Search Documents

By H. W. Reinhardt, J. W. Dally

A dynamic photoelastic analysis of stress waves interacting with a free surface is described. The free surface is that of a bench with a fixed bottom so common in quarry applications. The stress waves are generated by line charges of lead azide (Pb N,). Four models of identical geometry are investigated with the direction of detonation of the line charge varied between the four models. Dynamic photoelastic patterns are recorded and analyzed to indicate which method of detonating the line charge produced the largest magnitude of tension at the free surface. The mechanics of rock breakage by means of explosives has received considerable treatment by many investigators including Duvall, Obert, Broberg, Rinehart, and Langefors1-11 over the past two decades. Indeed in more recent years several texts12-15 have been written on the topic, treating a wide variety of subjects which are logically related to the modern technique of rock blasting. In rock blasting the chemical energy of a concentrated explosive contained in a relatively small diameter borehole is utilized to fragment the rock. The explosive is transformed into a gas with enormous pressures which exceed 10-5 bars18 This high pressure shatters the rock in the area adjacent to the borehole and produces dilatational and distortional stress waves which propagate radially away from the borehole. The state of stress associated with these outgoing waves produces a system of cracks which extend for a few feet from the borehole. The breakage produced in this manner is limited as the dynamic stress in the pulse attenuates markedly with distance. In the absence of a free surface, the stress wave propagates away from the source without further fracture. With a free face of rock near the drill hole, another mode of breakage occurs which is due to scabbing failure of the layer of rock adjacent to the free face. These scabbing failures are produced by the reflection of the incident waves and the conversion of compressive stresses into tensile stresses sufficiently large to fracture the rock. The detailed nature of the interaction of the stress waves with the free surface is complex and difficult to treat analytically. However, dynamic photoelasticity offers an experimental approach which gives a fullfield visual display of propagating stress waves and the reflection process. Applications of static photoelasticity to solution of problems related to mining technology have become relatively common (see, for instance, Refs. 17 and 18) with a plastic model loaded to produce a state of stress representative of that occurring in the workings of a mine. The application of dynamic photoelasticity is ex tremely limited. Tandanand and Hartman19 have used a multiple spark camera to study fracture in glass and plastic plates impacted by a chisel-shaped tool. This paper describes a dynamic photoelastic analysis of stress waves interacting with a free surface. The free surface is that of a bench with a fixed bottom so common in quarry applications. The stress waves are generated by line charges of lead azide (Pb-N6). Four models of identical geometry are investigated with the direction of detonation of the line charge varied between the four models. Dynamic photoelastic patterns are recorded and analyzed to indicate which method of detonating the line charge produced the largest magnitude of tension at the free surface. Experimental Procedure The model illustrated in [Fig. 1] was fabricated from a sheet of Columbia Resin CR-39 to represent a bench with a fixed bottom. Properties of the CR-39 pertaining to these dynamic experiments are listed in [Table 1]. Scribe lines on 1-in. centers are used to identify locations along the bench face. The bench height was 8 in., the burden was 3 in., and the overall dimensions of the sheet, 16 and 18 in., were large enough to eliminate reflections from nonessential boundaries during the period of observation of the dynamic event. To simulate a charge in a borehole, a groove 0.062 in. wide and 0.080 in. deep groove was cut into the sheet from one side. The lower end of the groove was 1 in. or 1/3 the burden distance below the bottom of the bench. The upper end of the groove was 3 in. or one times the burden distance below the upper level of the bench. The groove was packed with 60 mg of Pb No per in. of length, and ignited with a bridge wire detonator. Four different ignition procedures were used to examine the effects of detonation direction on the stress wave interaction with the free face of the bench. In Test 1 the line charge was ignited at the top and the line charge detonated downward. In Test 2 the line charge was ignited at the bottom and the charge burned upward. In Test 3 the charge was ignited in the center with the top half burning upward and the bottom half burning downward. Finally in Test 4 the line charge was ignited at both ends simultaneously. Sixteen high-speed photographs of the photoelastic fringe patterns representing the stress wave propagation were recorded for each of the tests. A Cranz-Schardin multiple spark gap camera 20,21 was operated at framing rates which were systematically varied from 110,000 to 250,000 frames per sec during each test.

Jan 1, 1972

By W. W. Vaughn, R. H. Barnett, E. E. Wilson

Soon after the search for uranium ores on the Colorado Plateau began in earnest, thousands of feet of drill core ranging from 1 1/8 to 2 1/8 in. diam became available for study. Although significant advances had been made in the technique of quantitative gamma-ray borehole logging, instrumentation was in the development stage, and complete reliance could not be placed on gamma-ray logs alone to interpret quantitatively the meaning of radioactivity in a drillhole. A method that would be faster than chemical analysis and still give reproducible and reliable results for various drill core sizes was desirable to provide additional information on the enormous footage of drill core being accumulated. A solid phosphor scintillation drill core scanner was designed and constructed. Basically the instrument was developed to measure radiation from a drill core which would not be clearly recorded by a gamma-ray logger using a Geiger tube as the sensitive element. Such data would be beneficial in constructing isorad maps to delineate ore-bearing zones. A calibration in the range 0.01 to 0.1 pct eU.,O, was provided; above 0.1 pct eU3O8 gamma-ray logs were available and were being used to calculate grade and tonnage of ore reserves. The core scanner, however, has been used to estimate equivalent uranium content of ore-grade materials containing as much as 2.2 pct eU3O8 with an accuracy of ± 10 pct, the sample being in the form of a BX drill core. Actually, an apparent calibration of eU3O8 vs counts per unit time is a straight line with a slope that is a function of the sensitive element and the geometry of the counting assembly. A true calibration that will show the expected departure from a straight line is due principally to the random nature of the pulse from a radiation source and the nonlinearity of the electron circuitry. Design and Construction: Three methods of detecting radioactivity were considered and applied in developing the core scanner now in use: 1) the Geiger tube, 2) liquid scintillation phosphors, and 3) solid scintillation phosphors. The desired sensitivity and long-term drift characteristics needed for this operation could be attained only by using solid scintillation phosphors. All three methods are discussed. Before scintillation counters were common, nine beta-gamma sensitive Geiger tubes 7/8 in. diam by 12 in. long were used, arranged to surround the drill core with tube axes parallel to the axis of the core. This arrangement of Geiger tubes was en- closed in a lead shield 1 in. thick, and provision was made to slide a 6-ft length of drill core manually into the counting chamber, one foot at a time. A count for each segment was taken with a scaler while the core remained stationary. The equivalent uranium content of the different sections of drill core could then be estimated with the aid of a calibration curve of counts per unit time vs percent equivalent uranium (eU). In rare cases the effects of the radioactivity concentrated in small areas within the core introduced errors in the readings made with the Geiger tube arrangement owing to the geometry of the measurement. The variability of counting rate due to a localized concentration of radioactivity in a spot in the wall of a drill core is illustrated in Fig. 1. This effect and the inherent low efficiency of the Geiger tube were considered major disadvantages of this counting arrangement. When liquid scintillation phosphors became available the core scanner in Fig. 2 was constructed to make a more accurate measurement of the equivalent uranium content of a sample. This instrument contains about 4 liters of liquid phosphor in a stainless steel coaxial cylinder 1 ft long, with inner and outer walls 0.060 in. and 0.125 in. thick, respectively. Four end-window type photomulti-plier tube with cathodes of 2 in. diam, immersed in the solution at right angles to the axis of the core, were used to observe light flashes in the phosphor. The liquid phosphor offered equal sensitivity to radiation originating at any point in the enclosure and represented geometrically the optimum in design. However, providing a semi-permanent leak-proof seal between the glass envelope of the phototube and the metal walls of the container proved to be a serious problem in constructing the equipment. The most effective seals were especially machined O-rings from sections of large tygon tubing. The tygon took a permanent set owing to cold flow characteristics and in most cases sealed completely. The light absorption characteristics of the liquid phosphor changed gradually with time, and after one month the counting rate had decreased to half the original value. The most sensitive liquid phosphor tested proved to be a solution containing 4 g of 2.5-diphenyloxazole and 0.01 g of 2-(1-naphthy1)-5-phenyloxazole per liter of toluene. With fresh solution in the chamber and with all photomultiplier tubes operating in parallel, the counting rate contributed by any one of the four photomultiplier tubes was about 85 pct of the counting rate from a single tube operated individually. From these observations it was concluded that owing to coincident loss and light attenuation within the liquid phosphor, the apparent sensitivity could not have been materially increased by additional phototubes. However, this approach to core

Jan 1, 1960

By Kun Li, Alex Simkovich

The rate of dissolution of alumina in carbon-saturated liquid iron has been studied experimentally in a system where alumina was in the form of a cylindrical rod immersed in an iron bath contained in a graphite crucible. Data obtained consisted of the concentrations of aluminum in the melt as a function of time. In the case of static experiments, the data are shown to agree with theoretical prdictions based on the diffusion of aluminum.. The rate of dissolution was greatly increased by the rotation of the alumina rod. It is concluded that the diffusion of aluminum from the alumina/metal interface is the rate-controlling step. In the past, thermodynamic investigations of systems encountered in ferrous process metallurgy have received widespread attention. More recently, considerable work has been devoted to the study of kinetics associated with these systems in an effort to determine their rate controlling mechanisms. The alumina-iron system is of great importance in ferrous metallurgy. Yet information concerning kinetics of reaction in this system is seriously limited. The present study was made in order to establish the rate-controlling step for dissolution of solid alumina in liquid iron. LITERATURE REVIEW A number of papers concerning dissolution of solid metals in liquid metals have been reported in the literat~re. Generally, for these simple systems, dissolution is controlled by mass transfer of the dissolving species. Complex systems involving dissolution of solid metal carbides and oxides in liquid metals and slags have been studied to a much lesser extent. Skolnick5,6 reported on the reaction between liquid cobalt and poly-crystalline cylinders of tungsten carbide, in which the cylinders were dissolved while being rotated about their longitudinal axes at various speeds and temperatures. As a result of unexpected preferential grain boundary attack by the liquid cobalt, large errors in the measured dissolution rates occurred because of loss of tungsten carbide grains to the liquid cobalt. Nevertheless, it was possible to establish that the liquid Co-W carbide reaction was not controlled by mass transfer. In a similar approach, cooper7 was able to show that artificial sapphire rods, (alumina single crystals) dissolving in lime-alumina-silica slags obeyed a mechanism of mass transfer control. Here, again, the rods were rotated at various speeds and temperatures, and the process was followed as a function of these variables. Forster and Knacke8 took a practical approach to reaction between slags and refractories. By blowing argon through refractory cylinders of silica, silli-manite, or dolomite and directing the gas to rise along the slag-refractory interface, it was possible to increase the rate of mass transfer. Although the method was admittedly crude, it nevertheless permitted an evaluation of the relative stabilities of refractories with respect to slag attack. Data were interpreted on the basis of mass transfer control. EXPERIMENTAL TECHNIQUE Apparatus. An illustration of the apparatus used in this study is shown in Fig. 1. The furnace consisted of a Morganite recrystallized alumina tube wound with a molybdenum coil. A secondary molybdenum heater was mounted around the upper half of the primary coil to aid in controlling the thermal gradient within the furnace. The primary heater tube was 3 in. in ID and 30 in. long. A reducing mixture of 95 pct N and 5 pct H was maintained around the heating elements. Thermal insulation was provided by alumina powder. The chamber within the primary combustion tube contained a boron nitride block near the top to assist in controlling the thermal gradient to the furnace and also to provide a bearing surface for the rotating graphite shaft. The outside diameter of the graphite shaft was $ in. A separate threaded graphite specimen holder was screwed into the end of the shaft. The holder contained a tapered hole drilled into the end to guide the oxide specimens as they were pressed into it for mounting. Additional guidance for the rotating graphite shaft was furnished by a water-cooled bronze bushing attached to the top of the furnace. A steel clamp was fastened to the upper end of the graphite shaft and rested on a thrust bearing; the shaft and clamp were driven by a dc motor through a set of gears. Two O-rings located immediately above the bronze bushing maintained a gas-tight seal about the graphite shaft. The lower half of the alumina tube housed the crucible and charge, which were placed on a 3/4-in. diam movable alumina support tube. With this arrangement, charges could be inserted into or removed from the furnace while the hot zone was maintained at or above 1000°C. To control the temperature of the furnace, the thermocouple was mounted inside the support tube and in contact with the crucible bottom. Stray electric fields in the furnace were of sufficient intensity to cause erratic indications by the thermocouple. By enclosing the thermocouple protection tube in a molybdenum sheath and grounding this shield, the problem was eliminated. Output of the thermocouple went to an automatic continuous balance controller. Procedure. A typical run was as follows. First, electrolytic iron was premelted in graphite crucibles and cast into graphite molds with the same configura-

