Search Documents

By N. A. Zeigler, T. D. Yensen

Aging is a term that connotes a slow change in properties under ordinary operating conditions. It can be accelerated by increasing the temperature and by mechanical straining. The magnetic properties of iron and some of its alloys have for many years been known to be subject to such a change. But the cause of the change was not recognized until a few years ago after mechanical age-hardening in non-ferrous alloys of the "duralumin" type had been definitely shown, by means of solid solubility curves, changes in electrical resistance and microanalysis, to be caused by slow precipitation of one or more constituents from the solid solution. The similarity between the solid solubility curves of such alloys (Fig. 1) and those for iron-nitrogen (Fig. 2) and iron-carbon alloys (Fig. 3) below the A1 transformation point suggested at once that magnetic aging of iron might be caused by a similar process; namely, by the slow precipitation of carbon and nitrogen in the form of Fe4N and Fe3C at ordinary temperatures because of a cooling rate that had been too rapid to permit complete precipitation according to equilibrium conditions. This suggestion has since been amply confirmed, particularly by the comprehensive work of Köster1 on the effect of nitrogen and carbon. Typical curves are shown in Fig. 4 for the relationship between aging temperature and coercive force. Previous to the establishment of the precipitation-hardening theory it was thought that the effect of nonmetallic elements like C, N, O, etc., on the magnetic properties of iron was greatest when these elements were retained in solid solution, atomically dispersed throughout the interstitial spaces of the iron lattice2. Based on the theory that magnetic properties are intimately connected with the regularity of the lattice, it now follows quite logically that a greater effect on this regularity may be obtained when the above impurities are precipitated as very fine particles of Fe3C, Fe4N, FeO, etc., more or less uniformly dispersed within the grains. But if this is the case, it follows also that there must exist a relationship between the size

Jan 1, 1935

By M. T. Hepworth, R. Schuhmann

Hydrogen solubility measurements were made on a series of Ti-O alloys, and a portion of the 800°C isotherm for the Ti-O-H system was determined. Activities of oxygen and titanium were calculated from the experimentally measured activities of hydrogen, using ternary Gibbs-Duhem integrations. The data were extrapolated to obtain thermo-dynamic properties of binary Ti-O solutions. The thermodynamic properties of a Ti-0 solid solutions thus determined were compared with properties predicted from interstitial models. The Gibbs-Duhem equation, long used to calculate activities in binary systems, has been extended. recently to ternary systems.1"3 When the activity of one component in a ternary system is known as a function of composition, the activities of the other two components can be calculated from that of the first by an integration procedure, provided certain restrictions are observed on the path of integration. Often the activity of one component in a ternary system can be measured readily, while the activities of the other two components are difficult to measure. Gibbs-Duhem calculations are especially applicable to such systems. The further possibility arises of developing a new and indirect approach to activity measurement in binary systems. The proposed new method for binary systems consists of adding varying amounts of a third component whose thermodynamic behavior is easily measured. The activity of this added component is determined experimentally as a function of composition. Then the activities of the two components of the original binary system are calculated by Gibbs-Duhem integrations and extrapolations. For the work to be described in this paper, the Ti-0-H system was selected as a promising system for evaluating this approach. Investigations of the Ti-H system have shown that the activity of hydrogen can be measured precisely with little difficulty. Pure titanium transforms from a low-temperature a form (hcp) to a higher-temperature /3 firm (bcc) at 882.5"~ The phase transformation occurs in a temperature range where hydrogen pressures are easily measured. Also, the diffusivity of hydrogen at temperatures above 600" or 700"~ is large enough that equilibrium can be reached within a reasonable time.

Jan 1, 1962

By W. H. Bassett

ABOUT twenty-five years ago the copper industry had outgrown the Lake Superior production. The electrolytic copper producers had- their process well in hand and the industry was well started in the use of electrolytic copper, both for working into wire and sheet, and for the making of alloys. The trade knew that some brands of copper made good material and that other brands did not. There came together a group of men-a refiner, a rolling-mill man, and a manufacturer of electrical machinery-each of them desiring to know why there was a difference and what .it was. Much investigation was done with wire bars and cakes which involved ' chemical analysis, metallographic examination, physical tests and electrical tests. To find the influence of different kinds of copper in making brass 200 carloads of ingots were made into ,eyelet brass under careful observation; each carload was carefully analyzed and tested before used. The alloy was made and rolled under standard conditions and the results recorded.

Jan 1, 1924

By A. W. Allen

Progress in the art of amalgamation in recent years has been negligible, partly because a copper plate, though it occupies extensive floor space, requires frequent attention and invites theft, is a simple piece of apparatus that gives visible results; partly because the impression prevails that amalgamation is too primitive a process to merit consideration by an up-to-date metallurgist. Both these objections can be overruled. Plates can and will be superseded by more efficient, theft-proof amalgamating units. Improved methods of application other than by settlement or by impact, neither of which is effective for removal of fine gold, will be developed, and the process given wider opportunity. As A. F. Crosse aptly remarked, neglect to utilize amalgamation when free gold can be recovered thus is comparable to the decision of a storekeeper to deal only with credit customers. Usual practice is to follow amalgamation with cyaniding. Rarely is attention paid to the possibility of reversing the order—of extraction of the fine gold first by cyaniding; the coarse, subsequently, by amalgamation. Such a sequence avoids the alleged hazard that amalgamated gold particles may be unacted upon by cyanide, whereas cyanided gold particles are readily amalgamated. Research has indicated, however, that the mercury coating on a gold particle is easily soluble in cyanide solution, probably because such coating is a compound of gold and mercury. Comparative-test results at one plant, of cyaniding a ground concentrate before and after amalgamation, showed total extractions of about 90 and 100 per cent, respectively, indicating high solubility of the mercury-gold particle, if present, or the advantage of preliminary impoverishment before cyaniding, or both. The contention that a loss in residue from an amalgamation-cyanidation plant is occasioned by presence of amalgamated gold that escapes action by solution is not supported by adequate evidence. The function and scope of mercury as an aid to extraction of gold by cyanide is deserving of attention. It is a powerful desulfurizer. To restrict an extractive process to a concentrated product is logical and economical, if preliminary concentration will segregate a high per-

