Search Documents

By Don L. Kendall, Martin G. Buehler

A simple physical model has been developed to fit carrier-removal data in silicon irradiated near room temperature with reactor spectrum neutrons. Commonly observed donor and acceptor defect energy levels are assumed to be introduced linearly with neutron fluence. The donor levels (in ev) are Ev + 0.16, Ev + 0.27, and Ev + 0.31 and the acceptor levels are Ec - 0.55, Ec - 0.40, and Ec - 0.1 7, where Ev and Ec are the valence and conduction band energies, respectively. The introduction rates of each level are adjusted to fit literature initial carrier-removal rate data. When normalized with respect to the Ev +0.27 level, the relative values of introduction rates are 5.3, 1.0, 3.1, 1.0, 2.0, and 20.0, respectively for the six levels indicated above. To fit p-f (hole concentration vs neutron fluence) and n-f (electron concentration us neutron fluence) data, the introduction rates are multiplied by a factor which preserves the relative values given above. This factor depends upon irradiation temperature, reactor energy spectrum, neutron fluence calibration, and oxygen content of silicon. An extensive study of the effect of neutrons on carrier-removal in silicon irradiated with reactor spectrum neutrons (E > 10 kev) has been given by Stein and Gereth1 (SG) and Curtis, Bass, and Germano' (CBG). They measured initial carrier-removal rates for both p- and n-type silicon over an impurity range typical of silicon devices. In this work, we attempt to fit a simple theory to this data to establish a usable relationship between hole and electron concentration, p and n, respectively, and neutron fluence f. The p-f and n-f relations are needed to assist in the design of neutron tolerant silicon devices and are needed to clarify presently used empirical resistivity-fluence relationships.3 Neutron damage in silicon produces a variety of defects ranging from simple point defects to defect clusters. For the purpose of this treatment, we assume that simple point defects dominate carrier-removal effects. In contrast to this view, stein4 has proposed that defect clusters are responsible for a significant portion of carrier-removal effects. In the following section, it is shown that the carrier-removal effect in n-type silicon with an electron concentration less than 1015 cm-3 can be explained adequately by assuming that the divacancy is the dominant defect and that its introduction rate is independent of the electron concentration. For electron concentrations greater than 1015 cm-= an additional acceptor defect center is needed, and for simplicity the A-center (vacancy-oxygen pair) has been chosen. Although the E-center (vacancy-phosphorus pair) can account for some of the results, the A-center was chosen because the E-center requires a more involved treatment which the presently available data do not justify. In p-type silicon three radiation-induced donor levels are assumed, namely the divacancy and two other centers of unspecified nature located at Ev + 0.16 ev and Ev to 0.31 ev. The donor divacancy at Ev + 0.27 ev is assumed to be introduced at the same rate in p-type as in n-type. However, this rate is too low to fit p-type initial carrier-removal data. The dominant centers in p-type silicon are assumed to be the Ev + 0.16 ev and Ev + 0.31 ev levels where the latter is not the divacancy. The introduction rates are chosen to fit initial carrier-removal rate data. Assuming that the introduction rates are independent of Fermi level, the ratio between them is fixed for subsequent p-f and n-f calculations. Using the same ratios, the initial carrier-removal rate data1,2 as well as p-f and n-f data1,5 can be fit provided the absolute value of the introduction rates are adjusted to account for irradiation temperature, reactor energy spectrum, neutron fluence calculation, and the oxygen content of silicon. THEORETICAL ANALYSIS This analysis is basically the same as that used by Hi116 to analyze electron damage in silicon except we express the degree to which an impurity level is ionized not in terms of the Fermi level, but in terms of carrier concentration. Landis and pearson7 have used the latter approach to analyze y-damage in silicon. Neutron-induced defects responsible for carrier-removal at room temperature are assumed to be simple point defects with no interaction between defects so that they may be represented by discrete energy levels. It is also assumed that no constituent of a defect complex is used up and defects stabilize shortly after irradiation. Defects are assumed to be introduced linearly with fluence according to the product Rtf where Rt is the defect introduction rate and f the neutron fluence. Taking into account the ionization of defects according to Fermi statistics, and considering charge neutrality where minority carriers are neglected, the n-f relation is where no is the preirradiation electron concentration. The parameter Nt is the electron concentration at which the ionized defect concentration is one-half the total defect concentration (Rtf) or where Et is the defect energy level. For silicon at 300°K, ni = 1.45 X 1010 cm-3 and Ei = Ev+ 0.542 ev which was determined using Ec — Ev = 1.11 ev and me* = 1.07 mo and mh* = 0.558m0. The spin degeneracy factor, which usually appears as a number multiplying the Nt/n term of Eq. [1], is taken as unity. In effect, this factor has been incorporated into the defect en-

