Search Documents

Jan 1, 1947

By R. L. Boyer, F. Morgan, M. Muskat

In the last several years a type of application of the fundamental permeability-saturation relationships has been developed which offers great promise in the interpretation of general reservoir performance.l-' In these developments no attempt is made to solve the detailed flow dynamics in actual oil and gas-producing systems. That problem still remains unsolved. But the permeability-saturation phenomena are incorporated in the analysis of oil-producing reservoirs in a manner such as to reflect the true role they play in controlling the gross reservoir behavior. In particular, they are introduced in such a manner as to determine the gas-oil ratio of the production at any stage of depletion, and, when combined explicitly with the so-called material-balance principles in differential form, one can obtain a differential equation describing completely the average pressure and gas-oil-ratio history of oil-producing reservoirs. While thus far they have been applied directly only to gas-drive reservoirs, they are susceptible of extension to partial water-drive systems, and qualitatively the same principles are of value in the study Of complete water-drive reservoirs. Moreover, in addition to giving a means for computing the variation of the reservoir pressure and gas-oil ratios vs. the cumulative production, they automatically lead to predictions of the gross ultimate recovery under natural depletion—and when properly extended they provide means for predicting the ultimate recovery under gas or water injection. The data required for such applications are the initial fluid distribution in the producing section, the pressure-volume-temperature data on the gas and liquid phases, and the permeability-saturation relationship for the oil-satuiated zone. The former can be obtained with reasonable accuracy by methods that are well established and available throughout the industry. The permeability-saturation data, however, involve techniques that are not at all standardized. In fact, only

Jan 1, 1947

By John J. McKetta, Donald L. Katz

Recent experimental data on the methane-n-butane-water system have been combined with data in literature to present: (I) the water-vapor content of natural gases, (2) the solubility of water in hydrocarbon liquids, and (3) the solubility of natural gases in water. Since water is present in petroleum reservoirs, experimental data regarding the behavior of water in hydrocarbon-water systems is of increasing engineering importance. A survey of literature (Table I) shows that data of these types are fragmen-tary at temperatures above the hydrate region. Experimental data taken from the authors' recent investigation of the meth-ape-n-butane-water system are presented.17

Jan 1, 1947

By E. A. Nichols

The purpose of this paper is to present a contour map that shows the geographical variation in geothermal gradients in the Mid-Continent and Gulf Coast oil fields. Introduction Frequently there arises the need for an estimation of the temperature to be expected at a predetermined depth in working with a wildcat or in deepening old production into 'Ones that are relatively unknown. The study of geothermal gradients in oil fields has been made previously by many investigators. One of the pioneers in this work was C. E. Van Orstrand,' who, as well as most of the other investigators, was primarily interested in the relation between temperature and depth in a given locality. Between 192r and 1929, American Petroleum Institute Research Project NO. 25 fostered a thorough study of earth temperatures.2 K. IV. French did some work in California, the results of which were published in 1939.3 In most of these studies, an attempt was made to correlate temperatures in a rather localized area. This paper will show the results of an attempt to correlate temperature gradients over rather large Portions of the Mid-Continent and Gulf Coast area. Method of Obtaining Data It has been found that for complete thermal equilibrium to be established in a well, the well must stand undisturbed for a period of several hours to several weeks, depending upon the way the well was produced prior to the shut-in period. The data in this paper were obtained in conjunction with running pressures in wells or in taking samples of reservoir fluid. Glass maximum recording thermometers were used in all but a few instances. The wells had been shut in from one hour up to I. or 12 days. Since the wells were normal producing wells, some of the shut-in times undoubtedly were insufficient for complete thermal equilibrium to have been established and the temperatures measured do not necessarily represent the true earth temperatures. They do represent the temperature conditions that will be encountered in a well that is producing oil. The data, therefore, can be useful in working with or in estimating the temperatures to be expected in producing reservoirs. In obtaining a thermal gradient by dividing the difference in temperature of the formation and the mean annual surface temperature by the depth of the formation, one is assuming the depth-temperature line to be straight. Depth-temperature curves are almost straight lines (Fig. I). However, on close examination of accurate data, a slight curvature usually is observed. Most wells yield curves with slopes increasing with increasing depth. These curves are believed to be from wells that are in and are underlain by sedimentary material. Some few curves have slopes that decrease with increasing depth. These curves are believed to be from wells that are underlain rather closely by granite or some other igneous rock. Most investigators

