Search Documents

By C. J. Frosch, J. A. May, H. G. White, C. D. Thurmond

Gallium phosphide epitaxial layers were grown from the vapor phase on undoped single-crystal galliurn arsenide substrates in silica tubes by an open-tube wet-hydrogen process. The epitaxial layers were grown over a range of water pressures at three substrate temperatures. Excess donor concentrations were determined by surface barvier capacitance measurelrzents without removing the layers from the substrates. The excess dmlor concentration, ND-NA, is fo~ind to vary approxilnately inversely with the pressure of water added to the hydrogen carrier gas. This is the relationship that would be expected for singly ionized silicon donors on gallium sites in extrinsic galliunz phosphide, with the silicon coming from the SiO generated by the reaction of hydrogen with the silica tube. An increase in the partial pressure of water in the hydrogen stream decreases the SiO pressure. The results indicate that ni, the intrinsic hole and electrmt concentration for gallium phosphide at the three substrate temperatures, is smaller than the concentration estimated from available data for the density of states effective masses and the energy gap. Mass-spectrographic measurements confirm that the dono?, introduced into gallium phosphide is silicon. The equilibrium concentrations of silicon in vapor-flown gallium Phosphide have been estimated from available thernzodynamic information that includes the solubility measurements of silicon in gallium phosphide in equilibrium with a gallium-rich liquid phase. Satisfactory agreement with the measured silicon concentrations is obtained. FROSCH1 has described an open-tube process for growing single-crystal Gap from the vapor phase by a GazO transport mechanism. The method depends upon the reaction of H20 in an H2 carrier gas with a heated source of polycrystalline Gap which provides the necessary vapor species. When the temperature of these vapor species is lowered, super saturation occurs and single-crystal Gap will deposit on a suitable substrate. Unintentionally doped single crystals of Gap grown by the wet H2 process in silica tubes are n type. Evidence is presented to show that the donor introduced is silicon, and that a qua si-equilibrium model accounts for the inverse dependence of the donor concentration on the water partial pressure and predicts the magnitude of the donor concentrations. Ainslie et al. experimentally showed a similar inverse relationship between the carrier density and oxygen pressure for GaAs. Emission-spectrographic analyses showed a decrease in the silicon concentration with increasing oxygen overpressure for GaAs. Cochran and Foster suggested the theoretical possibility of suppressing silicon contamination by using Ga20 generated by the reaction of gallium with water vapor. 1) EXPERIMENTAL The apparatus and procedures are essentially the same as those described by Frosch.' The apparatus consists of a 25-mm-ID SiO2 tube extending through a controlled high-temperature flat zone for the location of the polycrystalline Gap source and a downstream temperature gradient falling at a rate of about 14°C per cm. The latter provides the region of super saturation for the location of the single-crystal substrate. The partial pressure of water in the inflowing hydrogen stream, pA2, O was controlled by mixing me-tered proportions of dry H2 with H2 saturated with H2O vapor at 0°C. The total gas flows were about 200 cu cm per min in all experiments. The Gap sources were prepared by pulverizing boat-grown polycrystalline ingots to pass a 20-mesh sieve. The substrates were cut from an undoped single-crystal boat-grown GaAs ingot purchased from Monsanto. This ingot had a carrier concentration of about 1015 atoms per cu cm, a resistivity of about 5 ohm-cm, and a mobility of about 5000 sq cm per v sec at 25°C. Substrates with dimensions of 1 by 1 by 5 x lo-' cm were employed. The growth faces were chemically polished (111) arsenic faces. Epitaxial layers, at least 7.5 x 10-3 cm thick, were grown,. This required from 1 to 24 hr depending upon the Pii2Q values and the temperatures. In all of the runs, the source temperatures were 50°C higher than the substrate temperatures. Samples were prepared

Jan 1, 1968

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

THIS discussion is based on statistical manipulation and evaluation of operating data from several commercial blast furnaces which include rather wide variations in practice. We are concerned here mainly with the effect of manganese on sulphur elimination under operating conditions in commercial blast furnaces, but this effect can be outlined more clearly by including certain other variables such as silicon in iron and slag basicity, which are all in part interrelated. Real progress in understanding all of the smelting-zone reactions in this process requires more laboratory studies of slag-metal melts similar to those in the blast furnace, with careful control of oxygen pressure. Such experimental work is very difficult, however, and in the meantime, with the present scarcity of more fundamental data, such statistical trends and correlations as can be obtained from commercial furnace operating records may help in choosing laboratory test conditions as well as adding to our knowledge of operating procedures. We do not know just why the furnace data are apparently so irregular that statistical treatment of large numbers is needed to obtain accurate trends. It is partly caused, no doubt, by the many simultaneous variables present, but there are perhaps also such factors as: (I) the con- tinual swings up and down so characteristic of the process; (2) difficulty of accurate sampling of co-existing phases, as well as (3) the presence of both equilibrium and rate factors as a result of the fast driving of the process. In desulphurization, for example, the flow of thin layers of slag and metal over solid coke surfaces, plus the contact between droplets of iron falling through the slag layer in the hearth, would lead one to expect a close approach to equilibrium. On the other hand, most of the critical part of sulphur removal may occur in a short time and distance below the tuyeres with perhaps an accompanying decrease in oxygen pressure and the concurrent reduction of much of the silicon and manganese, so its degree of approach to equilibrium may be quite variable. Probable Chemistry of Desulphurization Since there is little evidence that sulphur atoms in the iron solution are aasociated to any appreciable extent in molecules such as MnS, the transfer from slag to metal and back again can be written:* (FeS) ? [FeS], [I] hence the following reactions seem most probable in the slag phase, (FeS) + (CaO) ? (CaS) + (FeO), [2] and, (FeS) + (MnO) ? (MnS) + (FeO). 131

