Chapter 1. Carbides
for high-temperature applications
The carbides comprise a particularly interesting family of compounds
in that they were the first man made
refractories. Contrasting oxides and silicates they are extremely
uncommon in natural world. Iron carbide, Fe3C and afterwards
TiC and WC were identified and extracted from steels in the mid 1800’s.
French scientist Moisson had synthesized a number of refractory carbides
in his arc furnace in1900 and studied their properties.
Characteristically speaking most of these carbides have good
electrical and thermal conductivity, high stability and high hardness.
These properties account for the principal applications of these carbides:
structural resistant to chemical reaction, uses in which wear resistance
is of major importance, and high temperature radiant-energy sources. The
brittleness of carbides however prevents its application as single-phase
materials in highly stressed structural applications, which in turn led
to the development of metal-bonded composites (cemented carbides or cermets).
The carbides vary widely in their chemical inertness, particularly with
respect to their attack by oxidising atmosphere. At sufficiently high temperature
all of them are attacked quite rapidly, although some are more resistant
than metals of comparable melting point.
Carbides may be grouped accordingly to the periodic classification
of the metal constituents. Alkali- metal –carbides are quite unstable in
nature. The alkaline –earth carbides include several high –melting compounds,
but all are strong reducing agents and hydrolyse rapidly. The carbides
of boron, silicon, and the transition metals are of greatest interest and
utility
1.1 Lattice structures
of carbides
The transition metal carbides have the NaCl type structure. Assuming
metal /metal contact, the atomic radii of the metals in the Group IV carbides
are only about 3% greater than those in the metal crystal from which they
are derived. This suggests that a close-packed structure in which the ratio
of the radii of the atoms, r/R, is geometrically fixed at 0.414,as shown
in Model A, (Fig 1a), describes these compounds. For group V carbides,
however the large difference between the metal atom radii in the metal
and in the carbide suggests that a model of metal /metal contact is less
open structure, shown in Fig 1b,describes their structures.
In Model B, Fig 1b, it is assumed that the main contact is
between metal and carbon and that the radius ratio is variable .If the
radius of the metal atom is taken to be the same as that in the parent
transition metal, the radii of the carbon atoms and hence the ratios r/R
for each carbide can be calculated from the lattice parameters The data
are shown in following tables. On the basis of this model, the effective
radius ratio is about 0.46 for Group IV carbides and about 0.58 for group
V carbides.
Figure 1
(Schematic representation of the carbide by a hard sphere model
of metal atoms (large circles) and carbon atoms (small circles))
Using a similar model, Hagg showed that the carbides have simple
crystal structures for radius ratio <0.59, but above this ratio,
the crystal structure changed. Here, it is proposed that the lattice becomes
progressively less stable as the radius increases above 0.414.
In measuring the radius ratios, the dimensions that should be given
to the radii of the different atom species are subject to some discussions,
particularly if charge transfer takes place in compounds to alter the atom
radii from those in parent crystals. However, from these models, which
utilise dimensions in the carbide crystals, it would seem that a close–packed
metal lattice such as shown in Fig 1a more closely describes the Group
IV carbides. If this is the case, the stability of the crystals may be
expected to arise from metal/metal interactions, consistent with Lye”s
model for the band structure and cohesion in TiC .As soon as the radius
ratio exceeds 0.414, the metal /metal inter atomic forces would be expected
to decrease sharply and the bond strength in<110> directions would be
reduced. Consequently, metal /non-metal interactions may be expected to
become a proportionally greater importance as the radius ratio increases.
The anisotropy factor 2C44/C11 –C12 represents
the ratio of the resistance to shear on a {100} plane in a <001> direction
to that of a {110} plane in <110> direction. For all the Group IV carbide,
the anisotropy factor is <1 while for stoichiometric TaC, it is >1,in
accord with the model discussed in Fig 1 and proposed for the structures
of these carbides. Additional evidence fir this picture is obtained from
the measurements on the photon spectra of tantalum carbide, which suggests
that the metal/non-metal force constants are greater than the metal/metal
interactions in this compound. There are also other indications that the
lattices of TaC and NbC are different from those of the Group IV carbides.
For example both are comparatively soft and both are super conductors at
high carbon-to-metal ratio that suggests that the lattice are somewhat
unstable.
Out that little information on the deformation characteristics and
strengths of single crystal of tantalum carbide is yet available, although
several workers have grown single crystals. Mechanical property measurements
have been made on the arrangement of substoichiometric polycrystalline
tantalum carbides made by carburising tantalum carbides or by hot pressing
powders.
Table 1. A Comparison
between radii of metal atoms in carbides and in the pure metal
| Carbide | Lattice Parameter
(A) |
Metal AtomicRadius In Metal (A) | Metal Atomic Radius In Carbide (A) | Expansion of Metal Atoms in Carbide (%) |
| TaC | 4.45 | 1.42 | 1.57 | 10.5 |
| NbC | 4.47 | 1.43 | 1.57 | 10.0 |
| VC | 4.16 | 1.31 | 1.47 | 12.5 |
| HfC | 4.64 | 1.59 | 1.64 | 2.50 |
Table 2.
Radius ratios of the carbides
| Carbide | Atomic Radius in Metal(A)
|
Radius of Carbon
site
(A) |
Radius Ratio |
| TaC | 1.42 | 0.81 | 0.57 |
| NbC | 1.43 | 0.80 | 0.56 |
| VC | 1.31 | 0.77 | 0.59 |
| HfC | 1.59 | 0.73 | 0.46 |
| ZrC | 1.61 | 0.74 | 0.46 |
| TiC | 1.47 | 0.69 | 0.47 |
1.2 Mechanical Properties
Tensile, compression, and flexure tests all have been made on self-bonded carbides but yield numerical data of only limited utility. In common with many other high-temperature materials, analysis and comparison of most mechanical property data available on carbides are made difficult by the extreme brittleness of these materials at ordinary temperatures. Specific results in many instances include not only a wide scatter in reported strength values but also an apparent inverse temperature dependence of strength. Such tests can only be used to study trends in the effects of compositional variables. They cannot be used to access qualitatively the intrinsic strengths of carbides, to compare results of different investigators, or to design engineering structures.