Jan 1, 1947

By F. Morgan, M. Muskat, R. G. Russell

Core experiments with radioactive tracers are reported in which an artificial radioactive material, radiovanadium, was used in aqueous solution in a study of the mobility of connate water. Sandstone cores saturated with active water were first oil-flooded to simulate virgin sand conditions in which the pores are filled with oil and connate water. They were then subjected to flooding with inactive water. Measurements were made of the activity of various portions of the effluent as well as the activity of the core after the flooding process was completed. In additional experiments the distribution of activity along the core was measured during the flooding process. In still other experiments direct flooding tests were made on cores fully saturated with active water. In all cases it was found that after injecting into the core a volume of inactive water equal to a pore volume, the effluent contained more than 80 per cent of the active content originally in the core. It is felt that these results imply a high degree of mobility of the connate water when subjected to flooding, although at least a partial contribution of the effect of tracer transfer within the core cannot be definitely excluded. The implications of this result with respect to water-flooding, core analysis and electric logging are discussed. A better approach to determining the connate water saturation of oil-producing rocks, by measuring the fluid content of core samples at the surface, is by the use of crude oil or oil-base muds as the coring fluid. Although water flooding is of considerable importance in the recovery of ail, comparatively little is known with certainty about the role played by the connate water during the operations. Two possibilities exist. Either the connate water is retained in the formation while the flood water sweeps past, or it is pushed along to a greater or lesser degree by the flood. Thus far no explicit recognition in terms of quantitative studies appears to have been given to the displacement of connate waters by the intruding extraneous waters, although such a possibility might be important in determining the efficiency of the oil -displacement process in water-flooding operations, especially in argillaceous rocks. On the other hand, it is a common observation that the first water produced with the oil as a result of water-flooding is definitely saline.*.' Similar questions regarding the mobility of the interstitial water arise in the quantitative description of the advance of edge-water drives. In the interpretation of the water saturations determined from core analyses, when coring with water-base muds, it is generally assumed that the drilling fluid merely adds to the connate water without displacing the

Jan 1, 1947

By Henry Schaefer

Solution by electric analogy of performance problems of reservoirs containing oil and gas has heretofore depended upon a process of successive approximations based on material-balance calculations, because of variation in fluid compressibility. By the addition or subtraction of electric condenser capacity, the analogy network can be adjusted automatically for changing compressibility due to evolution of gas in the reservoir. The addition of a condenser is accomplished by an electronic circuit, which is described; subtracting a condenser is done by merely opening the electrical circuit by means of a relay. Introduction V. Paschkis and H. D. Baker1 first developed in this country a method of applying the analogy of flow of electric current in noninductive circuits to solution of heat-flow problems and suggested its application for certain other problems, such as flow of fluids or mass transfer and chemical diffusion. Necessary details for applying the analogy to flow of fluids in porous media were developed by IV. A. Bruce;= he classified3 the types of analysis and predictions that can be made by the analogy in class I, when the reservoir contains oil but no gas for the period under consideration; in Class 2, when both oil and gas exist in the free state in the reservoir sometime during the period under consideration. Duplication of bottom-hole pressure history and predictions for class I reservoirs can be obtained directly by electric analogy. Until now, the solution of problems pertaining to reservoirs in class 2 was obtained by successive approximations along with material-balance calculations. The object of this paper is to present a method for simplifying the application of electric analogy to class 2 reservoirs, resulting in a solution of these problems as direct as is possible for class I reservoirs. This is accomplished by an electronic control circuit, which adjusts the analogy network for changing compressibilities of fluids contained in the reservoirs. Analogy Network In electrical phenomena the analogy to two-dimensional fluid flow in a porous media is that of current flow through an ideal conducting sheet. This sheet, perfectly insulated on both sides, has resistance and capacitance uniformly distributed throughout its area but is of negligible inductance. The need for having the surface boundaries perfectly insulated is that good correspondence may be had with a permeable fluid-conducting stratum bounded above and below by impermeable media. 111 actual application the ideal conducting sheet, which has distributed parameters, is replaced by a network having lumped values of resistance and capacitance—the resistances are series elements and the capacitances are shunt elements. The larger the number of lumped

