Fracture and fatigue of gamma based titanium aluminide intermetallic alloys.
This thesis presents the room temperature mechanical properties of two y-based titanium
aluminides. The fracture toughness, fatigue crack growth resistance (FeGR), and tensile
properties of duplex, fully lamellar, and near fully lamellar microstructures have been
assessed. The inter relationships between fracture toughness, FeGR, and tensile results were
Two alloys of nominal composition Ti-45AI-2Mn-2Nb (45-2-2) and Ti-48AI-2Mn-2Nb (48-
2-2) were used for mechanical property evaluation. The ingots were produced by plasma arc
cold hearth (PACH) melting at the IRe in Materials for High Performance Applications. The
material was obtained in both the as-cast and isothermally forged conditions (the cast
microstructure was near-fully lamellar, whilst the forged microstructure was fine grained
equiaxed-y). Subsequent heat treatments were used to produce fully lamellar type
microstructures. The colony size and lamella thickness was varied by altering the duration at,
and cooling rate, from the solution heat treatment temperature. The effects of microstructure
(equiaxed-y versus fully lamellar), lamellar colony size, lamella spacing, specimen thickness,
loading rate, and precracking on the mechanical properties have been determined.
For fully lamellar microstructures the measured fracture toughness and FCGR is critically
dependent on the micromechanisms of fracture. Increases in both properties can be achieved
by a fully translamellar mechanism of fraCture at the crack tip. Substantial decreases are
observed if the colonies fail by an interlamellar mechanism. For example, the fracture
toughness can be halved, and the slope ("m-value" from the Paris relationship) of the FeGR
curve can increase from 5 to 40 if the fracture surface exhibits a high proportion of
interlamellar failure. It was found that the fracture toughness decreased by approximately 1
MPa-V'm per 10% increase in the amount of interlamellar failure ahead of the precrack.
It was found that increasing lamellar colony size from 800-4000 Jlm gave little difference in
fracture toughness but larger colony sizes could lead to steeper FCGR curves due to a high
proportion of interlamellar fracture at the crack tip. Increasing the cooling rate during heat
treatment of fully lamellar microstructures led to a slight increase in fracture toughness and
tensile strength (due to finer lamella spacing). However, increased cooling rate could also
lead to slightly steeper FeGR curves due apparently to smoother colony boundaries which
cause an increased proportion of intergranular failure.
Increasing the specimen thickness (for a fixed lamellar colony size of 800 Jlm) from 4 to 14
mm led to an increase in the fracture toughness, and the "initiation" stress intensity factor
during FCGR testing. The KQ fracture toughness increased from 9 to 19 MPa.vm with
increasing specimen size. Because of the large colony size relative to the specimen
dimensions, the number of grains sampled in the process zone beneath the sharp crack is
critical. Increasing the number of grains reduces the influence that the low energy
interlamellar fracture mechanism can have on the measured fracture toughness (hence leading
to an increased KQ value). Measurements of crack length during the tests confirmed that
significant stable crack extension (due to interlamellar failure) was delayed in the thicker
The effect of loading rate was significant on the fracture toughness values of fully lamellar
microstructures, but there was little effect for the as-forged duplex microstructure. For the
fully lamellar microstructure the mean fracture toughness increased from 13 to 18 MPa.vm
with loading rates of 0.0025 to 10 MPa.vm.s-1 (higher toughness values were achieved at both
the very low and very high ends of the loading rates studied). Whilst for the duplex
microstructure the mean fracture toughness only varied by 0.7 MPa.vm over the same range of
loading rates. A test loading rate of between 0.05 to 1.0 MPa.vm.s-1 should be used to obtain
conservative fracture toughness values.
It was found that the fracture toughness values were slightly higher (50% increase) for fully
lamellar microstructures if tested with just a sharp notch (i.e. no precrack), but much higher
for duplex microstructures (an increase of 300%). This indicates that the fully lamellar
microstructure is more resistant to a sharp defect. Additionally, the specimens must be
precracked in-order to obtain valid fracture toughness values.
The FCGR curves for the fully lamellar microstructures were non-linear, and this was related
to varying proportions of interlamellar failure at the crack tip as it propagates (the more
interlamellar failure the steeper the curve). Indeed, the m-value increased from 5 for a
translamellar, to 40 for a predominantly interlamellar failure mechanism at the crack tip.
Decreasing the colony size and/or increasing the testpiece thickness reduced the extent of this
variation (a direct result of more grains being sampled in the process zone). The slope of the
FeGR curve showed no clear variation with decreasing lamella spacing.
Tensile testing revealed that the duplex microstructure had a higher proof stress and fracture
strength compared to the fully lamellar microstructures. This is a consequence of the finer
grain size of this microstructure (50 pm and >500 pm respectively). DecreaSing the lamella
spacing leads to an increase in the strength of fully lamellar microstructures.
The results indicate that a fine grained fully lamellar microstructure, with a fine lamella
spacing, would give the best combination of fracture toughness and fatigue crack growth
resistance. A higher fracture strength is found for fine grained duplex microstructures,
however such microstructures have extremely poor fracture toughness and fatigue crack