Accepted Manuscript

Interfacial characterization in ductile refractory metals reinforced

MoSi2 based laminated composites

Manoj Kumar Jain, Jiten Das, J. Subrahmanyam, S. Ray

PII: S0263-4368(16)30253-0
DOI: doi: 10.1016/j.ijrmhm.2017.04.003
Reference: RMHM 4443
To appear in: International Journal of Refractory Metals and Hard Materials
Received date: 9 May 2016
Revised date: 31 March 2017
Accepted date: 11 April 2017

Please cite this article as: Manoj Kumar Jain, Jiten Das, J. Subrahmanyam, S. Ray ,
Interfacial characterization in ductile refractory metals reinforced MoSi2 based laminated
composites. The address for the corresponding author was captured as affiliation for all
authors. Please check if appropriate. Rmhm(2017), doi: 10.1016/j.ijrmhm.2017.04.003

This is a PDF file of an unedited manuscript that has been accepted for publication. As
a service to our customers we are providing this early version of the manuscript. The
manuscript will undergo copyediting, typesetting, and review of the resulting proof before
it is published in its final form. Please note that during the production process errors may
be discovered which could affect the content, and all legal disclaimers that apply to the
journal pertain.
 ACCEPTED MANUSCRIPT

Interfacial Characterization in Ductile Refractory Metals Reinforced

MoSi2 Based Laminated Composites
Manoj Kumar Jain, Jiten Das#, J Subrahmanyam, S Ray*

Defence Metallurgical Research Laboratory, Hyderabad

*Indian Institute of Technology, Roorkee
#Corresponding Author: dasjiten@rediffmail.com

ABSTRACT

PT
Efforts to develop MoSi2 for high temperature structural applications have been hampered by
its extreme brittleness at temperatures below 1000 C. In the present work, toughening of

RI
MoSi2 at lower temperatures has been explored using the ductile refractory metals as
reinforcements. Model tri-layer laminated composites were made using Mo, Ta and Nb

SC
sheets sandwiched in between two layers of MoSi2 based matrix. As these refractory metals
are thermodynamically unstable with MoSi2, interfacial reaction layers were observed
NU
between MoSi2 and the refractory metal sheets. The main reaction product was identified to
be as M5Si3 (M = Mo, Ta or Nb) consuming a part of the thickness of the ductile metal
sheets. The Ta sheet reinforced laminated composite exhibited the minimum thickness of
MA

the reaction zone (~ 10 µm) suggesting that the Ta has better chemical compatibility with
MoSi2 as compared to Mo and Nb. Micro-hardness measurements made in the reaction
zone revealed that the reaction products in all the three laminated composites were hard and
D

brittle in nature. A quantitative idea of the thickness and micro-hardness values of the
E

interfacial reaction layers is very useful to give some important guidelines to obtain the
PT

enhanced mechanical properties in MoSi2 based laminated composites.

Keywords : molybdenum di-silicide, refractory sheet laminated composite, interface
CE

INTRODUCTION
AC

Silicides are a new class of materials, which offers the advantages of a ceramic as
well as of a metal. There is a large number of silicide compounds out of which, MoSi2 has
been proposed as a model material for high temperature structural applications due to its
unique physical and mechanical properties. MoSi2 has the potential of meeting structural and
oxidation requirements up to a temperature of 1600 °C. A generic definition of a ceramic is
that, it is a solid, iono-covalent, inorganic compound. MoSi2 deviates from this definition
because its atomic bonding is a mixture of covalent and metallic bonding [1]. It behaves like
brittle ceramics at room temperature while it behaves like metals at high temperatures,
undergoing plastic deformation and creep. MoSi2 has better ductility than competitive
ceramics for use at and above 1200 C [2].
 ACCEPTED MANUSCRIPT

Almost all components used for high temperature structural applications are
subjected to some degree of stresses at low temperatures, which could be catastrophic for
materials having very poor toughness at lower temperatures. Efforts to develop MoSi2 have
been hampered by its extreme brittleness at temperatures below 1000 C, coupled with
relatively low creep resistance. MoSi2 is also prone to a low temperature intergranular
oxidation problem known as pesting which disintegrates the exposed material along grain
boundaries rapidly. However, pesting behavior does not occur in materials of density greater

PT
than 98 %, crack or pore-free samples and stress free single crystals of MoSi2. Many
researchers addresses the Pesting problem by addition of Si3N4 [3], SiC [4] and Re [5]
addition.However, Pesting problem still be a big issue till now. For its effective use as a high

RI
temperature structural material, it becomes necessary to toughen the material at lower

SC
temperatures (within the MoSi2 brittle regime) while simultaneously improving the strength at
higher temperatures. To solve these twin problems MoSi2 matrix composites are being
developed with a variety of reinforcements.
NU
Ceramic reinforcements are used to improve the high temperature strength and
creep resistance. However, their effect on improving the room temperature fracture
MA

toughness is only marginal. Therefore, other types of reinforcements, e.g., ductile phases
need to be explored. Ductile phase toughening of MoSi2 was proposed originally by V. D.
Kristic [6, 7]. It was first investigated by Fitzer and Remmele [8]. However, there is very
D

limited work reported till date on the use of ductile reinforcements in MoSi2. Most of the
E

workers have used only Nb as the ductile reinforcement in MoSi2 [9-21].

PT

The aim of the present work is to explore the feasibility of using different ductile
phases as the reinforcements in MoSi2 based matrix. Three ductile refractory metals
CE

molybdenum, tantalum and niobium were selected as the reinforcements. As these

refractory metals are thermodynamically unstable with MoSi2, interfacial reactions are
AC

expected to occur between MoSi2 and the refractory metals. Model tri-layer laminated
composites were made using Mo, Ta and Nb sheets sandwiched in between two layers of
MoSi2 based matrix to carry out a comparative study of their chemical compatibility with
MoSi2. MoSi2 has a higher coefficient of thermal expansion (8.52x10-6/K) than all the three
refractory metals used in the present investigation (CTEs of Mo, Ta and Nb are 4.9x10-6/K,
6.5x10-6/K and 7.3x10-6/K, respectively). This may give rise to residual thermal stresses in
the composites during cooling from the processing temperatures to room temperature. The
residual stresses may affect the interfacial bonding and the resultant mechanical properties
of the composites. Therefore, SiC (CTE = 4.78x10-6/K) particles were added in MoSi2 matrix
to improve thermal compatibility between MoSi2 and the refractory metal sheets. The CTE
 ACCEPTED MANUSCRIPT

value of (MoSi2 + 20 vol% SiC) layer is expected to be 7.82x10-6/K as calculated by rule of

mixtures. SiC is thermodynamically stable with MoSi2 and has a lower coefficient of thermal
expansion than MoSi2.

The interfaces developed between MoSi2-SiC and various refractory metal sheets
were characterized for their microstructure. The thickness of the interfacial reaction layers
and the chemical nature of the various reaction products formed were determined using
techniques like optical microscopy, x-ray diffraction, scanning electron microscopy (SEM)

PT
and electron probe micro analyzer (EPMA). Micro-hardness measurements were made in
the reaction zone to evaluate the mechanical nature of the interfacial reaction layers. A
quantitative idea of the thickness and micro-hardness values of interfacial reaction layers

RI
may be very useful and can give some important guidelines to obtain the enhanced

SC
mechanical properties of the MoSi2 based composites.

EXPERIMENTAL
NU
The idea of laminating different metals and alloys to form a composite material
MA

(laminated metal composites) that exploits the good properties of the constituent materials
has been known from antiquity. The internal ‘interface’ between layers limits crack
propagation through several different mechanisms. The application of laminated metal
composites (metal – metal) to gun barrels for through thickness toughness is not new [22].
D

However, the idea of making laminated structures with alternate brittle ceramics and ductile
E

metal layers is relatively new.

PT

Processing
CE

Processing of hybrid laminated composites consisting of layers of brittle ceramic

particles and ductile metal sheets is extremely challenging. The component material layers
AC

are stacked together in a vacuum hot press and heated to a high temperature with
simultaneous application of pressure to facilitate diffusion bonding between the various
layers. Vacuum hot pressing is the most widely used process to consolidate the MoSi2
based materials.