Jan 1, 1947

By O. F. Thornton, D. L. Marshall

Results of the determination of the inter stitial water content of several core samples from oil reservoirs are presented. Data obtained by the capillary pressure method, which has recently been developed, are compared with water saturations measured in cores cut with oil-base mud and with values calculated from the electrical resistivity of the formation. The agreement between the capillary pressure and other methods, combined with the fact that similar agreement has been obtained by others, indicates that the capillary method properly applied yields results sufficiently accurate for most engineering purposes. In some cases, it may be possible to extend the utility of data obtained with the capillary pressure method by correlation of interstitial water saturation with more easily measured physical properties of the media. Introduction Several methods are in use for determining or for estimating the interstitial water saturation at given points within oil and gas reservoirs. The direct measurement of the water content of cores obtained with nonaqueous fluid in the hole is believed to be an accurate method under most conditions. The water saturation may be calculated by use of the resistivity curves of the electric log,' although this method is not applicable under conditions that exist in some reservoirs. Recently, a method has been developed and described3.' wherein the interstitial water saturation of cores is measured at a capillary pressure such that conditions existing in the reservoir are simulated. The purpose of this paper is to present data obtained by use of this latter method, and to compare the results obtained with determinations made by other means. Method The theory underlying the capillary pressure method applied to petroleum reservoirs has been discussed adequately in the literature by Garrison,2 Leverett,5 and Others.3,6,7 Briefly, and simply, capillary pressure is defined as the difference in pressure between two contacting fluid phases, such as oil and water. For the simplified sand, water and oil system illustrated in Fig I, the pressure difference can be stated as follows:6 where Pc is the capillary pressure in pounds per square inch, ? is the interfacial tension between oil and water in pounds per inch, r1 and r2 are the radii of curvature of the interface between oil and water in inches. Capillary pressure varies in a given reservoir with depth because of the greater density of the water phase as compared with oil or gas. The following equation expresses the difference in pressure between two points: p1- p2 - h?p/144 [2] where p1 — p2 is the capillary pressure difference in pounds per square inch, h is

Jan 1, 1947

By M. Muskat

A theory has been developed describing the behavior of wells and reservoirs producing by the action of bottom water drives. It is assumed that the pressures are maintained above the bubble point, and that the permeability to viscosity ratio of the water in the flooded zone is the same as for the oil in the oil-saturated pay. Accordingly, the treatment is based on the homogeneous fluid potential theory. The effect of the difference in density between the oil and water is neglected. .4fter deriving the proper potential distrihutions, the nature of the rise of the water-oil interface below the producing wells is calculated. The primary clean-oil-production phase is expressed in terms of the displacement efficiency, defined as the fraction of the pay flooded out by the rising water table at tlle time of first water entry. The variation of this displacement efficiency is calculated as a function of the well spacing, pay thickness, the ratio of the horizontal to vertical permeability, and the well penetration. The important parameter, other than the well penetration, is found to be the ratio of the well separation to the thickness of the oil pay multiplied by the square root of the ratio of vertical to horizontal permeability. It was found that the displacement efficiency continually decreases as this parameter increases, and for values of the latter >3.5, the efficiency varies inversely as the square of the parameter. It also increases with decreasing well penetration. Perhaps the most striking result of this phase of the analysis is the conclusion that in order to explain the delay in entry of water in wells producing by bottom water drives for periods longer than a few days normal production, with drawdowns large compared with the differential density head between the oil and water, it is necessary to assume that the vertical permeability is a very low fraction of the horizontal permeability. An analysis is also given of the production history after the water has broken through, on the assumption that the production mechanism and details of the flow distribution continue to be the same as during the phase of clean-oil production. As is to be expected, it is found that the water cut will increase at an accelerated rate as production proceeds. However, the initial differences between systems of different well penetrations or well spacings with respect to the total clean-oil production gradually decrease in terms of the variation of the water-oil ratio with cumulative production. In particular, by the time the water-oil ratio has reached values of the order of 5 or greater the cumulative recoveries will generally differ much less among various producing systems than their corresponding clean-oil recoveries. Introduction On the basis of elementary geological principles, one may classify effective water-drive fields in two groups. In the first, the entry of water into the oil-producing formation may be characterized as an edge-water encroachment. This term is used to suggest that the motion of the water proceeds largely in a direction parallel to the planes of stratification. This type of water intrusion occurs usually in thin producing zones and in strata lying along structural flanks of appreciable dip. In the second group—"bottom water-drive" fields—the water-oil interface lies in a plane of zero or slight inclination to the