Jan 1, 1970

By G. R. Belton, R. J. Fruehan

The Knudsen cell-mass spectrometer combimtion has been used to study the Fe-Al and Ag-Al liquid alloys. By application of the recently developed integration technique to the measured ion-current ratios, activities have been derived for the Fe-A1 system at 1600° C and for the Ag-Al system at 1340"C. The results are partially represented by the following equations: Internal consistency between the data on silver-rich and iron-rich alloys is demonstrated by application of the literature measurements on the distribution of aluminum between the nearly immiscible liquids iron and silver. The usual restrictions on the ratio of the mean free path of the escaping atoms to the orifice diameter of the Knudsen cell are shown not to be limiting in this technique. DESPITE the importance of a knowledge of the activity of aluminum in understanding deoxidation equilibria in molten steel, no direct studies have been made of activities in liquid Fe-A1 alloys at steel-making temperatures. Lower-temperature direct studies have, however, been carried out on aluminum-rich liquid alloys by Gross, Levi, Dewing, and Eilson' at 1300°C and by Coskun and Elliott' at 1315°C. Apart from phase diagram calculations by Pehlke, other determinations have been indirect and were made by measurement of the distribution of aluminum between iron and silver475 and combination of these data with extrapolated activities in the Ag-A1 system.~-% ecently, however, Woolley and Elliott have made a significant contribution by directly measuring heats of solution in the Fe-A1 system at 1600°C. The present authorslo have recently employed a Knudsen cell-mass spectrometer technique in a study of activities in iron-based liquid alloys. In this technique activities and heats of solution are determined from a series of measurements of the ratio of ion currents of the components; and since ion-current ratios are used, problems caused by changes in instrument sensitivity or cell geometry are overcome. Results obtained for the Fe-Ni system were found to be in excellent agreement with previous work, thus demonstrating the reliability of the method. The present paper describes a similar study of activities in the liquid Fe-A1 and Ag-A1 systems, this latter system being included in order that a meaningful comparison can be made with the above-mentioned indirect studies. INTEGRATION EQUATIONS A detailed derivation of the equations used to determine the thermodynamic properties from the measured ion current ratios has been given elsewhere;'' however it is useful to summarize them here. By the combination of the Gibbs-Duhem equation with the direct proportionality between ion-current ratios and partial pressure ratios, it was shown that for a binary system at constant temperature and pressure: where al is the activity of component 1 with pure substance as the standard state, N, is the atom fraction of component 2 in the solution, and I; and t'2 are ion currents of given isotopes of the components. The activity coefficient is given by: this latter equation being more suitable for graphical integration. Combination of Eq. [l] with the Gibbs-Helmholtz equation gives an expression for the partial molar heat of mixing: EXPERIMENTAL A Bendix Time-of-Flight mass spectrometer model 12! fitted with a 107 ion source and a M-105-G-6 electron multiplier, was used to analyze the vapor effusing from the Knudsen cell. The arrangement of the Knudsen cell assembly was essentially that of the commercial instrument (Bendix model 1030) but with several modifications. Instead of heating with a single tungsten filament, a cylindrical tantalum-mesh heater was employed. Up to 1400°C simple resistance heating was used but above this temperature electron bombardment between the tantalum mesh and the tantalum cell susceptor was necessary. The temperature was measured by means of a Leeds and Northrup disappear ing-filament type optical pyrometer sighted on an essentially black-body hole in the side of the cell. Details of the temperature control, temperature measurement, and in situ calibration of the optical pyrometer can be found elsewhere.I0 In the investigation of the Fe-A1 system the Knudsen cells were constructed of thoria crucibles with fitted thoria lids (Zircoa). The cells employed in investigating the Ag-A1 alloys were made up of high-purity alumina crucibles (Morganite) with lids of recrystal-lized alumina (Lucalox). The cells were 0.370 in.

Jan 1, 1970

By John F. Abel

"The most conservative persons I ever met are college undergraduates." Woodrow Wilson said that in 1905. No one is saying anything like it today. Chaos on the campus was a cliche of the sixties. Since 1905, education has expanded in every direction. Many more people are being educated for much longer periods of time. The school-leaving age has risen from 14 years to 20 years, on an average. Many are in school until after their 25th birthday. On the surface this situation appears all to the good. Our technical society needs a massive yearly infusion of semitrained (or educated, if you prefer) new blood. The efficiency of our economy and our standard of living are tied to high individual productivity which can be achieved only with an educated work force. Our democratic institutions are said to be founded on an educated electorate.

Jan 1, 1970

By M. J. Pool, R. K. Saxer, J. R. Myers, D. S. Dickson

ThE equilibrium phase diagram of the Mo-Cr system, described by anssen,' indicates that chromium and molybdenum form a continuous series of solid solutions above 1673K. The liquidus and solidus have been reported to have a minimum at about 12 at. pct Mo and 2133°K by Bloom and Grant2 and by Putnam, Potter, and Grant.3 X-ray studies by Trzebiatowski and ploszek4 and Baen and Duwez' revealed no indication of solid solution decomposition after long anneals at 1473", 1273", 923", and 873°K. However, Bloom and rant' by means of cooling and heating curve analysis of alloys containing up to 45 at. pct Mo found thermal effects at temperatures approximately 100" * 30°K below the liquidus. The only previous thermodynamic study of the Cr-Mo alloy system was made by Laffitte and Kuba-schewski6 who used the Knudsen effusion method at 1471°K. These investigators reported the possibility of a miscibility gap because they observed positive deviations from ideality in their chemical activity

