The effect of reaction products on the fracture resistance of a metal/ceramic interface

Solid state diffusion bonds between a Ta(Ti) alloy and Al₂O₃ have been used to study the effect of reaction products on the fracture resistance of a metal/ceramic interface. The reaction layer consists of brittle intermetallics, consistent with the TiTaAl ternary phase diagram. The interface fracture energy Γ_i was found to be dominated by periodic tunnel cracking of the reaction products, and was relatively insensitive to layer thickness, because of the self-similar nature of the periodic cracking. The contribution to Γ_i from plastic dissipation within the alloy was minimal because the intervening reaction layer inhibited yielding. A sequence of cracking observations was used to bound the unknown values of the fracture energy of the reaction products, the residual stress within the layer and the fracture energy of the reaction product/sapphire interface.     

The fracture resistance of metal-ceramic interfaces

The effects of interface structure and microstructure on the fracture energy, Γ_i, of metal-ceramic interfaces are reviewed. Some systems exhibit a ductile fracture mechanism and others fail by brittle mechanisms. In the absence of either interphases or reaction products, Γ_i is dominated by plastic dissipation (for both fracture mechanisms), leading to important effects of metal thickness, h, and yield strength, σ₀. Additionally, Γ_i is larger when fracture occurs by ductile void growth (for the same h and σ₀). A fundamental understanding now exists for the ductile fracture mechanism. However, some basic issues remain to be understood when fracture occurs by brittle bond rupture, particularly with regard to the role of the work of adhesion, W_ad. Interphases and reaction products have been shown to have an important effect on Γ_i. A general trend found by experiment is that Γ_i scales with the fracture energy of the interphase itself, wherein Γ_i tends to increase for the interphase sequence: amorphous oxides > crystalline oxides > intermetallics. However, there also appear to be important effects of the residual stresses in the interface (which influence the fracture mechanism and the layer thickness.     

Effects of plasticity on the crack propagation resistance of a metal/ceramic interface

Fracture experiments have been conducted on a gold/sapphire interface. The interface is found to fail by interface separation in a nominally “brittle” manner with a critical strain energy release rate, G_c ≈ 50 Jm⁻², substantially larger than the work of adhesion, W_ad ≈ 0.5 Jm⁻². Evidence of plastic deformation on the gold fracture surface, such as blunting steps and slip steps, suggest that plastic dissipation is the primary contribution to the measured G_c. Calculations suggest that the majority effect occurs in the plastic zone through the crack wake. The interface is also found to be susceptible to slow crack growth.     

Fracture toughness and fatigue crack growth in rapidly quenched Nb-Cr-Ti In situ composites

In situ composites based on the Nb-Cr-Ti ternary system were processed by rapid solidification in order to reduce the size of the reinforcing intermetallic phase. Two-phase microstructures with small Cr₂Nb particles in a Nb(Cr, Ti) solid solution alloy matrix were produced for several compositions that previous work showed to produce high toughness composites in cast materials. The fracture and fatigue behaviors of these composites were characterized at ambient temperature. The results indicate that the fracture resistance increases with a decreasing volume of Cr₂Nb particles. Fracture toughnesses of the rapidly solidified materials with their smaller particle sizes were lower than for conventionally processed composites with larger particles of the intermetallic compound. The fatigue crack growth rate curves exhibit steep slopes and a low critical stress intensity factor at fracture. The lack of fracture and fatigue resistance is attributed to the contiguity of the intermetallic particles and the absence of plastic flow in the Nb solid solution matrix. The matrix alloy appears to be embrittled by (1) the rapid solidification processing that prevented plastic relaxation of residual stresses, (2) a high oxygen content, and (3) the constraint caused by the hard Cr₂Nb particles.     

