Search Documents

By M. Bevis, A. F. Acton, P. C. Rowlands

Defor~nation twinning and transformation twinning modes most likely to be operative in Fe-Ni and Fe-Ni-C martensites have been determined using a new theory of the crystallography of deformation t~inning.~ This analysis shows that potentially important conventional and nonconventional twinning modes1 have been omitted in previous analyses. Discussion is given on the relevance of the predicted twinning modes to the lattice invariant shear associated with the martensite transformation in steels and to anomalous deformation twinning in Fe-Ni-C martensites. THE two most important criteria which appear to govern operative twinning modes in metallic structures1 are that the magnitude of the twinning shear should be small and that the twinning shear should restore the lattice or a multiple lattice in a twin orientation. The latter criterion ensures that the shuffle mechanism required to restore the structure in a twin orientation is simple. These criteria have been adhered to in the prediction of twinning modes2"6 in bcc and bct single-lattice structures with axial ratios in the range y = 1 to 1.09 as, for example, encountered in martensite occurring in steels. Refs. 2 and 3 in particular consider the martensite transformation in steels and the twinning modes in these cases relate to transformation twinning, and hence the lattice invariant shear associated with the martensite transformation. The list of twinning modes which can be compiled from these sources is incomplete and the ranges of magnitude of shear considered could be unrealistically small, particularly in the case of deformation twinning. The latter consideration is supported by the fact that twinning modes with magnitudes of shear large compared with the smallest shear consistent with a simple shuffle mechanism have been established in, for example, the single-lattice structure mercury7 and the multiple-lattice structure zirconium.&apos; In addition the anomalous deformation twins reported by Ftichrnan4 to occur in a range of Fe-Ni-C martensites still remain unexplained. It is clear that a comprehensive analysis of twinning modes likely to be operative in martensite In steels is required. The results of the application of a new theory of the crystallography of deformation twinningg to these structures are presented in this paper. The theory has been used to determine all shears which restore the lattice or a multiple lattice in a new orientation with magnitude of shear up to a required maximum. The orientation relationships between parent and twinned lattices are not restricted to the classical orientation relationships of reflection in the twin plane or a rotation of 180 deg about the shear direction. PREDICTED TWINNING MODES Twinning modes which restore all or one half of lattice points to their correct twin positions will be referred to as m = 1 and m = 2 modes, respectively. These modes are the most likely to describe operative modes in single lattice structures. The bcc m = 1 and m = 2 modes which have magnitudes of shear s in the range s < 2 and s < 1, respectively, have been given10 and are reproduced here in Tables I and 11. Detailed discussion of the crystallography of these modes and cubic modes in general will be discussed elsewhere (~evis and rocker, to be published). The four twinning elements Kl, &,ql,7)2 as well as the magnitude of shear s are given for each twinning mode, and the twinning modes are given in order of increasing shear. Two twinning modes are given in each row of the tables, the twinning mode Kl, Kz, ql, q2 and the reciprocal twinning mode with elements Kl = K,, Ki = Kl, q: = q2, and 17; = ql. The m = 1 and m = 2 twinning modes which describe twinning shears with small magnitudes of shear and simple shuffle mechanisms in bct crystals with -y = 1 to 1.09 are given in Tables I11 and IV, respectively. On increasing the symmetry of the tetragonal lattice to cubic, that is making y = 1, all modes listed in Tables 111 and IV must reduce to crystallographically equivalent variants of the modes given in Tables I and 11, respectively, or become twinning modes with both shear planes as symmetry planes in the cubic lattice and hence not considered in Tables I and 11. With the exception of this last type of mode only those tetragonal twinning modes which reduce to modes 1.1, 1.2, 2.1, and 2.2 of Tables I and I1 are considered in Tables 111 and IV. For values of y in the range -y = 1 to 1.09 the tetragonal modes and the corresponding cubic twinning modes have approximately the same magnitude of shear. The twinning modes listed in Tables 111 and IV are therefore by the criteria given above the most

Jan 1, 1969

By G. M. Willis, F. L. Hennessy

The oxygen pressure in equilibrium with silver and Ag2O-B2O3 melts has been measured between 800&apos; and 900°C, to obtain the thermodynamic properties of the liquid. The compound Ag20. 4B20:1 appears to exist in the liquid, which shows marked heat content and entropy effects. A KNOWLEDGE of the thermodynamic properties of binary liquid silicates, borates, and phosphates would be of considerable assistance in the interpretation of the behavior of multi-component metallurgical slags. However, the literature contains comparatively few studies of the thermodynamics of binary slags. The system Ag20-B,O, attracted our attention as it was known to give a single liquid phase,&apos;,&apos; in which high contents of silver could be obtained (up to 61 pct Ag according to Foex2). Further, it would be expected that the partial pressure of oxygen over melts in equilibrium with metallic silver could be used to determine the activity of Ag2O in the Ag,O-B,O, system. In many respects, it may be expected that the reaction of a basic oxide with boric oxide would be analogous to its reaction with silica. Liquid immiscibility frequently occurs in both borate and silicate systems. With B2O3 and SiO reaction with a basic oxide presumably involves a breakdown of the three-dimensional network of the acid oxide by reaction with oxygen atoms common to more than one silicon or boron atom. Ag2O-B2O3 was therefore investigated as a model of a slag system in the hope that its thermodynamic properties would assist in understanding those of other systems. Several methods for determining the activity of a component in a slag have been described in the literature. Chang and Derge" used high temperature electromotive force measurements to obtain the activity of SiO2 in CaO-SiO2 and Ca0-Al203-Si02 slags, but the cell reaction in their work is not clear. low has used rate of volatilization and vapor pressure measurements combined with phase diagrams to obtain activities in the systems KO-SiO,, Na,O-SiO, and Li,O-SiO," and PbO-SiO26 Taylor and Chipman7 extrapolated their results for the distribution of FeO between liquid iron and CaO (+Mg0)-FeO-SiOl slags to obtain the activity of FeO in the binary FeO-SiO2 system. In principle, one of the most direct methods for obtaining the activity of a metallic oxide in a phase is by comparison of the equilibrium oxygen pressure for the system metal-pure oxide with that of metal oxide-containing phase. Schenck and othersa have studied the stabilization of Ag2O on combination with other oxides (MO,) in the solid state by measurements of the oxygen pressure in systems of the type Ag-Ag,O-xM0,-MOy-0, (gas). Schuhmann and Ensio" have determined the activity of FeO in iron silicate slags in equilibrium with solid iron, using CO/CO2 mixtures to establish known partial pressure of oxygen. Although the method gives the activity of FeO without ambiguity, the slag is not a binary system, and interpretation of the results in terms of the hypothetical binary system FeO-SiO, is not possible. If a metal is solid at temperatures at which the properties of the slag containing its oxide are to be studied, this method has the considerable experimental advantage that the metal can be used as the container for the slag, and contamination by contact with refractories is avoided. In this work, crucibles for Ag2-B,O, melts were made from silver, and the liquid brought to equilibrium with definite pressures of oxygen gas. The oxygen pressure PO, thus fixes the activity of Ag20 in the liquid silver borate. For the reaction at a given temperature. is substantially constant, is directly proportional to the square root of the equilibrium oxygen pressure. Varying the oxygen pressure changed the silver oxide content of the liquid and it was possible to obtain the activity of Ag2O over a range of composition. Experimental Procedure In principle, the method consisted of bringing melts in silver crucibles or boats to equilibrium at a fixed temperature under a definite pressure of oxygen and analyzing the glass after solidification. Materials: B2O3 glass was prepared from A.R. quality boric acid by fusion in platinum. The silver