Jan 1, 1970

By P. H. Lindon, J. C. Billington

Fe-O melts containing 0.045 pct 0 were deoxidized with Mn-Si-A1 alloys. Product compositions were reluted to the melt and alloy compositions and were found to be most sensitive to the aluminum content of the alloy. Low residual oxygen contents could be obtained when aluminum oxide was present in the Products because of the reduction of silica and manganese oxide activities. Flotation of the Products from a quiescent melt was followed both by analysis of the oxygen content and metallographic measurement of inclusion concentration. MnO-SiO2-A12O3 products were found to float most rapidly when their composition was such that their viscosity may be expected to be low. Changes in the particle size distribution indicates that particle coalescence occurred and differences in the degree of coalescence are thought to be responsible for the different flotation rates observed between products 0f differing composition. Measured flotation rates were slower than those Predicted from a model based on Stoke's Law, although alumina flotation might be reasonably accounted for by this model. Interfacial effects between oxide particles and the melt are believed to be responsible for the discrepancy. It has been recognized that deoxidation products constitute a large proportion of the nonmetallic inclusions present in killed steel. The amount of oxide inclusions which originate as deoxidation products depends largely upon three factors. These may be summarized, according to P16ckinger1 as: 1) Amount of primary products remaining in the steel prior to cooling. 2) Residual dissolved oxygen content of the steel after deoxidation. 3) Amount of secondary products, formed during cooling and solidification, which remain entrapped in the solid steel. In a well-deoxidized steel containing residual aluminum and/or silicon, the equilibrium dissolved oxygen content is usually very low and so the maximum amount of oxide which may be produced as secondary deoxidation products is small in comparison with the amount of primary products. It may be seen, therefore, that the amount of indigenous nonmetallic inclusions may be minimized if a low dissolved oxygen content is achieved by deoxidation and if the primary deoxidation products are efficiently removed. Oxides which originate by reaction of the metal stream with the atmosphere during teeming are not considered in the present study. It is known that two or more deoxidizers may result in a lower equilibrium oxygen content when used in conjunction with one another than when any of the individual deoxidizers are used alone. Equilibrium studies by Hilty and crafts2 and by Bell3 have shown that manganese increases the effectiveness of silicon as a deoxidizer, and Walsh and Ramachandran4 relate this to a reduction in the activity of silica in the products as the manganese :silicon ratio in the steel increases. It was also shown by Herty's work on deoxidation of steel by silico manganese alloys,5 that there existed an optimum ratio of manganese to silicon which gave a minimum inclusion content. This ratio was in the range 4:l to 7:l and the (FeO-MnO-SiO2) products formed by such deoxidation practice were found to lie in a composition range having very low liquidus temperatures (1170 to 1250°C approx). The optimum manganese:silicon ratio was then explained by postulating that these fluid products were able to coalesce and that the larger particles formed floated out of the steel very quickly as predicted by Stoke's Law. The present work examines the effectiveness of various Mn-Si-A1 alloys as deoxidizers and their effects on the composition and removal of primary deoxidation products from a quiescent melt. EXPERIMENTAL TECHNIQUE Approximately 250 g of prepared Fe-O alloy, containing 0.045 to 0.055 pct O, were melted in an alumina crucible and deoxidized at 1550°C by plunging a thin steel cartridge containing the deoxidizer below the melt surface. A high frequency induction furnace supplying current at 8.5 kHz was used to heat a graphite susceptor, the interior of which had been machined to give a wall thickness of 0.85 in. to form a receptacle for the alumina crucible. The iron melt was essentially quiescent as the induced current was concentrated at the external surface of the graphite susceptor by the skin effect. A nonoxidizing atmosphere was maintained over the melt by passing a continuous stream of argon through the lid of the susceptor. The melt temperature was measured before deoxidation, and again at the end of an experiment by means

Jan 1, 1970

By Robert Sisselman

Saving time and energy at Consolidation Coal Co.'s Rowland mine is realized by exploiting its coal deposits via three mining methods. Surface, auger and underground techniques are uniquely combined with specialized transport vehicles to coax the coal from seams unduly protected by topography and lithology. These factors account for what is believed to be the swiftest and most economical means of working the Beckley, W. Va., property.

Jan 1, 1970

By W. R. Hardwick

Nuclear explosive fracturing has been suggested as a method to make hard rock amenable to extraction by a modified block-caving method. To evaluate this possibility, conventional and nuclear blasting practices are discussed, underground mining methods for large deposits and their costs are described, and the cost of fracturing ore with a nuclear explosive and subsequent mining by a modified caving system is estimated and compared with both open-stope and block-caving mining costs. While the direct cost of fracturing the ore in hard rock with a nuclear explosive is less than with conventional explosives, additional costs imposed by the use of nuclear explosives may make their use unattractive for deep mining.

