Search Documents

By D. A. Flanagan, P. B. Crawford

A study has been made of the feasibility of storing liquid meihane at low pressures in undergrohd caverns. Methane liquefies at — 258°F at atmospheric pressure. It is shown that the methane evaporation rates will rapidly decrease and cool the surrounding rock so that at the end of one month they would be between 50 Mcf/hr for a cavern of 25-ft radius and 700 Mcf/hr for a cavern of 100-ft radius. At the end of 10 years, the evaporation rates would be 18 and 100 Mcf/hr, respectively, for caverns of the same radii. The evaporation rates may be reduced by a factor of 2 to 10 by the application of insulation. The cost for mining the caverns is estimated to be $.75 to $1.25/Mcf of storage. This is substantially less than surface storage; it is believed to be safer and to result in lower maintenance, savings in space and savings in strategic materials. INTRODUCTION During the past few years, there has been an increasing interest in the economic feasibility of liquefying methane. Methane liquefaction is being considered for tanker transport; in addition, liquefaction is being reconsidered for shaving peak gas demands. Several articles have described natural gas liquefaction and the progress of the tanker in making trial runs from the United States to Great Britain to determine the feasibility of tanker transport. At the present time, pipelines are not designed to supply peak gas loads during extremely cold periods such as are oiten encountered in the North and Northeast. Gas is being stored in underground reservoirs en-route to its destination, but in many instances satisfactory storage in porous reservoirs has not been practical, especially along the Eastern seaboard where few petroleum reservoirs have been found. In England and other foreign countries, it is unlikely that satisfactory porous structures could be found, and it may be desirable to mine or excavate the rock to obtain storage. By storing the methane near the consumption point, product availability can be increased during periods of need. Methane liquefies at - 258OF at atmospheric pressure, and 1 cu ft of liquid methane will make about 600 cu ft of gaseous methane. Methane liquefaction for shaving peak gas demands was conducted in Cleveland, Ohio in the early 1940's. The Cleveland plant was designed to liquefy 4 MMcf/D of natural gas.' Development of several low-temperature processes has resulted in some improvements in the design of low-temperature storage vessels. Liquefied-methane storage vessels can be designed for various evaporation rates. The desired rate will depend on the reason for liquefaction. For shaving peak gas demands, the desired evaporation rates would be low, possibly between 0.2 and 1.5 per cent/day. For methane tanker transport and supply, the evaporation rates must be high and must equal the total daily rate of supply. It is apparent that the cost of storage facilities will depend on the objective and will vary appreciably. Surface storage of methane may cost between $3,000 and $20,000/MMscf, depending on the size and design. The failure of the surface storage vessels at Cleveland, with the resulting injury and loss of many lives, has caused considerable emphasis to be placed upon developing new, safer and improved methods of storing liquid methane. Storage of liquefied propane and butane in underground salt domes has been very satisfactory. The per-barrel cost of underground storage for any sizeable capacity is very small compared to the cost of surface storage for LPG mixtures. The larger the capacity of the underground storage, the less is the per-barrel cost. Numerous advantages appear to exist if liquefied methane could be satisfactorily stored underground, the principal advantages being increased safety, lower initial cost, lower maintenance cost, savings in space and savings in strategic materials. The purpose of this paper is to determine the feasibility of storing liquid methane underground in mined caverns. If a large spherical cavern is dug underground and is filled instantaneously with liquid methane, the surface of the sphere may approach the liquid methane temperature almost instantaneously. The temperature distribution for a sphere of radius r. in an infinite medium, initially at one temperature with spherical surface kept at T, from time t = 0, is given by Eq. 1.'

By J. C. Yannopoulos, N. J. Themelis

Ah electronic thermogravirnetric balance was used to measure the veductioiz rule o single cirpric oxide particles suspended in a stream of hydrogen. Very jzne thermocouples embedded in lie center and at the surface of the sphere recorded the variation of terw perature during reduction. In contrast to iron oxide reduction, where in most instances the rate of interface reactiokl is controlling, the heat- and mass-transfer phenotnena play a predolrinant role in the reduction of cupric oxide. The correlation of experitnental data or tnass lransfev lhrough the boundary layer at Reynolds nunlber 0-4 50 was as jollocs: ThE reduction of cupric oxide by hydrogen is represented by the equation: It is a highly exothermic reaction and the amount of heat released varies only slightly with temperature (-21,430 cal per g-mole at 400°C). Most of the published studies on this system have been conducted at temperatures in the range of 34 to 280c.l-&apos; The concensus of opinion has been that at temperatures below 200°C the reaction is preceded by an induction period during which there is very little or no reduction. The length of this period diminishes as the temperature of the reaction is increased1l5 and aseawa found that at 280°C reduction starts practically immediately when the oxide is exposed to hydrogen gas. The existence of an induction, or incubation, period has led experimenters to believe that the reduction of CuO by hydrogen is catalyzed by metallic copper.&apos; Boldurev and rmolaev found that mechanical additions of powdered copper to oxide did not affect its reduction, although metallic copper produced during the reduction seemed to have a catalytic effect. At least three other authors2&apos;5&apos;7 have reported that addition of copper had no effect on the rate of reduction. Pease and alor&apos; observed that water vapor had a strong inhibiting effect on the reduction at 1503C, but its effect was greatly diminished at 200 C. Pavlychenko and Rubinchik5 studied this reaction at 159 to 235 C and found that, once the reaction had been initiated, there was no inhibiting effect due to the introduction of water in the hydrogen stream; at temperatures above 183°C there was no inhibiting effect due to water vapor, either before or during reaction. The same authors noted that the rate of reduction was independent of pressure in the range of 200 to 700 mm Hg. aseawa investigated the kinetics of the CuO-Hz system at 160 to 280°C and estimated the apparent activation energy at 14,000 cal per g-mole. A similar value (13,500 1200 cal per g-mole) was reported by ond&apos; in the temperature range 148 to 216°C. Bond also reported that there was no incubation period and that water vapor had no inhibiting effect on reduction at temperatures above 190 C. The above studies have been conducted at temperatures below 280°C and the physical configuration of the system under investigation has not been defined adequately. However, it is well-known that the mechanism of gas-solid reactions depends on a number of in series physical and chemical rate prcesses.&apos;&apos; It is, therefore, essential to include among the experimental factors the geometry of the solid system and the mass-transport characteristics of the reducing gas. Scott studied the reduction of packed beds of cupric oxide cylinders by Hz-He streams of very low hydrogen concentrations at temperatures 400 to 600 C, pressure of 10 to 30 atm, and gas velocity from 0.18 to 5.8 ft per sec (3 < Rep < 30). The reduction rate was found to be mainly controlled by diffusion through the boundary layer and through the pores of the reduced copper phase. The following correlation was proposed: Scott in an earlier work studied the reduction of fixed beds of cupric oxide wire and found that the rate of flow of the gas stream through the bed had a controlling effect on the rate of reduction. The same mass-transfer phenomenon was observed by Bond and clark13 on a similar reduction system. The analogy between mass and heat transfer through the boundary layer which exists between a solid surface and the bulk stream stems from the similarity of the differential equations representing these phenomena

