloying additions have been tried out with a
view to improve the ductility of NiAl.1 Ti is known to
impart ductility to some phases with B2 structure.1 Ti
additions have been found to improve the fracture tough-
ness of NiAl.3 Additions of Ti beyond a certain limit lead
to the formation of the Ni2AlTi phase which can be used
for precipitation hardening of NiAl.4 Additions of B have
been found to be effective in improving the ductility of
some of the aluminides.5 With this point in view, the
changes in the properties of NiAl on B addition have
been examined by George et al.6 Composites containing
TiB2 in NiAl have also been examined in considerable
detail.7,8 However, studies involving micropyretically
synthesized NiAl containing Ti and B in amounts suffi-
cient for the formation of TiB2 as a second phase during
synthesis have not been carried out earlier.
In this study, the micropyretic synthesis of NiAl with
Ti and B additions was implemented and the effects of
these two diluents on the process and on the microstruc-
ture were analyzed. Diluents have been found to influ-
Many studies have been carried out to examine the
microstructural evolution during combustion synthesis
from elemental powders in the NiAl system.15,16 How-
ever, most of these studies have dealt with the formation
J. Mater. Res., Vol. 15, No. 1, Jan 2000
© 2000 Materials Research Society
63
http://journals.cambridge.org
Downloaded: 31 Mar 2015
IP address: 203.64.11.45

G.K. Dey et al.: Micropyretic synthesis of NiAl containing Ti and B
TABLE I. Purity and particle size of the powders used.
B. The adiabatic temperature (Tad), defined as the final
temperature attained by a system undergoing an adiabatic
reaction, is a good measure of the exothermicity of the
reaction and can be calculated by using the following
equation, assuming that the reactants are pure compo-
nents in their respective standard states:
Element
Purity
Powder size
NI
Al
Ti
B
99.7
99.5
99.0
95.9
3 ␮m
−325 mesh
−325 mesh
10 ␮m
T
⌬H =
⌬H
k͒ +
adC k͒ dT
,
͑
͑
p
͐
Α
Α
f,T
o
T
o
of a single-phase microstructure from the powders. In
this study, the mechanism of formation of a multiphase
microstructure from the powders were examined. The
mechanism of the evolution of this microstructure during
the synthesis from elemental powders was established by
stopping the combustion front and by carrying out a de-
tailed microstructural characterization of regions around
the stopped front.
k
k
where ⌬H is the enthalpy of the reaction, Cp is the heat
capacity of the product (k) and ⌬Hf,T is the heat of
formation of the product at temperature oTo. Since for an
adiabatic reaction, ⌬H ס 0, we have
T
−
⌬H
k͒ =
adC k͒ dT
.
͑
͑
p
͐
Α
Α
f,T
o
T
o
k
k
If the melting point of one of the products (␴), Tm, is
lower than the adiabatic temperature, then the above
equation takes a different form to include the enthalpy of
fusion (⌬Hm)17. In this case, the energy balance equation
will be given by
II. EXPERIMENTAL
High-purity powders (purity and size indicated in
Table I) were weighed accurately and mixed in a Spex
model 8000 ball mill for 20 min without the addition of
any liquid medium. Subsequently, the mixed powders
were pressed into rectangular bar specimens in a double
acting press at a pressure of 140 MPa. Induction heating
was used at one end of the pressed compacts for initiating
combustion in an inert atmosphere. In the case of speci-
mens preheated before combustion, the preheating was
effected in a furnace. The temperature of the specimen
could be monitored by using thermocouples. Very fine
holes were drilled in the green compact. The thermo-
couples were firmly affixed to these holes. The output
from the thermocouples was fed into a computer through
an analog to digital converter. The details of the setup
used for combustion are given elsewhere.14 X-ray dif-
fraction was carried out in a Siemens diffractometer
(model: D500) employing a Cu tube operating at 40 kV
and 30 mA. A Cambridge scanning electron microscope
(SEM), equipped with a Princeton Gamma-Tech detec-
tor, was used for the examination of the microstructure.
Specimens for transmission electron microscopy (TEM)
examination were made by dimple grinding followed by
ion milling in a Gatan ion mill. The thin foils were ex-
amined in a JEOL 2000 FX electron microscope and a
Philips CM 20 electron microscope equipped with an
EDAX microanalysis system.
T
−
⌬H
k͒ =
adC k, s͒ dT
͑
f,T
o
͑
p
͐
Α
Α
T
o
k
k°␴
T
+
mC ␴, s͒ dT + ⌬H ␴͒
͑ ͑
p m
͐
T
o
T
+
adC ␴, l dT
͒
,
͑
p
͐
T
m
where “s” and “l” correspond to solid and liquid states,
respectively. In the case where there is a partial melting
of one of the products, the melted fraction (fp) of ␴ can
be calculated by using the following equation:
T
−
⌬H
k =
͑ ͒
adC k, s dT + 1 − f
͒
͒
͑
p
͑
͐
Α
Α
f,T
o
p
T
o
k
␴
T
adC ␴, s͒ dT + f ⌬H ␴͒
͑
p
͑
m
͐
p
T
o
T
+ f
adC ␴, l͒ dT
.
͑
p
͐
p
T
m
This equation has also been used for the calculation of
the adiabatic temperature of off-stoichiometric alloys.18
The approach involves the estimation of the melting
point, the heat capacity, and the enthalpy of melting by
extrapolative methods. In the present study, the amount
of liquid formed was estimated for the following three
cases: (i) NiAl, (ii) NiAl with Ti, and (iii) NiAl with Ti
and B. The formation of Ni1−yAl1−yTi2y, assuming that
there was no formation of the Ni2AlTi phase, could be
represented by a single reaction of the type
III. RESULTS AND DISCUSSION
A. Theoretical estimation of adiabatic
temperature and melted fraction
(1 − y)Ni + (1 − y)Al + 2yTi ס Ni1−yAl1−yTi2y
.