Jan 1, 1954

By R. M. Hudson, P. E. Perry

Weight-gain data obtained by nitriding low-carbon sheet steel in an amrnonia CNH,) atmosphere indicated that the process obeyed a parabolic rate law. The calculated actization energy for nitriding in the range 964" to 1268°F agreed reasonably well with published data. At 1358"F, rate data indicated that the activation energy decreased. Weight-gain data obtained by uszng mixtures of NH3 -Nz at 1268°F containzng jrom 10 to 100 zol pct NH3 also obeyed a parabolic rate law. The rate of &apos;nitriding increased with an increase in the NH3 content of the gas Mixture. It is well-known that steel heated in gas mixtures containing ammonia (NH3) takes up much larger quantities of nitrogen than steel heated in nitrogen, both gases having a total pressure of 1 atm;&apos; this phenomenon can presumably be attributed to the catalytic decomposition of NH3 on the steel surface to furnish nascent (monatomic) nitrogen. This process was studied bv Brunauer. Jefferson, Emmett, and Hend-ricks at furnace temperatures of 752" and 831°F2 using mixtures of NH3 in Hz. Englehardt and wagner3 reported that, at a furnace temperature of 914°F and under their experimental conditions, both nitriding and denitriding were controlled by the rate of gas-metal reactions at a steel surface rather than by the rate of diffusion of nitrogen in iron. The present study was undertaken to obtain information on the kinetics of nitriding low-carbon steel strip at higher temperatures so that practical rates for short-time strip-annealing treatments could be estimated. Variables studied included time: temperature, and NH, content in the annealing atmosphere. Mechanical and chemical characteristics of steel nitrided in this manner will not be considered in the present article. MATERIALS AND EXPERIMENTAL WORK The samples used were from a commercial low-carbon steel, 0.0244 cm thick, in the cold-reduced condition. The chemical composition of this steel is given in Table I. Panels were cut to 5.1 by 17.8 cm, degreased in toluene, and weighed just before treatment. Four specimens were nitrided under each of the experimental conditions. A study was made of the nitriding rate of steel in a 100 vol pct ammonia atmosphere, 740 mm pressure, at five specific temperatures within the range 964" to 1358°F. The nitriding rates of steel in ammonia-nitrogen gas mixtures containing 10, 18, 26, 50, and 100 vol pct ammonia, 740 mm total pressure, at 1268°F were also determined. All atmospheres used were dried by successively passing them through drying towers packed with soda lime and with Linde Molecular sieve Type 4A. Quoted gas compositions refer to those entering the furnace. Specimens were held in the constant-temperature zone of a vertical annealing tube furnace for times of 14, 3, 5, 10, or 15 min. Gas flow rates were maintained at 3.8 cu ft per hr, which was nineteen volume changes per hour for the system used. The rate of flow was selected to provide a high level of free NH3 for cracking on the steel surface where the ammonia gas is most effectively used as a nitriding agent. The vertical annealing tube furnace consisted of a Hevi-Duty tube furnace with a 2 1/2-in.-ID mullite ceramic high-temperature tube. The constant-temperature zone (controlled within 10°F) was about 10 in. long. After each specimen was degreased, a hole was punched in one end, for attaching the specimen by hook to a chain so that it could be lowered into or raised from the high-temperature portion of the tube by means of a power-driven winch. A stainless-steel access port with O-ring seals was connected by suitable glass-to-metal seals to the cool upper portion of the furnace tube. After the weighed specimen was placed in the access port, the furnace tube was evacuated to approximately 10"3 torr, and then the system was flushed thoroughly with the atmosphere under study. When the gas flow rate and constant-temperature zone of the furnace were established, the specimen was lowered into the constant-temperature zone. The atmosphere flowed from the top to the bottom of the vertical furnace tube and was then vented. For all these runs, during the first 3 min of the time the specimen was in the constant-temperature zone of the furnace the specimen was heating up to the tempera-

Jan 1, 1970

By James R. Holden, Frederick E. Wang

Through both single-crystal and powder X-ray diffraction methods, ten A6B23-type compounds have been confirmed to exist between lanthanides (A) (plus scandium and yttrium) and manganese (B); A = Y, Nd, Sm, Gd, Tb, Dy, Ho, Er, Tm, and Lu. The formation of a compound of this type is shown to he extremely atomic size-sensitive; hence it can be classified as a "size-factorH compound. The "enveloping effect", a geometrical consideration observed in its crystal structure, is proposed as the reason for the A6B23-type compound being size-sensitive. The approximate ideal geometrical ratio of the radii R/r is 1.31 while experimentally A6B23-type compounds have a radius ratio lying within the range 1.2 to 1.4. FLORIO t al.&apos; characterized the structure of Th6MnZ3 as fcc, space group Fm3m, with 116 atoms in the unit cell. Since then, a number of isotypic binary compounds, and recently Gd n,,&apos; have been confirmed to exist. The fact that strontium and barium form A6Bz3-type compounds with magnesium strongly suggested the possible existence of Ba6Liz3. However, investigation3 showed the compound Ba6LiZ3 to be absent. Since both strontium and barium are group 11-a elements and are therefore "open metals",6 the nonexistence of Ba6LiZ3 can hardly be explained satisfactorily by valence-electron considerations. On the other hand, the consistent atomic-radius ratio, (R/r),* observed for the known A6Bz3-type compounds,3 strongly suggests that the formation of compounds of this type is atomic size-sensitive. Therefore, one is tempted to explain the nonexistence of Ba6LiZ3 entirely on the basis of the atomic-size difference between strontium and barium. However, this approach is not entirely without objection. Atoms are not rigid spheres and are known to vary in size within certain limits.7 Since the atomic-radius difference between strontium and barium (0.07 to 0.09A) is within these limits, it is reasonable to assume that the size difference would have a negligible effect on the formation of Ba6LiZ3. This view is further supported by the fact that the radius ratio, R/r, in other known "size-factor" compounds is observed to range widely—for example, from 1.08 to 1.45 for ABz-type compounds (C15, MgCuz type)&apos; and from 1.37 to 1.58 for AB5-type compounds (D2d, CaZn5 type).g The present investigation was undertaken in order to find a more satisfactory explanation for the non-existence of Ba6LiZ3 and, consequently, a better understanding of the nature of the A6Bz3-type compound. The primary objectives are to confirm the previous conclusion3 that the A6B23-type compound is indeed a "size-factor" compound and subsequently to determine the atomic-radius ratio range in which the A6Bz3-type compound can exist. In order to achieve these objectives, stoichiometric A6Bz3 alloys, where manganese (B) was alloyed with various lanthanide elements (A), were selected for investigation. The atomic-radius ratios of lanthanide elements with manganese range from 1.26 for Lu/~n to 1.46 for Eu/Mn. This radius ratio range includes and exceeds the range of all previously reported A6Bz3-type compounds—1.32 for Th/Mn&apos; through 1.38 for Sr/Li. Furthermore, the atomic-size difference between successive elements of the lanthanide series in order of atomic number) is of the order of 0.01A (europium and ytterbium are exceptions). The series of lanthanon-manganese alloy systems is ideally suited to a precise determination of the limits of allowable atomic-radius ratio for A6Bz3-type compound formation. EXPERIMENTAL PROCEDURE The lanthanide metals, in ingot form, supplied by Michigan Chemical Corp. (St. Louis, Mich.) and Nuclear Corp. of America (Burbank, Calif.), were guaranteed by the suppliers to be at least 99.9 wt pct pure (traces of silicon, calcium, and other minor constituents present on occasion, not to be more than 0.05 wt pct) as shown by spectrographic analysis. Manganese metal, in polycrystalline form, was redistilled from the commercial, chemically pure grade and was analyzed to be at least 99.95 wt pct pure. In all cases, the atom ratio between the two elements in each charge was A (rare-earth meta1):B (manganese) = 6:23 and a constant weight, 3 g, of

Jan 1, 1965

By Roland J. Lapee

A history of the progress made in copper refining in Montana is presented. The casting furnaces and the newly rebuilt electrolytic refinery are descmbed and operating details are given. Experiences with various addition agents, effects of rernoval of chlorine from the electrolyte, and effects of separan on electrolyte and on copper deposit are discussed. Observatzons are made on the effect of various impurities in anode copper and behavior of thezr salts in electrolyte and in slime, on treatment of slime for removal of copper, and on electrolyte puriification problems. The improved method for production of starting sheets is described. Attention is given to new materials for construction and to improvement in matrials handling ad quality control. COPPER refining at Great Falls, Montana, dates back to 1892. The original plant produced 65,000 lb. of cathodes per day at a current density of 16 amp per sq ft. In this plant anodes and cathodes were handled to and from the cells by means of hand-operated chain blocks and were moved about the plant by hand trucks. Cathodes were melted in coal-fired, reverbera-tory furnaces, were charged by hand, and refined copper was dipped by hand and cast in iron molds. In 1916, a new, modern plant with capacity to refine 18 million Ib. copper per month was built. Two refining furnaces were built, and each was provided with a twenty-mold Clark casting wheel. In 1922, the furnaces were converted to the use of pulverized coal: in 1923, to the use of oil; and in 1928, to the use of natural gas. In 1926, the plant was enlarged 50 pct. The use of pulverized coal for refining copper at Great Falls was discussed1 in a paper presented at the Salt Lake City meeting of the Institute in September 1925. Also, a comparison of the use of various fuels in copper refining furnaces was discussed 2 in a paper presented at the New York meeting of the Institute in February 1932. Prior to 1943 the cellroom was operated with twenty-five 720-lb. anodes and 26 cathodes per cell. Four cathodes, each weighing about 165 lb., were produced from an anode. In 1943 the weight of the anode was reduced to 460 lb., and two cathodes, each weighing about 190 lb., were produced from an anode. In 1949, a Billet Casting Wheel for the production of 3 in. diam phosphor deoxidized billets was built. A paper, presented at the New York Meeting of the Institute in February 1956, describes3 this plant and operation. In the electrolytic refinery, production is controlled by variation of the current density. In 1956, germanium rectifiers on a separate electrical circuit were provided for the starting sheet section so that current density could be maintained at any desired figure. Also in 1956, a program of modernization and enlargement of the Electrolytic Copper Refinery was started. This program, when completed, will raise the capacity of the Great Falls Electrolytic Copper Refinery approximately 33 pct, to 33 million lb. of cathodes per month. Furnace capacity for melting cathodes and anode scrap is ample to take care of the increased production from the electrolytic refinery. This fortunate condition came about primarily as the result of the three following changes: 1) The furnaces were lengthened 10 ft in 1922 when the fireboxes were found unnecessary for burning powdered coal. 2) Furnace life, and hence furnace capacity, was increased by the successful efforts of the Copper Refinery staff to develop a method for sanding furnace side walls and roof. 3) The natural gas being used has a very low sulfur content. As a result, it is possible to reduce time spent rabbling and poling and thus greatly increase tons per furnace charge. TANKHOUSE AND TANKS The Great Falls Electrolytic Refinery is 535 ft long by 252 ft wide. The cells are arranged in four crane bays in which are operated seven 10 ton Whiting Cranes. The old cells are in groups of ten, with circulation of electrolyte through five cells in cascade and with aisles between the groups of ten cells. The new cells are nested in groups of sixteen with no aisles, are all on one level, and have individual circulation to each cell. Present plans call for 1792 commercial cells and 128 stripper cells. To reduce confusion during the construction period, and to use the cranes, cars, and other equipment already on hand, the cells in the rebuilt section were designed to have approximately the same inside dimensions as the cells in the old part of the

