Search Documents

By Joseph J. Becker

RECENTLY the "oriented growth" hypothesis of recrystallization textures has been receiving considerable attention. According to this view, "whenever a new generation of grains is growing in a highly oriented matrix . . . one may assume . . . that from among available nuclei of a great variety of orientations those with suitable orientation relationship with respect to the matrix have by far the highest rate of growth (oriented growth) and will, therefore, predominate."' This statement includes primary recrystallization of single crystals and of polycrystalline materials showing strong deformation textures as particular cases. While the phenomenon of oriented growth undoubtedly does operate, and may be of importance in coarsening or secondary recrystallization, the writer feels that its extension to primary recrystallization is a different matter. Several aspects of this phenomenon may be pointed out which are difficult to rationalize on the basis of oriented growth. In recrystallizing deformed single crystals of alpha brass, Maddin, Mathewson, and Hibbard2 found that the twin composition planes of the recrystallized grains were parallel to the three operative slip planes they observed, and not to the fourth octahedral plane. The writer's current work seems to indicate further that when only one or two slip systems can be seen to operate, the recrystallized grains are aligned only with those slip planes actually observed. This behavior could hardly be accounted for by oriented growth, because on the latter basis the recrystallized texture of a deformed single crystal should certainly have the same symmetry as the crystal itself. Growth rates could not be expected to distinguish between crystallogra- phically equivalent directions, unless the presence of a visible slip system vastly increases the growth rate in the appropriate direction. This seems unlikely. These relationships were observed by the writer only when considerable care had been exercised before annealing to etch away the effects of cutting out the centers of tensile specimens with a jeweler's saw and to etch off the ends of compression specimens. When these.precautions were not taken, nu-cleation inevitably appeared to occur at the saw cuts or at the ends, and the resulting orientations never showed a [111] rotation but appeared quite random. The provision of many nuclei of many orientations did not enhance the correlation, but destroyed it. Recrystallization textures are sometimes strongly dependent on annealing temperature, which would imply that the anisotropy of growth rates is a suddenly varying function of temperature. This might be regarded as unlikely. Recrystallization textures can be strongly dependent on the degree of deformation. If textures are to be explained by selective growth from among nuclei of all orientations, it is difficult to see how the recrystallization process could have any dependence whatever on the manner or degree of deformation. It is felt that the above points should be considered before the oriented growth hypothesis is applied to primary recrystallization, with the necessary complete divorce of the mechanism of recrystallization from the details of the deformation process. References 1P. A. Beck and H. Hu: Reply to discussion of their paper on Recrystallization Texture and Coarsening Texture in High Purity Aluminum. Trans. AIME, 188, 1057; Journal oF Metals (Aug. 1950) TP 29083. 'R. Maddin, C. H. Mathewson, and W. R. Hibbard, Jr.: The Origin of Annealing Twins in Brass. Trans. AIME, 185, 655; Journal of Metals (Sept. 1949) TP 26763.

Jan 1, 1952

By M. T. Simnad, C. E. Birchenall, M. H. Davies

THE marker movements observed by Davies, Simnad, and Birchenall during the growth of magnetite on wustite have been misinterpreted. It is the purpose of this note to correct the original interpretation. In that investigation wustite specimens were prepared by oxidizing pure iron disks in water vapor until the iron was consumed. Argon was used as a carrier gas for the water vapor and on the complete conversion of the metal to its oxides ('FeO' plus a small amount (10 pct) of surface Fe3O) the water supply was bypassed and the specimens were annealed in argon alone in order to equilibrate the wustite with the overlying magnetite. After cooling, the magnetite was ground off, a crude magnetic test being employed to determine its complete removal. Radioactive silver streaks were then applied to the surface and the specimen was oxidized completely in water vapor to prevent hematite formation. After complete oxidation the markers were found to enclose about 80 pct of the magnetite with the other 20 pct lying outside the markers. This observation was interpreted to mean that oxygen transport was four times as great as iron transport during the growth of the magnetite. The conclusion would have been correct if, as in the case of wustite growing on iron, all the iron had started from one side of the marker and all the oxygen from the other side initially. Actually a material balance is required. For simplicity, take wustite to be exactly FeO, and magnetite to be Fe3O,. Then at the start of the reaction 3 FeO + O -> Fe,O., All of the iron and three-fourths of the oxygen of the final product will be within the marker inter- faces. The other fourth of the oxygen is in the gas phase outside the marker. From the experimentally determined final position of the markers simple calculation shows that one-fifth of the total iron has crossed the marker interface, while one-twentieth of the total oxygen has crossed in the reverse direction. The simplified result favors iron transport in the ratio four to one. However wiistite is not stoichiometric FeO. At 900°C, the temperature of the measurement, the composition is about FeO When the material balance for the same experimental result is repeated, assuming the product will be 1 mol of Fe3O4, it is found that 0.6 gram atoms of iron and 0.19 gram atoms of oxygen must have passed the marker in the same direction. (This is readily apparent from the reaction to form 1 mol of Fe3O4 SFeO The total oxygen supplied from the atmosphere is 0.61 gram atoms, but the iron which passes the marker interface, i.e., 0.6 gram atoms, requires 0.8 gram atoms of oxygen to convert it to Fe,O,.) If all the transport during magnetite growth is due to iron ion diffusion, then the marker would enclose about 85 pct of the final magnetite and the original measurement is in error by about 5 pct. This error is probably less than the uncertainty in locating the markers by autoradiography. The possibility also exists that the starting wustite was not quite saturated with oxygen and underwent further enrichment with consequent envelopment of the marker before magnetite caught up with the marker interface. In any case it appears that these measurements actually favor iron ion diffusion in magnetite overwhelmingly as the factor controlling its growth, although further experimental measurements are needed to eliminate the error in material balance described. Such studies are planned.

