Search Documents

By F. G. Smith

A FEW years ago, the writer encountered a, problem that, at first, seemed to be due to peculiar conditions affecting grain growth. Large cups made from heavy metal failed in the first drawing operation in the process of making seamless tubes. The metal, as cast, appeared to be sound and strong; it rolled successfully and withstood the first cupping operation, but on the first draw it failed, with a crystalline fracture. The failure appeared in the slightly worked part of the bottom of the cup, and microscopic examination revealed exceptionally large grains. Knowing the general conditions governing recrystallization of brass, which were later published by Matthewson and Phillips, an investigation was outlined with the idea of including various conditions under which new grains might form, and which would include the effect that a previous anneal might have upon the anneal following. These conditions were produced by annealing specimens of brass at five different temperatures, indenting them with a 10-inn. ball under four different pressures, and then reannealing at each of the five temperatures five specimens each of which had been annealed at one of the five temperatures.

Jan 8, 1919

By M. L. Holzworth, Joseph C. Kremer, Paul A. Beck, L. J. Demer

FOR alloys which are in practice heat treated to obtain increased strength, such as steels, duralumin, copper-beryllium, and others, the treatment usually involves heating to a relatively high temperature. At such temperatures the grain size of the product is largely determined by grain growth. Even in metals and alloys annealed merely to relieve work hardening, recrystallization is generally followed by a certain amount of grain growth. Thus, effective grain size control, which is often of great practical importance in manufacturing processes, depends on the grain growth behavior of the material used. Despite its importance, grain growth has been the subject of few investigations of a fundamental nature. The term "grain growth" has been used to designate a group of related, but distinctly different, phenomena. The two most important and best known of this group are illustrated by the following example. Curve a in Fig r shows the increase of the average grain size of a copper-beryllium alloy, containing 2 pct Be and 0.15 pct Co, with increasing annealing temperature. The time of annealing was 2 hrs. The grain size was quite uniform in all specimens. The results of a similar experiment with an alloy containing 2 pct Be and 0.31 pct Co are shown by curve b. It is seen that for any annealing temperature below T. the grain size was smaller than in the previous case. Grain growth became slower with the higher Co content. This "inhibiting" effect of the higher Co addition was associated with the appearance in the microstructure of fine, light blue particles of the CoBe phase. Up to temperature T. the structure of the inhibited alloy (curve b) was just as uniform as the structures represented by curve a. However, after annealing at T, or at higher temperatures, the inhibited alloy showed some extremely large grains (much larger than the grain sizes developed by the uninhibited alloy with low Co content) mixed with the small grains according to curve b. The presence of these mixed or "duplex" structures is indicated in Fig 11 by the vertical lines branching off curve b. The type of grain growth occurring in uninhibited Cu-Be alloys (curve a) has has been most frequently and best investigated in brass. This type has been designated as "normal grain growth," while the grain growth phenomena occurring in inhibited materials were, at times, referred to as "abnormal grain growth." However, since the latter term has been applied occasionally to the phenomenon of critical recrystallization, which also results in large grains although by an entirely different mechanism, in the present paper a different nomenclature has been adopted in order to avoid confusion. Typical of the grain growth in uninhibited metals and alloys is the continuous

Jan 1, 1947

By Floyd Kelley

THE literature of the last decade is rich with information relating to the cause and means of control of grain growth in pure metals, but is deficient concerning the role diffusion plays in grain growth, although during this period many new products that depend on diffusion for their properties and success have been developed. Among these are cementation products, produced by calorizing, chromizing, and carbonizing, or by mixing, pressing, and sintering powdered metals such as Syke's iron-molybdenum die material, Genelite, and Elcon Metal (copper, tungsten) etc. This is a fertile field for new developments, but before it can be util-ized to the fullest extent there must be a thorough investigation of the processes by which diffusion and grain growth take place. Grain growth plays no small part in the process and certainly the physical properties of these products depend largely on the grain size and state of equilibrium produced. The field of investigation does not end here, but may be extended to the effect of diffusion on impurities, segregations and dendrites in the broad field of ferrous and non-ferrous metallurgy. A number of years ago the author investigated several of these cases and noted, with striking regularity, the pronounced effect of diffusion in promoting recrystallization and grain growth. An example of this phenomenon is given in his paper on chromizing1 which contained photomicrographs of large radial or columnar crystals produced by diffusion of chromium in the solid state. Another manifestation of this effect was discussed by Austen,2 who showed the part hydrogen plays in the decarburization of steel and the effect of diffusion upon the grain structure.