Jan 1, 1947

By E. C. Patton

Jan 1, 1947

By Helmut Thielsch

In recent years an ever increasing amount of interest has been directed toward research studying the principles and various applications of sonic and supersonic waves. Though still in their early stages of practical development, sonic vibrations are already used in the emulsifi-cation and degasification of liquids. Metals exposed to sonic or ultrasonic vibrations while solidifying show a considerable amount of grain refinement. Often sound waves are applied instead of radiography to test engineering materials for internal cracks and flaws.' In the field of pure research these waves are used in dispersion and also in coagulation of particles or liquids in gaseous or, under certain conditions, in other liquid systems. In the settling and thickening of slurries considerable time is necessary to compress the solid particles and to squeeze out the liquid to obtain as low a dilution (liquid to solid ratio) as possible. Various methods have been developed to increase the amount of coagulation of the very fine solids (colloidal or semicolloidal) and to shorten the over-all time. The addition of small amounts of electrolytes and the warming of the suspension represent merely the more important techniques. Some of the effects of supersonic vibrations have been reported previously. From preliminary studies, therefore, it was hoped that by application of sonic vibrations it would be possible to reduce the time required to thicken such materials to a specific dilution, as well as to thicken to lower ultimate dilutions than could be obtained by quiescent settling alone. By increasing the volume thickened per unit time—i.e., improving the plant efficiency— great savings in equipment and time would be obtained for industrial processes. Also, faster settling rates should result if coagulation of particles could be increased. Experiments have been conducted for a period of several months in the Allis-Chalmers laboratory to study the effect of sonic vibrations on settling of particles of different sizes and concentrations. Apparatus The sedimentation apparatus used in these experiments is illustrated in Figs. I and 2. The essential features are the stationary field coil and the movable voice coil. Both windings are placed around a solid iron core. A leather diaphragm connected to the steel frame holds the voice coil in place. A brass rod is fastened to the voice coil with a metal disk connected to the other end. This disk is sealed to the lucite cylinder with a flexible rubber diaphragm. The lucite tube is 17 in. long; its inner diameter is 2½ in. and the wall thickness is ¼ in. The tube is calibrated on the outside in 10 c.c. divisions for a total of 1000 cubic centimeters.

Jan 1, 1947

By Grover J. Holt

In order to provide a clear picture of the development of the sink-and-float process of concentration as applied to the iron ores of Minnesota, a few pertinent facts should be brought out concerning the ores and the products of concentration. Iron ore is primarily a cheap commodity handled in immense quantities. The price of iron ore at present, based on a natural iron content of 51.5 per cent, varies from $2.55 to $2.70 per gross ton f.0.b. cars at the mine or washing plant in Minnesota. Last year 92,000,000 gross tons were shipped from the Lake Superior district (Minnesota, Michigan, and Wisconsin) to the furnaces. Concentrated iron ores have no greater value per unit of iron content than ores that can be mined and shipped direct from the shafts or open pits. For years only the higher grades of concentrating ores could be treated and sold in competition with the immense reserves of direct shipping ore available. These higher-grade "straight wash" ores required only the simplest treatment; namely, the washing out of fine sand from coarse ore particles. As the exhaustion of the " straight wash " ore is reached, rock is found replacing part of this sand, and the ore no longer responds to the former sizing treatment. Substantial tonnages of this marginal (jig ore) remained to be treated. Because of the limitations placed on concentrates by the furnaces—namely, low price of product and minimum amount of fines—iron-ore concentration oil has been and must continue to be a one-pass process. Only an increase in price or a change in structure of concentrates allowed by the ore consumer will encourage departure from this practice. Faced with the necessity of handling large tonnages, with the limitations mentioned, iron-ore research men have found their search for new and more efficient methods of treatment seriously handicapped and removed from the field of fine crushing. Sink-and-float treatment appeared to offer the best field in coarse concentration and has been proved by several plants to be well adapted to our problem. The use of ferrosilicon as a medium solid was decided upon. In our present practice, the ferrosilicon metal, carrying about 83 per cent iron, 15 per cent silicon and 2 per cent carbon and nonmagnetics, is crushed to pass 65 mesh and used without further sizing. When mixed with water the specific gravity of the medium may be controlled to within 0.02 per cent above or below the required gravity. Table I shows examples of a change in solid content of the medium for varying gravities. the I is based on clean ferrosilicon, whereas some contamination is always present in the mill circuit, tending to