Jan 1, 1967

By J. K. Elliott, P. H. Kelly

Many engineering functions such as surface metering work and laboratory compressibility check points involve the use of liquid densities of light hydrocarbon mixtures at various pressures and temperatures. However, at the present time, no simple reliable method exists for determining density variation, particularly if the composition of the liquid is unknown. Consequently, a study was undertaken to develop and present a simple and accurate method of predicting density variation of a light hydrocarbon liquid with pressure and temperature, knowing only the density of the liquid at some condition. The experimental liquid compressibility data from API Project 37 by Sage and Lacey&apos; have been considered to be accurate within 0.5 per cent and cover a wide range of pressure (14.7 to 10,000 psia), temperature (100" to 400°F) and molecular weight (up to 150). From these data, a set of liquid density curves, which relate density to pressure, temperature and molecular weight, was developed. These curves make it possible to predict density variation with pressure and temperature. Compared to extensive laboratory compressibility data on a complex, light hydrocarbon liquid, the use of the liquid density curves resulted in an average error of less than 0.5 per cent. Based on the results of this analysis, it is concluded that the set of liquid density curves developed from the data of Sage and Lacey provides an accurate and simple method for predicting the density variation of light hydrocarbon liquids when the density at some condition is known. These curves should be very helpful in many engineering calculations, particularly in the surface metering of light hydrocarbon liquids. INTRODUCTION Many situations arise in field and engineering laboratory work, such as reservoir engineering studies, check of experimentally determined laboratory data and orifice flow-meter formulas, where liquid density factors at various pressure-temperature conditions are required. Also, the need for accurate light hydrocarbon liquid information has become more important with the advent of miscible-type displacements for secondary recovery purposes in oilfield operations. Several reliable methods are available1 - "or determining the density of liquid hydrocarbons if the composition of the liquid is known. However, there is a definite lack of methods for accurately determining the variation of density when the composition of the liquid is unknown. The purpose of this study is to review the various methods for determining hydrocarbon liquid densities and to develop a simple and reliable method of determining variation in density of light hydrocarbon liquids with pressure and temperature when the compositio~n of the liquid is unknown. METHODS FOR DETERMINING DENSITY OF LIQUIDS OF KNOWN COMPOSITION Sage, Lacey and Hicks&apos; have proposed a method to predict the density of light liquid hydrocarbons by using partial molal volumes. Data are available on experimentally developed partial liquid volumes of hydrocarbons over a rather limited range of temperature, pressure and composition. The partial mold volume method has proved satisfactory for determining the density of some hydrocarbon liquids when the composition is known. Within the range covered in the Sage, Lacey and Hicks1 data, the results agree within about 3 per cent of the experimental values. Hanson mentions the limitation of this method to a composition range of approximately 10 per cent by weight of methane, which will not allow this correction to cover most low molecular weight-light hydrocarbon liquids. Standing and Katz2 studied data on light hydrocarbon-liquid systems containing methane and ethane at high temperature and pressure and have presented a method for determining liquid densities, assuming additive volumes for all components less volatile than ethane and using apparent densities for methane and ethane. The compressibility and thermal-expansion curves used by Standing are based on assumptions that compressibility of a hydrocarbon liquid at temperatures below 300°F is a function of the liquid density at 60°F and that thermal expansion of the liquid is affected little by pressure. The information required to use this technique with an example problem is furnished by Standing.&apos; Hanson eports an average error of - 0.5 per cent using the method of apparent densities in calculating liquid densities of several volatile hydrocarbon mixtures. However, as implied, the apparent density method is not applicable for liquid density calculations when the composition of the liquid is unknown. Watson- as presented a method

By K. H. Kilgren

At high formation pressures the distillate produced from a gas-condensate reservoir may be black in color. In this event the dense gas phase existing above the dew point is correspondingly dark. Volumetric phase data and an analysis of a reservoir fluid exhibiting these characteristics, together with a description of the visual equilibrium cell in which these observations were made, are presented in this paper. INTRODUCTION Previously the author, like many others in the oil and gas industry perhaps, tacitly assumed that the expressions black or dark oil system, crude oil system and bubble-point system were synonymous. Crude oil reservoir fluids are bubble-point systems and yield a black or dark stock-tank oil of relatively low API gravity. Conversely, a clear or amber colored trap distillate of high API gravity is assumed indicative of a dew point system or a gas-condensate reservoir fluid. This broad classification appears satisfactory for shallow reservoirs, but as the following study demonstrates, may be misleading when applied to deep reservoirs. Theoretically, there is no reason to exclude the possibility of producing a dark, low-gravity distillate from a gas-condensate reservoir. At sufficiently high values of pressure and temperature, heavy, dark-colored hydrocarbons may exist in the vapor state of a multi-component system. If enough dark-colored components are present in the reservoir vapor phase, the resulting condensate will be dark. The reservoir fluid investigated in the present study supports this contention. Stock-tank production was black in color and measured 29" APT gravity. From outward appearances, the liquid closely resembled a medium gravity crude oil. Experimental measurements proved the reservoir fluid was in reality a gas-condensate system. Volumetric phase data for this high-pressure system and a description of the visual cell in which the study was conducted successfully are presented in this paper. THEORY Phase behavior of a reservoir fluid can be predicted accurately with reference to a pressure-temperature phase diagram. If the reservoir temperature is lower than the critical temperature of the hydrocarbon fluid in place, bubble-point behavior will be observed. If the reservoir temperature lies between the critical and cricondentherm temperature, dew point behavior and retrograde condensation will occur. For reservoir temperatures above the cri- condentherm, only a single gas phase can exist in the reservoir regardless of pressure. Providing the composition of the reservoir fluid were known, it would be possible to predict the critical temperature and estimate the phase behavior from equilibrium relationships. However, the usual practice is to obtain a sample of reservoir fluid, subject it to varying pressures at the reservoir temperature and observe the phase behavior experimentally. The latter method was used to obtain the data reported here. WELL AND TRAPPING INFORMATION A summary of pertinent data for the well from which the reservoir fluid was sampled is presented in Table 1. This well is located offshore Louisiana. Except for the pressure which substantially exceeds hydrostatic pressure, the information does not appear unusual. Prior to the sampling program, the well was produced for 22 hours at a stock tank oil rate of 139 B/D. Average trapping conditions and gauging data for the six-hour test period that followed are summarized in Table 2. Samples of the first-stage trap gas and liquid were obtained during the latter portion of the test period. Ambient temperature remained 5 to 10F below trap temperature and presented no problem for sampling. Surface wind and moderate foaming of stock-tank oil presented some difficulty in obtaining accurate stock-tank gauges. SAMPLE ANALYSIS Compositions of the gas and liquid samples are shown in Table 3. The trap gas was analyzed by isothermal chromatography which revealed only a trace of heptane in the stream. The trap liquid was initially analyzed by low-temperature fractional distillation, yielding a bottom product of heptane and heavier components. Specific gravity of this fraction was measured and the mol weight was determined by freezing point depression. The hexane and lighter overhead gas collected during distillation was

Jan 1, 1967

By C. S. Matthews, M. J. F. Rosenbaum

The purpose of thir work wax to lcarn it~lzut infori~lation could he obtained from various typs of pilot water floods and to attempt to find the optunum pilot patter11, for a revervoir which had previously been depleted by a solution gas drive. The study was made in the laboratory with mathemetical methods a dynamic analog and a potentiotnetric analog. Results werp tested against the field llistorics of a nrrnlber of pilot water floods. At a reasonable valrre of currzulative injection, the total production rate for the one-injector five-spot should reach about 6.5 per cent of injection rate, and for a four-injector five-spot, about 9 per cent. Accurate estimates of ultimate recovery cannot be made on the basis of such snzall prorluction rates. However, with a pilot composed of nine ir1jector.s and 16 producers the production rate is approximately 50 pcr cent of injection rate at a reasonable value of camulative injection. Sonle inforn~ation for extended performance predictions might he obtained from such a large pilot. These conclusions were drawn on the basis of results obtained for unit mobility ratio, and a sturly using tlre potentiometric analog was made of the effect of other mobility ratios to determine the range of applicability of these predictions. For the four-injector, five-spot pilot with the ratio of production to injection rate (before water breakthrough) is about twice that for with it is about two-thirds; and with M0= 10, it is about one-third For high mobility ratios, it was found that the production rate increased considerably as water-cut increased. These result can be used to modify, qualitatively, the inter.pretntions based on curves for the unit rnobilit\. ratio CaSeS. It was found that the maximum ratio of production rate to injection rate obseriled in field pilot floods was of rhe scime order as that prerdicted by these methods. The time required to reach thisr maximum did not generally agree with the time predicted for a homogeti~orir reservoir. The differcrlce between predicted and observed time of response gives an indication of the permeability profile and of the condition of the producin,g wells. Pilot water floods of the pattern type are generally carried out in reservoirs which have been depleted by solution gas drive and are at low pressure. Under these conditions, oil and water can be considered incompressible. It is assumed that, as the water is injected, an oil bank forms ahead of it and that there is a distinct interface between the water zone (or bank) and the oil zone (or bank) and between the oil zone and the region ahead of the oil zone. It is further assumed that only gas is mobile in the unflooded (gas) region, only oil is mobile in the oil bank and only water is mobile in the water bank. The saturations and the mobilities associated with each zone are assumed uniform. We idealize our reservoir to be homogeneous, horizontal and of constant thickness. Effects of gravity within the producing layer are assumed negligible. If the actual time-dcpendent flow problem is approximated by a acries of steady-state problems. the potential and stream function in the oil bank and water hank satisfy Laplace&apos;s equation in two dimensions. We can therefore use a poteiitiometric analog of this system. Potentiometric models have yielded uscful results in this laboratory&apos; and clsewhere in the study of a variety of secondary recovery problems. For the case where M = I, we generally prefer to use theoretical mcthods as well as a simpler dynamic analog. Except where otherwise noted, the ratio, side of five-spot/wellbore radius. is taken to be 3,600. a figure which corrcsponds to a normal-size wellbore in a 10-acre well spacing. THEORETICAL EVALUATION OF VARIOUS PILOT PATTERNS, Mw0 = 1 <&apos;he theoretical models which we used to examine the performance of various pilots are shown in Fig. 1. Image theory was used to determine the ratio of production rate to injection rate as a function of the volumc of the flood. The ratio of production rate to injection rate was chosen because this is an easily measurable quantity which is characteristic of a pilot