Jan 1, 1928

By M. L. Samuels

DURING the period from 1910 to 1920, there was a lively interest in the subject of grain growth and many papers were published, followed by interesting discussions. Questions dealing with the fundamentals of grain growth-why and how it occurs-were treated in papers by Howe,1 Jeffries,2 Sauveur,3 Chappell,4 Beilby,5 Mathewson and Phillips,6 McAdam,7 Carpenter and Elam,8 Stead and Carpenter,9 Ruder,10 as well as others. Since that decade, much indeed has been done regarding the practical control of grain size, and it is now possible to order material having a size specified in the American Society for Testing Materials classification. Also, one hears the term "spliced boundaries" used in connection with certain tungsten lamp filaments" wherein it is important to prevent "sagging" between consecutive turns of the helix." This practical use of a certain knowledge as to the things that influence or even control grain growth should not be taken as an indication that the subject has been thoroughly explored and that nothing more can be learned from fundamental studies. It is with this thought in mind that a report of some observations and experiments on abnormal grain growth is believed justified. Abnormal grain growth is met with in heat-treating high-carbon steels in the austenitic range,13 in the hardening treatments of high-speed steels, 14.11 and in the low-temperature annealing of strained low-carbon steels as in sheet and wire material. Entirely aside from these instances of exaggerated grain growth, which some might consider as freakish, it is thought that abnormal growth is simply a case where the growth forces and the forces tending to prevent growth are almost, but not quite, evenly balanced. Information relative to grain-growth characteristics in general may be obtained from a study of abnormal growth conditions, which may be looked upon as vantage points. Reflection will show at once that if growth forces are great in comparison to inertia or resistance to growth, many crystals start to grow and, for that reason, none becomes very large, whereas if inertia is predominant no growth at all takes place. Under conditions in which

Jan 1, 1938

By W. E. Ruder

IT has been pointed out by Stead 1 that grains of considerable coarseness may be developed in steels containing from 3 to 5 per cent. of silicon, and in a previous paper 2 the present author has shown that under the proper conditions of annealing single grains having an area of 50 sq. cm. may be obtained. It is the purpose of this paper to discuss in more detail the conditions, promoting the growth of these grains, in the hope that the experiments described may throw some additional light upon the general problem of grain growth in metals. These silicon steels lend themselves particularly well to such experiments because of the ease with which the change in structure may be followed and because of the fact that the effect of the carbon present is largely obliterated by the high percentage of silicon. This forms a solid solution with the iron, which acts very much like a pure metal. This alloy has been found to have but one critical point, viz., Are. This point, determined by Roberts-Austen for Hadfield's 4 per cent. Si alloy, was found to lie at 703° C. Vigouroux3, in working with pure Fe-Si alloys, finds this point to rise from 750° C. for 0.5 per cent. Si to 860° for 6 per cent alloy, with a flat region in the curve from 3 to 4 per cent. Si. No accurate determination of this point has been undertaken by the present author, but it has been roughly determined to lie between 730° and 750°.

Jan 12, 1913

By Gerald Edmunds

CASTINGS of pure metals and many alloys usually have a coarse-grained structure characterized by long columnar grains throughout the main body of the casting. Frequently, the surface exhibits finer, some-times nearly equiaxed grains, and such grains may also compose other parts of the casting. This paper deals primarily with determinations of the grain-orientation textures of the columnar and surface grains of polycrystalline zinc castings. It includes some results upon a cadmium casting and a magnesium casting. A hypothesis is devel-oped to account for the observations on columnar grains and to use as a basis for predicting orientations in other cast metals. A general description of the position of the columnar grains in castings is that their long axes are approximately parallel to the direction of the thermal gradient during crystallization. Thus, in cylindrical castings the columnar grain axes tend to be radial, and in large flat castings where most of the cooling is through one or both of the large mold surfaces they tend to be perpendicular to the mold surfaces. At the edges and corners the direction of heat flow and there-fore the position of the grains is more com-plex. Actually observation shows that even in castings made in flat molds the cooling seems to be somewhat irregular, since columnar grains tend to emanate from certain areas of the cooling surface, indicat-ing that thermal contact was better there than in neighboring areas.