Effects of chromium on crack growth and oxidation in nickel aluminide

The influence of chromium additions on crack growth and oxidation have been examined in the nickel aluminide, Ni₃Al. Crack growth rates were measured in a chromium containing alloy as a function of stress intensity at temperatures between 600 and 760°C in air, together with rates of oxide film growth, and compared with previous measurements taken from Ni₃Al. The mechanisms of crack propagation and oxidation were investigated with a range of analytical techniques, including SEM, AES, XPS, SIMS, TEM and STEM. An addition of 8% chromium had a significantly beneficial effect on both crack growth resistance and oxidation resistance between 600 and 760°C. Low oxidation rates were associated with the formation of Cr₂O₃ together with Al₂O₃ at the metal/oxide interface, consistent with chromium acting initially as an oxygen getter, and promoting the formation of a protective Al₂O₃ layer, with little internal oxidation. It is proposed that chromium was also responsible for inhibiting oxygen access to and diffusion along grain boundaries at crack tips, modifying the mechanism of crack propagation from “step-wise cracking”, dominated by oxygen embrittlement (observed in the absence of chromium), to a more conventional creep crack growth process.     

Deformation Behavior and Damage Evaluation in a New Titanium Diboride (TiB2) Steel‐Based Composite

Deformation behavior and damage evaluation of a new composite steel has been investigated by means of in situ three‐point bend tests in the scanning electron microscope. The titanium diboride (TiB₂)‐reinforced steel composite is produced by in situ precipitation of the TiB₂ particles during eutectic solidification. This production process developed by ArcelorMittal leads to a steel composite with a significant increase in specific stiffness (>20%), and good strength/ductility compromise. The microstructures obtained consist of primary TiB₂ crystals surrounded by a eutectic mixture of ferrite and TiB₂ particles. The primary mode of damage is particle fracture and inhomogeneous plastic deformation in the matrix. In contrast with other production process, particle fracture was more common than interfacial debonding indicating that interfacial strength is not the limiting factor in damage accumulation and fracture in this composite. Crack growth occurred by particle fracture ahead of the crack tip, producing large microvoids, which then link up to the growing crack by ductile failure of the remaining matrix ligaments. The results suggest also that the cracks tended to avoid direct particle interactions.     

Effects of temperature and environment on fatigue crack growth in ordered (Fe,Ni)3 V-type alloys

The effects of long range order, test temperature, and test environment on fatigue crack growth of two (Fe, Ni)₃ V-type ordered alloys have been determined. Long range order suppressed crack growth at low and intermediate δK, but had less effect at high δK. Crack growth resistance of the ordered alloys decreased at 600 ‡C, but still compared favorably to that of several commercial high temperature alloys. Crack growth resistance of ordered material was decreased in the presence of hydrogen (precharged or gaseous), accompanied by a change in fracture path from transgranular to intergranular. Disordered material was nearly unaffected by hydrogen. The effects of long range order and hydrogen exposure on crack growth rates are discussed in terms of the characteristic superlattice dislocations in ordered alloys.     

The effect of interfacial reaction layer thickness on fracture of titanium-SiC particulate composites

Composites of commercial-purity Ti reinforced with 10 vol pct of SiC particles have been produced by cospraying and by powder blending and extrusion. Interfacial reaction layers have been studied by electron and optical microscopy and by Auger electron spectroscopy (AES) of fracture surfaces. The work of fracture has been measured as a function of reaction layer thickness for extruded and heat-treated composites. Material with very thin reaction layers (<∼0.1 μn) can be produced by cospraying, but porosity levels are relatively high (∼5 to 10 pct). Extruded material has been produced with a thin reaction layer (∼0.2 μm) and low porosity (<1 pct). It appears that the rate of reaction conforms with published parabolic rate constant data over a wide range of time and temperature. The reaction layer always consists of TiC and Ti₅Si₃, but the TiC grains tend to be larger than those of Ti₅Si₃. As the reaction layer thickness becomes greater than about 1 μm, the work of fracture falls sharply and the cracking pattern changes from one involving fracture of SiC particles to one in which cracking between the particles and adjacent reaction zones becomes predominant. It is suggested that the volume contraction accompanying this reaction, calculated at about 4.6 pct from density data, has a significant effect in promoting crack formation in these locations by generating radial tensile stresses across the interface. Thus, for this particular composite system, the important effect of a thicker reaction layer may be that it promotes the formation of an interfacial crackvia an effect on the local stress state, rather than itself constituting a larger flaw in the form of a through-thickness crack assumed to be present.     