By Peter Beardmore, Michael B. Bever

The effects of the temperature of deformation and the degvee of deformation on the annealing spectrum of a Au-Ag alloy have been determined by vesistance measurements. Specimens were deformed in torsion at 4°, 78K and room temperature up to strains nd/l of about 1.0 and were isochronally annealed. Discrete annealing stages similar to those found for pure metals were observed. Stage A occurred below room temperature; the decrease in reszstance associated with this stage increased as the temperature of deformation was lowered or the strain was increased. Stage B manifested itself as an increase in resistance; this is attributed to the restoration of short-range order promoted by vacancy or vacancy aggregate migration. At temperatures above 48S0K, recrystallization or Stage C occurred. Stages A, B, and C in the annealzng curves of this alloy may be considered to correspond to Stages II, Ill, and V, respectively, in the terminology used to describe the annealing spectrum of pure metals. ThE annealing behavior after deformation below room temperature has been thoroughly investigated on pure metals, in particular gold and copper. The occurrence of several discrete stages in the annealing spectrum of these metals has been established, but the interpretation of some of these stages in terms of specific mechanisms is still controversial.&apos; The annealing of alloys deformed below room temperature has been the subject of only a few investigations. In the investigation reported here, a Au-Ag alloy was annealed after deformation at and below room temperature. This continues a research program on the deformation and annealing behavior of Au-Ag alloys.2-8s EXPERIMENTAL PROCEDURE An alloy rod of composition 82.6 wt pct Au-17.4 pct Ag was reduced to wire, 0.0428 in. diam, by swaging and drawing. After a recrystallization treatment, the grain size was 0.031 mm, as obtained from the average linear intercept.7 The ends of wires about 6 in. long were flattened and sealed into short brass tubes with cold setting cement to form the grips of the torsion specimens. The equipment used for deformation has been described elsewhere.6 The specimens were deformed at room temperature, 78", and 4°K to various strains nd/l, where n is the number of turns, d the diameter of the specimen, and 1 its length between grips. The strain rate ?max was 0.1 min-&apos; at the sur- face of the specimen except where noted otherwise. The specimens were annealed isochronally for 30 min at intervals of approximately 30°K from the temperature of deformation to approximately 800°K, except that after deformation at 4°K the lowest annealing temperature was 78°K. A potentiometric method was used to measure the resistance: all measurements were made at 78°K with the exception of some measurements at 4°K mentioned below. Data obtained with a single specimen yielded a complete annealing curve: in several cases duplicate specimens were used to confirm the reproducibility of the curves. Since the surfaces of the specimens were rough after deformation, their diameters could not be measured accurately. The resistance increments after deformation and annealing are therefore reported as percentages of the resistance of the same specimen measured after completion of the annealing, that is, after full recrystallization. The corresponding resistivity values can be calculated by obtaining the effective diameter of a deformed specimen from the ratio of its resistance after recrystallization to the true resistivity, which was determined with a smooth specimen of known dimensions. This method of calculation was used to obtain data shown in several of the figures. RESULTS AND DISCUSSION The isochronal annealing curves after deformation to various strains at 4", 78K, and room temperature are shown in Figs. 1, 2, and 3, respectively. The resistance increment of the as-deformed specimens as a function of strain is shown in Fig. 4. For the specimens deformed at 4°K. values measured at 78°K are plotted with the exception of those for the as-deformed

Jan 1, 1970

By Craig R. Barreft, Oleg D. Sherby

The steady-state creep characteristics of pure polycrystalline copper were studied in the temperature range 400" to 950°C and in the stress range 400 to 7000 psi. Tests were conducted in dry deoxidized hydrogen and all specimens exhibited high ductility before fracture. Two activation energies for creep were established as a function of temperature. Above about T/T, = 0.65, the activation energy for creep was 48,000 cal per mole whereas between 0.5 and 0.65 T, it was 28,000 cal per mole. The observed temperature dependence of the activation energy for creep follows the same trend as exhibited by the self-diffusion coefficient of copper in copper single crystals and silver in silver single crystals. It is therefore believed that the rate-cmtrolling mechanism in creep of copper in the stress and temperature range studied is diffusion-dependent and is probably associated with a dislocation-climb process. The low activation-energy region is attributed to enhanced volume difFusion from the presence of dislocations which act as short-circuit paths for diffusion. THERE is extensive experimental evidence1,2 that the activation energy for creep, Q,, for pure poly-crystalline metals tested at temperatures above one half the absolute melting temperature (0.5 Tm) is a constant and equal to the activation energy for volume self-diffusion, Qsd This equality is generally interpreted as meaning that creep in this temperature range is controlled by a dislocation-climb process. The existence of high-temperature creep activation energies not equal to the respective self-diffusion values has been suggested both theoretically3"5 and experimentally.6-8 Two examples of apparent nonequality between Q, and Qsd have been reported for the cases of silver7 and copper,&apos; where activation energies for creep much lower than those for self-diffusion persist to rela- tively high temperatures (up to 0.6 T,). These observations suggest that dislocation climb may not be the only important process in the high-temperature deformation of these metals. That dislocation climb does indeed occur relatively slowly in these two metals due to their low stacking-fault energy has been suggested by many investigators.9-10 Although the creep behavior of copper has been studied extensively at low temperatures, the high-temperature activation energy for creep is not well-defined. The data of Feltham and Meakin8 indicated that Q, is temperature-dependent above one half the melting temperature, equaling a value of -30,000 cal per mole at temperatures up to 550°C and increasing to a value of -49,000 cal per mole at 700°C. No data has been reported above 700°C. The value of 49,000 cal per mole is very near the value of Qsd for copper reported by Kuper et al." as 47,100 cal per mole. In view of the lack of high-temperature creep data and the apparent temperature dependence of Q, for copper the present investigation was undertaken to determine the value of Q, over the temperature range of 400" to 950°C. It was hoped that such a study would clearly document the temperature dependence of Q, and yield significant information as to the nature of the rate-controlling creep mechanisms. MATERIALS AND PROCEDURES The material used in this investigation was high-purity OFHC copper. After rolling and recrystal-lization treatments spectrographic and chemical analyses showed only 0.0008 pct Ag, 0.0002 pctCa, 0.0002 pct H2, and 0.002 pct O2 as trace impurities indicating a purity of greater then 99.995 pct. Specimens were machined from cold-rolled sheet and the gage sections polished with 2/0 emery paper prior to annealing treatments. Annealing was done in both vacuum and dry deoxidized hydrogen atmospheres. No variation in creep strength was found as a result of the different annealing atmospheres. The final annealing temperature was either 700°, 800°, or 1000°C depending on the temperature range to be studied. These treatments resulted in average grain diameters of 0.03, 0.4, and 1.0 mm, respectively. The precise mechanical-thermal history of these