In this section, thermodynamic considerations have
been applied to estimate the adiabatic temperature and
the melted fraction for the synthesis of NiAl (i) in the
presence of Ti only and (ii) in the presence of both Ti and
The quantity Tad was determined by solving the energy
balance equation numerically. The corresponding tem-
perature expansion coefficients for the heat capacity
64
J. Mater. Res., Vol. 15, No. 1, Jan 2000
Downloaded: 31 Mar 2015
http://journals.cambridge.org
IP address: 203.64.11.45

G.K. Dey et al.: Micropyretic synthesis of NiAl containing Ti and B
TABLE II. Temperature expansion coefficients A–D for the heat ca-
TABLE III. Combustion temperature, adiabatic temperature, propa-
gation velocity, and ignition temperature of the different alloys.
pacity (Cp) in J/(deg mol) of different stoichiometric compounds based
on Ni–Al–Ti–B. (The temperature expansion is Cp ס A + B + C/T2
DT2.)
+
Combustion
temperature temperature
Adiabatic
Propagation
velocity
(m/s)
Ignition
temperature
(K)
B
C
D
Temperature
Alloy
NiAl
Ni49Al49Ti2
Ni48.5Al48.5Ti3
Ni46Al46Ti8
(K)
(K)
Compound
A
(×103)
(×10−5
)
(×106)
(K)
1825
1840
1860
1815
1920
1910
1900
1892
1875
1990
0.20
0.22
0.25
0.19
0.30
870
Al (fcc)a
Al (liq.)
Ni (fcc)
Ni (fcc)
Ni (liq.)
NiAl (B2)
NiAl (liq.)
Ti (hex.)
Ti (bcc)b
TiB2 (hex.)
TiB2 (liq.)
31.38
31.37
11.17
20.54
38.91
41.84
71.13
29.94
30.84
56.38
108.78
−16.40
−3.60
20.75
298–934
934
298–631
631–1728
1728–
298–1912
1912–
298–1155
1155–1943
298–3498
3498–
37.78
10.08
3.18
15.40
875
873
Ni45Al45Ti8B2
13.81
−1.63
was assumed that only the TiB2 phase formed. This as-
sumption was still valid because the amount of TiB
formed was very small in the preheated specimens and
this phase was absent in the unpreheated specimens.
6.57
−8.87
25.86
−17.47
1.34
6.44
−3.35
aface-centered-cubic
bbody-centered-cubic
B. Effect of addition of Ti and B on the
combustion process of NiAl
The velocity of the combustion front and the combus-
tion temperature in the case of NiAl have been deter-
mined in several studies.22,23 The effect of the sizes of
the Ni and the Al particles on the combustion reaction
and temperature has been well studied,24,25 and it has
been shown that during the synthesis of NiAl the peak
temperature as well as the combustion velocity decreases
with increasing Ni particle size. The temperature profile
pertinent to the synthesis of NiAl from Ni powder of
3-␮m size and −325 # Al powder is shown in Fig. 1. The
profile shows a solitary peak, the peak temperature being
1825 K. The adiabatic temperature for the synthesis of
NiAl in the absence of preheating is 1910 K,11 the melt-
ing point of the alloy. Figure 1 also shows the effect of
addition of Ti powder of size −325 # to Ni and Al pow-
ders of the aforementioned sizes on the temperature pro-
files generated during combustion. The temperature
profiles tended to become broader as more of Ti was
added, there being a small peak in the beginning in some
cases. The small peak, presumably, corresponded to the
formation of the NiAl phase, and the broad hump after
the peak corresponded to the dissolution of Ti in NiAl.
There was a change in the shape of the temperature pro-
file as more of Ti was added. It could be observed from
these profiles that the maximum temperature of the tem-
perature profiles of the synthesized alloys went up as a
result of Ti addition up to a Ti content of 3 at.%. The
velocity of propagation of the front also showed an in-
were taken from the literature19–21 and are given in
Table II. The following temperature expansion of heat
capacity was employed in the calculations:
C
T2
C T ͒ = A + BT +
+ DT2
.
͑
p
Since the Ti content in the product was not more than
3 at.%, the heat capacity of the product Ni1−yAl1−yTi2y
was taken, as a first approximation, to be the same as that
of pure NiAl. Our calculated values of Tad and fp are
given in Tables III and IV, respectively. A temperature
step of 1.5 K was employed for solving the energy bal-
ance equation, and the energy values were equated on
either side within ±5.5 kJ/mol.
Since the addition of Ti and B leads to the occurrence
of another combustion reaction, a separate equation
needs to be written for this reaction. The formation of
Ni1−yAl1−yTi2y−2B2 was assumed to occur through the
following steps:
1 − y͒Ni + 1 − y͒Al = 1 − y͒NiAl
,
͑
͑
͑
͑
͑
2y − 2͒Ti + 2B = TiB + 2y − 3͒Ti
,
͑
2
1 − y͒NiAl + TiB + 2y − 3͒Ti
͑
2
= Ni1−yAl1−yTi2y−2B2
.
The initial temperature of the reactants for the second
step was quite high because of the fact that the latter
could start only after NiAl had formed. The contribution
from the second reaction, however, was not very sub-
stantial because of the small volume fraction of the TiB2
phase formed. It may be mentioned here that microstruc-
tural examination of specimens containing Ti and B and
synthesized after preheating revealed the presence of
small amount of the TiB phase in addition to the TiB2
phase. However, in the thermodynamic calculations, it
TABLE IV. Melted fraction of the product phase as a function of
composition and preheat.
Alloy
NiAl
NiAl
Ni45Al45Ti8B2
Ni45Al45Ti8B2
Preheat temperature (K)
Melted fraction
473
623
473
623
0.5
0.57
0.55
0.7
J. Mater. Res., Vol. 15, No. 1, Jan 2000
65
http://journals.cambridge.org
Downloaded: 31 Mar 2015
IP address: 203.64.11.45

G.K. Dey et al.: Micropyretic synthesis of NiAl containing Ti and B
powder as a diluent to the reactants in the synthesis of
NiAl.22 (ii) Diluents that participate in the combustion
reaction either by going into solid solution in the product
or by generating another phase. Such reactants can be
referred to as active diluents. The role of active diluents
on the process of combustion has not been examined in
as much detail as that of passive diluents.27 One of the
endeavors in this study was to examine the effect of the
addition of active diluents on the synthesis process. Ti,
as an active diluent, can influence the process by forming
a solid solution with the product of the reaction. If the
amount of Ti is higher, then the Ni2AlTi phase may form.