Jan 1, 1962

By J. Chipman, G. G. Hatch

One of the important functions of the iron blast furnace is the desulphur-ization of pig iron before it enters the steelmaking furnaces. However, the increasing concentrations of sulphur in the metallurgical coke, source of approximately 90 pct of the sulphur present in the blast furnace charge, and demands for higher rates of production within recent years have increased the need for greater desulphurization within the iron blast furnace. Furnace operators are beginning to look for desulphurizing agents other than blast furnace slag to accomplish the desired degree of desulphurization. A considerable amount of work has been done on desulphurization outside the furnace with soda ash, calcium carbide and various synthetic slags. Whether the desulphurization of pig iron is accomplished wholly inside the furnace or partly inside and the remainder outside, will be determined by the economics involved. Regardless of which is the case, it is believed that it is necessary to have a better understanding of the physical chemistry of desulphurization by blast furnace slags. To this end, it is the object of the present investigation to attempt what is believed to be the first equilibrium study of the distribution of sulphur between liquid pig iron and a wide range of blast furnace slag compositions. Review of Literature There is a considerable amount of information in the literature concerning the desulphurizing power of iron blast furnace slags, the solubility of various sulphides in the slags, and the effect on desulphurization of temperature, of elements dissolved in the liquid iron, and of viscosity. However, there is nothing to indicate that the equilibrium distribution of sulphur between liquid iron saturated with carbon and iron blast furnace slags has been studied experimentally. Wentrupl has made probably the most detailed study of the desulphurization of pig iron to date. He considered that there are three distinct aspects involved, namely: 1. Desulphurization within the blast furnace (by lime and manganese). 2. Subsequent desulphurization by manganese. 3. The effect of subsidiary reactions on the desulphurization by manganese. The experimental work carried out by Wentrup was devoted mainly to obtaining a better understanding of how desulphurization by manganese was accomplished in the mixer and the ladle. Particular attention was given to the part played by carbon, silicon, and phosphorus associated with manganese in the iron, and the effect of temperature on desulphurization. The experimental results indicated that desulphurization by manganese is purely a process of crystallization of manganese sulphide. The addition of silicon to iron melts containing 3.5 pct carbon and less than 0.5 pct manganese had no noticeable effect on desulphurization, but with 1-2 pct manganese the silicon additions improved the desulphurization. Additions of phosphorus also resulted in improved desulphurizati011 by manganese, but the effect was not as marked as in the case of silicon. It was also found that desulphurization by manganese was further improved by lowering the temperature. In order to explain desulphurization inside the blast furnace, Wentrup considered the system iron, sulphur, calcium, oxygen, manganese. (silicon). The distribution of sulphur between the metal and slag was represented by the following equation: (SS) _ (S)Fe + (S)Ca + (S)Mn .... [S] = [s] [1] The parentheses and the brackets represent the equilibrium concentrations in weight per cent of the slag and metal constituents, respectively. Since FeS D (FeS) _ (FeS) (S)Fe LfeS - [FeS] [S] [2] (CaO) + S e (FeO) + (S)Ca _ (FeO)(S)Ca. (S)Ca _ (CaO) Kl = (CaO)[S] [S] ~K1(FeO) [3] Mn + S D (S)Mn (S)Mn (S)Mn K&apos; = [MnpT "1ST = *lIMnJ !4) Substitution of Eq 2, 3, and 4 into Eq 1 resulted in if = L- + *> (Sol + K^ (S) [51 Eq 5 was used to calcu1;lte -f^j and [S] at 1480°C for slags containing 30-50 pct lime, 0.1-2.5 pct iron oxide, 0-26 pct silica, 2 pct sulphur and iron analyzing 1.5 pct manganese. The value for LFaB at 1480°C was found to be equal to 4.5, based on the experimental work of Bardenheuer and Geller.2 The results of the calculations are shown in Table 1. Although the slags are hypothetical and do not represent the range of compositions found in ordinary blast furnace practice, the calculations indicate that lime is effective in controlling desulphurization only if the iron oxide and silica contents of the slag are kept low. Schenck3 did not claim K1 to be a true equilibrium constant, but an empirical value which varied with the silica content of the slag.

Jan 1, 1950

By M. H. Yoo

Possible interactions of the perfect dislocations of six slip systems or the c dislocation with the (10i2f (ioii), {ioIi}(ioiZ), {1122}(1123), and {1121}(ii26) type twins in hcp metals have been analyzed from the crystallographic and the energetic points of view. Twenty-six distinct types of possible interactions were identified, and those selected based on crystallographic constraints were examined for their energetic feasibilities by use of the anisotropic energy factors. No long-range elastic interaction exists for a dislocation when its Burgers vector is parallel to the twin interface. Under a suitable applied stress, a screw dislocation can cross slip at the twin interface. For basal mixed dislocations in cadmium and zinc, the interaction with {1012} twins is found to be attractive, indicating that incorporation of these dislocations into the twins is energetically feasible and that twin growth will result. On the other hand, the interaction between both basal and Prism mixed dislocations and the {1012} and (1121) twins is found to be repulsive in Mg, Co, Re, Zr, Ti, Hf, and Be. This indicates that under an applied stress a local stress concentration will develop due to a dislocation pileup at the interface, which may result in a site for either the nucleation of other twins or the formation of a crack, depending on the cleavage strength. WHEN a metal undergoes plastic deformation, a certain configuration of slip dislocations will result in a state of dislocation pileup against an obstacle. The stress concentration thus developed may enhance the process of twin nucleation and also twin growth. Furthermore, once formed and dispersed in the crystal, twins can act as effective barriers against slip dislocations. The degree of such mutual influence or interrelation between slip and twinning is generally known to be pronounced in the case of hcp, metals. It is also known that deformation by twinning occurs more commonly in hexagonal metals than in cubic metals. In fact, under suitable stress states, all hexagonal metals exhibit {1012) <1011> type twinning.&apos; In addition to this common type, deformation by (1151) <1126> type twinning occurs in zirconium, titanium, and rhenium, which show remarkable ductility.&apos; The importance of twinning during general deformation to the ductility of hcp polycrystals has been briefly discussed in recent review works.2&apos;3 The purpose of this paper is to analyze the interaction between slip dislocations and twins in the hcp structure and to discuss the nucleation and growth processes of twinning and the role of twinning in the <"°" noil) o, 1/3[112O] (OOO2) 1/3[1123] Fig. l—-Slip systems in hcp structure. ductility of hexagonal metals. The problem will be discussed from the geometric and the energetic points of view in a manner similar to that of the previous work on zinc.4 Since hcp crystals deform by several slip and twin systems, numerous interactions result as possibilities. The Burgers vectors of six slip systems and the c dislocation shown in Fig. 1 and the four twin systems listed in Table I are considered here. A complete tabulation of the possible interactions is followed by discussion of those that are more likely to occur on the basis of crystallographic constraints and energetic considerations. 1) CRYSTALLOGRAPHY OF TWINNING The crystallographic elements, K1, K2, n1, and n2, for the four compound twin systems are now well established.= A unit cell with the base vectors n1, and n2 is shown in Fig. 2 for each twin system. The unit cell before twinning is shown in solid line, and the corresponding unit cell after twinning is shown in dashed line. Also shown in Fig. 2 are the following crystallographic parameters: S is the plane of shear, d the interspacing of the twin habit planes K1,Ø Iis the acute angle between n1, and n 2, e is a numerical factor, and q is the number of K, lattice planes intersected by 17&apos;. These parameters can be expressed in terms of the axial ratio, y = c/a, as listed in Table 11. The macroscopic shear strain of twinning, s, and the magnitude of a "unit twin dis-l~cation,"4 bt, are also expressed in terms of y and given in Table 11. In Table 11, K1 and q1 are given in both Miller-Bravais and Miller indices. In double lattice structures, shuffling of atoms in addition to a homogeneous shear of the lattice is generally required if the original crystal structure is to be restored after twinning. The extent of current understanding on this problem of atom shuffling is per- Table I. Four Twin Systems in Hcp Structure

