See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/236045018

Oxyacetylene torch testing and microstructural characterization of tantalum

carbide

Article in Journal of Microscopy · March 2013

DOI: 10.1111/jmi.12028 · Source: PubMed

CITATIONS READS

15 2,278

5 authors, including:

Anish Paul Jon Binner

Ansaldo Energia University of Birmingham
18 PUBLICATIONS 1,041 CITATIONS 236 PUBLICATIONS 7,076 CITATIONS

SEE PROFILE SEE PROFILE

Bala Vaidhyanathan
Loughborough University
115 PUBLICATIONS 4,273 CITATIONS

SEE PROFILE

All content following this page was uploaded by Bala Vaidhyanathan on 19 September 2018.

The user has requested enhancement of the downloaded file.

Journal of Microscopy, 2013 doi: 10.1111/jmi.12028
Received 21 November 2012; accepted 15 February 2013

Oxyacetylene torch testing and microstructural characterization

of tantalum carbide

A . P A U L ∗, J . G . P . B I N N E R ∗, B . V A I D H Y A N A T H A N ∗, A . C . J . H E A T O N † & P . M . B R O W N †
∗ Department of Materials, Loughborough University, Loughborough, Leicestershire, UK

†DSTL, Porton Down, Salisbury, UK

Key words. Oxyacetylene, tantalum carbide, ultra-high temperature.

Summary superior oxidation resistance for Ta2 C and the formation of

an oxycarbide layer at the Ta2 C–Ta2 O5 interface, whereas
Tantalum carbide samples have been subjected to high-
no such oxycarbide was observed for TaC. Thermogravimet-
temperature testing at ∼2300◦ C using an oxyacetylene torch
ric analysis, at temperatures up to 1500◦ C, has been used to
to evaluate their potential for ultra-high temperature applica-
compare the oxidation behaviour of TaC, TaC+10 wt% TaB2
tions. While large samples cracked during the rapid heating,
and pure TaB2 (Zhang et al., 2008). It was observed that the
indicating their inability to withstand thermal shock, small
oxidation resistance of TaC was marginally improved by the
samples survived the severe test conditions. The oxidation
addition of TaB2 . Sciti et al. (2009) studied the oxidation of
products formed were characterized and found to comprise
TaC with separate additions of 15 vol% of MoSi2 and TaSi2 at
different phases of Ta2 O5 . The ultra-high temperature experi-
1600◦ C in static air, exposing the samples for 15 min. Both
enced by the samples resulted in the formation of many inter-
materials were found to undergo severe oxidation, which con-
esting microstructures, including the formation of submicron
siderably reduced the strength of the composites.
sized grains, which has not been reported previously in the
There are no reports on the ultra-high temperature testing
literature, as well as the expected evidence of melting.
of TaC ceramics using an oxyacetylene flame and the result-
ing morphology of the microstructure. This paper compares
Introduction the oxidation behaviour and microstructure at and around
the ablation centre of TaC samples subjected to oxyacetylene
Tantalum carbide, TaC, is a refractory carbide with one of
torch testing at temperatures up to ∼2300◦ C.
the highest known melting points (3985◦ C) and is consid-
ered as an ultrahigh temperature ceramic, UHTC (Andersson
et al., 2000; Tulhoff, 2000). There are several subcarbides of
Experimental methods
the Ta–C system with lower carbon contents; however, the
monocarbide, TaC, is the most widely used phase. Its main 100 × 100 × 10 mm TaC blocks were obtained from Cerac
application is in cemented carbide as its addition improves the Inc., Milwaukee, USA. According to the manufacturer, the
high-temperature fatigue strength, thermal shock resistance, sample was densified via hot pressing and had a purity of
hot hardness and resistance against cratering and oxidation 99.5%. The density of the sample was measured using the
(Andersson et al., 2000; Tulhoff, 2000). TaC is a potential can- Archimedes principle in water and was found to be 93% of
didate for ultra-high temperature applications such as leading theoretical. 15 and 32 mm diameter discs for oxyacetylene
edges and nose caps for hypersonic vehicles, rocket thrusters, torch testing were machined out of the ceramic block using
combustor linings, etc., because of its high melting point. electrical discharge machining (EDM).
There are only limited studies on the high-temperature A purpose built oxyacetylene torch test rig was used to per-
oxidation behaviour of TaC materials. Desmaison-Brut et al. form the high-temperature oxidation testing. A schematic of
(1997) compared the oxidation behaviour of hot isostatically the torch test facility is shown in Figure 1. Oxygen and acety-
pressed TaC and Ta2 C under isothermal and isobaric condi- lene in a 1.35:1 ratio was fed through a welding nozzle (No.13
tions in a dynamic flow of pure oxygen using a microbalance nozzle, BOC Industrial, Surrey, UK) to produce an oxidizing
at relatively low temperatures of up to 850◦ C. They reported flame. For testing, the samples were positioned in a water
cooled graphite holder that was then manually moved slowly
Correspondence to: Anish Paul, Department of Materials, Loughborough University, into the hot zone of the flame. The samples were held there for
Loughborough, LE11 3TU, UK. Tel: +44-1509-223154; fax: +44-1509-223949; 60 s and then the flame was extinguished allowing the sam-
e-mail: [email protected] ples to cool down naturally. During the test, the temperature

