R. G. Davies

The mechanical properties of ordered alloys

Abstract It has been our aim to demonstrate that the mechanical behavior of alloys that form superlattices can be understood in terms of the change in dislocation configuration with the degree of order. Fully ordered materials deform by the movement at relatively low stresses of superlattice dislocations, which consist generally of closely spaced pairs of unit dislocations. Since these dislocations are constrained to move as a group in order to preserve the ordered arrangement of the lattice, cross-slip is hindered. Long-range order thus results in a high rate of strain hardening, and in many cases to brittle fracture. Since cross-slip is thermally activated, however, at elevated temperatures the rate of strain hardening decreases and ordered materials become more ductile. The degree of order in superlattice alloys can be altered by thermal treatment (except for alloys ordered to the melting point) or by departing from the stoichiometric composition. Increases in the separation of superlattice dislocations brought about by partial order can explain the peak in yield stress manifested by most superlattice alloys near the critical ordering temperature, and the minimum in strength observed at the stoichiometric composition at low test temperatures. Another type of deviation from perfect long-range order is the antiphase domain boundary, which contributes to the strength primarily in the early stages of isothermal ordering. While preparing this review it has become apparent that certain areas of study have not received sufficient attention. In particular, single crystals of superlattices with structure other than L12, and to a very limited extent L20, have not been studied. Also, it has not yet been established whether sources emit unit dislocation or superlattice dislocations; careful electron transmission studies might resolve this question, which has direct bearing on theories of yielding in solid solution alloys. Studies of creep and diffusion in superlattice alloys have demonstrated a large increase in activation energy with order. It may be anticipated therefore that other diffusion controlled processes such as recrystallization and grain growth may be retarded by order, although there is virtually no information on this subject in the literature. It is in the latter category of diffusion controlled properties that ordered alloys appear to reveal the most promise for commercial applications. While the strength of superlattice alloys at low temperatures generally is low compared to other solid solution alloys, ordered alloys demonstrate superior strength at elevated temperatures because dislocation climb is restricted. However, in order for this property to be exploited it is necessary to utilize alloys with a high critical ordering temperature. One of the major obstacles to more widespread use of inter-metallic compounds and other ordered alloys is their extreme brittleness at low temperatures. The origin of this brittleness is now understood to lie in the restriction of cross-slip by long-range order and/or in the segregation of interstitial impurities to grain boundaries. Future research should now be directed towards means of ameliorating brittleness, with techniques such as ternary alloying additions, or by appropriate thermal processing to produce very fine grain sizes while avoiding contamination by interstitials.

Slip character and the ductile to brittle transition of single-phase solids

Abstract A transition from brittle to ductile behaviour of single-phase polycrystals is shown experimentally to be a consequence of a change in slip character rather than of a change in yield stress. In the case of a b.c.c. Fe-Co alloy, deformation by planar glide induces brittleness, whereas the alloy is ductile if it deforms exclusively by wavy glide. For a series of copper-base solid solutions loaded in mercury, the degree of embrittlement increases with planarity of glide. The Cottrell-Petch equation can account for such behaviour provided the appropriate parameter, k y, is re-considered in terms of the number of operable slip systems and the propensity for cross slip.

Influence of long-range order upon strain hardening

Abstract Shear stress-shear strain curves of ordered and disordered Cu3-Au single crystals have been obtained over the temperature range 77 to 500°K. Analysis of the data indicates (i) the strain hardening rate in stage II, θII, for the disordered condition is essentially independent of temperature, and (ii) θII for the ordered state increases continuously from 77°K to a maximum at 350°K. It is proposed that the increment of strain hardening due to ordering in Cu3Au type alloys arises from a source exhaustion mechanism; dislocations from a source cannot move large distances before being held up at barriers formed by partial cross slip of superlattice dislocations. The rate of formation of these barriers is thermally activated which accounts for the increase in θII with temperature.

The effect of interstitial solutes on the twinning stress of b.c.c. metals

Abstract Flow stress measurements at −196°C on Fe-4.8 at. % Sn-C alloys demonstrate that the presence of interstitial carbon markedly increases the stress for deformation by twinning. The flow stress for an alloy with 0.09 at. % carbon is 12.6 Kg/mm2 higher than the flow stress for a carbon-free alloy. This behaviour, which is an exception to the rule that twinning stresses are relatively invariant to solute content, was anticipated because the normal twinning mode in ferrite does not shear all interstitial atoms to proper octahedral sites. In confirmation of this viewpoint, the experimental results were found to be in reasonable agreement with a theoretical estimate of the change in twinning stress that should result from this crystallographic restraint.