Jan 1, 1970

By R. E. Miner, J. K. Jackson, T. L. Wilson

THE nature of the decomposition reactions producing hardening in aluminum-rich A1-Ag alloys is now fairly well understood. The silver is not uniformly dispersed even at temperatures above the solvus line: rather, it is clustered into zones and associated with dislocations.' During quenching, nearly all of the silver segregates into zones' and some is attracted to dislocations, producing stacking faults. Growth of these faults produces the semicoherent y precipitate.1 During aging, the larger zones and the y precipitates grow at the expense of the smaller zones. The large zones persist up to about 350°C; where they redissolve in favor of the y' and y precipitates.' The equilibrium hcp y precipitate occurs both by a cellular reaction at the grain boundaries1 and by growth of the Y' plates to a critical size where incoherency results. Damask and Nowick 4 observed an internal friction peak in an A1-6 at. pct Ag alloy at about 140°C (0.25 cps). The peak was found to exhibit a systematic instability. After aging at successively higher temperatures between 155' and 23Z°C, the peak first migrated to higher temperatures while increasing in height, then returned to lower temperatures and decreased in height. After aging at temperatures above about 200°C, the peak position remained constant while the height decreased very slowly. In quenched specimens, elastic after-effect measurements inhcated a rapid initial transient in peak position from low temperatures up to 155°C. The authors concluded that a stress induced growth and resolution of the clusters and the 7' was the most reasonable mechanism to account for the behavior. It would seem, however, very unlikely that both the clusters and the y' precipitate would provide damping contributions whose relaxation time and activation energies were identical, which is required for them to produce the same internal friction peak. The present investigation was motivated by this apparent anomaly and the development of a very low thermal inertia torsion pendulum5 capable of monitoring transient behavior. Low frequency internal friction measurements were made on four binary A1-Ag alloys nominally containing 5, 10: 15, and 20 at. pct Ag. The results of measurements made on a 10 at. pct Ag alloy in the quenched condition are shown in Fig. 1. There are two peaks apparent in the spectrum, one centered at about 160°C, the second at about 210°C. The two peak spectrum was observed in each alloy, but it was best resolved in the 10 pct Ag specimens. Because of the instability in the internal friction, 4 it seemed possible that the second peak was not a peak at all, but simply a manifestation of a single, sharply transient, peak. This effect has been observed in Au-Ni alloys.6 The existence of the higher tempera-

Jan 1, 1970

By K. Tangri, B. Swaroop

Simultaneous measurements of lattice (elastic) strain by X-ray line shift method and total strain with an electrical strain gage have been carried out on polycrystalli,ne nickel with the help of a specially designed tensometer attachment for the X-ray dif-fractometer. During the initial stages of deformation, the rate of increase in lattice strain closely follows the total strain until the plastic strain sets in. From then onwards, the two strains deviate from each other and with further increase in afiplied stress the rate of increase of lattice strain eventually decreases. Depending upon the mode of unloading, both compressive and tensile strains have been observed in nickel deformed up to 0.29 pct strain. These results have been explained on the basis of the effects of clustering of dislocations and also the production and behavior of point defects during loading and unloading, respectively. POLYCRYSTALS deformed plastically in a uniaxial tension test show residual lattice strains (hereafter referred to as RLS), which broaden the X-ray line profiles and shift their peak positions.''4 It is known that uniform straining of a crystal lattice (macrostrain) produced movement of ddfraction line peaks, whereas nonuniform straining (microstrain) causes line broadening.= Though the nature and origin of RLS in deformed metals is not yet clearly understood, these are believed to be due to the presence of some form of a locked-up stress system. Several possible detailed interpretations6'" of the stress system have been proposed by various workers and common to all these interpretations is the assumption that different parts of the aggregate have different tensile yield stresses; e.g., that a part A yields under a lower applied stress than a part B. Therefore, during the deformation process, the elastic strain in A will be less than that in B, and, after completion of deformation, B will tend to contract further than A but will be prevented from doing so by the restraining influence of A. Thus, when equilibrium is reached, A will be in compression while B is in tension. Although there is general agreement on the correctness of this argument, controversy still exists as to the exact nature of the parts A and B in a deformed metal. In an alternative approach to the creation of parts A and B, many other investigators have assumed that the RLS observed in unloaded specimens are in some way connected with the lack of proportionality between lattice strain and applied stress in the region above the yield stress." cullity13 has prepared a schematic summary of the results of previous workers, mainly Smith and Wood,2'11 which suggests that, above the elastic limit, the lattice strain may increase less rapidly with respect to strain, or may even decrease with increase in applied stress. If the applied stress is then decreased, the lattice strain would decrease along a line parallel to the loading line in the elastic region and thus produce a compressive residual lattice strain after unloading. At equilibrium, to balance these compressive stresses? there would have to be regions under tensile stress which may well be the grain boundaries or the substructure walls formed during deformation. The present investigation was undertaken to gain a better understanding of the nature and origin of RLS and also to experimentally verify the salient features of this hypothesis. EXPERIMENTAL Stress and strain determinations were made with a specially constructed tensometer attachment for the X-ray diffractometer, Fig. 1: which permits the following measurements simultaneously on a sheet specimen under uniaxial tension at various stress levels during loading and unloading: a) lattice strains, E=. in a direction perpendicular to the direction of pulling (x direction) from shifts in X-ray line peak positions; b) the total strain, ex, in the direction of pulling. with the help of an electrical strain gage affixed to the back of the specimen directly below the area irradiated by X-rays: and c) the applied stress, with the help of a load cell which consisted of a calibrated stainless-steel sample of dimensions identical to those of the specimen under investigation, and to which it was coupled. Because of its high stacking fault energy.I4 nickel