Stress-assisted hydrogen attack cracking in 2.25Cr-1Mo steels at elevated temperatures

Crack growth in 2.25Cr-lMo steels exposed to 3000 psi hydrogen has been investigated in the temperature range 440 °C to 500 °C, using modified wedge-opening loaded specimens to vary stress intensity. Under conditions of temperature and hydrogen pressure, where general hydrogen attack does not occur, the crack propagated by the growth and coalescence of a high density of methane bubbles on grain boundaries, driven by the synergistic influence of internal methane pressure and applied stress. Crack growth rates were measured in base metal, and the heat-affected zones (HAZs) of welds were tempered to different strength levels. The crack growth rate increased with material strength. Above a threshold of about K_l = 20 MPa√m (at 480 °C), the crack growth rate increased rapidly with stress intensity, increasing as roughly K_l ^6.5. Because of better creep resistance, stronger materials can sustain higher levels of stress intensity to drive crack growth and nucleate the high density of voids necessary for crack growth. Stress relaxation by creep reduces the stress intensity, and thus the growth rate, especially in weaker materials. The crack growth rate in the heat-affected zone was found to be substantially faster than in the base metal of the welds. Analysis indicates that K_l rather than C* is the appropriate crack-tip loading parameter in the specimen used here and in a thick-walled pressure vessel. The DC potential drop technique met with limited success in this application due to the spatially discontinuous manner of crack growth and limited crack-tip opening displacement. Formerly Graduate Student, Materials Science and Engineering Department, The Ohio State University     

The fabrication and failure of laminar ceramic composites

Rather than using fibres or other dense inclusions as a means of introducing weak interfaces to deflect cracks, a much simpler route has been developed where composite elements, in this case sheets, are made from silicon carbide powder. After coating, these sheets are stacked, compacted together and sintered without any pressure. It is shown that the graphite layers will deflect cracks preventing catastrophic failure whilst raising the apparent fracture toughness from 3.6 MPa✓m to 17.7 MPa✓m and the work required to break the sample from 28 J m⁻². Further crack growth through the sample occurs when the unbroken part reaches is failure strees as determined from an unnothched beam. This allows the apparent fracture toughness to be simply related to the failure stress and this has been confirmed by experiments where the strength of the beam has been controlled by adding artificial flaws. Crack growth is not much affected by the thickness of the interface except at low thicknesses where small gaps in the interface are more likely to occur, allowing the delamination crack to kink out of the interface and cross the next lamina. The thermal stability of the laminate up to 1500°C in air is limited only by the oxidation resistance of the graphite. The potential for the process is also discussed.     

Effect of yttrium on the oxidation behavior of cast Ni-30Cr alloy

Cast Ni-30Cr and Ni-30Cr-0.5Y alloys were oxidized at 1000°C in pure O₂ for various times, then were either furnace cooled to room temperature, or thermally cycled between 1000°C and different lower temperatures. The isothermal oxidation rate of the Ni-30Cr alloy was reduced by about a factor of 3.6 by the addition of 0.5% Y. Acoustic emission signals, which are generated by scale fracture events, were collected during isothermal oxidation, during continuous furnace cooling and during thermal cycling. These data showed, as others have shown, that the scale formed on Ni-30Cr-0.5Y was significantly more resistant to fracture than that on Ni-30Cr. This advantage of the Y-containing alloy was evident for comparisons based on equal oxidation times, and more importantly, at equal scale thicknesses. SEM and EDAX analyses show that continuous Cr₂O₃ scales were formed on both Y-bearing and Y-free alloys after a short time of oxidation (2 h), but after a longer period of oxidation and thermal cycling, a NiO or NiCr₂O₄ outer layer was found. This outer scale created a new interface with the Cr₂O₃ scale where thermal stresses will be generated during cooling due to the thermal expansion difference between Cr₂O₃ and NiO or NiCr₂O₄. Spallation at the inner scale/outer scale interface, as well as at the metal/scale interface, was observed. X-ray measurements of scale strains at equal scale thicknesses showed that the growth strains (at the end of the isothermal oxidation period) were larger on the Y-containing alloy, and that this alloy also sustained larger residual strains upon cooling to room temperature. Using a model based on elastic strain energy, estimates of the surface energy for scale fracture (a measure of scale adhesion) were significantly higher for the Y-containing alloy at equal scale thicknesses. Both the AE and the strain measurements are consistent with the proposal that Y improves the inherent strength of the metal/scale interface. The smaller rate of scale cracking for Y-containing alloys, combined with their slower scale growth rate, offers the further benefit of delaying the onset of NiO or NiCr₂O₄ overgrowth layers, which themselves may degrade the integrity of the scale.     