C 2013 The Authors

Journal of Microscopy 
C 2013 Royal Microscopical Society
2 A. PAUL ET AL.

modal with an average size of 2–3 μm. Porosity is clearly

visible as the ceramic was only 93% dense. From the XRD
pattern it can be seen that all the peaks correspond to the face
centred cubic structure.
The time temperature profiles for the 15 mm diameter
discs held at different distances from the nozzle during torch
testing are given in Figure 3. It can be seen that increasing dis-
tance between the sample and nozzle resulted in much lower
heating rates and peak front face temperature, as expected.
The heating rate for the first 5 s decreased from 330◦ C s−1 to
325◦ C s−1 to 275◦ C s−1 , whereas the peak front face temper-
Fig. 1. Oxyacetylene torch test set up. (1) Back face thermocouple, (2) ature decreased from 2300◦ C to 2190◦ C and then to 1990◦ C
water cooling, (3) sample holder, (4) sample, (5) guide rail, (6) protective as the distance increased from 10 mm to 15 mm and then to
insulation, (7) oxyacetylene torch, (8) neutral density filter, (9) thermal
20 mm respectively. It is well known that the gas flow rate and
imaging camera and (10) two-colour pyrometer.
gas flow ratio has a pronounced influence on the nature of the
flame and its temperature; these two parameters were kept
distribution was recorded using a modified infrared thermal constant and only the distance was varied. The oxyacetylene
imaging camera (Thermovision A40, FLIR Systems AB, Dan- flame has an inner cone and an outer feather and the peak
deryd, Sweden) and the peak temperature was recorded us- temperature is at the tip of the inner cone (Campbell, 1919).
ing a Marathon MR1SCSF two-colour infrared pyrometer Figure 4 shows a view of one of the samples at 15 mm from
(Raytek GmbH, Berlin, Germany). The back face temperature the tip of the nozzle, through a welding goggle, during testing.
of the samples was also recorded during testing using a K-type It can be seen that the inner cone of the flame is just touching
thermocouple connected to a data logger. the sample surface. The length of this cone was determined to
For testing the 15 mm diameter discs, the samples were be ∼16 mm so this sample was exposed to the highest temper-
placed in a carbon–carbon insert prior to fixing inside the ature. As well as a decrease in temperature, a decrease in gas
graphite sample holder whereas the 32 mm samples were flow velocity was also expected with an increase in distance
fixed directly within the holder. For the 15 mm diameter between the sample and the nozzle tip, though the magnitude
discs, distances of 10, 15 and 20 mm were used between the of this could not be quantified.
sample and the nozzle to alter the peak temperature and heat- Images of the samples after testing are shown in Figure 5.
ing profile. Four 32 mm diameter samples were also tested at The front faces can be seen to be covered with a white layer
10 mm distance, but three of them shattered within a few sec- indicating that the expected oxidation had occurred. The sur-
onds of moving into the flame, even before reaching the peak face oxide layer completely detached from the base material
temperature. for the sample tested at 20 mm from the nozzle tip, whereas
The surface microstructure of the TaC before and after there was some adhesion for the other samples, possibly due
oxyacetylene torch testing was characterized using a field the formation of liquid phases at the test temperature and
emission gun scanning electron microscope (FEGSEM, LEO their subsequent solidification on cooling. However, no signs
1530VP, Carl Zeiss SMT, Oberkochen, Germany) and the of surface erosion or cracking were observed for the 20 mm
chemical composition was examined using energy disper- sample, whilst erosion, melting and cracking were all observed
sive spectroscopy, EDS (EDAX, EDAX Inc., NJ, USA). The mi- for the other two distances. On cross-sectional examination,
crostructure and composition beneath the oxide layer was also the cracks were found to propagate into the bulk.
analysed after removing the oxide; the adhesion between the The response of the 32 mm diameter TaC discs to the flame
base layer and oxide was poor and the latter was easily re- was completely different due to the large thermal gradients
moved. X-ray diffraction patterns of the as-received TaC and that developed. Unlike the 15 mm diameter discs, which
the various oxidation products formed after the oxyacetylene were embedded in a carbon-carbon holder, these samples
torch tests were recorded with a Bruker D8 model diffractome- were held directly by the graphite sample holder using three
ter (Bruker AXS GmbH, Karlsruhe, Germany) using Cu-Kα graphite bolts. During the test, there was rapid heating and
radiation. The samples were cross sectioned after the oxyacety- severe thermal gradients. As the samples tried to expand, the
lene test, polished and characterized using FEGSEM and EDS. graphite bolts resulted in stresses being generated that resulted
in crack formation. Three out of four 32 mm TaC discs tested
shattered and fell out of the sample holder before reaching
Results
even 1000◦ C. While the fourth sample cracked, it didn’t fall
The microstructure, EDS spectrum and XRD pattern of out allowing the 60 s test to be completed, Figure 6. The cracks
the TaC before the oxyacetylene torch test are shown in originated from the centre of the sample, where the oxyacety-
Figure 2. It can be seen that the grain size distribution is uni- lene flame was focused, and propagated to the points where