The addition of B as well as Ti can change the energy
balance by giving rise to a reaction which results in the
formation of borides of Ti. It has been shown that the
addition of active diluents in very small quantities can
significantly influence the rate of the combustion proc-
ess.14 However, when the amount of the diluents is in-
creased considerably, the rate of the combustion reaction
has been found to reduce substantially. The reason for
this observation seems to be the following: the addition
of the diluents in very small quantities can lead to a
reduction in the green porosity and hence can improve
particle contact substantially. The particles of the dilu-
ents being of a different size and size distribution tend to
fill the interstices in the green compact. This leads to an
increase in the combustibility of the specimen due to
better particle contact. It is also believed that the addition
of diluents can increase the fluidity of the first forming
liquid thereby enhancing the combustibility of the al-
loy.28 However, the addition of diluents in large quanti-
ties can lead to a reduction in the exothermicity of the
combustion reaction itself. The observed increase in the
reaction velocity and the reaction temperature as com-
pared to binary NiAl on the addition of small amounts of
Ti could be ascribed to the aforementioned reasons.
On the addition of large amounts of Ti, the combustion
temperature came down because of the fact that the
reduction in the exothermicity of the reaction out-
weighed the increase in the combustibility due to ki-
netic factors.
FIG. 1. Temperature profile generated during the micropyretic syn-
thesis of the different alloys: (1) NiAl, (2) Ni49Al49Ti2, (3)
Ni48.5Al48.5Ti3, (4) Ni46Al46Ti8, (5) Ni45Al45Ti8B2.
crease as compared to that observed in the case of binary
NiAl. The concurrent increase in the reaction velocity
with the combustion temperature on the addition of small
amounts of Ti was consistent with the fact that the size of
the powder particles used was in the kinetic regime
where a higher combustion temperature always leads to a
higher combustion velocity.26 However, as the Ti addi-
tion exceeded 3 at.%, the peak temperature as well as the
velocity of the combustion front came down. The adia-
batic temperatures for alloys containing different
amounts of Ti were calculated, and these values are
shown in Table III. It could be seen that the adiabatic
temperature decreased with an increase in the Ti content.
It has been the general impression that the addition of
diluents reduces the adiabatic temperature based prima-
rily on reaction thermodynamics.26 The effect of compo-
sition on the combustion temperature has been examined
in the case of NiAl, and it has been observed that the
combustion temperature is the highest at the exact stoi-
chiometric composition.27 In this study, the variation in
the combustion temperature with additions of Ti to NiAl
showed a trend slightly different from that shown by the
adiabatic temperature (Table III). All the values of the
combustion temperatures were lower than the corre-
sponding values of the adiabatic temperatures. The in-
crease in the combustion temperature with the addition of
small amounts of Ti could be attributed to kinetic reasons
to be discussed in the next section.
The temperature profile obtained on adding Ti and B
to NiAl in amounts such that the stoichiometry became
Ni45Al45Ti8B2 is shown in Fig. 1. It could be seen that
the first peak temperature as well as the maximum tem-
perature of the hump following it increased with the ad-
dition of B. There was also a change in the shape of the
temperature profile. The addition of B led to higher tem-
perature of the hump because the hump, presumably,
represented contributions from the formation of the Ti
boride phases. This contention is based on the observa-
tions reported in the section on the mechanism of syn-
thesis of this alloy where it has been shown that the
formation of the Ti boride phases started after the for-
mation of the NiAl phase. In addition to an increase in
The effect of the addition of diluents on the rate of the
combustion process has been examined in some stud-
ies.25,26 The diluents can be divided into two categories
depending on their role in the combustion process. (i)
Those that do not participate in the combustion reaction
and play the role of enthalpy absorbers. These diluents
are passive diluents. These are mostly the products of the
combustion reaction, for instance, the addition of NiAl
66
J. Mater. Res., Vol. 15, No. 1, Jan 2000
Downloaded: 31 Mar 2015
http://journals.cambridge.org
IP address: 203.64.11.45

G.K. Dey et al.: Micropyretic synthesis of NiAl containing Ti and B
the temperature of combustion, an increase in the veloc-
phology from equiaxed to dendritic. In addition to the
morphology of the grains, preheating has also been found
to influence the extent of segregation of the alloying
additions and phase formation at the grain boundaries.14
In the present work, similar investigations were carried
out on NiAl containing Ti and B.
ity of the front was also observed. The adiabatic tem-
perature for the formation of the TiB2 phase is 3190 K.
However, it should be noted that the temperature of the
specimen would never rise to such a high value because
of the fact that the amount of the TiB2 phase forming
would be very small. Nevertheless, there would be a
noticeable contribution to the overall temperature of the
specimen due to this reaction. In the previous section, the
adiabatic temperature was calculated for the alloy in
which the formation of TiB2 also occurs. It could be seen
from Table III that the adiabatic temperature went up
because of the occurrence of the second combustion re-
action leading to the formation of the TiB2 phase.
Extensive studies have been carried out on the forma-
tion of borides.29,30 It has been observed that the forma-
tion of many of the borides, particularly those of the MeB
type, occurs in two steps. The first step is the formation
of the MeB2 phase.29 The MeB2 phase then reacts with
more of Me to yield the MeB phase. The formation of
MeB2 has been found to be associated with the occur-
rence of a single-stage thermal profile whereas the for-
mation of MeB is associated with a two-stage profile.30
In this study the profiles due to the formation of the Ti
boride phases and the NiAl phase were superimposed on
each other. As a result, it was very difficult to make out
the nature of the thermal profile originating due to the Ti
boride phase formation alone. The influence of the boride
phases on the overall profile was small because of the
small volume fraction of these phases. It may be noted
that specimens with and without preheating showed a
difference in the phase content in terms of the borides.