Jan 1, 1940

By E. V. Konopleva, H. J. McQueen, V. M. Khlestov

In addition to increasing grain boundary area during hot working of austenite, the dislocation substructures create potential sites for nucleation of ferrite both at original grain boundaries and at transition boundaries between deformation bands. This development gives rise to a marked increase in the rates of nucleation, of growth and of transformation to ferrite. The combined net effect is product grain refinement that requires less stringent cooling between finish rolling and coiling; hence determination of the above rates would advance process optimization. The important control parameters are the preheat temperature (degree of homogenization), finishing temperature, strain (accumulation near finishing) and the extent of recrystallization after finishing. These effects vary greatly with steel composition; nevertheless, some general rules have been formulated. Moreover, there is also potential for deformation of the refined ferrite. There are other thermo-mechanical processes such as warm forming of metastable alloy austenite to provide a dense substructure that is carried into the martensite on quenching. In other processes, the austenite may be deformed during cooling through the transformation to develop a dense substructure in the ferrite that is stabilized by fine carbides.

Jan 1, 2004

By R. G. Zaripova, N. D. Ryan, H. J. McQueen

The austenitic and ferritic stainless steels do not undergo any phase transformation that induces very fine grains in the product through thermomechanical processing as practiced in most other steels. Considerable grain refinement can be achieved by intense multistage deformation descending through warm to cold working; notably multi-axial forging of bulk material can attain high strains with diminished risk of cracking. Hot working to improve ductility yields finer grain sizes as the T is reduced and strain rate E raised; increased strain has limited effect due to the steady state regime. Ferritic steels develop subgrains with size diminishing as E rises and T falls with the advantage of retained strain hardening; recrystallization delivers grains several times the cell size. In austenitic steels, dynamic recrystallization (DRX) provides finer grains at higher E and lower T with a lower limit due to intergranular cracking. The DRX grains contain a strengthening substructure that can easily be retained by quenching; if static recrystallization is allowed to occur the grain size rises markedly. Multistage rolling with diminishing T is able to provide grain refinement to the same degree as straining at the finishing T yet provides better ductility. grains on recrystallization. As a result of the strain induced transformation to martensite, small subzero deformations of austenitic or duplex steels produce very high dislocation densities that yield marked grain refinement when they revert upon reheating.

Jan 1, 2004

By M. G. Maruma, L. A. Mampuru

"Mill liner wear is a major cost item in the mining industry and there is continuous research to prolong the life of the liners. Over the years it has become apparent that even though high-chromium white cast irons are highly efficient as abrasion-resistant materials, a combination of wear resistance and fracture strength remains difficult to achieve. Increasing the hardness of the high-chrome white cast iron (HCWCI), which improves the resistance to abrasion wear, is often accompanied by a deterioration in fracture strength. Operational conditions inside the mill require that the liner should be made of highly wear-resistant material with some fracture strength. Vanadium additions ranging from 0.2 to 3 wt% were made to HCWCI in an attempt to refine the microstructure. It was found that an increase in vanadium content promotes grain refinement. A content of 1.5-3 wt% gave the best results as measured by the maximum breaking strength.IntroductionOver the years it has become apparent that even though high-chromium white cast irons are highly efficient abrasion-resistant material, a combination of excellent wear resistance and fracture strength remains difficult to achieve. The exceptional wear resistance of highchromium white cast iron (HCWCI) is attributed to hard chromium carbides embedded in an austenitic/martensitic/pearlitic matrix. When the chromium contents exceed 12 wt%, interconnected MC3-type carbides of conventional cast irons are replaced by rodshaped and isolated M7C3 carbides (Powell, 1980), leading to an improvement in the impact toughness, ductility and fatigue resistance. However, the main M7C3 carbides in HCWCI are hard and brittle and can provide an easy path for crack propagation. This limits the applications of HCWCI, particularly in severe impact conditions. Most operations involving HCWCI are in crushing and grinding and the life of a part/liner is limited by its wear resistance. However, improving the wear resistance of HWCI comes at the cost of a significant reduction in fracture strength. Concerns about premature failure of HCWCI mill liners in the mining industries prompted a study to improve the wear resistance and mechanical properties of HCWCI."