The source and average particle size of the powders used in the present work are
given in Table 1 as they are reported by the manufacturers, while the source and thickness
of the refractory metal sheets are given in Table 2. The chemical composition of all the raw
materials was determined by analytical chemistry methods and is presented in Table 3.
While analysis of Fe, Si and Al is carried out by wet chemical method, the analysis of O, N
 ACCEPTED MANUSCRIPT

and C is carried out by LECO analyser. The oxygen content of MoSi2 powder has been
found to be about 1.3 wt%. The other principal impurities in MoSi2 are Fe (0.11 wt%), C
(0.08 wt%) and N (0.032 wt%). The oxygen in MoSi2 is known to be present in the form of
amorphous silica (SiO2) particles in free form and as thin films on the surface of MoSi2
powder particles. The main impurities in SiC powder are C (0.8 wt%), O (0.25 wt%) and N
(0.18 wt%). In all the metal sheets, the main impurities were found to be O and C.

The commercially pure MoSi2 and 20 vol% of SiC powders were mixed and wet
blended for 8 h in a ball mill. The average size of SiC particles was kept small (~ 2 m), so as to

PT
eliminate the thermally induced cracking in the composites. SiC particles have a tendency to
agglomerate due to the static charge present on their surfaces. Therefore, the SiC particles

RI
were given a de-agglomeration treatment prior to blending with MoSi2 powder. The de-

SC
agglomeration process involved treating SiC powder with a polar solvent n-butanol, which
neutralizes the static charge present on the particle surfaces.
NU
The laminated composites were made by stacking layers of MoSi2-SiC blended
powder and refractory metal sheets of Mo, Ta and Nb. A single refractory metal sheet of 0.5
mm thickness was sandwiched in between two layers of MoSi2-SiC blended powder. The
MA

amount of MoSi2-SiC blended powder on both sides of refractory metal foil was kept so that
a final thickness of approximately 2.5 mm of each layer of MoSi2-SiC is obtained after
consolidation by vacuum hot pressing as shown schematically in Fig. 1. The refractory metal
D

sheets were finish-polished through 15 m diamond paste on both the bonding surfaces.
E
PT

The blended powders were vacuum degassed prior to consolidation in the

temperature range of 1000 – 1100 °C for 1 h to remove any moisture and volatile matter. The
CE

degassing temperature was kept lower than the sintering temperature (1300 °C) of MoSi 2.
The vacuum hot pressing was carried out at 1600 °C for 1 h using a pressure of 27 MPa. The
high melting point and brittleness of the silicides necessitates using higher processing
AC

temperatures. The vacuum was maintained at around 1.33x10-2 Pa. A GCA, USA vacuum hot
press was used in the present work.

A typical time – temperature - pressure cycle used in the vacuum hot pressing of the
composites investigated in the present study is shown in Fig. 2. The hot pressed product was
a disc of 75 mm diameter with 5-6 mm thickness. The densities of the materials produced were
above 99 % of the theoretical densities.

Characterization
 ACCEPTED MANUSCRIPT

The microstructures of the MoSi2-SiC monolithic material and the laminated

composites prepared in the present work were studied using optical as well as scanning
electron microscopes (SEM). The conventional chemical etching was found not suitable for
MoSi2 based composites. The as-polished specimens were observed using ordinary as well
as polarized light. The use of polarized light successfully produced the microstructural
contrast revealing the grain size in the composite specimens. Gibbs et al. [23] had also
observed that with MoSi2, it might not be necessary to chemically etch the material since the
microstructure could be clearly viewed in polarized light due to anisotropy in its crystal

PT
structure.

A Leo 440i scanning electron microscope equipped with secondary and back

RI
scattered electron detectors as well as energy dispersive spectrometer (EDS) was used. The

SC
various interfacial reaction products were identified by x-ray diffraction (XRD) analysis and
analyzed for their chemical compositions by electron probe micro-analysis (EPMA) and
energy dispersive spectroscopy (EDS). The XRD analysis was carried out using the Cu Kα (
NU
= 15.4056 nm) radiation. The EPMA analysis was done using an accelerating voltage of
15 kV with a beam diameter of 1 m.
MA

The mechanical nature of the interfaces was characterized by micro-hardness tests.

Micro-hardness tests are conducted to determine the hardness over very small areas. A
D

MMT-3 MATSUZWA micro-hardness testing equipment having a load capacity of 5 kg was

used for this purpose. Micro-hardness measurements were made on metallographically
E

polished specimens with a Vickers diamond indenter using a load in the range of 50 g to 300
PT

g for 15 s. The variation in applied load for different laminated composites was necessitated
depending upon the thickness of the interfacial reaction layer. The applied load was selected
CE

to ensure that the size of the indentation mark generated falls within the interfacial reaction
zone and would not overlap with either MoSi2 matrix or the refractory metal sheet.
AC

Room temperature fracture toughness of MoSi2-SiC monolithic and various laminated

composites was evaluated by single edge notched beam (SENB) test method using a three-

point bend test specimen of the dimensions: span ( S ) = 40 mm, width ( W ) = 5 mm and
thickness ( B ) = 5 mm. The minimum length of the specimen was kept as 45 mm. The
toughness was measured in crack arrester mode. The schematic diagram of the test
specimen notched in crack arrester mode is shown in Fig. 3. The notch depth, a was
maintained between 0.4W and 0.45W, and was machined using a linear precision diamond
saw. The distance between the notch tip and the refractory metal foil was kept to be ~ 200
m. Fatigue pre-cracking was not performed. The crosshead speed was kept 0.5
 ACCEPTED MANUSCRIPT

mm/min. The crosshead speed was so chosen that the strain rate upon the
specimen was of the order of 1×10-4/s. The fracture toughness was calculated from the
peak load, Pmax of the load-displacement curve [24] and reported in terms of Kmax.

RESULTS AND DISCUSSION

Microstructural Characterization of Interfaces

(MoSi2 + 20 vol% SiC) / Mo Sheet Laminated Composite

PT
The optical microstructure of (MoSi2 + 20 vol% SiC) + Mo sheet laminated composite

RI
taken under polarized light contrast is shown in Fig. 4(a). This polarized light optical
microstructure reveals the MoSi2 grains with a grain size in the range of 5 – 8 m. The

SC
interfacial reaction layer is indicated by the arrow marks. Figure 4(b) is an SEM back
scattered electron (BSE) image showing the interfacial reaction layers on both sides of the
NU
Mo sheet as indicated by arrow marks. The thickness of the interfacial reaction layer was
measured to be approximately 50 m.
MA

The x-ray diffraction pattern of (MoSi2 + 20 vol% SiC) + Mo sheet laminated

composite as shown in Fig. 5, reveals the peaks of MoSi2, SiC and pure Mo. A few peaks of
Mo5Si3 were also detected but most of the Mo5Si3 peaks were not visible above background
D

because of their small height. X-ray diffraction data suggests that the interfacial reaction
layer consisted of mainly Mo5Si3 phase. Si diffused from MoSi2 and reacted with pure Mo to
E

form Mo5Si3 at the interface.

PT

Figure 6(a) exhibits a back scattered electron (BSE) image of (MoSi2 + 20 vol%
CE

SiC) + Mo sheet laminated composite. The x-ray mapping using EPMA of a particular region
demarcated by rectangular box in Fig. 6(a). The chemical composition obtained by EPMA
analysis (Table 4) revealed the presence of Mo5Si3 and Mo2C phases in the interfacial
AC

reaction layer (marked as A and B, respectively in the BSE micrograph). Mo2C was found in
discrete form within the Mo5Si3 layer as well as in the form of a thin and continuous layer at
pure Mo \ Mo5Si3 interface. The Mo-Si binary phase diagram [25] suggests that Mo5Si3 does
not co-exist with pure Mo. Therefore, the formation of Mo2C was necessitated and it
separated Mo5Si3 from pure Mo. Appreciable non-planar morphologies at Mo / Mo2C and
Mo2C / Mo5Si3 interfaces are revealed. The non-planar growth fronts of the Mo5Si3 and Mo2C
layers may reflect anisotropy in diffusion.

Formation of Mo2C is not expected as a result of reaction between MoSi2 and SiC as
both are highly stable with each other thermodynamically throughout the entire temperature
 ACCEPTED MANUSCRIPT

range as depicted in Fig. 7. A possible reaction between MoSi2 and SiC resulting in the
formation of Mo2C can be described as follows:

2MoSi 2  SiC  Mo2 C  5Si (1)

The variation of free energy change with temperature for the above reaction, ΔG (per
mole of SiC) is plotted in Fig. 7 and found to be positive for the entire temperature range.