Jan 1, 1970

By John W. Burd

A system is described for the simultaneous deposition of epitaxial layers on as many as eight substrates. A high degree of uniformity of both physical and electrical characteristics is achieved in the films. Variation of film thicknesses is consistently less than ±10pct within a wafer and from wafer to wafer within a run with the variation typically on the order of 55 pct. Composition variation of GaAs1-x PX layers within a wafer and from wafer to wafer within a run is consistently less than 51 pct. Electrical evaluation of the films by several techniques indicates excellent doping uniformity within a wafer and from wafer to wafer within a run. Mobilities for lightly doped GaAs films at 300°K are consistently >6000 cm2 v-1 sec-1 and mobilities > 7000 cm2 v- 1 sec-1 are regularly attainable. Techniques for the preparation of material with carrier concentrations from 1 x 1015cm-3 to 1 x 1019 cm-3 n-type and 5 x 1016 to 5 x 1018 cm-3 p-type are discussed. METHODS for the preparation of 111-V compounds by vapor phase reactions have been extensively reported in the literature.1-6 Almost all of the apparatus described for these various methods are suitable for processing one or at the most a very limited number of wafers simultaneously. With the recent rapid advances in the use of vapor grown GaAs for microwave oscillators and GaAs1-xPx as visible light emitters the requirements for these materials are steadily increasing. In order to satisfy these requirements it is necessary to move from a laboratory scale apparatus to one which is capable of processing a large number of wafers simultaneously. Desirable features would be a high degree of uniformity among the wafers and good reproducibility from run to run. The apparatus to be described fulfills these requirements very well. DISCUSSION The various methods reported in the literature can be classified under three headings: 1) closed tube, 2) open tube, and 3) the close-spaced method. Of these three the open-tube method is the most amenable for scale-up to a manufacturing process. It is the most versatile and the various operating conditions can be more precisely controlled than with the other two methods. A number of chemical reactions may be used to achieve vapor-phase growth of 111-V compounds. Sev-era1 of the more generally used reactions are shown in Fig. 1. All of these reactions have the following points in common: 1) generation of a volatile group III(Ga) species by the reaction of the transport agent (halide or HC1) with either Ga or GaAs, 2) introduction of the Group V(As and/or PI component, 3) a method of adding dopant, if desired, and 4) a region in which deposition from the vapor will occur and form as a single crystal epitaxial film on the substrates. The laboratory scale reactors permit the hot re-actant gases to flow into the relatively cooler deposition zone and pass successively over the several substrates which are arrayed along the long axis of the tube parallel to the gas flow. With this arrangement the composition of the reactant stream is continually changing as solid material is deposited on each successive substrate. As a result of this changing gas composition the reaction driving force also changes from substrate to substrate and the degree of uniformity of layer thickness, doping level, and so forth, is poor. This effect can be partially overcome by imposing a controlled temperature gradient along the deposition region to compensate for change in gas composition. However, even when this is done variations in layer thickness on the order of 30 to 40 pct are common and as high as 50 pct are frequently experienced between adjacent wafers in the tube. To expand this arrangement to a large number of wafers would only increase the nonuniformity from the first to last wafer in the line. From the above discussion the two undesirable features of changing gas composition and temperature gradient become evident. A reactor system which eliminates or minimizes these undesirable features is one in which the apparatus is mounted vertically as shown schematically in Fig. 2. The vertical mounting permits the disposition of a number of substrates on a suitable support so that all wafers are at the same vertical height in the furnace and hence at essentially the same temperature. By using only a single row of wafers the reactant gas mixture passes over only one substrate in its path through the reactor. Thus the two undesirable features of changing gas composition and temperature gradient are minimized. An additional design feature which further minimizes temperature variations is rotation of the substrate holder. Rotation serves to integrate any radial temperature gradient existing around the resistance heated furnace. A photograph of a reactor assembly at the completion of a run is shown in Fig. 3. MATERIAL PREPARATION Apparatus. Although any of the several chemical systems shown in Fig. 1 are adaptable for use in this apparatus the one generally used is System 2, the hydride synthesis system. This system has been de-

Jan 1, 1970

By Frank C. Kelton

It is perhaps not widely realized that extraction and saturation processes carried out on oil well core samples alter the properties of these samples to varying degrees. On the other hand it is felt by some that quick-freezing of core samples increases their permeability and porosity significantly. Accordingly, laboratory tests were carried out on 49 pairs of horizontally adjacent samples in order to differentiate between the effect of quick-freezing per se on permeability and porosity of the samples, as distinguished from the effect of the identical saturation treatment on permeability and porosity of the companion samples. Also, additional field data were obtained on comparison of frozen vs unfrozen companion samples. LABORATORY INVESTIGATION OF FREEZING us SATURATION EFFECTS Procedure The samples used in these tests were two-cm cubes cut in horizontally adjacent pairs from cores from eight Gulf Coast and Mid-Continent wells, which cores had not previously been frozen. These samples were extracted with carbon tetrachloride, dried, and air permeabilities run in the conventional manner. They were then evacuated and saturated with brine of 25,000 ppm sodium chloride content, and porosities determined by gain in weight. The samples were partially desaturated by evaporation down to an average brine saturation of 68 per cent. One sample from each pair was quick-frozen by covering with dry ice after wrapping in a single layer of paper, and allowed to remain frozen for about two hours; the companion sample from each pair was not frozen. After thawing the frozen sample, all samples were immersed in tap water overnight in order to leach out most of the brine. Air permeabilities were re-run, and the samples were again saturated with brine to determine a second porosity value. For purposes of averaging of data, the samples were grouped according to four permeability ranges, from 0 to 10, 10 to 100, 100 to 1,000, and 1,000 to 3,840 md. Average permeability and porosity changes for the frozen vs the unfrozen adjacent samples are shown in Table 1. Discussion As may be seen from Table 1, the averages of the per cent permeability increases for the quick-frozen samples ranged from 3.8 to 12.9 per cent among the four permeability groups. The average changes among the four groups of unfrozen companion samples ranged from a decrease of 0.2 per cent to an increase of 9.3 per cent. There was no particular correlation of these changes with magnitude of permeability; however, the increase for each group of frozen samples paralleled the increase for the corresponding unfrozen samples. The differences between the two sets of values are believed to be a valid indication of the effect of the quick-freezing in itself, since the treatment of the two samples in each pair was identical except for freezing. The permeability changes which are strictly the result of the quick-freezing are shown in the sixth column of Table 1. These range from a decrease of 0.9 per cent to an increase of 4.0 per cent; the overall weighted average is 1.2 per cent, as compared to an average increase of 6.8 per cent caused by the saturation treatment of the samples not frozen. The average porosity changes are in general smaller than the changes in permeability, and range from a decrease of 2.3 per cent to an increase of 3.3 per cent. The overall weighted average change ascribed to the quick-freezing is 1.0 per cent of porosity. Many factors can contribute to the changes in permeability and porosity observed when subjecting cores to the simple processes used in these tests. Such are: hydration and swelling of clay, adsorption of ions, changes in surface structure and wettability, expansion and compression effects due to ice formation, shrinking and cracking, leaching of salts and colloids, displacement of particles resulting in either blocking or enlarging of pore openings. Whatever particular mechanisms are involved. however, it is apparent not only from this study but also from other investigations in the literature' not directly concerned with quick-freezing, that the effects produced by commonly used extraction, saturation and drying techniques may be of considerable magnitude The results of this study indicate that for the particular samples and techniques used, such effects are of the order of five to six times the effect of quick-freezing. insofar as changes in permeability are concerned. It may be argued that these samples might not include extremely shaly material where the effect of freezing upon permeability may be much greater. However, had such material been available for these tests, it would undoubtedly have been very susceptible also to alteration by the extraction and saturation treatment used. To investigate this point further, the individual sample data were re-grouped according to the magnitude of the average per cent permeability increases for the pairs of samples, irrespective of permeability. The results