Jan 1, 1964

By C. F. Gates, H. J. Ramey

Recent literature shows that pronounced increases in oil recovery can result from the use of miscible systems in recovery operations. This literature also points out certain problems associated with maintaining miscibility, e.g., compositional changes resulting from alterations in pressure and temperature or from zone dilution. The purpose of the work described in this paper was to study miscible zones for possible use in waterflooding operations and examine the manner in which these zones break down. The fluid system selected for study was oil—terNary-butyl alcohol—water, which, because of its narrow range of miscibility, is ideally suited for investigation in short, linear systems. The laboratory-observed effects on oil recovery and zone stability of a number of variables are reported. Among the variables investigated were miscible zone size, viscosity ratio, length of travel, and interstitial water saturation. Results obtained with this system show that interstitial water adversely affects recoveries because of premature phase break. Salinity of the water further aggravates Ais situation. This effect, although pronounced for this system, would probably occur in any fluid system containing an alcohol. As expected, viscosity ratio has a decided influence on recovery efficiency and zone stability. The results also show that use of a suficieno zone size, even though the system has poor phase characteristics, yields higher oil recoveries than are obtained from straight water floods. Length-of-travel studies showed a square root relationship between zone growth and path length for favorable viscosity ratios. No such clear-cut dependence was observed under unfavorable viscosity conditions. INTRODUCTION As new oil reserves have become increasingly more difficult and expensive to find and develop, the oil industry has devoted more and more time and money to finding more efficient methods for exploiting known reserves. The increasing number of water floods is one manifestation of this attempt to improve oil recoveries from existing fields. Since considerable quantities of oil are by-passed by the waterflood process, it is only a partial answer to the problem. The search continues. Of late, several proposed processes have received the critical attention of investigators. Among them are gas drives, in situ thermal reactions and miscible displacements. The present paper is concerned with miscible displacements in conjunction with water floods. Recent literature1- reports that, under proper conditions, oil recoveries from LPG floods, propane sweeps, condensing gas drives, etc., approach 100 per cent of the oil in place in the region contacted by the flood. This literature also points out some of the problems associated with these processes, one of the main ones being the maintaining of miscibility. Compositional changes resulting from pressure and temperature variations or from zone dilution give rise to phase breaks and loss of miscibility, the conclusion being that, after loss of miscibility, much of the beneficial influence of the zone is lost.&apos; The results of the subject work show that, if sufficient zone material exists ahead of a phase interface, considerable improvement in oil recovery can be obtained. In order to increase our understanding of the behavior of miscible systems as they degenerate into immiscible processes, a laboratory system having rapid deterioration characteristics was needed. The fluid system oil—ter tiary-butyl alcohol (TBA)—water possesses these properties and was the system investigated. Although these fluids are different from those generally employed in miscible studies, an understanding of their behavior can be applied to other miscible slug processes. However, under conditions where mass transfer maintains a miscible system through multiple contacts, e.g., enriched gas drive or high pressure gas drive, the observed behavior of alcohol zones would not be applicable. The variables of interest were zone size, viscosity ratio, length of travel, initial water saturation, rate and core characteristics. Of these, only the effects of the following on zone breakdown and oil recovery are reported in detail: (1) zone size, (2) path length, and (3) initial water saturation. The effects of viscosity ratio and core characteristics are discussed in conjunction with other variables. Although initial water saturation is not envisioned as important to the phase behavior of hydrocarbon systems, e.g., LPG, it is important in alcohol systems. The properties of the fluids and cores used are given in Table 1. As previously stated, this fluid system was selected because of the limited range of miscibility

By F. Bonaccorso, G. Venturello, C. Antonione

A study of the effect of small amounts of interstitial impurities on recovery and re crystallization in high-purity iron (99.995 pct) has been undertaken. This paper gives results on the effect of carbon, introduced in small-dosed amounts @om 0.0005 to 0.0086 wt pct) by heating iron specimens of high purity m a static atmosphere of CO + ,. The materials prepared in this way, cold-worked 80 pct and subjected to a series of isochronal and isothermal annealings, were submitted to examinations by X-rays, micrographs, and hardness tests. It was observed that effect of carbon is remarkable in the sense of blocking recovery of mechanical properties up to the temperature at which re crystallization begins. On the contrary, carbon has a negligible effect on primary re crystallization temperature, when compared with the known effect of substitutional impurities. This is in agreement with the high mobility of interstitials. In effect, only a slight decrease, -20 pct, of the grain-boundary motion rate was noted, due to the interaction between the grain boundary and the carbon atoms. On the other hand, in the samples in which the carbon content is above the solubility limit at the temperature at which re crystallization occurs, a slight increase of nucleation frequency is noted due to the presence of precipitated carbides. ThE effect of purity on recovery and recrystalli-zation phenomena has been known for a long time; however, recently, new attention has been given to the problem since new methods for obtaining metals with extremely low impurity contents have become available. Most of this research work essentially concerns the effects of impurities which cause precipitates or give rise to substitutional solid solutions. The works of Bolling and winegard1 and Aust and utter&apos; on lead, and Vandermeer and orddon&apos; on aluminum, 01sen4 on nickel, and Abrahamson and Blakeney on iron should in particular be referred to. On the effect of this type of impurities a quantitative theory has been formulated by Detert and Lucke6 and has lately been discussed critically by Cahn7 and Gordon and vandermeer.&apos; Interstitial solid solutions in iron have not as yet been studied. As the study of the effect of interstitials is of great interest both from a theoretical and practical standpoint, it was deemed useful to examine the effects of carbon, nitrogen, and boron on recovery and primary recrystallization of iron. There already is some work by Chaudron et al.,1-" on the effect of interstitials on iron; their work, however, mainly concerns secondary recrystallization. The present paper refers in particular to the effect of carbon. EXPERIMENTAL PART Preparation of Materials. The pure iron used for this research was obtained from FeCIS recrystal-lized and purified by extraction with isopropylic ether. From the ferric chloride purified in this manner, hydroxide was precipitated with a solution of very pure ammonia, and, by calcination in pure sintered alumina crucibles, oxide was obtained. Reduction of the oxide to iron sponge was performed in sintered alumina tubes with very pure hydrogen at 650°C; at the end of the operation temperature was increased to 900°C. Specimens for the experiments were obtained by sintering the sponge at 1480°C in pure Hz after a pre sinter ing treatment at 900° C. It is important to note that the treatment at 1480°C in Ha produces a further purification from more volatile elements such as zinc, cadmium, arsenic, lead, and tin. Details on the preparation and characteristics of this type of very pure iron are given in a previous work.&apos;&apos; Only the complete analysis performed by neutron-activation methodsz3 is given here, Table I. Some of the specimens prepared in this way were carburized in the 0 region with very low amounts of carbon by treating them at 700°C in a static atmosphere of Ha containing a definite amount of CO. The set-up used is described in Fig. 1. A gas-tight quartz tube containing the specimen to be carburized and an internal friction control specimen, after being evacuated, was filled with Hz

Jan 1, 1963

By J. F. Martin, J. E. Friedline, L. M. Melnick, G. E. Pellissier

A method has been developed for determining gases in steel in which the gases are extracted by vacuum fusion and analyzed by mass spectrometry. This method is especially applicable for determining small amounts of oxygen and nitrogen. The lower limit of detection for both oxygen and nitrogen is 0.0001 pct with an average deviation of about 0.0001 pct for oxygen and 0.00008 pct for nitrogen. Hydrogen analyses by this method are in good agreement with analyses by vacuum tin-fusion. A complete analysis of all gases extracted from a steel sample can be accomplished in 40 min. It has long been known that certain of the gaseous elements exert particular effects on the physical and mechanical properties of steel. With the recent advent of new experimental steelmaking practices it is now possible to obtain steels with extremely low contents of such elements as oxygen and nitrogen. Present methods for determining these two gases are of insufficient sensitivity as applied to steels made by these newer practices, and therefore a new technique was considered necessary. A review of the literature showed that the conventional vacuum-fusion method1 was the procedure most suitable for modification to obtain a sensitive method for determining the gaseous elements present in steel simultaneously. However, for trace amounts of gases it is impractical to increase the sample size because of the physical limitations of the apparatus and the increased degassing time which would be necessary. Also, when only one of the desired gases is present in trace amounts, quantitative separation by conventional vacuum fusion is almost impossible. Since small analytical mass spectrometers are now commercially available, it was decided to use this instrument in conjunction with conventional vacuum fusion for both qualitative examination and quantitative determinations of the gaseous elements in steel. This technique has already been used with success for the analysis of copper and lanthanum.2"4 APPARATUS A vacuum-fusion unit was first designed and built, Fig. 1. This unit was designed especially for the determination of small amounts of nitrogen and/or oxygen in steel since an acceptable method (vacuum tin-fusion)5 was already in use for the determination of hydrogen in steel. The vacuum system was con- structed of Pyrex glass throughout. Mercury check valves were used in place of stopcocks in the high-vacuum section of the system. Ferromagnetic samples are normally introduced into the vacuum system through a mercury lift. In case of a break in the lift, the mercury is prevented from entering the furnace by a self-locking, mercury safety check valve, Fig. 2. This check valve, developed at our laboratory, is a modified T/S ball-and-socket joint placed in the sample inlet system between the mercury and the furnace. A sudden entry of mercury or air pushes this valve into the closed position, completely blocking the top of the mercury lift. The check valve is held in the closed position by the vacuum of the apparatus. A separate entrance tube also was placed in the system, so that samples which would amalgamate with mercury, or which are nonmagnetic, could be introduced into the furnace. A slightly modified Guldner-Beach furnace is used in this system.8 The optical flat was connected to the system by a ground-glass joint to facilitate cleaning, a quartz delivery tube was substituted for Pyrex, and the crucible was used uncovered. Fusion takes place in a high-purity graphite crucible insulated by — 200-mesh high-purity graphite powder. The crucible set (crucible and funnel) is suspended in a quartz crucible. The furnace envelope and the sample inlet system are exhausted by 2 single-stage, mercury-diffusion pumps in series. The first of these is a high-speed pump, which will work only with a very low backing pressure. This pump is used to exhaust the gases from the furnace as rapidly as possible. The second mercury-diffusion pump is used as a backing pump for the first; the capacity of this pump is much lower. However, it will work efficiently against a back pressure of 3 mm of mercury. To assure positive pumping, a modified Naughton and Uhlig condenser7 is used. With the use of proper valving the second pump is also used as a circulation pump in the catalyst system.