1. Microstructure without preheat
Effect of diluents on the microstructure of alloys
i. Effect of Ti
The solubility of Ti in equiatomic NiAl is about
10 at.%.31,32 If this limit is exceeded, the formation of the
Ni2AlTi phase is expected to occur. The structures of the
NiAl and the Ni2AlTi phases are very similar. Ni2AlTi is
also known as the Heusler or BЈ phase and forms by
further ordering on the Al sites of the CsCl (B2) type
NiAl structure brought about the Ti atoms.4 In this study,
x-ray diffractograms of specimens containing 8 at.% Ti
could be indexed in terms of the B2 structure with a
lattice parameter a ס 0.290 nm, which was slightly dif-
ferent from that of unalloyed NiAl. Ti is likely to sub-
stitute for Ni as well as the Al in the NiAl lattice.33 In
either case, the substitution would lead to change in the
lattice parameter because of the differences in the sizes of
the atoms. In this study, the amount of Ti added was less
than that required for the formation of the Ni2AlTi phase.
The microstructure of the alloy containing Ti as the al-
loying addition was found to comprise of a single phase.
The morphology of the grains in alloys containing up to
3 at.% Ti continued to be equiaxed like in binary NiAl.
However, in alloys which contained higher amounts of
Ti, the grains were dendritic in appearance. The change
in microstructure with Ti content could be attributed to
the increase in the volume fraction of the melted phase.
C. Microstructures and phases
It has been shown in recent studies that the extent of
preheating can significantly influence the microstructure
of micropyretically synthesized alloys in terms of the
morphology of the grains as well as the nature of the
phases present.14 Calculations have shown that preheat
affects the melted volume fraction as evidenced from the
thermodynamic energy balance (Table IV). An increase
in the melted fraction is likely to influence the solidifi-
cation process as well. It has been experimentally ob-
served that preheating increases the velocity of the
combustion front. These factors, viz., a higher amount of
liquid and a higher velocity, will have an effect on the
microstructure of the alloy. A detailed account of how
the extent of preheat influences the aforementioned pa-
rameters in the case of NiAl with Fe, Cr, and V additions
has been given in a recent publication.14 In that work, the
microstructures developing with and without preheating
of the specimen before combustion have been examined
in considerable detail and it has been observed that in-
creasing amounts of preheating transforms the grain mor-
ii. Effect of Ti and B
Addition of B to NiAl has been carried out with a view
to strengthen the grain boundaries.6 B atoms go to the
interstitial positions in the lattice.34 In the alloys contain-
ing Ti and B, some amount of B goes to the NiAl lattice
and the remaining participates in the formation of the Ti
boride phases. A typical x-ray diffraction pattern ob-
tained from a specimen containing Ti and B is shown in
Fig. 2(a). It could be seen that this diffractogram shows
peaks form the matrix B2 phase and the TiB2 phase. The
(1010) peak is the most intense peak from TiB2 as per the
x-ray powder diffraction data.35 The very small intensity
of this peak suggested that the amount of the TiB2 phase
was very small. This was consistent with the TEM ob-
servations which indicated that this phase was present as
small particles at the B2 grain boundaries. The presence
of the TiB phase could not be detected by x-ray diffrac-
tion in the specimens synthesized without preheating.
J. Mater. Res., Vol. 15, No. 1, Jan 2000
Downloaded: 31 Mar 2015
67
http://journals.cambridge.org
IP address: 203.64.11.45

G.K. Dey et al.: Micropyretic synthesis of NiAl containing Ti and B
FIG. 2. (a) X-ray diffractogram from the Ni45Al45Ti8B2 phase in the unpreheated condition, (b) microstructure of the Ni45Al45Ti8B2 specimen
in the unpreheated condition, and (c, d) EDS spectra from the grain interior and the grain boundary, respectively.
The unpreheated specimens of NiAl with Ti and B
additions showed a multiphase microstructure compris-
ing a matrix of the NiAl phase [Fig. 2(b)]. The grains of
this phase did not have an equiaxed morphology as seen
in the case of unalloyed NiAl. Instead, the grains ap-
peared to have a spherical morphology indicating that
the images corresponded to sections of large dendrites.
SEM along with energy dispersive spectroscopy (EDS)
was used to determine the amount of Ti in the grains
of unpreheated specimens. It could be seen that the pres-
ence of Ti in NiAl was significant but not to the extent
expected from the composition of the alloy. This was,
presumably, due to the formation of the TiB2 phase
which consumed some amount of available Ti. Figures
2(c) and 2(d) show typical EDS spectra from the grain
interior and the grain boundary, respectively.
It is interesting to note that the morphology of the
grains was not equiaxed. Instead, these were curved
boundaries giving the grains a spherical shape. Spherical
grain morphologies could be obtained in binary NiAl as
the material was preheated. It appeared that the effect of
adding Ti and B on the microstructure of micropyreti-
cally synthesized NiAl was the same as that obtained by
preheating of the binary alloy. Preheating would add an
68
J. Mater. Res., Vol. 15, No. 1, Jan 2000
Downloaded: 31 Mar 2015
http://journals.cambridge.org
IP address: 203.64.11.45

G.K. Dey et al.: Micropyretic synthesis of NiAl containing Ti and B
additional enthalpy to that obtained by the combustion
b ס 0.30 nm, and c ס 0.445 nm. These values are very
close to the reported lattice parameters of the TiB
phase.35 The presence of the TiB and the TiB2 could be
confirmed by electron diffraction. The lattice parameter
of the NiAl phase in the unpreheated specimens
(0.290 nm) was larger than that in the preheated speci-
mens (0.285 nm). This was because of the fact that in the
unpreheated specimen the postsynthesis solidification
occurred at a very rapid rate, leading to a supersaturation
of the NiAl lattice with Ti and B atoms. In the case of the
preheated specimen, the extent of melting was much
larger, and as a consequence of this, solidification oc-
curred at much slower rate, allowing Ti and B atoms to
segregate to form the TiB2 phase. Similar observations
have been made in the case of NiAl with Fe, Cr, and V
additions, where it has been seen that in the preheated
specimen the microstructure shows signs of segrega-
tion.14 A comparison of the volume fraction of the pores
in the unpreheated specimens with that in the preheated
specimens indicated that the latter had a smaller number
of pores. However, the average pore size in the former
was smaller.