Jan 1, 1954

By P. Greenfield, P. A. Beck

RECENT work by J. Darby, Jr. in this laboratory has shown that the reported vanadium-rich limit of the s phase in the Co-V system at 1200°C is in error. The appearance of a second phase in s actually is restricted to alloys containing over 68 atomic pct V. The error in the original value is due apparently to a reaction that takes place along the numerous cracks in the s specimens between the alloy and the impurities present in the He + 8 pct H atmosphere in the annealing furnace. Such a reaction and the consequent appearance of a second phase can be suppressed if the alloys are sealed under vacuum in Vycor tubing. The new value of 68 atomic pct V for the vana-dium-rich limit of the s phase in the Co-V system effects slightly the s compositions calculated from average values of N, and N,. for various s phases. The standard deviation between experimental mean compositions and those calculated from the assumption that s is characterized by a constant number of 3d + 4s electrons per atom (N,), is slightly reduced from 9.1 to 8.8 pct. The maximum divergence for the V-Mn s phase is reduced from 15.5 to 14 atomic pct. The standard deviation calculated from constant electron vacancy numbers given by N.. = 4.66 (Mo + Cr + V) + 3.2 (Mn) + 2.2(Fe) + 1.71(Co) + 1.6(Ni) is now greatly reduced, as the maximum divergence for this correlation previously occurred in the V-Co system. This maximum divergence is reduced from 7.6 to 2.3 pct, and the standard deviation for all s- phases is reduced from 4.1 to only 2.9 pct. The correlation for electron vacancy numbers given by N. 5.66 (V when with Ni or Co) + 4.88 (V when with Mn or Fe) + 4.66 (Cr + Mn) + 3.3 (Mn) + 2.66 (Fe) + 1.71 (Co) + 0.61 (Ni) is made slightly worse, and the standard deviation increases from 2.8 to 3.9 pct. These changes, however, do not affect the conclusions of the paper.' The deviation between calculated and experimental s compositions is much less for either one of the electron vacancy correlations than for the correlation with the number of 3d + 4s electrons.

Jan 1, 1955

By H. L. Sauder, J. L. Newman, C. Waddell

An instrument capable of measuring subsurface flow rates is described. The instrument is self-contained and may be run on piano wire line. It detects flow by means of an impeller suspended between two torsion wires. The force of the well fluid striking the impeller causes the impeller to rotate, exerting a torsional force to the wires. This force is determined by recording the angle of rotation of the impeller on film using a battery-driven, clock-controlled camera. The impeller is sensitive to flow in either direction and the same instrument can be used to measure injection as well as production rates. Changes in direction of flow, such as might be caused by thief zones, are measurable and are indicated by a reversal of the direction of rotation of the impeller. Adjustment of the sensitivity of the instrument to measure a wide range of flow rates is accomplished by the use of different size torsion wires. An umbrella-type fluid trap, which contacts the casing or wellbore, diverts the flow through the flow tube. The fluid trap remains closed while rurrning in the hole and can be opened at any point in the well. The instrument with the trap closed is 1 3/4 in. OD and may be run through 2 in. tubing. Successful flow profiles have been made on wells with flow rates ranging from 60 BID to 4,000 B/D at surface pressure up to 4,500 psi. The instrument is designed for high-pressure and high-temperature operations. INTRODUCTION The need for an instrument that will measure flow rates in a wellbore is as old as reservoir engineering itself. A number of such devices have been designed within the past few years. Some of these have been described in the literature1,2,3,4,5,6 One of these devices1. a hot-wire anemometer type, made many successful jobs in dry gas producing areas. Difficulty was encountered when this flowmeter was used in wells producing liquids and development work was initiated to overcome this difficulty. The present flowmeter is the result of this work and provides an instrument with a wider range of application in the industry. Consider- able development work has been done on the tool involving many changes in the mechanics of recording the flow readings, however, the basic principle of flow detection has remained the same since its inception. PRINCIPLE OF OPERATION The basic principle of operation of the flowmeter is the utilization of a portion of the kinetic energy of flowing well fluid to impart a rotational force to an impeller suspended between two torsion wires. This impeller, Fig. 1, is a screw-type propeller with the blades having a pitch angle of 40. As the well fluid passes through the instrument it strikes the impeller. The resulting force rotates the impeller and twist\ the torsion wires to a balance point at which the torque imparted to the wires balances the force required for the impeller to deflect the well fluid. The impeller is symmetrical and a change in direction of the flow results in a reversal of the direction in which the impeller is deflected. Thus the direction of the flow as well as its velocity is indicated. A dial is attached to the impeller by means of a hollow tube through which the upper torsion wire extends. The angular position of the impeller is recorded by photographing the dial at frequent intervals. In any given set of torsion wires the modulus of shear and polar moment of inertia remain constant and the deflection varies directly with the applied torque. The deflection imparted to the torsion wires also varies as the square of the fluid velocity. Further, the velocities represented by any two given deflections hear the following relationships: and Where V = velocity of flow stream and O ;= deflection. This is the basic relationship which is used to calculate the relative velocities represented by a given set of deflections where the velocity represented by any one of the deflection readings is a known quantity. For a given torque, the deflection of the torsion wire varies inversely as the fourth power of the wire diameter. Wire sizes ranging from ,005 in. in diameter to .022 in. in diameter can be used. The range of

Jan 1, 1957

By William L. Russell

A new instrument for measuring bulk and grain vol-urnes is described, and compared in accuracy with two modifications of Nutting's method, All three methods are reasonably fast, and in all three the probable error in the determinations of bulk volume can be kept down to about 0.1 to 0.2 per cent. INTRODUCTION The new pycnometer described in this paper is designed to give speed without sacrificing accuracy. Some time ago Russell1 described a somewhat similar appa-ratus; however, the instrument described in the present paper is more accurate. Nutting' designed the pycnom-eter shown in Fig. 1 (A and B), and Pirson3 has shown points like those of Fig. 1-D. The procedure described in this paper for using Nutting's pycnometer has been modified from that used by Nutting. CHOICE OF LIQUIDS The best liquids for measuring grain volumes are hydrocarbons or chlorinated hydrocarbons which evaporate slowly. Trimethylbenzene or mesitylene, triethylben-zene, carbon tetrachloride, and tertalin have been used for this purpose. Any of these liquids or mercury may be used for measuring bulk volume. PREVENTING ERRORS DUE TO INVASION OF LIQUID Errors due to the invasion of the porous, permeable bulk samples by the liquid of measurement may be prevented or greatly reduced by using mercury, by applying an impervious coating, or by soaking the bulk samples in the liquid of measurement until bubbles cease issuing from them. Errors are produced only by the invasion of the samples which occurs between movement when the samples are weighed in air and the moment the weight of liquid displaced by the sample in the pycnometer is determined. No error is produced by varying degrees of filling of the pore space of the samples by the liquid of measurement. The excess liquid should be wiped from the samples immediately before the weight in air W1, is determined, and they should be placed in the pycnometers immediately. OPERATION OF NEW INSTRUMENT The new pycnometer is shown in Fig. 2. It consists of two glass parts which fit together at a ground glass joint. The spaces J and D are connected by fine glass tubes E, and the space J must be larger than the space D by at least the volume of the largest sample to be tested. The projections H afford a grip for twisting when the two parts tend to stick together. The position of the sample in mercury is shown by I, and in other liquids by M. F is a permanent mark on the glass, and G is a tem-Some marker. In order to determine the position at which G should be placed, invert the instrument so that A is at the bottom, and fill space D. Place the cub K