Jan 1, 1949

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

THIS discussion is based on statistical manipulation and evaluation of operating data from several commercial blast furnaces which include rather wide variations in practice. We are concerned here mainly with the effect of manganese on sulphur elimination under operating conditions in commercial blast furnaces, but this effect can be outlined more clearly by including certain other variables such as silicon in iron and slag basicity, which are all in part interrelated. Real progress in understanding all of the smelting-zone reactions in this process requires more laboratory studies of slag-metal melts similar to those in the blast furnace, with careful control of oxygen pressure. Such experimental work is very difficult, however, and in the meantime, with the present scarcity of more fundamental data, such statistical trends and correlations as can be obtained from commercial furnace operating records may help in choosing laboratory test conditions as well as adding to our knowledge of operating procedures. We do not know just why the furnace data are apparently so irregular that statistical treatment of large numbers is needed to obtain accurate trends. It is partly caused, no doubt, by the many simultaneous variables present, but there are perhaps also such factors as: (I) the con- tinual swings up and down so characteristic of the process; (2) difficulty of accurate sampling of co-existing phases, as well as (3) the presence of both equilibrium and rate factors as a result of the fast driving of the process. In desulphurization, for example, the flow of thin layers of slag and metal over solid coke surfaces, plus the contact between droplets of iron falling through the slag layer in the hearth, would lead one to expect a close approach to equilibrium. On the other hand, most of the critical part of sulphur removal may occur in a short time and distance below the tuyeres with perhaps an accompanying decrease in oxygen pressure and the concurrent reduction of much of the silicon and manganese, so its degree of approach to equilibrium may be quite variable. Probable Chemistry of Desulphurization Since there is little evidence that sulphur atoms in the iron solution are aasociated to any appreciable extent in molecules such as MnS, the transfer from slag to metal and back again can be written:* (FeS) ? [FeS], [I] hence the following reactions seem most probable in the slag phase, (FeS) + (CaO) ? (CaS) + (FeO), [2] and, (FeS) + (MnO) ? (MnS) + (FeO). 131

Jan 1, 1949

By Charles H. Grant

Some of the limiting factors relative to explosive crushing of rock and ways to overcome a few of these problems are presented. Relationships between borehole diameters, bench heights, and spacings, along with a review of the influence geometry has on energy as these are changed, are discussed. Efficiency in use of explosives and the decay of energy as it moves through rock and is absorbed and dissipated, is described, along with fragmentation as a function of spacings and energy zoning, etc. Communications are one of the major problems encountered. In an effort to provide a better understanding of the use of explosives, it is necessary to take a little different view of what explosives are, how to look at them as tools to fragment rock, and some of the problems encountered in doing so. First, take the explosive: although there are many factors involved, consider these as being reduced to only two — shock-strain imparted to the rock by the high early development of energy, and the gas effect which is a combination of heat, moles of gas formed, rate of formation of these gases which develop pressures, etc. First, consider shock energy by itself and assume there is no gas effect in the reaction. Fig. 1 illustrates a block or cube of rock, in the center of which is detonated an explosive charge which is 100% shock energy. Tensile slabbing would be seen on the surface and probably the cube of rock would generally hang together even though microcracks were formed. If the situation is reversed and an explosive whch has no shock energy and only gas effect (Fig. 2) is considered, the cube of rock would act as a pressure vessel and contain the pressure from the gas effect until it exceeded the rock-vessel strength; then the rock would break in a few large pieces. If these two kinds of energy are put together and the area of shock-strain around the explosive (Fig. 3) is considered, the two energies will be seen working together to furnish broken rock. The gas effect applies pressure to the microcracks formed from the shock energy to weaken the rock-pressure vessel and propagate these cracks to break the rock apart. It not only will be broken more finely, but will break apart at a lower pressure than the gaseffect case, since the shock energy has first weakened the rock vessel. Although tensile spalling from the shock-strain imparts momentum to the rock, the main source of displacement comes from the gas effect. The term "rock" is being used to mean any material to be blasted. These energies are absorbed by the rock in different ways. First, classify rock into two main categories: "elastic" and "plastic-acting." Elastic rock should be thought of as rock which can transmit a shock wave and is high in compressive strength, such as granite or quartzite. Since this elastic rock transmits a shock wave well, it makes good use of the shock energy from the explosive-forming cracks, etc., for the gas effect to work on. Plastic-acting rocks are rock masses which are relatively low in compressive strength and absorb shock energy at a much faster rate, thereby making poor use of the shock energy by not developing as extensive a cracked zone for the gas effect to work on. Rocks of this type are generally softer materials such as some limestones, sandstones, and porphyries. For the most part, the shockenergy part of the explosive reaction is wasted in plastic-acting rock, leaving most of the work to the gas effect. Since the ratio of gas effect to shock energy is different in different explosives, it is easy to understand why some explosives perform well in elastic rock and poorly in plastic-acting rock, and vice versa. Some of the most difficult blasting situations arise when mixtures of plastic-acting and elastic rock are encountered (Fig. 4). Fig. 4 shows an example of granite boulders cemented together with something like a decomposed quartz monzonite which is plastic-acting. The elastic granite boulders will transmit the shock-strain within itself, but when this shock tries to move through the monzonite to the next boulder, its intensity is absorbed by the monzonite and little shock-strain is placed on the adjoining boulder. In addition to this loss by absorbtion, shock reflection at the surface of the boulder will effect tensile spalling. The net effect is poor breakage of the boulders which do not have drillholes in them as they simply will be popped out with the muck. The same is true (Fig. 5) when layers and joints make

Jan 1, 1970

By R. O. Day, R. S. Dean

In constructing a theory of the flow and hardening of metals, v necessarily make use of such phenomena as seem to be universal1 observed in metals. It is, therefore, a matter of concern to the con structors of such mnemonic systems if one or two of the group classific as metals show anomalous behavior. Such egregious phenomena observed in pure or nearly pure iron in the reaction of its mechanic properties to low-temperature annealing after cold work and to increase temperature of testing. In short it has been generally observcd tha iron, as an outstanding exception among metals, increases its hardnes and strength by low-temperature annealing after cold work,' and alt by increase of testing temperature to the range of 150" to 300" C. The investigation here recorded was made then with the object ( ascertaining if similar phenomena were observed in high-purity iro and, if not, to the presence of which impurities these phenomena coul be traced. Experimental Procedure Three kinds of iron were taken for a study of hardening by rehcatin after cold work: 1. Vacuum-melted electrolytic iron; melted in an Arsem furnace in a magnesia crucible. That this iron was of high purity is indicated b a magnetic permeability above 20,000. 2. Electrolytic iron melted in air in a small laboratory furnace. 3. Armco iron, commercial material. These materials in the form of forged bars, annealed at 980" C. in vacuum, were reduced 30 per cent. in thickness by cold rolling and the reheated to temperatures varying from 100' to 500' C. The Rockwe hardness number was measured after each treatment. The results and shown in Fig. 1, curves 1, 2 and 3.