In the case of the NiAl alloy with Ti and B additions,
the specimens preheated to 473 K before synthesis
showed a multiphase microstructure with the NiAl phase
as the matrix phase [Fig. 3(b)]. The grains of this phase
had a spherical morphology, and at the boundaries of
these grains precipitates of a second phase could be seen.
Some of the grains of the NiAl phase showed faults
running across the length of the grain. In many grains the
faults were found to be running in two directions.
Figure 3(c) shows a TEM micrograph showing a faulted
region. Electron diffraction patterns from these regions
did not show any additional feature as compared to those
obtained from regions having no faults. Figure 3(d)
shows a TEM micrograph of the grain boundary region
of a specimen preheated at 473 K. The presence of par-
ticles of various morphologies at the grain boundaries
could be seen. Electron diffraction studies indicated that
some of these particles were of the TiB2 phases whereas
others were of the TiB phase. A typical electron diffrac-
tion pattern containing reflections from each of these
phases is shown in Fig. 3(e). The disposition of the spots
could be seen to be consistent with the hexagonal struc-
ture of the TiB2 phase and the orthorhombic structure of
the TiB phase. The morphology of the TiB2 particles
resembled that of TiB2 precipitates forming in the Ti
aluminides with B additions.37 No well-defined crystal-
lographic orientation relationship could be observed be-
tween the boride phases and the NiAl matrix. A mottling
of the matrix phase could also be seen [Fig. 3(f)]. This
mottling was, presumably, due to the presence of Ti in
the matrix phase.
reaction. This would result in a higher volume fraction of
the melted product which, as a consequence, would so-
lidify into a dendritic product. In the case of NiAl with B
and Ti additions, the effect of the alloying was to intro-
duce additional enthalpy through the formation of the Ti
boride phases. It could be inferred that, due to the higher
heat input, the microstructure changed to the spherical or
the dendritic form as compared to the equiaxed micro-
structure seen in the binary NiAl material. The extent of
liquid formation in the case of NiAl with Ti and B ad-
ditions would be larger than that in the case of binary
NiAl (Table IV); this is a factor which influenced the
microstructure of the alloy.
2. Effect of preheating on microstructure and nature
of phases
It has been shown by Lakshmikanta et al.26 that the
initial temperature of the compact prior to ignition has an
effect on the combustion velocity as well as the combus-
tion temperature. Both these quantities increase with an
increase in the initial temperature. The increase in the
initial temperature, however, is not additive in determin-
ing the final combustion temperature. This is primarily
because of the fact that the specific heat of the product is
higher than that of the reactants and the effect of the
thermal conductivity on the solution is enhanced because
of the steeper temperature gradients.36 Since the initial
temperature affects the combustion temperature and the
melted fraction, the cooling rate of the specimen in the
post combustion regime gets affected and this exerts an
influence on the microstructure. The initial temperature
also influences the velocity of the combustion front. In
the following section, the effect of preheat on the alloy
microstructure has been examined. The increased veloc-
ity of the combustion front ensures that the entire speci-
men is taken to a higher temperature rather quickly. In
order to examine the effect of preheat, combustion of the
specimens was carried out after heating these to either
473 or 623 K.
X-ray diffractograms obtained from the samples pre-
heated at these two different temperatures were similar in
appearance and showed that the major phase forming
was the phase having the B2 structure. A representative
x-ray diffractogram from a specimen preheated to a tem-
perature of 623 K is shown in Fig. 3(a). This diffracto-
gram shows the presence of the B2-NiAl phase as the
major phase. In addition, the presence of the TiB phase
could also be established since the major peaks expected
from this phase35 were found to be present. The volume
fraction of this phase, as confirmed by electron micros-
copy, was very small, and thus, the peak intensities were
found to be weak. The TiB phase has an orthorhombic
structure, and the following lattice parameters could be
determined for this phase in this study: a ס 0.602 nm,
In the specimens preheated to 623 K before combus-
tion, the microstructure had a dendritic appearance
J. Mater. Res., Vol. 15, No. 1, Jan 2000
Downloaded: 31 Mar 2015
69
http://journals.cambridge.org
IP address: 203.64.11.45

G.K. Dey et al.: Micropyretic synthesis of NiAl containing Ti and B
FIG. 3. (a) X-ray diffractogram from the Ni45Al45Ti8B2 specimen obtained after synthesis of preheated specimen (preheating temperature 473 K).
(b) Microstructure of the Ni45Al45Ti8B2 specimen obtained after synthesis of preheated specimens (preheating temperature 473 K). (c) TEM
microstructure showing the faulted region. (d) TEM micrograph showing the presence of particles of a second phase at the grain boundaries. (e)
electron diffraction pattern from the TiB2 and the TiB phases (TiB2 zone axis [2243] and TiB zone axis [011]; the 100 reflection has appeared
in the later due to double diffraction). (f ) TEM micrograph showing mottling of the matrix phase. (g) Complete dendritic appearance of the
microstructure in a specimen preheated to a temperature 623 K.
70
J. Mater. Res., Vol. 15, No. 1, Jan 2000
http://journals.cambridge.org
Downloaded: 31 Mar 2015
IP address: 203.64.11.45

G.K. Dey et al.: Micropyretic synthesis of NiAl containing Ti and B
[Fig. 3(g)], with the dendrite boundaries decorated with a
alloys led to an enhancement in the combustion tempera-
ture which went beyond the melting point of the alloys,
pushing the reaction in the category of high-temperature
synthesis.
second phase. The dendritic matrix phase was not found
to have any faults in them. It could be inferred from these
observations that the extent of preheating was largely
responsible for changing the morphology of the phases.
This was also related to the fact that the Ti and B contents
in the matrix phase depended on the extent of preheat
used. The formation of the TiB phase would be more
likely in the preheated specimen because the combustion
peak temperature was higher than that in the case of the
unpreheated specimen. The rate of cooling of the speci-
men after the combustion wave had passed could be as-
certained from the cooling part of the combustion curve.