Jan 1, 1958

By T. M. Geffen, R. E. Gladfelter

Lipson1 has recently presented a technical note wherein theoretical considerations were used to demonstrate. that the adsorption us saturation relation for a linear absorption method is not necessarily a single valued function of saturation.1 It was concluded that the absorption us saturation relationship is subject to the same hysteretic effect as is the flow behavior between the drainage and imbibition systems. It is the purpose of this note to show, by experimental results, that for the manner in which fluids are distributed in porous materials during fluid flow, the relative X-ray absorption us saturation relationship is single-valued regardless of the saturation history. The design of the apparatus used to measure relative X-ray absorption of cores undergoing fluid flow tests is similar to that described by Morgan, McDowell and Doty. A few modifications have been incorporated in the design to increase sensitivity and simplicity of measurement. The sensitivity is such that for usual types of cores a recording pen movement of approximately 12 linear in. is occasioned between the recording of the core dry and when it is completely full of an absorbing liquid, a solution of approximately five per cent by volume of iodo-benzene in a close-cut hydrocarbon fraction, or a solution of 50,000 ppm sodium iodide in water. Two types of experiments were conducted, one in which the liquid was inside the core, and the other with the liquid outside the core. Tests with the liquid inside the core were run to cover the saturation range encountered in both the drainage (gas drive) and imbibition (liquid flood) flow systems. It is pointed out that the same core was used, but the tests were conducted at different times with the recording mechanism balanced at a different position. Therefore, the absolute values of relative X-ray intensity at any saturation are not the same in each test. In the second type of test the same core was used, but this time its pore space cmtained only gas, and various thicknesses of oil contained in plastic cells were placed in front of the dry core. As before. the relative X-ray ah- sorption was measured and compared to the quantity of oil in front of the core. The arrangement of the fluids with respect to the path of X-rays in these two types of tests represent fluid distributions of a wider difference than can be expected between that established in a core just due to imbibition and drainage saturation flow histories. Thus any deviation from a straight line relationship of the log of the relative X-ray intensity us saturation calibration should be readily apparent. Results of the laboratory tests are given in Fig. 1. A schematic diagram showing the position of the fluids with respect to the path of the X-ray beam is given for the two cases investigated. In the case where the oil existed outside the core (Case B), the log of the X-ray intensity (log, I/I,) is related directly to the cell thickness (amount of oil in X-ray beam path). This correlation is in full agreement with X-ray theory. In the case where the oil or water existed inside the core (Case A) which covers both drainage and imbibition types of flow processes, again the relationship between absorption and saturation (amount of oil in X-ray beam path

Jan 1, 1952

By M. K. Yen

The distortion of crystal structures referred to as "biege gleitung",1 "local curvature"2-4 and "deformation bands"5,6 is believed to be an unavoidable characteristic of deformed single crystals. The presence of asterism on Laue photograms of Al single crystals either deformed by pure shear7 or extended in a condition relatively free from the grip effect2 indicates strong evidence substantiating this point of view. The type of localized bending described in connection with the bending test of Al single crystals by Yen and Hibbard8 could conceivably exist either microscopically or submicroscopically in all types of deformation. The purpose of this note is to extend this idea to the postulate of a hypothetical "microbending site," as applied to the mechanism of slip in a general sense. The mosiac structure, as postulated by Darwin,9 is that crystals were actually composed of small blocks which have a linear dimension 10—4to 10-6 cm on the edge and deviate from each other by an angle of about one-tenth of a degree. The following discussion is based on the assumption that when slip is propagated from one mosaic block to the adjacent one, bending of the lattice at the region of their boundary is produced in order to maintain the continuity of the gliding process along a certain slip plane. Consider a two-dimensional diagram of two ideal mosiac blocks on the cross-section perpendicular to the slip plane and containing the active direction of slip. When an external force is applied, the localized bending or "microbending site," would be generated at the region as illustrated in Fig 1A at point B. A schematic diagram of a "microbending site," generated at a mosaic boundary, is illustrated in Fig IB. It may be seen that once the curvature of the bending site reaches sufficient magnitude, the interaction force encountered by pairs of atoms symmetrical with respect to the center of the bending is equal but opposite in sign. Consequently, the overall energy required to displace the atoms near the center is approximately zero.10"12 Based upon the finding that the misalignment between two mosaics is 0.1 degree,9 the assumption is made that "microbending sites" may possibly exist in the region where one plane will have one more row of atoms than an adjacent plane. It is probable that the origin of the bending site will be where the distance between two mosaic boundaries is "2d" as shown in Fig 1A at point B, where " d" is the interatomic distance. Thus, the distance between two "microbending sites" is in the range of about 600 interplanar spacings. In the case of aluminum, the spacing between two octahedral planes is about 2.33 A. The distance between two bending sites will be approximately 2.5 X 10"5 cm. It is interesting to note that according to Woods13 these values approach closely the size of the domains of a cold-worked metal whose minimum

Jan 1, 1950

By R. P. Carreker, J. C. Fisher

Cheep tests are normally constant load tests. Such tests approximate some types of service conditions and therefore are justified from the engineering point of view. Coupled with this consideration is the advantage of simplicity inherent in constant load tests, as contrasted with the comparatively complex requirements of a constant stress test. It is not surprising that the vast majority of creep tests reported are of the constant load type. Despite the enormous amount of accumulated creep test data, plastic flow is by no means completely understood. The widespread use of constant load tests is at least partly responsible for this lack of understanding. Any attempt to analyze deformation must involve the more fundamental quantity stress, rather than load. It is possible to predict constant load behavior from known constant stress data, but the reverse is much more difficult. When a metal specimen elongates, as in a creep test, the cross section decreases, maintaining constant volume. Thus For a constant stress test the force on the specimen must be decreased as the cross section decreases, that is, this force must be inversely proportional to the elongation of the specimen. An-drade has described three methods of obtaining the constant stress condition,l-3 each of which is relatively complex. The most recent proposal,³ suggests a major simplification which maintains the simplicity of the constant load test while approximating quite closely the desirable condition of constant stress. Andrade's device is admittedly more accurate but the method described below has definite advantages Consider a specimen in the form of a "V" supported at its ends. If a weight be applied at the vertex of the angle, the stress in each leg of the "V" If this stress causes an elongation ?1 in each leg of the "V," the angle decreases from a. to a smaller angle a1. The quantity sec — thus decreases with specimen elongation while the quantity 1/A increases. By selecting proper initial conditions it should be possible to hold S/So very nearly constant at unity. The proper initial conditions are realized when a0 is made equal to approximately 90°. Fig 1 compares the ratios of S/S0 for a0 = 85, 90, 95, 100°, with a0 = 0, that is, a straight speci- men, for true strains up to e = 0.10. It may prove convenient to measure X, the movement of the vertex, rather than the true elongation 1. Contour lines of X/lo are also plotted in Fig 1. The method for obtaining constant stress described above is especially suited to the testing of small wires, but may easily be extended to rods of any diameter through the use of a suitable grip which serves as the vertex joining two identical rods forming the legs of the "V." References 1. E. N. da C. Andrade: On the Viscous Flow in Metals, and Allied Phenomena. Proc. Roy. Soc., A, 84, 1, (1910-11). 2. E. N. da C. Andrade and B. Chalmers: The Resistivity of Polycrystalline Wires in Relation to Plastic Deformation, and the Mechanism of Plastic Flow. Proc. Roy. Soc., A, 138, 348, (1932). 3. E. N. da C. Andrade: A New Device for Maintaining Constant Stress in a Rod Undergoing Plastic Extension. Proc. Phys. Soc., 60, (3). 304, (March, 1948).