Jan 1, 1953

By N. J. Grant, B. C. Giessen, Paul Predecki

ALLOYS of gold, silver, and aluminum with elements of the groups BII, BIII, BIV, and BV were prepared by a rapid quenching technique (splat) and were examined by X-ray diffraction. Five new intermediate phases were found and will be described briefly herein. For the gold and silver systems, the concentration ranges having an electron/atom ratio e/a of 1.4 to 1.5 ("3/2 Hume-Rothery phases") were studied primarily. Master alloys were prepared from high-purity metals (99.9+ pct or better) by melting either in evacuated fused silica capsules or by nonconsum-able-electrode arc melting in an argon atmosphere. Small pieces, 20 to 50 mg, of each alloy were blast-atomized to form a splat, by a technique similar to that described by Duwez and Willens.1 The technique used for this study is described in detail in Ref. 2; it utilizes a resistance-heated graphite crucible with a small hole at the bottom, directed toward a metal substrate or quenching plate. The prepared alloy rests over the fine hole, through which it is expelled by an explosion shock wave in the form of fine droplets (1 to 50 µ) of molten metal onto a copper or silver substrate, which is maintained at about -190°C. The resulting very high cooling rates (see Ref. 2 for quantitative measurements) can prevent the process of nuclea-tion and growth in many instances, resulting in the formation of metastable phases. The splat particles were transferred to a GE-XRD5 diffractometer and maintained at -190°C, where they were examined with CuKa radiation. The samples were then allowed to warm to room temperature or were heated to higher temperatures until the equilibrium structures formed. Of fifteen alloy systems considered, nonequi-librium structures were encountered in six; these are described below and summarized in Table I. In the system Au-Sb a metastable £ phase (A3 type, hcp, a = 2.898 + 0.002A; c = 4.731 * 0.004A; c/a = 1.633) was found in the concentration range Au + 13 to 15 at. pct Sb. This phase is isomorphous with the stable phases in the systems Au-Cd, Au-In, and Au-Sn, all at an average e/a ratio of 1.4 to 1.5. The concentration range of one-phase metastable was deduced from the small amounts of supersaturated gold solid-solution phase present in the splat product. It was found that ? could also be retained by splatting onto a substrate held at room temperature: however, decomposed into the equilibrium phases Au + AuSb2 after heating to 200°C for 1/2 hr, or on holding the powdered splatted alloy at 20°C for several months. Calorimetric measurements will be made in an attempt to decide the question whether ? is metastable at all temperatures or whether it is a stable phase at low temperatures. There is evidence that another phase, possibly also close-packed but with a different stacking sequence, can be obtained by rapid quenching of alloys with a different antimony content. Klement, Willens, and Duwez3 reported the existence of an amorphous phase on quenching Au-Si alloys (25 at. pct Si) to - 196°C. They found that on heating to room temperature another phase of unknown crystal structure was formed. This was confirmed (see Table I); however, the new crystalline phase, designated as ?, could also be formed simply by rapid quenching to room temperature, and even was found to exist already in the as-cast Au + 20 at. pct Si alloy. It was found that ? decomposed into Au + Si on the specimen surface at room temperature. This behavior, and the question whether or not there is an equilibrium-temperature region for ?, have not yet been resolved. It is probable that ? (Au + 20 to 21 at. pct Si) is cubic of the -brass type (D81-3) with a = 9.60, + 0.01A and N = 52 atoms per cell [compare 6 (CU-Sn)4]. Except for two very weak lines, the powder pattern of about thirty lines could be indexed on this basis; however, a determination of the atom positions has not yet been attempted. For Au-Ge the C phase was observed at about 21 at. pct Ge as reported by Luo et at.5 Lattice parameters a = 2.876A, c = 4.73,A, c/a = 1.64 were found. In the Au-Pb system, formation of a ? phase was not observed, but in the lead-rich region at 75 at. pct Pb, broad peaks belonging to an amorphous phase were found. The maximum diffracted intensity occurred at 28 = 32.4 deg which is about 1 deg larger than the position of the (111) line of lead (Cuka). For Ag-Pb, an amorphous phase analogous to the one found in the Au-Pb system was observed; this metastable phase exists probably at about 75 at. pct Pb. Since no lead-rich alloys were tested, all alloys consisted of silver + amorphous phase at -190°C. In A1-Ge alloys, line-rich and complex powder patterns were obtained at about 30 at. pct Ge; they bear similarities to those of aluminum and germanium, but are of lower symmetry; the existence of more than one intermediate phase is possible. The authors are grateful to the Kennecott Copper Corp. for Fellowship support, and ARPA (Contract

Jan 1, 1965

By B. F. Oliver, C. W. Haworth

Pure tin and Sn-0.5pct Pb ingots have been frozen unidirectionally from the base. For quiescent melts that were initially undercooled, a transition from lower eqlciaxed structure to an upper columnar structure is found in the alloy ingots. Columnar to equi-axed back to columnar transitions are observed in superheated alloy ingots, but no such equiaxed band is observed impure tin. The reproducible equiaxed band is associated with a thermal undercooling followed by a recalescence. This undercooling is <5"C, whereas the critical (maximum obtainable) under-cooling for both the pure tin and the alloys used is -20°C. A similar undercooling is observed at the same position in the pure tin ingots, although in this case no clear transition in structure can be seen. The structure of the pure tin ingots is either entirely columnar or mixed columnar-equiaxed. A consideration of the detailed thermal history of the ingots indicates that the ingot macrostructures are determined by the occurrence of a local therlnal undercooling in conjunction with nuclei multiplication and transport mechanisrris. GENERALLY it is found that a pure metal ingot solidifies so as to produce an entirely columnar structure. Frequently an alloy ingot is found to have a columnar outer zone and an equiaxed central portion. Early systematic work to examine the factors controlling the formation of the equiaxed structure was reported by Northcott&apos; who showed that, for copper alloys frozen unidirectionally with a given ingot practice, the alloying element influenced the length of columnar crystals and the extent of the equiaxed structure. Northcott showed that alloys with a wider freezing range more readily produced the equiaxed structure. The nucleation process can be important in producing equiaxed structures; frequently an alloy which readily solidifies with an entirely columnar structure will produce an entirely equiaxed structure when a nucleating agent is added to the melt.&apos; The formation of the equiaxed structure was attributed by Winegard and chalmers3 to the presence of constitutional supercooling; that is, a region of liquid in front of the growing solid could have a temperature below its equilibrium liquidus temperature. Thus, with a small enough temperature gradient in the liquid, it was suggested that the presence of constitutional supercooling may be sufficient to bring about the nuclea-tion necessary for the formation of an equiaxed structure. Although this explanation is plausible, and may be relevant in many ingots, Walker has described an experiment&apos; for which constitutional supercooling seems to be an unlikely cause of nucleation. A Ni-20 pct Cu alloy, repeatedly undercooled more than 50"C, was crystallized and found to show the typical colum-nar-equiaxed structure. The separation between the liquidus and the solidus for the alloy is 40°C. Thus, in this experiment the nucleation required for the formation of the equiaxed structure must have come about in some other way than by the nucleation catalysis constitutional supercooling hypothesis. Chalmers has suggested more recently5 that nuclei (in a typical ingot) are present immediately after pouring and are prevented from redissolving by the constitutional supercooling effect. More recently Uhlman, Seward, Jackson, and ~unt&apos; have shown direct evidence using ice and organic materials that freeze dendritically that the "remelt mechanism" may be an extremely effective crystal multiplication process during the freezing of ingots under conditions involving dendritic growth. JSlia" experimentally demonstrated the detachment of dendrite arms. chernov14 has analyzed the dendrite arm detachment process as a coarsening phenomena driven by the minimization of interphase area. Katta-mis and ~lemings" working with undercooled steel melts give evidence supporting this mechanism. Mechanisms of dendrite arm detachment such as those assisted by convection are believed to be the origin of the macrostructures obtained in this study. This study makes no attempt to distinguish the relative contributions of these mechanisms. The object of the present work was to obtain accurate temperature measurements during the solidification of an ingot and to correlate these measurements with the formation of equiaxed grains in the resulting ingot structures. Similar previous work is very limited. The measurements carried out by Northcott are neither sufficiently extensive nor sufficiently accurate for any interpretation. Plaskett and winegard7 carried out experiments on A1-Mg alloys in which they observed values of the temperature gradient, G, in the liquid and rate of freezing, R (for a given alloy solute content Co), at the transition from a columnar to an equiaxed structure. They reported that equiaxed crystals were produced at values of G/G approximately proportional to the solidus composition. Similar experiments using Pb-Sn alloys carried out by £111011" showed a linear relation between G/R and the solidus composition. However, the thermocouples were in the mold wall rather than in the melt and, in one case, ingot surfaces were examined. There is ambiguity in the meaning of the values of G and R measured in all these experiments. APPARATUS AND EXPERIMENTAL PROCEDURE Alloys were prepared by induction melting 99.999 pct Sn and 99.999 pct Pb to form a Sn-0.5 wt pct Pb alloy in air in a graphite crucible and casting into a cylindrical graphite mold 6 in. long, 1 in. in diarn , and with a & in. wall thickness. This mold was mounted on a copper base through which cooling water could be

Jan 1, 1968

By Winston A. Wong, John J. Jonas

Commercial purity aluminum was deformed by extrusion over the temperature range 320° to 616°C and the strain rate range 0.1 to 10 per sec. Flow stresses and strain rates were calculated from the experimenLa1 ram pressures and speeds. The stress-strain rate-lemperature relationship in extrusion was found to be similar to that in creep. Extrusion, torsion, compression, and creep data extending over ten orders of magnitude of strain rate and over two orders of magnitude of stress were correlated by a single creep equation. It was concluded that hot-working is a thermally activated process, in which the rate-controlling mechanism is either the climb of edge dislocations or [he motion of jogged screw dislocations. The microstructural changes observed during extrusion were consistent with the proposed deformation mechanisms. ALTHOUGH great progress has been made in understanding the technology of extrusion, very little is known about the actual deformation mechanisms operating during flow. Previous accounts describing extrusion have indicated that the relationship between ram speed (V), pressure (P), and temperature (T) can be given as follows:1 V = apb and P = A&apos; exp(-AT). In these equations, a and b are constants which depend on temperature, A&apos; is a constant which depends on ram speed, and A is a "coefficient" with a different value for each metal. Although these equations have fairly wide application, they do not contribute much to a fundamental understanding of the deformation. Furthermore, extrusion has not hitherto been considered as a thermally activated rate process. This lacuna is surprising because hot-working is similar to high-temperature creep in several respects. There is, in fact, a fair body of experimental evidence suggesting that the material response under hot-working conditions is similar to that occurring under creep conditions, in spite of the many orders of magnitude difference in strain rate.2"4 Since creep has been extensively analyzed in terms of dislocation mechanisms, the comparison of hot-working to creep is useful, for it can suggest the possible deformation mechanisms operating during hot-working. In this paper, the hot extrusion of aluminum will be examined from the point of view of thermally activated deformation mechanisms, such as operate during creep. EXPERIMENTAL PROCEDURE The experimental procedure consisted of extruding commercial purity aluminum* over a range of ram velocities and temperatures at constant die reduction by the direct method. Details of the experimental equipment have been published elsewhere.5 Extrusion was carried out at each of the following billet temperatures: 320°, 376°, 445°, 490°, 555°, and 616°C at the following constant ram speeds: 0.002, 0.008, 0.02, 0.1, and 0.2 in. per sec.* All results were obtained using a square-shouldered die with an extrusion ratio of 40:1, giving a reduction in area of 97.5 pct. The ram force was the dependent variable, and was measured by means of strain gages on the ram and was plotted as a function of ram travel. The sequence of events before making an extrusion was duplicated before each run so as to minimize as much as possible variations in experimental conditions. For example, after the equipment had been assembled, the billet was allowed to heat up to temperature inside the insulated container. Once the container attained the desired temperature, a period of 1/2 hr was allowed to elapse before the extrusion was made. This time was found to be required to allow the billet to reach a steady-state temperature, as determined from previous tests. When all was ready, extrusion was carried out without interruption; that is, the billet was upset and extruded in one operation. EXPERIMENTAL RESULTS AND DISCUSSION The two usual experimental approaches for investigating high-temperature deformation exhibit an important common feature. In the first approach, which corresponds to creep, a constant stress (or load) is applied to the material at constant temperature and the resultant strain is recorded against time. After an initial transient stage, a state of constant strain rate exists (secondary creep), in which a steady-state condition is established which is sensitive to variation in either applied stress or temperature. In the second approach, a constant strain rate is applied and the resultant flow stress is recorded. This corresponds to the situation in hot torsion or hot compression, where it is observed that, for a constant test temperature, there is an initial rise in stress to a steady value which is maintained up to very high strains. In tests of this type, a steady-state region is also established in which the stress is sensitive to variation in either the strain rate or the temperature.3,4,6-16 In both types of tests, therefore, a steady-state region is established after an initial transient. In the case of hot-working this region may be called steady-state hot-working, and it is analogous to steady-state creep with which it has many common features. Stress Dependence of the strain Rate in Extrusion. In order to assess the stress dependence of the strain rate under extrusion conditions, and to compare it to that of creep, as well as of hot torsion and hot compression, the extrusion data were analyzed according to power, exponential and hyperbolic sine creep equations.