An analysis of slag stratification in nickel laterite smelting furnaces due to composition and temperature gradients

The density of FeO-MgO-SiO₂ and FeO-Fe₂O₃-SiO₂ based slags has been analyzed in terms of smelting of lateritic ores for the production of ferronickel. The density of these slags decreases with increasing MgO, SiO₂, and Fe₂O₃ contents as well as with increasing temperature. During the electric furnace smelting of calcined and prereduced garnieritic ores, the slag temperature decreases from the upper layer down toward the slag/metal interface. Together with precipitation of either olivine or silica, this leads to the formation of a dense and stagnant slag layer at the slag/metal interface. For limonitic ores, the use of deep electrode immersion and high currents leads to slag reduction and increased slag temperatures toward the bottom part of the slag layer. The reduction of Fe₂O₃ to FeO increases the slag density. In this manner, it may be possible to maintain a hot slag layer in the region of the slag/metal interface, without buoyancy-induced flow.     

Intermetallic compound layer growth at the interface of solid refractory metals molybdenum and niobium with molten aluminum

The growth mechanisms and growth kinetics of intermetallic phases formed between the solid refractory metals Mo and Nb and molten aluminum have been studied for contact times ranging from 1 to 180 minutes at various temperatures in the range from 700 to 1100°C. The growth of the layers of the resulting intermetallic phases has been investigated under static conditions in a saturated melt and under dynamic conditions using forced convection in unsatured aluminum melts. The Nb/Al interfacial microstructure consisted of a single intermetallic phase layer, Al₃Nb, whereas two to four different phase layers were observed in the Mo/Al interface region, depending upon the operating temperature. It was found that, in a satured melt, the intermetallic phase growth process was diffusion-controlled. The parabolic growth constants of the first and second kind and integral values of the chemical diffusion coefficients over the widths of the phases were calculated for both Mo/Al and Nb/Al systems. It also was found that the AlNb₂ phase grew between the Nb and Al₃Nb phases after consumption of the saturated Al phase. Similarly, the AlMo₃ phase grew between the Mo and Al₈Mo₃ phases with diminishing of all the other existing compound phases. In an unsaturated melt, the intermetallic phase layer grows at the solid surface while, simultaneously, dissolution occurs at the solid/liquid interface. This behavior is compared to the growth mechanisms proposed in existing theories, taking into consideration that interaction occurs between neighboring phases. It was found that the intermetallic phase, Al₈Mo₃, adjoining the base metal, was not bonded strongly to the base metal Mo and was brittle; its hardness also was larger than that of the layer near the adhering aluminum and the adjacent phases.     