Jan 1, 1947

By R. R. Knuckey

Having been associated with the operation of the de Vooys process for coal, which has treated 13,000,000 tons per annum, and recognizing the process as of value in ore sorting, Huntington, Heberlein and Co. Ltd. (Great Britain) in 1936 made an examination of all other existing methods. In none of the pioneer applications had recognition been given to the necessity for maintaining certain conditions that are essential for sorting certain ores if a truly efficient separation by gravity alone is to be assured. Consequently, in 1936-1937 Huntington, Heberlein and Co. Ltd., with Pearson, employing certain ideas connected with the de Vooys application, commenced the development of a separator that would allow conformance with optimum conditions in the zone where separation is effected. Later, Holt and Wade added to the value of the original application by Wuensch, Rakowsky and others, whose conceptions are being successfully operated today. Meantime, Trostler had enhanced the value of the original H-H. process by improvements that are now embodied in operating plants, and today the process incorporates an improved apparatus and a stable medium that produce the optimum conditions for separation. These conditions are: 1. Elimination of disturbing currents in the separating zone; 2. Reduction to a minimum of medium-density differences throughout the separating zone; 3. Stable medium of minimum viscosity to permit the unretarded fall of mineralized particles; 4. Removal of products of separation with minimum disturbance to the separating zone. Design of Separator The H-H. sink-and-float separator consists of a pyramidal tank in which the separation actually takes place, connected at the bottom with an inclined bucket elevator. The tank and the elevator are filled with medium to within a few inches of the top of the tank. Medium is fed into the separator at a variety of points, according to the type of ore being treated, and it is withdrawn from the separator at two points: (1) together with the float product, by means of a variable-speed addle, at the surface of the medium, and (2) by flowing over an adjustable weir cut in the side of the elevator casing. The prepared ore is fed onto the surface of the medium. The intermediate paddles, which revolve at a slow speed of 2 to 5 r.p.m., move the floating ore across the bath from feed end to float-discharge end. The ore, which is heavier than the medium, sinks below the sweep of the addles and reaches the boot of the bucket elevator, which discharges it without disturbance of the separating zone. These arrangements for withdrawing medium from the separator allow the use of minimum flows, which connote minimum

Jan 1, 1947

By C. Y Garber

Interest in the sink-float plant operated by the Bunker Hill and Sullivan Mining and Concentrating Co. at Kellogg, Idaho, has recently been largely confined to the preparation and reconditioning of the galena medium. The ease and simplicity of these operations in a concentrating mill producing a flotation lead concentrate has occasioned enough comment to possibly justify a brief description. The subject is treated from an operator's viewpoint only; no effort is made to discuss it in a theoretical or technical manner. Plant at Bunker Hill The sink-float plant was first operated in December 1941. It was built for the purpose of quickly increasing mill capacity with a minimum of equipment. It is the Huntington-Heberlein sink and float process using galena medium and is operated under license from the Sink and Float Corporation, New York city. Lead-zinc ore from the Bunker Hill mine is crushed to pass a 1¼ in, screen and is then washed and separated on vibrating screens having 4½-mm. round-hole punched plates. The screen under-size, amounting to 33 per cent of the mill feed, is dewatered and sent to the ball-mill feed bins. The screen oversize goes to the sink-float plant, where 50 per cent of the oversize, or 33 per cent of the mill feed, is float, which is washed free of medium and disposed of as a tailing. The remaining 33 per cent is sink, which, after washing, is crushed to pass 36 in. and joins the under-size in the ball-mill feed bins. The normal plant capacity is 1700 tons per 24 hr. About 1.7 per cent of the total lead and 3.3 per cent of the total zinc in the feed is lost with the sink-float tailings (Table I). More complete descriptions of the H. H. process and Bunker Hill sink-float plant are given in several recent articles, listed at the end of the paper. Preparation OF Medium at Bunker Hill New medium is made intermittently as needed, usually 12 hr. in each 6 days. A small portion of the current mill flotation concentrate is diverted from the lead cleaner cells to a I-in. vertical sand pump and pumped to a two-cell 36-in. Agitair flotation machine, where NaCn and ZnSO4 are added and the grade of concentrate is improved as shown in Table 2. Tailing from this operation is returned by gravity