C 2013 The Authors

Journal of Microscopy 
C 2013 Royal Microscopical Society
 OXYACETYLENE TORCH TESTING AND MICROSTRUCTURAL CHARACTERIZATION OF TaC 3

Fig. 2. (a) SEM, (b) EDS and (c) XRD of TaC before torch testing.

it was held by the bolts; highlighted by the arrows. One way The formation of a white layer was also observed on the
to avoid this cracking in the future is to eliminate any stresses 32 mm diameter sample front surface, together with straw
from the sample holder. The thermal gradients on the 32 mm coloured, frozen liquid phases with a clear melt boundary. X-
disc are shown in Figure 7 and the peak back face temperature ray diffraction patterns of both the white surface layer and
was recorded to be ∼1050◦ C. the solidified molten structure are shown in Figure 8 (a) and
(b); the layer comprised α and β phases of Ta2 O5 , with the
strongest peak corresponding to that of the α. Similarly, the
frozen droplets were also found to comprise mainly α, with
small amounts of β also being detected.
Cross-sectional analysis of the 15 mm discs was performed
to compare the thickness of the oxide layers formed and the
chemical compositions, Figure 9. An increase in oxide layer
thickness was expected with an increase in temperature, but
a decrease was actually observed at the highest temperatures
due to the melting of the Ta2 O5 and its removal by the gas flow.
This clearly highlights the disadvantage of using this type of
material at ultra-high temperatures when a high-velocity gas
flow is present. EDS analysis of the frozen droplets and the
oxide layer indicated the presence of Ta and O, whereas the
bulk, unaffected material continued to reveal only Ta and C.
Fig. 3. Time-temperature profiles of TaC discs subjected to oxyacetylene The surface of the disc after oxyacetylene torch testing can
flame at 10, 15 and 20 mm from the nozzle. be divided into three separate zones, each with a different

C 2013 The Authors

Journal of Microscopy 
C 2013 Royal Microscopical Society
4 A. PAUL ET AL.

Fig. 4. View of sample through a welding goggle during oxyacetylene

torch testing; the inner cone of the flame is clearly visible. The sample was
15 mm from the nozzle tip.
Fig. 6. 32 mm diameter TaC sample after an oxyacetylene torch test. The
graphite sample holder is also seen and the arrows indicate the points
where the sample was gripped using the bolts. Inset picture shows the
TaC disc before testing. The numbers indicate the three different zones
observed after torch testing. (1) Flame tip zone, (2) flow zone and (3) outer
zone.