Fracture mechanics of Ti/Al2O3 interfaces

Measurement of mixed mode interfacial fracture toughness in Ti/Al₂O₃ bimaterial couples has been accomplished by four-point bending tests of sandwich specimens with symmetrical cracks for different processing temperatures and thicknesses of Ti interlayers. Fracture surfaces and sample cross sections were analyzed by SEM, XPS, EDAX, EPMA and AES. The interfacial fracture toughness measured with four-point bending tests increases with increasing applied bonding temperature up to 950°C. Above this temperature toughness decreases. The deterioration of toughness is found to be due to an intermetallic phase (Ti₃Al) produced by diffusion at high temperature. The interfacial fracture toughness also increases when the Ti interlayer is thicker. The measured critical strain energy release rate (or interfacial fracture energy) ranges from 10 to 45 J/m². This is much larger than the estimated true work of adhesion which is about 2 J/m². This is because of plastic energy dissipation in the Ti interlayer during the fracture process.     

Subcritical crack growth at bimaterial interfaces: Part II. microstructural effects on fracture resistance of metal/ceramic interfaces

The flexural peel technique was used to study the fracture resistance of two model A1₂O₃/A1 interfaces. The bimaterial interface was formed by bonding high-purity A1₂O₃ with molten Al-5 pct Cu alloy under pressure. The specimens were then heat treated so that the Al-Cu alloy reached peakaged and extended-overaged conditions. The fracture resistance curve was established for two interfaces with either the peak-aged or overaged Al alloy. The fracture resistance of the interface with the peak-aged Al-Cu alloy was higher in terms of both the initiation and peak toughness. While the peak toughness of the interface scaled with the yield strength of the metal, the initiation toughness differed by a factor of 8. The difference in the initiation toughness is discussed in terms of the disparity in the interfacial microstructure.     

An investigation of the plastic fracture of AISI 4340 and 18 Nickel-200 grade maraging steels

The mechanisms of plastic fracture (dimpled rupture) in high-purity and commercial 18 Ni, 200 grade maraging steels and quenched and tempered AISI 4340 steels have been studied. Plastic fracture takes place in the maraging alloys through void initiation by fracture of titanium carbo-nitride inclusions and the growth of these voids until impingement results in coalescence and final fracture. The fracture of AISI 4340 steel at a yield strength of 200 ksi (1378 MN/mm²) occurs by nucleation and subsequent growth of voids formed by fracture of the interface between manganese sulfide inclusions and the matrix. The growth of these inclusion-nucleated voids is interrupted long before coalescence by impingement, by the formation of void sheets which connect neighboring sulfide-nucleated voids. These sheets are composed of small voids nucleated by the cementite precipitates in the quenched and tempered structures. The sizes of non-metallic inclusions are an important aspect of the fracture resistance of these alloys since the investigation demonstrates that void nuclea-tion occurs more readily at the larger inclusions and that void growth also proceeds more rapidly from the larger inclusions. Using both notched and smooth round tensile specimens, it was demonstrated that the level of tensile stress triaxiality does not effect the void nu-cleation process in these alloys but that increased levels of triaxial tension do result in greatly increased rates of void growth and a concomitant reduction in the resistance to plastic fracture.     

Intergranular fracture in an Al-15 Wt Pct Zn alloy

Crack propagation rates for the intergranular fracture of an Al-15 wt pct Zn alloy tested in air, distilled H₂O, and 0.5M NaCl have been studied using a double cantilever beam specimen. This technique has been applied to two different types of specimens: polycrystals with large equiaxed grains and bicrystals. The susceptibility to intergranular fracture in air and 0.5M NaCl increases as the volume fraction of G.P. zones in the matrix increases. The microstructure which is most susceptible to fracture exhibits intense planar slip traces while the more resistant microstructures have a more diffuse surface slip pattern. The dependence of cracking rates on the microstructure is explained in terms of the blocking of inhomogeneous plastic flow in the matrix by grain boundaries, resulting in locally severe stress concentrations at the boundaries. Bicrystal tests substantiate these results and show that orientations favoring partial continuity of slip bands across grain boundaries are less susceptible to stress corrosion cracking than arbitrarily oriented boundaries with no slip continuity. WILLIAM J. KOVACS, formerly Graduate Student, Carnegie-Mellon University, Pittsburgh, Pa. This paper is based upon a thesis submitted by WILLIAM J. KOVACS in partial fulfillment of the requirements of the degree of Doctor of Philosophy in Metallurgy and Materials Science at Carnegie-Mellon University.     