Jan 1, 2016

By N. Dahata

Earlier research has indicated that an addition of titanium master alloy to aluminum alloys, results in improved fillability, grain refinement and finer pores. B206 alloy differs from other alloys used in automotive industry, with lower titanium level. This paper examines the effect of alloy chemistry (copper and titanium levels) and grain refinement (A1-5`)oTi- 1%B grain refiner) on grain size, microstructure and microporosity in B206 aluminum alloy. In this study, characterization of grain size, microstructure and microporosity was carried out using light optical microscopy (LOM) and scanning electron microscopy (SEM). Results have shown that the addition of 0.05 wt % titanium resulted in grain refinement of B206 alloy. With an increase in titanium to 0.25 wt further grain refinement, much reduced pore size and uniform distribution of pores resulted. These results suggest that B206 with proper addition of titanium has a promise as an alloy for automotive casting.

Jan 1, 2007

By D. G. Eskin

"It is well known that it takes some time for the solid phase to completely dissolve upon melting, especially inside the defects of insoluble particles, e.g. oxides. Until then the oxides remain active solidification substrates in the case of subsequent solidification. It is also known that ultrasonic melt treatment causes grain refinement through activation and dispersion of solidification substrates (one of the mechanisms) and also accelerates the dissolution of solid metal in the melt. In this study we combine these effects and demonstrate that the introduction of an alloy rod into the matrix melt of the same composition results in significant grain refinement, this effect being increased by the ultrasonic vibration of the rod. The achieved grain size is comparable to that obtained by a standard Al–Ti–B grain refiner. All samples were cast using a standard TP-1 mould to enable correct comparison. The effect of the temperature range of the rod introduction, as well as the effect of ultrasonic vibrations are discussed.INTRODUCTION The fact that by introducing solid metal into the melt one can achieve structure refinement is rather well known. As early as in the 1930s–1950s a series of papers and patents have been published, showing the benefits for grain refinement when dissolving solid metal in the melt before solidification (Danilov and Neimark, 1938; Scheil, 1956). Later on this way of structure refinement was developed further and dubbed “suspension casting” or introduction of “internal chills” (Madyanov, 1969; Zatulovsky, 1981). The solid metal was added in a form of cut wire, cut sheet, or powder, with the resulting grain refining, elimination of columnar grain structure in ingots and castings of steel, copper and aluminium alloys. The underlying mechanisms have been suggested as (1) rapid cooling of the melt due to the latent heat consumption upon melting of the solid metal with the resultant melt undercooling and (2) introduction of many solidification substrates in the form of crystal fragments and active non-metallic inclusions (Danilov & Neimark, 1938; Balandin, 1973). A number of patents have been filed where either the same solid alloy or a master alloy with additions (e.g. Ti for Al alloy) is introduced in the amounts up to 50% (typically less than 10%) into the melt close to the liquidus temperature (Schmidt, 1966; Talbot & Soller, 1973, Krupp, 1974; Bondarev, 1979). The main reason that this technique is not widely used in industry is the possible incomplete dissolution of the solid parts introduced into the melt with ensuing inhomogeneous as-cast structure and potential defects. Also the selection of the temperature range where the technology works the best is not clear."

Jan 1, 2018

By T. Billotte, G. Tirand, J. Zollinger, V. Robin, B. Rouat, D. Daloz

The grain structure prediction of metallic alloys during a welding process is often an issue. In addition to the welding parameters and the alloy nominal composition and properties, the base metal grain structure and chemical homogeneity also play an important role in the formation of the final microstructure. In this paper intensive characterisation of solidification structure in gas tungsten arc welding of Ni base alloy 600 are described. It is shown that in transverse section of the weld, while the grain structure shows some texture following constrained solidification growth, grain density in the weld is similar or greater than grain density in the heat affected zone. This suggests that simulation of the welding structure needs, in order to be predictive, to take into account : epitaxial growth, secondary nucleation events and 3D growth inside the weld pool.