Jan 1, 1950

By John J. McKetta. Jr., Louis B. Lesem, George H. Fancher

The proration of oil produced in the field frequently is based partially or entirely upon the gas-oil ratio of wells. The measurement of the gas-oil ratio is one of the more important field tests in regulatory and pro-ration work, and the test always should be conducted according to standardized methods and procedure. Obviously, the gas-oil ratio and the volume of gas produced by a well depend upon many factors but should be independent of the method of measurement and of the devices used to measure gas and oil. Consequently, the volume of gas accompanying a barrel of oil produced by a well may be measured by any reliable and accurate device or instrument. Frequently either a critical flow prover or an orifice well tester is used for this purpose, and for a particular well the same rate of flow of gas should be obtained regardless of whether a critical flow prover or an orifice well tester is employed in the test. In Texas, when using either instrument, either Capacity Table 1 or 5' is employed in making the necessary computations. If the tables are used, a discrepancy always is found whenever the two instruments are compared by extrapolation to the same conditions of flow. Clearly, Tables 1 and 5 must be at fault in some respects. The orifice well tester and the Bureau of Mines type of critical flow prover are essentially the same instrument; both devices utilize a square-edged orifice 1/8 in. in di-ameter as the primary element, and both freely discharge gas to the atmosphere. Tables for the orifice well tester' have been published in the ranges of 0 to 15 in, of water and 0 to 40 in. of mercury (Hg) differential in pressure. Coefficients for the critical flow prover have been published for differentials in pressure greater than 75 psia. An extrapolation of either differs from the other set of data as much as 18 per cent at some points. No immediately obvious reasons for the discrepancy was found, and data available from the literature were insufficient to effect reconciliation. Consequently, a series of experiments was performed to check the available data and to determine discharge coefficients for the two devices in an overlapping range of differential pressure. The correlating equation used in preparation of tables such as Tables 1 and 5' is the so-called hydraulic equation, Q = C1 vh .....(1) The tables cover orifice sizes from 1/8 through 1 1/4 in. A second set of tables, for use with greater differentials in pressure, apply only to the 3/4, 1, and 1 1/4-in. orifice plates over a pressure range of 0 to 40 in. of Hg. The correlating equation used in preparation of the tables is Q = C2 vH(29.32 + 0.3H)/G . (2) Each of these equations is valid only for the range to which it has been applied, and neither equation is valid for extrapolation. Theoretical equations for flow through an orifice are based upon assumptions of fractionless flow and an emergent jet the size of the orifice. A multiplier, C,, called the discharge coefficient, is inserted into the theoretical expression to compensate for both frictional losses and the contraction of the jet experienced in the plane of the orifice. Buckingham3 shows that C, should vary with the ratio of upstream to downstream pressure for flow of a compressible fluid at any average linear velocity through the orifice less than the velocity of sound. Work published by the National Advisory Committee for Aeronautics4 indicates that the variation continues into the so-called critical flow region until the vena contracta coincides with the plane of the orifice. The NACA work, however, does not indicate a leveling off of the coefficient. The work of the Bureau of Mines5 for the critical flow prover was based on differential pressures greater than 75 psi and indicates that the discharge coefficient, Cd, is constant in this region at a value of about 0.86.

Jan 1, 1958

By D. W. Fuerstenau, L. M. Berlioz

In 1957, Charles1 first presented an analysis of a relationship between the energy expended in comminution and the extent of size reduction achieved. Subsequently, Schuhmann extended and interpreted Charles' approach to comminution. According to their analysis, the energy, E, expended in comminuting unit mass of fairly uniformly-sized material is related to the size modulus of the product, k, by where a is the distribution modulus (slope of the log-log cumulative size distribution plot) and C is a constant, which is primarily determined by properties of the solid. If there is any fundamental basis for the E vs k relationship, E must necessarily be concerned with the useful work for the size reduction process and not necessarily the work input, as discussed in detail by Harris.3 Charles and numerous other investigators tested this equation only by relating k to the grinding time, assuming expended energy to be directly proportional to time, but never directly measuring power. In our laboratories, we recently constructed a special batch ball mill with which the power could be accurately determined with torque measuring apparatus.4 With this device, a direct test of the Charles relationship could then be made for ball milling. 5 EXPERIMENTAL A stainless steel mill, 11.5 in. long and 10 in. in inside diam, lined with eight low-silhouette lifter bars, was used for these experiments. The grinding media consisted of 30 kg of single sized, 1-in. diam, stainless steel balls, which occupied about 50% of the total mill volume under static conditions. The void fraction of the static ball load was approximately 0.4, and for the dry grinding conditions used, the corresponding interstitial volume was filled when about 3300 g of the dolomite feed material was present. The 7x9 mesh dolomite used as feed was obtained as a natural, uncrushed product thus as- suring a random distribution of feed particle shapes. Charge weights of 660, 1320, 1980, 2480, 3300, 3960 and 5280 g, which correspond to 20%, 40%, 60%, 80%, l00%, 120%, and 160% of the interstitial load, respectively, were used. Grinding times of 20, 40, 60, 80, 100, 150, 200 and 300 revolutions were considered, with the mill operating at 54 rpm (approximately 60% of critical speed). The tests were made sequentially, in that the size distribution obtained after a certain grinding interval was utilized as the feed for the subsequent period of grinding. Agreement with control experiments where the mill ran continuously for the entire interval was excellent. Torque measurements were made utilizing a Baldwin-Lima-Hamilton SR-4 torque pickup in the manner that has been described previously.3 Standard wet (to obtain the -400 mesh material) and dry sieving procedures were followed using the Tyler full screen series. RESULTS The Gaudin-Schuhmann size distributions obtained in all of these experiments are characterized by distribution moduli (slopes), a, of about 0.66 and are representative of normal dry batch ball milling of a material like dolomite. As an illustration of typical