However, the thermodynamic calculations revealed that SiC is not stable in the

PT
presence of pure Mo. The following possible reactions between Mo and SiC are considered
and the calculated free energy change data is plotted against the temperature as exhibited

RI
in Fig. 8 (Equations 2, 3, 4 and 5 are depicted as Reactions 1, 2, 3 and 4, respectively).

2Mo  SiC  Mo2C  Si

SC
1. (2)
2. 5Mo  2SiC  2Mo2C  MoSi2 (3)
3. 11Mo  3SiC  3Mo2C  Mo5 Si3 (4)
NU
4. 5Mo  SiC  Mo2C  Mo3Si (5)

It is evident from Fig. 8 that ΔG (per mole of SiC) is negative for all the reactions
MA

except the reaction 1 for the entire temperature range. ΔG for reaction 1 also shows a
decreasing trend with increase in temperature. This clearly establishes that SiC is not stable
with pure Mo and results in the formation of Mo2C. The data also indicates that the reaction 3
D

resulting in the formation of Mo2C with Mo5Si3 is most feasible at temperatures above 700 °C.
E

This explains the resulting microstructure in the reaction zone as a result of chemical
interactions between MoSi2 matrix and Mo sheet in the presence of SiC. The results in the
PT

present investigation are consistent with the observations of van Loo et al. [26].
CE

(MoSi2 + 20 vol% SiC) / Ta Sheet Laminated Composite

The optical microstructure and the SEM back scattered electron image of (MoSi 2 +
AC

20 vol% SiC) + Ta sheet laminated composite are shown in 9(a) and Fig. 9(b), respectively.
The optical microstructure reveals a uniform distribution of SiC particles in MoSi 2 matrix. The
interfacial reaction layer is indicated by an arrow mark. The x-ray diffraction pattern of (MoSi2
+ 20 vol% SiC) + Ta sheet laminated composite is presented in Fig.10, which exhibits the
peaks of MoSi2, SiC and pure Ta. A few peaks of Ta5Si3 and Ta2Si were also detected but
most of the Ta5Si3 and Ta2Si peaks were not visible above background because of their
small height. The x-ray diffraction data suggested that the reaction products formed at the
interface are Ta5Si3 and Ta2Si due to Si diffusion across the interface.
The thickness of the interfacial reaction layer in (MoSi 2 + 20 vol% SiC) + Ta sheet
laminated composite was found to be approximately 10 m which is much smaller than the
 ACCEPTED MANUSCRIPT

thickness of the reaction layer in (MoSi2 + 20 vol% SiC) + Mo sheet laminated composite.
This suggests that the diffusivity of Si in Ta and Ta5Si3 is smaller than the diffusivity of Si in
Mo and Mo5Si3.
The quantitative chemical analysis as determined by energy dispersive spectroscopy
(EDS) taken at two different locations within the reaction zone (indicated by arrow marks in
Fig. 9(b)) is presented in Table 5. EDS analysis revealed the presence of Ta5Si3 and Ta2Si
phases at the locations marked as A and B in Fig. 9(b). The back scattered electron image
could not resolve Ta5Si3 and Ta2Si phases clearly owing to their similar average atomic

PT
numbers. The average atomic numbers of Ta2Si and Ta5Si3 phases are 68.75 and 67.97,
respectively.

RI
The EPMA images of (MoSi2 + 20 vol% SiC) + Ta sheet laminated composite are

SC
shown in Fig. 11. Figure 11(a) is a back scattered electron image while Figs. 10(b) and 10(c)
exhibit x-ray maps of Mo and Ta respectively. The x-ray map and the line scan of Si are
shown in Figs. 10(d) and 10(e), respectively. The interfacial reaction zone was enriched with
NU
Si as well as Mo and Ta as exhibited by x-ray maps of these elements. Apart from Si
diffusion from MoSi2, some Mo and Ta have also diffused across the interface. Mo dissolved
MA

in newly formed Ta5Si3 phase forming (Ta, Mo)5Si3 while Ta dissolved in Mo5Si3 phase
formed as a result of Si depletion within MoSi2 region. However, the x-ray mapping and the
line scan of Si could not clearly resolve Ta5Si3 and Ta2Si phases separately. A very small
D

difference in the average atomic numbers could be the reason for these two phases not
getting resolved in back scattered electron images under SEM as well as in EPMA.
E
PT

The Ta-Si binary phase diagram [27] shows three stable Ta silicide phases, TaSi2,
Ta5Si3 and Ta2Si. The possible chemical reactions forming each one of these silicides as a
CE

result of MoSi2 reacting with Ta are:

1. 5MoSi 2  14Ta  7Ta2 Si  Mo5 Si3 (6)

AC

2. 15MoSi2  35Ta  7Ta5 Si3  3Mo5 Si3 (7)

3. 10MoSi2  7Ta  7TaSi2  2Mo5 Si3 (8)

The standard free energy change (per mole of MoSi2) for the above reactions is
plotted against the temperature as shown in Fig. 12 (Equations 6, 7 and 8 depicted as
Reactions 1, 2 and 3, respectively). As seen in this figure, ΔG values are negative for all the
three reactions considered. However, the phase TaSi2 (Reaction 3) seems to be less stable
as compared to Ta5Si3 and Ta2Si phases. The free energy change for reaction 3 increases
with increasing temperature in contrast to a decreasing ΔG with increasing temperature for
reactions 1 and 2. ΔG values for reactions 1 and 2 exhibit similar trends with only marginal
 ACCEPTED MANUSCRIPT

difference. Accordingly, Ta5Si3 and Ta2Si phase layers developed in the reaction zone in
(MoSi2 + 20 vol% SiC) + Ta sheet laminated composite (as existing in the Si-rich side of Ta-
Si phase diagram). Tortorici and Dayananda [28] have also reported the development of
three phase layers corresponding to Ta2Si, Ta5Si3 and (Mo, Ta)5Si3 in the diffusion zone of
MoSi2 + Ta disk diffusion couple annealed at 1600 °C for 8 h.

(MoSi2 + 20 vol% SiC) / Nb Sheet Laminated Composite

The SEM back scattered electron image and the x-ray diffraction pattern of (MoSi2 +

PT
20 vol% SiC) + Nb sheet composite are presented in Figs. 12 and 13, respectively. The SEM
microstructure exhibits more or less a uniform distribution of SiC particles. The

RI
microstructure reveals the presence of an interfacial reaction layer between MoSi2 + 20 vol%
SiC matrix and Nb sheet. The thickness of the interfacial reaction layer was measured to be

SC
approximately 20 m. The main reaction product between MoSi2 and Nb was identified as
Nb5Si3 by quantitative electron probe micro analysis. The x-ray diffraction data as presented in
NU
Fig. 14 does not reveal any Nb5Si3 peaks although the peaks of MoSi2, SiC and pure Nb are
clearly identified. This could be due to the small height of Nb5Si3 peaks which are not visible
above background.
MA

Figure 15 is a combination of EPMA back scattered electron (BSE) image as well as x-

ray elemental maps of Si, Mo and Nb for (MoSi2 + 20 vol% SiC) + Nb sheet laminated
D

composite. The quantitative analysis of chemical composition of the reaction zone is

E

presented in Table 8. The typical microprobe data obtained at locations indicated by A, B, C

and D in Fig. 15(a) correspond to MoSi2, Mo5Si3 (with Nb in solid solution), Nb5Si3 (with Mo in
PT

solid solution) and pure Nb phases, respectively.

CE

The relatively large and irregular shaped dark particles present in MoSi2 are of SiC
while smaller and rounded dark particles are of SiO2. The SiC and SiO2 particles are present
inside the reaction zone close to the MoSi2. This suggests that the interfacial reaction layer has
AC

grown inside both, the MoSi2 based matrix and the Nb sheet. The block arrow in Fig. 15(a)
indicates the probable original interface between MoSi2 based matrix and the Nb sheet prior to
any chemical interactions. The reaction layer towards the Nb sheet is much thicker than that
towards the matrix.