Jan 1, 2015

By C. W. Phillips, R. S. Busk

WITH most cast metals the grain size may vary within wide limits, depending upon the conditions at the moment of freezing. These conditions are subject to control in magnesium-base alloys, by proper melting and superheating techniques, enabling production of quite uniform and fine-grained castings, almost independent of section thickness. However, such variations of grain size as do exist, in common with all metals, are reflected in corresponding changes in mechanical properties. It is the intent of this paper to present data giving the relationship between grain size and mechanical properties. Data are also included on the combined effects of grain size and microporosity. In addition, a short discussion of factors influencing the grain size of sand castings is included. MEASUREMENT OF GRAIN SIZE By grain size is meant the statistical distribution of volumes occupied by the individual crystallites in a metal section. Since this is almost always impossible of direct measurement, several methods of estimating and expressing the size have been evolved. Most exact methods make use of the microscope. This involves study of a single plane cut at random through the metal. What is seen in the microscope are polygons, each the cross section of a grain. The diameters or areas of these polygons are then measured and the grain size is expressed as an average diameter, number of grains per unit area, average area per grain, or, rarely, average volume per grain. Rutherford, Aborn, and Bain3 have discussed extensively the errors involved when using a plane section to estimate volume. They clearly point out that the tendency is to underestimate the true grain size because of the nature of the measurement. The A.S.T.M.. recommends that measurement be made by comparison with a standard chart, by means of Jeffries' planimetric method' or Heyn's intercept method. The Society recommends that the notation be in terms of the number of grains per unit area (as is the grain-size number used for steel) or as an average diameter, expressed in inches or millimeters (as for copper and brass). The method of measurement used for this study was comparison with a standard chart, which has been described by P. F. George.', The chart was constructed by photographing at i00 diameters a uniform specimen with an average grain diameter of 0.003 in. The average diameter of this specimen was determined by numerous measurements, using both the Jeffries and Heyn methods. The other grain sizes were then obtained by appropriate enlargement of the master sample. Careful comparison of the chart method with other methods on identical samples indicates that though the differences be-

Jan 1, 1945

By R. A. Masumura, C. S. Pande, D. Agassi

"Grain size has a different effects on the critical current density (Jc) of high and low transition temperature (Tc) superconductors. In high Tc superconductors, the critical current density Jc across grain boundaries depends sensitively on the misorientation 8. It was pointed out by Pande that this dependence is directly related to dislocation structure of the grain boundaries i.e., it is due to dislocation cores and stresses. These ideas are developed further in this paper leading to a relation between Jc and 8. The analytical result is compared with the experimental results on bicrystals and used to explain the difference in behavior between high and low Tc superconductors.IntroductionThe grain size distribution and its evolution with time and temperature is important in determining the properties of high and low transition temperature (T c) superconductors. The distribution is a convenient function to characterize the processing of a high Tc material during sintering. For example, it is known that the maximum density of ceramic materials is obtained when the grain growth is reduced but grain boundaries are still mobile. For practical superconductors, the processing parameters are selected to achieve optimum properties: highest possible T c and maximum critical current (Jc). In terms of an average grain size, the requirements for optimum T c and Jc are sometime in conflict. In A15 type superconductors and even in NbTi, larger Jc have been measured in smaller rather than larger grain sized material. Conversely, in a high Tc materials like YBa2Cu307, higher Jc have been found in larger grain sized or single crystal specimens.Extended defects such as grain boundaries in low T c superconductors can be effective pinning agents for magnetic flux lines, thereby increasing the transport critical current density of these materials [1]. On the other hand, the transport critical currents in polycrystals of high Tc superconductors such as YBa2Cu307 is found to be relatively small (much smaller than those in single crystals of the same material) [2]. It is now known that grain boundaries in high Tc polycrystals can in many cases form ""weak links"" reducing the critical current density across them. This phenomenon can be studied more precisely by experiments on bicrystals, where there is only one grain boundary. It is found that the critical current density is quite sensitive to the misorientation angle of the bicrystal [3,4]. Although the original experiments were conducted on YBa2Cu307, other Tc superconductors have been shown to give the same result [5]."