Jan 1, 1968

By A. G. Metcalfe

THE failure to detect any change in the thermal properties of cobalt at the Curie point would appear to indicate that the apparatus may have a limited sensitivity. The Curie temperature has been determined by direct methods to be 1121+3°CP and 1131°C.3 The specific heat has been determined many times'.' and weighted values are given in Table I. The allotropic transformations in iron and titanium, used by the authors to test the experiment method, involved a volume change of the order of 3 pct on going from the close-packed to the body-centered-cubic lattice. These energy changes are likely to be much larger than those involved in a diffusionless transformation from a hexagonal to a cubic-close-packed lattice. However, it has been pointed out that thermal observations are unreliable in the detection of phase changes when these are near a Curie point.' James 0. McCaldin and Pol Duwez (authors' reply)—The sensitivity of the present technique is limited to first-order transformations, so that gradual changes in specific heat, e.g., in the vicinity of Curie points, are not detected. Furthermore, the existence of first-order transformations with latent heats much smaller than those occurring in the known allotropic transformations of iron and titanium is, we agree, still pending.

Jan 1, 1956

By C. A. Fothergill

An improved blotter model is described which makes use of a plastic sheet as base for the blotter field and pins to represent the input wells. The model is simple to construct and adapt and is particularly useful for demonstration purposes. Operational details are given. INTRODUCTION The blotter model was first applied to the investigation of oilfield problems by Wyckoff, Botset and Mus-kat,' primarily in order to demonstrate the flood patterns which could be expected from various arrangements of input and output wells in waterflooding projects. In recent years the model has been largely superseded for analog studies by the potentiometric and electric analog models which are more refined and versatile, since they can simulate the form, and, in certain cases, permeability variations which it is not possible to do with the blotter model. However, the latter is still used for analog studies where the reservoir can be represented as being uniform in thickness and permeability, and where a visual record of the flood advance is required. Ramey andNabor,2 for example, have applied the model to the investigation of areal sweeps in oil recovery by in-situ combustion. Because of its limitations, the blotter model can be used in general for only very simple arrangements of input and output wells, but within these limitations it has a definite advantage over all other types, in that it provides a picture of the advancing flood using the simplest and most easily constructed apparatus, by contrast with the gelatin model, for example, which requires more elaborate equipment. Moreover, the experiments can be quickly repeated and modified if necessary, whereas considerable time is required to prepare a gel field, compared with that required for the preparation of sheets of conducting blotting paper. For these reasons the blotter model can be very useful for demonstration and teaching purposes, enabling the student to experiment for himself with different arrangements of input and output wells. A model has been constructed for this purpose which is simple to operate and adapt. The model (Fig. 1) consists of a base platform or box, approximately 12-in. square, on top of which are screwed three lathes of 1/2 -in. square section to form three sides of a square 6 in. in length. This forms the support for a sheet of Cobex, transparent plastic, thickness .003 in., on which is carried the blotting paper representing the field. The plastic is flexible and relatively soft, and can be attached to the support by cellulose tape so that it can be easily removed when required. The input wells are represented by brass drawing pins pressed through the plastic sheet from the underside until the points are just projecting through the top; these are connected in the circuit as shown in Fig. 2. The sheet of blotting paper of required shape and size is saturated with a solution of 0.2 n-potassium sulfate to which has been added a saturated alcohol solution of phenol-phthalein at a strength of 10 cc/liter of stock solution. The surplus water is removed and the blotting paper placed in position on the plastic sheet making contact with the drawing pins and the sheet (which is an insulator). The anode consists of a second sheet of blotting paper cut as shown and saturated with the same solution. This is suspended from a laboratory stand, being connected to the conducting wire by means of a crocodile clip and making contact with the horizontal blotting paper at the projecting point. This is the method used by Ramey and Nabor and which permits the hydroxyl ions or colored front to reach the anode. As shown in Fig. 2, the electrical voltage used is 220-v DC and a shunt is placed across the terminals to step down the voltage as required. Further adjustment can be made as necessary using a lamp board in series with the shunt to obtain the

Jan 1, 1958

By Jr. F. M. Perkins.

Capillary forces play a controlling role in water-drive displacement processes both in laboratory experiments and in actual reservoirs, but their quantitative importance may be quite different in the two cases. Because of the importance of conducting laboratory experiments which are representative of field conditions, it is necessary to understand exactly the role of capillary forces in the displacement process. Though a number of experimental investigations related to this subject are contained in the literature, there appears to be a lack of information pertaining to unsteady-state experiments in water-wet media. This experimental study was conducted to obtain additional laboratory data to clarify further the role of capillary forces in both the macroscopic and microscopic flow of oil and water in porous materials. THEORETICAL CONSlDERATlONS The capillary pressure is defined as the difference in pressure between a continuous oil phase and a continuous water phase in a porous material.' The magnitude of this pressure difference depends on the interfacial curvature and the interfacial tension. The interfacial curvature is determined by the geometry of the pore spaces, the wettability of the rock surfaces, and the quantity of cach phase present. Capillary forces are involved in a water-drive displacement process in that they exert a controlling influence on the microscopic fluid distribution which in turn is reflected in the saturation or macroscopic flow behavior. Microscopic Fluid Distribution Because of the microscopic nature of the displacement of oil by water, it is necessary to consider the flow and the fluid distribution in individual pores. On this microscopic scale the capillary Forces, which act over a distance of one or two sand grain diameters, control the distribution of oil and water under static equilibrium conditions. When an external force is applied to the fluids, such as in a water-injection experiment, the applied forces tend to distort the oil-water interfaces. However, in most fine-grained, water-wet sands, the applied pressure difference across one or two grain diameters is usually several orders of magnitude less than the capillary pressure difference. These con-siderations lead to the theory' that even during flow the capillary forces continue to control the microscopic distribution of oil and water within the pores of a porous material for all practical reservoir and laboratory flow rates. This concept of capillary forces controlling the microscopic distribution of fluids has been substantially verified by other investigators3-7 who have found a lack of dependence of relative permeability and residual oil saturation on rate of fluid injection. Macroscopic Distribution The microscopic influence of capillary forces cannot be observed easily and only the effect on the macroscopic or average saturation can be detected. The saturation, of course, is really the point of interest. During a water flood, large differences in saturation at the flood front cause large capillary pressure gradients. This, in turn, causes water to advance ahead of the flood front thereby reducing the capillary pressure gradient in this region. The result is that in homogeneous porous media capillary pressure gradients tend to cause a diffuse displacement front. At low rates in laboratory columns, the front may extend over the entire column length. When the advancing water first reaches the outflow face of a core, the water, which is the wetting phase, cannot be produced because the pressure in the water just inside the core is lower than the pressure in the oil-filled space around the outflow face. This difference in pressure is equal to the capillary pressure for the water saturation existing at the outflow face. Water, therefore, accumulates at the outflow end of the core which causes a reduction in the capillary pressure. Because the capillary pressure does not vanish except at the residual oil saturation,7,8,9 water will not be produced until the residual oil saturation exists at the outflow face. This entire effect,"' which is called the "boundary effect", results in a region of relatively high water saturation near the outflow face. At low rates of injection in a short column, this region of high water saturation may extend over a considerable portion of the column. The influence of capillary forces on the macroscopic flow of oil and water have been described by Leverett.10 For unidirectional, viscous flow in the absence of gravity segregation, the expression in dimensionless form for the fraction of water in the flowing stream, f, is