Jan 1, 1969

By M. Metzger, J. C. Bilello

Microyielding in 99.999 pct Cu occuwed in two distinct parabolic microstages and was substantially indeoendent of grain size at the relatiz~ely large grain sizes stzcdied. The strain recouered on unloading was a significant fraction of the forward strain and was initially higher in a copper-coated single crystal than in poly crystals. Results were interpreted in terms of cooperative yielding and short-range dislocation motion activated otter a range of stresses, and a formalism was given for the first microstage. It was suggested that models involving long-range dislocation motion are more appropriate for impure or alloyed fcc metals. THERE are still many unanswered questions concerning the degree and origin of the grain size dependence of plastic properties. In the microstrain region, a theory of the stress-strain curve proposed by Brown and Lukens,&apos; based on an exhaustion hardening model in which the grain boundaries limit the amount of slip per source, accounted for the variation with grain size of microyielding in iron, zinc, and copper.&apos; This theory assumes N dislocation sources per unit volume whose activation stress varies only with grain orientation. Dislocations pile-up against grain boundaries until the back stress deactivates the source, which leads to a relationship between the axial stress and the strain in the microstrain region given by: where G is the shear modulus, D the grain diameter, a the flow stress, and a, is the stress required to activate a source in the most favorably oriented grain.3 If this or other grain-boundary pile-up models are correct, then the reverse strain on unloading would be much larger for a polycrystalline specimen than for a single crystal. Also, the microplasticity would become insensitive to grain size if this could be made larger than the mean dislocation glide path for a single crystal in the microregion. These questions are examined in the present work on polycrys-talline copper and a single crystal coated to provide a synthetic polycrystal. EXPERIMENTAL PROCEDURE Tensile specimens 3 mm sq were prepared from 99.999 pct Cu after a sequence of rolling and vacuum annealing treatments similar to those recommended by Cook and Richards4-6 to minimize preferred orientation. Grain size variation from 0.05 to 0.38 mm was obtained by a final anneal at temperatures from 310" to 700°C. Dislocation etching7 revealed pits on those few grains within 3 deg of (111). For all grain sizes dislocation densities could be estimated as -107 cm per cu cm with no prominent subboundaries. The single crystals, of the same cross section, were grown by the Bridgman technique with axes 8 deg from [Oll] and one face 2 deg from (111). An anneal at 1050°C produced dislocation densities of 2 x 106 cm per cu cm and subboundaries -1 mm apart in these single crystals. A Pb-Sn-Ag creep resistant solder was used to mount the specimens, with a 19 mm effective gage length, into aligned sleeve grips fitted to receive the strain gages. All specimens were chemically polished and rinsed8 to remove surface films just prior to testing. The synthetic polycrystal was made by electroplating a single crystal with 1 µ of polycrystalline copper from a cyanide bath. Mechanical testing was carried out on an Instron machine using two matched LVDT tranducers to measure specimen displacement, the temperature and the measuring circuit being sufficiently stable to yield a strain sensitivity of 5 x 107. At the crosshead speeds employed, plastic strain rates were, above strains of 10¯4, about 10¯5 per sec for polycrystalline specimens and 10-4 per sec for the single crystals. Plastic strain rates were an order of magnitude lower at strains near l0- &apos;. A few checks at strain rates tenfold higher were made for reassurance that the initial yielding of polycrystalline copper was not strongly strain-rate dependent. Test procedures followed the general framework outlined by Roberts and Brown.9,10 An alignment preload of 8 g per sq mm for polycrystals, and 2 to 4 g per sq mm for single crystals, was used for all tests. These gave no detectable permanent strain within the sensitivity of the present experiments; although at these stress levels, small permanent strains are detectable in copper with methods of higher sensitivity.11 12 stress and strain data are reported in terms of axial components. RESULTS General. The initial yielding is shown in the stress vs strain data of Fig. 1. For polycrystals, cycle lc, the loading line bent over gradually without a well-defined proportional limit, and almost all of the plastic prestrain appeared as permanent strain at the end of the cycle. The unloading curve was accurately linear over most of its length with a distinct break indicating the onset of a significant nonelastic reverse strain at the stress o u, indicated by the arrows. The yielding in subsequent cycles, Id and le, had the same general character. The single crystal behavior, shown to a different scale at the right of Fig. 1, was different in that initially the nonlinear reverse strain was unexpectedly much greater than for polycrystals. It should be noted that these soft crystals had a small elastic

Jan 1, 1970

By W. F. Chiao

With the application of dislocation damping theory an attempt was made to determine whether the generation and extension of dislocations is inherently more difficult in a brittle steel than in a ductile steel. A ductile steel was compared with a brittle stee1 by simultaneously measuring the ultrasonic attenuation and velocity during tensile test, and the density of free dislocations and their mean loop length were then calculated as a function of strain. The results showed that in the ductile steel there was always a large generation of dislocations and great extension of loop length occurring at some stage within the early plastic region. In contrast, the brittle steel showed very little or no such sudden changes in dislocation dynamic states after the onset of plastic deformation. Furthermore, a strong temperature dependence of dislocation dynamic states was also observed in the ductile steel and a hypothesis was suggested that a thermally activated process of dislocation rearrangement could occur at higher deformation temperatures. The activation energy of dislocation rearrangement at room temperature was estimated as about 2030 cal per mole.C. DUCTILITY is an indispensible property in the application of engineering materials, especially steel. During the past two decades the theoretical and experimental approach to the understanding of flow and fracture of metals has been constantly undergoing changes and progress." while the fracture behavior of metals can be influenced by many factors such as chemical Composition,3 second-phase particle mor-phology,4 and dislocation arrangement,5 it is now a general belief that the fundamental understanding of the ductile-brittle fracture phenomena of solid materials must stem from the study of dislocation dv-namics developed under stress conditions.6,7 Most of the traditional ductility tests, such as Charpy impact test, slow bend test, and tensile fracture test, cannot by themselves reveal directly the mechanisms of ductile to brittle transition of materials. In the experimental investigation of tensile ductility it would be ideal to be able to study directly the dynamics of dis-locations in a bulk specimen during the process of deformation. Since the ultrasonic pulse technique is the only satisfactory method for studying dislocations and the fine details of deformation characteristics in metals in the course of a tensile test, it would appear that a comparative study of ultrasonic attenuation changes during tensile tests of metallic materials exhibiting different ductility might be very informative. So far no work comparable to this study has appeared in the literature. Recent progress in both theory and experiment has indicated the feasibility of studying the dislocation mechanisms of ductility behaviors by ultrasonic measurements during tensile test. Granato and Lucke8 have developed a quantitative theory that enables the calculation of dislocation density and their average loop length from the measurements of ultrasonic attenuation and velocity, and several investigators, including Chiao and Gordon,9&apos;10 have shown that simultaneous ultrasonic measurements can be successfully made during a tensile test. Furthermore, many investigators11-13 have repeatedly proposed in the past several decades that deformation and fracture are mutually self-exclusive, and that the ability or inability of a material to deform plastically, i.e., to generate dislocations, is a major factor in determining whether the material will be ductile or brittle. Thus, in the present work an attempt was made to determine whether the generation and extension of dislocations is inherently more difficult in a brittle steel than in a ductile steel. This article is principally concerned with the study of the relation between the propagation of ultrasonic waves and tensile deformation in a steel series which displays quite different toughness at room tempera-turk. changes in attenuation and velocity of ultrasonic waves have been measured as a function of strain during the deformation process. The results have been interpreted in terms of the vibrating string model for dislocation damping as developed by Granato and Lucke, and it has been found that some of the more subtle predications of the model are in good agreement with the experiments. This would be especially meaningful because most of the previous experiments in testfying the model were carried out with single crystals of high-purity materials and little work has been done with polycrystalline steel alloys. EXPERIMENTAL PROCEDURES AND RESULTS Specimen Materials. The tensile specimens used throughout this experiment were of two compositions selected from a series of Fe-Mo-0.77 pct Mn-0.22 pct C steels prepared for a ductile-brittle fracture transition study. One steel contains 0.21 pct Mo and the other 1.03 pct Mo. These two compositions were chosen for the present study because they possess quite different toughness properties at room temperature. The 0.21 pct Mo steel is quite ductile while the 1.03 pct Mo steel is rather brittle, as measured by the standard Charpy impact test. The alloys had been prepared by vacuum induction melting and chill casting in steel molds. The ingots were hammer-forged into 1/2-in.-sq bars from which tensile specimen blanks were cut. These blanks were first normalized under argon atmosphere at 1700°F and then reaus-tenitized and isothermally transformed at 1050°F to a bainitic microstructure. The chemical compositions, heat treatments, hardness measurements, and Charpy transition temperatures of the two steels are listed in Table I.