Figure 11 (a) shows the formation of large columnar grains

growing in a radial direction. These will have been formed as
a result of the high-velocity flame pushing the molten oxide
Fig. 5. Photographic images of the TaC discs after oxyacetylene torch
out of zone 1. A higher magnification view of these columnar
testing at (a) 20 mm, (b) 15 mm and (c) 10 mm distance from the nozzle
tip.
grains indicates the presence of very fine, dense looking grains,
Figure 11 (b), (c) and (d). EDS analysis (not shown) detected
only Ta and O in these fine grains.
morphology, Figure 6. Zone 1 is at the flame tip, where the
Near to the melt boundary, but within zone 2, the colum-
flame directly impinged the sample and hence the tempera-
nar grains were oriented in random directions because of the
ture was highest. Zone 2 is the flow zone; the temperature
rapid cooling, as shown in Figure 12. The segregation of a
experienced was lower and the molten products formed in
secondary phase was identified on some of the frozen droplets
zone 1 were swept across by the high-velocity gas flow from
and EDS analysis confirmed the presence of Fe, which is be-
the torch nozzle. Zone 3 experienced the lowest temperature
lieved to have come from the impurities present in the starting
and no melting was observed. A clear melt boundary is visible
material, Figure 13, although no Fe was detected on the EDS of
between zones 2 and 3.
the starting material, Figure 2 (b). It is assumed that the very
The detailed microstructural analysis of the sample after
low quantities of Fe present in the starting material were uni-
testing revealed several interesting features. The white oxide
formly distributed and during ultra-high-temperature testing
layer formed on the surface of the sample in zone 1 was found to
they melted and coalesced in the Ta2 O5 droplets. A similar type
be highly porous, Figure 10 (a). This is believed to be due to the
of segregation was reported by Sautereau & Mocellin (1974)
escape of the CO/CO2 gases formed at high temperatures and
after the thermal etching of sintered TaC and NbC samples at
the porosity allows easy diffusion of oxygen towards the bulk of
1400◦ C. They reported that the impurities present in the start-
the sample. The microstructure beneath the white oxide layer
ing powders coalesced and formed droplets, predominantly at
in zone 1 is shown in Figure 10 (b). This region is not as porous
the grain centres, during the thermal etching process. The use
as the surface layer, indicating only a partial oxidation and the
of Fe as a sintering aid during the hot pressing of TaC has also
individual grains were observed to have small pores within
been reported (Scholz, 1963).
them. EDS analysis of this region detected mainly Ta and O. If
the sample was held at the testing temperature for longer, it is
Discussion
believed that these grains would also have developed porosity
similar to that of the surface layer. Oxidation of tantalum carbide has been observed to proceed
Many interesting microstructures were observed in zone with the formation of a porous, nonprotective oxide scale.
2, the area between the flame tip and the melt boundary. Melting of the scale is observed when the test temperature is


C 2013 The Authors

Journal of Microscopy 
C 2013 Royal Microscopical Society
 OXYACETYLENE TORCH TESTING AND MICROSTRUCTURAL CHARACTERIZATION OF TaC 5

Fig. 7. Thermograph showing the thermal gradients developed during the oxyacetylene torch testing of a 32 mm diameter TaC disc at 10 mm from the
nozzle. (a) Before cracking, (b) immediately after cracking and (c) before extinguishing the flame.