Jan 1, 1958

By Robert L. Slobod

In much of the work reported in the literature on long cores. true connate water value, probably have not been obtained because of insufficient flow of 011 to attain equilibrium. A -.satisfactory method for establishing connate water in long cores has been developed in which oil flow is maintained until no additiona1 water is displaced. INTRODUCTION Long cores have been used by sev-erals investigators in an effort to evaluate the efficiency of oil recovery for various producing schemes. The effects of many variables such as rate of displacement. pressure drop, volume of displacing medium, gas-oil ratio, inter-facial tension, and others have been investigated, and the merits of gas drive have been compared with water drive. In the older experiments no connate water was used. but in the more recent stndies the philosophy of the experiment calls for a core containing oil and connate water (irreducible water) as the starting point of the experiment. The method generally used to establish this condition in long cores involves dynamic' displacement wherein one or more phases are forced through the core. flow being maintained at a constant rate or at a constant pressure until a steady state is reached wherein there is no change of saturation with time. It is one purpose of this paper to point out that in much of the work reported in the literature true connate water values probably have not been obtained. The reason for this failure is pointed out. and a method for obtaining more reliable connate water values on long cores is presented. It is impossible at this time to completely evaluate the effect on the conclusions in the earlier papers resulting from the failure to obtain true connate water in the long cores, but some data are presented which indicate that serious errors may he present in many cases. MATERIALS AND APPARATUS Cores cut front outcrop sandstones (Strawn. Torpedo, and Elgin sandstones) with permeabilities of the order of 300 md were used. The major experiments were made with a core 18-in. long and 2-in. in diameter. Cores of the same diameter but only 12-in. long were used in a few tests? and in some work a core I-in. in diameter and 11-in. long was used. These materials were all somewhat bentonitic. 'To reduce the effect of such constituents, one of the cores (Torpedo No. 10) on which the most work was performed was acid extracted at the start of the experiment. Brine containing 50,000 ppm chloride and naphtha (Stoddard's Solvent) were used as the water and oil phases, respectively. The core was mounted in a rubber sleeve which forms the inner lining of a brass cylinder. In this arrangement2 the rubber sleeve is forced against the core by applying pressure to water in the annular space between the rubber sleeve and the brass cylinder wall. The core is therefore sealed in rubber so that all fluids introduced at the input end of the core flow through the core. With this equipment the core may be easily removed for inspection and analysis. Also, other cores may be introduced for study with no extra time for preparation such as is required when plastic mounting is used. When pressure or conductivity measurements are desired at intermediate points along tile length of the core, the brass cylinder comprising the outer jacket of the core holder is cut into section, of the desired length. Each section is then made up as an individual unit with its own rubber sleeve (see Fig. 1). The individual units are then butted up against a ring of plastic or other insulating material which carries an electrical contact for resistivity measurements and an opening which can be used for the measurements of fluid pressures. The individual units are forced tightly against the plastic ring separators and held in place with several tie rods. Inlet and outlet end butts of the same diameter as the core and several inches in length are provided to control the introduction and withdrawal of fluids and to complete the arrangement for sealing the core within the rubber sleeve. Fluid flow is maintained either by applying pressure to the liquid confined in a reservoir or by the use of a gear pump. The latter is preferable where continuous flow at a moderate rate over a long time interval is desired since a small volume of liquid can be readily recirculated using filters to prevent plugging of the core. EXPERIMENTAL The oven-dried core is weighed and is then completely saturated with brine

Jan 1, 1950

By R. Greenwood

The advent of the space age and its promise of interplanetary flight has prompted new ideas for propulsion systems that will allow maximum energy with minimum fuel weight. The use of cesium as the source of energy for such rocket flights has been found desirable. Theoretical efficiency of a rocket fuel is measured in terms of specific impulse, defined as the number of pounds of thrust per pound of propellant consumed per second. Specific impulse is usually expressed in seconds (the number of seconds for which one pound of fuel would produce one pound of thrust) although it is really a measure of total energy, not of power. The most sophisticated chemical fuels have a specific impulse of 300 to 400 sec; liquid hydrogen heated by nuclear reactor has a specific impulse of 1000 to 1200 sec. Specific impulses as high as 20,000 sec, however, are theoretically attainable with the ion rocket, in which free cesium ions are accelerated across a voltage drop and ejected at high velocity.' Characteristics of the system have been sketched by Boden.2 The desirable properties of an ion propellant are: 1) low ionization potential, 2) low boiling point, and 3) high atomic weight. Alkali metals best fulfill these requirements, and Table I indicates the superiority of cesium. The use of heavy complex ions such as UCl3 is ruled out by their instability and other disadvantages.' Despite its enormous superiority in specific impulse, the ion rocket is inherently low-powered, capable of very small acceleration. Application of ion propulsion will be limited to large rockets and spaceships on missions requiring powered flight for weeks or months and employing a separate chemical propulsion system for escape from the earth's atmosphere where power requirements are large but of brief duration. For example, Kraemer and Larson3 considered the case of an unmanned 25,000-lb pay-load placed in a 300-mile orbit by chemical rockets, proceeding thence by ion propulsion to the vicinity of Mars, where it would go into orbit once again. Calculations provide for 8500 lb of cesium for such a voyage. The question of whether such quantities of cesium will be readily available to the space program does not appear to have been considered by geologists and mining engineers. Present Production and Uses Literature on industrial applications of cesium is scanty. The metal is used in scintillation counters, photocells, infra-red detection devices, and as a "getter" in low-voltage vacuum tubes. Production is sporadic and total figures are not published, but the annual consumption is measured in pounds rather than in tons. The few pollucite deposits known, small as they are, can easily supply the present limited demand for gram or pound lots of cesium. Until recently the price of cesium was $2 to $4 per gram, but according to Chemical Engineering, cesium salts are now available at $13 to $27.50 per lb. Geochemical Distribution As a large alkali ion, the cesium ion accompanies potassium in geological processes and tends to be concentrated in the latest crystallizing fractions of potassium minerals in much the same manner as rubidium. Most of the cesium in igneous rocks, therefore, is in the micas and potash feldspars of granites, and especially in granitic pegmatites. The cesium content rises markedly in the micas of the late by-drothermal replacement stage in complex pegmatite~.' Typically, cesium is only 0.5 to 0.01 times as abundant as rubidium in potassium minerals. Rb', with a radius of 1.47 A, is close enough in size to K' (1.33 A) so that rubidium forms no minerals of its own. Cs+ is so large (1.67 A) that only small concentrations can be taken up in potassium minerals; when cesium is present in solution in larger concentration it forms the mineral pollucite, Cs3-Na2A17Si17O48.3H2O, which is known only in pegmatites. In the weathering of igneous rocks, cesium is dissolved, carried to the ocean, and apparently in-