Jan 1, 1970

By D. J. Mack, R. E. Reiswig

Six examples of the peritectoid transformation were selected from the literature and studied by the method of isothermal transformation. The kinetics and mechanisms of five of the examples are presented as TTT diagrams and photomicrographs. The exist-enc- of a peritectoid in the sixth case is doubtful. ALTHOUGH the peritectoid transformation per se has been known for many years, no precise published data exist concerning the kinetics or mechanisms involved in transformations of this type, except for the brief treatment by Rhines, et al. 1,10 Bearing in mind the fact that investigations of recent years are uncovering more and more peritectoids and suspected peritectoids, a thorough study of the well-established peritectoids appeared to be in order. It was for this reason that a study of the kinetics and morphological mechanisms of six binary peritectoids was undertaken. The six peritectoids selected from the literature for study were those reported at 7.02 wt pct Al-Ag, 26.0 wt pct Sb-Cu, 30.5 wt pct Sb-Cu, 32.3 wt pct Sn-Cu, 8.35 wt pct Si-Cu, and 21.2 wt pct Al-Cu. These selections were based on availability and purity of components, ease of preparation and heat-treatment, and estimated reliability of the available equilibrium diagrams in the regions of interest. EXPERIMENTAL PROCEDURE The alloys used in this investigation were induction melted in electrode-grade graphite and chill-cast in cast-iron split molds. In all cases, the alloys were so brittle that they could easily be broken into samples weighing 1 or 2 g. Chemical analyses showed that the alloys used were close to the respective peritectoid compositions reported in the literature and that the impurity levels were low in all cases. Metallographic examination showed uniform distributions of phases in all samples, indicating uniformity of composition in the samples studied. Isothermal transformation studies were carried out in fused-salt media, using the familiar inter-rupted-quench method. Uniformity of temperature in the salt baths was maintained by continuous stirring with a stainless-steel agitator. On the basis of actual observations of the temperature fluctuations, the estimated temperature control was + 10C for the Ag-Al and Cu-Sb alloys and ±30C for the Cu-Sn, Cu-Si, and Cu-Al alloys. The accuracy of all temperature measurements was estimated to be ±1°C. It was found necessary to mount metallographic specimens of the Ag-Al alloy in cold-curing methyl methacrylate, since the temperatures encountered in mounting in bakelite or lucite caused an appreciable degree of transformation to the ß phase. For the other alloys, wood-flour-filled bakelite mounts were used to avoid extraneous X-ray diffraction lines during the later examination of the metallographic specimens on a Norelco Geiger-counter d if f r ac tomete r. In the X-ray diffraction procedure, agreement between the published diffraction patterns and those obtained in this study was good. This was particularly important for phase identification, since the literature contained little in the way of micrograph description in some cases. Etching of the silver-aluminum alloy for metallographic examination was done by swabbing with either of the following reagents: 1) 10 g CrO3, 1 g (NH4), SO2, 0.5 g NH4NO3, 100 ml H2O, or 2) 10 ml NH4OH, 1 ml 20 pct KOH, 4 ml 3 pct H2O2, 5 ml H2O. The other alloys were etched with the usual bichromate etchant: 2 g K2Cr2O7, 1.5 g NaC1, 8 ml conc. H2SO4, 100 ml H20 (swabbed vigorously). EXPERIMENTAL RESULTS A) The Ag-Al peritectoid at 7.02 wt pct Al— The phase equilibrium involved in this peritectoid is shown in Fig. I.2 The phase boundaries in the vicinity of the peritectoid were most comprehensively established by Hume-Rothery, et al,3 who placed the equilibrium temperature at 448 °C and the equilibrium compositions of the a ß&apos; and y phases at 6.11, 7.02, and 7.24 wt pct Al, respective The alloy used in this study analyzed4 6.95 wt pct A making it slightly hypoperitectoid according to the accepted equilibrium diagram. The rate of the transformation a+ y — ß&apos; varies rapidly with degree of undercooling below the equilibrium temperature, passing through a maximum in the vicinity of 350°C. Thus the TTT dia-

Jan 1, 1960

By W. A. Tiller, W. C. Johnston

A high-temperature electromagnetic stirrer is described in which heating and stirring are accomplished by independently controlled power sources. The appavatus is suitable lor use at temperatures up to 1700°C in a variety of ambient atmospheres. Some typical examples of the homogenizatimz capabilities of the system are given. THERE are few processes in solidification that are not markedly affected by motion in the melt during freezing. In many instances, the mechanisms are diffusion-controlled, and the transport in the melt may be greatly accelerated by deliberately stirring the melt. In zone-refining, stirring1 assists the removal of rejected impurities from the interface, so the process proceeds at a faster rate. The transition from a planar to a cellular interface is caused by constitutional undercooling in the melt ahead of the interface: and stirring delays its onset. Stirring is valuable for homogenization of melts: and chemical reaction with sluggish kinetics may be accelerated. Finally, it has been observed that grain refinement is related to motion in the melt. Fine grain castings are usually produced by the addition of catalysts to the -melt,&apos; catalysts which are thought to act simply as hetereogeneous nucleation centers. Even here motion is important. Richards and Rostoker 5 applied ultrasonic vibration to a solidifying A1-Cu alloy which had been innoculated with a catalyst and found that the grain diameter fell linearly with the amplitude, the peak acceleration and the power input to the melt from the transducer. Finally, mechanical and electrical stirring alone have been used to generate a fine-grained structure.6,7 Johnston ef a1.&apos; have carried out a series of systematic investigations of grain refinement by electromagnetic stirring in a number of low melting point alloys. They found, for example, that the number of grains per unit volume in Pb-Sn alloys could be increased several orders of magnitude by stirring an undercooled melt at the moment of recalescence. In general, a relation AT .H = constant prevailed for a given grain size, where AT was the undercooling of the melt and H the field strength. In more recent work, deliberate homogeneous nucleation of slightly undercooled melts established that the mechanism of refinement must be one involving crystal fragmentation and subsequent multiplication, rather than a "shower" of nuclei effect.9 It is the purpose of this note to describe a stirring device suitable for use up to 1700°C. At low temperatures mechanical stirring and direct-current methods are feasible, but at high temperatures the problem of a protective atmosphere and of electrode corrosion rules out such procedures. The most convenient method for high temperatures is to use externally generated ac fields for both stirring and heating. With rf induction heating alone, considerable stirring and agitation can be achieved, but in general the penetration of field into the melt is small, and the stirring cannot be controlled independently of the heating. In the present experiments, separate power sources of different frequencies for heating and for stirring were used. A susceptor design was chosen so that the 450 kc rf heating field was completely absorbed in the susceptor. The stirring frequency, 400 cps, hereafter called the af field, was chosen so that a high penetration of the melt proper was achieved. EXPERIMENTAL APPARATUS The apparatus, Fig. 1, consists of a quartz tube and end plates, surrounded by an rf induction coil and six equally spaced af stirring coils, four of which are shown in full and a fifth in section. Each af stirring coil is a transformer of which the secondary is a single-turn water-cooled copper loop and the primary is composed of two 10 amp-117 v Variac cores as shown. These cores are cooled by forced air, as each of the six pairs will carry maximum currents of 15 amp for short periods. Each set of Variac windings are connected in series, but opposite sets are connected in parallel with a three-phase 400 cps 400-v source. By properly phasing the coils in this way, a rotating field is produced. Capacitors C1, C2, and C3 in Fig. 2 are used to match this inductive load to the generator. Fig. 3 shows a cutaway view of the quartz tube. The sample (1 in. diam by 1 in. high) is placed in a tapered alumina crucible. An axial W-26 pct Re thermocouple, enclosed by a protection tube, is provided. The cruci-