Jan 1, 1970

By H. S. Wang, Y. P. Gupta

Diffusion of zinc and germanium in liquid silver is measured in the temperature range 975" to 1400°C by use of the capillary-reservoir technique. The ex-perinzental results are discussed in terms of the fluctuation and the "hole" theories of liquid diffusion, assunzing the solute as a point charge equal to the difference in the valences of the solute and solvent atoms. The additional coulombic interaction term due to the excess solute charge is evaluated in terms of the Thomas-Fenni approximation. The theoretical and experimental results agree with both the fluctuation and "hole" theories of liquid diffusion. No perceptible effect on solute diffusion in liquid silver of the size difference between the solute and solvent atoms/ions is apparent. DIFFUSION in liquid silver of electropositive solutes indium, tin, antimony, and homovalent gold has been previously reported by Leak and Swalin1 and Gupta,' respectively. These solutes have nearly the same atomic size but differ principally from one another in excess valence (+2, +3, +4, respectively). The observed differences in the temperature dependence and the rate of solute diffusion were explained in terms of an added coulombic interaction between the solute and the solvent atoms. In the case of diffusion of gold in liquid silver, the data were found to be consistent with an effective valence for gold of -0.85, as proposed by Le Claire.3 Recently, Wang and Gupta4 have reported a study of diffusion of ruthenium, iron, and cobalt in liquid silver. The results of diffusion of these solutes and their temperature dependence were found consistent with the hole theories and the fluctuation theories of liquid diffusion, if ruthenium, iron, and cobalt were diffusing in liquid silver with relative valences -2, -1.3, and -0.5, respectively, instead of -3, -3, and -2 as suggested by the relative position of these elements in the Periodic Table of Elements. Furthermore, the predicted valences for these solutes were observed to be consistent with the sparsely available information on the magnetic properties of some of the dilute alloys of copper and silver with these solutes. In theoretical calculations of solute diffusivity, the solute was assumed as a point excess charge equal to the difference in the valences of the solute and solvent atoms. Both the assumption of a point excess solute charge and the use of the Thomas-Fermi approximation to calculate the potential of a solute charge are less precise for the case of electronegative solutes. Therefore, it was not possible to infer from the data on ruthenium, iron, and cobalt an effect, if any, on solute diffusion of the relative differences in either the Period or the size of the solute and solvent atoms or the solute atoms themselves. A study of the diffusion of Period 4 elements, specifically zinc and germanium was, therefore, undertaken since zinc and germanium, respectively, have relative valences of +1 and +3 in liquid silver and the data obtained on diffusion of these solutes could be compared with the available data on diffusion of Period 5 electropositive solutes, indium, tin, and antimony in liquid silver. From such a comparison, it was hoped to learn something about the effect, if any, of the ion size or the structure of the ion core on impurity diffusion in liquid silver. EXPERIMENT The diffusion of tracer zinc and germanium in liquid silver was studied by the capillary reservoir technique, modified adequately for use at high temperatures. In principle, capillaries containing liquid silver

Jan 1, 1970

By A. B. Baldwin

"Uranium and thorium mineralization occurs at several localities in eastern Quebec along the north shore of the Gulf of St. Lawrence. The main occurrences are near Sept-Iles, Baie Johan Beetz, Romaine and St-Augustin river. Mineralization occurs mainly as disseminated uraninite, uranothorite, uranothorianite, monazite and minor zircon and allanite. Yellow to orange oxidation products are observed along fracture planes of the uranium-bearing deposits to a depth of a few feet. Magnetite, biotite and smoky quartz have a distinct tendency to accompany the radioactive minerals. The host rocks include white to pink pegmatite, pink granite and also, in the case of the Romaine and St-Augustin River occurrences, cream to grey plagioclase-rich, well foliated paragneiss and migmatite. Radioactivity near secondary structures, such as fractures and joint planes, indicates some reconstitution of radioactive minerals.The North Shore of Quebec is the northeast extension of the Grenville sub-province. It is also a uranium-thorium metallogenic province."

Jan 1, 1970

By Robert M. Doerr

The research objective was to determine the effect of vanadium alloy additions on the oxidation behavior of cobalt-nickel Alloys. Eighteen alloys were reacted with 0 2 at 0.5 arm and 1,351°, 1,389°, and 1,429° K. The alloy compositions consisted of Co-Ni master alloys (0, 22, 50, 70, 91, and 100 wt pct Ni) with V additions of 0, 0.9, and 2.7 wt pct. Large early deviations from, or changes in, the parabolic rate constant were encountered, but the reaction rate in most cases declined approximately in accordance with the parabolic rate law. Relatively, the vanadium degraded the oxidation resistance most at the highest nickel contents. In some cases, melting of the oxide was inferred from microscopic evidence and X-ray diffraction measurements disclosed the presence of low-melting vanadates.