Jan 1, 1961

By J. A. Jaekel, W. C. Aitkenhead

Uranium deposits in the Spokane Indian Reservation, as well as those around Mt. Spokane, are essentially low grade, much of the ore containing less than 0.2 pct U3O8. The Mining Experiment Station of the Division of Industrial Research, State College of Washington, has been engaged in intensive research on the amenability of these low grade ores to froth flotation. The results: successful flotation of autinite, chief mineral constituent. At the outset of this work the goal was a concentrate of 1 pct U3O8 with a 90 pct recovery from ores containing less than 0.2 pct U3O8. Most of the work has been done on argillite ore from the Midnight mine on the Spokane Indian Reservation. The goal has not been attained using this ore, but samples of the granite ore from Mt. Spokane yielded successful results. For example, a concentrate containing 11.2 pcl U3O8 was produced from a Mt. Spokane high grade ore containing 1.27 pct U3O8 with a recovery of 97.8 pct. Another Mt. Spokane ore yielded a concentrate of 5.0 pct U3O8 from an ore containing 0.13 pct U3O8. with a recovery of 85 pct. This same ore gave a recovery of 93.5 pct when the grade of concentrate was reduced to 2.0 pct. It has been concluded that a successful method for floating autunite has been developed and that the mediocre results from the Midnight argillite ore are probably caused by the presence of some other uranium mineral or minerals less amenable to these reagents. The experimenters tested a third type of Washington ore, found on the Northwest Uranium Mines Inc. property on the Spokane Indian Reservation. This is a conglomerate of pebbles and small boulders of partially decomposed granite and is shot through with autunite. Its characteristics lie between those of the Midnight ore and the granite ore from the Spokane district. It responds better than the ore from Midnight but not as well as that from Mt. Spokane. As the fatty acids are the only type of collectors showing promise, investigation has been concerned with these acids and the optimum conditions for their use. The first method for treating the argillite ore from the Spokane Indian Reservation made use of Cyanamid's R-708 as a collector, a tall oil product described as a substitute for oleic acid. Although the investigators proved that R-708 is a collector for autunite when mixtures of autunite and silica sand are used, results on the ore were mediocre. Tests of other fatty acids revealed that the solid fatty acids of the saturated series are collectors for autunite and that their collecting power increases with the length of the carbon chain. The even carbon members of the whole series were tested from the 10 carbon acid (capric) to the 22 carbon acid (be-henic). The least expensive collector, stearic acid (18 carbon), proved to be a good one, so this was used in most of the tests. In first attempts with stearic acid, the collector was dissolved in various hydrocarbons and the solutions were added to the flotation cell. Cyclohexane, gasoline, fuel oil, kerosene, and other solvents were tried. Small amounts of high grade concentrates could be brought up, but recoveries were low. Finally emulsions of stearic acid were tried. It was discovered that stearic acid alone has little collecting power except when conditioning is carried out at high temperature. When hydrocarbon solvents were also present, it proved to be an excellent collector. An example of one emulsion that proved satisfactory for some ores is given as follows: 1 part stearic acid by weight, 1 part sodium oleate by weight, 1.2 parts kerosene by weight, 100 parts water. In some successful tests part of the stearic acid was replaced by oleic acid. The emulsions were made by agitating the stearic acid and sodium oleate together with hot water, then adding the kerosene and agitating while cooling. In the five tests reported in Table 1, 650 g of ore were ground with 650 cc water in a laboratory rod mill. The pulp was filtered to eliminate excess water and the ground ore transferred to a stainless steel beaker for conditioning at high pulp density. In most of the tests sodium hydroxide was added to the conditioner during agitation, then the collector emulsion, and finally the sodium silicate. The amount of alkali was adjusted to give a pH of 8.5 to 9.0 in the flotation cell. After conditioning the pulp was transferred to a laboratory flotation cell and the test completed in a normal manner. It is interesting to note that a deposit of high grade concentrate forms on the conditioning agitator and in the conditioning vessel, and at times on the agitator of the flotation cell itself. A few grams of concentrate running as high as 4 pct U3O8 were recovered from the conditioner when Midnight ore containing less than 0.2 pct U3O8 was treated. In the examples given in Table I this conditioner concentrate is calculated as part of the total concentrate. The authors have not yet fully explored the possi-