The maximum high temperature flexure strengths that have been reported for carbides have been obtained on hot-pressed B4C (78,000psi at 1500oC and 28,000 at 2000oC). Aluminium additions are reported to increase the strength of hot-pressed B4C.
The principal variables found to affect mechanical properties
of carbides are density, grain size and stoichiometry. The pronounced effects
of small amount of residual porosity after sintering are seen in data on
TiC as shown the Figure2.
Figure 2
(Effects of relative density on bending strength. After Glaster
and Ivanick)
Increase in grain size with increasing density in sintering
may reduce low-temperature strength and the usual accelerated grain growth
during the final stages of sintering acts to prevent complete densification,
so that the deleterious effects of these factors on strength properties
are compounded. Typical results of these effects on room –temperature bend
strength are shown in the Figure3 from Cr3C2 .The
same investigators found, however, a negligible effect of grain size (6
to 18 m) on the 1300oC bend strength.
Figure 3
(Combined effects of grain size and density on strength
in Cr3C2. After Hamjjan and Lidman)
Structural and chemical variations also affect mechanical properties. These include deviations from stoichiometry, dissolved impurities and the presence of second phase, principally free carbon. These factors are frequently interrelated and thus it is difficult to separate their effects. For example in the Figure 2, the grade 2 TiC had lower free carbon and higher combined carbon than grade 1, but which factor is more responsible for the strength improvement cannot be stated certainly.
Kovalskii and Makarenko and Cadoff et al have studied the effects
of combined carbon content on the room-temperature micro-hardness of TiC.
They find maximum hardness at compositions close to stoichiometric and
steep decrease in hardness with decreasing carbon content as shown in Figure
4. Whether this result is due solely to an increase in the concentration
of lattice defects with departure from stoichiometry is not certain.
Figure 4
(Effect of carbon content on the hardness
of TiC. After Cadoff, Nielson, and Miller)
Few studies have been made of solid-solution alloying effects in
self-bonded carbides. Kieffer and Kolbl have examined such effects in pseudo-binary
carbide systems with room-temperature micro-hardness tests but find only
small positive deviations from mixture rules as shown in figure 5.
Figure 5
(Effects of solid solution alloying on micro-hardness. After Kieffer
and Kolbl)
On the other hand, some moderately large effects have been observed is systems of limited solid solubility. Dissolved interstitials are generally believed to exert an important influence on properties, but few explicit data are available. Cadoff et al. studied the effects on TiC of solution alloying with TiO and found that oxygen increases both hardness and brittleness. Softening must ultimately result, however, from large TiO additions. The brittleness of carbides is readily appreciated but difficult to determine quantitatively.
The rate of increase of electrical conductivity with densification
of a cold compact has also been used as a measure of brittleness. No successful
correlations have been established so far between the brittleness or toughness
of carbides and fundamental physical properties or structures.
1.3 Chemical stability at high temperature
Many applications of carbides require some resistance to decomposition or chemical attack at a free surface. In vacuum and inert atmospheres, the high melting points and low vapour pressures give the carbides of group IV and V exceptional stability. Where the vapour pressure of carbon is higher than that of the metal, a loss of carbon has been observed close to the melting point. Hydrogen does not react appreciably with carbides in Groups IV, V, and VI at 2200oC to 2500oC although it promotes carbon removal from TaC close to the melting point and is known to be dissolved in vapour-deposited TiC.
The transition-metal mono carbides are most stable than are the mono nitrides at high temperatures. Group VI carbides are unaffected by nitrogen. Dry nitrogen does not react with TaC close to its melting point. However in TiC the solid solution in equilibrium with carbon contains increasing amounts of nitrogen as the temperature or total pressure in reduced.
The transition–metal carbides are attacked severely by halogens at a very high temperature. Boron Carbide (B4C) reacts with Cl2 at about 700oC, SiC will erode about 2mm in 100 hours at 800oC in Cl2.
Oxidation is less rapid for the carbides than for the corresponding nitrides but is more rapid than for the borides. Attack by oxygen and water increases on going from Group IV to Group VI. The carbides exhibit greatest stability when a dense, protective oxide film is formed and a parabolic-rate oxidation curve is obtained. This can be achieved in cases where the metal has a high affinity for oxygen and the rate may be similar to, or slightly less than, that of the pure metal (titanium, zirconium) In other case formation of CO gas can cause rupture of the oxide film, and there is much more rapid rate of oxidation than for the pure metal. The ability of the oxide to sinter into a protective layer as temperature is increased has been one explanation for the existence of intermediate temperature ranges in which oxidation is at a minimum.
Stability of carbides in contact with molten slags and fused salts is generally poor. Silicon carbide is dissolved in NaOH at about 3mm/100hr at 500oC. It is attacked by eutectics containing copper, nickel, chromium and iron oxides. Boron Carbide, B4C reacts with transition –metal carbides to form borides. It is also attacked by liquid cobalt, nickel and iron.
In regard to resistance to liquid metals, nickel, cobalt, chromium,
and silicon attack TiC, although eleven other liquid metals have little
effect close to the melting point. At 1800oC, reaction also
occurs with iron, zirconium and titanium.