Microcracking in ferrous martensites

The relative susceptibility to plate microcracking of Fe-1.0 to 1.8 wt pct C alloys containing 0 to 5 wt pct Mn, 0 to 8 wt pct Cr, and 0 to 15 wt pct Ni has been investigated. In the binary Fe-C alloys, the crack area per unit volume of martensite increases with increasing carbon content in the range from 1.0 to 1.3 wt pct in agreement with previous work. However, above 1.4 wt pct C, the specific crack area decreases as the carbon content increases due to a decrease in martensite plate length which in turn arises mostly from a habit plane transition from {225}γ to {259}γ. Indeed, when results from all alloys are considered, it is found that the dominant variable affecting microcracking is the martensite plate length. A direct, but less important, influence of carbon content is also found but any effects resulting directly from other solutes (Ni, Cr, and Mn) are negligible. The impingement model for microcracking suggested by Marder, Benscoter, and Krauss8 is examined in further detail. In this way, a rationale is developed for our observations which are in accord with fracture behavior of macroscopic samples.

The temperature dependence of the flow stress of the γ′ phase based upon Ni3Al

The temperature dependence of the ordinary flow stress, the microplastic yield stress, and the transient creep responses ofγ′ have been studied with respect to the effect of alloying additions, slip line topography, and dislocation structure. The increase observed in the flow stress with increase in temperature may be attributed to a change in the mechanism controlling the flow stress. An exhaustion hardening process at low temperatures appears to be supplanted by a debris hardening process at high temperatures. This transition arises from an increased propensity for {100} slip as the temperature is raised. Solute additions affect the temperature dependence of the flow stress probably by altering the tendency for {100} slip.

Nickel-rich portion of the Ni-Al-Nb phase diagram

The solubility of Nb + Al in the γ solid solution decreases markedly with decreasing temperature; thus alloys can be prepared that are γ at 1200°C and yet contain 50 pct γ’ precipitate after aging at 800°C. Thermal stability of the γ’ precipitate is related to the lattice mismatch between the γ and γ’ phases; the smaller the mismatch the lower is the interfacial elastic energy and the more stable is the γ’. Upon aging certain alloys at 800°C a γ’ growth interface other than the normal (100)γll(100)γ’ is observed. The maximum solubility of the niobium in γ’ is ∼7 at. pct; the width of the γ’ field increases with increasing niobium content but it is essentially independent of temperature. Replacing aluminum by niobium in γ’ gives hardnesses of up to 400 Dpn.

Ordering reactions in unfamiliar systems

Metallurgical understanding of order-disorder transitions is founded largely upon observations of alloys comprised of common metals and familiar (simple) crystal structures. Historically, the principal assumptions have been that only first and second nearest neighbor atoms in the crystal interact and therefore that the requisite atom movements necessary to effect transformation are small. These assumptions have been useful since they account, at least qualitatively, for the behavior of most “classical” ordered alloys. Because of the increasing importance of unfamiliar substances in several technologies, we consider in this survey the degree to which the above concepts can be applied to less-familiar materials or crystal structures. Three types of systems are examined: 1) alloys of metallic elements, at least one constituent of which is in the less-common category; 2) interstitial ordering in dilute alloys; and 3) complex systems involving chalcogenides, oxides, and intercalation compounds. Primary emphasis is given to describing transitions and transformation mechanisms not encountered in common alloys, and to pointing out possible experimental advantages of metallurgically unfamiliar systems for studies of mechanisms and kinetics.

Influence of austenite and martensite strength on martensite morphology

The martensite morphology and austenite flow strength have been determined in a variety of ferrous alloys chosen so that the austenites were paramagnetic, ferromagnetic, substitutional strengthened, and interstitial strengthened. It is demonstrated that two of the most important variables in determining the habit plane (and thus morphology) of martensite in a given alloy are the resistances to dislocation motion in austenite and in ferrite (i. e., martensite). In the wide variety of alloys where martensite with a {259}γ habit plane was observed, the austenite flow strength atMs is greater than 30,000 psi. At lower austenite strengths, either {225}γ or {111}γ habit planes are found depending on the resistance to dislocation motion in ferrite. Thus, {225} martensites are not always found as part of the spectrum between {111} and {259} martensites but only in the cases (e. g., interstitial strengthening) where ferrite is preferentially strengthened relative to austenite. All of the observations are consistent with the idea that the habit plane observed in a given alloy is the one involving the minimum plastic work for the lattice invariant shear.

Austenite ferromagnetism and martensite morphology

The Curie temperature of the austenite, the martensite-start temperature, and martensite morphology have been determined in a series of nil-carbon Fe−Ni and Fe−Ni−Co alloys. For these alloys, austenite ferromagnetism aboveMs is a necessary, but not sufficient, condition for the formation of lenticular rather than packet martensite. In contrast to Fe−Ni alloys where lenticular martensite only forms below ≈O°C, some of the Fe−Ni−Co alloys transform to this structure at temperatures up to ≈200°C. The results support the hypothesis that the resistance of austenite to plastic deformation affects the habit plane and thus morphology of the martensite which forms.