Jan 1, 1959

By N. De Nevers

A calculation method has been developed for carbonated water flooding. This method takes into account the effects of oil viscosity reduction and oil swelling, due to carbon dioxide transferred to the reservoir oil from the carbonated flood water. It shows the effect of changes in carbonation pressure and carbonated water slug size. The method is based on a Buckley-Leverett-type linear flow model. The mathematical approach is similar to that developed by Welge, et al, for enriched gas drives. However, it contains a very important advance over the previous papers in its solution to the problem of injection of a "slug" of carbonated water rather than continuous injection of fluid of constant composition. The method also shows the chromatographic nature of the transport of Co 2 in a carbonated water flood. Sample calculations indicate that viscosity reduction is the most important source of incremental oil recovery. Swelling contributes to a lesser extent. Raising the carbonation pressure increases recovery, but this effect diminishes as pressure is increased. increasing the slug size also increases recovery, but again this effect diminishes as slug size is increased. INTRODUCTION In recent years there has been considerable interest in carbonated water flooding as a secondary recovery method. Several carbonated waterflood field tests are now in progress.1&apos;3 Yet there is no published method for predicting the results of a carbonated water flood. This paper presents a method for predicting the oil recovery for ; carbonated water flood. In a carbonated water flood. a carbonated water slug is injected into the reservoir. Then this slug is driven forward by injecting plain water behind it. The field operator must choose the carbonated water slug size and the CO2:water ratio in the slug. The calculation method presented here shows the effects of both slug size and CO2:water ratio on oil recovery. The calculation assumes that the addition of carbon dioxide to flood water increases the oil recovery by two mechanisms: 1. Carbon dioxide transfers from the flood water to the oil, thereby reducing the viscosity of the oil. At the lower viscosity, the oil flows more readily and is more easily swept out by flood water. 2. The carbon dioxide which transfers to the oil phase causes the oil phase to swell. Assume that the residual oil saturation (volume per cent of pore space) is the same whether the oil is carbonated or not. Then the same volume will contain less stock tank oil if it is carbonated, due to oil swelling. The plain water driving the carbonated water slug forward will strip the CO2 out of this residual oil., causing this oil to shrink. This swelling-followed-by-shrinking leaves behind a lower residual oil saturation than would have been left behind by a plain water flood. Since some oils swell to almost twice their original volume at reasonable CO2 pressures, this effect may be large. Two additional benefits sometimes claimed for carbonated water flooding are: 1. Adding CO2 to flood water greatly increases the water injectivity in some cases.1 (Injectivity = BPD of water injected/ft of sand x differential bottom-hole pressure, psi.) CO2 forms a weak acid in water; injecting carbonated water is, therefore, somewhat like acid treating. If a formation responds to acid treating, it will probably have a higher injectivity for carbonated water than for plain water. Although the increased injectivity increases the profitability of a carbonated water flood over a plain water flood, it does not change the production curve based on total fluid injected. Therefore, the method presented here does not show any effect of increased injectivity. 2. It is claimed in one patent4 that carbon dioxide will react with some constituents of the oil to form detergents. These detergents are expected to increase the oil recovery. There is little data available to support this idea. The calculation method presented here takes no credit for extra oil recovery due to the formation of any such detergents.

Jan 1, 1965

By B. H. Caudle, L. H. Silberberg

Reservoir depletion by natural water drive is typified by the movement of water from an aquifer into the adjacent oil-bearing formation. Prior studies of this type 01 water movement have generally neglected the resistance to flow in the aquifer and its effect on the movement of water into the oil bearing zone. A method for designing and operating scaled models of such reservoir systems is presented. Experimental data on a model of an edge-water-drive reservoir are shown and discussed. INTRODUCTION Most petroleum reservoirs derive at least a part of their productive capacity from water influx. This water may be injected from the surface or it may come from an aquifer adjacent to the oil zone. In either case, the reservoir engineer must be able to estimate the advance of the water as a function of either the elapsed time or the fluid produced. This paper describes a type of fluid flow model which includes the effect of viscous and gravitational forces in the reservoir and the surrounding aquifer. Natural water influx can be divided into three general (but widely overlapping) categories according to the direction of flow in the aquifer. These are shown schematically in Fig. 1. Fig. l(a) illustrates an edge-water-drive mechanism. In this case, water advances updip along the stratum but the movement of water is essentially horizontal and very little of the oil is actually underlain by water. The bottom-water drive (Fig. l(b)) is characterized by a thick aquifer underlying the oil zone. The water movement is generally vertical in the aquifer. The third category is illustrated in l(c). Referred to as the "thin oil column", this type of oil accumulation consists of oil over water in a relatively thin stratum. Fluid movement is horizontal in both the oil and water zones, except close to the producing well. One characteristic which all three of these types have in common is that part of the water flow takes place in the water invaded region of the oil reservoir while the remainder of the water flow occurs in the aquifer. The resistance to the flow of water usually will not be the same in these two regions. This occurs because the porous rock normally contains only water in the aquifer, while the water-invaded region always contains microscopic globules of bypassed oil (residual oil) which interfere with the flow of water through the rock. The production history of any type of a water-drive system is a function of two phenomena — where the water goes, and how it displaces the oil in the area invaded by the water. It is usually impractical to study these two phenomena at the same time. In most cases one type of analysis is used to predict the gross movement of water in the reservoir, while a different type of analysis is employed to determine the amount of oil to be recovered from the gross volume contacted by the water. The portion of the oil reservoir which is invaded by water is mainly a function of the resistances to fluid flow in the several parts of the flow system. This gross water influx in the three types of natural water drive typically results in water cusping to the production wells either from the side or from below (water coning). In 1747, Muskat described a greatly simplified mathematical model for predicting the area invaded by water in a bottom-water-drive reservoir.1 This model used the Laplace equation and suitable boundary conditions to describe the isopotentials and streamlines in the flow system. The use of this model assumes that (1) the reservoir rock is homogeneous in nature, (2) the oil and water mobilities are equal, (3) there is no oil flowing in the water invaded region, and (4) external forces — such as gravity — do not affect the flow. This method has been used for both natural water drives and water injection projects. In this, as in most models, the method of images is used to reduce the size of the model necessary to describe the reservoir. The assumption of equal viscosities and permea-

Jan 1, 1966

By D. W. Reed, R. L. Reed, Tracht

Laboratory experiments on the reverse combustion of tar sands in a linear adiabatic system have shown that a highly upgraded oil can be produced from an exceedingly viscous, immobile oil. The dependence on the air-injection rate of peak temperature, combustion-zone velocity, oil recovery, air-oil ratio, residual coke and oil, fuel burned and distribution of product gases is shown graphically. Eflects of initial temperature, oxygen concentration, oil saturation and heat loss are discussed. Experiments bearing on the coking properties of heavy oils are mentioned and some results exhibited. Field application of the process hinges on the existence of adequate air permeability and the rate of reaction under reservoir conditions. INTRODUCTION It has been established that oil can be recovered from underground reservoirs by means of at least two fundamentally distinct methods involving in situ combustion of a certain fraction of the oil. Characteristic of both of these known methods is the production of oil from one or more wells by means of hot gases formed when a high-temperature reaction zone is advanced through the reservoir. In both cases, the reaction zone is created by heating certain of the wells to a sufficiently high temperature prior to the introduction of air, and the zone is maintained and advanced through the reservoir by appropriate control of the air-injection rate. In the first of these methods, which is called "forward combustion",&apos; the combustion zone advances in a direction which is generally the same as that of the air flow; whereas in the second method, "reverse combustion",&apos; the combustion zone moves in a direction generally opposite to that of the air flow. Forward combustion, on the one hand, is an ideal combustion process in the sense that a minimum of the most undesirable fraction of the oil is consumed as fuel in the form of coke, a clean sand is left behind and generated heat is used as efficiently as possible. However, the applicability of forward combustion is limited. Since the products of combustion, vaporized oil and connate water must flow into relatively cold regions of the reservoir, there is an upper limit on the viscosity of oil which can be moved by this process in a practical and economical fashion.&apos; On the other hand, it is characteristic of reverse combustion that the vaporized oil and water together with the products of combustion are produced through sand which is already hot and has had its mobile liquid content eliminated. This means there is no upper limit on oil viscosity; indeed, the oil may be an entirely immobile semi-solid. However, fuel for the process is an intermediate fraction of the original oil, and the most undesirable fraction remains on the sand surface as a substantial deposit of coke. Since this coked material is not burned during reverse combustion, it represents energy which is available for the production of oil but is not used for this purpose. It follows that one can expect economics to be somewhat less attractive with reverse combustion than with forward combustion. Nevertheless, it is a process which is designed for reservoirs where forward combustion is impossible and, as such, has become a subject of experimental and theoretical investigation. In this paper, only experiments made with tar sands are discussed. DESCRIPTION OF THE PROCESS We proceed, then, to consider the process of reverse combustion in greater detail. Fig. 1 illustrates a temperature profile defining a combustion zone which moves from right to left when air flows from left to right. In Zone 1, the temperature is the initial reservoir temperature, and the tar sand is as yet unaltered. This statement must be modified to the extent that physical properties of the oil may be changed by low-temperature oxidation at reservoir temperature. As air passes into Zone 2, which has been warmed by conduction, it assists in vaporization of the very light ends (if there are any), and oxidation occurs at a significant rate. In this region, there is almost no production of carbon monoxide or carbon dioxide because predominantly addition-type reactions take place with the formation of oxygenated compounds such as aldehydes and acids together with water. The hydrocarbon-enriched and slightly oxygen-depleted gas stream enters Zone 3 where