Jan 1, 1935

By T. B. Nolan, R. N. Hunt

In terms of present metal prices, analysis of extant records of the Eureka district indicate past production of the magnitude of $200,000,000 in recovered silver, lead, and gold. Production to date has been thoroughly oxidized ore from massive replacements, fissures, manta, and pipe-like bodies-in large part from the massive Eldorado (middle Cambrian) and Hamburg (middle to late Cambrian) dolomite horizons. In the productive area of about 6 square miles carbonate rocks comprise 80 per cent of the section, which includes early Cambrian to late Ordovician horizons. Four-fifths or more of the district's production came from a block of Eldorado dolomite in Ruby Hill that was moved in over younger horizons on strands of a regional thrust fault. The lower 500 to 600 feet of the block was much fractured and brecciated and became host to the ores. The Ruby Hill fault, a late normal fault, limits the Ruby Hill block of dolomite on the north for its entire length of nearly a mile and terminates the downward extent and exploitation of its ores. It introduced an hiatus of some 40 years in the district's development. Drilling of the past 20-odd years, followed by underground work, has now exposed the down-faulted extension of the Ruby Hill block of Eldorado dolomite and its contained ores, dropped 1800 feet in a dip shift to depths beneath the surface of 2000 to 3000 feet. The ores at depth are the sulfide counterpart of those mined in early years on the footwall side of the fault. The ores of the district may be attributed to the same source as the intrusive quartz diorite outcropping in a limited way south of Ruby Hill but found at depth elsewhere in drill holes. The long north-south axis of the district, its anticlinal structure, its major faults, and the distribution of the intrusives and, in their relation to the surface, the ores themselves, all suggest that erosion has not cut deep enough to expose the extent of intrusives and the ores consequent upon them.

Jan 1, 1968

By R. F. Mehl, G. T. Horne

Diffusion coefficients and mobilities were determined as functions of concentration in the a phase of the Cu-Zn system. Use was mode of incremental diffusion couples to determine the Kirkendall effect at various concentrations. Darken's analysis was used to calculate the individual diffusion coefficients and mobilities from these data. The general diffusion coefficient is a single-valued function of the concentration in this system to within the limits of accuracy of the experimental methods used. The form of the various functions (diffusion coefficients and mobilities) of concentration is the same in every case: it is essentially the same as the usual D vs c curve. AS Darken' has pointed out, the knowledge that the movements of different atoms diffusing in the same lattice need not proceed at the same rate is old. Ionic lattices are the classical examples; application of the generalization to lattices with metallic bonding was much more slowly accepted. It remained for Kirkendall'' to show clearly that metallic diffusion need not be symmetrical with respect to both species. Several investigators" have confirmed the initial observation of a "Kirkendall effect:" marker movement accompanying diffusion. They have extended it to other systems and conditions. The theoretical interpretations have been many; ne of the first and simplest of these is that of Darken"' which is essentially mechanism independent. Darken has shown that the Kirkendall experiment permlts a calculation of the individual diffusion coefficients and, coupled with activity data, offers a means of measuring the mobilities of the separateatoms in a binary diffusion process. There is no published information on the effects of concentration and temperature on the individual diffusion coefficients or on the mobility values in binary solid solution; this paper provides such information for zinc and copper in a brass. For purposes of clarity, the overall D value obtained, for example, from the Boltzmann-Matano solution to Fick's law, will be designated as the "general diffusion coefficient;" and the diffusion co- efficients for the separate diffusing species will be designated as the "individual diffusion coefficients." Experimental Procedure The authors have accepted Darken's analysis and have measured mobilities on this basis. The findings of da Silva and Mehl are accepted: that the Matano interface, defined as the origin of coordinates for the Boltzmann-Matano solution to Fick's law, coincides with the original weld interface; that the "shift" of a marker is the distance between the Matano interface and the marker; that the observed porosity is too small sensibly to affect this result; and that the shift is a linear function of the Matano area, the area defined as the absolute magnitude of the area on either side of the Matano interface. Alloys: The a-phase of the Cu-Zn system was chosen for investigation. The measurements were made on incremental couples for two separate reasons: l—it has been questioned whether D, the general diffusion coefficient, is a single-valued function of concentration (a comparison of the D-values for couples with small concentration range, e.g., 10 vs 20 pct Zn, against those from a wide concentration range, e.g., 0 vs 30 pct Zn, will answer that question); and 2—on the Darken analysis, the mobility can be determined only at the concentration at which the marker is observed to lie (this is but one concentration in a single couple) but choosing a series of incremental couples in the range between 0 and 30 pct Zn provides mobility values at several concentrations, thus giving mobility as a function of concentration. Four alloys were chosen; nominally, pure copper, copper with 10 pct Zn, copper with 20 pct Zn, and copper with 30 pct Zn. This allowed six different types of diffusion couples to be prepared, promising mobility determinations at six different concentrations, but having a minimum range of 10 pct Zn. The copper was an especially prepared pure copper supplied through the courtesy of the Ameri-