It could be observed that, in the case of the preheated
specimen, the specimen remained at higher temperatures
for much longer periods as compared to the unpreheated
specimen (Fig. 4). This slower cooling led to the devel-
opment of a microstructure totally different from that
seen in the case of the unpreheated specimen. The slow
cooling allowed a longer time for diffusion to occur and
promoted the formation of equilibrium phases. It has
been observed that, in the case of NiAl alloyed with Fe,
Cr, and V, preheating of the specimen leads to the for-
mation of phases at the grain boundaries which are absent
in the unpreheated specimen.14
The process of synthesis in the case of any alloy can be
categorized as high- or low-temperature synthesis de-
pending upon where the combustion temperature lies
with respect to the temperature of melting of the alloy.38
In this study, it could be seen that the temperature of
combustion was lower than the melting point of the al-
loys when no preheating was used. The adiabatic tem-
perature came down with the addition of the alloying
elements, but so did the melting point. Small alloying
additions as well as rise in the initial temperature of the
D. Mechanism of synthesis
With Ti and B additions to NiAl, the evolution of the
microstructure involved the formation of a boride phase
in addition to the NiAl phase. In order to study the
mechanism of microstructure formation during the syn-
thesis process in the specimen, it was necessary to freeze
the reaction front. The combustion front could be frozen
by several methods. Though the extent of the reaction
can be altered by changing the process conditions like the
green density and the preheating temperature and some
intermediate compounds forming during the synthesis
can be isolated, this approach fails to provide a detailed
understanding of the mechanism of formation of the final
product.38 Quenching the reaction front during its pas-
sage through the sample is a very powerful tool for study-
ing the mechanism and the different stages of synthesis.
This approach has been used by Lebrat et al.39 to study
the formation of Ni3Al from its alloying constituents. In
the present study, the front could be made to stop by
sandwiching a part of the specimen between two copper
plates. The details of this arrangement are given else-
where.14 In order to examine the sequence of microstruc-
ture formation from the powder particles, a detailed
microstructural characterization of the different regions
in a specimen where the reaction front has been stopped
was carried out. The microstructure of the specimen was
examined away from the front on the powder side, at the
front, and finally on the intermetallic side of the front.
Figure 5(a) shows the frozen combustion front with the
arrow indicating the approximate position of the front.
Depending upon the extent of reaction, the specimen
could be divided into three distinct regions—unreacted
region, partially reacted region, and fully reacted region.
Whereas there was a well-defined boundary between
the fully reacted and the partially reacted regions,
the boundary between the unreacted and the partially
reacted regions was not so distinct. The microstructures
of these three different regions are described in the
following.
1. Unreacted region
Figure 5(b) shows a representative micrograph of the
unreacted region at a considerable distance from the
boundary. The outline of the Al particles was very dif-
ficult to establish. This, presumably, was due to the fact
that the soft Al particles had undergone a significant
amount of deformation. Al particles were found to be
present around almost every single Ni particle with the
FIG. 4. Temperature profiles generated after synthesizing the speci-
mens (1) without preheat and (2) with a preheat of 673 K.
J. Mater. Res., Vol. 15, No. 1, Jan 2000
Downloaded: 31 Mar 2015
71
http://journals.cambridge.org
IP address: 203.64.11.45

G.K. Dey et al.: Micropyretic synthesis of NiAl containing Ti and B
FIG. 5. (a) Micrograph showing the frozen combustion front with the arrow indicating the approximate position of the front. (b) Micrograph
showing the unreacted region at a considerable distance from the boundary. (c) Micrograph showing the fully reacted region and part of the
partially reacted region. (d) Micrograph showing a region where the molten Al has undergone some reaction with the Ni particles forming Al-rich
intermetallic phase. (e) EDS profile from a region shown by an arrow in (d). (f ) The microstructure of a region closer to the stopped reaction front
(unconsumed Ni particles along with Al-rich phase mixture can be seen in this micrograph). (g) EDS profiles from a region shown by an arrow
in (f). (h) NiAl region showing the presence of the unreacted Ti particle.
72
J. Mater. Res., Vol. 15, No. 1, Jan 2000
http://journals.cambridge.org
Downloaded: 31 Mar 2015
IP address: 203.64.11.45

G.K. Dey et al.: Micropyretic synthesis of NiAl containing Ti and B
result that very few Ni particles touched each other. The
The EDS profile shows, in addition, that the Al-rich in-
termetallic compound had Ti in solid solution. Also seen
in this region are some large particles of Ni; the molten
Al had undergone reaction only on the surface of these
particles. Some of the Ti particles could be seen to be
maintaining their distinct identity. The Al-rich phases
formed appeared to have been quickly consumed on
moving toward the fully reacted region, to yield phases
with higher Ni contents. Ti continued to be in solid so-
lution in these intermediate intermetallic compounds.
Figure 5(f) shows the microstructure of a region closer to
the stopped reaction front. Unconsumed Ni particles
along with an Al-rich phase mixture could be seen. The
phase mixture was found to comprise an NiAl3 phase and
an Al-rich phase containing Ni and Ti [indicated by an
arrow in Fig. 5(f) and in the EDS profile shown in
Fig. 5(g)]. As the temperature increased, the molten Al
reacted further with Ni and the Al-rich phases to yield
phases containing higher amounts of Ni. Still closer to
the fully reacted region, NiAl regions could be encoun-
tered [Fig. 5(h)]. These regions were not found to be
surrounded by unreacted Ni particles. However, the oc-
currence of unreacted Ti particles could be readily seen
around the NiAl grains on the partially reacted side of the
specimen in these regions. These observations lent fur-
ther support to the contention that NiAl formed before
the formation of TiB2 commenced. The NiAl grains were
surrounded by a region in which Ti was more or less
uniformly distributed on the reacted side of the specimen.
This was the region where the formation of TiB2 had just
started. The heat generated by this reaction could lead to
an increase in the volume fraction of the liquid NiAl
formed and change the morphology of the grains from
equiaxed to dendritic.