Jan 1, 1970

By G. S. Cole, H. Biloni, G. F. Bolling

An experimental investigation of sovlze low .melting point alloys sJtows that a substvucture of isolated depressions can be the first manvestation of constitutional supercooling on solid-liquid interjaces veuealed by decanting. Electron-tni cvop vobe and wletallo gvaplic esanzinations, in tlze bulk belzind the interjace, oj the segregation associated with these isolated areas substantiate tlzei&apos;v depressed nature, since a solute of ko < 1 is enriched, and a solute of ko > 1 depleted. In contrast, the pox structuve, a set of projections often veported in the literature, leaves no trace oj. segvegation. These obserl;atims, accovlrpanied by a brief review of recent literature, point to inconsistencies between experirrental obsevvation and the idea that the fornzation of a projection is a causal step in the development of a cellular substructure. An argument is presented to show instead how it is plausible for substantial depvessiom to form in the pvesence of constitutional supercooling at dislocations threading the solid-liquid interjace. THE development of constitutional supercooling during growth from the melt leads to the formation of the cellular solidification substructure. This well-founded association between structure and instability has been basic in understanding cellular substructure and micro segregation; however, the initial formation of structure seems unclear. Rutter and Chalmers,&apos; in definitive experiments and theory, noted that in the presence of constitutional a planar interface might break down: "resulting in the formation of a small projection on an initially plane or uniformly curved interface." That is, the breakdown from a planar to a cellular interface was implied to be initiated via a projection into the unstable liquid. Later, Walton et (11. found that a structure of isolated projections, termed "pox", appeared at solid-liquid interfaces decanted under growth conditions near the onset of constitutional supercooling; the pox were taken as the indication of the instability promoted by the supercooling. Tiller and Rutter4 in their extensive work studied the shape transitions at decanted interfaces which were generally observed to proceed as— pox, "irregular cells", elongated cells, regular (hexagonal) cells, and so forth. The pox varied in size from lo-&apos; to 1CT4 cm, and tended to disappear as cells increased in number and regularity, but as noted,4 the first real array of cells did not seem to be a development from the pox. In fact these authors implied a lack of connection because they stated that the pox are denser on "irregular cells", and as cell boundaries increase in number (i.e., the cells become smaller) there is less need for the pox which do dis- appear. Thereafter, most authors dealing with either experiment or theory have accepted the reality of pox and have used them as a criterion for the onset of constitutional supercooling. In contrast, Spittle, Hunt, and smiths have now suggested that pox are irrelevant artifacts comprised of such things as entrapped oxide. This proposal invokes the observations of weinberg6 and chadwick7 each of whom have shown that the act of decanting leaves a residual liquid on a decanted interface; the remnant solid layer of the order 10 p may thus contain particles that might have been transported from the external surfaces, or elsewhere, during decanting. With the incentive of this suggestion,= some further experiments and a reexamination of the literature have been conducted, in order to question the validity of pox as evidence of an instability and to examine the initial development of the cellular substructure. 1) EXPERIMENTS Single crystals of zone-refined tin (-99.9999 pct) were grown from the melt in a controlled fashion with various, small concentration additions of lead and antimony, for which ko < 1 and > 1, respectively. The crystals were decanted at conditions near the onset of constitutional supercooling and were thus appropriate for observation of slight perturbations. It was possible to observe two types of small departure from smooth or "planar" interfaces in both cases of lead or antimony additions. Some were projections and others, if in regular array of any type, were depressions. The crystals were etched with suitable reagents progressively dissolving the decanted interface surface; projections left no record, but depressions were continuously associated with spotlike areas contrasting with the rest of the interface. Traverses were made with the beam of an electron microprobe across the regions of contrast; with lead addition the persistent spots were lead-rich, and with antimony addition the persistent spots were antimony-poor. This is consistent only with a dominant role for depressions, because if the projections had left spots but were incorrectly catalogued, a reversed observation should have been made; that is, the Pb(ko < 1) should have been depleted and the Sb(ko > 1) enriched. In the work of Cole and inegard, and elewhere, regular arrays of structure associated with the initial stage of instability have been shown, in photographs and represented as pox or projections. We believe this to be erroneous, by inference, since whenever a regular array was observed, in the present examination, it consisted of depressions, regardless of the nature of the solute, ko 1. Fig. 1 is reproduced8 as an ideal example of the possible optical illusion involved; the observer can satisfy himself from the distribution of illuminated areas that the markings are depressions. Fig. 2 from the present investigation is an interference photograph of an interface similar to that in Fig.

Jan 1, 1967

By William M. Boorstein, Robert D. Pehlke

The rates of solution (of hydrogen in liquid pure iron and in several liquid binary iron alloys were meas-ured using a constant volume technique. The rates of absorption and desorption were found to be equal un-der all experimental conditions. increasing concen-trations of S, Si, or Te decrease the rate of hydrogen uptake but additions of Al, B, Cr, Cu, or Ni have no measurable effect up to concentrations normally en-countered in steelmaking practice. No relation ship was found between the effect of an alloying element on the equilibrium solubility of hydrogen in liquid iron and its effect on the solution rate constant. Mathe-rnatical analysis of the data indicates that under the present experimental conditions the rate of reaction of hydrogen with liquid iron is controlled by transport of gas solute atoms in the metal phase. Comparison of the present resuts with data on nitrogen taken un der similar conditions establishes that the hydrody-nurnic conditions which exist near the surface of a metal bath are best approximated mathematically by a surface renewal model for the case of rapid in-ductive stirring and by a boundary layer model for more quiescent melts. HYDROGEN has long been recognized as being a detrimental constituent in steel. If dissolved in the molten metal in excess of its solid solubility, hydro-gen can be evolved during solidification and cause bleeding or porosity in ingots and castings. In the solid metal, lesser amounts play a definite role in causing other defects such as hairline cracks, blisters, and embrittlement. For significant refinements to be made in metallurgical procedures designed to control or eliminate hydrogen from liquid iron or steel dur-ing processing, available equilibrium solubility data must be supplemented with reliable fundamental in-formation pertaining to the kinetic factors involved in the transfer of hydrogen to or from the metal. The scarcity of such information in the literature prompted the present investigation. PREVIOUS RESEARCH Whereas much of the existing data on the solution kinetics of gases such as nitrogen were obtained during the course of thermodynamic investigations, the solu-tion rate of hydrogen has been found too rapid to be accurately determined by conventional solubility meas-urement techniques. Consequently, little work on hy-drogen solution kinetics has been reported in the lit-erature. Carney, Chipman, and crant1 attempted to study the rate of solution and evolution of hydrogen from liquid iron by employing a newly devised sampling method. Although no significant quantitative data could be obtained, it was observed that the rate of solution was approximately equal to the rate of evolution of hy-drogen from the melt. Karnaukov and Morozov2 stud-ied the rate of absorption and Knuppel and Oeters3 the rate of desorption of hydrogen from molten iron by measuring pressure changes with time in a constant volume system. Karnaukov and Morozov determined the hydrogen pressures over their inductively stirred melts with the aid of a McLeod gage and therefore, were forced to work at pressures not in excess of 40 mm of Hg. Their experimental data conformed to a mathematical correlation based on diffusion control: and the rate coefficients calculated on this basis were shown to be independent of the initial absorption pres-sure. These authors reported the solution rate of hy-drogen to be eight-to-ten times higher than they had found for nitrogen in a previous study. They also re-ported that under identical conditions, hydrogen dis-solves somewhat more slowly in iron-columbium alloys than in pure iron. Knuppel and Oeters found that the desorption of hydrogen from pure iron at 1600°C was controlled in all cases investigated by diffusion in the metal bath as long as bubble formation was sup-pressed. This was substantiated by Levin, Kurochkin, and umrikhin4 who studied the kinetics of hydrogen evolution from liquid (technical) iron while applying a vacuum. Salter5 measured the rate of hydrogen ab-sorbed by iron buttons, arc-melted by direct current, as a function of hydrogen partial pressure in a hy-drogen-argon atmosphere. A carrier gas technique was used for analysis of the hydrogen absorbed. The initial rate of absorption was found to increase di-rectly with the square root of the partial pressure of hydrogen. EXPERIMENTAL METHOD Because of the rapid uptake and evolution of hydro-gen by iron-base melts, a constant volume technique was devised in order to obtain meaningful kinetic data over the entire course of the solution process. Apparatus. A schematic view of the experimental apparatus is given in Fig. 1. The hydrogen-liquid iron reaction system consisted of a gas storage bulb con-nected to a meltcontaining reaction chamber through a normally-closed solenoid valve. The gas storage bulb, an inverted 250 ml round-bottomed Pyrex flask was joined to the inlet port of the solenoid valve by a glass-to-metal seal. A more detailed illustration of the reaction chamber is shown in Fig. 2. The design of the Vycor reaction bulb was essentially that de-scribed by Weinstein and Elliott6 with the exception of a shorter, larger diameter gas inlet for this kinetic study. In position, the reaction bulb was closely by an eight-turn coil of water-cooled copper tubing which, when energized by a 400-kc oscillator, provided the inductive heating source. The walls of the bulb were maintained relatively cool by circulating cold water along their outer surface, thus preventing