Jan 1, 1929

By Arnulf Muan, W. C. Hahn

Activities of "FeO" in "FeO"-MgO solid solutions have been determined in the temperature interval 1100" to 1300"C by equilibrating oxide samples with pure metallic iron in atmospheres of known oxygen partial pressures. The end members "FeO" and MgO form a complete solid solution series. The system shoztls a considerable positive deviation from Raoult's law. A large amount of data has been accumulated during the past 50 years on stability relations existing among various crystalline and liquid phases in oxide systems of importance in metallurgical processes. By comparison, very little is known about the thermodynamic properties within the various homogeneous phases present in the systems, particularly the crystalline phases. The present paper deals with determination of activities as a function of composition in solid solutions of wüstite ("FeO") and periclase (MgO). Such solid solutions ("Magnesiowustites") have simple sodium chloride structure and are known to be continuous between the end members at elevated temperatures (> 1000°C) and strongly reducing conditions such as prevail in contact with metallic iron. Wiistite has a defect structure in which the ratio of oxygen to iron varies and never has as low a value as 1:1 corresponding to the simplified formula FeO. Within the range of stability of this phase, the composition changes between a maximum oxygen content of approximately 25.6 wt pct (corresponding to the formula FeO.83O) and a minimum oxygen content of approximately 23.2 wt pct (corresponding to the formula Fe0-96O). The activity data to be reported in the present paper are based on using as standard state the wüstite which is in equilibrium with metallic iron at a particular temperature, i.e., "FeO" of minimum oxygen content at each temperature. For sake of simplicity, the formula FeO will be used to represent the wustite phase in the remainder of this paper. Consider the reaction 2 FeO solid solution = 2 Femeta1 + Ozgas in which FeO dissolved in another oxide dissociates to metal and gas. The activity of FeO in the solid solution, relative to "pure FeO" in equilibrium with metallic iron as a standard state, is then expressed as Here Po, and po2 * are the partial pressures of oxygen of the gas phase in equilibrium with metallic iron and the oxide solid solution or "pure FeO", respectively. The equation holds only if the metal phase is essentially pure Fe. The oxide combinations used in the present investigation fulfill this requirement. Magnesia is so much more stable than FeO that for all practical purposes aFe = 1 in the metal phase present in all mixtures used in this investigation. I. EXPERIMENTAL METHOD 1. General Procedure. The experimental method used was analogous to that developed by Foster and Welch1 and used recently by Hahn and Muan.2 It consisted of four steps: a) Determination of the equilibrium between FeO and Fe. b) Determination of d-spacing change with composition in oxide solid solutions (FeO-WO). c) Equilibration of various members of the solid solution series with a metal phase and a gas phase of known oxygen partial pressure. d) Determination of compositions of solid solutions resulting from step (c) by means of data obtained in step (b). 2. Starting Materials. General Aniline and Film Corp. carbonyl iron powder, type HP, was used as the source of metallic iron. A representative analysis was as follows: Fe, 99.6 pct; C, 0.01 to 0.04 pct; 0, 0.1 to 0.3 pct; N, 0.00 to 0.05 pct. Other elements were present as traces only. Wiistite was prepared from "Baker Analyzed" Reagent grade Fe2O3 of reported analysis as follows: Assay (Fe2O3), 98.8 pct; insoluble in HC1, 0.02 pct; not precipitated by NH,OH, 0.02 pct; Zn, 0.003 pct. Production of wüstite was achieved by a two-step

Jan 1, 1962

By H. Wilton Clark

VARYING in thickness and in number from place to place, coal seams in the Canadian Rockies also range in pitch from nearly horizontal to vertical, sometimes with overturns. Over the entire coal-bearing area there are considerable differences of rank in coals of the same geological age, and there are marked differences in ash content and wash-ability characteristics. Correlation of seams at mining operations within a few miles of each other has often proved impossible. These factors influence mining methods, and, of course, production results. The oldest coal formations of western Canada are of Lower Cretaceous age, as the carboniferous sediments are marine and contain no coal. Coal is also present in formations of Upper Cretaceous and Tertiary ages. The rank of coal varies from semi-anthracite in some operations in the Kootenay and Luscar formations of the Lower Cretaceous in the Rockies to lignite in the Tertiary fields of the Saskatchewan prairies. Fig. 1 shows the general extent of the formations. The pitching seams, chiefly Lower Cretaceous, occur in the western belt of the Rocky Mountain foothills and the eastern slopes of the Rockies themselves. (The road to Tent Mt. strip pit, elevation 7000 ft, is shown on p. 832.) Formations extend from the U. S.-Canadian boundary for several bundred miles in Alberta and continue for a similar distance in British Columbia. Present geological estimates show a probable reserve of the order of 25,000 million tons of coal ranging from high volatile bituminous to semi-anthracite, with a large percentage of coking coal, constituting one of the major world reserves of that type. It should be noted that although the quoted estimate probably errs on the conservative side, the question of access to the seams in a mountainous terrain will always be a problem; a wide divergence may exist between actual reserves and the quantities of coal economically recoverable. An excerpt from the Canadian government Report of the Royal Commission on Coal 1946 states that conditions favorable to coal formation were intermittent and these intervals relatively brief in comparison with periods in which no coal was formed. Conditions were favorable to the growth or accumulation of vegetation in one area, while fresh water sediments and marine shales were being deposited in other areas. During periods of emergence the coal deposits were subjected to erosion or were covered by coarse sands and gravels from the mountains, whereas during submergence they were covered by fine clays deposited in embayments of the sea. During some of the periods of coal formation, volcanic activity deposited beds of fine volcanic ash and dust with the coal-forming vegetation. Coal deposits reach their greatest development in the mountains and thin rapidly to the east into the plains area, where they are deeply buried beneath younger sediments. For example, in the Fernie area of southeastern British Columbia there are in places 22 seams having an aggregate thickness of 150 ft in a stratigraphic interval of 3500 ft, whereas at Cole-man in the Alberta Crowsnest area the measures are only 800 ft in thickness and contain a maximum of five seams aggregating about 47 ft of coal. At Belle-vue, 10 miles further east, the measures are reduced to 430 ft with only three seams aggregating about 37 ft of coal. Subsequently the tremendous forces involved in the upthrust of the Rocky Mountains produced great displacement of the coal-bearing formations and at the same time a change in geologically young coals from low to high rank. The floor or footwall of most seams consists of carbonaceous shales, which are almost as adaptable to bending as the coal seams themselves. The seams show very few cleats or cleavage planes and have been so weakened structurally that as mined 50 pct or more will pass through a ¼-in. screen, seldom