The x-ray maps of Si, Mo and Nb as shown in Figs. 14(b), 14(c) and 14(d) respectively,
reveal extensive diffusion of Si into Nb has taken place as compared to the diffusion of Mo and
Nb across the interface. Si diffused far deeper into the Nb sheet than Mo. Diffusivity of Si is
expected to be higher than diffusivity of Nb owing to its lower melting point and a lower atomic
 ACCEPTED MANUSCRIPT

size as compared to Nb. Faster Si diffusion than Nb diffusing in the opposite direction may
introduce Kirkendall porosity on the MoSi2 side as a result of vacant sites left behind by Si.
However, non-stoichiometry is not possible in MoSi2 as MoSi2 is a line compound. The
depletion of Si is compensated by some amount of MoSi2 converting into Mo5Si3. Therefore,
Kirkendall porosity is not likely to be present in these composites.

The line scans of Mo and Nb are shown in Figs. 15(b) and 15(c), respectively in
combination with a BSE image taken at a higher magnification and shaded dark to reveal the

PT
SiO2 particles at the interface (Fig. 16(a)). Both the line scans indicated the presence of
concentration gradients for Mo and Nb within the reaction zone. The troughs observed in line
scan of Mo indicate the scan passing through the SiC particles present inside MoSi2.

RI
The thermodynamic calculations were carried out exploring the formation of various

SC
silicide phases in the reaction zone between MoSi2-SiC and pure Nb. The Nb-Si binary
phase diagram [29] exhibits only two stable silicides, NbSi2 and Nb5Si3 at room temperature.
NU
A third phase Nb3Si is stable between 1770 to 1980 °C. The possible chemical reactions
between MoSi2 and Nb resulting in the formation of Nb5Si3 and NbSi2 are as follows:
MA

1. 15MoSi 2  35Nb  7 Nb5 Si3  3Mo5 Si3 (9)

2. 10MoSi2  7 Nb  7 NbSi 2  2Mo5 Si3 (10)
3. 5MoSi 2  7 Nb  Nb5 Si3  2 NbSi 2  Mo5 Si3 (11)
D

The standard free energy change for the above reactions with respect to temperature
E

is plotted in Fig. 17. ΔG (per mole of MoSi2) is much more negative for equation 9 (depicted
PT

in Fig. 17 as Reaction 1) leading to the formation of Nb5Si3 as compared to reactions 2 and 3

(equations 10 and 11) which result in the formation of NbSi2 as one of the reaction products
CE

at all temperatures. The microstructure in the reaction zone obtained in the present study
also exhibits the presence of Nb5Si3 along with Mo5Si3 with no evidence of NbSi2 phase. The
AC

formation of solid solution (Mox, Nb1-x)5Si3 from mixing of the pure components can be
described by the following reaction:

x Mo5Si3 + (1-x) Nb5Si3  (Mox, Nb1-x)5Si3 (12)

However, no data on the free energy of mixing for the above reaction is available.

As exhibited in Fig. 15(b) and 15(c), extensive diffusion of Si and Mo has taken place
into pure Nb. Si seems to have diffused more rapidly and uniformly than Mo. This is in
agreement with the Si and Mo diffusivities in Nb. The diffusivities of Si and Mo in Nb are given
by the following equations [30, 31]:
 ACCEPTED MANUSCRIPT

DSi  0.5110 2 exp  48000 RT  at T = 700 – 1500 C (13)

DMo  92 exp  122000 RT  at T = 1725 – 2182 °C (14)

If the above equations are used to calculate approximate values of diffusivities of Si and
Mo in Nb at the hot pressing temperature of 1600 C (although, equations 13 and 14 are valid
only between 700 – 1500 °C and 1725 – 2182 °C, respectively), the resultant diffusivity of Si in
Nb at 1600 °C is higher by four orders of magnitude than diffusivity of Mo in Nb. This clearly
indicates that Si atoms diffuse much more rapidly than Mo atoms in Nb.

PT
Figure 18 exhibits an isothermal section of Mo-Si-Nb ternary phase diagram at 800

RI
°C [32]. Isothermal sections of ternary phase diagrams contain three different kinds of
regions, namely, single-phase, two-phase and three-phase regions. In Fig. 18, there are four

SC
distinct single-phase regions, which are depicted as black areas. These are (i) Nb and Mo
dissolving completely in each other as a body centered cubic (bcc) solid solution (ii) Mo3Si
NU
phase region (iii) a continuous solid solution from Nb5Si3 to Mo5Si3 and (iv) NbSi2 phase
(hP9) having a crystal structure different from MoSi2 (tI6). Between single-phase regions are
the two-phase regions, which contain tie lines whose ends are the compositions of the
MA

saturated phases in an equilibrium co-existing mixture. A narrow two-phase region lies

between the disilicides. The three-phase regions are the triangular regions in which three
phases are in equilibrium with compositions given by the corners. Each side of the triangle is
D

a limiting tie line of the adjacent two-phase field.

E

The analysis of interfacial reaction layers in composites requires the use of these
PT

diagrams. Savitsky et al. [33] have reported that the isopleth through MoSi2 / Nb shows two
four-phase equilibriums at 1740 °C and 1730 °C associated with MoSi2 / Nb vertical section.
CE

They also found that the relevant phase fields remain essentially unchanged below 1730 °C.
Therefore, an attempt has been made to analyze the composition path in the present study
AC

on the basis of the isothermal section at 800 °C which is the only ternary phase diagram of
this system available in the literature [34].

In (MoSi2 + 20 vol% SiC)/ Nb sheet laminated composite, it was found that the
interface consists of a Nb rich (Nb, Mo)5Si3 layer and a Mo rich (Mo, Nb)5Si3 layer as shown
in Fig. 15(a). Based on the microprobe data taken across the reaction layers and plotted on
the ternary triangle, a possible graph of the diffusion path is obtained as shown in Fig. 19. It
runs from the pure Nb corner into the single-phase bcc region (dissolving a very small
amount of Si) and then jumps across the two-phase field to Nb5Si3. Sharp interface between
pure Nb and Nb5Si3 layer (Fig. 15(a)) indicates that the composition path runs along the tie
 ACCEPTED MANUSCRIPT

line in the corresponding two-phase region. The jump occurs at positions in the sample
corresponding to the location of the boundary between the phases. The path then runs
across the Nb5Si3 - Mo5Si3 single-phase region. The non-planar interface between MoSi2 and
Mo5Si3 phases as revealed in Fig. 15(a) suggests that the composition path crosses the
corresponding two-phase region by cutting across the tie lines and therefore, the diffusion
path is further plotted passing through the Mo5Si3 – MoSi2 two-phase region before reaching
MoSi2.

PT
It is revealed from Fig. 19 that the diffusion path zigzags across the line connecting
the two compositions of the two reacting materials satisfying the condition that the path must
lie on both sides of the connecting line, which contains the average composition, to conserve

RI
the mass within the reaction zone. The route that the path takes across the ternary triangle is

SC
related to the identity of the fast diffusing species. The ends of the diffusion path, near the
MoSi2 end and the Nb end, both point either toward or away from the Si corner. This fact
indicates that the Si is the fastest diffusing species in MoSi2-SiC/ Nb composite system.
NU
As seen in Fig. 15(a), the interface between Nb and Nb5Si3 is relatively planar while the
interface between Mo5Si3 and MoSi2 is non-planar. Tortorici and Dayananda [28] have also
MA

reported the development of Nb5Si3 and (Mo, Nb)5Si3 layers in the diffusion zone in MoSi2 +
Nb diffusion couple annealed at 1600 °C for 16 h with a planar interface between Nb and
Nb5Si3. They also reported that the diffusivity of Nb in (Mo, Nb)5Si3 layer was larger in
D

magnitude than the diffusivity of Mo. The melting point of Nb is lower than Mo and therefore,
E

Nb is expected to diffuse faster than Mo.

PT

The thickness of the interfacial reaction layers and the various interfacial reaction
products formed in the laminated composites studied in the present work are summarized in
CE

Table 9.
AC

The melting points and atomic sizes of various elements present in the composite
systems studied in the present work are given in Table 8. Si has the lowest melting point and
the smallest atom size among all the diffusing species and therefore, is likely to have the
highest diffusivity among all the elements present in these composites. At the processing
temperature, Si is diffusing at a temperature greater than its melting point.