Jan 1, 1968

By B. M. Larsen, T. E. Brower

A perspective is presented of the steel plant sulphur distribution and elimination problem from coal to liquid steel ready for teeming, giving distributions of sulphur over a range of coke sulphur content, and some methods of sulphur control, in the blast furnace, external desulphuriza-tion between blast furnace and open hearth, distribution between fuel, slag, and metal, and methods and limitations of control of sulphur in the open hearth furnace. AS a part of the 1951 AIME symposium on sulphur in steelmaking, it was thought that a discussion of the distribution of sulphur throughout the whole series of operations, from coal and ore to finished steel ingots, might have some value in giving a perspective on the whole problem. The following discussion is an attempt to present such an overall picture. The order is that of the actual plant operations, beginning with a very brief consideration of the coking process. Sulphur in Coal and Coke Since by far the largest source of sulphur entering the steelmaking cycle is in the coal used to make coke for the blast furnace, it would seem reasonable to eliminate some of it, either from the coal, or the coke, or during the coking process. This has appeared impracticable up to the present, at least, for two main reasons: the low activity of the organic sulphur in either coal or coke, and because of price limitations involved in treating a low cost material such as coke. A variable portion, usually M or less, of the sulphur is present in coal in the form of pyrites or similar compounds, and a large part of this sulphur may be removed in the coal washery. Most of the sulphur, however, is normally present as "organic" sulphur, intimately associated with the coal structure. Its distribution prevents any separation by mechanical means. Its low activity makes improbable &apos; any rapid chemical removal, although hydrogen will remove sulphur from both coal and coke. Thus, prolonged recirculation of coke oven gas in the coking process would tend to leave a smaller percentage of the total sulphur in the coke residue. Table I shows a typical distribution of sulphur from coal into products in the coking process. As the sulphur in the coal increases, the sulphur in the coke tends to increase in about the same proportion. Sulphur in the Blast Furnace The best picture of the situation in the blast furnace is provided by a sulphur balance of raw materials entering, and of products leaving, the furnace. The difficulties in accurate weighing and sampling of the variable solid materials entering this process, and the number of hours required for the raw materials to descend through the furnace under variable operating conditions, make it difficult to obtain an accurate balance. However, balances made over periods of weeks or months tend to average out some of these uncertainties. Table I1 presents three typical sulphur balances similar to a number that the writers have calculated. In most of these the slag volume calculated from the sulphur balance is, in some instances more, and in other instances less, than the value corresponding to the best input and output balances of the other slag constituents (lime, silica, alumina, etc.). Probably the greatest source of error in these calculations is the sulphur content of the slag. Despite some possible inaccuracies the balances of Table II show rather definitely the following points: 1—That 87 to 95 pct of the total sulphur input is in the coke and 95 to 97 pct of the total sulphur output is in the slag. Also, that if any sulphur leaves the furnace with the gas it is relatively small, amounting to a possible 1 pct or less. 2—At the lower sulphur coke level of 0.86 pct the total amount of sulphur charged is 15 Ib of sulphur per ton iron increasing to 26 lb per ton at the higher sulphur, intimately associated with the coal struc-rare burdens containing sulphur-rich ores will the total sulphur burden fail to be nearly proportional to the content in the coke used. 3—The 7 to 9 pct of the total sulphur input from the limestone of furnaces B and C is due to the relatively high sulphur content of the stone, 0.226 and 0.265 pct, respectively. In the case of furnace A, the sulphur content of the limestone was only 0.06 pct which resulted in only 3 pct of the total sulphur input coming from this source. It is rather interesting to compare the sulphur balances of a typical ferromanganese furnace with

Jan 1, 1952

By O. F. Kammerer, W. E. Miller, D. H. Gurinsky, J. Sadofsky, J. R. Weeks

Zirconium and titanium inhibit solution mass transfer of steels by liquid bismuth, mercury, and lead. It is shown that in bismuth and mercury, these adsorb on the surface of the steels and subsequently react with nitrogen and possibly carbon from the steels to form inert, adherent surface layers of ZrN, TiN, or TiN + Tic. Data are presented which describe the condition under which thase deposits form. These inhibitors decrease the solution rate of iron into bismuth, and require a higher supersaturation for precipitation of iron from bismuth. USE of the low-melting heavy metals (bismuth, lead, mercury, and their alloys) as coolants has been limited because solution mass transfer of steels occurs in these liquids; i. e., iron dissolves in the hot sections of the heat transfer circuit and deposits in the colder sections. The rate of solution of iron and the temperature coefficient of solubility are sufficiently great to cause complete or partial stoppage by the deposition in the coldest section of a closed circuit in finite time, even though the actual solubilities are extremely low. In the development of the mercury vapor turbine by the General Electric Co., Nerad and his associates1 discovered that the addition of as little as 1 ppm Ti or Zr to magnesium-deoxidized mercury reduced the mass transfer of ferrous alloys by mercury to a negligible amount. Reid2 reported that titanium was detected chemically on the surface of steels contacted with this mercury alloy in amounts varying from 2.0 to 2.6 mg per sq in., the greatest amount being found in the hottest portion of the circuit. Reid stated that the titanium forms the intermetallic compound Fe2Ti by reaction with iron on the surface of the steels. This compound was presumed to be highly insoluble in mercury. More recently, El-gert and Egan3 have reported a greater than 100-fold reduction in the rate of mass transfer of a 5 pet Cr steel by liquid bismuth upon the addition of titanium (in excess of 50 ppm) and magnesium (350 ppm) in the liquid metal, during experiments performed in thermal convection loops* over the temperature differential 700° to 615° C. Also, Shep-ard and his associates&apos; have reported that the addition of titanium to liquid bismuth and Pb-Bi eutec-tic produced a marked decrease in the rates of solution of both iron and chromium from type 410 steel capsules under static conditions. This inhibiting effect increased with repeated reuse of the capsules. Tests performed in this laboratory under carefully controlled conditions have shown that the addition of zirconium and magnesium, or titanium and magnesium, to liquid bismuth or lead greatly reduces the rate of mass transfer of chromium alloy steels and carbon steels in thermal convection loops with a maximum temperature of 550°C.5-9 The present paper will review the data obtained to date at this laboratory on the behavior of iron and steels in contact with liquid bismuth alloys containing titanium or zirconium, and will attempt to explain the role of the above additives in reducing solution mass transfer. Reaction between the Zirconium or Titanium Dissolved In Liquid Bismuth and an Iron or Steel Surface Reaction between Zirconium Dissolved in Bismuth and the Surface of Pure Iron-—A small pure iron crucible (analyzed by the supplier to contain 0.8 ppm N was contacted with bismuth containing approximately 0.1 pet Mg and varying amounts of a radioactive zirconium tracer. The crucible was then inverted at the temperature of contact. The thin residual layer of adherent bismuth was dissolved in cold, concentrated nitric acid. The crucible surface and the solidified bismuth were then analyzed for radioactive zirconium. An analysis of the activity loss on the crucible surface and the weight loss of the crucible during the nitric acid treatment showed that the acid treatment removed the zirconium that had originally been dissolved in the adherent bismuth, but not any zirconium that may have reacted with the crucible surface. The crucible was then pickled in warm aqua regia to remove all surface activity, hydrogen-fired at 600°C, and recontacted with a new liquid alloy. The results of the experiments contacted 1 hr at 450°C show, Fig. 1, a Langmuir-type adsorption with an adsorption free energy of approximately 17 keal per g atom Zr.5 This deposit was estimated to contain 1 atom of zirconium for each 7 to 8 iron atoms on the crucible surface, assuming a surface roughness factor of the pickled crucibles to be five. Increasing the temperature to 520°C caused consi-