Jan 1, 1955

By G. H. Kesler, J. E. Oberele, C. E. Dryden, J. H. Oxley

Heat-transfer rates were measured in a model of a multifilament vapor-deposition bulb fo the preparation of high-purzty metals. Local transfer coefficients for heat transfer frow the filaments to the circuates process stream were determined as a functiol of gas flow rates and bulb, inlet, and filament, geowzetvies. These results a-e then convertted to n mass-t7,ansfer basis to provide a method of correlating vapor-deposition rates and to predict the performance of commercial urzits. The firm1 cor?,elating equation for mass transfer was found to be- wheve k, is the mass-transfer coefficient; D, is bze gas difficsivity : P is the system pressure; R is the gas constant; T is the gas temperature; Df, DB, and Di ave tlze diameters of the filament, bulb, and inlet, respectively; k is the gas conductivity; C, is the specific heat of the gas, M is the wzolecular zleight of the gas; Npr . N,,, NRe a7'e the Prandtl, Schmidt, and Reynolds numbem of the system, respectively; LB is the height of the bulb; and SB and Sf are the total suyiace area of the bulb and filaments, respectively. ONE of the more conventional methods of preparing very high-purity metals is the vapor deposition of the metal upon heated filaments. The reactions which can be employed in such a process generally fall into two classes: 1) Pyrolysis of a volatile metal compound. Generally because of their instabilities, metal iodides are used; however, other metal halides and hydrides have been employed. These processes are frequently carried out under vacuum conditions, and the filaments are maintained at high temperatures. 2) Hydrogen reduction of a volatile metal compound. A metal chloride is normally the preferred feed for these processes, but other halides are sometimes used under special conditions. These hydro- gen-reduction processes are usually operated at atmospheric pressure. Many of the current processes for the production of semiconductor materials are actually vapor-deposition processes. However, the vapor-deposition process itself is relatively old, and almost all metals can be prepared in a state of very high purity by the use of vapor-deposition techniques. Loonam has recently given an excellent review of the iodide process for metal deposition,' and Owen has started some detailed studies with a carbonyl system.' A more general treatment of all vapor-deposition processes has been outlined by Powell, Campbell, and Gonser.3 Despite the fact that vapor-deposition techniques have been employed since 1890, there is a surprising lack of fundamental information regarding the transport characteristics of even the simplest deposition system, the heated filament. The purpose of the investigation described in this paper was to extend some of the earlier data obtained at Battelle to predict deposition rates and the uniformity of deposition in large-diameter, multifilament deposition bulbs under forced convection conditions.4'5 A number of other investigators have already presented a simplified treatment of the deposition process under non-flow conditions, i.e., pure diffusion.6"9 To obtain information on the transport characteristics of a deposition bulb, a model of a typical plant deposition unit was constructed, and local heat-transfer coefficients from the filaments to a circulating air stream were determined under various conditions of air-flow rates and bulb, inlet, and filament geometries. These results were then converted to a mass-transfer basis, and consequently provided a method to correlate experimental vapor-deposition rates and to design commercial deposition units. EXPERIMENTAL WORK Description of Model. The apparatus which was used in this work was a scaled model of one type of deposition bulb. It was constructed of Lucite and was easily altered to give several combinations of diameter of bulb liner, inlet size and location, number of filaments, and outlet geometry. The model is shown in Figs. 1 and 2. The bulb shell was made octagonal for convenience in construction and observation, and the bulb liners were cylindrical. Air, introduced through one of several different jet inlets, was used as the test fluid in these model studies. Filaments for deposition of metal were simulated by metal rods 1/4-in. in diam. Two of the rods could be heated by passing electrical current through them ¦ while the remaining rods could not be heated. This procedure minimized the amount of heat which was

Jan 1, 1962

By E. S. Machlin, Morris Cohen

The martensite reaction in single crystals and polycrystals of 70 pct Fe-30 pct Ni alloys is shown to be autocatalytic in nature, producing bursts of transformation during cooling. The temperature of the first burst of transformation, called Mb, occurs below M, in these alloys. Experiments were devised to test the athermal embryo and strain embryo theories of martensite nucleation. The results indicate that internal strains, either within the virgin austenite or around existing martensitic plates, control the nucleation process in these alloys. Furthermore, the growth of martensitic plates is not limited by the attainment of an elastic balance with the austenitic matrix, but by the occurrence of plastic deformation at the martensite boundaries which interferes with the propagation mechanism. IN an investigation of the martensitic habit in single crystals of a 69 pct Fe-31 pct Ni alloy,' it was observed that about 25 pct of the austenite transformed during subatmospheric cooling within the time-interval of an audible click. This event proved quite spectacular: The shock wave sent out from the specimen freely suspended on a thread in the refrigerating liquid was occasionally sufficiently intense to shatter the Dewar container and to separate the toluene column in the immersed thermometer. The Present investigation was undertaken to determine- the kinetics and mechanism of this "burst" type of martensitic reaction. The analyses of the alloys studied are given in Table I. The composition of the single crystal specimens is designated by alloy A, while the polycrystal-line specimens were made of alloys B and C as noted in the text. The single crystals were prepared in a vacuum furnace, using a modified Bridgman technique. Most of these crystals were homogenized by holding for 24 hr at about 1300°C just after solidification. However, it may be emphasized here that the degree of homogenization was not a controlling factor in the subsequent experiments, inasmuch as specimens having different degrees of homogenization yielded the same results. All of the single crystals were fully austenitic as slowly cooled to room temperature. An illustration of the burst phenomenon is given in Fig. 1, which shows oscillograms of electrical resistivity and temperature vs. time during the continuous quenching of 1/16 in. wire specimens (alloy B) in a dry ice and acetone bath at —77°C. There are at least two observable bursts in this case, as indicated by the sharp decreases of resistance accompanying the sudden formation of substantial quantities of martensite. The thermal arrest during the quench probably corresponds to the larger burst. Usually the bursts are followed by more or less progressive transformation during continuous cooling. It will also be noted that the resistance continues to decrease after the specimen has reached the bath temperature. This isothermal change denotes the formation of martensite at constant temperature, and will be the subject of another paper. Examination of fiducial scratches on the surface of a transformed single crystal has shown2 that the scratches in adjoining nonparallel martensitic plates are usually bent in opposite directions, as though one plate forms in such a way as to relieve the matrix stresses set up by the adjacent plate. This, together with some of the results described in ref. 1, Table I. Compositions of Alloys Studied, in Percent Alloy Ni C N Mn Si P S Cr A 31±0.3 0.048 0.027 0.003 0.56 0.007 0.002 B 29.5±0.2 0.036 0.02 0.19 0.09 0.008 0.006 C 19.99 0.52 0.37 0.47 0.010 0.015 0.04 led to the tentative concept that a cooperative action exists which provides the impetus for much of the transformation that appears during the burst. The following series of experiments were performed in order to test this idea. Cooperative Nature of the Burst Two adjacent disks, Va in. thick x % in. diam, were cut from a single austenite crystal of alloy A using a jeweler's saw. One of the disks was then cut into 15 parts. Then 12 of the latter pieces and the second disk were austenitized (stress relieved) at 600°C for 30 min and water quenched to room temperature. The temperatures at which the first burst of transformation appeared were determined for