Jan 1, 1970

By Richard O. Carlson

Ions of p31 and B 11 were implanted in natural insulating diamond macles. The thin (-0.4µ) layers showed sheet resistances of 107 to 1011 ohm per sq and activation energies of 0.17 to 0.34 ev above room temperature. However, no Hall effect could be measured, indicating that mobilities were less than I to 10 sq cm per (v-sec). Such low mobilities may be due to excessive scattering due to radiation damage in the bombarded layer or to very high concentrations of active compensating impurities. Annealing did not materially change the room -temperature properties, but the temperature was limited to -800°C by surface graphitization, the latter process probably being accelerated by lattice damage in our diamonds arising from the ion bombardment. A weak n-type thermoelectric power was detected after p31 irradiation, but it cannot be presumed that we have made the phosphorus ions active as donors in diamond. The natural diamonds have a high and uncontrolled concentration of impurities and, when coupled with the radiation damage and graphitization problems, would appear to seriously limit the quality of semiconductor that we can presently achieve by ion implantation in diamond. ALTHOUGH diamonds are usually thought of as insulating in terms of their electrical conductivity, it was found about 15 years ago that natural semiconducting diamonds do occur rarely, and these were designated as type IIb diamonds. These p-type semiconducting diamonds were found to be dominated by an impurity level 0.37 ev from the valence band. Evidence today based on correlation of the concentration of acceptor states from Hall effect measurements with the impurity concentrations determined by neutron activation analysis point to aluminum as the dominant acceptor impurity.' The compensating donor is believed to be nitrogen, which has a donor level 1.6 ev above the valence band.' However, only a small fraction of the total nitrogen content in the diamond is electrically active. An infrared absorption band at 7.8 p and ultraviolet absorption near 4 ev have been associated with nitrogen, the former providing a quantitative measure of nitrogen content.3 The nitrogen content is -l020 cm-3 in insulating or type I diamonds, but is less than 10'' in natural semiconducting diamonds. Much of this nitrogen is distributed in platelets oriented in (100) planes and not atomically dispersed in the diamond lattice.4 Semiconducting diamonds have been deliberately formed by incorporating impurities into the graphite charge in the high-pressure apparatus used to form diamonds.5 These manufactured semiconducting diamonds are always p-type, and while some show the 0.37 ev level due to aluminum, most such samples have been ascribed to impurity banding effects.6 The best natural or manufactured semiconducting diamonds have hole mobilities near 1500 sq cm per (v-sec) at room temperature, and those with the 0.37-ev level can be analyzed to reveal an acceptor concentration of 3 to 8 X 1016 cm-3 and a donor concentration 3 to 10 times lower.' NO bulk n-type diamonds have ever been reported, but the electron mobility has been measured as 1900 sq cm per (v-sec) by irradiating a diamond with ultraviolet to excite electrons out of the deep nitrogen donor level or other levels.7 New hope for the formation of n-type diamond has emerged from the ion-implantation method whereby desired impurities are introduced into a crystal lattice by bombardment with a high-energy beam of the impurity ion. It was found in the case of silicon that the usual donors such as phosphorus and arsenic and acceptors such as boron and gallium could be implanted into silicon. Wentorf and arrow' produced semiconducting layers on diamonds by an ionic bombardment in a glow discharge at potentials of about 2 kv. The observed typeness from thermoelectric probing seemed to depend on the atmosphere gas (nitrogen, argon, or hydrogen) rather than on the electrode material, but the nature of the conduction process in the thin damaged surface layer is completely unknown. A Russian group under Vavilov has attempted high-energy ion implantation in natural diamonds using boron and lithium ions9 and later also phosphorus, aluminum, and carbon ions.10 Their papers claim n-type layers from lithium, phosphorus, and carbon implant and a p-type layer from boron and aluminum implant, though the methods of type determination are not described in detail. Under government contract support, the Ion Physics Corp. has studied ion implantation in several semiconductors. While the bulk of their study1' was devoted to the irradiation of silicon, they did carry out a short study on boron and phosphorus implantation into natural diamonds. They did observe surface conducting layers but did not determine the typeness of the layers. For our experiments, phosphorus and boron were chosen as the dopant ions because of their respective donor and acceptor behavior in germanium and silicon. Moreover, they are the lightest mass dopants of the shallow level donors and acceptors from columns 111 and V of the Periodic Table (excluding nitrogen which is already present in diamond), and will have the largest depth penetration into the diamond lattice. EXPERIMENTAL Diamonds. As the target diamonds for our ion implantation study, we chose commercially available macles, which are flat, twinned diamond crystals. By scanning through the stock of a wholesale distributor