Jan 1, 1959

By C. K. Gupta, P. K. Jena

Niobium (columbium) metal in the form of a button has been produced by calciothermic reduction of niobium pentoxide using sulfur as the heat booster. In these experiments with 50 g of niobium pentoxide per batch, the influence of percentage of calcium in excess of the stoichiometric, the percentage of sulfur by weight of the niobium oxide, and the outer-wall temperature of the bomb on the quality and yield of the metal has been studied. The maximum yield on this scale was 78 pct using 50 pct excess Ca and 20 pct of S and at a bomb-wall temperature of 600°C. In experiments conducted on 500 g of niobium pentoxide, reduction with 40 pct excess Ca and 20 pct of S, and at a bomb-wall temperature of 500°C, the.yield of the metal improved to 82 to 84 pct. Some of the reduced-metal samples have been electron-beam melted and the button melts thus obtained have been observed to be extremely ductile and capable of cold reduction to 98 pct without intermediate annealing. NIOBIUM, because of its high melting point (2468oC), its ductility, its resistance to corrosion especially in liquid metals, its high-temperature mechanical properties, and its relatively low capture cross section (1.1 barns) for thermal neutrons, has potentialities as a structural material in the field of nuclear engineering. Niobium and its alloys are also gaining importance in high-temperature technology as components in aircrafts, jet engines, and missiles, electronics, chemical engineering, and surgery. In 1904 Weiss and Aichel 1 produced niobium by reduction of its pentoxide by misch-metal. Since then a large number of investigations have been carried out for the extraction of niobium from its compounds. The various routes through which the metal niobium has been extracted from its compounds can broadely be classified into the following groups: a) reduction of its halides, oxides, and fluo-salts with metals like calcium, magnesium, sodium, and aluminum;2-10 b) reduction of its halides with hydrogen and oxides with carbon or carbides;11-18 c) electrolytic winning from its halides and fluo salts;19-25 and d) thermal decomposition and disproportion ion of its halides26-30 Among the processes in commercial use at present are sodium reduction or fused-salt electrolysis of potassium niobium fluoride, carbide or carbon reduction of the oxide, and hydrogen or metallothermic reduction of the chlorides. In 1907 Von Bolton2 prepared niobium by reducing niobium pentoxide with aluminum. Dennis and Adam-sona attempted reducing the oxide with magnesium powder in argon atmosphere and the resulting metal contained as much as 5 pct O. Mondolfo8 claimed a method for the reduction of the oxides with aluminum. In the case of both magnesium and aluminum reduction, it has been stated31 that by-product oxides such as MgO or A12O3 would involve a major separation problem. Further, in the case of alumino-thermic reduction, the solubility of aluminum in niobium and intermetallic formation have to be reckoned with. In 1922 Bridge3 produced niobium by calcium reduction of niobium pentoxide. Dennis and pdamson&apos; obtained the metal by a similar method with an oxygen content of 0.2 pct. In most of the cases of metallothermic reduction of niobium oxides mentioned above, the metal niobium is obtained in the form of powder. In an attempt to produce massive niobium metal, lock&apos; used iodine as the thermal booster in the bomb reduction of niobium pentoxide by calcium. Massive niobium metal was obtained with a maximum yield of 75 pct containing 0.1 pct 0. Joly32 carried out experiments on calcium reduction of niobium pentoxide on a 2-kg scale using sulfur as the heat booster in a sealed bomb filled with argon. In his experiments, the reaction was initiated by passing a current through niobium spiral wire embedded in the charge. In a typical run consisting of 2000 g of niobium pentoxide, 3100 g of calcium, and 800 g of sulfur (S/Nb2O5 = 3.3), he obtained massive metal in the form of a very uneven mass with a yield of -57 pct. With lower amounts of sulfur and calcium, powder niobium mixed with small metal globules was obtained and the yield was also poor. From these experiments he concluded that this method is not a suitable one for industrial development considering the quality and yield of metal obtainable and the possible hazards with higher ratios of sulfur to niobium pentoxide in the -~charere. In the present work, a detailed investigation on the calciothermic reduction of niobium pentoxide in a stainless-steel bomb in presence of sulfur as the heat booster was undertaken on a laboratory scale.

Jan 1, 1964

By Edna A. Dancy, Gerhard J. Derge

The specific conductance of FeOx,-CaO melts in contact with iron was found to decrease from 200 ohm-1 cm-1 for FeO, to 40 ohm-1 cm-1 for a melt containing 26.3 pct CaO at 1400°C. The temperature coefficient was positive at all compositions, but became smaller at high CaO contents. Current efficiencies for electrolysis increased from 2.5 pct in FeOx to 17.3 pct at the high CaO composition, indicating a change from predominantly electronic conduction to conduction with a substantial ionic contribution. It was shown that Ca++ ions as well as Fe++ ions carry the ionic current. A subsidiary investigation on the apparent effect of atmospheres of argon, helium, and nitrogen on the electrical conductivity showed that this could be correlated with surface temperature losses, which varied with the thermal conductivities of the gases and resulted in precipitation of metal by the reaction 3 Fe++ = 2 Fe+++ + Fe. The work described in this paper is offered as a contribution to the general fund of knowledge concerning metallurgical slags. Measurement of electrical conductivity and electrolysis are comparatively trouble -free methods for investigating molten materials, but, although these methods had been used for complex slags, it was not until the work of Bockris et al.1 that the approach of examining simple binary slag systems was employed, and CaO-SiO2, MnO-SiO2, and Al2O3-Si9 were studied. Two groups have performed work of particular relevance to the present investigation. Inouye, Tomlinson, and chipman2 studied the conductivity of wustite as a function of temperature and of the addition of 5 mol pct of a number of oxides, including CaO. They concluded that molten FeOx in equilibrium with iron is a semiconductor. Simnad, Derge, and ceorge3 demonstrated the ionic nature of liquid iron silicate slags and also concluded that, although the conductivity of FeOx in equilibrium with iron is predominantly electronic in nature, there is a small ionic contribution. The work reported here on FeOx,-CaO slags consists of three main parts, namely, the determination of the specific conductance over a wide composition range, an investigation into the nature of the conductivity through current-efficiency measurements over the same composition range, and an attempt to identify the current-carrying ions, as well as a subsidiary investigation on the apparent effect of the nature of the inert atmosphere on the conductivity. EXPERIMENTAL Materials. The slags, varying in composition from FeOx to 27 pct CaO, were prepared by heating reagent- grade Fe2O3 in an ingot iron crucible with a suitable amount of CaCO3 and, in some cases, powdered iron, in air. This prefused material was then used for the runs. At the end of each run the cell was removed from the furnace and quenched by immersing the bottom half in water. After crushing, the slags were analyzed for calcium and total iron by the usual wet methods. The oxygen content was obtained by difference. Specific Conductance: Apparatus and Method. Fig. shows the experimental setup, with the conductivity cell and leads of ingot iron. The standard four-probe method for measuring high conductivities was used. In this, the potential drop across the unknown resistance is compared with the potential drop across a known resistance connected in series, i .e., same current through both resistances. Thus there are both current and potential leads to the center electrode and to the crucible, which acts as th other electrode. Both ac and dc circuits were available for the measurements; they have been described in earlier work performed in this laboratory.4,5 The geometry of the cell was such that the center electrode was equidistant from the bottom and sides of the crucible. This ensured that the current path was the same irrespective of the magnitude of the conductivity of the material in the cell. Cell constant were measured with KC1 or NaCl solutions, which have considerably lower conductivities (0.0013 to 0.25 ohm-&apos; cm) than the slags, and this precaution in design made sure that the determined cell constants applied to the cells with contents of any conductivity. The cell-constant determinations were made with the ac measuring circuit to prevent polarization. The four-probe method eliminates lead resistance but not the resistance of those parts of the center

Jan 1, 1967

By Harry A. Lipsitt, Douglas Y. Wang

A fatigue study in completely reversed axial tension-compression has been perforried on high-purity titanium and on three high-purity alloys of titanium. The alloys each contain approxi7nately 0.75 at. pct of a single interstitial element; carbon, nitrogen, and oxygen, respectivley. The results corroborate a previously published theory which proposed that strain aging under alternating stress was responsible for the fatigue limit behavior of certain alloys. The present data indicate that in these alloys an increasing strain-talline aging effect under alternating stress is provided by oxygen, carbon, and nitrogen, respectively. CURRENT research on the nature of the fatigue limit in metals suggests that the presence of a fatigue limit in metallic materials is a manifestation of strain aging that occurs under alternating stress.lm5 A comprehensive theoretical model based on the above hypothesis has been developed to explain the existence of a fatigue limit.&apos; This model also provides increased insight into several other fatigue phenomena as under stressing, overstressing, and coaxing effects. The theory, as well, provides equal understanding for those cases where no real fatigue limit is observed. Briefly, this theory assumes that the S-N curve for a pure metal is a smooth function of the applied stress, and the effect of adding an element that is soluble (or forms a precipitate) in the pure metal is simply to shift the S-N curve to the right. If the added element confers the power to strain age, the result is a further shift of the S-N curve, this time upward and to the right. Since strain aging is not expected to be a strong function of stress, and since damage per cycle is known to be quite stress dependent, it is to be expected that there will be some limiting lower stress at which the strengthening due to strain aging will balance the damage due to crack propagation. This stress is the fatigue limit. The position of the fatigue-limit knee was thought to be a function of the magnitude of the strain-aging effect on both the finite and infinite life portions of the S-N curve. Although the strain aging hypothesis seems to be reasonably valid for bcc materials,2&apos;6 it needed to be tested for both fcc and cph metals. This report is the first of a series concerning the fatigue-limit behavior of titanium with varying amounts of the interstitial solutes (C, N,, and 4) that are known to cause static strain aging in titanium. Yield-point effects have been reported for polycrys-talline high-purity titanium alloys containing either carbon, nitrogen, or oxygen.7&apos;9 These effects were observed at testing temperatures in the range 100 to 300&apos;. In addition yield-point and strain-aging effects have been reported for single crystals of titanium containing 0.1 wt pct C plus N.&apos; These yield points were observed over a wide temperature range, but no room-temperature aging occurred. Aging at 180&apos; was required to cause the return of the yield point. The magnitude of the yield phenomena in titanium containing interstitials is not expected to be as large as is observed in bcc metals because of several factors. Titanium has a very high chemical affinity for oxygen and nitrogen. The thermodynamic stability of solutions of oxygen or nitrogen in titanium is recognized. Lattice parameter measurements of titanium containing arbon, oxygen,1° or nitrogen" show that the "c" parameter is expanded more than the "a" parameter, but that up to about 2 wt pct this results in an insignificant change of the axial ratio &apos;c/a." Ehrlich" has shown that the sites occupied by interstitial atoms in titanium are spherically symmetrical and therefore a lattice expansion, at a constant c/a ratio, results in a simple dilation of the interstitial site. Such a dilation involving no shear has been shown to react only with edge components of dislocations.13 This causes only a weak pinning action. Shear stresses would be anticipated locally when only one of the two interstitial positions was occupied. The carbon atom will cause a symmetrical distortion of the lattice whereas the oxygen and nitrogen atoms have, in addition, the previously mentioned chemical affinity of titanium for these elements. These factors will result in a considerably smaller reduction of free energy upon the association of interstitial atoms with dislocations, and therefore a much weaker pinning than has been observed for the bcc metals. These considerations would lead to the hypothesis that of the interstitial elements considered here carbon would cause the strongest pinning effect in titanium where the amount of interstitial in solution is constant. This hypothesis will be borne out in the analysis of the present results.