Jan 1, 1952

By Paul Rhedey

Various commercial pelroleum cokes were heat-1,reated at temperatures between 500° and 1500°C, in a nitrogen atmosphere, in laboratovy induction furnaces. The rate of tenlperature rise was varied betzveen 10" and 300°C per min, or the green cokes were flash-calcined, orv a combination of heating rates was used. Changes in the rate of heating had only negligible effects on the degree of' calcination as characterized by mean crystallite thickness and the chemical properties of the calcined coke. The physical structure of the cake was however significantly affected when rate of heating during calcination exceeded 50°C per min over the temperature range of 600° to 900°C. Surface-accessible porosily increased with the .rate of temperature rise, and this was accompanied by a change in pore size distribution. Source and properties of the green coke also had an influence on the structure of the calcined coke. The evidence presented suggests a similar mechanism of porosity development in petroleuiiz coke during calcination in industvial equiplnenl, such as rotary kilns. An increase in surface accessible povosity incveased the pitch binder requirement when the coke was used as aggregate in Soderberg paste. A correlation was established between calcined coke porosity and paste binder requiremenl. ManY results have been published on the changes of properties of petroleum coke during calcination, such as chemical composition, real density, electrical resistivity, crystallite and pore structure. The correlation of these properties with temperature of calcination and time at maximum temperature has been rather well established in both laboratory and pilot plant experiments. Surprisingly little attention has been given however to the effect of calcination conditions, such as rate of temperature rise or furnace atmosphere, on the chemical and structural properties of the calcined coke. It has been observed that petroleum coke, when calcined in industrial equipment, acquires higher porosity and lower real density than those attainable in laboratory furnaces at apparently identical calcina- tion temperature and soaking time. This paper describes a study of the effect of rate of heating during calcination on calcined coke properties using green petroleum cokes of different volatile matter, hydrogen, and sulfur content. An attempt was made to correlate the changes in coke structure with the flowability of anode paste of the type normally used in aluminum reduction cells. EXPERIMENTAL Petroleum Cokes Used. In the study of release of volatile matter and sulfur during calcination and the effect of rate of heating on calcined coke properties two delayed cokes of different sulfur contents were used. The results of analysis of the green cokes are given in Table I. In the study of the effect of flash calcination thirty-six commercial petroleum cokes from twelve different refineries were used with a range of properties shown in Table 11. Calcination Conditions. Calcination experiments were carried out in a laboratory induction furnace. In each run a 200-g sample of dry green coke sized to 10 by 65 Tyler mesh was calcined in a graphite crucible. The crucible containing the sample was placed in the middle of a stack of eight others filled with metallurgical coke to reduce temperature gradients within the sample. Temperature was measured by two Pt, Pt-10 pct Rh thermocouples and controlled by a Celectray instrument. The two couples generally agreed within 5°C. In the study of volatile matter and sulfur release the samples were heated to temperatures in the range of 500" to 1500°C at a rate of 10°C per min in a nitrogen atmosphere and held at the final temperatures for 30 min. In the study of the effect of heating rate on calcined coke properties the desired rates between 10' and 300°C per min were obtained by manually adjusting the power input. Flash calcination was carried out by dropping 100 g of the green coke into the graphite dish preheated to the calcination temperature. Because of the small heat capacity of the furnace the coke was introduced in 20-g portions at a time. For this purpose a 2-in.-long nipple between two 1-in. gate valves installed on the top flange of the furnace served to provide a gas seal while feeding coke to the furnace. It was estimated that the temperature of the coke reached that of the furnace at a rate of approximately 1000°C per min. Holding time at final temperature was also 30 min. Calcined Coke Proper- Determined. Porosity was determined on 20 by 35 Tyler mesh samples using an Aminco-Winslow mercury pressure porosimeter with an operating range of 1.8 to 3000 psi absolute pressure (100 to 0.05 p pore diameter range).' Apparent density was obtained by the mercury poro-simeter. It represents a particle density of the 20 by

Jan 1, 1968

R. M. Raymond, New York, N. Y.—The filling and drawing down of the overhead material was done at considerable depth, which is not the usual method in which it starts at the surface.. R. J. Mechin.— It is a little more than 1000 ft. at El Bordo. The vein above had been caved practically to the surface and some of the ore had been excavated, using underhand and overhand square-set methods. This was far enough away from the shaft to prevent any bad effects. When the sill floor was taken out, to insure the old cave material following down well, we put chutes in the square sets for several weeks until we were sure that the material from above was moving down; on each level above, several crosscuts were driven into the old fill by which we could see the material coming down and thus be sure it was following the first. We had to dynamite it very little because it was dry and dusty. The only trouble was that timber starting from the old square sets would get across the chute and hang up. R.. M. Raymond.—HOW many floors did you operate at once? R.. J. Mechin.—We would operate one floor through the whole length and prepare for the second. F. W. SperR, Houghton, Mich.—Could you have worked it in steps? R.. J. Mechin.—Yes; but working it all on one floor was much better for ventilation. The passage between steps would tend to become heavy and blocked and, as the passages from the mining floor down are usually filled with ore, there would be no way for the air to pass up. By means of doors and fans, air was driven to the extreme end of the orebody through the stope and up the other end; but it could have been worked in steps. F. W. SpeRR.—D the following-up material ever become so heavy that you could see weight on the caps before you had to take out the posts?