Si diffusion is believed to occur by several mechanisms including lattice diffusion,

grain boundary diffusion, dislocation diffusion and surface diffusion. The diffusivity is
generally higher along the grain boundaries, dislocations and the free surfaces like pores
and micro-cracks. The grain boundary regions are not as closely packed as the crystal. The
 ACCEPTED MANUSCRIPT

activation energy for grain boundary diffusion is appreciably less than that for lattice
diffusion. However, the grain boundary diffusion is more important at lower temperatures.
The dislocation density in ceramics and intermetallics is appreciably lower than metals. The
diffusion of Si through such regions in these composites is not likely to be controlled by
dislocation diffusion mechanism. In the initial stages of consolidation, Si can diffuse through
free surfaces. However, in the fully consolidated composite, Si diffusion is most likely to
occur by vacancy mechanism through its own sub-lattice. At a processing temperature of
about 1600 °C, high concentration of vacancies is expected. The energy of a Si atom on a

PT
Mo site is expected to be higher than on a Si site. However, if this energy difference is small
enough, some Si atoms may reside in Mo sites. As MoSi2 is a stoichiometric compound, if Si

RI
atom diffuses by moving into adjacent vacant Mo site, there will be a high probability of the
Si atom returning to its own vacant Si site on its next jump. Therefore, it may be preferable

SC
that Si atom diffuses only through its own sub-lattice.

In laminated composites consisting of alternate layers of (MoSi2 + 20 vol% SiC) and

NU
Mo, Ta and Nb sheets, the total thickness of the interfacial reaction zone was found to be as
50 m, 10 m and 20 m, respectively. This is consistent with the reported values of Si
MA

diffusivity and activation energies as reported by Tortorici and Dayananda [28, 34] for the
development of reaction zones having different phases in these laminated composites. The
main reaction product was identified to be M5Si3 type of silicide (M = Mo, Ta, Nb) in all the
D

laminated composites studied in the present work. The diffusivity of Si (Dsi) in Ta5Si3 at 1600
°C was reported as 7.1x10-17 m2/s, which is one order of magnitude lower than the diffusivity
E

of Si in Nb5Si3 (Dsi = 5.4x10-16 m2/s). They reported activation energy of 210 kJ/mole for the
PT

growth of Mo5Si3 layer as compared to the activation energy of 265 kJ/mole for Nb5Si3
phase. Lower the activation energy for diffusion, higher is the diffusivity. Tortorici and
CE

Dayananda [28, 34] have also found smaller thickness of the diffusion layers in ternary
couples (MoSi2-Nb, MoSi2-Ta) than those for MoSi2-Mo binary couples.
AC

Evaluation of Micro-hardness of Interfacial Reaction Layers

Interfaces play an important role in determining the mechanical properties of the

composites. In laminated composites synthesized and studied in the present work, interfacial
reaction layers of varying thicknesses consisting of different reaction products were
observed. The micro-hardness values measured for interfacial reaction layers in (MoSi2 + 20
vol% SiC) matrix reinforced with Mo, Ta and Nb sheets laminated composites are presented
in Table 9. Each data point in Tables 9 is an average of minimum 5 tests. The variation in
hardness data was found to be within ±5 % of the average value for all the composites.
 ACCEPTED MANUSCRIPT

The thickness of the reaction zone was 50 m in case of Mo sheet laminated composite
while it was 20 m in Nb sheet laminated composite. Ta seems to have better chemical
compatibility with MoSi2 as it exhibited only a 10 m thick reaction zone after processing under
similar conditions. However, the micro-hardness data presented in Table 9 reveals that the
silicide phases formed as a result of chemical interactions between MoSi2 and Ta sheet are
harder than the phases formed at the interface in Mo and Nb sheet reinforced laminated
composites. Micro-hardness measurements were made using indentation loads in the range of
50 g to 300 g for different laminated composites. The variation in applied load was necessitated

PT
depending upon the thickness of the interfacial reaction layer. The applied load was selected to
ensure that the size of the indentation mark generated falls within the interfacial reaction zone

RI
and would not overlap with either MoSi2 matrix or the refractory metal sheet.

SC
It has to be noted that in all the three laminated composites, the reaction zone consisted
of more than one phase. The micro-hardness data presented here is indicative of the average
hardness of the reaction zone in each case. Micro-hardness measurements of individual
NU
phases formed within the same interfacial reaction layer of a composite could not be carried out
in the present work with the existing micro-hardness testing equipment. To obtain indentation
MA

marks well within the individual phases, indentation loads much smaller in magnitude would be
required.
D

The micro-hardness values were also measured inside the matrix layer and metal
sheets in all the three laminated composites as presented in Table 9. The micro-hardness
E

values measured inside the MoSi2 + 20 vol % SiC matrix were found to be lower than the
PT

macro-hardness values measured with indentation loads of 10 – 30 kg [35]. The discrepancy

in hardness values measured by different load levels could be due to a much smaller size of
CE

indentation mark generated with small loads. The contribution to the hardness by SiC
particles is expected to be more with higher loads generating large size indentation marks.
AC

SiC is much harder than MoSi2. The hardness value of SiC is reported to be ~ 25 GPa [36]
as compared to the hardness value of pure MoSi2 (9 - 10 GPa). However, the micro-
hardness values of refractory metal sheets inside the laminated composites were higher
than the hardness values measured with higher load levels [35]. For lighter loads (< 300 g),
the small amount of elastic recovery becomes appreciable and with very small indentations
produced at light loads, the error in locating the actual ends of the indentation becomes
greater. Both these factors have the effect of giving a higher hardness reading, so that it is
usually observed that the micro-hardness number increases as the load is decreased below
about 300 g.
 ACCEPTED MANUSCRIPT

Figures 20(a) and 20(b) are the optical micrographs showing a typical micro-
hardness indentation mark inside the interfacial reaction layer in (MoSi2 + 20 vol% SiC) + Mo
sheet laminated composite taken under ordinary light and polarized light, respectively. The
interfacial reaction layer is demarcated by arrow marks. The microstructures also exhibit the
indentation marks taken in the matrix layer as well as inside the Mo sheet. The indentation
mark inside the MoSi2-SiC matrix is encircled to differentiate it from SiC particles present in
the matrix. A barrelled indentation mark is observed in Mo sheet. Mo, unlike Ta and Nb, has
a large contribution to the yield stress below 0.5 Tm due to thermal activation of slip. This

PT
could account for its higher flow stress than Ta or Nb. In this case, the diagonal
measurement produces a low value of the contact area so that the hardness numbers

RI
obtained are high.

SC
Figure 21(a) exhibits the typical micro-hardness indentation marks inside the
interfacial reaction layer, inside MoSi2-SiC matrix and inside the pure Ta sheet in (MoSi2 +
NU
20 vol% SiC) + Ta sheet laminated composite. The interface is depicted by arrow marks.
The thickness of the interfacial reaction layer in Ta sheet laminated composite was only 10
m resulting in difficulties in the measurement of its micro-hardness by ensuring that a
MA

hardness indentation mark is generated within the thin reaction zone without overlapping
with MoSi2-SiC matrix or pure Ta sheet. This could be achieved only with a load as low as
50 g. The small indentation marks inside the reaction zone and MoSi2-SiC matrix are
D

enclosed by rectangles in Fig. 21(a) for the purpose of easy identification and clarity. The
E

same indentation mark is encircled in Fig. 21(b), which was taken at a higher magnification.
PT

In contrast to Mo sheet, Ta sheet exhibits a pincushion indentation mark due to sinking-in.

The pincushion indentation is the result of sinking-in of the metal around the flat faces of the
pyramid. This condition is observed with annealed metals and results in an overestimate of
CE

the diagonal length. For annealed metals, work hardening of the surface during polishing can
also influence the hardness results.
AC

Similarly, Figs. 21(a) and 21(b) are the optical micrographs showing micro-hardness
indentation marks inside the interfacial reaction layer, matrix layer and in pure Nb sheet in
(MoSi2 + 20 vol% SiC) + Nb sheet laminated composite taken at different magnifications.
Here again, the Indentation marks inside the interfacial reaction layer and in the near-by
matrix are encircled for easy identification.