Jan 1, 1970

By G. A. Vissac

The drying of fine coal involves special techniques, which are discussed and analyzed. Types of dryers employing these techniques are described. Calculations are presented for new methods of dealing with the entrained dust that is always present in fine coal drying operations. NEW conditions, new requirements, and new methods have increased the demand for more efficient and more economical methods of drying fine coal. Dewatering of larger sizes may reduce the surface moisture to 8 or 9 pct. It is more difficult, however, to dewater sizes below 1/4 in., and some filter cakes still contain as much as 20 or 25 pct moisture. Increased freight rates and stricter consumer specifications have resulted in a demand for further reductions in moisture content. This can be obtained only by heat drying. Most modern methods of heat drying disperse or spread the mass of coal to be dried, in an atmosphere of dry hot gases. The more intimate the contact between coal particles and hot gases, the quicker and more efficient the drying operation will be. Two different techniques are generally employed, using either a fluidized condition or an entrained condition of the coal to be dried. Fluidized Condition Fluidization of a body of sand was defined and explained by Fraser and Yancey in a paper published in 1926.&apos; This condition was artificially obtained and maintained by proper regulation of the rate of air flowing through the sand body. "The sand bath &apos;boils&apos; uniformly on the surface," they write, "and feels like a fluid." The fluidization technique was also described and analyzed by Steinmetzer2 in connection with the operation of an air cleaning table. His main conclusions are as follows: "Fluidity is, for the particles involved, the possibility of motion with minimum friction. . . . Then fluidity requires the introduction of various forms of energy capable of neutralising frictions. Two solutions can be used— air and/or mechanical motions (such as the shaking motion of the carrying deck of the air table). The combination of mechanical and air energy will give the widest margins of size ratios and of bed thickness, translated in capacity per unit area of the carrying table." Richardson and Langston3 have indicated results obtained with a dryer working with a fluidized bed. They used a vertical tube type of dryer, however, without the assistance of any mechanical energy, and without any lateral motion of the fluidized bed. The capacity of such a dryer is too limited for practical applications, since the speed of the acceptable air currents is held to the speed of fall of the particles involved. Capacities as low as 182 Ib of coal per hr per sq ft of dryer area are indicated. As stated by Richardson: "A basic limitation to a fluidised bed dryer is that the velocities of the gas must be held within a definite range; with velocities of 10 ft per second, all coal minus 6 mesh in size will be entrained, and the operation is then similar to that of a Flash dryer." A fluidized bed must be virtually static. The coal particles simply kept in suspension offer a minimum resistance to the flow of gases, insuring the most favorable conditions for rapid evaporation of surface moisture. However, very wet fine coal, i.e., over 12 pct of surface moisture, will be delivered in the forms of mud balls, or as a soggy, sticky mass, almost impossible to disperse, sticking and acting as a wet blanket on the deck. Strong currents of gases and wide deck perforations will be required to punch holes in the wet mass and gradually loosen and fluidize it. The mechanics of fluidizing a bed of coal in a gas medium for the purpose of obtaining the most efficient drying condition are entirely similar when the fluid used is water and the purpose is to break up and distend a bed of coal to be cleaned so that perfect stratification according to densities will be insured. Purely mechanical energy is used in the basket-type jig, water pulsations in the piston and in the Baum-type jigs. A combination of mechanical motion and of air pulsation offers the most efficient and favorable conditions. Entrained Condition The most critical factor to be considered in the design of a dryer employing the entrained condition technique is the speed of the hot gases to be circulated in the drying column. With insufficient gas velocity, excessive amounts of the largest sizes will drop to the bottom of the dryer column without being thoroughly dried. On the other hand, high gas velocity will cause degradation, dust losses, and high power consumption. Figs. 1 and 2, reproduced from Hanot,4 show the relative importance of speed and temperature for various sizes of particles. It can be seen, for instance, that to maintain in unstable equilibrium particles of 1/4-in. size in a gas current at 500°C, a speed of 30 meters per sec, or 6000 fpm, will be required. For % -in. particles an almost prohibitive speed of 45 meters per sec, or 9000 fpm, will be necessary. In practice, maximum gas velocities of 3000 fpm are recommended; since power increases as the cube of the velocity, it can be seen that beyond certain limits such dryers would not be economical. If the particles were moving at the same speed as the hot gases they would remain in the same

Jan 1, 1954

By W. J. Slatosky

W. 0. Philbrook (Cairiegie Institute of Technologyogv—Mr. Slatosky has presented an interesting and constructive paper that represents another step along the way of converting steelmaking from an art to a science. I am confident that his computer will be practical and successful and that with a very few months of experience it will provide a significantly better degree of control than his record of 65 pct of heats within range obtained with the slide-rule calculator . A paper such as this, with a lot of symbols and condensed mathematics, is difficult to comprehend quickly. Since I have had an opportunity to study it carefully, perhaps my evaluation of its validity and accomplishments will save time for others. Mr. Slatosky has correctly used standard principles of stoichiometry and heat balances, which are available to anybody, but he has also brought to them two original contributions: 1) He has developed from operating data some empirical relations for predicting the final FeO content of the slag (at 0.5 pct C end-point) as a function of slag basicity, lance height, and scrap, ore, and scale in the charge. This improves the accuracy of prediction of temperature or scrap requirement compared with assuming an arbitrary, constant FeO content at the end of each heat. There is no assurance yet that exactly the same relations will hold for other furnaces or practices, but similar correlations can be expected. 2) He has combined calculations that are ordinarily carried out laboriously as a number of individual steps into a single, simple linear equation that can readily be fed into a machine. This involved a tremendous amount of painstaking detail work as well as the imagination to see the possibility and work out the steps. While his particular Eqs. [3] and [6] are valid only for the furnace design, charge weight, and blowing time used at Aliquippa Works, only a few numerical values have to be changed to adapt it for other conditions. In order to arrive at a useable solution, Mr. Slatosky had to make some basic assumptions about the process that are similar to those used by others. He neglected variation in some process variables and assumed an arbitrary average value for waste gas analysis and temperature for want of more exact information at the present time. All of these judgments are clearly stated. In addition, some thermody-namic data presently available are not adequate for the job, notably in relation to heats of formation and sensible heat in slag, and some expedient has to be adopted to get around the difficulty. Other people might prefer slightly different judgments about these details and hence obtain somewhat different numerical solutions. This is not of serious importance, however, because the errors accumulate in the "heat loss" term and are largely self-compensating for a constant heat time. Although the extended Eq. l(a) in Appendix I was set up as a rate equation originally, for convenience in using an analogue computer as stated in the paper, the time dependence was removed by later mathematical manipulations and assumptions about the process. The final result is an integration of element and energy balances from initial to final states; this procedure is as legitimate here as in any other form of heat-balance calculation. The formal handling of stoichiometry and thermochemistry appears to be correct, and it is assumed that any arithmetical errors would have come to light in applying the calculations to furnace practice. Mr. Slatosky&apos;s approach is not necessarily unique, in that other people might start with apparently different equations or prefer another form of final equation for another type of computer. However, he has presented an accomplished result that appears to be a theoretically sound and practically useful way of applying scientific principles and rapid computation for better control of steelmaking. His success will undoubtedly encourage himself and others to improve on the mathematical model and its use as better informatioq becomes available. John F. Elliott (Massachusetts Institute of Teck-t2ology)-The last comment by Mr. Richards that a calculator is quite unnecessary for an L-D operation ?-equi??es a rebuttal. The L-D furnace is a very high capacity process which places a premium on close control. When one is making steel at rates between 100 and 200 tons per hr, one cannot afford the luxury of an extra 5 or 10 min at the end of a heat correcting for an error that should never have been made in the first place. Mr. Slatosky&apos;s paper is a very sound application of the simple principles of stoichiometry and the energy balance. It is a satisfactory and valuable start, but only the start of the development of methods of control for this process. An analysis of the process shows that it should be very suitable to control by a computer. This is especially the case when various grades of steel are to be made. In fact, it would seem that the organizations who are planning new and bigger installations of L-D vessels should consider carefully the advantages that would stem from computer control of a vessel with the operator present to do little more

Jan 1, 1962

By George W. King

The yield strength of a dispersion-hardened W-3.8 vol pct Tho,alloy, in both the recovered and recrys-tallized condition, was investigated and cornpared with that ofrecrystallized pure tungsten over the temperature range of 325" to 2400°C. It is deduced that the Orowan mechanism is obeyed in the recrystallized alloy. In the recovered alloy, a further enhancement of the yield strength results from the retained substructure which is stable up to temperatures in excess of 2700°C. Temperature and strain rate cycling tests were also performed, and the apparent activation energy for the deformation process was derived. This activation energy, - 3 ev, for the recovered and also the recrystallized alloy was about the same as that for re crystallized pure tungsten. However, the activation volume of the recovered alloy, -10-2 cu cm, was about an order of magnitude lower than that of the recrystallized alloy or pure tungsten. This fact accounts for an enhancement oj- the temperature dependence of the yield stress of the recovered alloy. A dislocation velocity exponent of about 4 to 13 was deduced frorn the strain rate cycling tests , which is in good agreement with values reported for tungsten single crystals. VARIOUS theories have been developed to explain the enhanced yield strength of a metal containing a dispersed second phase of small hard particles. These theories are thoroughly reviewed by Kelly and Nicholson.&apos; The theoretical models can be separated into two types. The first type assumes direct interactions between moving dislocations and dispersoids. One of the most widely investigated models for this mechanism is the bowing out of dislocations between the dis-persoids and their subsequent pinching off in order to bypass the obstacles. This is the well-known Orowan mechanism.&apos; The second type is an indirect effect of the dispersion because of its ability to stabilize to high temperatures the substructure introduced by cold working. In this instance, the increment in the yield strength is expected to be inversely proportional to the square root of the subgrain diameter. In the present work, a quantitative study was made of the strengthening effect caused by a thoria dispersion in a recrystallized W-3.8 vol pct Thoz alloy over the temperature range 325" to 2400°C. The results are compared with the increment predicted for the Orowan mechanism based on the calculations by ~shb~.~ In addition, the alloy was tested in the recovered state so that any additional strengthening resulting from the substructure produced during fabrication could be measured. The respective contributions can be separated in this manner, provided that the particle size distribution of the dispersion remains the same in both the work-hardened and the recrystallized state. Particle size distribution measurements showed that this condition was met in the present work. I) EXPERIMENTAL PROCEDURES A) Material Production and Fabrication. The alloy investigated is essentially the same as that reported much earlier by ~effries,~ who also found the strength of tungsten to be improved by the thoria dispersion. The procedure for producing the alloy consisted of mechanically blending a thorium nitrate solution in proper concentration with tungsten oxide (WO3) powder, followed by hydrogen reduction to metal powder. After reduction, the dispersed second phase is present as thoria (Thoz). The pure tungsten powder used for comparison was produced in the same manner except that the thoria doping step was omitted. The powders were consolidated by cold pressing and self-resistance sintering in hydrogen. The resulting ingot had a cross section about 0.6 sq in. and a density about 93 pct of theoretical. The ingot was swaged to 0.174-in.-diam rod at temperatures varying from 1650°C initially to -1200°C near final rod sizes. Two intermediate recrystallization anneals were employed during fabrication. Analysis of the swaged rods is reported in Table I. B) Electron Microscopy Techniques. Carbon extraction rrPxcas prepared by a technique reported by ~00&apos; were used to quantitatively evaluate the thoria particle size and distribution. Electron nlicrographs of extraction replicas were taken at 20,000 times but were then enlarged two to three times in printing. The areas photographed were randomly selected. A Zeiss Particle Size Analyzer (Model TGZ3) was used to count and measure the sizes of all particles present on the print. About three thousand particles were counted in determining a distribution curve. Electron transmission microscopy was used to determine the effect of annealing on the substructures of the materials. Thin foils were produced by a two-stage thinning process. The rods were first ground on emery paper to ribbons about 10 mils thick and then a jet of 5 pct KOH was used to electrolytically reduce a portion of the cross section of the ribbon. Final perforation was achieved by immersing the specimen in a 5 pct KOH solution and electrolytically polishing at a current density of about 0.3 amp cm-&apos;. The foils were examined with a Hitachi HU-11A electron microscope. C) Tensile Testing. Tensile testing was performed in an Instron Testing Machine equipped with a radiation-type vacuum furnace which operates at about 1O"S torr at temperatures as high as 2400 °C. The same furnace was used for annealing the tensile specimens.