Jan 1, 1956

By R. A. Pallanch

IN his recent paper, "The Controversial Art of Flotation," (Mining Technology, March, 1944) E. H. Rose states that "flotation is a science in so many variables that only art can blend them." This statement is apt. It contains meat. In other words, a valid scientific reason is behind every flotative effect. Some of these reasons we understand, but much the greater number of them still escape us. This does not prevent us from "going places," as familiarity and experience with the problems result in an increasingly intelligent and artful handling of these effects. Hypotheses are built up by the individuals concerned to explain certain happenings. Many of these hypotheses have to be abandoned when placed under the glaring scrutiny of actual trial. A few survive the severest of tests, come through unscathed, and are elevated to the rank of theories, something susceptible of certain chemical and physical explanation, under the conditions of the particular plant concerned. If these theories continue to hold water under all the variations existing in most milling plants, they gradually become accepted as facts in the minds of the plant operators. But-and here is a big BUT-these so-called facts may fall down completely in another plant working under an entirely different set of conditions. In other words, each plant tends gradually to build up its own facts, which may or may not be facts at another operation.

Jan 1, 1945

By Luiz C. Correa da Silva, Robert F. Mehl

An experimental study of the movement of markers in the systems Cu/a-brass, Cu/Sna-solid solution, Cu/Ala-solid solution, Cu/Ni, Cu/Au, Ag/Au, employing many types of markers and a variety of temperatures. Marker movement is confirmed and undoubtedly associated with the manner in which atoms move during diffusion. EVERAL years ago Smigelskas and Kirkendall' •S reported measurements on the interdiffusion of copper and a-brass and on the movement of inert markers placed at the original join. These markers were observed to move toward the a-brass side of the join, and from this it was inferred that zinc diffuses in a-brass more rapidly than copper. Similar results were reported a little earlier for the diffusion of solvents into high polymer solids.' This has attracted much attention, particularly from those interested in the mechanism and the theory of diffusion: Seitz3 has employed it in supporting the vacancy theory of diffusion; from the data Darken4 has calculated separate diffusion coefficients for copper and zinc, developing a "phenomenological" theory of diffusion; and Bardeen5 has employed it in constructing a general theory, including therein the work of Seitz and Darken, as has Le Claire. These matters have recently been reconsidered in a sympo~ium.7 Some reason has existed for questioning the work of Smigelskas and Kirkendall.' Their technique consisted in electroplating copper on a-brass, upon the surface of which Mo-wires had been secured, and then diffusing, measuring the extent of diffusion along the normal to the join, and measuring the marker displacement in the same direction by a comparator determination of the distance between Mo-wires on opposite sides of the electroplated brass sample. It has been suspected that the copper plate might have been porous, and that the original interface might have been an imperfect join, both contributing to the transfer of the volatile zinc by vapor transfer;' it has also been questioned whether the marker movement might not in a measure be a property of the marker itself; etc. These legitimate questions require experimental answers; the present paper offers such answers. Moreover, Smigelskas and Kirkendall studied the effect only in the Cu-Zn system and at one temperature; the attention which the work of Smigelskas and Kirkendall has commanded recommends similar

Jan 1, 1952

By Robert D. Pehlke, Robert G. Blossey

The solubility of nitrogen in liquid cobalt and a number of dilute cobalt alloys Was measured in the temperature range 1550" to 1700°C at nitrogen pressures to 1 atm. At 1600°C and 1 atm nitrogen pressure, the solubility in pure cobalt is 0.0047 wt pct [3.8 cu cm (stp) per 100 g Co]. The standard free- energy change for the solution reaction 1/2 N,(g) = gout pct in Co) is iF - 10,000 + 5.24 T cal per g atom N. The alloying elements Cv, Cu, Mo, Nb, Tu, W, and V increase the nitrogen solubility, while Al, Fe, Ni, and Si decrease the nitrogen solubility. THE solubility of nitrogen in cobalt was studied previously by Sieverts, Busch and Dodd,2 and Schenck, Frohberg, and Graf.3 Sieverts investigated the solution below the melting point of cobalt, and Busch and Dodd worked with the liquid metal; however, neither study found any solubility for nitrogen in cobalt. Schenck et al, reported a solubility of 0.005 wt pct N in cobalt at 1600°C. Seagle et al. 4 reported residual nitrogen in cobalt after electron-beam melting, although no reasons were ventured for the presence of this interstitial impurity. This investigation was undertaken to clarify these inconsistencies, and to provide additional data on atomic interactions in dilute metallic solution. EXPERIMENTAL PROCEDURE This study was well-suited to the Sieverts' method, which has not changed significantly since its introduction by Sieverts. 5 The apparatus was of the same design as that described by Humbert and Elliott, 6 with a reaction bulb similar to that reported by Weinstein and Elliott. 7 The buret was not water-jacketed, but the measured gas volumes were corrected for small changes in room temperature. The melt temperature was measured by a disappear ing-filament optical pyrometer sighted through a prism onto the melt surface. Temperature calibration was made based on a measurement of the melting point of pure cobalt taken as 1495°C. The difference between the reciprocal of the observed and the true temperatures was assumed to be a constant, neglecting the changes in emissivity with small additions of alloying elements. The total error in gas-volume measurement caused by accuracy limitations in reading the mercury level in the buret, variation in ambient temperature, and variation in melt temperature was about 0.5 cu cm. This represents approximately 10 pct of the total solubility, and is a larger relative error than in previous studies6-' because of the small nitrogen solubilities encountered in the present investigation. The source and purity of the materials used are listed in Table IA. Recrystallized alumina crucibles were used throughout the investigation. Examination of the crucibles after solidification of the melts showed no reaction or penetration. Alloying of the cobalt melt was accomplished with the dropping technique described by Humbert. 6 The investigation was started using cobalt from African Metals Corp. in shot form, but this material was not of sufficient purity and did not yield consistent solubility results. Sherritt Gordon Mines Ltd. cobalt was then used throughout the investigation. In experiments involving alloying elements which could react with oxygen dissolved in the melt, the cobalt was further refined by annealing at 1400°C in a dry-hydrogen atmosphere. This treatment, described by Woolley and Pehlke, 8 markedly reduced the oxygen content. The oxygen and nitrogen analyses of the cobalt used in this study are presented in Table IB. RESULTS Hot Volume. The hot volume was measured prior to each run and was 50 to 55 cu cm (stp) at 1600°C. It increased linearly with increasing pressure and decreasing melt temperature. Pure Cobalt. Measurements of the solubility of nitrogen in pure cobalt and the temperature coefficient of solubility are summarized in Table 11. At 1600°C and 1 atm nitrogen pressure, the solubility was found

Jan 1, 1967

By Colin G. ink

OUTSTANDI'NG in progress among the less familiar 'metals during 1936 is the electrolytic production of 99.9 per cent manganese meta1 readily and many quantity. Strictly speaking, manganese should not' be classed as 'a rare metal, since there is 800 times as much manganese in the earth's crust as there is zinc or lead. But manganese' as a metal is rare, just as titanium metal is rare though titanium is five times as abundant in the earth's crust as manganese.

Jan 1, 1937

By Herbert Hoover

YOUR committee asks that I speak today on the relations of the engineering profession to public affairs. That takes in a lot of ground. This being a cheerful occasion, I will assume that I should exclude the engineer's function as a taxpayer. This being a nonpartisan occasion, I will assume that I should not refer to his voting emotions in 1940. Perhaps what was in the minds of your committee was the application of engineering thought to the confusions of the times. That is also a lot of ground. I fear, however, its boundaries disappear in mirages and rainbows on all sides. Now any entry of the engineer into public affairs involves his breaking into the fields of the sociologists, the economists, and the politicians. However I don't know why, seeing the difficulties these professions are having, they should not welcome a little help.

Jan 1, 1939

By T. Y. Yan

SUMMARY Laboratory high pressure column tests have shown that the presence of 1-20 ppm of aluminum ion effectively prevents permeability loss during uranium leaching with leachates containing sodium carbonate. If added after permeability loss has occurred, aluminum ion can restore the permeability to nearly its original value. No deleterious effect was observed on uranium leaching performance and the technique should be quite compatible with all field operations. INTRODUCTION The recovery of uranium values from underground deposits by in situ leaching or solution mining has become economically viable and competitive with conventional open pit or underground mining/milling systems (Merrit, 1971). In situ leaching processes are particularly suitable for small, low-grade deposits located in deep formations and dispersed in many thin layers. Many such ore bodies occur along a broad band of the Gulf Coastal Plain (Eargle et. al., 1971). The advantages of the in situ leaching processes have been reviewed (Anderson and Ritchi, 1968). In the in situ leaching process, a lixiviant containing the leaching chemicals is injected into the subterranean deposit and solubilizes uranium as it traverses the ore body. The pregnant lixiviant or leachate is produced from the production well and is then treated to recover the uranium. The resulting barren solution is made up with the leaching chemical to form lixiviant for re-injection. Upon completion of the leaching operation, the formation is contaminated with leaching chemicals and other species made soluble in the leaching operation and has to be treated to reduce the concentration of these contaminants in the ground water to levels acceptable to the regulatory agencies (Witlington and Taylor, 1978). Restoration is accomplished by injecting a restoration fluid, which could be the fresh water or water containing chemicals, into the formation. As it traverses the leached formation, the restoration fluid picks up the contaminants and is then produced at the production well. This produced water is either disposed or purified for recycle. In both phases of operation, formation permeability or well injectivity is one of the most important parameters which determines the viability of the in situ leaching process. Low formation permeability limits production rates, leading to uneconomical operations. The formation is said to be sensitive if there is a sharp loss of permeability on contact with water and other fluids. Many uranium bearing formations, for example, the Catahoula formation of the Texas Coastal Plain, contain significant amounts of clay minerals which are water sensitive. Serious permeability losses can occur when the pH and chemical composition of the lixiviant is significantly different from that of the formation water. Jones has investigated the influence of chemical composition of water on clay blocking of permeability (Jones, 1964) and Mungan studied permeability reduction through changes in pH and salinity of the water (Mungan, 1965). Various mechanisms of permeability damage have been proposed and reviewed (Jones, 1964; Mungan, 1965; Gray and Rex, 1966; and Veley, 1969). When large amounts of swelling clays are present, a significant fraction of the flow channels in the formation can be reduced due to swelling. However, in most cases, swelling need not be the main cause of permeability losses. Particle dispersion and migration or clay sliming can be more important causes for formation damage. Clay particles entrained in the moving fluids are carried downstream until they lodge in pore constrictions. As a result, microscopic filter cakes are formed by these obstructions, plugging the pores, effectively restricting fluid flow and reducing the formation permeability. Moore found that as little as 1-4 percent clays present in a fine grained sandstone could completely plug the formation if they are contacted by incompatible injected fluids (Moore, 1960). It has been found that injection of NaHC03/Na2CO3 lixiviant into formations with significant clay content often leads to loss of formation permeability and well injectivity. To alleviate this problem a change of the lixiviant composition to KHC03/K2CO3 has been proposed. At present, however, many in situ leaching operations employ NH4HC03/(NH4)2C03 mixtures as a source of carbonates. This approach has been successfully used in South Texas by Mobil, Intercontinental Energy, Wyoming Minerals and U.S. Steel, etc. The use of ammonium carbonates solutions, however, contaminates the formation and requires a time-consuming restoration operation. The other approach to reduce the permeability loss is to pretreat the sensitive formation with chemicals which prevent clay dispersion and migration. Such chemicals include hydroxy-aluminum (Reed, 1972 and Coppel et. al., 1973), hydrolyzable zirconium salts (Peters and Stout, 1977), hydrolyzable metal ions in general (Veley, 1969) and polyelectrolyte polymers (Anonymous). Still another approach, is to minimize the "shock" caused by sudden injection by gradually changing the chemical composition of the injected fluids from that of the formation water. THE APPROACH Since permeability loss can be an important factor limiting the efficiency and economic viability of the in situ leaching process, a study was initiated on

Jan 1, 1982

By O. D. Gonzalez, R. A. Oriani

The migration of hydrogen and of deuterium dissolved in a iron and in nickel induced by an applied electrical potential has been measured over a range of temperature. In all cases the intevstitial solute migrates to the cathode. The deduced charges of transport are positive, independent of or slightly dependent on temperature, and markedly larger for the beavier mass. These results are interpreted as showing predominating momenlum transfer between electron holes and the acticated complex of the elementary diffusion act. THE migration of components of a metallic system under the action of an applied electrical potential is called electromigration or electratransport. This phenomenon has been known since 1861 and is becoming of increasing interest to metal scientists because of the opportunity it offers to explore the connections between mass transport and electron transport. Until recently, reliable data have been very scarce because of the difficulties inherent in the measurements. In addition, theoretical understanding has been either nonexistent or rudimentary and at present this situation is only moderately improved. Whereas the prevalent notion had been (and is still widely held) that electromigration directly reflects the state of ionization of a component in its equilibrium state in the metal, it is now becoming clear that the major, if not the sole, factor is the momentum transfer between the charge carriers, themselves in motion because of the applied field, and the metallic components in their activated state. It becomes important therefore to obtain reliable data over a range of temperature on electrornigration in systems where only one component moves, that is, either in pure metals (vacancy motion) or in interstitial solid solutions. Our choice of systems was dictated by our measurements' of the Soret effect in solid solutions of hydrogen and deuterium in iron and nickel. Indeed, the main reason for making measurements on electromigration at all is that we suspected a mechanistic connection between thermal diffusion and electromigration. In our former paper' we proposed that what is common to both phenomena is the momentum transfer between the charge carriers in the metal and the atom as it jumps from one equilibrium position to another. In Soret diffusion, however, there is in addition the dissipation of the energy of the jumping atom via pho-nons. We have since found that similar ideas for this interconnection have been developed by Fiks.2 In the present paper we will discuss in more detail our understanding of the mechanism of electromigration itself, but we turn first to a presentation of the experiments and results. EXPERIMENTS AND RESULTS Let us consider a metallic system in which only one component, m , can diffuse at the temperature in question, and to which are applied simultaneously an electrical potential gradient, grad cp, and a chemical potential gradient, grad p,. One may write3 for the material flux Jm and the electronic charge flux Je by applying the condition, grad = 0, to Eq. [I] and evaluating grad(um/T) for the isothermal case and for activity coefficient independent of composition over the relevant composition range. ere, c is the concentration and D the diffusivity of the mobile component. The experimental technique that we have used to measure Z* for the isotopes of hydrogen individually dissolved in each of two transition metals has been developed from a method used by Wagner and Heller.4 (A somewhat similar method was also used by Herold.5) A wire or thin rod of the metal under study connects two volumes of gaseous hydrogen at identical pressures; hence grad um = 0. A direct current is made to pass through the specimen, serving to maintain the specimen at the desired temperature and also to produce the electromigration of the hydrogen dissolved in the metal in equilibrium with the gaseous hydrogen. Under these conditions, the flux of hydrogen from one volume to the other is given, from Eq. [I] for grad( um/T) =0by

Jan 1, 1968

By C. W. Sherman, John Chipman

IN the mathematical statement of the law of mass action, the activity of each substance consumed or produced in a reaction is used to obtain a numerical constant which is characteristic of the equilibrium of that reaction at the given temperature. The activity in many cases is equal to, or proportional to, the concentration. If this is the case, any units denoting concentrations, such as weight percentage or mol fraction, may be substituted in the equilibrium expression. It is even permissible to use mixed units provided only that each unit is that used in defining the activity of the corresponding substance. The indiscriminate use of concentration units in writing equilibrium constants sometimes leads to very erroneous results. On the other hand, what appear to be contradictory results from similar investigations often might be resolved if the activity is the quantity that is used to determine the equilibrium constant for the reaction in question. De-sulphurization serves as an important illustration. Sulphur in lron and Steel The control of the sulphur content of the finished product is one of the oldest and broadest problems in the metallurgy of iron and steel. In general, this control must be exercised through the action of a slag, and for this reason experimental studies of the distribution of sulphur between slag and metal have assumed considerable importance. Under slags of the type found in the basic open-hearth furnace the ratio (% S)/[% S] does not in general exceed about 8. Under blast-furnace conditions, however, de-sulphurization ratios of 50 to 100 are readily obtained, and under laboratory conditions using similar slags, the ratio may be as high as 400. The fact that blast-furnace metal is much more readily desulphurized than is molten steel rests upon two distinct phenomena. The first of these is connected with the slag reaction. In the open-hearth furnace, the equilibrium in the following reaction places a definite limit on the sulphur content of the slag: S + 0-- = 0 + S- - where the symbols underlined indicate elements dissolved in the liquid metal, while slag constituents are represented as ions. The high oxygen content of the open-hearth bath prevents this reaction from proceeding far enough to effect good desul-phurization. In the blast-furnace process, however, the oxygen activity is limited by the presence of carbon in the metal and in the coke bed and the controlling reaction becomes: S_+O- + C_=S- + CO(gas) [2] The second phenomenon which contributes to the easier desulphurization of blast-furnace metal is the effect of high concentrations of other elements. These elements, notably carbon, silicon, and phosphorus, increase the thermodynamic activity of sulphur several fold. As a result of this increase, the escaping tendency of sulphur from the metal phase is increased and its removal facilitated. Previous Studies It becomes, therefore, a matter of importance to determine the effects of various elements upon the activity of sulphur in the melt. The experimental approach to this problem has been made through the equilibrium between sulphur in the melt and gaseous atmospheres containing hydrogen and hydrogen sulphide. The reaction and its equilibrium constant are written as follows: S + H2(gas) =H2S(gas);K1 = -- [3] Pn2 • as where the symbol a, represents the activity of the sulphur in solution in the metal. The understanding of the chemical behavior of sulphur in steel and pig iron rests on a firm knowledge of the binary system Fe-S. A number of studies of the equilibrium have given closely agreeing results from which the activity and free energy of the dissolved sulphur may be found. The recent paper by Sherman, Elvander, and Chipman1 includes a discussion and quantitative comparison of the results of several of the significant researchesr-" on dilute solutions of sulphur in iron. The results of this study indicate that the relationship between the ratio Ph,s/p+ and the percentage of sulphur in solution is not a linear one. This

Jan 1, 1953

By John Chipman

THE equilibrium between titanium in liquid iron and titanium oxides has been studied by Hadley and Derge.' They have shown that a minimum occurs in the oxygen content of the metal between 0.1 and 1.0 pct Ti, but its exact level was not fixed. The oxide phase in this minimum range was principally TiO, containing some iron oxide. On the other hand, for concentrations above 5 pct Ti the oxide phase was TiO and the oxygen content of the metal increased rapidly with increasing titanium content. From their observations and the free energy of Ti0 and TiOz reported by Kelley and ah,' it is possible to derive approximate values for the activity coefficients of titanium and oxygen in liquid iron-titanium alloys and the conditions for equilibrium in steel deoxidation. Neglecting possible departures from the stoichio-metric composition, the condition of equilibrium with Ti0 may be expressed as follows: TiO= Ti_ + C) log K = log NTl + log [% O] + log yTi + log fQ Here as a matter of convenience the titanium concentration is expressed in mol fraction while that of oxygen is in weight percent. The activity coefficients are yTi and fo. The value of log fo approaches zero as the concentration of titanium approaches zero, while that of log yTi approaches a limiting value of In Fig. 1 the sum of log NTi + log [% 0] is plotted for mol fractions of titanium between 0.05 and 0.50. The smooth extrapolation gives a value of —4.15 which corresponds to log K — log y:i . We therefore have the equation: According to the data of Kelley and Mah,' the free energy of formation of Ti0 at 1900°K is -82,400 cal. The free energy of solution of half a mol of oxygen in iron, taking 1 pct 0 as the standard state, is -29,300 cal. From these we write: From these equations the activity coefficient of titanium at infinite dilution in iron is found to have a value of 0.011 or log y°. = -1.96. This is only an approximate value but is certainly better than the estimated 0.05 in Basic Open Hearth Steelmaking To obtain approximate values for the activity coefficient at intermediate concentrations, it is as- sumed that log yTi(l is constant and equal to -1.96. This corresponds to a value of E;: = 3 In yTi/aNTi = 9.0 which is not unreasonable We now have the value for the activity of titanium and the percentage of oxygen in equilibrium with TiO. It is therefore a simple matter to calculate the activity coefficient of oxygen which is shown in Fig. 2. The effect of titanium is rather large, being comparable to that of vanadium, but not so large as has been reported for aluminum. The slope of the line at low concentrations corresponds to E: = aln yo/ aNTi=-0.37. At concentrations below 1 pct Ti the deoxidation constant and oxygen content for equilibrium with TiO, can be calculated from the foregoing data and the free energy of formation of TiO,, which according to Kelley and Mah is -144,600 cal at 1900°K. The free energy of solution of oxygen has been given; that of Ti to form a dilute solution with 1 pct Ti as the standard state is RT ln (0.5585 y:i/47.9) = -34,800 cal. From these data we find (for concentrations in wt pct): TiO, = Ti + 2 -0; hFTgO0= + 51,200 cal log if =-5.89 The interaction coefficients, concentrations being expressed as wt pct, are e;: = alogfTi/a[% Ti] = +0.048 and e:' = a logfo,/a[ojTi] = -0.187. The calculation gives the following results which, in view of the impurity of the oxide phase, must be regarded as approximate:

Jan 1, 1961

By Milton E. Wadsworth

The reduction of a metal in an aqueous solution may be defined in terms of its source of electrons. In electrolytic reduction electrons flow as a result of applied voltages. In other systems electron sources are reducing gases or metals of lower electron affinity. In the latter two cases it is increasingly evident that reduction is electro-chemical in nature in that electrons flow through one or more conducting phases even though the discharge reactions may not be rate controlling. Electrode processes impmtant in reduction and metal deposition are reviewed and extended to contact reduction by precipitation (cementation) on mme electropositive metals. Contact reduc- tion Processes are ancient in origin and represent techniques which have had successful application for hundreds of years. In spite of this, little has been understood in terms of the dynamic properties of such systems or of important side reactions. In recent years an increasing amount of laboratory work has been carried out on contact reduction and an attempt is made here to summarize the important findings. Undoubtedly the interest aroused recently may be attributed to improved reactor design which has demonstrated on a commercial basis accelerated metal recovery by taking advantage of important kinetic and chemical parameters. of gold cyanide (Au(CN)-), respectively. The most common method for the reduction of ions in solution is by electrolysis. The impure metal is made the anode and the ions are reduced at the cathode. Leach liquors containing reducible ions may similarly be reduced electrolytically using an inert anode. The latter process is referred to as electrowinning. In recent years hydrogen has been used successfully as a gaseous re-Metallurgical Society of the American Institute of Mining, Metallurgical and Petroleum Engineers (AIME). He was also a member of the Executive Board of the Metallurgical Society (AIME). In 1957 Dr. WADSWORTH was given the AlME Best Paper Award as coauthor of a paper on the roasting of sphalerite. On July 1, 1968 Dr. WADS-WORTH started an 24 month assignment in the Philippine Islands to assist in the development of graduate metallurgical engineering at the University of the Philippines. In addition to graduate teaching and curriculum development he will establish research programs to develop extraction processes for metals production in the islands. The project is supported by the Ford Foundation through Educational Projects, Incorporated.

Jan 1, 1970

By G. Osuch, G. Derge, G. M. Pound

Using a new dternating-current potentiometer circuit and a specially designed four-terminal cell, the specific conductance of molten Cu2S-FeS mattes was measured as a function of temperature, from the liquidus to 1500°C, over the complete range of composition. The high conductivities, about 1500 ohm-I cm-l for FeS and 100 ohm-l cm-l for Cu,S, indicate that the conduction is electronic rather than ionic. Molten FeS has a negative temperature coefficient of specific conductance, resembling metallic conduction. Molten Cu,S has a positive temperature coefficient, resembling semiconduction. The binary roughly follows an additive rule of mixtures with respect to both magnitude and temperature coefficient of specific conductance. Metallic bonding in the liquid is postulated to explain these phenomena. MUCH has been learned in the past about the nature of liquids and the ionic or molecular species in solution by means of electrical measurements. Thus, dielectric constants','2 have given information about molecular liquids such as water and benzene. Measurements of dielectric constant usually are impossible in electrically conducting liquids, such as aqueous solutions of ionic salts and molten ionic salts. However, measurements of electrical conductance and ionic transference have provided much knowledge about the latter systems.a-" In recent years, the ionic nature of certain molten metallurgical slags has been established by Derge and Martin7 through electrical conductance and electrolysis measurements. Chipman, Inouye, and Tom-linsonq ave studied the electrical conductance of molten FeO and report a high specific conductance of about 200 ohm-' cm-' (compared with 4 ohm" cm-' for an ordinary ionic liquid such as molten NaCl) and a positive temperature coefficient of conductance. They interpret these results in terms of p-type semiconduction by analogy to the situation in solid FeO.Y imnad and Derge" have studied cell efficiency in the electrolysis of molten FeO-SiO, systems and conclude that ordinary ionic conductance increases with SiO, content. Very recently, interest has been revived in the electrical conductance of liquid metals and liquid metallic solutions. Scala and Robertson1' report a close resemblance between the liquid and solid states with respect to thermal, structural, and compositional relationships. Molten sulphides have not received a great deal of attention. Bornemann and von Rauschenplat" measured the specific conductance of molten Cu2S as a function of temperature with a four-terminal cell using direct current. A high specific conductance and a positive temperature coefficient were found in that investigation." Using a two-electrode apparatus, Savelsberg" electrolyzed various molten sulphide mixtures. He concluded that pure molten Cu,S and FeS were electronic conductors but that the mixtures exhibited some ionic conduction. In the present investigation, the specific conductance of the industrially important Cu-Fe sulphide mattes was measured as a function of temperature and composition in order to investigate the mode of electrical conduction and the structure of these molten mattes. An alternating-current circuit was used to eliminate the effect of any possible electrode reactions. Apparatus The Conductance Cell: Due to the high specific conductance of the systems studied (10' to 10" ohm-' cm-'), the classical two-terminal cell and Wheat-stone bridge apparatus could not be used. A four-terminal cell was developed in order to eliminate lead resistance, and an ac potentiometer circuit was designed to give rapid and sufficiently accurate measurements of the cell resistance. A diagram of the conductivity cell is given in Fig. 1. The molten matte is contained in a dense alundum crucible, and spectrographic graphite rods that dip into the molten matte serve as the four conductance terminals. Two of the graphite rods on opposite sides of the cell serve as current-carrying leads, and the other two graphite rods are null-current probes that detect the potential drop across the cell. These graphite rods are contained in silica tubes, and the lower constricted portions of the two silica tubes define the column of liquid whose electrical resistance is being measured. The electrical resistance of the broad ex-

Jan 1, 1956

By Paul H. Price

The history of the salt industry in West Virginia dates back nearly two hundred years; however, the history of salt as an important raw material for the chemical industry is much more recent. The earliest record of West Virginia salt was in 1753, when a raiding party of Shawnee Indians attacked frontier settlements in Virginia taking captive a number of the pioneers. One of these, who later escaped and returned to her friends, described how the Indians camped at a salt spring on the banks of the Kanawha River and evaporated the water to obtain a supply of salt to take back to Ohio with them. Settlement of the Kanawha Valley began in 1774, that same year the Battle of Point Pleasant ended the Indian's power in the Ohio Valley; and settlement of the Kanawha Valley progressed rapidly with a consequent increase in the importance of the Kanawha Licks as a source of salt. In those days the evaporation of the brine was a family matter for which the family wash-kettle was commandeered.

Jan 1, 1949

By W. Geo. Waring

THE growing need of better methods for the recovery of zinc and other elements from complex sulfide ores has suggested an inquiry respecting a possible group separation of the elements by the aid of volatilization at more or less elevated temperatures. It 'is evident that if the zinc content of such ores can be completely separated, for instance, as pure oxide, by a volatilization 'method, there should be no difficulty in recovering the other elements by smelting and refining. Zinc burning" as applied to zinkiferous ores has advanced but little in practice since the invention of the wetherill grate-furnace process in 1852, as may be seen by reference to papers on the subject.' In other words the thermochemical reaction employed in effecting the volatilization of the zinc is essentially one of reduction by means of carbon, whether the ores treated contain their zinc in an oxidized form, or as the sulfide. The zinc vapors, liberated from the ores at a temperature above that required for the reduction of ZnO by C (about ,1100° C.) or for the volatilization of ZnS per se, (about' 1300°-1400° C.) are subsequently burned to ZnO or to ZnO + SOz, by contact with' atmospheric ,oxygen in the upper or posterior parts of the furnace chamber, or in a so-called "combustion chamber" after leaving the .furnace proper.

Jan 1, 1925

By R. E. Pray, M. C. Fuerstenau, J. D. Miller, B. F. Perinne

Quartz cannot be floated with potassium amyl xanthate as collector at any pH. Complete flotation is achieved with certain minimal additions of amyl xanthate and Pb from pH 5.8 to 8.5 and with amyl xanthate and Zn from pH 7.5 to 8.1. The active species of these metal ions responsible for activation are shown to be PbOH and ZnOH No flotation was effected when Cu and Mg were added as activators. Previous work1,2,3,4,5,6,7 has shown that quartz may be floated in the presence of sulfonate or soap when the following two criteria are met: 1) some portion of the added metal ion has hydro-lyzed to its first hydroxy complex, and 2) a precipitate of the metal-collector has formed. Since sulfonate and soap, two substantially different anionic collectors, conform to these criteria, this phenomenon may well be expected to be of a general nature. In this view then, it would appear that any anionic collector, including xanthates, should also float quartz when these two criteria are met. The object of this paper is to demonstrate the conditions under which quartz will respond to flotation in the presence of potassium amyl xanthate. EXPERIMENTAL MATERIALS AND TECHNIQUES Quartz was prepared by leaching the sized sample (48 x 150 mesh) with HC1 until no iron could be detected in the leach liquor. Conductivity water, made by passing distilled water through an ion exchange column, was used in the investigation. Potassium amyl xanthate was selected as collector. Purification of this reagent was effected with the following method: 1) a laboratory sample of potassium amyl xanthate was dissolved to saturation in hot ethyl alcohol and filtered several times while still hot, 2) the filtrate was cooled, during which time potassium amyl xanthate crystallized out of solution; the xanthate was then separated from the solution by filtration, 3) the amyl xanthate was dissolved in fresh hot ethyl alcohol and filtered, 4) step (2) was repeated, 5) the crystalline potassium amyl xanthate was washed with ethyl ether five times, and 6) the xanthate was dried for 24 hours under continuous vacuum. Inorganic impurities are insoluble in hot ethyl alcohol whereas dixanthogen is soluble in ethyl ether. Five-gram charges of 48x150 mesh quartz were conditioned for 3 min in 125 cc of solution containing the reagents and were floated with the small-scale cell and procedure described in previous papers.478 Reagent grade HCl and NaOH were used for pH adjustment, while reagent n-amyl alcohol was added as frother. Frother addition is necessary with this technique, but it should be mentioned that data, similar to those shown in Fig. 2 for 1 x10-4 mole per liter Pb++ and 1 X 10-4 mole per liter AX-, were also obtained with a Hallimond cell in the absence of frother. EXPERIMENTAL RESULTS Clean quartz could not be floated with amyl xanthate at any pH. Flotation was achieved under certain conditions, however, when various metal ions were added to the system. The activating ions used were lead, zinc, copper and magnesium, added as reagent grade chloride salts. Lead as Activator: The effect of Pb++ as activator was examined at five different additions of amyl xanthate (Fig. 1, 2 and 3). When 1x 10-5 mole per liter xanthate was added in the presence of 1 x 10"4 mole per liter Pb ++, no flotation was obtained at pH 6.5 or below, while a recovery of about 60% was effected at pH 7. When the pH was increased to 7.8, no flotation was again obtained, Complete depression was noted from pH 7.8 to about 10, at which point a recovery of about 30% was achieved. At about pH 11, the precipitate of lead amyl xanthate which was present below this pH dissolved and complete depression occurred. When 2.5x 10-5 mole per liter xanthate was added with the same lead addition, no flotation was obtained at pH 5.8 and below, whereas a recovery of

Jan 1, 1965