Jan 1, 1962

By F. M. Morris

A large reserve of thick coal in southwest Virginia was developed by Clinch-field Coal Co. in 1957-1958 to produce a nominal rate of 1500 tph raw coal. Operation features coal cleaning in transit. Refuse removed averages 35 pct. Evolution of plant from initial conception to completion is discussed, selection of means applied is explained, and performance to date us design expectation is described. The long-range plan calls for ultimate handling capacity of 40,000 tpd raw coal with anticipated clean coal capacity of 26,000 tpd on a three-shift basis. From a materials handling viewpoint, the Moss No. 3 operation is principally of interest as an ensemble. Generally speaking, it uses time-tested equipment and ideas but some of these are employed on a scale that may be new in the industry. At present about 20,000 tpd of raw coal are being handled. This is expected to increase to 40,000 tpd as soon as business conditions justify it. The purpose of this paper is to discuss the evolution of the plant as to materials handling practices, to describe briefly the equipment and methods used, and to comment on performance in certain areas. The subject divides itself naturally into four phases: 1) operations at the mines, 2) transportation from mines to plant, 3) raw coal handling into the plant, and 4) transportation of refuse away from the plant. OPERATIONS AT THE MINES The coal reserves for Moss 3 are in southwest Virginia where Dickenson, Russell, and Buchanan Counties come together. This area contains about 15 square miles and over 100,000,000 tons of coal. Here the No. 4 (Tiller) and No. 5 (Jawbone) seams of the Norton Formation lie so closely together that for practical purposes, they constitute one seam of coal. This seam, which is called the Thick Tiller, varies from 10 to 18 ft in thickness and underlies Sandy Ridge, a mountain cresting between 2400 and 3300 ft above sea-level. At 18-ft seam height (which is considered to be the maximum practical mining height) the parting between seams will be about 3 1/2 ft thick and each bench of coal will be about 7 ft thick. Depending on the amount of impurities in the seams Drover. total reject in this height coal may approximate 50 pet by wt. It will average about 35 pet for the property as a whole. The first move in developing this resource was made in 1953 when contour maps of several square miles around Duty, Va., including all the known outcrop, were made by photogrammetry. At this time, it was felt the prospective operation would be served by the Clinchfield Railroad and a photogrammetric route survey was made by this railroad from Haysi to Duty. Study of the resulting maps indicated only one site—adversely owned—which might accommodate the size washing plant to be erected. Water resources of a dependable nature seemed nonexistent. In 1954 bulk washability tests were made on the Tiller Bench at the Moss No. 1 preparation plant. The tests indicated this portion of the seam, mined separately, would wash to 4 pet ash with good recovery. Also in 1954, development of Moss No. 2, south of Sandy Ridge, was begun. This mine is in the Tiller Seam where it is about 100 ft below the Jawbone Seam. The reasons for developing a mine in the normal Tiller Seam before tackling the Thick Tiller seemed compelling: Railroad service could be established quickly, communications were better (though not good!), more was known of the seam (it had been mined in the years 1911 to 19241, and there was no essential property to be acquired. After some legal skirmishing, the Norfolk & Western Railroad was granted the right to serve the new mine. Three decades earlier the old mine had been served by the Clinchfield Railroad. The event which triggered active development of Moss 3 was the Appalachian Power Co.&apos;s decision in 1956 to build a 450,000-kva power plant on Clinch River at Carbo. This solved the problem of marketing the steam coal which inevitably must be a product of a mine in the Thick Tiller. Management promptly decided to build the preparation plant at Carbo where an excellent site was owned; where railroad service existed; where telephone service could be obtained; and where roads, bridges, and water supply were tolerable.

Jan 1, 1961

By C. R. Knopp

Gas-oil relative permeability ratio is an important relationship in oil reservoir predictive calculations. A correlation has been developed from 107 gas-flood k/k tests on Venezuelan core samples. The correlating parameter is based on restored-state water saturation tests and&apos; is applicable to both consolidated and poorly consolidated sandstone reservoirs. Data of the correlation show that there are no distinguishah1e differences between the mass-data groupings for the two c1assifications A procedure is recommended for running .sufficient relative. permeability analyses to compute a geometric mean of the sample group. The geometric mean is more representative of the total core, and probably the entire reservoir. For example, while only one in four of the k,,/k,., test curves agreed closely with the resultant correlation of this report, the geometric mean curves of the 16 suites (three samples or more). showed good agreetment ill three cases out of four. INTRODUCTION The gas-oil relative permeability ratio is an important, fundamental relationship in most oil reservoir predictive calculations. Predictive calculations are made to estimate future reservoir production characteristics and ultimate oil recovery. The k1,/k2, relationship is specifically needed to relate the surface gas-oil ratio to the reservoir oil and gas saturation, and to calculate the relative movement of these phases within the reservoir whenever some of the more complex driving mechanisms are present. Laboratory k1/k2, tests are not generally run as a routine analysis. Consequently, k1/k2 data often are not available when needed because the cost of laboratory work could not be justified or the need for such data had not been properly anticipated. When laboratory k1/k2, data are available, they are often very difficult to interpret. For example, wide divergence is sometimes shown in a family of k1,/k1, tests representative of the producing horizon in a single well. With these considerations in mind, a study was made to determine if a relationship might exist between the k1,/k2, curve and some other simple laboratory test criteria. The most probable k1/k2, curve correlation for Venezuela described in this paper is the result of the investigation. The presented correlation defines the most probable gas-flood k,,/k,, curve through the medium of air-water capillary displacement and centrifuge water saturation tests. The laboratory procedures of these tests are. relatively simple, and inexpensive; test data should be. widely available- from routine analysis. DATA AVAILABLE, LABORATORY METHODS The report correlation utilized 107 gas-Hood k1/k2, tests run on sandstone cores of Venezuelan reservoirs. Table 1 is a general tabulation of data pertinent to the tests, while Table 2 summarizes the data. Thetests include 96 from Western Venezuela and 11 from Eastern Venezuela. Eighty-two- of the 107 test samples were sandstones that varied from poorly consolidated to-unconsolidated; 25 were consolidated. The average sample porosity was 26.7 per cent and the average permeability was 1,121 md; these values typify the better sandstone reservoirs of&apos; Venezuela. The Welge gas-flood technique,&apos; based on fundamental Buckley-Leverett frontal displacement theory, was introduced in about 1952 and is widely accepted in the industry. The laboratory procedure is relatively simple, rapid, and can be performed on small core samples. While there have been some minor variations in sample preparation and laboratory procedure in the tests used for the correlation, these tests can be generally summarized as follows. The core sample was first sol vent-extracted and dried. Connate-water saturation was restored by the oil-flushing or evaporation-blow down methods. At the beginning of gas flood the hydrocarbon pore volume was completeiy saturated with the test oil phase. Unsteady-state gas-oil displacement then began with the injection of nitrogen or helium. while the displaced oil and gas phases were incrementally metered at the out-flow face. From the test data, the k,,/k,, curve was calculated by the Welge method.&apos; The individual oil and gas relative permeabilities were also calculated." CORRELATING PROCEDURES In attempting to establish a basis of correlation, we found that broad mid-range sections of 105 of the 107 k,,/k,, test curves could be closely duplicated by a straight line. Only two curves did not show a degree of linearity in this region. Correlation-curve definition parameters were subsequently developed from this observation of consistent mid-range linearity. Possible correlating variables were limited to the physical properties measured on core samples that (1) were widely available as common test data and (2) could be easily and cheaply obtained through future laboratory work. The more obvious possibilities were porosity, permeability and

Jan 1, 1966

By Charles R. Keyes

A COMPLETE transformation has taken place in the boras industry during the year 1908. A most remarkable factor in this radical change in method of producing the crude borates has been its removal from the realm of industrial chemistry to the field of mining. With the development of extensive deposits of borate-minerals interstratified in thick sequences of Tertiary clays and sands, their winning becomes a strictly mining-enterprise, of the same kind and of the same certainty as the digging of coal or iron-ore. The major supply of the world&apos;s production of commercial borax now comes from the United States. While formerly all of the boric salts of commerce were laboriously extracted from the waters of saline lakes of the arid regions, or from the bottom-salts of desiccated ponds, it later was largely slowly leached from ancient desert shales and clays. During these periods the borax industry was a very hazardous and expensive vocation.