The Effect of Second Tempering on Hydrogen Embrittlement of Ultra-High-Strength Steel

The objective of the present study is to enhance the resistance of hydrogen embrittlement (HE) via second tempering at 250 °C for 30, 60 and 120 minutes. Although second tempering results in a higher saturated hydrogen content for the second tempering specimens during a slow strain rate test (SSRT), it effectively reduces HE susceptibility. As the second tempering time increases, dislocation density decreases. In contrast, the size of the cementite and Mo_yC_x precipitates increase slightly. The density of Mo_yC_x precipitates increases, whereas the density of cementite remains approximately the same as the second tempering time increases. Regarding second tempering specimens, the volume fraction’s increase in Mo_yC_x precipitates, which acts as a hydrogen trap with high binding energy, plays an important role in reducing the HE susceptibility, and the decrease in dislocation density can also improve HE resistance. In addition, the growth of the interface of the cementite and matrix disperses more hydrogen, which could enhance HE resistance. The result also reveals that the cementite and matrix interface is a type of low-binding-energy hydrogen trap without plastic deformation, whereas the strain interface with interfacial dislocations is a type of high binding energy hydrogen trap under plastic deformation.

     

Cyclic fatigue-crack propagation along ceramic/metal interfaces

The integrity of ceramic/metal joints is investigated under mechanically applied cyclic stresses using double-cantilever-beam, and compact-tension, sandwich test specimens. Specifically, fatigue-crack propagation rates for interfacial cracks are characterized over a range of velocities from 10⁻⁹ to 10⁻⁴m/s for glass/copper and alumina/aluminum-alloy interfaces tested in moist air. Compared to corresponding (stress-corrosion) results under sustained loading, it is found that true interfacial cracks in glass-copper joints are significantly accelerated under cyclic loads. In addition, crack extension force (G) thresholds for interfacial crack growth under cyclic loads are some 46% lower than under sustained loads and are typically over six times lower than the interface toughness (G_c). For the alumina/aluminum-alloy system, conversely, fracture never occurs in the interface; under monotonic loading cracking progresses near the interface in the ceramic layer whereas under cyclic loading failure may occur either in the ceramic or in the metal. Based on a comparison with fatigue-crack growth data in bulk alumina and bulk aluminum alloys, it is found that near interfacial crack-growth rates in the metal are much lower than those of the bulk ceramic and show a far higher dependency on the range of G than behavior in the bulk metal.     

Near-Neutral pH Stress Corrosion Cracking Susceptibility of Plastically Prestrained X70 Steel Weldment

The application of strain-based design for pipelines requires comprehensive understanding of the postyield mechanical behavior of materials. In this article, the impact of plastic prestrain on near-neutral pH stress corrosion cracking (SCC) susceptibility of welded X70 steel was investigated with a slow strain rate tensile (SSRT) test. Generally, plastic prestrain reduces the SCC resistance in various welded zones. The SCC susceptibility of the test materials can be put in the following order: heat-affected zone (HAZ) > weld metal (WM) > base metal (BM). Fractographic analysis indicates that there are two cracking modes, mode I and mode II, during SSRT tests. Mode I cracks propagate along the direction perpendicular to the maximum tensile stress, and mode II cracks lie in planes roughly parallel to the plane where the maximum shear exists. The SCC of the BM is governed by mode I cracking and fracture of the HAZ, and the WM is dominated by mode II cracking. Damage analysis shows that the detrimental impact of plastic prestrain on the residual SCC resistance cannot be evaluated with the linear superposition model. A plastic prestrain sensitivity, a material constant independent of plastic prestrain, is proposed to characterize the susceptibility of SCC resistance to plastic prestrain, and it increases with the SCC susceptibility of the steels. The enhanced SCC susceptibility caused by plastic prestrain may be related to an increase in yield strength. The correlation of the ratio of the reduction in area in NS4 solution to that in air (RA _SCC/RA _air) with the yield strength is microstructure dependent.