Jan 1, 1994

By W. Stumpf

Mill log analyses and complementary hot deformation studies were carried out on the as-cast structure and on a hot rolled structure of the same low carbon steel (SAE 1006) that had been processed according to the direct or hot charge route in a compact strip plant. The main softening mechanisms during hot rolling were found to be complete dynamic recrystallization during each pass above about 1000°C for the as-cast structure with an austenite grain size of 277µm and, secondly, incomplete static recrystallization after each pass with a finer austenite grain size of about 26 µm and tested below 1000°C. A two-stage hot working mechanism has also been found by others on a similar steel also processed by the hot charge route and this may appear to be a characteristic of this type of material. Constitutive constants determined from hot working studies were combined with other constants from the literature and incorporated into a grain size development model that correctly predicts the final ferrite grain size in the hot rolled strip steel. The predicted incomplete static recrystallization during the last three finishing passes in the plant appears to result in some retained strain in the hot rolled austenite as it transforms to ferrite on the run-out table. Analyses with the model predict relatively little sensitivity of the final ferrite grain size to process and product variables such as temperature and strain per pass for the last three passes and also the starting austenite grain size in the slab. Keywords: Low carbon steel, hot rolling of strip steel, rolling mill log analyses, austenite grain size, ferrite grain size, grain size modelling.

Jan 1, 2003

By Božidar Šarler, Umut Hanoglu

A rolling simulation system has been developed to calculate the macroscopic rate of deformation, stress, strain, and temperature fields by using a novel meshless method. The model is based on two-dimensional traveling slice assumption. The slices are aligned towards the rolling direction. The initial size of the billet, bloom or slab and the geometry of the grooves at each rolling stand, the distances between the rolling stands, and continuous or reverse rolling can be arbitrarily considered. The purpose of this paper is to complement these development by predicting and visualizing also the grain size of steel over the strand cross-section for each position in the rolling mill. The grain size models have been in one-way coupled to the macroscopic calculations. First the critical strain for dynamic recrystallization is being checked. If the calculated strain is higher than that, considering also the time for meta-dynamic recrystallization, dynamic or meta-dynamic grain size calculations are carried out. If the strain is too low then only the static recrystallization is considered together with its required time. Finally, austenite grain size and ferrite grain size are calculated respectively as the steel cools down. The listed physical models are based on collection of empirical relations found in literature. The user of the system is free to choose or create the model he wants with his own parameters. The simulation system has been coded as a user friendly computer application for an industrial use based on C# and .NET framework and runs on regular PCs. The computational time for a rolling simulation is usually less than one hour and can thus be straightforwardly used for optimization of the rolling mill design in a reasonable time.

Oct 1, 2019

By Chidchai Loyprasert, R. A. Stoehr

"A computer model has been developed to predict grain structures in alloys solidified under unidirectional flow using a combination of three techniques -- solution of the equations for flow of an incompressible fluid bounded by curved surfaces by the SOLA-SURF algorithm, calculation of heat flow by a finite difference method, and simulation of grain formation by a two-dimensional cellular automaton (CA) model. The changing rigid surface has been taken into account. Heterogeneous and homogeneous nucleation are considered, and the growth kinetics of a dendrite tip are evaluated by the Kurz-Giovanola-Trivedi (KGT) model. These calculations were applied to solidification of an Al-4.5% Cu alloy flowing over a chill on the bottom of a channel, a system for which experimental results were available. Using a variety of flow rates and superheats, agreement between the computed and experimental results were very good including transitions between columnar and equiaxed grains, the aspect ratio of the grains, and the inclination of the grains to the flow.IntroductionModeling of casting and solidification processes has developed over the years along two distinct tracks: (1.) modeling of macroscopic heat transfer and fluid flow for design purposes, and (2.) modeling of microscopic processes such as nucleation and growth in order to improve the scientific understanding of the phenomena involved. The present work is an attempt to combine the macroscopic and microscopic approaches for the solidification of an alloy flowing over a chill. The specific case chosen is that of an aluminum-4.5 % copper alloy flowing linearly over a water-cooled copper chill for which experimental results were available over a range of flow rates and superheats. For the macroscopic calculations, the SOLA-SURF algorithm for solution of the twodimensional fluid flow equations was combined with the finite difference method (FDM) for calculating heat transfer. For the microscopic calculations, heterogeneous and homogeneous nucleation were considered, and the growth kinetics at the dendrite tips were evaluated. The cellular automaton (CA) technique was used to apply the microscopic kinetics to the temperature and .flow conditions."