Jan 1, 1970

By R. Grierson, L. J. Bonis

The high-temperature Strength oF TD nickel has been observed to be dependent upon the previons thermal and mechanical history of the material. Variations in both the level and the anisotropy of strength have been observed. 01 this paper- these variations are correlated with the storing of annealing resistant elastic strain energy in the matrix of the TD nickel. An x-vay line -broadening tecknique is used to measure the maLrTis elastie strain. THE inclusion of a finely dispersed second phase into a ductile matrix has long been recognized as an extremely effective method of strengthening the matrix both at high and at low homologous temperatures. It has been found, however, that the factors which determine the high-temperature strength are not the same as those which are important at low temperatures. Below 0.5 Tm the size and distribution of the second phase particles are of prime importance in determining the strength,&apos;)&apos; while above this temperature the strength is mainly dependent upon the previous thermal and mechanical history of the alloy,3-7 This paper is primarily concerned with explaining the response of the high-temperature mechanical strength of one of these alloys (DuPont&apos;s TD nickel) to various thermo-mechanical treatments. It will be shown that this response is not associated with the occurrence of any form of dislocation substructure within the matrix of the alloy. It has been found, however, that a correlation does exist between the elastic strain level in the matrix and the previous thermomechanical history of the alloy and that the observed changes in elastic strain level parallel the measured changes in high-temperature strength. It therefore must be concluded that variations in high-temperature strength are a direct result of the variations in elastic strain level. MATERIAL TD nickel contains approximately 2 vol pct of Tho2 in an unalloyed nickel matrix. It is formed, as a powder, by a chemical technique and this powder is compacted to form ingots which are then extruded to give 21/2-in.-diam rod. Rod of smaller diameter is prepared from the as-extruded rod by swaging. In the studies reported in this paper, 1/2-in.-diam rod was used. This rod received an anneal of 1 hr at 1100°C prior to being used in any of these studies. EXPERIMENTAL TECHNIQUES Two methods were used to examine the structure of the nickel matrix of the TD nickel. These were: 1) transmission electron microscopy; 2) the analysis of the position and profile of X-ray diffraction lines obtained using the nickel matrix as the diffracting media. To prepare thin foils for electron-microscopical examination, slices of TD nickel approximately 0.050 in. thick were cut from the as-received 1/2-in.-diam rod. These were then chemically polished down to 0.045 in., rolled to 0.009 in., given a predetermined heat treatment, and thinned, using a modified Bollman technique, to provide the foils for observation. All observations were carried out at 100 kv, using a Hitachi HU-11 electron microscope. Specimens of the undeformed rod were prepared by grinding down the 0.050-in.-thick slices to approximately 0.015 in. and then thinning chemically and electrolytically to give the thin foils. The X-ray specimens were prepared by rolling 0.375-in.-thick rectangular blocks down to 0.075 in. The surfaces of the rolled material were ground flat, chemically polished to remove the layer disturbed by the grinding, and given a predetermined anneal in an inert atmosphere. They were then ground lightly to check their flatness and given a final chemical polish prior to being examined. The X-ray diffraction line profiles were measured using an automated Picker biplane diffractometer. A special specimen holder was built to allow a more accurate and reproducible positioning of the specimen. The line profiles were determined by carrying out intensity measurements at intervals of either 1/30 deg or 1/60 deg over a range of 3 deg on either side of the nickel peaks of interest. A piece of pure nickel which had been recrystallized to give a large grain size was used as a standard to give the X-ray line profile generated by a strain-free matrix. The analysis of the X-ray diffraction line profiles is a modification of that due initially to Warren and Aver-bach8and has been described elsewhere.3 This analysis gives a measurement of two parameters associated with the structure of the nickel matrix. These parameters are: 1) the size of the coherently diffracting domains within the nickel matrix; 2) the magnitude of the elastic strains in these domains. Both of these parameters are first determined in terms of a Fourier series. These series are obtained from other Fourier series which describe the measured profile of the X-ray diffraction lines. Thus, for both the coherently diffracting domain size and the elastic strain level, it is possible to plot Ft (the Fourier coefficient) against t (the term in the Fourier series), where t can be expressed in terms of a distance L and the Fourier coefficient Ft(S) (associated with elastic strain level) can be expressed in terms of the root mean square strain (e2)1/2. Thus a plot of (F 2)1/2 vs L can be obtained. Plots of this type are shown graphically in Figs. 6 and 8. Interpretation

Jan 1, 1968

By Nicholas J. Grant, Joseph T. Blucher

High-purity aluminum and an Al-10 pet Zn alloy zvere tested in axial fatigue from 80" to 900oF, at struzn vales of 5 and 150 pct per min, at a strain amplitude of 1 pcl. Cycles to failure were recorded as well as the load per cycle during the entive test. Several grain sizes were examined in each material. Examination was made of modes of deformation, initiation and growlh of&apos; cracks, and vecovery mechanisms such as srbgrain formation and boundary migration. Strain rate effects on cycles to failure are first observed ahoi&apos;e 50O0F, the highev vate vesulting in longer lije. Crack initiclion at room temperature may be truns-or iutercrystalline but fructures are transcrystalline. Abore 600&apos;F, crack iniliation and growth ave largely inlercvystalline. Boundary wzigratiotz to 45-deg positions is observed above 70Oo F, and fractrrves are a combination of grain bol~ndary voids and cvacks. It is only in recent years that studies of deformation and fracture which prevail in fatigue at elevated temperatures have attracted significant attention.&apos; Of such studies considerably less attention was given to high strain-low strain rate fatigue. Moreover, the majority of high-temperature fatigue studies were performed at conventional machine speeds (1000 to 10,000 cpm). As it is well-demonstrated in uniaxial creep-rupture series, at high strain rates, even at high temperatures, metals undergo work hardening with little or no attendant recovery or recrystallization thus the nature of deformation and fracture which is observed is similar to that encountered at lower temperatures.&apos;-" Thus, for example, fatigue testing of a stainless steel at 750°F does not involve high-temperature deformation processes,2 and might more correctly be termed "fatigue testing at an elevated temperature". It was the purpose of this work to study deformation and fracture in fatigue as a function of low strain rates and temperature, selecting conditions which would result in grain boundary sliding, migration, fold and subgrain formation, and intercrystalline cracking in high-purity aluminum and a high-purity A1- 10 pct Zn alloy. Grain size was an additional variable. Extensive studies of the deformation and fracture behavior of these aluminum materials in simple creep had been done in the authors&apos; laboratory, and were to serve as a basis of comparison for the observed effects in fatigue:&apos;-&apos;&apos; the range of the creep test temperatures was 80° to 1150oF. MATERIALS AND EXPERIMENTAL PROCEDURE The compositions of the 99.99 pct pure A1 and the A1-10 pct Zn alloy are shown in Table I. Button-head specimens, with a liberal fillet, of 0.20 in. diam and of gage length 0.40 in. were machined from wrought bar stock. The ratio of 2:l gage length to diameter was selected after preliminary tests showed that a shorter length gave a shorter life, probably due to end effects, and after evidence of buckling in longer gage length specimens. After machining, the specimens were chemically polished to remove the worked outer layer, and were subsequently heat-treated to stabilize the selected grain sizes. Both the high-purity aluminum and the A1-10 pct Zn alloy were heat-treated to produce grain diameters of approximately 0.5 and 2 mm in each case. These grain sizes are referred to in the text as fine and coarse grain, respectively. One lot of the high-purity aluminum was heat-treated to produce a still coarser grain size in which the cross section was occupied by 2 to 3 grains. This structure is referred to as very coarsegrained. After heat treatment, the specimens were again electropolished. To avoid complications of both stress and strain gradients in the cross section of the specimen, a hydraulic, axial fatigue machine was designed and built. A button-head specimen, 1/2 in. diam at the head, was firmly gripped in a split-type holder free of any play in the grips. The test temperatures varied from 80" to 900°F. The strain amplitude in all of the reported tests was 1 pct for a total strain amplitude of 2 pct. The strain range was set by precision micrometers and measured by a precision dial gage. Constant strain rates of 5 and 150 pct per min were selected so that high-temperature type deformation and fracture would occur in the higher-temperature tests5,6 The strains and strain rates must be regarded as nominal values because they are based on the original specimen dimensions, which changed significantly as a result of necking and crack propagation, as can be observed from Fig. 8. For the elevated-temperature tests, a thermocouple was inserted into a well in the head of the specimen; the selected temperatures could be maintained with less than ± 5oF fluctuation during the entire test. To avoid changes in grain size before the test, specimens were heated to the test temperature in less than 15 min; similarly, they were cooled to room temperature after fracture with an air blast to avoid or minimize recovery or recrystallization. During the fatigue tests, load vs strain curves were recorded by a strain gage load cell for each fatigue cycle. In addition, the maximum values of load amplitude were recorded for the entire test.

Jan 1, 1968

By Kiyoshi Azuma, Hiroshi Kametani

The variation of the dissolution behavior of a ferric oxide with calcining temperature has been investigated. Samples were prepared by thermal decomposition of ferric hydroxide, nitrate, oxalate, and sulfate at low temperature, followed by the calcination in the temperature range between 600" and 1200°C. The samples of eight series and a fine crystalline sample of hematite were dissolved in 1 N hydrochloric acid at 55.2°C and the results are represented on double-log graphs for convenience. It is confirmed that all dissolution courses follouj either the accelerated process or the parabolic process except in the special case of the crystalline hematite which dissolced in accordance with the uniform dissolution of a particle. Examinations of the physical properties of the oxide powders revealed that the surface area measured by the permeability method is strikingly relevant to the dissolution behavior of the oxide. In the previous paper,&apos; detailed data were presented on the effect of the kind of acid, the solution temperature, and the concentration of acid on the dissolution of two ferric oxides. It was also shown that these sam ples dissolved in strikingly different ways. The present investigation was carried out on the dissolution of various calcined samples prepared from various ferri salts by various methods to ascertain the course of dissolution. Pryor and Evans2 pointed out a change of the dissolution rate at around 700°C for a series of calcined ferric oxides prepared from the hydroxide. Several papers374 reported also the dissolution of ferric oxide samples. It seems, however, that a systematic account of the relationship between the dissolution behavior and physical properties of the oxide has not yet been given. This paper presents the variation of the dissolution of the oxide in relation to the calcining temperature and the change of physical properties of the calcines. EXPERIMENTAL Raw materials were prepared by precalcination of ferric hydroxide, thermal decomposition of ferric nitrate, oxalate, and sulfate, and aerial oxidation of ferric chloride vapor, at as low a temperature as possible. The products were crushed, ground, if necessary, and sieved with a 100-mesh Tylor screen prior to calcination, after which the specimens were dissolved in acid solution. The following is a detailed description of the preparation of the samples. Sample H. About 500 g of ferric chloride (guaranteed reagent) were dissolved in 5 liters of deionized water and filtered. Ferric hydroxide was precipitated by addition of the minimum amount of ammonium hydroxide solution, and the precipitate was washed continuously till chloride ion was not detected by silver nitrate solution, and then filtered. The filter cake was dried at 120°C for a week and ground, and the -100 mesh portion was used. Sample S. Ferric sulfate (guaranteed reagent) was pyrolytically decomposed in a crucible at 700°C for 24 hr and the product was sieved. In this case the following calcination was carried out at temperatures over 700°C. Sample B. Commercial ferric oxide (guaranteed reagent). About 15 kg of ferric nitrate were decomposed in a furnace maintained at 800°C for 2 hr. The actual temperature of the decomposition was not measured. The product was crushed and sieved, and the -100 mesh portion was used. Sample N. About 50 g of ferric nitrate (guaranteed reagent) were decomposed in a beaker in a sand bath until a red-brown dense solid was produced. This product was crushed and sieved, and subjected to complete decomposition at 500°C. The precalcined product was again sieved and used. Sample N2.5. Since the decomposition temperature was not controlled for sample AT, a different sample was prepared in a temperature-controlled furnace. The subscript represents the decomposition at 250°C. The product was treated in the same manner as sample N. Sample Nc. Under atmospheric pressure it is prac-tically inevitable that ferric nitrate hydrate melts to form a brown liquid at about 50°C before pyrolysis. For this reason, the salt was first slowly heated under reduced pressure (about 10-3 mm Hg measured in a trap refrigerated by dry ice-alcohol) to achieve dehydration without melting. About 5 hr were required for the dehydration and the partial decomposition. Then the temperature was elevated to 500° C in air for complete decomposition. The relatively porous product was sieved and used. Sample Ov. About 200 g of ferric oxalate hydrate (extra pure) were dehydrated under reduced pressure (as described above) followed by thermal decomposition at 500°C for 6 hr in air. The decomposition of this salt was accompanied by liberation of carbon monoxide, by which the ferric salt was initially reduced to a black powder. The powder changed in turn into brown ferric oxide as the gas liberation decreased and reoxidation predominated. The product consisted of sparkling fine particles passing through a 100-mesh screen. However it was ground and sieved as for the other samples. Sample D. Commercial fine powder for magnetic tape purposes. The preparation was as follows.5 Ferric chloride vapor and preheated excess air were mixed and passed into a reaction tube where oxidation took place at 450°C. The fine powder formed was collected in a cottrell chamber. The product was vacuum-degassed at 450°C for 1 hr and sieved.