Jan 1, 1970

By A. S. Venkatadri, H. B. Bell

The equilibrium sulfur distribution between molten iron and Ca0-Mg0-Al203 slags containing iron oxide was investigated at 1550°C. The results were used to derive the iron oxide activities at low iron oxide concentrations in the slag by combining the sulfide capacity data obtained from gas-slag work with the free energies of both the sulfur solution in iron and the iron oxide formation in slag. The derived ferrous oxide activities were compared with values based on Tem-kin's kin's and Flood's ionic models. One difficulty in using these models is that the nature of the aluminate ion in slag is uncertain. Nevertheless, such indirect methods, in particular, those described in the present paper, are of value because of the difficulty of measuring small amounts of oxygen in liquid iron in equilibrium with slag. It is shown that these methods confirm the consistency of thermodynamics data on liquid iron and slags. It is well established that decreasing the iron oxide activity in the slag increases the desulfurization of molten iron at constant slag basicity. This effect is most pronounced at the very low iron oxide activities, characteristic of blast furnace slags. Yet a precise quantitative determination of the significance of low iron oxide contents in slag in blast furnace desulfuri-zation is not possible for the following reasons: a) difficulty of separation of iron "shots" from the slag, and b) errors in chemical analysis of small amounts of iron oxide in slags. In view of these obstacles, one must resort to indirect methods of calculating iron oxide activities. EXPERIMENTAL TECHNIQUE The apparatus for providing the sulfur equilibrium data has been described previously1 and was similar to that used by ell' in connection with the study of slag-metal manganese equilibrium. The procedure consisted of: a) melting about 50 g of Armco iron in a magnesia crucible in a platinum furnace, b) adding a mixture of about 15 g of lime-alumina slag and varying amounts of Fe2O3 and CaS, and c) maintaining the temperature at 1550°C for more than an hour in an atmosphere of argon to enable the sulfur equilibrium to be attained. Several melts were made using lime-alumina slags with basic composition 55, 50, and 45 pct lime. During the experiment the temperature was controlled manually using a Pt/10 pet Rh-Pt thermocouple. After the experiment, the Power was shut off and the flow rate of argon was increased to freeze the melt as quickly as possible. The analysis of sulfur in the metal was carried out by the oxygen combustion method3 using uniform drillings from the top and bottom of the metal button. After crushing and grinding and removal of any iron particles with the aid of a hand magnet, the slag was analyzed for sulfur by the CO2 combustion method.4 The E.D.T.A. method was employed for the analysis of lime5,6 and magnesia,= the ceric sulfate method7 for the analysis of slag iron oxide, and the perchloric acid dehydration method5 for the analysis of silica. The remaining amount was taken to be Al2O3 precipitation with ammonium hydroxide in several preliminary melts had confirmed the propriety of using this simple procedure. RESULTS The activity of iron oxide in binary, ternary, and more complex slags has been the object of numerous investigations, and the two experimental methods for its determination are: 1) Equilibrating the metal with the slag in question and measuring the oxygen content of the metal. The ferrous oxide activity is then given by aFeO L%OJSat where [%0]sat is the oxygen content of the metal in equilibrium with pure iron oxide slag. This method was used by Chipman et al.8,9 2) Equilibrating the slag in iron crucibles with known partial pressures of H2/H2O or CO/CO2 mix-tures.10-12 This method is limited to temperatures between 1265" and 1500°C. The very low oxygen content of the melts in this investigation made it impossible to derive the ferrous oxide activity by the first of these methods. Therefore, the iron oxide activities were computed by means of: Sulfide capacity data from the gas-slag work" Temkin's concept14 Flood's approach15 a FeO from Sulfide Capacity. The method of calculating the aFeO involves the sulfide capacity of the slag (c,), the sulfur distribution coefficient (Ls), the free energy of dissolution of sulfur in iron, and the free energy of formation of iron oxide in the slag. Bell and Kalyanram13 have investigated the sulfur absorption characteristics of lime-alumina slags containing magnesia by the Carter-Macfarlane method16 (based on comparing the sulfide capacity of the slag in question with that of a standard slag of unit lime activity) and have derived lime activity values. The relation between sulfide capacity and their lime activity a'CaO is given by: Cs= 3—: Xa'CaO at 1500°C

Jan 1, 1970

By Fred Leighton

The Bureau of Mines has developed rock noise source location techniques that may be used to obtain precise and practical source locations from micro-seismic field data. These techniques enable one to monitor all or any part of a rock structure and map progressive stress buildup or rock movement. Thus, the microseismic method of determining instability in rock structures is becoming an invaluable engineering tool for the mining or civil engineer engaged in rock excavation and construction.

Jan 1, 1970

By Ivan L. Nichols

Tailings dumps at Mercur, Utah, contain about $7 million worth of gold in 3.5 million tons of mill tailings assaying about 0.07 ounce of gold per ton. Commercial mining and milling at Mercur ceased in 1942. To determine whether improved technology was now available that might permit economic reprocessing of the Mercur dumps, the Bureau of Mines probed for a low-cost method that would be applicable to these tailings. A composite sample prepared for laboratory studies assayed 0.062 ounce gold per ton. About 10 percent of the gold occurred as a water-soluble cyanide. Conventional cyanide leaching dissolved 32 percent of the gold, and hot, strong cyanide solution dissolved 42 percent. Results were not improved by pretreatment with acids, bases, or oxidants. After expensive preroasting, gold recovery by cyanidation was increased to 65 percent. The simple procedures were ineffective, and more effective procedures were too costly for economical processing of Mercur tailings.

Jan 1, 1970

By J. K. Conibear

Since the 1948 introduction of hydraulic fill at the Sudbury Ontario Mines of the International Nickel Company of Canada, 32,000,000 tons of water-borne sandfill and 420,000 tons of portland cement addition have been placed in underground workings. Fill method mining, currently conducted at six of the ten operating mines, accounts for 28 percent of the total mines underground production. Both classified mill tailings and screened alluvial pit sand are used as fill for cut-and-fill primary mining, undercut-and-fill pillar recovery, and backfilling of blasthole and shrinkage stopes. Cemented fill, consisting of a lean mixture of portland cement, is used in areas where a stabilized fill is required to reduce dilution and aid in the recovery of adjacent pillar ore.

Jan 1, 1970

By M. M. Fine

A process for simultaneously reducing and indurating iron are pellets was developed by Bureau of Mines laboratory research and confirmed on a continuous pilot plant basis. Moist iron are concentrate was balled, mixed with solid reductant, and heated at about 1,100° to 1,150° C. On a small scale, a variety of concentrates and fuels were utilized successfully. The pilot rotary kiln tests were limited to natural magnetite, synthetic magnetite, lignite, and anthracite. In both cases, partially reduced pellets with excellent physical properties were regularly produced with average analyses of at least 80 percent total iron and 60 percent or more metallic iron. These and other prereduced pellets were smelted in the Bureau's experimental blast furnace facility, Bruceton, Pa. With prereduced spheres comprising 85 percent of the burden, the coke rate was approximately half that with standard oxide pellets and production of hot metal was increased by 75 percent. Economic appraisal of the cost of making hot metal indicates that reduced pellets could be more profitable than any other iron raw material in both investment capital and daily cost of operation.