Jan 1, 1927

By M. C. Fuerstenau, D. W. Fuerstenau

Studies of the collection of alkali and ammonium halides utilizing vacuum flotation techniques and contact angle measurements show that ionic size controls the flotation of techniquesthese halides with amine salts measurementsas collector. Contact angles of air bubbles on sylvite in saturated brines were withaminemeasured salts asascollector.a function of such variables as collector addition, length of collector chain, and pH of the brine. No contact occurs between halite and an air bubble in brines containing dodecylammonium acetate as collector. LONG-CHAINED aliphatic amine salts have been used for the separation of sylvite (KCl) from halite (NaCl) by flotation.1,2 It is puzzling how these two minerals, which are so similar chemically and crystallographically, can be separated by this method. Gaudin" has postulated that the difference in floatability of halite and sylvite with salts of primary amines depends on ionic size: In the case of amine flotation, the cation would attach itself to the chloride. I have a speculation there, which I cannot prove, that the ammonium group, that is the —NH3 group in the amine, floats potassium chloride because the dimensions of this grour, as it has been measured in other compounds is almost identically the dimensions of the potassium ion, quite different from the sodium ion, and so it fits where potassium had been, in place of it and not attached to it. Apparently, because an aminium ion (RNH3+) is much larger than a sodium ion, it cannot fit into the lattice of halite. Taggart also has speculated that ionic size may control the floatability of sylvite.4 The object of this experimental investigation has been to test this hypothesis and to study what controls the adsorption of cationic collectors at the surface of sylvite. Since collection is to be approached from the viewpoint of ionic size, the ionic radii that are of interest in this work are presented in Table I. The values of the ionic radii of the ions listed in Table I, except NH4+, are those given by Pauling." Several different values for the radius of the ammonium ion have been given, but that of Goldschmidt6 seems to be preferred. The radius of the charged head of a dodecylammonium ion is assumed to be the same as that for the ammonium ion. Little experimental work has been reported in the technical literature concerning the separation of sylvite from halite by flotation. Guyer and Perren studied the separation by flotation of 50 pct binary mixtures of NaCl, KC1, NH,Cl, NaNO3, KNO3, K2SO4, and Na,SO, using either oleic acid or a sodium sul-fonate as collector.' It is possible to measure floatability under actual flotation conditions where all three phases, air- water-mineral, are present by vacuum flotation tests and contact angle measurements.9 Both of these techniques were used in the experimental approach in this paper. Experimental Method and Materials The vacuum flotation tests were run with glass-stoppered pyrex graduated cylinders. Twenty-five ml graduates were used to test the floatability of all salts studied except rubidium and cesium salts. For each test distilled water containing the desired collector concentration was saturated with the salt to be floated. Sufficient salt (—48 mesh) was added to leave about 2 ml of solids in the bottom of the graduate. After the graduate had been agitated several minutes to saturate the solution with air, a vacuum was applied. If the salt were floatable in the collector solution, the gas bubbles attached themselves to the particles, and the particles floated to the surface. In determining the floatability of the expensive Rb and Cs halides, the experiments were run in 10 ml graduates with about 11/2 ml of collector solution initially. Contact angles were measured in the usual manner except that the solutions had to be previously saturated with the mineral to avoid dissolution of the crystal. Solutions for studying contact angles were made by adding the desired amount of collector to a saturated brine, giving the collector concentration in molarity. The mixture was agitated until dissolution of the collector was complete, with the exception of those concentrations greater than about millimolar. At these high concentrations complete dissolution of the collector was impossible. The face of the mineral to be tested was a freshly cleaved crystal of halite or sylvite. The mineral was placed in the brine and conditioned with collector for at least 15 min, which was found to be long enough to obtain a maximum value for the contact angle. The temperature remained constant during each experiment. The experiments were run at 24°C ±2°C. For contact angle measurements, a crystal of halite from Carlsbad, N. M., was used. Several samples of sylvite were used in this work: a crystal of sylvite from Stassfurt, Germany; a crystal from Carlsbad, N. M.; and a crystal of chemically pure potassium chloride. Saturated brines were made from reagent grade chemicals and distilled water.

Jan 1, 1957

By N. M. Langdon

Spokane Meeting, September, 1909.) THE following corrections to the paper of Mr. Langdon (Bulletin No. 34, October, 1909, pp. 919 to 940), received after the publication of the paper in the Bulletin, will be incorporated in the paper as finally published in the forthcoming Vol. XL. of the Transactions, but are here presented for the information of members:' Page. 921. Table I., furnace M2, item 11. For " 8801 " read "0.8801." 922. Table II., furnace C, item 2, under gas. For " 0.011 C " read " 0.011 0." Furnace G, item 13, under gas. For " 5.074 air " read " 3.074 air." Item 15, 114.269 " should be " 4.289." Furnace H, item 2, under iron. For " 0.04 " read ".0.01." 923. Table II., furnace J, item 1, under iron. For " 0.95 " read " 0.93." Furnace N2, item 6, under gas. For " 0.66 CO," read 0.066 C02." 924. Table III., furnace C, item 17. For " 379 " read " 397." 927. Table IV., furnace C, item 4. For " 0.0124 " read 0.0424." Furnace G, item 15.. For " 4,060 " read " 4,066." 928. Line 2. Items 19 and 20 of heat-supply, should be items 20 and 21. 930. Line 3. Expunge remainder of paragraph after the word " supplied," and substitute the following : "As compared with the fusion-zone, the loss by radiation n C was 25.5 per cent. of the total (2,278 units) supplied to that zone, while in D it was 22.5 per cent. of the total heat (1,824 units) supplied to that zone. Finally, as compared with the reduction-zone, C lost by radiation 36.8. per cent. of the 1,580 units supplied to that zone,

May 1, 1910

By N. A. D. Parlee

The diffusion rate of hydrogen in liquid iron has been measured by a gas-liquid metal diffusion cell technique. The diffusion cell was formed by immersing an alumina tube containing hydrogen gas at 1 atm in a bath of stagnant liquid iron. The change in the composition of the melt in the cell was determined by measuring the rate of absorption of the gas in the cell. The appropriate solution to Fick's second law was used to examine the data and calculate diffusivi-ties. The absorption of hydrogen in stagnant pure liquid iron has been found to be diffusion-controlled. The results show that the chemical diffusion coefficient, D, of hydrogen in pure iron in the range of 1547" to 1726°C can be represented by the following Arrhenius relation: D(sq cnz per sec) = 3.2 x X exp(- 3300 i 1800/RT) where the uncertainty in the activation energy corresponds to the YO pct confidence level. Oxygen in the melt (above 0.015 pct 2) increased the apparent rate of absorption of hydrogen. The importance of diffusion data on liquid metals for predicting the rates of certain metallurgical processes has been recognized for a long time. Moreover, these data are much needed to test and develop theory for diffusion in liquid metals. Despite this practical and theoretical interest, however, relatively little reliable information about diffusion in liquid metals is available in the literature. This is particularly true for gas components such as hydrogen, oxygen, and nitrogen in liquid metals, where almost no data on chemical diffusion coefficients are to be found. This is probably due to a multitude of experimental difficulties particularly associated with high-temperature melts. In an effort to fill this gap in information, a research program was undertaken to study the diffusivities and rates of solution of gases in liquid metals. This paper presents the results of a study of the diffusion of hydrogen in liquid iron. EXPERIMENTAL METHOD Two methods for the study of the kinetics of dissolution of gases in liquid metals are being employed in this laboratory. Both involve the measurement of the volume of gas absorbed by the melt as a function of time and as such both avoid the uncertainties involved in chemical analyses of quenched samples for relatively small amounts of gas. In the first method, the gas dissolves in an inductively stirred melt and, in the absence of a slow surface reaction, the results are often interpreted in terms of mass transport across a liquid "boundary layer" between the homogeneous gas phase and well-stirred part of the melt. Other interpretations of the results of such experiments have also been described in the literature.1'5 In the second method a gas-liquid metal diffusion cell is used.' The gas dissolves in a cylindrical column of stagnant liquid metal and, in the absence of a slow surface reaction, the results are interpreted in terms of a non-steady-state diffusion solution to Fick's second law. The weakness of the first method is that while it gives information on the mechanism of absorption by stirred melts it yields an overall rate constant which even in the simplest cases depends on the nature and the thickness of the "mass transport layer" or "boundary layer". It yields no values of diffusion coefficients. The second method was used in this research because in many cases it is possible to determine the diffusion coefficient of the gas component in the liquid metal. In this research it has been utilized to measure diffusion coefficients of hydrogen in liquid iron. The apparatus used was essentially the same as that described by Mizikar, Grace, and par lee but certain modifications have been introduced to meet the elevated temperatures and special conditions of this research. Fig. 1 is a schematic drawing of the apparatus and Table I gives the identification of various parts in this figure. The diffusion cell, shown in detail in Fig. 2, was formed by immersing an impervious alumina tube (hereafter called absorption tube) in a bath of pure liquid iron contained in an alumina crucible. Two types of tubes were used, Morganite triangle RR and McDanel AP35. The crucible was contained in a vertical impervious alumina combustion tube (32 mm ID by 914 mm long) which was closed at both ends by water-cooled brass heads employing O-ring compression seals, Fig. 1. A protection tube enclosing a Pt, 5 pct Rh-Pt, 20 pct Rh thermocouple was introduced through the lower end of the combustion tube