Jan 1, 1947

By Harry L. McNeill

The Selective Media Process was developed by the author in a laboratory provided by The Steams-Roger Manufacturing Co. It is the outcome of an investigation into the possibilities of coarse beneficiation by sink and float without using the conventional heavy media. Commercial work to date has been limited to the treatment of minus 3/8-in. iron ore at the Hill Trumbull washing plant of the Cleveland Clifis Iron Co., at Calumet, Minnesota. Two cut away prints of the machine are shown (Figs. I and 2). A rotary conical impeller is mounted within a stationary cone (Fig. 2). This impeller is adjustable vertically by a 12-ton screw jack. The speed of the impeller can be varied from 14 to 32 r.p.m. by the Reeves variable-speed motor pulley mounted on the 7½-hp. motor drive. Attached to the bottom of the stationary cone is the automatic control valve for sink product, actuated by a ¼-hp. gear motor. The sink product is dewatered in a conventional drag driven by a 7½-hp. motor and discharged onto an inclined conveyor containing from 6 to 10 per cent moisture. Two plants are operated at the Hill Trumbull property. All the ore is first treated in a typical washing plant (Fig. 3). when treating "wash ore" a finished concentrate can be produced by the log washers and Akins classifiers, as the separation involves only fine silica (minus 48-mesh) and free iron. However, when "jig ore" is led to the washing plant the coarse

Jan 1, 1947

By A. W. Faheneald, Lewis S. Prater

By taking advantage of the fundamental difference between screening and classification—namely, that specific gravity has no effect on screening but is one of the important factors in classification—a combination of the two may be used as a process of mineral concentration. Although this fact is not new, so far as is known it has had little or no commercial application. The scheme is not presented now as a process for producing finished products but rather as a preliminary operation rejecting part of the barren material from an ore before it is subjected to more costly treatment. That is largely the function of the new heavy-media concentration processes that have been widely accepted. Original installation costs for the present sink-float processes are relatively high, however, which confines their use to operations where large tonnages are handled. Sink-float has been successful on only the relatively coarse (approximately plus 10-mesh) portion of an ore, and it is in this same range that classification screening can function most satisfactorily. Theoretically, there is no upper limit to the size of particles that can be handled in a sink-float operation, but the degree of liberation in the very large pieces usually limits this. With classification screening, particles from % to % in. would be the largest that could be handled, but the lower range could be extended somewhat below 10 mesh. In order to take full advantage of the difference in specific gravities of the minerals being separated, sorting in the classifier should be under as extreme hindered-settling conditions as possible. Laboratory tests have been made on a galena ore and on an artificial mixture of quartz and magnesite to determine the separations that may be expected. Although more than one scheme may be followed, in these tests the separations were made by cutting the underflow from the classifier as a concentrate and then screening the overflow. The screen oversize was rejected as a tailing. Classification was done in a hindered-settling teeter column and screening was done by hand, using the standard 4 2 series. The accompanying flowsheet gives the testing procedure followed.

Jan 1, 1947

By W. M. Bertholf

In 1945 tests were made in the coal washery of the Colorado Fuel and Iron Corporation, at Pueblo, Colo., to determine the value of the Humphreys spiral. So far we have demonstrated that it is of definite value to us in rewashing a portion of our table middlings; and we are fairly certain that this is not necessarily the only place in which it can be used to advantage. Because of manpower and material shortages, experimental work has been discontinued for the present. Three of the spirals are being installed for use in rewashing the middlings, and testing and experimental work will proceed when this installation is completed. All evidence secured so far points to the fact that the spiral call make an effective separation of heavy material below 8-mesh; but that it is ineffective in removing larger, heavy material. This, of course, will limit its use to fine coal or a classified feed containing no heavy material larger than 8-mesh. In considering the results of our tests, it should be noted that the feed was our table middling, and that any real separation of this material is a "moral victory)" as there is little material that could be properly called coal and practically no heavy rock, the consequence being that previous attempts to clean the middlings have not been at all successful. Several articles dealing with the spiral and its use in treating heavy minerals have already been published,' but for the benefit of those who have not seen these we include a brief description of the spiral and of the way it operates, taken from Bulletin 4 of The Humphreys Investment Co.: The Humphreys Spiral Concentrator consists of a spiral conduit of modified semicircular cross section. Ore pulp is fed to the top and, as it flows downward, the heavier particles concentrate in a band along the inner side of the pulp stream (Fig. I). Outlets for removal of products are located at the lowest points in the cross section of the conduit. Wash water, added at the inner edge of the stream, flows outwardly across the concentrate band. The width of concentrate band removed at the outlets is controlled by adjustable splitters. Concentrate is usually removed from the upper portion of the spiral and middling from the lower portion. Tailing is discharged from the lower end. Tests are usually made in closed circuit, since this provides a maximum of information with a minimum amount of sample. Fig. 2 shows the spiral set up for closed-circuit testing, and gives a good idea of the general arrangement of such an installation. The small cyclone separator is used to provide clean wash water without introducing additional water into the circuit.