Jan 1, 1970

By R. G. Vardiman, M. R. Achter

THERE has been interest recently in the conditions under which dislocation pipe diffusion may be observed. In order to extend the measurements to temperatures below Tm/2, where pipe diffusion becomes more important, a chemical sectioning technique was developed for silver single crystals which makes possible the removal of layers of 100A or less. The procedure is briefly as follows. Silver single crystals of 99.995 pct purity were cut in the form of discs t by 4 in., electropolished, and electropolated with Ag110 to about 1000A thickness. The thickness was determined by calibrating the activity with the weight of heavier platings, up to 10,000A. Both diffused and undiffused samples were sectioned in a two-step process. Reaction with a solution of iodine in alcohol formed a thin, even layer of AgI, which was then dissolved in a KCN solution. The reaction rate of the iodine solution varied slightly over a period of time, so that the thickness of the stripped layers, while quite even and regular for a given specimen, ranged from 65 to l00W for different specimens. This variation was checked by using a thickly plated standard with each run. Layer thickness was determined from undiffused samples by the point where the layer activity drops to half the maximum. This maximum intensity also provides, with appropriate corrections, the initial concentration per layer co. Because of the small penetration distances, it is necessary to use the exact solution of Fick's law for a source of finite thickness.' For a plating of thick- where c is the concentration at depth x. flj the error function, D the apparent diffusion coefficient, and t the time. The diffusion coefficient was obtained from the concentration-penetration data by extrapolating to x = 0 where c/co = erf h/2(Dt)1/2 Enough points were taken to impart confidence in the accuracy of the intersection, generally about ten layers. Four temperatures were used as in Table I. For best accuracy, t and h were chosen to give values of the parameter (Dt)1/2/h between 1.0 and 1.7. A typical concentration-penetration curve is shown in Fig. 1. The diffusion coefficients calculated from Eq. [I] are plotted in Fig. 2. For comparison, the high-temperature data of Tomizuka and sonder 2 which go down to 630°C are plotted as the solid line 2(a) from their expression D = 0.40 exp (-44,090/RT). Also for comparison, the lowest temperature point of Hoffman and Tumbull3 is plotted. It is apparent that our data diverge appreciably from the high-temperature extrapolation by 373°C. A better straight line fit for the low-temperature portion is obtained by subtracting the extrapolated volume diffusion coefficient from the apparent. or measured coefficient, and plotting the difference, curve (h) of

Jan 1, 1970

By J. Serra, M. Turpin, R. Alberny

A new method is proposed to obtain automatically an unbiased estimate of the interdendritic spacing A. It is shown that the structure can be built by a random distribution of a rectangular basic unit. The usual definition of the interdendritic spacing yields a value 1 which, coupled with A, gives the shape of the basic rectangular unit. A study of the relationship between dendrite arm spacing and solidification rate was conducted as a part of a research program on dendritic solidification. For any comparison between theory and experiments to be meaningful, it is necessary to measure reproducible and clearly defined spacings. Fig. 1 shows clearly that even in a fairly simple structure such measurements are difficult. if we compare photos (a) and (b), we can hardly assess which one shows the larger spacing. Moreover, the quantity to be measured can be defined in several ways. On a small scale, dendrites seem to be distributed in a very regular and ordered way. However, on a larger scale, this regularity no longer exists. The usual way to measure spacings is to pick up some rows of dendrite stalks; such measurements are obviously biased since less orderly regions are not taken into account. A more objective method is to section the picture with randomly oriented straight lines and average the measured interdendritic spacings. The purpose of this paper is to show that such measurements, which can be conducted automatically, yield objective and reproducible results which are directly related to characteristic features of the dendritic structure. I) THEORETICAL ANALYSIS 1-1) Mathematical Morphology and Function of Co-variance. Matheronl and Serra2 enlarced- the definition and geometrical meaning of the notion of covariance to encompass problems in mathematical morphology. A brief review of part of their theory can be stated as follows. Let A' be an isolated particle, and let us consider a group of two scanning points h is the vector joinlng these points, h its modulus, and a its direction. Let us also define a boolean function, k(x), whose value is 1 if x lies inside the particle and 0 if it 1s outside. The total number of events in which both points are inside A' IS the covariance Ii',(h): is the measure of the intersection of A' with the particle ALh obtained by translating A' of -h Obviously K,(O) is the volume of the particle A', or its area if A' is a two-dimensional body. The average value of K,(h), when the direction of a vector h of constant modulus is rotated, is: An important property of the covariance is illustrated on Fig. 2: if the shape of a particle shows some regularity along a direction a, K,(h) generally presents a maximum. We now consider a group A, made of many particles A' distributed in space; then k(x) is taken as the realization of a stationary random f~nction.~ The volume fraction (or fraction of area if the problem is two-dimensional) occupied by A equals the average value q of k(x): meaning "average value of"). We also replace

Jan 1, 1970