Jan 1, 1969

By R. Bainbridge

THE Consolidated Mining and Smelting CO. of Canada Ltd. has operated a lead smelter at Trail, B. C., for many years. In order to take advantage of metallurgical advances, as well as to improve materials handling methods, this company, commonly known as "Cominco," commenced planning a program of smelter revision and modernization some years ago. The first stage of this program involved the design and construction of a new blast furnace gas cleaning system. The selection of equipment, the design of facilities, and preliminary operating details of this system will be dealt with in this paper. The essential problem was to clean and collect 100 tons of dust daily from 153,000 cfm* (12,225 lb per min) of lead blast furnace gas which varied in temperature from 350º to 1100°F. Because it was desired to collect the dust dry, either a Cottrell or a baghouse cleaning plant was to be selected. Comin-co&apos;s many years of experience with both systems provided a background for choosing the most satisfactory installation. All information pertinent to the two methods of dust recovery was carefully investigated, and it was decided to replace the existing equipment with a baghouse. Very briefly, the reasons for this decision were as follows: 1—A baghouse installation would be practical because the SO2 content of the gas was low and corrosion would not be a problem if the baghouse operating temperatures were held sufficiently above the dew point. 2—Variations in the physical characteristics of fume and dust, which are inherent in this blast furnace operation, should not substantially affect the operating efficiency of a baghouse. 3—For the same capital cost, metal losses (stack and water losses) would be appreciably less in a baghouse. 4—A baghouse would be easier to operate, and would not require the use of highly skilled labor. 5—Operating and maintenance costs of a bag-house would be lower. 6—The only available space for reconstruction was relatively small, and not suited to a Cottrell installation. Once the baghouse system was decided upon, detailed design of the installation was begun. Baghouse Design Gas Cooling: Before the required capacity of the baghouse could be determined, the method of cooling the gas to the temperature necessary for bag-house operation had to be chosen. The problem confronting the design engineers was how best to cool 153,000 cfm of gas from a temperature ranging from 350°F to brief peaks of 1100°F, down to 210°F, the maximum safe baghouse inlet temperature. A survey of existing blast furnace gas temperatures in the outlet flue showed that the normal range was as given in Table I. The obvious choices of cooling method were: 1— cool completely by the addition of tempering air; 2—utilize a heat exchanger; 3—cool by radiation; and 4—cool with water spray in conjunction with the admission of tempering air. The advantages and disadvantages of the various cooling methods were: Air Addition: To cool completely by the admission of tempering air involved tremendous volumes, Fig. 1. For example, to cool 1 lb of blast furnace gas at 450°F requires 1.84 lb of air at 80°F or 1.60 lb at 60°F. As it is necessary to design for peak conditions, it can readily be seen that volumes of tempering air in the order of 1,500,000 cfm would have to be handled. Using the normal design figure of 2.5 cu ft per sq ft of bag area, a baghouse installation comprising some 600,000 sq ft of filter cloth would be necessary. Such design requirements would be prohibitive, not only from a standpoint of capital expenditure, but also because of space limitations. Heat Exchanger: The utilization of a heat exchanger was given serious consideration. A horizontal tube unit using air as the medium to cool the required volume of blast furnace gas from 400" to 250°F was investigated. Cooling above 400°F would be done by water spray, and below 250°F by admission of tempering air. The estimated capital cost of such a unit was found to be prohibitive. From an operating standpoint, there was considerable doubt as to whether the soot blowing equipment provided would effectively keep the dust from building up on the tube surface. The performance of heat exchangers operating on dusty gas in other company operations had not been too favorable. Radiation Cooling: Although somewhat cumbersome, gas cooling by radiation from &apos;trombone&apos; tubes or other similar equipment (cyclones) is employed in many metallurgical operations. Such an installation was also considered. However, calculations showed that an installation much larger than the space available would be required to handle the gas volume involved. For example, to cool 153,000 cfm of blast furnace gas from, say, 600&apos; to 250°F (i.e., remove in the order of 58,500,000 Btu per hr with heat transfer rates varying from 1.1 Btu per sq ft per hr per OF for the higher temperature ranges to 0.88 Btu per sq ft per hr per OF for the lower ranges) would need a cooling area of some 175,000 sq ft.

Jan 1, 1953

By J. J. Frawley, W. J. Childs

The dynamic nucleation of supercooled bismuth and Bi-Sn alloys has been studied over a frequency range of 15 to 20,000 cps. For low-frequency vibration, a minimum vibrational energy was required for enhancement of nucleation. Above this critical energy, the dynamic supercooling was less than static supercooling showing that vibration promoted nucleation. The amount of dynamic supercooling continued to decrease with increasing vibrational energy until a minimum or threshold value was reached. This minimum value of supercooling for nucleation remained constant joy all further increases in vibrational energy. For higher frequencies, similar results were observed. This behavior has been related to the necessity of cavitation for dynamic nucleation. When a liquid is cooled to a temperature below its equilibrium melting point, the solid phase is more thermodynamically stable. However, for solidification to occur, a two-step process, nucleation and subsequent growth of the solid phase, must occur. When a liquid is supercooled, that is cooled below the equilibrium melting point, the controlling process for solidification to begin is the rate of nucleation. Once nucleation has occurred, the solidification process is controlled by the rate of growth. Nucleation can be induced by two factors: either by a catalyst or by the use of mechanical shock. Numerous investigators1-4 have studied the effect of nucleation catalysis but much less systematic study has been made of nucleation by mechanical shock waves. The influence of vibrations on grain size in castings and ingots has been studied by many authors but no clear understanding of the mechanism or accurate prediction of the effect has been presented.5 It would be intuitively expected that the further the departure from equilibrium (i.e., the greater the supercooling), the easier it would be to induce nucleation. This has been quantitatively demonstrated both by walker6 and later by Stuhr,7 that the greater the degree of supercooling the easier it is to nucleate by a shock wave. Stuhr also attempted to obtain the mechanical energy required for nucleation of bismuth as a function of supercooling. He vibrated a crucible containing supercooled metal at low frequencies and various amplitudes and noted the corresponding dynamic supercooling obtained. The amount of supercooling was inversely proportional to the mechanical energy applied. Limitation of his experiment was the problem of the confinement of the liquid in the crucible without splashing and minimizing other unwanted modes of vibration. Tiller et al.8,9 did similar work on tin and Sn-Pb alloys using an electromagnetic stirring device. Their conclusions were that the magnitude of the magnetic field strength did not affect the amount of undercooling at which nucleation was initiated. While conclusive experimental results have been lacking to explain this effect of mechanical vibration on inducing nucleation, a number of theories have been proposed. Two of these theories are discussed below. 1) The Change in Melting:- Point Locally Due to the Change in Pressure (Clapeyron Equation). According to Vonnegut10 the most plausible explanation for the nucleation of a supercooled melt by cavitation is the effect of changing the melting point by a change in pressure. For materials where the volume decreases on solidification, an increase in pressure raises the melting point; for materials which expand on solidification, the melting point is raised for a decrease in pressure, i.e., rarefaction. Using the Clapeyron equation, the melting point of a metal can be calculated as a function of pressure. If it is assumed that the equation can also be used to calculate the temperature of nucleation of a supercooled melt as a function of pressure (i.e., the temperature of heterogeneous nucleation will increase with pressure at the same rate as the melting point), the amount of supercooling required for nucleation will be constant at all pressures as shown in Fig. 1. It is obvious that an isothermal change which results in an increase in melting point results in an equal increase in supercooling. This increase in supercooling may now be sufficient for nucleation. A pressure of 80,000 atm was calculated, using the Clapeyron equation, as the pressure required to increase the temperature of nucleation of nickel by 200°C. According to Lord Rayleigh,11 this very large pressure could be generated for a very brief period of time by the collapse of a cavity. This pressure wave is radiated in all directions from the collapsed cavity. If the temperature of the melt is slightly below its equilibrium melting temperature at atmospheric pressure, stable growth can follow; that is, once nucleation occurs, growth becomes the main driving force of the solidification process. This proposal has been extended to water which expands on freezing by assuming that nucleation occurs during rarefaction following the pressure pulse. This negative pressure pulse should follow immediately after the positive pressure pulse with its magnitude approaching the critical tensile strength of the liquid. The negative pressure developed during this period would raise the melting point of water and thus promote nucleation. Hunt and jackson12 have suggested this for water. Similarly, it could be postulated that bismuth which also expands on freezing could be nucleated during the negative pressure pulse. 2) Nucleation by a High-pressure Phase. An extension of the Clapeyron equation to systems where density decreased on freezing at atmosphere pressure has been proposed by Hickling.13 The phase diagram for water initially shows the well-known decrease in

Jan 1, 1969