Intermetallic alloys have generated increasing interest over the past decade as possible replacements for titanium- and nickel-based superalloys currently used in aerospace applications. Attractive properties include higher melting temperatures and lower densities which provide improved specific creep strength for elevated temperature structures in high performance engines. However, as most intermetallic compounds are brittle due to their complex crystal structure, they generally suffer from poor room temperature fracture resistance. To improve their low intrinsic toughness, extrinsic toughening techniques that invoke crack-tip shielding mechanisms are often used in alloy design and microstructural development. Such mechanisms, which include crack bridging via ductile or brittle reinforcements, primarily act behind the crack tip and locally screen the crack from the applied far field driving force. Examples of materials toughened in this manner are various ceramic- and intermetallic-matrix composites such as Co/WC, Al/Al2O3, Nb/MoSi2, Al or Mg/glass, Nb or TiNb/TiAl, Mo or Cr/NiAl, and Nb/Nb3Al, that incorporate ductile (or brittle) reinforcements in the form of particulates, fibers, or laminates. Despite the efforts in improving the toughness of brittle materials using ductile metal reinforcements, work in several intermetallic systems has shown that the toughening induced by the ductile phase can be severely degraded under cyclic loading which will likely be present in many structural applications.
In the present study, we examine the fracture and fatigue behavior in a model Nb-reinforced Nb3Al composite, where toughening is achieved by the addition of ductile Nb layers having layer dimensions in the tens to hundreds of microns. Focus is placed on the effect of reinforcement morphology, laminate orientation (Figure 1), and reinforcement layer thickness. Results are compared with earlier studies which involved Nb/ Nb3Al composites with Nb as ~20 micron particulate reinforcement formed in situ by powder metallurgy, or as 2 micron thick (magnetron sputtered) layers to form a microlaminate.
The R-curves, plotted in Figure 2 for the divider orientation, illustrate that the addition of high aspect ratio ductile reinforcements leads to significant improvements in the fracture behavior of Nb3Al. Compared with the intrinsic toughness of only 1 MPa(m)^0.5 for monolithic Nb3Al, the addition of 20 vol% of 50 micron thick Nb layers resulted an initiation toughness of ~5 MPa(m)^0.5 m and a maximum toughness (on the R-curve) exceeding 10 MPa(m)^0.5. The 125 and 250 micron Nb laminates yielded even better properties having initiation toughnesses of 6.7 and 7.1 MPa(m)^0.5, respectively, and corresponding maximum toughnesses over 15 and 20 MPa(m)^0.5 m. The high toughness achieved by these laminates is a result of crack tunneling in the intermetallic phase between the Nb layers. Cracks tunneled ahead almost the entire crack growth range leaving extensive bridging zones behind the crack tip to shield the far-field applied stress intensity.
Note that the present laminates contain only 20 vol% Nb, yet they exhibit higher toughness than Nb particulate toughened Nb3Al composites containing twice as much Nb (Fig. 2b). Moreover, they displayed comparable or superior toughness to a microlaminate consisting of 50 vol% of 2 micron thick Nb layers in Nb3Al having initiation and steady-state toughnesses of only ~6 and 10 MPa(m)^0.5, respectively. For the latter material, quasi-static crack extension occurred only over the first 200 microns of growth; in contrast, the present laminates exhibited stable cracking over crack extensions of 3 mm or more. It is clear that the laminates with coarser Nb layers require far less reinforcement to achieve equivalent or superior fracture toughness properties; indeed, compared with particulate or microlaminate reinforcement, the crack initiation toughness, the maximum crack growth toughness, and the extent of subcritical cracking prior to instability are generally all enhanced.
