The heavy alloys are tungsten based two-phase composites used in applications requiring
high density. The alloys are liquid phase sintered from blended elemental powders. After
sintering, the microstructure consists of a rounded tungsten phase (typically 50 am in
diameter) surrounded by a matrix phase containing dissolved tungsten.
The typical chemical composition ranges from 80 to 98% W with either
Ni-Cu, Ni-Fe or Ni-Fe-Co additions. Understandably, the mechanical properties are
variable with microstructure, chemistry and processing. Yield strengths in excess of
500 MPa are fairly common, however, ductility and toughness tend to be unpredictable.
Generally, the Ni-Fe alloys exhibit superior mechanical properties and a 7:3 ratio of
nickel to iron is observed to be optimal.
In spite of numerous studies on the heavy alloys dating back to the 1930’s, there
is still uncertainty as to the sources of toughness variation. Considering the large
number of parameters associated with this material, the observed variability in toughness
is not surprising.
Generally, the factors influencing toughness can be divided into three categories. First
are those factors, which produce differing results between studies such as composition,
sintering temperature, test geometry, sintering atmosphere, and heat treatment. Second
are those factors, which give differing properties between similarly processed heats
such as density, pore size, impurities and particle size. Third are the factors, which
contribute to property variations within a single heat of heavy alloys such as thermal
and gravitational gradients. All these factors are interrelated. Hence, studies aimed at
optimizing specific properties like toughness must be performed carefully to avoid
confusing results from the other factors.
Many previous studies have optimized mechanical properties of the heavy alloys through
either rapid quenching or slow cooling from temperatures above 1000°C.
Some researches gave specific attention to cooling rate effects and increased tensile
elongations obtained with slower cooling rates. The proposed explanations for the cooling
rate sensitivity include intermetallic phase formation, matrix phase saturation, hydrogen
embrittlement, altered ductile-brittle transition temperature, and impurity
Most likely each of these proposed processes can contribute to the embrittlement. How
dominance shifts with alloy composition, material purity, and material processing is
unclear, however. In wrought tungsten, brittle intergranular failure is commonly
associated with impurity segregation. Similarly, segregation of impurities is a possible
cause of embrittlement in the heavy alloys as well.
It is probable that toughness variations associated with heavy alloys represent several
effects. The obvious contradictions among investigations cannot be resolved without
greater experimental detail. The purpose of this study was to determine the cooling
rate effect on toughness of the-95 W-3.5 Ni-1.5 Fe alloy. Past experience on this alloy
demonstrated considerable heat-to-heat variation in toughness. Hence, post sintering
anneals up to 20 hours at temperatures of 1000°C with an air cool are used to minimize
the variations. In this condition, the ductility and toughness are improved.
Material for this investigation, 95 W-3.5 Ni-1.5 Fe, was fabricated from blended
elemental powders. The tungsten was minimum 99.9% pure with a Fisher subsieve size
between 3 to 4 μm, and a mean sedimentation size of approximately 7μm. Both the
nickel and iron powders were carbonyl types (INCO and GAF, respectively) with minimum
purities of 99.5%, and an average size less than 10μm.
The powders were blended for 30 minutes without lubricant or binders and loaded into
polyvinyl chloride bags. The bags were evacuated, sealed, and cold isostatically pressed
at 200 MPa. The compacts were induction sintered in the liquid phase at
1470±5°C for two hours in a dynamic hydrogen atmosphere with a subsequent
solid state 1350°C, 0.5 hour vacuum anneal followed by an air cool from 1000°C.
The resulting material had a density of 18.15 g/cm3 (≈99.9% of
theoretical), a total impurity content of less than 500 ppm by weight, and mean tungsten
grain size of 43±16 μm.
Nominal mechanical properties for 95W-3.5Ni-1.5Fe heavy alloy:
• Yield strength
• Ultimate tensile strength
• Elastic modulus
• Reduction in area
• Charpy impact energy
The ductile to brittle transition with decreasing test temperature has previously been
noted for the heavy alloys. The tungsten phase is more temperature dependent, and hence
there is a shift to tungsten cleavage at lower temperatures. Additionally, the heavy
alloys have more tungsten-tungsten interfacial area and less matrix phase (which acts
to arrest crack growth) as the tungsten content increases. Thus, the 95 W-3,5 Ni-1.5 Fe
is more sensitive to test temperature than the 90 W-5 Ni-5 Fe, 90 W-7 Ni-3 Fe, and
85 W-10.5 Ni-4.5 Fe alloys. Hence, the observed test temperature effect on impact energy
is attributed to the lower matrix phase content and larger interfacial area found with
the 95 W alloy.
In the absence of other changes, it would be expected that decreases in hardness in
simple systems would be associated with increases in toughness. Thus, since the micro
hardness changes are small, they indicate that mechanical properties of the matrix are
not a factor in the toughness variations with cooling rate.
The cooling rate effect on toughness is attributed to interfacial segregation; rapid
cooling from a post-sintering anneal resulted in improved toughness. Several possible
explanations exist for the toughness sensitivity to cooling rate. These include impurity
segregation to interfaces, compositional and heat treatment effects on the matrix phase
and tungsten grain chemistries, hydrogen embrittlement of the matrix phase, formation
of intermetallic compounds, changes in the defect (pore) structure, and a ductile-brittle
transition temperature close to room temperature.
In wrought tungsten there is a strong impurity effect on ductility. The segregation of
impurities to interfacial areas on slow cooling would be more detrimental to toughness as
the matrix content is decreased. Thus, the 95 W alloy would be expected to be more
sensitive to cooling rate than the lower tungsten content alloys.
From these findings it is concluded that impurities are responsible for the observed
toughness variations with cooling rate in 95 W-3.5 Ni-1.5 Fe. Microstructural features
are essentially unchanged by the differing heat treatments. Furthermore, variables such
as composition, hardness, and density do not explain the ductile-brittle toughness
transitions with test temperature and cooling rate. Past suggestions of intermetallic
formation and matrix phase aging are rejected for this system.
In the 95 W alloy there is a large amount of interfacial area. The tungsten-tungsten
grain boundaries are known to be embrittled by impurities. In the present case the role
of impurities is very strong. Slow cooling promotes interfacial segregation of impurities;
thus, the fracture path is predominately along the tungsten-tungsten and tungsten-matrix
boundaries. The impurity content correlates with the impact energies, showing the
detrimental role of impurities on toughness. Thus, the 95 W alloy exhibits the highest
toughness when rapidly cooled from a homogenization temperature of approximately
1000°C. On the other hand, slow cooling gives a decreasing impurity solubility
coupled to a high diffusive mobility.
Consequently, the material is embrittled by impurity segregation to interfacial
boundaries. Past conflicting reports concerning the cooling rate effect are probably due
in part to different impurity contents. Based on these findings, it is probable that high
purity heavy alloys will exhibit high toughness and less sensitivity to cooling rate.
However, the sensitivity to test temperature as demonstrated in this study cannot be
totally eliminated through use of higher purity material. The ductile-brittle transition
with test temperature is due to the differing flow stress and ductility dependencies on
temperature for the two alloy components. Hence, lower toughness is expected at lower