>1900◦ C. The major reactions taking place during the oxida- 900–1300◦ C, depending on the oxygen partial pressure (Chen
tion of TaC are given by equations 1 and 2 (Desmaison-Brut et al., 2010). No such oxides were observed after oxyacetylene
et al., 1997). torch testing, though they might have been formed during the
heating stages.
4TaC + 7O 2 → 2Ta2 O 5 + 4CO (1)
The indexing of the X-ray diffraction pattern of the low-
temperature form of Ta2 O5 has always confused crystallog-
2TaC + 9/2O 2 → T a 2 O 5 + 2CO2 (2)
raphers. Most of the intense diffraction lines can be indexed
The formation of other oxides, such as TaO2 and Ta2 O3 , using an orthorhombic unit cell; however, many of the weak
has also been reported at intermediate temperatures of around lines have not been interpreted unambiguously. The rea-
son for this is that the position of the weak lines is depen-
dent on the nature of the heat treatment and the amount
of cation and anion impurities (Roth et al., 1970). The sit-
uation is made more complex in this study because of the
rapid heating and cooling involved. It has also been reported
that the high-temperature polymorph (α-Ta2 O5 ) undergoes
several phase transformations on quenching. Although it is
triclinic at room temperature, it is monoclinic at intermedi-
ate temperatures and tetragonal at the temperature at which
it is stable (Waring & Roth, 1968). XRD patterns of the
oxide layers from this study indicated the presence of both
α and β-Ta2 O5 . The crystal structure of the α-phase was ob-
served to be triclinic. β-Ta2 O5 is a low-temperature phase
with a transition temperature of 1320 ± 20◦ C to the α phase,
which is stable at high temperature (Lagergren & Magnéli,
1952). Reisman et al. (1956) reported a slightly higher trans-
formation temperature of 1360 ± 5◦ C after heating Ta2 O5
using an oxy-gas flame and also suggested that this trans-
formation is sluggish but completely reversible. Quenching
of the molten phases resulted in a straw coloured, crys-
talline Ta2 O5 that had a similar structure to the frozen
droplets from this study. For Reisman et al. (1956) how-
ever, their X-ray diffraction pattern of the solidified struc-
ture revealed the presence of just the α-phase whereas β-
phase was also observed in this study. The presence of
the former is expected to be due to the rapid heating and
cooling and the sluggish transformation behaviour. An in-
crease in the α-phase intensity for the frozen droplets is
possibly due to the higher temperature experienced by the
droplets and the minimal time available for this phase to
Fig. 8. X-ray diffraction patterns on the 32 mm diameter sample after
revert to β due to rapid cooling. The presence of ε and δ-
oxyacetylene torch testing. (a) White layer formed on the surface and phases (Izumi & Kodama., 1979), and many high-pressure
(b) solidified molten structure. phases (Filonenko & Zibrov, 2001; Zibrov et al., 2000), is also

C 2013 The Authors

Journal of Microscopy 
C 2013 Royal Microscopical Society
6 A. PAUL ET AL.

Fig. 9. Cross-sectional images from directly below the flame tip from the 15 mm diameter samples subjected to oxyacetylene torch testing at three different
distances from the nozzle, (a) 20 mm (b) 15 mm and (c) 10 mm. The EDS patterns in (d) were recorded from locations 1, 2 and 3 in Fig. 9b.

reported in the literature, but was not observed in this work. at around 1800◦ C (1785 ± 30◦ C for β-Ta2 O5 and 1872◦ C
Even though the melting point of TaC is around 3985◦ C, it for α-Ta2 O5 (Reisman et al., 1956)), which is well be-
is reported to undergo oxidation at temperatures as low as low the oxyacetylene torch test temperatures used in this
800◦ C (Zhang et al., 2008) and the resulting oxides melt study.

Fig. 10. FEGSEM image of (a) white surface oxide layer, (b) partially oxidized grains below the surface layer, (c) the EDS pattern taken on the partially
oxidized grains in (b).


C 2013 The Authors

Journal of Microscopy 
C 2013 Royal Microscopical Society
 OXYACETYLENE TORCH TESTING AND MICROSTRUCTURAL CHARACTERIZATION OF TaC 7