Jan 1, 1968

By Parke A. Dickey

Potential differences between strata of shale and sandstone have been recognized for about years, and they form the basis of the electrical logging of oil wells. Hitherto these potentials have been ascribed solely to electrochemical reactions caused by the disturbing effects of the water in the drill hole through which they are generally measured. Observations by the author in wells empty of water and in a mine shaft suggest that potential differences between sandstones and shales are natural and specific properties of the rocks, although they are modified in measurement by the disturbing effects of the water in the drill hole. The cause of these potentials probably is to be sought in the relative polar adsorptive capacities of quartz and clay particles. Introduction The existence of potential differences in the surface of the earth has been known for some time, and their measurement has facilitated the discovery of geological discontinuities such as ore bodies. The technique of measurement has been well established.'-a In measurements in oil wells, it has been found that the potential of shales with respect to some arbitrary reference is very nearly constant in any one well, at least below the zone to which meteoric water has gained access. Sandstones and some other types of sedimentary rocks have a difference of potential with respect to shale that is negative and may attain a value of 200 millivolts or more. The difference of potential, as usually measured commercially with water in the hole, is distributed over about one foot vertically at the top and bottom contacts of the sandstones. The sharpness of the curve at the formation contacts can be improved by using a short electrode that is close to the wall of the hole. These potentials generally have been ascribed to the disturbing effects of the water in the drill hole. Potentials have been found, however, in wells in which no water was present. In the course of experiments in electrical logging in Pennsylvania, carried out by the Pennsylvania Geological Survey, observations were made of several phenomena that hitherto have not been described, to the author's knowledge, in the published literature. The potential differences across the contacts of sandstone and shale are called in this paper "natural potentials," as distinguished from the potentials caused by electroendosmosis (electrofiltration) and concentration differences, which will be discussed subsequently. Potentials in Cored Well In the spring of 1942, the author measured the potentials in an exploratory well (Day 13, near Grand Valley, Pa.) which had been cored with diamond tools. The potentials were first measured with fresh water in the hole. A null balance potentiometer was used in this and all the other measurements described in this paper. One nonpolarizing porous pot electrode, containing copper sulphate solution and crystals surrounding a pure copper elec-

Jan 1, 1944

By Parke A. Dickey

Potential differences between strata of shale and sandstone have been recognized for about years, and they form the basis of the electrical logging of oil wells. Hitherto these potentials have been ascribed solely to electrochemical reactions caused by the disturbing effects of the water in the drill hole through which they are generally measured. Observations by the author in wells empty of water and in a mine shaft suggest that potential differences between sandstones and shales are natural and specific properties of the rocks, although they are modified in measurement by the disturbing effects of the water in the drill hole. The cause of these potentials probably is to be sought in the relative polar adsorptive capacities of quartz and clay particles. Introduction The existence of potential differences in the surface of the earth has been known for some time, and their measurement has facilitated the discovery of geological discontinuities such as ore bodies. The technique of measurement has been well established.'-a In measurements in oil wells, it has been found that the potential of shales with respect to some arbitrary reference is very nearly constant in any one well, at least below the zone to which meteoric water has gained access. Sandstones and some other types of sedimentary rocks have a difference of potential with respect to shale that is negative and may attain a value of 200 millivolts or more. The difference of potential, as usually measured commercially with water in the hole, is distributed over about one foot vertically at the top and bottom contacts of the sandstones. The sharpness of the curve at the formation contacts can be improved by using a short electrode that is close to the wall of the hole. These potentials generally have been ascribed to the disturbing effects of the water in the drill hole. Potentials have been found, however, in wells in which no water was present. In the course of experiments in electrical logging in Pennsylvania, carried out by the Pennsylvania Geological Survey, observations were made of several phenomena that hitherto have not been described, to the author's knowledge, in the published literature. The potential differences across the contacts of sandstone and shale are called in this paper "natural potentials," as distinguished from the potentials caused by electroendosmosis (electrofiltration) and concentration differences, which will be discussed subsequently. Potentials in Cored Well In the spring of 1942, the author measured the potentials in an exploratory well (Day 13, near Grand Valley, Pa.) which had been cored with diamond tools. The potentials were first measured with fresh water in the hole. A null balance potentiometer was used in this and all the other measurements described in this paper. One nonpolarizing porous pot electrode, containing copper sulphate solution and crystals surrounding a pure copper elec-

Jan 1, 1944

By R.J. Anderson

Roughly, a crack in a casting may be considered, for the moment, to be due to fracture of the alloy resulting from the stress set up by the contraction in volume on passing from the liquid to the solid state. Cracks may be external or surface, internal or deep-seated, or incipient or so-called strain cracks. Draws may appear as depressions, unevenness, and "pinching in" of the surface of a casting. Roughly, draws are due to contraction in volume and are governed largely by the same factors as cracks. A casting may be cracked, however, without being drawn. Shrinkage is a foundry term and refers to the contractio~l in volume of a metal or alloy on passing from the liquid to the solid state. Patternmaker's shrinkage is the total contraction in volume (including liquid, solidification, and solid shrinkage) that occurs when an alloy is poured into a mold. The contraction in volume of aluminum is one of the principal factors affecting the occurrence of cracks, as will be shown later, and the question of shrinkage and its control requires careful study in any treatment of cracking in castings. In the case of sand and die castings, cracks have been attributed to a number of causes, but in the case of most light aluminum alloys, they are due to the large contraction in volume on freezing and the lack of its regulation or control. The most important factors are: (1) Contraction in volume; (2) composition of alloy; (3) quality of melting charge; (4) design of casting; (5) method of molding; (6) hardness of ramming; (7) method of gating; (8) hardness and characteristics of cores; (9) chills; (10) risers; (11) melting temperatures; (12) furnace used for melting; (13) pouring temperatures; (14) inclusions in alloys; (15) hot shortness of alloys; and (16) physical properties of alloys at high temperatures. Of these, the first, second, third, and eleventh to sixteenth involve ascertainable metallurgical facts; the remainder are largely matters of opinion, judgment, and technical knowledge of men experienced in the production of castings.