Jan 1, 1960

By Robert J. White

Strong's equation for calculating bottom-hole pressure reduction due to gas cutting of drilling mud is corrected, resulting in a simpler equation which is easier to use. Use of the equation is illustrated by examples. It is shown that reduction of weight of a 10 lb/gal mud by 50 per cent at the surface due to gas cutting produces only n 3 per cent decrease in hydrostatic head at the bottom of a 5,000-ft well. It is also shown that at the surface 5 volumes of air per volume of mud are required to reduce the apparent mud weight from 8.5 to 6.5 lh/gal in a 3,000-ft well. Illustrative graphs are included to show the effect of various degrees of gas cutting on apparent mud weight at depths of 2,000, .5,000 and 10,000 ft and to show variation of apparent mud weight, with depth for a 100 lb/cu ft (13.37 lb/gal) mud gas cut to 90 lb/cu ft (12.03 lb/gal) at the surface. INTRODUCTION In 1938, Strong' published the following equation for calculating reduction in bottom-hole pressure due to gas cutting of drilling mud h = 1/D [p + pn/100 In p + p(1-n/100)/p(1- n-100)] where: h = depth in feet P = pressure in atmospheres at depth h due to mud column only D = hydrostatic pressure in atmospheres of a col-umn of uncut mud 1 ft high p = back pressure at well-head (in atmospheres) n/100 = fraction by volume of gas in mud at well-head at back pres-sure p This equation was recently found to contain an error due to an incor-rect expression for variation of vol-ume fraction of gas with pressure. Values calculated by the original equation are close to the correct values for small percentages of gas for which it was derived but become increasingly erroneous as the gas content increases and are completely unusable at the very high gas con-tents employed in aerated mud. DERIVATION The nomenclature used is the same as in Strong's paper. The pres-sure dP exerted by a lamina of thickness dh at a depth h ft is ex-pressed by Eq. 1 where x = volume fraction of gas at depth h dP = D (l - x)dh . . . (1) When we express x in terms of the wellhead per cent gas n, the well-head pressure p and the pressure P at depth h, the result is Eq. 2. x = (np/p+p)/(100-n) + (np/P + p) When this is substituted in Eq. 1 and rearranged to separate the var-iables the result is dh = 1/D [dp + np/100-n dp/(P + p]. Integration between appropriate lim-its results in or if the wellhead is open so that p = 1 atm and we rearrange Eq. 4, the result is loss in head = hD — P = n/100-n In (P + 1) ...(5) Eq. 5 can be solved easily by successive approximation using hD as a first approximation for P on the right-hand side and calculating a first value for loss in head. This can then be used to get a new value of P using the left-hand side of the equation. Values converge very rap-idly even for quite large values of per cent by volume gas n. EXAMPLES The effect on bottom-hole pressure of 50 per cent gas cutting of the mud in a 5,000-ft well using 10 Ib/ gal mud will be calculated, we have: h = 5,000 ft 10 X 7.48 144 X 14.7 = 0.0353 atm/ft hD = 176.5 atm n = 50 Substituting these values in Eq. 5 and using hD as a first approxima-tion of P on the right, we have, approximate loss in head = 50/50 In (176.5 + 1) = 2.30 log 177.5 = 5.17 atm. We calculate a new value of P as follows:

Jan 1, 1958

By R. E. Collins, Paul B. Crawford

Mathematical equations have been derived to show the effect of the slippage phenomomenon (Klinkenberg effect) on unsteady-state gas flow through porous media. The assumption is made that the slippage may be represented by an equation of the type It is pointed out that most of the existing data on unsteady gas flow may be corrected for slippage by a simple change in definition of the dimen-sionless groups. INTRODUCTION Several solutions to the various problems associated with the unsteady-state flow of gas through porous media have been recently presented in the literature.','" The cited publications have resented solutions for flow under ideal conditjons. The principal assumptions involved include: (a) the density of the gas obeys the perfect gas law; (b) the viscosity of the gas is constant; and (c) the gas permeability is constant. Few, if any, actual gases under reservoir conditions are believed to conform with the above, but the ideal solutions do provide the basis for making improved estimates on the performance of gas reservoirs. The purpose of this paper is to show the performance characteristics when assumption (c) is corrected for the slippage phenomenon or the Klinkenberg effect. This is believed to be a step nearer to reality. MATHEMATICAL DEVELOPMENT Case 1. Gas Flow — Ideal Conditions The equations describing the flow of gas under ideal conditions have been clearly presented in the cited literature; however, for completeness of this work they are repeated here. The basic differential equation describing the flow of gas is Equation (6). It may be derived as follows: The continuity equation, the perfect gas law and Darcy's law are combined to give: Manuscript received in the Petroleum Branch office Aug. 3, 1963 or If — is constant, Equation (5) is rewritten to form the basic Equation (6). or for the linear case A dimensionless equation is obtained by dividing through by (pl)z and making the substitutions shown below: The resulting equation is A similar procedure for the radial case leads to the following equations: where and Equations (11), (12), and (13) are those used by Bruce, et al., and Jenkins and Aronofsky for their computational work.

Jan 1, 1953

By W. F. Baldwin

The timing circuit generally employed in Carter Reservoir Behavior Analyzers consists of a synchronous motor driven cam which actuates a microswitch. A mechanically actuated timer of this type has several inherent disadvantages, among which are the following: 1. The first time interval may not be of uniform length, depending on the exact position of the cam at the conclusion of a previous run. 2. The positioning adjustments on the microswitch lever arms are critical and require periodic readjustment. 3. The mechanical system is subject to wear which will ultimately impair its accuracy. It was therefore considered desirable to devise an alternate method which would eliminate these difficulties without introducing others that would be equally troublesome. This note describes the design of an electronic timing device which has been incorporated in the analyzer constructed by the Field Research Laboratories of Magnolia Petroleum Co. The electronic timer circuit is built around a simple neon glow tube relaxation oscillator, the complete circuit diagram of which is shown in Fig. l. The operation of the circuit is as follows: Condenser C, is charged to the critical firing voltage of the neon glow tube VR-75 by battery B1 through resistors R1 and R2. When the neon glow tube fires, the condenser C, is discharged, The charging cycle then repeats itself,

Jan 1, 1950

By W. G. Pfann

WHEN a molten zone traverses a long, solid ingot in a level, open boat the ingot becomes tapered. While the taper may be slight after one zone pass, it can be appreciable after repeated zone passes, even to the extent of causing an overflow of matter at the end of the boat. It will be shown here that such matter transport by a molten zone arises from the volume change on melting and that it corresponds in magnitude and direction to the magnitude and sign of the volume change. It will be shown also that the mechanism of matter transport is basically similar to that of solute transport by a molten zone, which has been discussed elsewhere,1,2 and that the two processes can be described by similar equations. Finally it will be shown that matter transport can be prevented by inclining the ingot at a critical angle. Contraction on melting results in forward transport, that is, transport in the direction of zone travel; expansion on melting results in reverse transport. The way in which forward transport arises from the contraction on melting may be seen as follows: At the starting end of a solid horizontal ingot of uniform rectangular cross section of height h., form a molten zone of length 1, as shown in Fig. 1. The height h of the liquid will be h. where a is the ratio of density of solid to that of liquid. Advance the molten zone, at constant length, a distance dx. The volume of solid melted will be hdx; the volume of solid frozen will be Izdx, because the height of the freezing solid will tend to be equal to that of the liquid. The resultant change, dv, in liquid volume will be the difference in these quantities multiplied by the density ratio. Hence the following equation may be used for a zone at any distance x from the origin: which is valid in all but the last zone length in the

Jan 1, 1954