Looking at Figs. 19, 20 and 21, it is revealed that the sizes of indentation marks are
similar in MoSi2 + 20 vol% SiC matrix and inside the reaction zones. This suggests that the
various interfacial reaction products are also hard and brittle like the matrix. These are
 ACCEPTED MANUSCRIPT

formed at the expense of ductile metal sheets reducing the effective volume fraction of the
ductile phases. The fracture toughness values of the laminated composites as measured by
a three-point bend test [35] are presented in Table10. The Ta sheet resulted in the highest
improvement in fracture toughness of MoSi2 based laminated composites. Ta also results in
the smallest reaction zone thickness with MoSi2. In fact, it is interesting to observe that the
improvement in room temperature fracture toughness decreased as the thickness of the
reaction zone increased with different refractory metal sheets. As interfacial reaction layer
thickness increases, more ductile metal is consumed and converted into brittle phases.

PT
However, the role of interfacial reaction products in determining the mechanical
properties of the laminated composites is not fully understood at present. Future work is

RI
clearly needed in this area. The toughness of the MoSi2 reinforced with ductile refractory

SC
metal sheets is likely to be governed by not only the combined effects of intrinsic properties
of the metal sheet, sheet thickness, thickness and mechanical nature of the interfacial
reaction products but also on the residual thermal stresses generated during high
NU
temperature processing.

4. CONCLUSIONS
MA

A variety of reaction products are formed in ductile refractory metal sheets reinforced
D

MoSi2 based laminated composites studied in the present work. The formation of various
phases in the interfacial reaction zone has been analyzed based on the thermodynamic and
E

kinetic principles. The following conclusions have emerged from the present investigation:
PT

1. In ductile refractory metal sheet reinforced MoSi2 based laminated composites,

CE

interfacial reaction layers of varying thickness were observed to form between MoSi2-
SiC matrix and the metal sheets due to extensive diffusion of Si into Mo, Ta and Nb
sheets during high temperature processing.
AC

2. The interfacial reaction products were identified to be mainly silicide phases

consuming a part of the thickness of the ductile metal sheets. The main reaction
product was found to be as M5Si3 (M = Mo, Ta or Nb).

3. The Ta sheet reinforced laminated composite exhibited the minimum thickness of the
reaction zone of only 10 µm suggesting that the Ta has better chemical compatibility
with MoSi2 as compared to Mo and Nb. The highest improvement in room
temperature fracture toughness was also observed in case of Ta sheet laminated
composite.
 ACCEPTED MANUSCRIPT

4. The micro-hardness measurements made in the reaction zone established that the
interfacial reaction products in all the three laminated composites were hard and
brittle in nature.

ACKNOWLEDGEMENTS

The authors are thankful to Dr S V Kamat,, Outstanding Scientist & Director, DMRL
for kindly giving permission to publish this work and DRDO for funding this work. The

PT
authors also extremely grateful to Dr. G Madhusudan Reddy, Outstanding Scientist for his
tremendous help in carrying out micro-hardness measurements of the samples. Technical

RI
help from various groups in carrying out the microstructural, EPMA, XRD and mechanical
characterization is gratefully acknowledged.

SC
REFERENCES
NU
1. M Alouani, RC Albers and M Methfessel, “Calculated Elastic Constants and
Structural Properties of Mo and MoSi2”, Phys. Rev. B, 43 (1991) 6500-6509.
MA

2. K Badrinarayanan, AL Mckelvey, KT Venkateswara Rao and RO Ritchie,

“Fracture and Fatigue Crack Growth Behaviour in Ductile Phase Toughened
MoSi2: Effects of Nb Wire vs. Particulate Reinforcements”, Metallurgical and
Materials Transactions A, 27A (1996) 3781-3792.
D

3. Mohan G. Hebsur, “PEST RESISTANT MOSIZ MATERIALS AND METHOD OF

E

MAKING” United States Patent, Patent Number: 5,429,997 Date of Patent: Jul.
PT

4, 1995,

4. Jin-Kook Yoon, Kyung-Hwan Lee, Gyeung-Ho Kim, Jun-Hyun Han, Jung-Mann

Doh and Kyung-Tae Hong(2004), ”Low-Temperature Cyclic Oxidation Behavior of
CE

MoSi2/SiC Nanocomposite Coating Formed on Mo Substrate”, Materials

Transactions, Vol. 45, No. 7 pp. 2435-2442.

Michael J. Maloney, Donald Anton, Dilip Shah, “PEST RESISTANT

AC

5.
MOLYBDENUM DISILICIDE TYPE MATERIALS”, United States Patent, Patent
No.: US 6,265,080 Date of Patent: Jul. 24, 2001.

6. VD Krstic, PS Nicholson and RG Hoagland, “Toughening of Glasses by Metal

Particles”, Journal of American Ceramic Society, 64 (1981) 499-504.

7. VD Krstic, Philosophical Magazine, 48 (1983) 695.

8. E Fitzer and W Remmele, “Possibilities and Limits of Metal Reinforced Refractory

Silicides Especially MoSi2", Proceedings, 5th International Conference on
Composite Materials (ICCM-V), Eds. WC Harrigan Jr, J Strife and AK Dhingra,
AIME, Warrendale, PA, (1985) 515-530.
 ACCEPTED MANUSCRIPT

9. L Xiao, “Study of the Flow Behaviour of Constrained Ductile Phases – I.

Experiment”, Developments in Ceramic and Metal Matrix Composites, Ed. K
Upadhyaya, The Minerals, Metals and Materials Society, (1991) 115-124.

10. L Xiao, “Study of the Flow Behaviour of Constrained Ductile Phases – II.
Modeling”, Developments in Ceramic and Metal Matrix Composites, Ed. K
Upadhyaya, The Minerals, Metals and Materials Society, (1991) 359-369.

11. L Xiao and R Abbaschian, “Interfacial Modification in Nb / MoSi2 Composites and

its Effects on Fracture Toughness”, Materials Science and Engineering, A155
(1992) 135-145.

PT
12. L Xiao and R Abbaschian, “Role of Matrix / Reinforcement Interfaces in the
Fracture Toughness of Brittle Materials Toughened by Ductile Reinforcements”,
Metallurgical Transactions A, 23A (1992) 2863-2872.

RI
13. DE Alman and NS Stoloff, “Effect of Ductile Phase Reinforcement Morphology on
Toughening of MoSi2”, Intermetallic Matrix Composites II, Proceedings Materials

SC
Research Society Symposium, 273 (1992) 247-252.

14. DE Alman and NS Stoloff, “The Effect of Nb Morphology on the Cyclic Oxidation
Resistance of MoSi2/20 vol% Nb Composites”, Metallurgical and Materials
NU
Transactions A, 25A (1994) 1033-1040.

15. DE Alman and NS Stoloff, “The Effect of Nb Morphology on the Fracture

Behaviour of MoSi2/Nb Composites”, Metallurgical and Materials Transactions A,
MA

26A (1995) 289-303.

16. DE Alman, JA Hawk, JJ Petrovic, CH Henager and M Singh, Wear of

Engineering Materials, Ed. J Hawk, ASM International, Materials Park, Ohio,
D

(1998) 51.
E

17. DE Alman, JA Hawk and AV Petty (Jr), Proceedings Materials Research Society
Symposium, 350 (1994) 195.
PT

18. DE Alman, KG Shaw, NS Stoloff and K Rajan, “Fabrication, Structure and

Properties of MoSi2 Base Composites”, Materials Science and Engineering, A155
CE

(1992) 8593.

19. KT Venkateswara Rao and RO Ritchie, “Fracture and Fatigue Considerations in

the Development of Ductile Phase Reinforced Intermetallic Matrix Composites”,
AC

Fatigue and Fracture of Ordered Intermetallic Materials I, Eds. WO Soboyejo, TS

Srivatsan and DL Davidson, The Minerals, Metals and Materials Society, (1994).

20. KT Venkateswara Rao, GR Odette and RO Ritchie, “On the Contrasting Role of
Ductile Phase reinforcements in the Fracture Toughness and Fatigue Crack
Propagation Behaviour of TiNb/  -TiAl Intermetallic Matrix Composites”, Acta
Metallurgica et Materialia, 40 (1992) 353-361.

21. KT Venkateswara Rao, WO Soboyejo and RO Ritchie, “Ductile Phase

Toughening and Fatigue Crack Growth in Nb Reinforced MoSi2 Intermetallic
Composites”, Metallurgical Transaction A, 23A (1992) 2249-2257.

22. RW Hertzberg, Deformation and Fracture Mechanics of Engineering Materials,

Wiley, New York, (1983) 328.
 ACCEPTED MANUSCRIPT

23. WS Gibbs, JJ Petrovic and RE Honnell, “SiC Whisker-MoSi2 Matrix Composites”,

Ceramic Engineering and Science Proceedings, 8 (1987) 645-648.