Ti particles also were surrounded by Al particles. The
distribution of B could not be established because of its
small amount and the limitation of the EDS system being
used. It could be expected that B was distributed uni-
formly. The knowledge of the spatial distribution of the
different elements helps in ascertaining the types of re-
actions and the sequence of their occurrence. For in-
stance, it was very likely that Ni and Al particles reacted
first because these were in contact with each other almost
everywhere. Similarly, Ti was likely to react with Al
because the particles of Ti were surrounded by those of
Al. The reaction between Ni and Ti was less likely to
occur because of the fact that particles of these two ele-
ments were rarely in contact. Again, as Al was also the
lowest melting element present, it was likely to melt and
spread around and encircle even those particles with
which it was not in contact to begin with. The reactions
most likely to occur were, therefore, those between (i) Ni
and Al and (ii) Ti and Al. Out of these two, the reaction
between Ni and Al is associated with a higher adiabatic
temperature.15 Depending upon the distribution of B, the
following reactions could also be expected to occur: Ti–B,
Ni–B, and Al–B. Out of these, the reaction between Ti
and B was the most exothermic and has the highest adia-
batic temperature.16 Considering the two most readily
occurring reactions in the two groups, one could see that
even though the adiabatic temperature for the Ti–B re-
action was higher than that for the Ni–Al reaction, the
former had a lower initiation temperature and hence was
likely to start first. The aforementioned points have to be
kept in mind and compared with the actual sequence of
events occurring in the partially reacted region to be
discussed in the next section.
2. Partially reacted region
3. Fully reacted region
As mentioned earlier, although a well-defined bound-
ary seemed to exist between the fully reacted and the
partially reacted regions [Fig. 5(c)], a similar well-
defined boundary could not be seen between the partially
reacted and the unreacted regions. This was, presumably,
because of the fact that Al melted and spread across to a
considerable distance and surrounded the particles of Ni
and Ti. In this region of the specimen, the entrapment of
Ni and Ti particles was limited to the coating of these
particles with Al without the formation of any new phase.
However, on approachment of the fully reacted region,
the formation of the second phase particles could be seen
to commence. Figure 5(d) shows a region where the mol-
ten Al had undergone some reaction with the Ni particles
forming an Al-rich intermetallic phase. Figure 5(e)
shows the EDS profile from such a region. The stoichi-
ometry of this region did not correspond to that of any of
the known phases shown in the Ni–Al phase diagram.
In this region the microstructure was very similar to
that of a fully combusted specimen. The grains closer to
the front were smaller than those away from it. This was,
presumably, because of the fact that these grains did not
coarsen as the temperature of this region fell sharply on
account of the front being stopped.
On the basis of the observations made in this study, the
following sequence of microstructure formation could be
envisaged. The Al particles melted first and surrounded
the Ni, Ti, and B particles. Initially, Al-rich phases
formed and dissolved some amount of Ti and B. As the
temperature increased, the molten Al reacted with Ni and
the Al-rich phases to form phases containing higher
amounts of Ni and small amounts of Ti and B. With
further increase in the temperature, the Ti–B reaction
started and generated considerable amounts of heat. The
initiation of this reaction is likely to influence the kinet-
ics of formation of the NiAl phase because of the asso-
J. Mater. Res., Vol. 15, No. 1, Jan 2000
Downloaded: 31 Mar 2015
73
http://journals.cambridge.org
IP address: 203.64.11.45

G.K. Dey et al.: Micropyretic synthesis of NiAl containing Ti and B
ciated increase in temperature. The effect would have
been more substantial had the amount of TiB2 formed
been larger.
higher and higher amounts of Ni, ultimately leading to
the formation of the NiAl phase. Ti and B initially got
incorporated in the Al-rich phases and were later passed
on to the phases which became progressively richer in
Ni. The TiB2 phase formed only after the formation of
the NiAl phase was complete.
A comparison can be made between the mechanism of
synthesis of the single-phase NiAl alloy and the alloy
studied in the present work where the final microstruc-
ture had more than one phase. It has been observed that,
in the case of NiAl, the first stage in the synthesis is
melting of Al. The molten Al reacts with Ni to form
Al-rich phases first. These then react with more of Ni,
leading to the formation of phases which contain higher
and higher amounts of Ni such as NiAl3 and Ni2Al3. It
may be noted that the mechanism of formation of NiAl
from elemental powders reported in several studies14,15 is
very similar to the one described here for the two-phase
alloy containing Ti and B additions. This is not surprising
in view of the fact that the addition of Ti alone or along
with B was not found to influence the ignition tempera-
ture of the compacts (Table III) of these alloys. This
indicated that the mechanism of formation of the NiAl
phase remained similar and that this was the first phase to
form. Ti and B got incorporated in the intermediate
phases as these formed. The formation of the TiB2 phase
occurs only after the NiAl phase formation was nearing
completion.
REFERENCES
1. R.D. Noebe, R.R. Bowman, and M.V. Nathal, Int. Mater. Rev. 38,
193 (1993).
2. V.I. Levit, I.A. Bul, J. Hu, and M.J. Kaufmann, Scr. Mater. 34,
1925 (1996).
3. W.A. Kayseer, R. Laag, J.C. Murray, and G.E. Petzow, Int. J.
Powder Met. 27, 43 (1991).
4. R. Darolia, J. Met. 3, 44 (1991).
5. C.T. Liu and C.L. White, Scr. Mater. 20, 1613 (1986).
6. E.P. George and C.T. Liu, J. Mater. Res. 5, 754 (1990).
7. T. Cheng and H.M. Flower, Acta Met. Mater. 42, 1399 (1994).
8. D.J. Whittenberger, R.K. Viswanadham, S.K. Mannan, and
K.S. Kumar, J. Mater. Res. 4, 1164 (1989).
9. H.P. Li and J.A. Sekhar, J. Mater. Sci. 30, 4628 (1995).
10. J. Subramanyan and M. Vijayakumar, J. Mater. Sci. 27, 6249
(1992).
11. K.A. Philopt, Z.A. Munir, and J.B. Holt, J. Mater. Sci. 22, 159
(1987).