could be due to the large difference in testing temperatures

and the more oxidizing conditions.
Gaseous CO/CO2 is produced as a by-product during the
oxidation of TaC. The escape of these gases produces continu-
ous porosity in the oxide layer, as reported by Courtright et al.
(1991). The pores act as pathways for both the inward trans-
port of oxygen and further outward diffusion of CO, preventing
CO build up and catastrophic delamination of the oxide layer
under static oxidation conditions. The delamination of the ox-
ide layer observed in this study was again partially due to the
rapid heating rates involved, these will have limited the time
available for the gasses formed to escape. Further factors are
outlined below.
The pores present in the grains beneath the fully oxidized
Fig. 11. (a) SEM image showing the formation of columnar grains. surface layer will have been created by the partial internal
(b) Higher magnification view of the area highlighted by the rectangle in oxidation of the TaC grains. Similar observations were made
(a). (c) and (d) are higher magnification images of the submicron grains. by Courtright et al. (1991). A well-defined carbide-oxide
boundary layer indicated the reaction front. The oxidation re-
action taking place at this interface can be assumed to control
the rate of oxidation. Similar interfacial oxidation reactions
were also reported for other high-temperature carbides such
as HfC and ZrC with an oxygen gradient from the oxide layer
to the carbide. The diffusion of oxygen into the carbide limits
the rate of the reaction (Shimada, 2001). A change in the rate
controlling mechanism at ∼1800◦ C has also been reported
in the literature. The lower temperature kinetics can be
Fig. 12. SEM images showing the random orientation of grains at the explained by assuming that oxide growth is controlled by O2 ,
melt boundary. CO and CO2 gaseous transport through the interconnected
pores in the oxide. This process became less important at
increasing temperatures as the oxide sintered and the effective
The XRD results are not completely in agreement with those volume available for gaseous diffusion was reduced (Cour-
reported after the oxidation of C-TaC-C composites at temper- tright et al., 1991). During the oxyacetylene torch testing, the
atures up to 1400◦ C (Chen et al., 2010), where the formation oxides formed melted making it more difficult for the O2 to
of a hexagonal structure (δ-Ta2 O5 ) was reported at tempera- diffuse through the viscous liquid phase compared to diffusing
tures below 900◦ C. This then transformed to an orthorhombic through the interconnected porous oxide. Also, near the tip of
(β-Ta2 O5 ) structure above this temperature. This difference the flame, the melt formed was pushed away from the flame

Fig. 13. FEGSEM image showing the formation of a secondary phase on the solidified droplets. EDS detected the presence of Fe along with Ta, O and C.

C 2013 The Authors

Journal of Microscopy 
C 2013 Royal Microscopical Society
8 A. PAUL ET AL.