Consistent with prior studies, the initiation toughness did steadily increase with layer thickness for the arrester orientation (FIGURE 3) with values of approximately 7.3, 9, and 10.4 MPa(m)^0.5 for the 50, 125, and 250 micron thick Nb layers, respectively. The maximum toughness was ~20 MPa(m)^0.5 or more for all layer thicknesses and was generally higher than that of the divider laminates for a given reinforcement layer thickness. Extensive bridging was also responsible for the significant improvement in maximum toughness of the arrester laminates. The 50 and 125 micron Nb laminate appeared to just reach equilibrium bridging zone sizes of ~3 and 5 mm respectively, but the 250 micron Nb laminate crack showed no sign of an equilibrium zone length and remained fully bridged.
Large scale bridging conditions - when the bridging zone is large relative the crack length and the uncracked ligament - are expected to apply to the laminates in both orientations due the large bridging zones. Large scale bridging causes the R-curve to become geometry dependent, but modeling techniques do allow predictions of small scale bridging R-curve behavior to be extracted from the measured R-curves. Predicted values of ~17, 22, and 33 MPa(m)^0.5 were calculated for the 50, 125, and 250 micron Nb laminates from the constrained flow stresses obtained from the large scale bridging analysis.
It can be seen in FIGURES 4 and 5 that the addition of Nb as a laminated phase in either orientation dramatically improves the fatigue crack growth resistance of the brittle Nb3Al intermetallic by ~3-7 MPa(m)^0.5, depending on the composite layer thickness and orientation. It can also be seen (FIGURE 4a) that fatigue thresholds are improved by ~133% to 233% using the high aspect ratio reinforcing layers; this demonstrates that laminates are far more resistant to the cyclic loading induced degradation in shielding, compared to the Nb particulate reinforced Nb3Al composites examined earlier. It should again be noted that these improvements have been achieved using a reinforcement volume fraction which is half that of the particulate composite. Based upon these comparisons, laminate reinforcements are clearly the best choice for improving the fatigue crack growth properties of brittle materials when forming composites by the incorporation of a ductile second phase.
The arrester orientation showed a fatigue threshold of ~6.5 MPa(m)^0.5 that was essentially independent of laminate thickness and showed a more consistent improvement compared to the divider thresholds which varied with Nb layer thickness. However, the slope of the fatigue curve in the arrester laminates decreased with increasing Nb layer thickness implying improved fatigue crack growth resistance. The arrester orientation shows a fatigue threshold for all layer thicknesses was which is very similar to that observed for the Nb metal processed under similar conditions (FIGURE 5); however, the slope of the fatigue curves decreases with increasing layer thickness. This behavior appears to be closely related to the change in crack-layer interactions, with increasing applied stress intensity range. Metallograhic investigation revealed that at lower stress intensity range levels the crack impinges on the Nb reinforcing layer and proceeds to fatigue through the metal layer prior to renucleation in the Nb3Al intermetallic. As a result, bridging ligaments in the crack wake become evanescent, and the fatigue threshold of the arrester laminates appears to be dominated by the Nb metal fatigue properties. In contrast, higher applied stress intensity range cause crack renucleation in the Nb3Al across the Nb layer, and the subsequent development of a bridging zone in a specimen because the maximum applied stress intensity approaches or exceeds the monotonic crack initiation toughness. This bridging zone then acts to shield the crack tip leading to improved crack growth properties.
FIGURE 4a also shows that there is a significant improvement in the fatigue crack resistance of the divider orientation with increasing reinforcement layer thickness. Fatigue thresholds increased from ~3.5 to 5 to 7 MPa(m)^0.5 for the 50, 125, and 250 micron Nb laminates, respectively, for load ratio of 0.1. Higher mean stresses, achieved using a load ratio of 0.5, appear to degrade the cyclic crack growth resistance of these composites and reduce the thresholds by about 30% to 2.5, 3.5 and 5 MPa(m)^0.5, respectively. Replotting the fatigue data in terms of Kmax (FIGURE 4b) narrows the spread of the data with respect to mean stress variation which indicates that the monotonic fracture mechanism of the brittle Nb3Al intermetallic has a significant influence on the cyclic crack growth process. This is supported by fitting the fatigue curves to a modified Paris equation
da/dN ~ (Delta K)^n * Kmax^p
where the exponents for each layer thickness are given as follows: