Search Documents

By Eugene A. Mizikar

A mathetnatical model of heal transfer in continuously cast steel slabs is described. The model, consisting of a unidimensional transient conduction equation and boundary condition equations, has been pvogrammed for computer solution. Temperatuve and solidification profiles calculated for a 6-in. slab, being cast under several conditions of secondary cooling, are presented and compared. Calculated solidification profiles are in agreement with reported expevinle~ztal values. For the mold zone, the predicted slab shell thickness can be described by: Resuilts of the study indicate that multibank spray cooling followed by radiant cooling should be employed when solidifying thick slabs with minimum surface temperature of 1600°F. Under these conditions, a 6-in. slab can be expected to solidify in about 8.3 min. Computer results also indicate that radiant cooling can replace spray cooling during solidification of the final 30 pct of slab with little increase in overall solidification time. CONTINUOUS casting has come to the forefront of the steel industry in recent years because of economic advantages resulting from increased yields and elimination of several processing steps. Considerable work, however, remains to be done regarding modifications to the process and operating procedures. Both a lack of operating plants in this country and the experimental difficulties associated with direct measurements on moving castings have made determination of solidification rates and temperature distributions difficult. An alternate approach is to mathematically simulate heat transfer in a continuously cast section, and then calculate the temperature distributions as a function of the controllable variables of the process. Simulation of heat transfer during solidification requires that a nonlinear mathematical problem be solved. As pointed out by Ruddle,' there are two mathematical approaches to the problem- the analytical approach and the numerical approach. While the analytical approach is certainly the more elegant of the two, it does require a number of inexact assumptions because of the complexity of the problem. For example, noteworthy analytical treatments of heat transfer in continuous casting have been developed by Savage Hills,3 and pehlke4 only by making one or more simplifying assumptions such as invariant thermophysical properties, constant heat-transfer coefficients in the mold, and linear temperature profiles in the shell. Simplifying assumptions such as these can introduce considerable uncertainty in the validity of results calculated with analytical solutions. Numerical solutions, which are considerably more versatile, appear to be better suited for solving solidification problems. Complex variations in the boundary conditions and variable thermophysical properties can be handled readily with this technique. Whereas numerical computations can be long and tedious when done by hand, results can now be obtained quite rapidly with the use of either the digital or analog computer. Several numerical models of heat transfer in continuous casting have been published. In 1963, Adenis, Coats, and Ragones published a numerical model used to calculate temperature distributions in direct-chill-cast magnesium billets. More recently, Donaldson and Hess 6 presented results obtained with a numerical computer model of heat transfer in continuously cast steel billets. In the present study, a model of unidimensional heat transfer in continuously cast slabs is presented. The method of solution on a digital computer is also included. Calculated temperature distributions and solidification profiles for various schemes of secondary cooling along with attempts to verify the model are also discussed. MATHEMATICAL MODEL The schematic representation of the slab continuous casting process in Fig. 1 illustrates that the slab passes through three distinct zones of cooling. Accordingly, the mathematical model consists of three parts: 1) solidification in the mold; 2) solidification in the spray cooling zone ; 3) solidification in the radiant cooling zone. Heat-Transfer Equations. The model was developed bymaking a heat balance on a horizontal slice of slab over the time period required for the slice to proceed from the liquid metal meniscus in the mold to the cutoff station. As shown in Fig. 2, the imaginary slice extends from the center line to the surface of the slab. As the slice moves downward, heat is conducted from the center line to the surface of the slab at a rate governed by the surface boundary condition and thermophysical properties of the metal in the slice. The following partial differential equation describes the conduction of heat in a medium moving at velocity U in direction Z:

Jan 1, 1968

By B. C. Allen

Liquid surface tensions of copper and 18 Group IV-A to VIII transition metals (Ti, Zr, Hf, V, Cb, Ta, Mo, W, Re, Ru, Rh, Pd, Os, Ir, Pt, Fe, Ni. Co) have been measured by the static pendant-drop and dynamic drop-weight methods on the same rod sample. The drops were formed by electron-bombardment heating in high vacuum. Application of necessary corrections enabled determination of surface tensions to k2 pct and agreement between methods to * 1 to 4 pct for all the metals studied. Evidence is presented that the liquid surface was well screened by metal vapor, suggesting negligible adsorption of gaseous impurities and that the surface tensions measured are characteristic of the metals involved. Correlations between liquid surface tension and melting point, molar volume, heat of vaporization, and atomic number are presented and discussed. Temperature coefficients of surface tension were calculated using Eötvös' law. SURFACE tension and interface energies of metals are the result of incomplete atomic coordination resulting in a force perpendicular to the boundary, tending to minimize its area. These energies are important in many phases of metallurgy, including microstructure, sintering, joining, electronic emission, and lubrication, which in turn affect many physical and mechanical properties of metals. Liquid surface tension refers to the interface between the liquid and its own vapor or nonreactive atmosphere, and is considered a physical property of the liquid. Since equilibrium shapes can be readily obtained with liquids, the units of surface tension (force per length) and interface energy (energy per area) are interchangeable. Thermo-dynamically, the surface tension ?LV is given by the change in free energy F with surface area A at constant pressure P, temperature T, and composition N: for commercially important refractory metals, vanadium, columbium, tantalum,13,14 molybdenum,14 and tungsten,l5 and none for the rare platinum-group metals and rhenium. Measuring the surface tensions of transition metals is difficult because of their high reactivity and melting points. Of the many techniques available,16-18 the pendant-drop 19-21 and drop-weight methods17,22 are considered superior because they can be modified to eliminate persistent sources of contamination such as supports and capillaries necessary in the popular sessile drop, capillary rise, and maximum-bubble-pressure techniques. The necessary modification is to melt a drop on the tip of a vertical rod, which provides support through a solid of the same composition.13,15 The object of the work was to study the surface-tension behavior of copper and transition metals, having reasonably low vapor pressures. Liquid surface tension of each pure metal was systematically determined by using a combination of the static pendant-drop and dynamic drop-weight methods on the same rod. Liquid drops were formed by electron-bombardment heating in high vacuum. EXPERIMENTAL WORK As indicated in Table I, the metals studied were high purity and generally were obtained in rod form. The exceptions were titanium and zirconium, which were machined from crystal bars, and molybdenum (tot 11, osmium, and ruthenium, which were sintered and arc cast into rods. All the metals were ground or swaged to desired sizes between 1- and 7-mm diam, and centerless ground round and smooth to a finish better than 50 µ in., rms. Each rod was thoroughly cleaned with steel wool, degreased, acid etched, and dried. In a typical run, the rod was vertically clamped in a modified floating-zone electron-bombardment furnace designed after Calverley23 and Carlson.24 The specimen was outgassed and maintained positive at several kilovolts in a dynamic vacuum of 10-5 to 10-7 mm. The bottom end was enclosed by a tantalum pillbox and heated by electrons emitted from a hot concentric tungsten filament mounted on a movable bracket. The bombardment power was slowly raised until melting was observed. Power requirements ranged from 30 w for copper to 1300 w for 4-mm tungsten. The stabilized and out-gassed drop of near-maximum size was photographed at 4. 1X on panchromatic film at desired time inter-

Jan 1, 1963

By E. D. Levine

A study was made to delineate the dislocation mechanistns controlling prismatic and basal slip at low temperatu.ves in titanium containing approximately 100 ppm interstitial impurities. Mechanical tests were performed on single crystals prepared from electron-beam zone-refined crystal bar by a thermal cycling technique. Measurements were made of critical shear stress and activation volume as functions of temperature. The study revealed that, for prismatic slip, two overlapping thermally activated processes aye rate-controlling below 300°K. From absolute zero to 220°K, reasonable agreement is obtained between experiment and the mechanism which involves overcoming 01. the Peierl's stress by the formation of double kinks in dislocation lines lying in close-packed directions. Between 220°and 300°K prismatic slip is controlled by a difierent thermally activated process which is characterized by a stronger temperature dependence of critical shear stress. Above 300OK, prismatic slip is controlled by an athermal process. Basal slip is controlled by a thermally activated process between 195' and 400°K. An approximate analysis suggests that it is the Peierl's mechanism that is rate-controlling in this range. Below 195°K, the occurrence of mechanical twinning obscures the features of basal slip. Above '400 X, basal slip is controlled by an athermal process. TITANIUM, an hcp metal with a less-than-ideal axial ratio of 1.587, deforms primarily by {1010}( 1210) slip at low temperatures.'-3 The critical shear stress is strongly affected by the presence of interstitial impurities. For example, churchman3 reported that a decrease in the combined nitrogen plus oxygen content from 0.1 to 0.01 pct decreased the critical shear stress for prismatic slip from 9.19 to 1.4 kg per sq mm at room temperature. Basal slip, however, is less strongly affected by interstitials; Churchman's data for {0001}( 1210) slip show a smaller decrease in critical shear stress from 10.90 to 6.3 kg per sq mm at room temperature, for the same decrease in interstitial content. In the present study, an attempt has been made to delineate the dislocation mechanisms controlling prismatic and basal slip in titanium. In view of the strong effects that impurities may have on at least some of these mechanisms, experiments were performed on material of reasonably high purity, so that comparisons could be made with investigations on material of both lower and higher impurity contents. EXPERIMENTAL PROCEDURE Single crystals were prepared from 1/4-in.-diam electron-beam zone-refined crystal bar, furnished by Materials Research Corp. The analysis of the zone-refined material is given in Table I, indicating that the interstitial content is of the same order as the purer of the two materials investigated by Churchman. The Vickers diamond pyramid hardness number of this material was 85.6. This value agrees quite well with the correlation between hardness and interstitial content proposed by Ogden and Jaffee.4 On the basis of hardness, the material investigated in this study is, therefore, intermediate in purity between that studied by Churchman (Vpn = 97) and that studied by Spangler and Herman (Vpn = 62.4).5 Single crystals were grown by thermal cycling through the a-ß transformation. Two-in.-long by 1/2-in.-diam bars were heat-treated in vacuum for 4 hr at 1160°C. The temperature of the furnace was then lowered to 865°C, and the specimens were maintained at this temperature for 7 days. This procedure was repeated several times, resulting in the growth of individual crystals up to 14 in. in length. Specimens for testing in direct shear along either {l010}(1210) or {0001}(1210) were obtained by electrical-discharge machining. The specimens were in the form of cylinders 1/4 in. in diam by approximately 1/2 in. in length, with a reduced section 1/8 in. in diam and a in. in length. All specimens were electropol-ished subsequent to electrical-discharge machining. The ends of the test specimens were cemented in stainless-steel collars. Shear tests were conducted in an Instron Tensile Tester at various temperatures from 4.2° to 508°K. Measurements were made of critical shear stress and activation volume. The latter measurements were performed by measuring the change in flow stress upon cycling the strain rate between ?1 = 2.62 x 10-3 per sec and ?2 = 2.62 x 10-4 per sec. Activation volume, I)*, is defined as u* =ßkT [11

Jan 1, 1967

By Byron M. Bird

JIGGING is the stratification of a mass of solid particles in upward pulsations of water or in alternating upward and downward pulsa¬tions. The stratification usually is effected in a rectangular open¬top container having a perforated bottom through which the pulsations of water are applied to the materials to be separated; in separating coal, to a mass of coal and refuse. Jigging usually is a continuous process involving three steps: (1) the reception of the raw coal, (91) the stratification of the coal and refuse so that the refuse settles to the screen plate, and (3) the overflow of the cleaned coal and the withdrawal of the refuse collected on the screen. The jig is the most universal washer ever devised. It is handling in practice a wide range of closely sized coals and ores, mixed sizes of coal from 8-in. to 0 through %-in. to 0, and mixed sizes of ores down to %3 O-in. to 0 and even smaller. It is treating materials ranging in specific gravity from that of gold at 19 to that of coal at 1.30. It is handling feeds in which the proportions of high-gravity materials range from 90 per cent or higher, as for some iron ores, to 10 percent or less for some coals. It is handling feeds containing all shapes of particles-round, rectangular, and flat. It is handling simple separation problems and difficult separation problems. Thus, it is an almost universal separating device. One might add that it has been one of the most misunderstood types of washers, from both a theoretical and a practical point of view. Although con¬siderable progress has been made in coal jigging in the past 10 years, the jig probably has more unexplored possibilities than any other type of washer. In this chapter the object has been to treat the jigging process in a practical manner and to give information that will help the reader to improve his particular operation. Inasmuch as performance data from operating plants would contribute little of value, none has been given. Instead, the emphasis is upon jigging principles as a foundation for attacking the problems that arise. These principles are in reality tenta¬tive conclusions from research in the field of jigging during the past 10 years and most of this chapter should be regarded as a progress report on an investigation still under way.

Jan 1, 1943

By Royal Jackman

Two comprehensive papers have appeared regarding the forms of copper that occur in smelter slags, one by Frank E. Lathe1 and the other by C. G. Maier and G. D. Van Arsdale.2 These authors comment on other published results and draw conclusions from their own investiga-tions. The development of X-ray examination of minerals and improve-ment in high-power microscopic methods make it possible today to get additional evidence by means that were not available to earlier investi-gators. The present work was confined to reverberatory slags obtained from five different plants. Some definite conclusions were reached regarding the forms of copper present in these slags. In most cases these agree with some of the findings of Maier and Van Arsdale but new methods of attack were used in the investigation.

Jan 1, 1933

By R. M. Brick

THE etched microstructures of cold-worked alpha brasses, after reduc-tions in excess of about 20 per cent, exhibit dark lines or markings, which have been termed "deformation bands," "etch bands," and "strain markings." The terms will be used synonomously in this report. The markings are similar in their general appearance to etch figures found in cold-worked iron, aluminum and other cubic metals but in certain respects may differ noticeably. At reductions of from about 20 to 50 per cent in brass, they appear as straight dark lines of irregular spacing and at random directions in different grains, but apparently intimately associated with definite crystallographic planes, usually {111}, in each crystal. At higher reductions, the markings on the rolling plane tend to appear in a direction transverse to the rolling direction in all crystals but may bend away at considerable angles, particularly near grain boundaries. Samans,1 on the basis of X-ray and micrographic studies of two single crystals of brass, rolled to a 50 per cent reduction, showed that the lines of deformation were the edges of thin mechanical twins parallel to octahedral planes. Barrett and Levenson2-4 have made comprehensive studies of the deformation of iron and of aluminum. They show that in these two metals the deformation markings originate by division of the original homogeneous crystal into banded areas differing in orientation by an amount that increases with the degree of deformation. The markings thus cannot be twin bands, and Barrett, in a review of this subject,2 attributed Samans' results on brass to an accidental choice of deformation that left the bands in approximately a twinned relationship. On the other hand, Mathewson5 has suggested that in some cases mechanical twinning offers a better rationalization of the coverage of some (111) pole figures of worked brass than the assumption of band rotation into sym-metry positions. Considering next a deformed and annealed alpha brass, the most striking aspect of its microstructure is the multiplicity of twin bands.

Jan 1, 1940

By Henry S. Rawdon

The structure of wrought iron as usually described by metallographists and workers in metal in general is that of a fairly pure iron. Impurities, if present, are usually considered as being in solid solution in the crystals of the ferrite matrix or as forming part of the ever present "slag streaks." Attention is herein directed to one type of these dissolved impurities; to the detection of such impurities, particularly in low-grade irons, and to their possible influence on the physical properties of the metal. The attention of the Bureau of Standards was first directed to wrought iron of the peculiar and unusual characteristics to be described later, in material which had failed in service and was submitted for test. The examination was extended to other grades of wrought iron, to see whether such features are of common occurrence in this class of material. MICROSTRUCTURE OF WROUGHT IRON Usual Structure The general microstructure of wrought iron is so well known that a detailed description of it here is needless. There is shown in Fig. 2 the structure of a sample of Swedish iron which illustrates well the matrix of ferrite crystals in which are embedded the slag threads so characteristic of the puddling process and the subsequent working of the material. The usual commercial product contains varying amounts of the impurities commonly associated with iron and steel, i,e,, manganese, phosphorus, silicon, and sulphur together with small amounts of carbon. Of these, the larger part of the phosphorus and some of the silicon present are held in solid solution in the ferrite while most of the remaining impurities, other than carbon, exist in the in closures of slag. Aside from the discontinuities introduced by the slag streaks the ferrite matrix has the microstructure and appearance of a pure metal. No definite orientation of the various crystals or grains is apparent; deep etching of the specimens reveals no intracrystalline features (i.e., within the grains) other than those noted in " pure " metals in general (etching pits).

Jan 1, 1918

By J. Szekely, J. W. Evans

A formulation is given for radiant heat loss from the surface of molten steel held in a ladle and numerical solutions are presented for the resultant integrodif-ferential equations. The results are shown on plots of the net rate of heat loss against time, with the geometlry, emissivities, thermal properties of the walls, and Preheat as parameters. It was found that the rate of the vadiant heat loss from the exposed metal surface was comparable to the conductive losses from the Portion of the metal in direct contact with the walls. The radiant heat loss was markedly affected by the geometry and by the thermal properties of the walls the net radiant heat loss from the steel could be drastically reduced by providing a top cover and also by the provision of preheat. The results of this work are thought to be relevant to heat loss calculatious concerned with vacuum degassing and with operatzons in which molten metals are held in ladles. WhEN molten steel is held or transferred in ladles heat losses inevitably occur. The accurate assessment of these heat losses is of importance in process design considerations, as the temperature of the metal may have to meet rigid specifications in certain stages of the processing sequence. As shown in Fig. 1, the principal mechanisms responsible for heat loss are: i) Conduction into the walls and ii) Thermal radation from the top metal surface (possibly augmented by natural convection) The conductive heat loss may be readily calculated as much information is available in the literature on both the general theory1'' and on its particular applications to ladle heat loss calculations. Evaluation of the radiant component of the heat loss is less straight-forward. In two recent papers it has been shown that the presence of a slag layer 2 to 3 in. thick would prevent any significant heat loss for periods up to an hour,5'" however, should bare metal be exposed, constituting the radiant surface, substantial heat losses could occur. This latter case is of considerable practical interest in various degassing operations and in the vacuum processing of certain specialty steels, where the top surface of the metal is not covered by a protective slag layer. Under these conditions a substantial radiant flux is emitted by the top metal surface; part of this flux leaves the system, part of it falls in the ladle walls. The radiant flux incident on the ladle walls is in part absorbed and conducted away, and in part reflected and re radiated. The purpose of the paper is to present a general formulation of the problem, the resultant solution of which should enable computation of the net radiant heat loss from the metal surface. The principal variables, the effect of which will be investigated, include initial metal temperature, geometry, thermal and radiative properties of the wall, and various modes of preheat. FORMULATION Consider a cylindrical enclosure, shown in Fig. 2, with a diameter D, height H, and wall thickness W, the cover at the top of the cylinder having a wall thickness L. If the temperature of the base (representing the top surface of the metal) is kept constant, say, Tb, the problem is to calculate the net rate of heat loss from the base as a function of time, geometry, the initial temperature distribution, and the thermal properties of the system. As customary, the formulation will consist of the mathematical statement of the conservation of thermal energy within the system.

Jan 1, 1970

By C. D. Thompson, D. C. Violetta, C. A. Czako

The mechanics of the continuous sintering process are briefly reviewed and the application of this process for the beneficiation of principal minerals used in making cement is shown. Pilot plant tests using pyrobeneficiation techniques were conducted with mixtures of limestone and shale in various ratios. Results are illustrated and show the percentage removal of carbon, carbonates, sulfur, alkalies, and volatiles as a function of the limestone-shale ratio. Tests with varying quantities of solid fwls, added to mixtures containing limestone and shale in a 4 to 1 ratio, were made to attain optimum elimination of the volatile constituents. Experimental results show that 90 pct of these constituents were removed during the calcination sintem'ng process. The advantages of using the beneficiated material for making cement are discussed with regard to prepam'ng, handling, and clinkering operations. A sintering process for treating raw materials or kiln waste products which contain high amounts of alkaline materials is described. Tests were performed with a high alkaline waste dust mixed in a 1 to 1 ratio with raw cement meal. The data from these tests and others using different halides for aiding alkali removal are evaluated by the alkali content of the cement clinker. Test results showed that the Dwight-Lloyd process could be applied with facility in beneficiating raw materials for cement making. Substantial quantities of deleterious volutile constituents were removed by calcining and sintering techniques. The continuous sintering process has been applied for a number of years as a means of beneficiating minerals for subsequent conversion to merchantable products. The process was developed and first used for nonferrous metallurgical applicatfons. However, it has risen to prominence for ferrous metallurgical applications where last year in the U.S. and Canada nearly 40 million tons of iron ore were treated by grate firing processes. Through the years of process development many investigators have realized the versatility of the sintering process and have directed experiments toward the use of it for applications with nonmetallic minerals. These have included calcining limestone, carbonizing coal, expanding shale, and burning cement raw meal. Today commercial sintering plants produce cement clinker from mixtures of limestone and shale, lightweight aggregate from shale or clay, and calcined products from phosphate shale. The mechanics of the downdraft sintering process have been presented by many authors. In general practice the materials to be agglomated are mixed as moistened fines with a solid fuel and this mixture is spread uniformly on the grates of the sintering machine. The charge is then momentarily directed under an ignition furnace. After ignition a continuous downdraft of air promotes combustion of the fuel and through heat transfer particles become partially fused into a coherent mass. This is cooled from continued draft induced after combustion. Many investigators have shown that distinct zones are formed during the process of sintering iron ores. At the midway point of the sintering process these are prominent and are, respectively, from the top of the bed to the bottom, comprised of a cooling zone of sintered material, a combustion zone of sinter formation, a calcination zone, a drying zone, and a zone of unreacted materials. Completion of sintering occurs near the discharge end of the machine where the firing zone approaches either the grates or grate dressing material called hearth layer. Here the peak draft temperature usually can be noted in the windboxes. The zonal pattern has been illustrated with the use of graphs which show thermal gradients as they progress through the bed of materials during sintering. Fig. 1 shows, in simplified form, thermal gradients existing when clinkering cement on a sintering machine. The thermal gradients illustrated are for periods of initiation, midway and completion of the clinkering process. The occurrence of these zones has importance for the beneficiation of cement raw materials. A research program was performed at the Dwight-

Jan 1, 1961

By Paul M. Tyler

Mineral raw materials, because they are essential to our industrial prosperity and military strength, must be made available in substantial quantities. regardless of cost. Variations in the cost of procuring minerals, as compared with the cost of obtaining other commodities, and changes in the cost of production of different minerals relative to one another, will naturally influence the quantities of a given mineral item that can be marketed readily. However, in the last analysis, the cost of winning minerals from the ground usually constitutes only a small fraction of the value of the end-use products and services derived from them. The miner's share of the cost of a typewriter, telephone circuit, or even a modem office building, is almost insignificant. Far more important to the ultimate consumer, therefore, than any minor saving in cost is the assurance of ample supplies of the minerals needed to aliment our expanding economy. To create a successful metal mining operation involves ( 1 ) the discovery and systematic development of a mineral deposit, (2) extraction of the ore, (3) the mechanical and metallurgical processing to eliminate worth- less or harmful impurities, and (4) the profitable disposal of the product. The first step of the technological process belongs chiefly in the realm of the geologist, the second is a mining engineer's job, the third is mainly up to the millmen and extractive metallurgists; whereas the fourth and equally important step poses a commercial problem that has aspects in common with the marketing of other products. Ore finding is far from being an exact science. There is no mathematical criterion for evaluating the efficiency of prospecting and exploration. Lady Luck shares the honors or assumes part of the blame along with the prospector and geologist. Marketing like- wise involves factors that seemingly defy engineering analysis. Although the other steps can be measured with some degree of mathematical accuracy, even they tend to vary so widely in basic principle, as well as in

Jan 1, 1959

By John Griffen

THEORY HYDRAULIC classification as explained by Rittinger and others was largely restricted to conditions wherein the free-falling velocities of the particles were conceived as governing the separations effected (see chapter 9). The laws governing classification under free-falling conditions have been summarized by Chapman and Mott as fo1lows:l Nomenclature r represents, in a particle of irregular shape, the round-hole screen through which the particle would pass, or some other linear dimension of the particle equivalent to d, the diameter of a sphere. s, the specific gravity of the particle. 1, the specific gravity of the water used as a medium. K, a constant. W, the velocity of the water. In an Upward Current of water 1. A particle will move upward if [V > K dr(s - 1)] and, by suit- ably regulating the speed of the water current, the larger and heavier particles may be made to fall downward against the current, leaving the smaller, lighter particles in suspension in the current. 2. Particles such that [- = -] cannot be separated in an upward- current classifier with any one velocity of current. Coal particles having a specific gravity of 1.35 cannot be separated from refuse particles of [2.5 - 1 2.5 sp. gr.] unless the ratio of diameters is less than [ 4.3 to 1, 1.35 – 1] and the raw coal must be screened to size limits within this ratio if they are to be separated. Cone-shaped vessels or vessels of other shapes that allow the use of different velocities of current at different points in the flow are used to modify this limitation.

Jan 1, 1943

By J. R. Mihalisin

Phase changes in as-cast alloys in the Ni2 CoCr-Ni, Ti-NSAl system containing 0.1 wt pct C were studied after heat treatment at 1650°F. It was found that a vegion of CrCo-type s was developed in this system after I1000 hr. The tendency for s formation was dePendent upon the aluminum and litanium content, the s -forming tendency generally becoming more pro-nounced as the latter two elements are increased. s lended to form in interdendritic regioues and was associated with solutioning- of- carbide phases. Electron- vacancy computations for prediction of s formation on these alloys mere in sufficient agreement with what was observed experimentally to indicate that the method should be of value in assessing a-forming potential in cast alloys. However, the correlation of electrotz- vacancy numbers may not be as favorable in cast as cormpared to wrought alloys. THE increasing use of high-temperature alloys for longer times at higher temperatures has resulted in a real need for obtaining information on the phase changes taking place in these alloys since these changes often have a direct effect on properties. This paper reports on a study of the phase changes in Ni-Cr-Co-Al-Ti alloys after prolonged heating at 1650°F. The alloys were selected from the pseudo-ternary system Ni2CoCr-Ni3Ti-Ni3Al and this system was chosen because it is the basis of many of the newer cast commercial superalloys. The attempt here was not to define accurately the boundaries of s formation in the NizCoCr-Ni3Ti-Ni3Al system but to delineate the phase transformations taking place upon heating alloys covering a wide range of chemistries in this system. An addition of carbon was made to all alloys to further simulate commercial alloys. It was hoped that the information generated in this manner would be of more direct use to superalloy designers than an equilibrium diagram approach since commercial alloys generally have outlived their usefulness when equilibrium is attained. EXPERIMENTAL PROCEDURE The alloys were vacuum-melted from carbonyl-nickel pellet, vacuum-grade chromium, electrolytic cobalt, high-purity aluminum and titanium, and spec-trographic-grade carbon.

Jan 1, 1968

By E. R. Dean

A wire-feed mechanism has been employed to fabricute metal alloy film resistors to various sheet resistivities on oxidized silicon substrates. The effect of several thousand hours storage in air at elevated temperatures on the resistance and temperature coefficient of resistance is presented. Unprotected films of- sheet resistivity between 100 and 500 ohms per sq fabricated at 300°C substrate tenzperatures were unstable when stored at 150°Cfor extended periods. The higher sheet resistivity films exhibited the greatest instability; however, even the 100 ohms per sq films drijted excessively for device application. When stored at still higher temperatures, the normalized resistance increases to maximum, then decreases until a minimum value is obtained, then finally increases in resistance until open. The use of a protective overcoating of SiO has had considerable benejicial effects on the film stability, so that 250 ohms per sq films deposited on 300°C substrates are now stable after 1000 hr storage at 250°C and possess a temperature coefficient of resistance less than 200 ppm per C. The use of a low substrate temperature during depositon (100°C) enables the preparation oj resistors with very low tenperature coefficients of resistance (10 to 20 ppnl per 'C). However, these films are less stable than their higher substrate temperatuve counterparts. During extended storage, the resistance of the protected films always decreases with lower substrate films exhibiting larger normalized resistance decreases. This resistance decrease is accompanied by a linearly related increase in the temperature coefficient of resistance. The electrical behavior of these films may be explained by postulating That the structure of the films is in the transition region between thick continuous films and ultrathin island structure films so that the conductivity is the restlt of both electron scattering and tunneling-activated charge carrier creation between neighboring grains. The annealing behavior when thermally aged is the result of defect anneal, grain growth, and selective oxidation. TheRE is a considerable interest in thin metallic films for use as resistor elements in microelectronic circuits.1,2 These resistor films must be stable when exposed to elevated-temperature storage or operating ambient, possess low temperature coefficients of resistance (hereafter referred to as T.C.R.), and be of a high enough sheet resistivity to be useful. The resistivity and T.C.R. of a thin metallic film are determined by the structure and thickness of the film. Hence the stability of the film when exposed to stress would depend on the stability of the structure. The resistivity of a thin film is the result of the electrons' interaction with the lattice vibrations (electron-phonon scattering), scattering of electrons due to impurities, defects, and grain boundaries and specula reflection from the film surfaces.3' All of these effects serve to reduce the electron mean free path and result in a resistivity higher than the ideal bulk. In ultrathin films possessing essentially an island structure consisting of a planar array of aggregates separated by a few to a few tens of angstroms, conduction is dominated by a combination of tunneling and activated charge carrier reation. Thin films would be characterized by relatively low resistivity, positive T.C.R., and reasonable structure stability whereas ultrathin films possess unstable structures, negative T.C.R., and exhibit high resistivity. Feldman6 has investigated films intermediate between the two regions and postulates that the resistivity of a film in the transition region is composed of two linearly additive parts: that due to the resistance of the grains and that due to the gaps. The grain resistivity would represent the scattering of the conduction electrons by defects, surfaces, and phonon interaction while the gap term represents the tunneling contribution. For gold and platinum films, he found as the film structure becomes progressively less continuous, corresponding to thinner films, the absolute value of the T.C.R. becomes progressively smaller until negative values are observed with respective zero T.C.R.'s at about 35 and 150 ohm per sq, respectively. Recently the author7 investigated the behavior of a multicomponent alloy of nickel, chromium, and iron and found the T.C.R.'s of these films decrease with decreasing substrate temperature during deposition. The lower T.C.A. associated with lower substrate temperatures was attributed to an increased contribution to the total resistivity from the negatively temperature-dependent gap tunneling since it is well-known that thin films deposited on low-temperature substrates consist of a higher density of smaller grains than the same sheet resistivity films deposited on higher-temperature substrates. Under conditions of thermal anneal, the author found that, when various sheet resistivity films fabricated at 300°C substrate temperature are exposed to extended storage in air at 300°C, the thinner films increase in resistance until open, while the thicker films

Jan 1, 1967

By F. Keller, J. D. Edwards

THE anodic treatment of aluminum presents problems of scientific as well as of commercial interest.1-3 Of particular interest is the fact that, during the anodic oxidation process, the oxide continues to form at the metal-oxide interface under any oxide previously formed. This has led to speculation as to the mechanism involved in the formation of the relatively thick oxide coatings that are used commercially for decorative or protective purposes. Furthermore, the ability of certain types of oxide coatings to adsorb dyes and other substances has stimulated research to determine the actual structure of these adsorptive coatings. It has been found that anodic oxide coatings on aluminum are composed essentially of aluminum oxide. They are formed by the action of oxygen ions penetrating to the metal surface during the electrolytic oxidation treatment. These coatings can be formed in a number of different electrolytes, as for example, those which contain sulphuric acid, chromic acid, oxalic acid, or boric acid. These are the commonest and most useful electrolytes employed commercially. Except for certain specific conditions of formation, the coatings in general have been found to be amorphous alumina, as far as can be determined by X-ray or electron-diffraction methods. Anodic oxidation processes can be arranged in three rather general classes if they are grouped in relation to the solvent action of the electrolyte on the coating.4 In the first class, the electrolyte has little or no solvent action on the coating that is formed. In general, coatings produced under such circumstances are nonporous and non- adsorptive. In the second class, the electrolyte exerts an appreciable solvent action on the coating. These coatings are porous and adsorptive. Finally, for the third class, the electrolyte tends to dissolve the coating about as rapidly as it is formed. This action produces electrolytic brightening or anodic polishing of the aluminum surface and at most leaves only a very thin film of oxide. The nonporous and nonadsorptive type of oxide film is represented by the coatings that are formed when solutions of boric acid are employed as the electrolyte. It is especially significant that these impervious films are formed in electrolytes that exert little or no solvent action on the coating. Where a boric acid electrolyte is used, the coating tends to form rapidly, with the result that the flow of current is soon reduced to substantially zero. This indicates that the growth of the coating has stopped. As a rule, the thickness of the coating is roughly proportional to the voltage employed for formation and the coatings of this type are exceedingly thin. In an electrolyte of this type, the coating has a high resistance when the aluminum is made anode; consequently, any current that may flow (leakage current) is very

Jan 1, 1944

By S. R. Seagle, R. Schuhmann, N. A. Parlee

Rates of CO evolution and CO absorption were measured for liquid-iron alloys containing from 0.15 to 4.4 pet C, using a modified Sieverts apparatus. The alloys were held in alumina crucibles, so that both crucible-metal and gas-metal reactions occurred simultaneously. The data are interpreted on the hypothesis that the C-O reaction rate is controlled by oxygen diffusion across the boundary layers at the gas-metal and crucible-metal surfaces. CARBON-monoxide evolution from liquid Fe-C-0 alloys is a key reaction in steelmaking processes, both in the steelmaking furnace and in the ingot mold. Also, this reaction probably is an essential step in the desulfurization of iron under blast-furnace conditions, for which carbon is the principal reducing agent or deoxidizer present in the iron. The thermodynamic properties of the solutions of C and 0 in Fe and the equilibria of these solutions with gaseous CO and CO, have been investigated in some detail and are reasonably well understood. However, while these equilibrium data have accumulated in the laboratory, other data have accumulated to show that equilibrium conditions are often not achieved under plant operating conditions. Thus, an understanding of the rate and mechanism of the carbon-oxygen reaction not only has theoretical interest, but ultimately may assume considerable practical importance. Reaction Mechanisms and Rate Equations The over-all reaction under consideration is C (in liquid Fe) + 0 (in liquid Fe) + CO (gas) [I] Although this appears to be a straightforward heterogeneous reaction, involving just two phases (liquid metal and gas), a wide variety of reaction mechanisms has been proposed. Different authors have expressed divergent views as to the nature of the rate-controlling steps. Accordingly, a brief discussion of the possible reaction steps will be given. 1) Homogeneous Reaction Within Liquid Iron— Feild' and Jette2 have applied chemical reaction-rate theory as if the C-0 reaction were a homogeneous, second-order reaction occurring within the liquid-metal phase. This mechanism necessarily produces CO "molecules" dissolved in liquid iron. In order to yield CO gas, the dissolved CO must nucleate CO gas bubbles, and the dissolved CO also must be transported to the bubble surface where it enters the gas phase. However, this mechanism is inconsistent with present concepts of the structure of liquid-metal solutions. 2) Reaction at Gas-Metal Interface—At the surface of contact between gas and metal, a second-order reaction between dissolved C and 0 may occur to produce a CO molecule which enters the gas phase. This reaction presumably involves an "activated complex" structure in the interface. Some such surface mechanism appears essential because the carbon and oxygen do not occur in the same species in the two phases and therefore must react in some way at the surface. At liquid-iron temperatures, calculations based on reaction-rate theory indicate that the surface reaction must proceed extremely rapidly and thus cannot be a rate-determining step in the C-O reaction under ordinary conditions.3 ccordingly when the over-all reaction is measurably slow, the surface reaction may be considered at equilibrium; that is Cc* Co* = m'Pco [2] in which Co* and Co* are the concentrations of carbon and oxygen (weight per unit volume of metal), respectively, in the liquid-metal phase at the interface, Poo* is partial pressure of CO in the gas phase at the surface (atmospheres), and m' is the mass-action constant for Reaction [I]. Since activities of C and 0 are not strictly proportional to concentrations, m' is not a true thermodynamic equilibrium constant but varies with carbon content.' 3) Mass Transport—If the reaction is to proceed at the gas-metal surface, as described above, dissolved carbon and oxygen in the liquid iron must come to the surface and gaseous carbon monoxide must move away from the surface. Details of the mass-transport mechanisms depend on the kind of system under consideration. For a stirred metal bath in contact with a CO gas phase, the transport of C and 0 to the surface can be described by 1 Do Mols CO evolved per sec = — 1/16— Do/do Agm (CO-Co*) = 1/12 Do/do Agm *(Co-Co*) [3] In this equation, Do and D, are the diffusion constants for oxygen and carbon in liquid iron, 6, and 6,. are the boundary-layer thicknesses, A,,,, is the

Jan 1, 1959

By H. H. Kellogg, T. R. Ingraham

Three anhydrous zinc sulfates have been identified. They are: ZnSO,(a), stable below 1007°K; ZnS04(/3), stable above 1007OK; and ZnO.ZZnSO,. The decomposition pressure of each sulfate has been measured and the following relations calculated from the results: The measured data have been combined with the known themzodynamic properties of ZnS, ZnO, Zn, SO,, and SO, to yield univariant and bivariant phase diagrams for the system Zn-S-0. The application of these diagrams to problems of roasting zinc concentrates is discussed. PRESENT knowledge of the equilibrium chemistry of the system Zn-S:O, so essential to scientific control of zinc-blende roasting, is deficient in several respects. A basic sulfate of zinc has been reported by many investigators, but its composition is in doubt and its thermodynamic properties have not been measured. Our scanty knowledge of the properties of the normal sulfate at roasting temperatures is based on old and inexact studies. A crystal transformation in the normal sulfate, involving a sizable heat effect, has been noted by several investigators but has never been quantified. The studies reported herein were designed to fill these gaps in knowledge and to make possible a reasonably complete thermodynamic description of the system Zn-S-0 at roasting temperatures. The results make possible a far more exact understanding of zinc roasting chemistry than was hitherto possible. EXPERIMENTAL The principal experimental technique employed was the measurement of the total gas pressure developed from decomposition of either ZnSO, or ZnO2ZnS0, in a closed system. The apparatus employed was essentially the same as that described by Warner and lngrahaml and Flengas and tngraham.' The unusual feature is a flexible Pyrex bellows used to separate the reaction gas mixture (S03, SOz, and 02) from the mercury in the manometer and thus prevent corrosion of the mercury by SO3. The bellows operates without evidence of hysteresis and, barring accidents, may be used for several months. The response of the bellows varies from manometer to manometer, but is generally of the order of 0.6-cm displacement for a l-cm change in pressure. The response for all manometers has been linear over the range 0 to 1 atm. To prevent condensation and polymerization of sulfur trioxide, the bellows-manometer was maintained at 100.0" * 0.5" by a circulating bath of silicone oil. The connecting tubing between the manometer and the furnace was maintained at 100.0" * 0.2"* also, with a stirred air bath. The open end of the manometer was connected with a chamber which was maintained at a known pressure. The pressure could be varied so that readings of the height of the mercury column, measured with a cathetometer, could be made with both rising and falling mercury column. By using the average of the readings, any error due to sticking of the mercury column was minimized. The sample was placed in a 2-in.-long platinum boat formed into the shape of a half-cylinder. The sample was prepared as a fine powder and pressed against the inner wall of the platinum boat. A silica thermocouple well, covered with a platinum sleeve, was placed lengthwise along the center of the boat. The temperature of the sample was read on a calibrated Pt-10 pct Rh theirnocouple. The temperature profile of the furnace was constant within ± 0.2 C over the length of furnace occupied by the sample. The horizontal furnace used to contain the sample and the quartz reaction tube was electrically heated, with the control thermocouple placed 1/4 in. from the windings. A Brown Pyro-O-Vane temperature controller with a 3-sec response time was used to maintain the furnace temperature. Before beginning a run, the apparatus, including the bellows manometer, was thoroughly washed in sib with acid and distilled water to remove any

Jan 1, 1963

By A. W. McReynolds

One characteristic of plastic deformation which distinguishes it from elastic strain is the essential inhomo-geneity of plastic strains. Elastic strain varies continuously through a material, and average relative displacements of initially adjacent atoms are only small fractions of their initial spacing, (strains of the order of 0.01 or less). On the other hand, plastic flow corresponds to the appearance of discontinuities in strain of the lattice, such as dislocations or slip bands, where local strain, on an atomic scale, is several orders of magnitude higher. These discontinuities are visible on a microscopic scale as the familiar slip lines (Fig 1). In spite of this obvious microscopic inhomogeneity, however, macroscopic measurements almost invariably show a smooth curve of stress vs. strain (Fig 2b) even if measurements of linear strains be made to an accuracy of one part in 107. This macroscopic homogeneity of strain indicates that the discontinuities in strain on slip planes occur in increments too small or too slow to be recorded individually, and further that they occur sufficiently independent of one another so that the small increments add at random to a smooth stress-strain curve. The present paper describes observations of plastic strain in aluminum of commercial purity and in high purity Al-Cu alloys, where there exists a strong coupling between slip in various regions of the specimen such that once initiated it spreads rapidly through a large volume. The total effect is that of relatively large, rapid, and regularly spaced steps of strain followed by periods of only elastic strain. Fig 2a illustrates the type of "stair-step" stress-strain curve which results. The properties of this cooperative slip phenomenon will be described further in the section on results: in par- ticular it will be shown that each step corresponds to the propagation of a wave of plastic deformation through the specimen. Some interpretations of the mechanism by which it occurs will be made in the following section. Although the type of plastic wave phenomena to be described has not previously been reported, there are numerous cases of related effects in the plastic yielding of metals: YIELD POINT PHENOMENA The most familiar of such effects is the "yield point" observed in low carbon steels, brass, duralurninum, and the like. It consists in the sudden termination of the elastic portion of the stress-strain curve by a large plastic strain. Since the usual tensile machine is such that yielding of the specimen relieves the load, the resulting curve is as shown in Fig 3. As the strain continues, deformation occurs at a lower stress for some time, then follows a rising curve, but with no further sudden yielding. This effect has been observed in brass by Sachs and Shojil and later by many others. Edwards, Phillips and Jones2 made extensive studies of the effect in steel, and of the role of various alloying elements. Although there seems to be fairly clear evidence that the yield point is caused by a hardening of the material by precipitation of impurities, no satisfactory explanation for the sudden yielding has been given. Winlock and Leiter3 have shown that the strain. level of the upper yield point depends strongly on the rate of loading, the yield point increasing by almost a fac- tor of two as the strain rate goes from 0.002 in. per in. per min. to 4.4 in. per in. per min. This effect would seem to imply an incubation period before yielding is initiated at a certain stress. On the other hand, by going to very slow loading rates, Edwards, Phillips and Jones2 showed that the yield point does not become lower and eventually disappear as might be expected, but, on the contrary, begins to rise at loading rates below about 25 Ib per in. per min. becoming much higher than at rapid loading rates. STRAIN AGING If, instead of continuing straining of a specimen after occurrence of a yield point, the load is removed and the specimen aged, resumption of the test results in occurrence of another yield point as shown by the dotted curve of Fig 3. The new yield stress is generally higher than the previous maximum applied stress. This hardening of the material by straining and subsequent aging is undoubtedly related to quench age-hardening resulting from the aging of a specimen quenched from high temperature. Since neither effect is observed in pure metals, it is generally accepted that quench-aging in all cases is the result of hardening by precipitation of a supersaturated alloying element, and that strain-aging is probably a similar precipitation, accelerated by disruptions of the lattice by previous strain. Pfeil4 has shown that strain-aging does not occur in iron from which all of the carbon has been removed, but that only a very small carbon content, around 0.003 pct, is necessary to cause strain-aging. In accord with this observation is recent work by Dijkstra5 in this laboratory showing that the solubility limit of carbon in iron is extremely low, less than 0.001 pct at 400°C. Edwards, Phillips and Jones2 have shown that the strain-aging effect is also removed by the addition of small quantities of elements such as Mo, Mn, Ti, and the like, which readily form carbides. Their results demonstrate the

Jan 1, 1950

By N. H. Emmons

An interesting development of copper blast-furnace construction has been brought about in adapting the blast-furnace to be a "burner" for sulphuric acid making. When the Tennessee Copper Co. first decided to construct an acid-plant, the standard type of brick-top furnace, supported by structural steel, was the only one in use at its plant. All the tests for temperature and strength of gas had been made on these furnaces, and had been satisfactory. When, however, the acid-plant was put on the flue, and the flue had to have a damper put in it to force the gas into that plant, troubles began. The furnaces were fairly tight, and for a few months the work was quite satisfactory. Fig. 1 shows the old type of furnace-top used at the start. It was soon found that this top would not stand the temperatures, particularly when dampered back, as was necessary to force the gas into the Glover towers. The structural steel, of which the furnace skeleton was constructed, warped and twisted so badly that it was not safe to use the furnaccs. This effect is shown by Fig. 2, which is a photograph of old No. 4 furnace, after the brick-work had been removed. It became apparent that a new type of top mas necessary and the engineering force was started on the design of a new furnace-top, with the result that a low top, with brick-lined flues at each end below the feed-floor, leading to the main concrete dust-chamber, was constructed. This furnace, No. 7, is shown in Fig. 3. The charging-doors are made of cast-iron sections, three sections to a door, and three doors to a side. A door to the furnace was made of east copper, but the heat was great that the copper bent readily, due to the jarring of its

Jan 1, 1911

By D. Kuhlmann-Wilsdorf

A new theory of work hardening is developed which rests on only a few simple principles and is applicable to a wide variety of materials and dislocation structures. It explains, qualitatively, the general characteristics of the three-stage shear stress-shear strain curve of fcc metals, as well as the most important features connected with easy glide. The value of, the work-hardening coefficient in stage 11, is derived quantitatively, from first principles, and is shown to be insensitive to many changes in detailed dislocation behavior. The low rate at which energy is stored during mechanical working in stage 11, the dependence of slip line length on stress, and the so-called Cottrell-Stokes law are explained. Although the theory is primarily developed for fcc metals and alloys, it is largely applicable also to some other materiuls, and in particular apparently to poly crystalline simple steels in theh linear range of hardening. FOR pure fcc metals and alloys, after many different pretreatments and under a great variety of testing conditions, the well-known three-stage work-hardening curve is commonly observed. Moreover, the work-hardening coefficient in stage 11, 011, bears an almost constant relationship to G, the modulus of rigidity, such that K = G/Brr" = 300, within a factor of about 2 either way. The very fact that the three-stage curve and the value of K are so very persistent, can hardly be understood except on the assumption that some quite simple phenomena are responsible. In this light all available theories of work hardening must be judged unsatisfactory: Each of them is based on some quite specific dislocation model, while it is known that metals with similar values of K may have widely different dislocation structures, for example, dislocation tangles in the case of pure fcc metals, and piled-up groups of dislocations in a-brasstype alloys. Thus, while a particular theory might conceivably represent a special case with a reasonable degree of accuracy, it cannot possibly have illuminated the underlying principle. Another severe criticism applying to the presently most widely discussed theories is that they employ empirical or even quite unexplained parameters in order to arrive at a numerical value of K. This indicates that some vital aspect of the work-hardening process has remained unexplained. In the present paper, a new theory of work-hardening is developed which does not suffer from the above shortcomings. It is derived from first principles, and is founded on the realization that the underlying causes for the three-stage work-harden- ing curve and for the persistence of the numerical value of O,, must be few and simple. As a first step it is shown herein that the experimental oklservations on "easy glide" can be explained on the assumption that in this stage dislocations multiply, beginning in a number of restricted area:;, and from there penetrate into crystal regions still substantially free of glide dislocations until a quasi-uniform dislocation distribution has been established. During "easy glide", the resistance to the movement of the foremost dislocations, advancing into still largely a dislocation-free regions, is believed to control the flow stress. This resistance is determined by various factors, among them the reaction of the dislocations against bowing-out into loops, which is due to their line tension. Without such bowing-out no dislocation multiplication would occur, but the corresponding contribution to the flow stress stays practically constant all through stage I. According to the present theory, stage TI begins as soon as there are no more areas left into which newly formed dislocations have not yet penetrated, i.e., as soon :is a quasi-uniform dislocation density has been attaned. Regardless of how, in detail, the dislocations should happen to be arranged at the end of "easy glide", whether in pileups or tangles, on one glide system or on several, the stress necessary to overcome the line tension of the segments must now increase, because the dislocation density increiises while the average free dislocation length decreases. It is argued that all other contributions to the flow stress stay at about the same level as in stage I, and that the said increase in the stress necessary to bow out dislocation segments into loops is responsible for the major part of work hardening in stage 11. Stage TII of the stress-strain curve is explained by largely follouing the ideas of Seeger and his co-workers; however, it is pointed out that the start of stage m may on occasion indicate the onset of "conservative climb" instead of cross slip. Indirect evidence indicates that linear work hardening persists even under the profuse action of cross slip, but at a lower rate. Only when, in addition to cross slip, climb operates extensively, does the work-hardening rat,. drop sharply, ultimately down to zero. BRIEF SURVEY OF EXPERIMENTAL EVIDENCE The stress-strain curve of crystals of many different substances are qualitatively similar. No significant permanent plastic deformation takes place below the so-called "critical resolved shear stress". As the applied stress is raised beyond this level. yielding occurs with little or no work hardening and

Jan 1, 1962

By E. W. Dewing

The nature of solid solutions of sodium in non-paphitic carbon at temperatures near 1000°C has been investigated by an electrolytic technique. The activity coefficient is found to vary strongly with the heat-treatment temperature of the carbon, and an analysis of the energy terms involved suggests that this is due to variations in the Fermi level. The diffusion coefficient for sodium is in the range implying that the sodium atoms (or ions) are not bound to specific sites on the carbon, in agreement with lack of evidence for specific compounds in the activity -composition curves. In the light of These results phenomena occurring in aluminum reduction cell linings are dis-cussed. It is concluded that sodium migrates by diffusion through the carbon lallice and not the pores, and that A14C9 is formed within the lining hy The excess of NaF invariably found in used linings is formed by this reaction and by air oxidation: There has recently been much interest in the swelling of carbon which occurs when it is made cathodic in an aluminum electrolysis cell, and there is general agreement that the phenomenon is due to the penetration of sodium atoms, derived from the reaction on the carbon lattice. (NaF and AlF3 are constituents of the fused-salt electrolyte.) The low-temperature lamellar compounds of potassium with graphite are well-known, and one compound between sodium and graphite has been reported.4 There is, however, virtually no information on the basic chemistry of the interaction of sodium with nongraphitic carbon at high temperatures, although Dell3 has made some fundamental studies of rates of penetration. At the time the present work was carried out these were not available. Rapoport and samoilenkol had measured rates of physical expansion in a dilatometer, using the fractional increase in length of a carbon specimen after electrolysis in a cryolite bath for 2 hr as an index of susceptibility to cathodic swelling. Work here with a similar apparatus showed that expansion (of a 1.5-in.-diam cylindrical specimen) ceased in less than 20 hr, and that the final expansion was much more reproducible than the value at some intermediate time. This suggested that an equilibrium was being set up, and that a thermodynamic approach was required. The results obtained are not as accurate as had been hoped, but they are the only ones of their type available and have enabled useful conclusions to be drawn. The experimental work was divided into two parts: an investigation of the activity-concentration diagram for sodium in carbon, and a rough determination of the diffusion rate. For the former, most of the experiments were made with petroleum coke even though this is not a form of carbon used in cell linings. It has the advantages, however, that it is a homogeneous material, that it is low in ash, and that its structure is reasonably well understood. Anthracites and coal cokes gave results of the same type as obtained with petroleum coke, but the reprodu-cibility from specimen to specimen was extremely poor. EXPERIMENTAL 1) The Activity-Composition Diagram for Sodium in Carbon. The petroleum coke was ground, pressed at 40,000 psi into plates 19 by 9 by 2 mm without added binder, and heat-treated. The sodium activities were established by an electrolytic technique. The carbon was made anodic in a cryolite bath with an aluminum cathode, and a controlled potential, E, applied so that the sodium activity at the anode was given by ln Na = 1n asa - EF/RT [21 where as, is the sodium activity generated by Reaction [I]. The relative activity is sufficient for many purposes and is simply given by The value of Na not at present known, but when the thermodynamics of the Na-A1 system have been worked out it can be derived from the equilibrium

Jan 1, 1963