Oct 1, 1909

By A. L. Slaughter

The Homestake mine, located in western South Dakota, was discovered in 1876. The first reported production was in I 878. Total production through 1965 is 6,554,249 troy ounces of silver and 27,961,276 troy ounces of gold, worth $987,000,000 at present prices. Precambrian sedimentary formations at Lead total 20,000 feet. The Poorman Formation, the oldest, is chiefly ankerite carbonate. The Homestake Formation is iron-magnesium carbonate. The younger formations are argillaceous with some quartzites. The Cambrian Deadwood Formation unconformably overlies the Precambrian. All are cut by Tertiary intrusives. Sideroplesite schist of the Homestake Formation was converted to cummingtonite schist in the zones of higher-grade metamorphism. The Precambrian rocks show the effects of extreme isoclinal folding. A second stage of folding ( cross folds) may have been caused by the Tertiary intrusives. All folds plunge to the southeast. The Homestake ore bodies are pipe-like replacements of the Homestake Formation in zones of cross folding. It is believed that dilatation of the schists has increased permeability in these zones. Chloritization of the schists is the most widespread effect of the mineralization which is divided into four stages. Principal minerals of the first stage are quartz, chlorite, and arsenopyrite; in the second stage, quartz, ankerite, and pyrrhotite. In the third stage only pyrrhotite was deposited. The fourth stage followed the intrusions of Tertiary dikes and deposited mostly pyrite and calcite. Gold accompanies all stages and is always the last mineral to be deposited. It is most closely associated with arsenopyrite of the first stage but can be interpreted as having been deposited during the fourth stage and superimposed on the products of the earlier stages.

Jan 1, 1968

By Edmond S. Miksch

The morphology and growth rate of ice grown in supercooled water are markedly affected by convection of the water. A dendritic ice sheet with water flowing past the dendrite tip exhibits deflection of the dendrite spine toward the upstream direction, accelerated growth of secondary spines on the upstream side, and suppression 01" secondaries on the downstream side. Flow causes a thin disk to become corrugated, the corrugations being parallel to the flow. Flow causes an increase in the growth rate of the ice. IF a small seed of ice is placed in slightly supercooled water, it grows initially as a disk with the axis of the disk parallel to the c axis of the ice crystal.&apos; It appears that growth in the c direction is inhibited, while growth occurs with equal facility in all directions in the basal plane. It appears that growth rate as a function of interface temperature is very nearly equal for all directions similarly oriented relative to the c axis. Subsequent to the discoid growth, branching occurs and dendrites develop. The primary dendrite spines grow parallel to the six equivalent a axes of the crystal. Secondary, and in some cases, tertiary spines also grow in these directions. The highly anisotropic: dendritic growth indicates that there must be some finite difference of growth rate for different directions in the basal plane. These differences must be quite small, however, since if they were large, the discoid mode would not occur. Hence, it was expected that the dendrite growth direction could be deflected from the crystallographic direction by perturbation of the thermal field at the dendrite tip. Experiments were done in which perturbation of the thermal field was accomplished by causing the supercooled water to flow past the dendrite tip. The flow direction was perpendicular to the c axis of the crystal, and was perpendicular to the primary spine of the unperturbed crystal. Supercooled water was supplied by a 15 cm cubical copper box, which was refrigerated and insulated so it could cool water to temperatures down to —2oC, and maintain the temperature constant to O.l°C for the duration of a run. A lucite channel of rectangular cross-section, 12 by 24 mm, was mounted horizontally in the box, with one end open so water could flow into it, and the other end leading out through the wall of the box to a length of tubing leading: to an orifice plate 123 cm below. cnunMn s MIKSCH, MemberAiMe, formerly Research Associate, Research Associate, Fig. l(b)—Vertical section of apparatus A horizontal cross-section is shown in Fig. l(a), and a vertical section is shown in Fig. l(b). Means for observing the ice are shown in Fig. l(a). Collimated light passes through the ice sheet and forms an image on the ground glass screen. A lucite box containing dry air was employed to prevent condensation. Ice of the desired orientation was provided by the seed holder shown in Fig. l(b). This device was made of lucite, and had a reservoir in which ice could be frozen by placing it in a freezer. This reservoir com-unicated with the outside only through a narrow gap between two sheets of lucite. The seed holder (after ice had been washed off its exterior) was immersed in the tank of supercooled water, the narrow gap terminating in a slot in the top of the flow channel. Ice growing through a thin water layer in the gap between the two sheets of lucite was oriented so its c axis was perpendicular to the sheets. This orientation is an application of a growth phenomenon2 which causes ice growing along a smooth solid surface in supercooled water to orient itself so its c axis is normal to the surface. After the ice grew into the channel, further orientation was necessary, since the direction of the a axis

Jan 1, 1970

By W. C. Leslie

W. C. Leslie (Edgar C. Bain Laboratory for Funda mental Research)—This investigation, with its com bination of high-purity metal, calorimetry, and metallography, will serve as a model for annealing studies for a long time to come. The evidence for a competition between recovery and recrystalliza-tion is particularly convincing. The demonstration of edge-nucleated recrystallization was so clear that it may obscure the necessity for other types of "nucleation" sites; otherwise, single crystals of aluminum could not be recrystallized after cold work. At least three types of sites have been found to operate in iron: 1) high-angle boundaries, probably in the same manner as in aluminum, 2) matrix sites, 3) inclusions or second-phase particles. In each instance, the cells developed during cold work and perfected during recovery are the "nucle for recrystallization. Observations by ourselves and others of the dislocation structure of cold-worked metals and of the structural changes that occur during annealing have led us to doubt the validity of the basic assumption of the Lucke-Detert theory, and of the modification developed by Gordon and Vandermeer, that during recrystallization of a metal containing foreign elements in solid solution a high-angle boundary is ad vancing into a cold-worked matrix having a random distribution of solute atoms. We contend that an es sential part of the cold-working and annealing of solid solution alloys is the trapping of solute atoms by the low-energy "wells" afforded by individual dislocations and by dislocation arrays. If cells are formed during cold work, the cell walls are stabilized by this concentration of solute atoms. Thus, solutes inhibit recrystallization not only by decreas-ing the rate of high-angle boundary migration, but also by preventing the formation of high-angle boundaries by subgrain growth or coalescence, i.e., solutes can inhibit both "nucleation" and growth. In addition to stabilizing the dislocation Structure, the solute-dislocation interactions may produce clusters of solute atoms which are not as readily assimilated into the recrystallized grains as are individual solute atoms. There is an additional complication. The dislocation structure formed upon cold working a dilute solid solution may not be the same as that formed during cold working of the pure solvent metal. The solute can either enhance or minimize the formation of cells during cold working. Both effects have been observed in binary alloys of iron. The annealing process is inextricably connected with the mechanisms of strain hardening and solid solution hardening. Although the correlation between experimental boundary migration rates and the predictions made from the theoretical equations are impressive, it should be mentioned that in at least four other instances (Pb-Ag and pb-Au,15, Fe-Mn,16, Fe-Mo)17 the agreement was rather poor. In brief, while it is probable that the migration of high-angle boundaries is held back by solutes in the manner described, this is only a partial explanation of the total effect of solutes upon recrystallization of cold-worked alloys. Paul Gordon and R. A. Vandermeer (author&apos;s reply)— The authors first wish to thank Dr. Leslie for his compliments of this work. It is perhaps worth pointing out that some of Dr. Leslie&apos;s remarks are directed toward subject matter covered in two earlier papers in our work.18,19 We agree with Dr. Leslie&apos;s main contention that the exact nature of the metal structure in the vicinity of a boundary should have an effect on the boundary migration. The theory of Lucke and Detert does not take such detail into account, nor does it account for orientation effects on the migration rates. We assume that since our data checks the theory so well, the effects of detailed structure are statistically averaged out by the polycrystalline techniques used just as are the effects of orientation factors.

Jan 1, 1963