front by the high-velocity gas flow hitting the sample, expos- References
ing fresh surfaces for oxidation and influencing the rate of the
Andersson, K., Reichert, K. & Wolf, R. (2000) Tantalum and tantalum
reaction. The formation and escape of any intermediate oxides compounds. Ullmann’s Encyclopedia of Industrial Chemistry, pp. 1–15.
such as TaO2 and Ta2 O3 will also have led to the generation Wiley-VCH Verlag GmbH & Co. KGaA, New York.
of more porosity, helping the rapid inward diffusion of oxygen Campbell, L. (1919) Chapter II: operation. Oxy-Acetylene Welding Manual,
and accelerating the oxidation process (Chen et al., 2010). In pp. 27–38. John Wiley and Sons, Inc., New York.
addition to gas build up, the difference in coefficient of thermal Chang, L.L.Y. & Phillips B. (1969) Phase relations in refractory metal-
expansion between the oxide and the underlying carbide can oxygen systems. J. Am. Ceram. Soc. 52, 527–533.
also play a role in delamination. The CTE of Ta2 O5 is reported Chen, Z., Xiong X., Li G. & Wang Y. (2010) Mechanical properties and
to be around 2.9 × 10−6◦ C−1 from room temperature oxidation behaviors of carbon/carbon composites with C-TaC-C multi-
to 500◦ C (Moldovan et al., 2004) whereas that of TaC is interlayer. J. Mater. Sci. 45, 3477–3482.
Courtright E.L., Prater J.T., Holcomb G.R., Pierre G.R.S. & Rapp R.A.
6.29 × 10−6◦ C−1 at room temperature (Tulhoff, 2000).
(1991) Oxidation of hafnium carbide and hafnium carbide with addi-
The molten oxides recrystallized on cooling forming colum-
tions of tantalum and praseodymium. Oxid. Met. 36, 423–437.
nar grains. The formation of the submicron grains is expected Desmaison-Brut M., Alexandre N. & Desmaison J. (1997) Comparison of
to be due to the formation of tantalum suboxides, including the oxidation behaviour of two dense hot isostatically pressed tanta-
TaO2 and Ta2 O3 at intermediate temperature of oxidation, lum carbide (TaC and Ta2 C) materials. J. Eur. Ceram. Soc. 17, 1325–
probably during the heating up stage. The formation temper- 1334.
ature of some of these suboxides are reported to be <600◦ C Filonenko, V.P. & Zibrov I.P. (2001) High-pressure phase transitions of
(Chang & Phillips, 1969). It is believed that some of these ox- M2 O5 (M = V, Nb, Ta) and thermal stability of new polymorphs. Inorg.
ides will have vaporized at the ultra-high temperatures being Mat. 37, 953–959.
used (Chen et al., 2010) and then deposited at regions of rel- Izumi, F. & Kodama H. (1979) A new modification of tantalum (V) oxide.
J. Less Comm. Metals 63, 305–307.
atively lower temperature on the sample surface. They would
Lagergren, S. & Magnéli A. (1952) On the tantalum–oxygen system. Acta.
subsequently undergo sintering, leading to the formation of
Chem. Scand. 6, 444–446.
dense-looking submicron grains on the surface. The presence
Moldovan, M., Weyant C., Johnson D. & Faber K. (2004) Tantalum oxide
of molten phases is expected to act as nucleation sites for the coatings as candidate environmental barriers. J. Therm. Spray Technol.
growth of these grains. 13, 51–56.
Reisman, A., Holtzberg F., Berkenblit M. & Berry M. (1956) Reactions of
the group VB Pentoxides with alkali oxides and carbonates. III: thermal
Conclusions and X-ray Phase diagrams of the system K2 O or K2 CO3 with Ta2 O5 .
J. Am. Chem. Soc. 78, 4514–4520.
Even though TaC has the highest melting point amongst all
Roth, R.S., Waring J.L. & Parker H.S. (1970) Effect of oxide additions on the
the potential materials for UHTC applications, the results of the polymorphism of tantalum pentoxide. IV. The system Ta2 O5 -Ta2 WO8 .
oxyacetylene torch testing proved its inability to fulfil the re- J. Solid State Chem. 2, 445–461.
quirements for such applications. TaC was unable to withstand Sautereau, J. & Mocellin A. (1974) Sintering behaviour of ultrafine NbC
the high thermal shocks experienced during the test, leading and TaC powders. J. Mater. Sci. 9, 761–771.
to severe cracking. It oxidized at temperatures well below the Scholz, S. (1963) Some new aspects of hot pressing of refractories. Special Ce-
melting point of the carbide and the oxides melted at tem- ramics 1962, Proceedings of a Symposium Held by the British Ceramic
peratures below 1900◦ C rendering these materials unsuitable Research Association, 293–307.
for applications requiring ultra-high temperature (>2000◦ C) Sciti, D., Silvestroni L., Guicciardi S., Fabbriche D.D. & Bellosi A. (2009)
resistance. Moreover, the oxide layer formed was porous and Processing, mechanical properties and oxidation behavior of TaC and
HfC composites containing 15 vol% TaSi2 or MoSi2 . J. Mater. Sci. 24,
not completely adherent, offering little or no additional protec-
2056–2065.
tion at elevated temperatures. The oxides were found to com-
Shimada, S. (2001) Interfacial reaction on oxidation of carbides with
prise the low- and high-temperature stable phases of Ta2 O5 . formation of carbon. Solid State Ionics 141–142, 99–104.
The formation of submicron grains was expected to be due to Tulhoff, H. (2000) Carbides. Ullmann’s Encyclopedia of Industrial
the volatilization, deposition and subsequent sintering of the Chemistry, pp. 555–582. Wiley-VCH Verlag GmbH & Co. KGaA,
oxide phases formed at intermediate temperatures. New York.
Waring, J.L. & Roth R.S. (1968) Effect of oxide additions on the polymor-
phism of tantalum pentoxide (System Ta2 O5 –TiO2 ). J. Res. Nat. Bur.
Acknowledgements Stand. Sect. A. 72A, 175–186.
Zhang, X., Hilmas G.E. & Fahrenholtz W.G. (2008) Densification, me-
The authors thank the UK’s Defence Science and Technology chanical properties, and oxidation resistance of TaC-TaB2 ceramics.
Laboratory (DSTL) for providing the financial support for this J. Am. Ceram. Soc. 91, 4129–4132.
work under contract number DSTLX-1000015267 as well as Zibrov, I.P., Filonenko V.P., Sundberg, M. & Werner P.-E. (2000) Struc-
the US Air Force Research Laboratory’s Materials and Manu- tures and phase transitions of B-Ta2 O5 and Z--Ta2 O5 : two high-
facturing Directorate for on-going collaborations. pressure forms of Ta2 O5 . Acta Crystallog. Sec. B 56, 659–665.


C 2013 The Authors

Journal of Microscopy 
C 2013 Royal Microscopical Society

View publication stats