A D V A N C E D M A T E R I A L S & P R O C E S S E S | S E P T E M B E R 2 0 2 2 2 4 composition and temperature are controlled well, but turbulence and air entrainment is overlooked. Thus, the numbers, areas, and thicknesses of bifilms will vary, and the conditions of freezing will orient and concentrate the bifilms by pushing them ahead of the advancing grains (grains cannot grow through the “air layer” of bifilms). In general, therefore, bifilms concentrate at grain boundaries, becoming generally intergranular in equiaxed solidification conditions. Fatigue resistance is generally studied on beautifully polished small samples of metal, and tested on machines for which stress, strain, and rate of cycling is highly controlled. Fatigue in real engineering structures is often rather different, especially in the size of the component and the size of defects which it can harbor. Thus, in real components, fatigue failure occurs in conjunction with pre-existing defects which are too large to be considered in laboratory conditions. Although wellknown statistical systems are used in an attempt to allow for the distribution of sizes of defects so that the laboratory results are as representative as possible of larger samples, this approach is limited by the substantial changes in the morphologies of bifilms in different casting conditions. A recent study[4] of a series of engineering components has illustrated that the percentage contribution of bifilms to the overall area of the fracture seems often to be well in excess of the contribution from the fatigue process. With each bifilm, the propagation of the crack is expected to occur across the whole extent of the bifilm, which is often a distance commensurate with the grain size. In contrast, the stepwise propagation of the fatigue component will occur only in those regions where a convenient bifilm is absent. Thus, some grains will exhibit fatigue striations; but others will be clear of striations, appearing to be some type of cleavage, often called quasi-cleavage, to mask ignorance of what it really is—because it will certainly not be cleavage. This author is suggesting that these regions are bifilms, hiding in plain sight. Furthermore, the total area of bifilms can be large, sometimes as high as 99%, with only 1% of the failure being due to fatigue[4]. In many so-called fatigue failures, it is hard to find any striations among the quasi-cleavage facets. The facets are bordered by ductile shear steps, in which the crack finds its new alignment on arrival at each new bifilm. CREEP There has been no work so far to demonstrate that creep is affected by bifilms. Even so, there exists a critical piece of evidence that offers an explanation for the excellent creep resistance of nickel superalloy single crystals, which otherwise lack a cogent explanation for their astonishing good properties[1]. The traditional superalloy polycrystalline casting suffers from bifilms trapped in randomly oriented grain boundaries as grains nucleate randomly in the liquid, then grow, pushing the bifilms ahead until the grains mutually impinge, trapping the bifilms. All this happens within the few minutes required for freezing. In contrast, the single crystal grows very slowly from its base, in an upward direction, once again pushing bifilms ahead. In this way the single crystal does a good cleaning exercise, pushing bifilms out of the top of the casting and so achieving a low density of bifilms in the body of the casting. The traditional explanation that the eliminated transverse grain boundaries were weak was true in the sense that their bifilm content rendered them weak; but grain boundaries themselves are now proven to be immensely strong, practically as strong as the matrix. It follows that if superalloy polycrystals could be cast with low bifilm contents, it would not be necessary to grow a single crystal to achieve the desired creep properties. However, of course, the single crystal might in any case be desired for quite different reasons such as increased resistance to certain vibratory modes. INVASIVE CORROSION: PITTING It is relevant to question: why does a metal surface, particularly when protected by a natural passive oxide, tend to corrode only in highly localized regions? In general, the answer is because a foreign inclusion is present. The question then becomes, why is the inclusion present? The real answer in most cases investigated so far, and perhaps true for all cases, is that a bifilm is present, which has provided the favorable substrate permitting the inclusion to form—recalling the high probability that inclusions form only on bifilms. The bifilm forms a natural pathway for ingress of a corrodent, and the various inclusions and second phases that decorate the bifilm will aid the corrosive attack because of their different chemical potentials. Once again, it is necessary to observe that the bifilm is present first, and the etch pit merely constitutes the localized corrosion hollow at its emergence at the surface. The etch pit is therefore the witness to the prior presence of a bifilm crack and its emergence at this precise location (Fig. 3). This conclusion is diametrically opposed to the traditional view, used, erroneously in the view of this author, to explain a number of fatigue failures of helicopter drive shafts in which the fatal fatigue crack was assumed to be Fig. 3 — A steel blade from a steam turbine exhibits (preexisting) bifilm cracks that have led to the development of an etch pit where the bifilms have met the surface[2].
RkJQdWJsaXNoZXIy MTYyMzk3NQ==