Jan 1, 1947

By E. H. Rose

The question is often pondered whether the flotation process is still an art or has become a science. The fact is that flotation is a science in so many variables that only art can blend them. It is demonstrable that no less than one hundred variables are involved in any flotation operation, most of them immeasurable on a routine basis. The behavior of every one of them is always governed by orderly and unchanging law, as no man can doubt. There is nothing truly indefinite about that behavior; the only indefiniteness is in our knowledge of it, either in general or from hour to hour in the circuit. In dealing with a simultaneous equation in, say, one hundred variables, all not only affecting the outcome to some degree but many influencing each other as well, it is small wonder that the mineral-dressing profession presents a case history of debate and controversial opinion. The statement of the subject of this symposium—a symposium on flotation machines—seems to delimit and define the matters to be considered and at least the general nature of the conclusions sought. Nevertheless, in such a meeting it is not possible to meditate upon all the variables that do in one way or another affect performance of flotation machines. To attempt it would be much like trying to operate a hundred-ring circus. Perverse variables being what they are, one could expect the elephant in ring 36 to squirt a trunkful of water at the tightrope performer in ring 37 to ruin that artist's act, just as our learning to depress pyrite in the early 1920's ruined the act of the formerly faithful old Callows. Honest Disagreements Selective flotation dates from 1922 or 1923, when several plants established their regular flotation operations on a selective basis. Since then, two decades of efforr have gone into the development of better and better flotation machines. Many have come and gone. The competition has simmered down to perhaps a dozen machines, each one of which is staunchly defended by some operators as being the best for their particular jobs, and as scornfully dismissed by at least a few others elsewhere as being positively the poorest for theirs. These disagreements are honest disagreements, honestly arrived at after plenty of sweat. In many a flotation plant, one machine has been tried out against another under conditions as fair as they could possibly be made. Sometimes the results are on a par but more often one machine shows a consistent and unmistakable margin of superiority over the other. In equally honest and enlightened tests, machine A may win over machine B in one place, and machine B over machine A by as great a margin in another. This paper purposes to identify some of the more prominent reasons why this is so,

Jan 1, 1947

By Gustav S. Preller, A. M. Gaudin

For the past 20 years it has been generally accepted that the flotation process is made possible as a result of the action of certain chemical substances on the surface of the mineral particles. In fact, it has been suspected, but never demonstrated, that this action produces on the mineral particles a layer of chemosorbed ions that is of the order of magnitude of a monolayer; that is, of a layer one ion deep. It is the purpose of this paper to present certain initial results on the relationship of mineral surface area and collecting-agent requirements. Up to about eight years ago, all considerations involving the surface area of finely divided solids were limited by the fact that there was not in existence any general method for determining the extent of such surfaces, especially on mixed materials. However, the method invented by Brunauer, Emmett and Teller1 has changed the situation by providing a general procedure (BET method). This procedure has already been applied to mineral dressing problems2-4 and has now received confirmation as to accuracy by an entirely independent method.' This confirmatory or "absolute" method uses a very tedious calorimetric procedure and is not suitable for routine determinations; but still another method6 now is available to relate the results obtained by the BET method to those obtained calorimetrically. This Harkins-Jura method has the advantage that it does not require an assumption concerning the parking area per molecule of adsorbate (nitrogen). In the original work by Brunauer, Emmett and Teller the parking area per nitrogen molecule was calculated to be either 13.8?2 or 16.2?2. according as the nitrogen was assumed to be closely packed as in crystalline nitrogen or loosely packed as in liquid nitrogen. The initial experimental work from this laboratory2 supported the 13.8?2. figure; that work was done on glass spheres of measurable geometric area and it was based on the assumption that the glass spheres were truly smooth; any deviation from this assumption would, of course, have made the calculated parking area higher. Harkins and Jura's results suggest that the parking area is not the same for all substrata; furthermore, the average of the 87 measurements cited indicates a "preferred" parking area for the nitrogen molecule of about 15.2?2., with extreme values at 13.7 and 16.9?2. The average of 21 determinations made in connection with this study7 gave a parking area of 14.3 ± 0.27?². In view of this result and of the Harkins and Jura