By A. E. Dwight

Lattice parameters were determined for eighteen equiatornic alloys of the CsCl-type structure, ten of which were previously un-reported. It was found that fomation of the CsCl-type structure in binary alloys of the transition elements is largely dependent on position of the elements in the periodic table. The relative size of the two elements was not found to be a controlling factor. A recent paper by Beck, Darby, and Arora1 corre-lates the occurrence of CsC1-type ordered structures with the position of the constituent elements in the periodic table for the first long period. It was also suggested that a definite increase in relative bond strength between unlike atoms occurred when, in binary alloys of iron-group elements, the other component is changed from a chromium-group element, to a vanadium-group element, to titanium. A later paper by Philip and Beck2 noted that the lattice contraction increased in the order CrFe, VFe, and TiFe. It was also noted by Philip and Beck2 that the lattice contractions of CsC1-type alloys decreased in the order: TiFe, TiCo, and TiNi, which is an apparent reversal of the contractions expected from the position in the periodic table. It was suggested that the increasing lattice contraction is an indication of increased stability, i.e., greater A-Bbond strength. The present investigation was carried out to determine whether the relation of the position in the periodic table to the formation of the CsC1-type structure was also correct for alloys involving the second and third long-period elements. A systematic search was made for CsC1-type structures among equiatomic alloys and for those found, the lattice contraction was determined. EXPERIMENTAL TECHNIQUE The elements Y, Gd, Ti, Zr, Hf, V, and Cb are designated the A group and the elements Mn, Re, Fe, Ru, Os, Co, Rh, Ir, Ni, Pd, Pt, Cu, Ag, and Au are designated the B group. Equiatomic alloys were prepared for 57 AB combinations. The alloys were arc melted in a multicrucible furnace3 in buttons ranging from 5 to 20 g. Chemical analyses were not made, as the charge weights agreed closely with those of the buttons after melting. The alloy buttons were homogenized at 800° to 900°C. Metal log raphic and X-ray specimens were prepared and heat treated at temperatures from 600° to 1200°C. Specimens for X-ray diffraction were usually ground to a powder in an agate mortar; however, needle-shaped solid specimens were used when the alloy was sufficiently ductile to permit their preparation. Diffraction patterns were taken with a Straumanistype Debye-Scherrer camera using filtered Cu or Co radiation. The lattice parameters were obtained in A by plotting the calculated a0 values against the cos29/sin 0 + cos2?/? function and extrapolating linearly to ?= 900. Metallographic control specimens were polished on cloth wheels with diamond paste and etched with various phosphoric and nitric acid reagents. RESULTS The eighteen equiatomic alloys listed in Table I gave evidence of a cubic structure with two atoms in the unit cell, although two of these cubic structures exist only at elevated temperatures and transform to a tetragonal structure on quenching. Nine of these eighteen alloys gave diffraction patterns with super-lattice lines showing that the structure is of the CsC1-type. The lack of superlattice lines in patterns of the other nine alloys may be attributed to the small difference in atomic scattering power of the components. Metallographic study indicates the occurrence of nine narrow single-phase fields at the AB composition. Any or all of these nine may also have a CsC1-type structure. The VFe alloy was found to have a CsCl-type structure by Philip and Beck2 through the use of CrKa radiation (for which the scattering factor of V is anomalously low), whereas the Cu radiation used

Jan 1, 1960

By Ben-Zion Weiss, Melvin R. Meyerson

Chromium diffusion coatings on commercial Armco iron lead to carbide precipitation at the grain boundaries in and below the coatings. High compressive stresses are introduced into the coating and, as a result, tensile stresses are introduced into the base material below the coating. In coated samples, fatigue cracks form at the grain boundaries below the coating after only a limited portion of the total lifetime (5 to 10 pct). Residual tensile stresses and a stress concentration caused by precipitated carbides seems to be responsible for early crack nucleation. Stage I of Propagation may be divided into two substages: I-a in which the crack Propagates along grain boundaries, and I-b in which the crack propagates along slip boundaries according to a shear mode. In uncoated samples, the cracks form at slip bands after 40 to 50 pct of the total lifetime. In coated samples the Propagation process takes longer than in uncoated samples because of the moderate rate of crack extension until the crack breaks through the coating. The chromium-diffusion coating causes little if any increase in fatigue life. CHROMIUM diffusion is one of the most popular processes used to apply coatings to iron alloys. This popularity stems from the fact that the diffusion treatment is comparatively easy to carry out, and it improves certain surface properties such as corrosion and wear resistance as well as some other properties. Very little effort has been devoted to the investigation of the effect of chromium-diffusion coatings on mechanical properties and particularly on fatigue properties. Since the chromium coating usually represents a small fraction of the total volume of the base material it is generally assumed that the macro properties are dependent principally on the base material rather than the coating. Insofar as fatigue is concerned, there are indications that the fatigue life of the chromium coating is not less than that of the basic material and the fatigue properties are generally not reduced by it.&apos; But chromium diffusion causes drastic structural and chemical changes in the region of the material surface2-4 which introduce additional residual stresses. Such changes must affect fatigue crack nucleation and may sometimes affect the initial stages of propagation. Furthermore, it is conceivable that the secondary stages of the crack propagation could be influenced by some irreversible structural changes in the core of the material, which stem from the long-time heat treatment at high temperatures required to apply the coating. This paper analyses the structural and compositional changes produced in commercial Armco iron by a chromium-diffusion coating and the effect of the coating on fatigue crack initiation and propagation. EXPERIMENTAL PROCEDURES The investigations were carried out with commercial Armco iron. The chemical composition of the iron, the number of specimens tested, the grain size produced during chromizing and heat treating, and the thickness and microhardness of the chromized layers are shown in Table I. Samples were chromized by a gas diffusion process. The temperature and time of the diffusion coating process were chosen according to a previous grain growth study. It had been found that at a temperature of 970°C and a holding time of between 3 and 9 hr there are practically no changes in grain size. To describe the grain size more accurately the ASTM method5 was supplemented by a statistical analysis of the mean volume diameter as outlined by Fullman6 and The The mean volume diameter was deter- mined from D = p/2m where D is the mean volume diameter, and m is the mean reciprocal value of grain diameters as seen on the micrograph. The shape and the size of the asymmetrical sample, see Fig. 1, were chosen so as to facilitate the study of fatigue crack initiation. Before coating, the samples were machined and the surface of the groove was polished. After chromizing, the two side surfaces were machined and polished so as to retain the coating only on the top surface of the groove. Residual stress in the chromized layer was measured on the top of the grooved section of the specimen after the diffusion process. The inclined incident X-ray beam procedure was used.9&apos;10 The computations were performed with the initial assumption of a zero surface normal stress component. An electron microprobe analyzer was used to obtain the relative amounts of chromium and iron con-