24. MK Jain, J Subrahmanyam and S Ray, “Development of Mo and Ta Sheet

Reinforced (MoSi2 + 20 vol% SiCp) Matrix Laminated Composites” Advanced
Materials Resarch, 585 (2012) 306-310.

25. AB Gokhale and GJ Abbaschian, “Molybdenum-Silicon”, in Binary Alloy Phase

Diagrams, Vol. 2, Ed. TB Massalski, American Society for Metals, Metals Park,
OH, (1986) 1631-1632.

26. FJJ van Loo, FS Smet, GD Riek and G Verspui, “Phase Relations and Diffusion

PT
Paths in the Mo-Si-C System at 1200 °C”, High temperatures-High Pressures, 14,
(1982) 25-31.

ME Schlesinger, “The Si-Ta (Silicon-Tantalum) System”, Journal of Phase

RI
27.
Equilibrium, 15 (1994) 90.

SC
28. PC Tortorici and MA Dayananda, “Interdiffusion and Diffusion Structure
Development in Selected Refractory Metal Silicides”, Materials Science and
Engineering, A261 (1999) 64-77.
NU
29. ME Schlesinger, H Okamoto, AB Gokhale and R Abbaschian, Journal of Phase
Equilibria, 14 (1993) 502.

30. F Roux and A Vignes, Rev. Phys. Appl., 5 (1970) 393.

MA

31. E Fitzer and FK Schmidt, “Si Diffusion in Me5Si3 Phases of the Metals Nb, Ta, Mo
and W to 1700 C”, High Temperature High Pressure, 3 (1971) 445-460.
D

32. EM Savitskiy, VV Baron, MI Bychkova, SA Bakuta and EI Gladyshevskiy,

Russian Metall., 2 (1965) 91.
E

33. EM Savitskiy, VV Baron and MI Bychkova, Trudy Inst. Met. Baikova, Acad. Naut
PT

SSSR, 12 (1963) 179.

34. PC Tortorici and MA Dayananda, “Diffusion Structures in Mo vs. Si Solid-Solid

CE

Diffusion Couples, Scripta Materialia, 38 (1998) 1863-1869.

35. MK Jain, J Subrahmanyam and S Ray, “Development of Mo and Ta Sheet

Reinforced (MoSi2 + 20 vol% SiC) Matrix Hybrid Laminated Composites”, DMRL
AC

Technical Report DRDO-DMRL-CCG-032-2012, (2012) 21.

36. AK Bhattacharya and JJ Petrovic, “Hardness and Fracture Toughness of SiC

Particle Reinforced MoSi2 Composites”, Journal of American Ceramic Society, 74
(1991) 2700-2703.
 ACCEPTED MANUSCRIPT

MoSi2 Layers

Metal Foil

PT
RI
SC
NU
MA

Fig. 1 Schematic illustration of stacking and bonding of different layers in making of a tri-layer
D

laminated composite
E
PT
CE
AC
 ACCEPTED MANUSCRIPT

Apply Pressure
27 MPa

Consolidation
1600 ℃ / 1h

PT
RI
Heating
Degassing
~ 10 ℃ / min
1100 ℃ / 1h

SC
NU
Cooling
MA

Heating
~ 15 ℃ / min
E D
PT

Fig. 2 A typical time-temperature-pressure cycle used in the vacuum hot pressing of

composites investigated in the present study.
CE
AC
 ACCEPTED MANUSCRIPT

P
B

Metal Foil

MoSi2 W
a

PT
S

RI
P/2 P/2

SC
Fig. 3 Schematic diagram of the SENB fracture toughness test specimen in
crack arrester mode
NU
MA
E D
PT
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
b
NU
MA
E D
PT

Fig. 4 Microstructures of (MoSi2 + 20 vol% SiC) + Mo foil laminated composite (a)

Polarized light optical microstructure [24] and (b) SEM back scattered electron
CE

(BSE) image showing interfacial reaction layers on both sides of the Mo sheet
AC
 MoSi2 (002)
AC
MoSi2 (101)
CE
SiC (111)
PT
MoSi2 (110)
E Mo (110)
D SiC (200)

MoSi2 (112) MoSi2 (103)

MA

2 (Degree)
MoSi2 (200)
Mo (200), SiC (220)
NU
MoSi2 (202)
SC
MoSi2 (211)

MoSi2 (006), SiC (311)

RI
Mo (211), SiC (222)
ACCEPTED MANUSCRIPT

MoSi2 (213)
PT
Mo5Si3
MoSi2 (116)
Mo (220)
Fig. 5 X-ray diffraction pattern of (MoSi2 + 20 vol% SiC) + Mo foil laminated composite.
 ACCEPTED MANUSCRIPT

B Mo2C
A

PT
b

RI
SC
NU
MA

c
E D
PT
CE

d
AC

Fig. 6 (a) EPMA back scattered electron image of (MoSi2 + 20 vol% SiC) + Mo foil
composite. (b) High magnification back scattered electron image of the area
surrounded by a rectangle, (c) x-ray map of Si and (d) x-ray map of carbon.
 ACCEPTED MANUSCRIPT

300
Free Energy Change (kJ/mole)
250

200

150

PT
100

RI
50

SC
0
200 400 600 800
NU 1000 1200 1400
Temperature (K)

Fig. 7 Standard free energy change (ΔG vs. temperature) for a possible chemical reaction
MA

between MoSi2 and SiC.

D

20
E
Free Energy Change (kJ\mole)

0
PT

Reaction 1
-20 Reaction 2
Reaction 3
CE

-40 Reaction 4

-60
AC

-80

-100

-120
200 400 600 800 1000 1200 1400 1600
Temperature (K)

Fig. 8 Standard free energy change (ΔG vs. Temperature) for the possible chemical
reactions between SiC and pure Mo resulting in the formation of Mo2C.
 ACCEPTED MANUSCRIPT

PT
RI
SC
b
NU
MA
E D
PT

Fig. 9 (a) Microstructures of (MoSi2 + 20 vol% SiC) + Ta foil laminated composite (a) optical
microstructure and (b) SEM back scattered electron image.
CE
AC
 MoSi2 (002)
AC
MoSi2 (101)
CE
SiC (111)
PT
Ta (110)
E MoSi2 (110)
D MoSi2 (112) MoSi2 (103)
MA

2 (Degree)
Ta (200)
MoSi2 (200), SiC (220)
NU
Ta5Si3, Ta2Si
MoSi2 (202)
SC
MoSi2 (211)
Ta (211), SiC (311)
RI
MoSi2 (006)
ACCEPTED MANUSCRIPT

MoSi2 (213) SiC (222)

PT
MoSi2 (204)

Ta (220)
MoSi2 (116)
Fig. 10 X-ray diffraction pattern of (MoSi2 + 20 vol% SiC) + Ta foil laminated composite.
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
DE
PT
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D

Fig. 11 EPMA images of (MoSi2 + 20 vol% SiC) + Ta foil laminated composite (a)
E

back scattered electron image (b) x-ray map of Mo (c) x-ray map of Ta (d) x-
ray map of Si and (e) line scan of Si
PT
CE
AC
 ACCEPTED MANUSCRIPT

Free Energy Change (kJ/mole) 0

-25
Reaction 1
-50 Reaction 2
Reaction 3
-75

PT
-100

RI
-125

SC
-150
200 600 1000 1400 1800 2200
NU
Temperature (K)
MA

Fig. 12 Standard free energy change (ΔG vs. Temperature) for the possible chemical
reactions between MoSi2 and pure Ta resulting in the formation of tantalum
silicides.
E D
PT
CE
AC

Fig. 13 SEM back scattered electron image of (MoSi2 + 20 vol% SiC) + Nb foil
laminated composite.
 Fig. 14
MoSi2 (002)
AC

composite.
MoSi2 (101)
CE
SiC (111)
Nb (110)
PT
E MoSi2 (110)
D MoSi2 (112)
MoSi2 (103)
MA

2 (Degree)
Nb (200)
MoSi2 (200)
NU
SiC (220)
MoSi2 (202)
SC
MoSi2 (211)
Nb (211)
MoSi2 (006), SiC (311)
RI
ACCEPTED MANUSCRIPT

MoSi2 (213) SiC (222)

PT
MoSi2 (204)