12. N.A. Martirosyan, S.K. Dolukhanyanand, and A.G. Mershanov,
Combust. Explos. Shock Waves (Engl. Transl.) 19, 569 (1985).
13. V.V. Boldyrev, V.V. Alekshandrov, M.A. Korchagin, B.P. Tolo-
chko, S.N. Guesenko, A.S. Sokolov, M.A. Shermonov, and
N.Z. Lyakhov, Dokl. Akad. Nauk. USSR 259, 1127 (1981).
14. G.K. Dey and J.A. Sekhar, Met. Mater. Trans. 28, 905 (1997).
15. V.V. Aleksandrov and M.A. Korchagin, Combust. Explos. Shock
Waves (Engl. Transl.) 23, 557 (1987).
16. V.V. Alekshandrov, M.A. Korchagin, B.P. Tolchonko,
S.N. Gusenko, A.S. Sokolov, and M.A. Sheromov, Combust. Ex-
plos. Shock Waves (Engl. Transl.) 19, 30 (1984).
17. D.C. Dunand, J.L. Sommer, and A. Mortensen, in Processing and
Fabrication of Advanced Materials for High Temperature Appli-
cations II, edited by T.A. Srivatsan and R.A. Ravi (TMS, War-
rendale, PA, 1993), p. 635.
18. D.C. Dunand, J.L. Sommer, and A. Mortensen, Met. Trans. 24A,
2161 (1992).
19. O. Kubaschewski and C.B. Alcock, in Metallurgical Thermo-
chemistry, International Series on Materials Science and Technol-
ogy, edited by G.V. Raynor (Pergamon Press, Oxford, United
Kingdom, 1989).
20. R. Hultgren, P.D. Desai, D.T. Hawkins, M. Gleiser, and K. McKelly,
Selected Values of Thermodynamic Properties of Binary Alloys
(ASM, Metals Park, OH, 1973).
21. L. Kaufman and H. Bernstein, in Computer Calculation of Phase
Diagrams, Refractory Materials, A Series of Monographs Vol. 4,
edited by J.L. Margrave (Academic Press, New York, 1970).
22. H.P. Li and J.A. Sekhar, J. Mater. Res. 8, 2515 (1993).
23. H.P. Li and J.A. Sekhar, Mater. Sci. Eng. A160, 221 (1993).
24. H.P. Li and J.A. Sekhar, in Proceedings of the First International
Conference on Advanced Synthesis of Engineered Materials,
San Francisco, CA, edited by J.J. Moore, E.J. Lavernia, and
F.H. Froes (ASM International, Materials Park, OH, 1993), p. 25.
25. H.P. Li and J.A. Sekhar, J. Mater. Res. 10, 2471 (1995).
26. Hj. Feng and J.J. Moore, Prog. Mater. Sci. 39, 275 (1995).
27. Y.S. Naiborodenko and V.I. Itin, Combust. Explos. Shock Waves
(Engl. Transl.) 11, 734 (1995).
IV. CONCLUSIONS
On the basis of the observations made in this work, the
following conclusions could be drawn:
(1) The microstructure of the micropyretically synthe-
sized quaternary alloy comprised the NiAl and the TiB2
phases in the unpreheated specimens. The former was the
matrix phase where as the latter was present as a grain
boundary phase. No specific orientation relationship
between these phases was observed. The microstruc-
ture was different from that of binary NiAl because of
the additional heat generated by the reaction between
Ti and B.
(2) The extent of preheat had a major influence on the
microstructure and the nature of the phases present in the
micropyretically synthesized Ni45Al45Ti8B2 alloy. In
preheated specimens of this alloy, the formation of the
TiB phase occurred in addition to that of the TiB2 phase.
(3) A thermodynamic analysis of the process of syn-
thesis was carried out, and it could be inferred that, in the
alloys having Ti and B, the adiabatic temperature as well
as the melted fraction would be higher. This inference
was found to be consistent with the microstructural ob-
servations.
(4) The mechanism of synthesis of the two-phase mi-
crostructure involved the formation of the NiAl phase
first. This phase formed through a sequence starting with
the formation of the Al-rich phases. Ni reacted with the
Al-rich phases to yield progressively the phases having
74
J. Mater. Res., Vol. 15, No. 1, Jan 2000
Downloaded: 31 Mar 2015
http://journals.cambridge.org
IP address: 203.64.11.45

G.K. Dey et al.: Micropyretic synthesis of NiAl containing Ti and B
28. Z.A. Munir and U. Anselmi-Tamburini, Mater. Sci. Rep. 3, 277
34. T.G. Neih, J. Wadsworth, and C.T. Liu, Scr. Mater. 22, 1409
(1989).
(1988).
29. A.G. Mershanov and I.P. Borovinskaya, Dokl. Akad. Nauk. USSR
204, 429 (1972).
35. Powder Diffraction File, Alphabetical Indexes, International Cen-
ter for Diffraction Data (1994).
30. A.A. Zenin, A.G. Mershanov, and G.A. Nersisyan, Dokl. Phys.
Chem. 236, 973 (1980).
36. M.G. Lakshmikantha, A. Bhattacharya, and J.A. Sekhar, Met.
Trans. 23A, 23–34 (1992).
31. W.J. Boettinger, L.A. Bendersky, F.S. Biancaniello, and J.W. Cahn,
Mater. Sci. Eng. 98, 273 (1988).
37. M.E. Hyman, C. McCullough, C.G. Levi, and R. Mehrabian, Met.
Trans. 22A, 1647 (1991).
32. A. Raman and K. Schubert, Z. Metallkde. 56, 99 (1965).
33. R.S. Polvani, W.S. Tzeng, and P.R. Strutt, Met. Trans. 7, 33
(1976).
38. A. Verma and J.P. Lebrat, Chem. Eng. Sci. 47, 2179 (1992).
39. J.P. Lebart, A. Varma, and P.J. McGinn, J. Mater. Res. 9, 1184
(1994).
J. Mater. Res., Vol. 15, No. 1, Jan 2000
Downloaded: 31 Mar 2015
75
http://journals.cambridge.org
IP address: 203.64.11.45