Jan 1, 1970

By C. Gatlin, N. E. Garner, A. Podio

Experimental data from single chisel blows on Leuders limestone are presented. A pressure chamber, similar in design to well known microbit drilling chambers, was utilized to impose variorcs stress states on the sample. Confining pressure of zero to 10,000 psi and borehole pressures from zero to 5,000 psi have been used in the studies. Pore pressure was zero and the rock samples dry in all instances. Force-displacement records and visual examination of the craters indicate that the mode of failure depends on both confining and borehole pressure in certain ranges, ranging from brittle, through a transition, to near plastic. The mode of failure is reflected in the observed blow force and energy data, as well as the shape of force-penetration curves. INTRODUCTION A previous paper&apos; presented initial data from the project on which this paper is the second report. While the first paper dealt with impact studies on synthetic rocks, the literature cited there is also pertinent to this paper, but will not be listed again in order to avoid needless repetition. This paper presents experimental data from single chisel impacts on limestone at certain simulated wellbore stress states. Specific variables investigated were combinations of borehole and confining pressure, crater geometry, a narrow range of impact velocity and the forces required to drive the chisel into the specimens under the varying conditions. EXPERIMENTAL PROCEDURE The same basic apparatus described in our earlier paper, except for the pressure vessel, was used in this study. Fig. 1 shows the complete experimental system. Fig. 2 shows details of the pressure cell, which is quite similar to the well known microbit drilling chamber. The confining (overburden) pressure system provides independent pressure control over the sample except for two, 21/4 in. diameter areas located on the sample ends. These two areas provide a pore pressure entrance on the bottom and a striking surface on the top where borehole pressure can be controlled. The ends are isolated from the confining pressure by O-ring seals. The borehole and pore pressures are related through the striking surface; they will be the same unless a seal is deposited on the borehole surface, either a mud cake or some other impermeable membrane. In these studies pore pressure control was not used, as the rocks were dry and unsaturated in all cases. Rock samples were prepared from 4-in. cubes as illustrated by Fig. 3. The circular disks of adhesive vinyl plastic protected the impact surface from the resin coating (Scotchcast No. 2) and were removed just prior to each test. This procedure insured a fresh uncontaminated surface for the impact. Fig. 4 shows the dimensions and loading of the sample, including the concentrations around the "borehole" periphery (O-ring seal). The rock used in this study was Leuders limestone, of Permian age (Leonard), quarried near Leuders, Tex. Geologically, Leuders limestone is a light gray, fossili-ferous limestone (oolitic foraminifera1 biosporite). Its fossil content is 80 to 90 per cent and consists primarily of calcitornellid, ostracods, pelecypods and oolites with the remaining part intraclasts. There are hematite or limo-nite rims on some of the intraclasts. The porosity is approximately 20 per cent, but the permeability is less than 1 md. There is no apparent orientation in the structure. Triaxial tests have been conducted to determine the variation of the physical properties of Leuders limestone in three orthogonal directions&apos;, and it is known to be unusually isotropic and uniform. Physical parameters for this rock are: uniaxial compressive strength = 9,700 psi; F = 32 and C = 2,700 psi; where + and C are Coulomb equation values (t = C+u tan +). Impact tests were run under the combinations of confining and borehole pressures shown in Table 1. Samples were dry with atmospheric pore pressure in all cases. Tests, in which elevated borehole pressures were applied, used a single layer of household Saran-Wrap to prevent horehole fluid invasion into the sample. This proved to be simple and reliable. The blunt wedge used in all the tests was 0.75 in. long and had an included angle of 60" with a 0.05-in. flat. In addition, a sharp 60" wedge was used for all confining pressures at zero borehole pressure. The chipper speed

By E. A. von Hahn, T. R. lngraham

The rates of cementation of pd11 on electropolislzed copper cylinders were studied in aqueous perchloric acid solutions at pdII concentrations from 0.02 to 0.1 mM. The cylinders were rotated at various high speeds to reduce the diffusion layer to a minimum thickness. In 0.1 M HClO4 solutions, the cemented palladium is dense, adherent, and shiny. The rate data indicate that there are two stages in the deposition. In the initial stage, the rate of cementation is first order with respect to the pd11 concentration. The second stage is much slower. The first stage is consistent with rate control by the diffusion of pd11 ions to the copper surface, and/or chemical reaction at the surface, and the second stage is consistent with rate control by the diffusion of copper ions from the cop-per surface through the deposit of cemented palladium, out to the main body of solution. In 0.001 M HClO4 solution, only the first stage is evident and the rate is more rapid than at higher acidities. This rate enhancement is attributed to PdOH+ ions that predominate at low acidity and aye more reactive than unhydrolyzed pdII. The deposit is porous and loose. All of the cemented deposits are Pd-Cu alloys rather than pure palladium. Activation energies are 9.5 kcal per mole in 0.1 M HClO4 and 7.4 kcal per mole in 0.001 M HClO4. CEMENTATION or displacement reactions occur between aqueous solutions of metal salts and immersed metals according to the general equation: where M1 is electrochemically the more noble metal. Although these reactions find considerable application in metals recovery processes (e.g., the cementation of copper192 or gold3), in electrorefining processes (e.g., solution purification before electrolysis4,5), and in plating and metal finishing processes,6 few studies have been made of the kinetics of such reactions.7-9 The rates of cementation reactions will depend on one or more of the following factors: a) chemical reactions at the metal-solution interface; b) ionic transport to or away from the interface;&apos;-&apos; and c) the character and adherence of the cemented deposit on the substrate metal,6a,10 because the deposit will inhibit the rate of transport of dissolving ions (M2m+) into the solution. In this investigation, the kinetics of the early stages of cementation were studied to obtain an understanding of the reaction mechanism under conditions in which the deposit was sufficiently thin that its inhibiting effect could be disregarded. To achieve this, very dilute solutions in Mn+ were used. Cementation rates are often fast in the initial stages and transport-contr01led.9 To find conditions under which transport control would shift to chemical control, very high stirring velocities were used to minimize the diffusion-layer thickness. Of the many possible cementation reaction systems, the palladous perchlorate-copper system was chosen for this investigation because it was believed to involve simple ion-for-ion exchange. In addition, there are no interfering side reactions, such as the reduction of hydrogen ions, and anion effects are usually absent in perchlorate solutions. EXPERIMENTAL Materials. Reagent-grade chemicals and redistilled water were used in all experiments. Palladous perchlorate stock solutions (0.02 M) were prepared by dissolving 99.5 pct Pd sponge (Johnson Matthey and Mallory Ltd.) in concentrated nitric acid, adding concentrated perchloric acid, evaporating twice slowly (to prevent PdO precipitation) to a small volume, and diluting in a volumetric flask. To prevent gradual hydrolysis of PdII, the stock solutions were made 0.4 M in HC104. Tests for pdIV (Ref. 11) and C1- were negative. Copper strips were cut accurately from 0.025-in. electrolytic tough pitch sheet, Sample 1 (Mines Branch stock or Canadian Copper Refineries Ltd., Montreal), or from oxygen-free, high-conductivity sheet, Sample 2 (American Metal Climax O.F.H.C. brand, sold by Utility Brass and Copper Corp., Brooklyn). The strips were annealed for 1 hr at 335°C in a stream of purified nitrogen (Union Carbide, Linde Division) which had