Jan 1, 1923

By R. J. Anderson

Roughly, a crack in a casting may be considered, for the moment, to be due to fracture of the alloy resulting from the stress set up by the contraction in volume on passing from the liquid to the solid state. Cracks may be external or surface, internal or deep-seated, or incipient or so-called strain cracks. Draws may appear as depressions, unevenness, and "pinching in" of the surface of a casting. Roughly, draws are due to contraction in volume and are governed largely by the same factors as cracks. A casting may be cracked, however, without being drawn. Shrinkage is a foundry term and refers to the contractio~l in volume of a metal or alloy on passing from the liquid to the solid state. Patternmaker's shrinkage is the total contraction in volume (including liquid, solidification, and solid shrinkage) that occurs when an alloy is poured into a mold. The contraction in volume of aluminum is one of the principal factors affecting the occurrence of cracks, as will be shown later, and the question of shrinkage and its control requires careful study in any treatment of cracking in castings. In the case of sand and die castings, cracks have been attributed to a number of causes, but in the case of most light aluminum alloys, they are due to the large contraction in volume on freezing and the lack of its regulation or control. The most important factors are: (1) Contraction in volume; (2) composition of alloy; (3) quality of melting charge; (4) design of casting; (5) method of molding; (6) hardness of ramming; (7) method of gating; (8) hardness and characteristics of cores; (9) chills; (10) risers; (11) melting temperatures; (12) furnace used for melting; (13) pouring temperatures; (14) inclusions in alloys; (15) hot shortness of alloys; and (16) physical properties of alloys at high temperatures. Of these, the first, second, third, and eleventh to sixteenth involve ascertainable metallurgical facts; the remainder are largely matters of opinion, judgment, and technical knowledge of men experienced in the production of castings.

Jan 1, 1923

By L. W. Kesler

Kansas produced 41,966,773 bbl. of oil in the year 1927, thereby taking fourth place among the oil-producing states of the Union. The daily average production decreased from 121,609 bbl. in January to 110,244 bbl. in December. Due to excessive flush production and the resulting decline in the price of crude during the year, operations, both in proved and outlying districts, were reduced materially from those of 1926. There were 2338 completions in 1926, but only a total of 1333 in 1927. Regardless of this decrease in the total number of completions, Kansas made a slight gain in total production for the year, producing 620,262 bbl. more than in 1926. This is accounted for, in part, by the comparative long life of the settled production from the Bartlesville sand areas to the east of the Nemaha granite ridge, and by prolific production in eastern Summer County from horizons new in their importance in the state, productive here at two localities along the axis of the granite ridge (Table 1 and Figs. 1, 2 and 3.) For convenience in further consideration, the state is divided into Eastern and "Western" Kansas. Special emphasis will be placed on those areas in the state of which but little is generally known, and which have not been given detailed consideration in previous publications.

Jan 1, 1928

By Seymour G. Epstein

Using a modified capillary-reservoir technique, electromigration of dilute solutions of silver in liquid bismuth has been measured with relatively good precision. The effects of experiment duration, temperature. current density, and silver concentration have been investigated. Chemical diffusion coefficients of silver in liquid bismuth were also measured at temperatures ranging from 300°to 600°C. The effective valence of silver in bismuth was calculated as a function of temperature from a comparison of electric mobilities and diffusion coefficients. In an alloy containing 1 wt pct Ag, the electric mobility of silver increased from 1.65 x X sq cm per v-sec at 300°C to 2.38 x 10-4at 600°C; the diffusion coefficient in-creased from 3.0 to 17.5 x 10-5 sq cm per sec over the same temperature range. Calculated values of effective valence decreased from 0.28 at 300°C to 0.10 at 600°c. ALTHOUGH there has been considerable progress in the theoretical treatment of electron-transport properties in liquid metals, the mechanisms of mass-transport phenomena in liquid metals are still not clearly understood. This is especially true for the case of electro migration, alternatively termed elec-trotransport or electrodiffusion, which is mass transport induced by an applied electric field. A thorough knowledge of the mechanisms governing electromigration will considerably aid in understanding the electronic state of an impurity atom in a liquid metal. Electromigration has been long observed and studied in liquid metals and alloys. However, no satisfactory theory to explain the observed effects has been developed. The disparity in the present theoretical treatments, summarized by Verhoeven,1 has been attributed to the difficulty in describing the momentum transfer between electrons and ions and to the lack of precise measurements in liquid metals, especially at elevated temperatures. Prior to the last decade most of the electromigration studies were qualitative, in that only the direction of migration of alloy constituents was determined with any degree of reliability. These results have been summarized by schwarz,2 who was able to measure the electric mobility of several elements in mercury. More recently electromigration in liquid metals has been measured by a variety of techniques, with a varying degree of success. Mangelsdorf3,4 made precise measurements on dilute amalgams at room temperature, using a technique relating changes in composition to changes in alloy resistivity, but was unable to make measurements at elevated temperatures. Using a modification of this technique, Verhoeven and Hucke5 measured electromigration and resistivities in liquid Bi-Sn alloys. However, the resistivity technique is not sensitive to small composition changes and is impractical for very dilute alloys. Verhoeven and Hucke6 also measured electromigration of several solutes in liquid bismuth by adapting the capillary-reservoir method for determining diffusion coefficients in liquids. The reported precision of these measurements was poor, however, primarily due to the large uncertainties in the chemical analyses to determine concentration changes. Belashchenko7 has summarized the recent Russian investigations of electromigration, many of which, however, are of questionable accuracy. The present study was initiated to develop an experimental technique which would be both versatile and sufficiently sensitive to yield precise measurements of electromigration of solutes in dilute metal alloys at elevated temperatures. Using this technique, a comprehensive study of silver in liquid bismuth was made to determine the effects of experimental variables on the apparent electric mobilities. Chemical diffusion coefficients of silver in liquid bismuth were also measured over the same temperature range for comparison with the electromigration data. EXPERIMENTAL TECHNIQUE Electromigration Measurements. Electromigration of silver in liquid bismuth was measured by the capillary-reservoir technique, similar to that used by Verhoeven and Hucke,6 but modified to yield more

Jan 1, 1967