Jan 1, 1999

W. E. RUDER, Schenectady, N. Y. (written discussion?).-To many members of the Institute, papers like this one by Prof. Jeffries, and that on "Reversal of Inheritance" by Prof. Howe, may seem highly academic. Aside from their direct application to all heat-treating processes, to my mind, the establishment of a physical fact and the correlation of facts into a definite law are in themselves ample justification for immense effort. A law, once established, cannot be ignored by those who carry the work along to practical ends. In studying both papers, it seems to me that each author is more concerned with defending a theory than with explaining the individual results as published by various investigators. Take, for example, my sample of pressed electrolytic iron, frequently referred to by both. Prof. Howe, defending inheritance, says that the grains on the second (1300° C.) heating inherited their size from the mother austenite. Prof. Jeffries, defending reversed inheritance, elabor-.

Jan 4, 1918

ZAY JEFFRIES (written discussion*).-I have read with much interest Mr. Ruder's discussion of Professor Howe's paper, "The Supposed Reversal of Inheritance of Ferrite Grain Size from that of Austenite." I am yet of the opinion that my first explanation' of Mr. Ruder's samples is substantially correct. In that explanation I did not intend, however, to convey the impression that "reversed inheritance" was the general rule nor even of common occurrence but rather that." Non-inheritance" was the general rule. Mr. Ruder suggests that a " perfectly formed" grain is an indication of inheritance. The grains in my Fig. 52 are perfectly formed and in fact more nearly equiaxed than those in Mr. Ruder's Fig: C,3 yet the grains in the former are known not to be inherited but to have been formed by the partitioning of larger austenite grains. On close examination of Mr. Ruder's Fig. C the very shapes of the grain suggest partitioning. Again, the grain size (25 to 30 grains per square millimeter) is at least 10 or- more likely 20 times smaller than might be expected of the austenite from which these grains were .formed under the given time and temperature conditions. Mr. Ruder's newer results on electrolytic and Armco iron are most interesting and instructive. His results fit in. nicely with my conclusions 5, 6 and 10.4 His photographs, however, indicate that the thickness of the pieces of sheet metal are such that abnormal grain growth is promoted. He indicates that equilibrium grain size after heating to 950° C., with furnace cooling, is about the same as his Fig. 1, above. I estimate this to be approximately one-half -grain per square millimeter, or 40 times as large as I have obtained with the same material (Armco iron) after heating Y4-in. bars to 1300° C. for 2 hr. and cooling in the furnace. The effect of thickness of sample on grain size is one which needs more attention.

Jan 3, 1918

By Cyril Stanley Smith

THE art of metallography is mature and the forms in which various micro-constituents appear are well known. Investigations almost without end have disclosed the importance of the exact manner of distribution of phases on the physical properties and usefulness of an alloy. Surprisingly, however, relatively little attention has been paid to the forces that are responsible for the particular and varied spatial arrangements of grains and phases that are observed. Like the anatomist, the metallurgist has been more concerned with form and function than with origins. In this paper it is proposed to develop the simple concept that many microstructures result from an attempted approach to equilibrium between phase and grain interfaces whose surface tensions geometrically balance each other at the points and along the lines where they meet. From this follows a number of principles which may prove to be of interest to the metallographer and of practical use in explaining failures and in designing alloys for particular service. In general, the discussion will be limited to structures obtained after full annealing. The simple principles do not apply to structures resulting from Widmanstatten mechanisms in precipitation, or to other structures not in local equilibrium and in which lattice coherency is a determining factor. INTERFACIAL TENSION Much has been written by physical chemists about the free energy of surfaces, which manifests itself as a measurable surface tension. Two- and three-phase interfaces in fluids seek a condition of minimum energy and approach a geometrical configuration in which the various forces are in vectorial balance.* Measurement of the angles established between the interfaces when in equilibrium provides a quantitative value for the relative value of the surface forces involved. The case of two fluids meeting at the surface of a solid (of great importance in connection with the flotation of minerals) has often been studied, but there have been few investigations of the equally definite interface between two or more crystals of a single solid phase, or the more complex cases where crystal grains of more than one phase are involved. The principles are best illustrated by considering first the case of two and three immiscible liquids. If, as in Fig I, a liquid, 2, is dropped into a less dense liquid, I, it will form a sphere-if gravity and viscosity can be ignored-for in this shape the interface has the smallest possible area. If these two phases should together encounter a third phase in which they are both immiscible, the geometry will change to become that which gives the lowest total energy for all three interfaces. Thus, if the spherical drop of liquid, falling through r, eventually reaches the interface between liquids 3 and I (we are assuming that the density of 3 is greater than 2, and 2 than I), it will assume the shape of a lens, the angle

Jan 1, 1948