Jan 1, 1967

By J. M. Tobin, L. H. Cadoff, L. M. Adelsberg

The reaction between liquid zirconium and graphite at temperatures above 2000 °C has been investigated. The reaction products were found to be carbon-saturated zirconium metal and ZrC which formed between the graphite and the metal. Parabolic growth behavior was observed for the ZrC Phase at all temperatures of this investigation. The parabolic growth constant at temperatures between 2000° and 2860°C was measured to be 1.83 exp 84,300/RT) sq cm per sec. The reaclion mechanism was proposed to be the rapid carbon saturation of the liquid metal and the formation of ZrC at the metal-carbide interface with diffusion of carbon through the ZrC, the "rate - determining step" of the reaction. The concentration-independent diffusion coefficient of carbon in ZrC, (DZrC), was expressed as 0.95 exp? 78,700/RT) sq cm per sec. This value mas calculated using the temperature-invariant ZrC phase fields proposed in the literature. The [ZrQiq)] — [ZYQiq) + ZrC] phase boundary over the temperature range 2000° to 2800°C was determined and the ZrC + C eutectic temperature was found to be 2890° ± 50°C. ThE Group IV and VB transition-metal refractory carbides are of interest because of their high melting points, high temperature strength properties, and relative inertness in certain corrosive environments. The present-day understanding of these materials, however, is limited by the general unavailability of accurate and reliable physical and chemical property data. This is due primarily to the difficulties associated with the preparation of suitable, high-density, high-purity carbide samples, and the achievement and control of uniform high-temperature environments. Accurate measurement of temperature is also an important factor limiting the reliability of reported data. The Zr-C reaction was selected for investigation because of the high melting point (>34000C) and favorable nuclear properties of the reaction product, ZrC. The direct reaction method afforded an opportunity to obtain kinetic data on the fully dense carbide. In this paper, layer-growth techniques were used to estimate the diffusivity of carbon in ZrC and to investigate the phase equilibria in the Zr-C system at temperatures above 2000°C. EXPERIMENTAL PROCEDURES AND DATA Crystal bar zirconium (99.9 pct Zr), purchased from the Nuclear Materials and Engineering Corp., and ZTA graphite (99.9+ pct) were used in this study. The analyses of the materials are listed in Table I. A schematic of the high-temperature carburization apparatus is shown in Fig. 1. The sample was a zirconium metal charge in a 1/2 -in.-ID graphite crucible which was capped with a tightly fitting graphite stern. The crucible and stem were designed to closely approximate black-body conditions. The graphite crucible was packed in lampblack (outgassed 1 hr at 2100°C) which provided insulation and thermal stability to the system. The inert atmosphere was maintained by a constant flow of high-purity argon or helium gas. The nozzle-diffuser section on the top flange was sufficient to prevent back-diffusion of air into the system. Chemical analysis of the carburized samples revealed oxygen and nitrogen concentrations of less than 44 and 20 ppm, respectively. The sample was heated by induction with a Westing-house 5 kw-450 kc power source. Temperature was measured with a Milletron two-color pyrometer which was sighted into the crucible by reflection from an overhead front surface mirror. Optical losses due to glass absorption and light reflection at air-glass interfaces were minimized by the employment of the dynamic-gas seal. It was found that no correction was required for the front surface reflector. The pyrometer was calibrated periodically against a U.S. Bureau of Standards tungsten ribbon secondary standard. Temperatures inside the crucible were also calibrated against the Zr-ZrC eutectic (1850°C)1 and the Nb-Nb2C eutectic (2335°C)2,3 temperatures and agreed to within ±10°C of these values. The temperature was controlled by manually adjusting the power input; variations of

Jan 1, 1967

E. A. Loria (Product Metallurgical Engineer, Crucible Steel Co. of America, Pittsburgh)—In this interesting paper, our introductory work was quoted. We would like to call attention to our sequel paper on the experimental determination of oxygen in cupola-melted cast iron,20 which was not mentioned. Vacuum-fusion oxygen values (as well as hydrogen and nitrogen) were reported for nine heats of cast iron melted in the Battelle 10-in. cupola under normal operating practice and under oxidizing conditions. The oxygen analyses ranged from 12 to 68 pprn compared to the author&apos;s computed range of 10 to 80 ppm. The average amount of oxygen found in our irons was about 20 pprn and changes in the silicon content of the iron from 1.32 to 2.35 pct had no consistent effect on the oxygen content of the iron. The gas determination specimens were poured in split steel molds that produced a clean pin, 3/8 in. diam and 2 in. long. Because freezing was almost instantaneous, the pins were entirely white iron (nongraphitic). In the early stages of the investigation, the pins were transferred to a mercury-filled trap system immediately after pouring. This was done to collect gas evolved between pouring and analysis. However, it was found that during storage for 4 weeks gas evolution was negligible. Because the vacuum-fusion analysis was usually completed within 4 days of pouring, pins from later heats were not stored in the mercury-trap system. We found some evidence that cast iron picks up oxygen during long storage, because of rusting. Earlier work by the British Cast Iron Research Association has shown that cast irons may be stored for a long time without significant change in their oxygen content. The practical significance of this study (and our own) would be in the improvement of cast-iron quality. Has the author investigated this aspect and reached any conclusions on the effect of oxygen on the mechanical properties of cast iron? The second phase of our study was to determine the properties of the test bars poured simultaneously with the gas analysis specimens. We realize that there may be complicating factors attendant in this procedure.21 Results from many test specimens measuring chill depth, transverse flexure and deflection strength, spiral fluidity, and sensitivity to hardness of gray irons ranging from 12 to 68 pprn oxygen showed that the lowering of transverse strength was the only significant undesirable effect of high oxygen content. A statistical study of the chill test results21 showed that the iron containing 22 to 46 pprn oxygen had forced chill depths that were 2/32 in. below the expected value from their composition, and irons containing less than 16 ppm oxygen had forced chill depths averaging 1/32 in. greater than the expected chill depth. Higher oxygen contents, within the range of 12 to 68 pprn did not increase forced chill depth. With the wedge tests, there was a good linear relationship between carbon equivalent of the irons and their chill depth. The results indicated that oxygen contents below 50 ppm in the iron did not affect chill depth. With 50 to 70 ppm oxygen in the iron, oxygen appeared to have a slight graphitizing tendency. These results are in disagreement with the common belief in gray iron foundries that "oxidized irons" produce high chill depths. It would be appreciated if the author would comment on this subject. Gustaf Ostberg (author&apos;s reply)—In Fig. 1 the legend of line I should read 2 pct C, 1 pct Si. The author wishes to thank Mr. Loria for calling attention to his later work, which was published after the present paper was concluded. The range of oxygen contents quoted seems to agree well with the author&apos;s values. The lack of response to variations in silicon content is probably due to the fact that the oxygen content in most cases was below the saturation level. The absence of temperature dependence, even in the case of saturation, is understandable if the difficulty in formation and escape of the deoxidation products is taken into account.

Jan 1, 1960

By L. D. Mullins, E. B. Elfrink, W. L. Wahl

In the past few years several articles and papers presenting results of solution gas-drive depletion calculations have appeared in the lit-erature. Such calculations are of interest to the oil industry, for investment decisions must often be made before much is known about a reservoir. At other times, an estimate of the possible benefits to be realized from alternate production methods is desirable, and theoretical depletion calculations can serve as a floor or reference level from which to work. In any case, an estimate of ultimate oil recovery based upon engineering data is commonly required. An engineer confronted with the problem of obtaining, for a specific reservoir system, an estimate of ultimate oil recovery by solution gas-drive depletion usually will be forced to perform the calculations himself. This is despite the quantity of data in the literature. Rarely will either experience or the literature provide results from a reservoir system similar in all important respects to the one under consideration, and calculated results are not so plentiful that satisfactory interpolation procedures can be devised. Performing the calculations, however, is a tedious, time-consuming task unless an electronic computer is available, and, in practice, time and manpower are not always available for this purpose. A quick, simple, consistent method was needed for reducing the uncer- tainty in estimated oil recovery from solution gas-drive reservoirs when only minimum information about the reservoir system is available. PROCEDURE Method of Calculation The usual requisite assumptions were made so that the material balance equation could be used to calculate data for the charts. The following assumptions were made: (1) the reservoir is homogeneous and isotropic; (2) oil recovery is due entirely to solution gas drive and neither a gas cap nor a water drive nor gravity drainage is present; (3) the initial reservoir pressure is the bubble-point pressure of the reservoir fluid; (4) initial total liquid saturation is 100 per cent of pore space; (5) interstitial water saturation remains at the initial value as the reservoir pressure declines from the bubble-point pressure to atmospheric pressure; (6) equilibrium gas saturation is 5 per cent of pore space; and (7) oil and gas saturations are uniformly distributed throughout the reservoir at all times. There are no saturation gradients due to a wellbore, nor is the geometry of the reservoir system considered. The material balance equation was written in the form of a differential equation&apos; which was integrated to determine the change in oil saturation for an assigned pressure drop. Formal integration was not possible, so recourse was made to the Runge-Kutta method&apos; of numerical integration. All computations were performed on IBM equipment. Numerical integration yielded the change in oil saturation within the reservoir as the pressure. declined from the bubble-point pressure to atmospheric pressure. The initial oil saturation minus the change in oil saturation yielded the oil saturation at atmospheric pressure. The oil originally in place was obtained by dividing the initial oil saturation by the initial formation volume factor (differential liberation) while the oil in place at atmospheric pressure was obtained by dividing the final oil saturation by the formation volume factor at atmospheric pressure. U1timate oil recovery, expressed as a percentage of the initial oil in place, was obtained by dividing the difference between the oil initially in place and the oil in place at atmospheric pressure by the oil initially in place and multiplying by 100. PVT Data Charts were based upon 135 solutions to the material balance equation. PVT properties of the reservoir fluids were variables in this equation and had to be known as functions of pressure. The required PVT data might have been obtained from either actual reservoir fluid systems or correlated data. However, correlated PVT data were developed and employed in the calculations for the following reasons: (1) it is doubtful that 135 sets of PVT data could have been obtained for the values of variables investigated in this study, and (2) results of recovery calculations from randomly obtained PVT data could not be correlated as well as rssults from selected PVT data. The PVT data used to develop the