Nb (220)
MoSi2 (116)
X-ray diffraction pattern of (MoSi2 + 20 vol% SiC) + Nb foil laminated
 ACCEPTED MANUSCRIPT

PT
d
d

RI
SC
NU
MA

Fig. 15 EPMA images of (MoSi2 + 20 vol% SiC) + Nb foil laminated composite (a)
back scattered electron image (b) x-ray map of Si (c) x-ray map of Mo and (d)
x-ray map Nb.
E D
PT
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
E D
PT
CE
AC

Fig. 16 EPMA images of (MoSi2 + 20 vol% SiC) + Nb foil composite (a) BSE image
with matrix area shaded dark revealing the SiO2 particles present at the
interface (b) x-ray map of Mo and (c) x-ray map of Nb.
 ACCEPTED MANUSCRIPT

-50
Free Energy Change (KJ\mole)

-100

PT
Reaction 1

-150

RI
Reaction 2

Reaction 3

SC
-200
NU
-250
200 600 1000 1400 1800 2200
MA

Temperature (K)

Fig. 17 Standard free energy change (ΔG vs. Temperature) for the possible chemical
reactions between MoSi2 and Nb resulting in the formation of niobium
D

silicides.
E
PT
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
Fig. 18 Mo-Si-Nb isothermal section at 800 °C [33]. The composition widths of the
compounds are schematic.
MA
E D
PT

*
*
CE

*
*
*
AC

*** * * * * * *
*
*
*
*
*
Fig. 19 Schematic diffusion path observed for the reaction between Nb and MoSi 2 at
1600 °C superimposed on the ternary triangle of Fig. 18 at 800 °C.
 ACCEPTED MANUSCRIPT

PT
RI
b

SC
NU
MA
D

Fig. 20 Micro-hardness indentation marks inside the interfacial reaction layer, in the
matrix layer (encircled) and inside the Mo foil in MoSi2-SiC + Mo foil laminated
E

composite as seen under (a) ordinary light and (b) polarized light.
PT
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
Fig. 21 Optical micrographs showing micro-hardness indentation mark inside the

SC
interfacial reaction layer in MoSi2-SiC + Ta foil laminated composite taken at
different magnifications. The indentation mark inside the reaction layer is
encircled in (b) while the same is enclosed by a rectangle in (a).
NU
MA
E D
PT
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
b
MA
E D
PT
CE
AC

Fig. 22 Optical micrographs showing micro-hardness indentation mark inside the

interfacial reaction layer in MoSi2-SiC + Nb foil laminated composite taken at
different magnifications. The Indentation marks inside the interfacial layer and
in the near-by matrix are encircled for the ease of identification.
 ACCEPTED MANUSCRIPT

Table 1 Source and the average particle size of powders used in making MoSi2 based
laminated composites

S No. Powder Source Average Particle Size, (m)

1 MoSi2 H. C. Starck, Germany 6
2 SiC Superior Graphite Co., USA 2

Table 2 Source and the thickness of the refractory metal foils used in making MoSi 2

PT
based laminated composites

S No. Refractory Metal Foil Source Thickness, (mm)

RI
1 Mo Midhani, Hyderabad, India 0.5
2 Ta NFC, Hyderabad, India 0.5

SC
3 Nb NFC, Hyderabad, India 0.5
NU
Table 3 Chemical analysis of as received MoSi2, SiC powders and Mo, Ta, Nb foils
(The amount of elements are in wt%)
MA

S No. Material O C Si Fe N Al
1 MoSi2 1.3 0.08 0.11 0.032 <0.02
2 SiC 0.25 0.8 0.03 0.18
D

3 Mo foil 0.003 0.004

E

4 Ta foil 0.0127 0.003 <0.002 <0.0025

5 Nb foil 0.0213 0.005 0.01 <0.0025
PT
CE

Table 4 Composition of the phases identified in interfacial reaction zone of (MoSi 2 +

20 vol% SiC) + Mo sheet laminated composite as determined by EPMA. All
compositions are given in atom percent
AC

Elements
Location Phase
Mo Si C
A 62.11 37.89 - ~ Mo5Si3
B 68.42 - 31.58 ~ Mo2C
 ACCEPTED MANUSCRIPT

Table 5 Composition of the various phases identified in (MoSi2 + 20 vol% SiC) + Ta

foil laminated composite as determined by energy dispersive spectroscopy
(EDS). All compositions are given in atom percent

Elements
Location Phase
Mo Si Ta
A 17.98 38.40 43.62 ~ (Ta, Mo)5Si3
B - 32.26 67.74 ~ Ta2Si

PT
Table 6 Electron probe micro analysis (EPMA) of the interfacial reaction layer in
(MoSi2 + 20 vol% SiC) + Nb foil laminated composite (Fig. 14). All

RI
compositions are given in atom percent

Region Mo Si Nb Phase

SC
A 34.368 65.632 0.000 ~ MoSi2
B 57.401 38.400 4.199 ~ (Mo, Nb)5Si3
NU
C 1.351 39.137 59.512 ~ (Nb, Mo)5Si3
D 0.000 0.314 99.686 ~ Nb
MA

Table 7 Characteristics of the interfacial reaction layers in various laminated

composites
D

Interfacial Reaction Layer

E

Laminated Composite Thickness

Reaction Products
(m)
PT

(MoSi2 + 20 vol% SiC)

+ 50 Mo5Si3, Mo2C
Mo Sheet
CE

(MoSi2 + 20 vol% SiC)

+ 10 Ta5Si3, Ta2Si
Ta Sheet
(MoSi2 + 20 vol% SiC)
AC

+ 20 Nb5Si3
Nb Sheet

Table 8 Melting points and atomic sizes of elements forming various phases present
in the laminated composites

Property Si Mo Ta Nb
Melting Point (°C) 1414 2623 3020 2469
Atom Size (nm) 11.7 13.6 14.3 14.3
 ACCEPTED MANUSCRIPT

Table 9 Micro-hardness data for (MoSi2 + 20 vol% SiCp) + Mo, Ta and Nb foil
laminated composites

Microhardness
Laminated (VHN)
Composite Matrix Foil Interfacial Reaction Layer
(MoSi2 + 20 vol% SiC)
+ 1358+47 241+10 1298+65
Mo Sheet
(MoSi2 + 20 vol% SiC)
+ 1287+60 193+10 1814+85

PT
Ta Sheet
(MoSi2 + 20 vol% SiC)
+ 1333+65 145+5 1553+75
Nb Sheet

RI
SC
Table 10 Fracture toughness values of Mo, Ta and Nb foil reinforced (MoSi2 + 20 vol%
SiC) matrix laminated composites

Fracture Toughness, Kmax

NU
Laminated Composite
(MPa√m)
(MoSi2 + 20 vol% SiC)
5.2
Monolithic
MA

(MoSi2 + 20 vol% SiC)

+ 15.7
Mo Sheet
(MoSi2 + 20 vol% SiC)
+ 20.5
D

Ta Sheet
(MoSi2 + 20 vol% SiC)
E

+ 17.1
PT

Nb Sheet
CE
AC
 ACCEPTED MANUSCRIPT

Highlights
A variety of reaction products are formed in ductile refractory metal foils reinforced
MoSi2 based laminated composites studied in the present work. The formation of
various phases in the interfacial reaction zone has been analyzed based on the
thermodynamic and kinetic principles. The following conclusions have emerged from
the present investigation:

1. In ductile refractory metal foil reinforced MoSi2 based laminated composites,

PT
interfacial reaction layers of varying thickness were observed to form between
MoSi2-SiCp matrix and the metal foils due to extensive diffusion of Si into Mo,

RI
Ta and Nb foils during high temperature processing.

SC
2. The interfacial reaction products were identified to be mainly silicide phases
consuming a part of the thickness of the ductile metal foils. The main reaction
NU
product was found to be as M5Si3 (M = Mo, Ta or Nb).

3. The Ta foil reinforced laminated composite exhibited the minimum thickness

MA

of the reaction zone of only 10 µm suggesting that the Ta has better chemical
compatibility with MoSi2 as compared to Mo and Nb. The highest
improvement in room temperature fracture toughness was also observed in
D

case of Ta foil laminated composite.

E
PT

4. The micro-hardness measurements made in the reaction zone established

that the interfacial reaction products in all the three laminated composites
CE

were hard and brittle in nature.

AC