Jan 1, 1947

By M. D. Hassialis, A. F. Taggart

ark' observed some years ago that collector-coating reactions with xanthates and with fatty acids clearly follow paths parallel to those prescribed by the familiar mass-action law for reactions in dilute solutions. He proposed, therefore, the logical hypothesis that they are governed by the solubility-product principle. However, he went astray in the application of his proposal to collector coating by overlooking the fact that coating involves a heterogeneous reaction. A mistaken concept of the meaning of the word adsorption also entered. Wark thought that the concentration of metallic ion that must be used in setting up the solubility-product equation for the coating reaction was that which could be obtained by solution of the mineral. Since this is very low for sulphide ores, and since the amount of the collector ion necessary for coating produces an extremely low concentration, the resulting ion-concentration product, under conditions when coating occurs, is much lower than that which would indicate a saturated solution of the collector-coating compound. Hence he was forced to the position that the solubility product governing the coating is much smaller than the standard solubility product. To justify this position it was necessary to postulate a coating mechanism that did not obey the ordinary laws of chemical reaction. This Wark did by adopting the concept that adsorption is a property of matter, rather than merely a more or less ephemeral local concentration thereof. Thus he says (p. 142 of ref. I): "Adsorption is not a purely physical property of the solid surface, for the specific nature of the adsorbed molecule or ion is important." The result was that he destroyed the great potential utility of his postulate by tying it to this erroneous concept in the words (pp. 142, 143): "It will be obvious that the adsorption solubility product is less than the normal solubility product." The writers have found quantitatively in the case of galena and potassium ethyl xanthate, and semiquantitatively in the case of covellite and the same collector, that the coating reaction does follow the mass-action law to the extent that coating occurs when the ordinary solubility product of the coating compound is exceeded, and that coating does not take place under the reverse conditions. In the course of the investigation leading to the establishment of this fact it was necessary to work with xanthate concentrations so low (down to 0.0004 mg. per liter), that very long conditioning times, and bubble-contact times of as much as 16 hr. were necessary sometimes to follow the reactions. The results have led to some illuminating conclusions as to necessary collector concentrations in normal flotation operations, and the action of the frothing agent in connection therewith' The experimental work comprised accurate determination of the standard solubility product of lead ethyl xanthate, development of a much less laborious

Jan 1, 1947

By A. F. Taggart, Nathaniel Arbiter

The chemical reaction occurring in collection of nonmetallic minerals with amine-type collectors was early postulated by students of flotation phenomena to be metathesis between the mineral and the collector, involving exchange between the amine-bearing (aminium) ion and the cation of the mineral, and formation at the mineral surface of an oriented film of the amine salt of the mineral anion. Qualitative evidence of exchange has been obtained on several occasions at Columbia. Thus 1000 C.C. of a solution containing 495 mg of laurylamine hydrochloride was shaken 17 min. with 100 grams of barite and ground in a porcelain mill to pass a 200-mesh screen; the mixture was then set aside for 48 hr. Filtrate therefrom analyzed 476 mg. per liter of laurylamine hydrochloride. There was no substantial loss of chloride ion (78.4 mg. per liter in the original; 74 in the filtrate). No successful analyses were concluded for barium ion. Earlier, confessedly rough work with scheelite indicated a solution gain in Ca++ equivalent to about one-half the aminium ion lost. Real quantitative roof of the exchange, in the form of an established stoichiometric balance, is unknown to the authors. The data of the present paper, however, obtained from experiments designed on the basis of the exchange hypothesis, constitute strong evidence of its truth, and at the same time point the way to economic utilization of the expensive nitrogen-bearing collectors in concentration of the low-priced nonmetallic minerals. The basic postulate underlying the experiments performed was that, if the collecting reaction is of the type R+ + Y- + MX = RX + M+ + Y- where R = aminium ion, Y = the anion of the amine salt, and M and X are, respectively, the cation and anion of the mineral, the reaction should obey qualitatively the implications of the mass-action and solubility-product laws. This is to say that coating and collection should be aided by the addition to the solution of either ion (R+ or X-) of the collector-coating compound, and hindered by the addition of either of the ions M+ or y-. Experimental work comprised determination of (,) the relative solubilities of the laurylammonium salts of a variety of nonmetallic-mineral anions; and (21 the contact angles' of various minerals in solutions containing different concentrations of the postulated controlling ions. Semiquantitative flotation tests were run With barite as checks on the diagnostic character of the contact angles. The amine

Jan 1, 1947