Search Documents

By Ivor Jenkins

Processes employing 'controlled at-mospheres in the heat-treatment of metals and alloys are now well established on an industrial scale, and the general principles involved and the advantages to be gained in the use of the technique have received considerable publicity over a period of years. During this time the greatest emphasis has been on the establishment of the process in industry, the main effort being devoted to development work on various types of atmospheres, their useful fields of application and the design of suitable heat-treatment furnaces in which they are to be used. The various types of controlled atmospheres are fairly well standardized, but there still remains a considerable field for the investigation of related problems, in connection with both the gas-generating equipment and the surface chemistry of the heat-treatment of metals, especially in complex controlled atmospheres. This paper describes certain investigations of this nature that were carried out in England up to 1939. The development of processes employing controlled atmospheres has been asso-ciated to a very large extent with the heat-treatment of steels, both because of the major importance of steel in industry and the many diverse and dificult prob-lems of surface chemistry involved when steels of such a wide range of composition are heated in the presence of active gases. Controlled atmospheres such as disso-ciated and burnt ammonia or charcoal gas are used in the heat-treatment of low-carbon and high-carbon steels, but from certain points of view a suitable atmosphere generated by the partial combustion of city gas is to be preferred. Both the ammonia derivatives are comparatively expensive, and even if the burnt ammonia is regenerated, thereby reducing operating costs, the capital expenditure of the gas plant may make it prohibitive for the smaller type of furnace installation. Charcoal-gas generators are not without their operating difficulties especially those associated with the quality of the charcoal, clinker formation and the removal of carbon dust from the generated gas. City gas has the advantage of being comparatively cheap and "on tap" in most if not all industrial centers. The use of raw city gas as a protective atmosphere in the heat-treatment of steels has been described,' but it presents certain difficulties because of its complex composition and its explosive nature when mixed with air. These factors, together with the tendency toward sooting at elevated temperatures, arising mainly from the hydrocarbon gases present, all militate against its use as a protective atmosphere in industrial furnaces. The following analysis, based on the analyses of city gas in different localities in Great Britain, indicates the wide limits existing

Jan 1, 1948

By Ivor Jenkins

Processes employing 'controlled at-mospheres in the heat-treatment of metals and alloys are now well established on an industrial scale, and the general principles involved and the advantages to be gained in the use of the technique have received considerable publicity over a period of years. During this time the greatest emphasis has been on the establishment of the process in industry, the main effort being devoted to development work on various types of atmospheres, their useful fields of application and the design of suitable heat-treatment furnaces in which they are to be used. The various types of controlled atmospheres are fairly well standardized, but there still remains a considerable field for the investigation of related problems, in connection with both the gas-generating equipment and the surface chemistry of the heat-treatment of metals, especially in complex controlled atmospheres. This paper describes certain investigations of this nature that were carried out in England up to 1939. The development of processes employing controlled atmospheres has been asso-ciated to a very large extent with the heat-treatment of steels, both because of the major importance of steel in industry and the many diverse and dificult prob-lems of surface chemistry involved when steels of such a wide range of composition are heated in the presence of active gases. Controlled atmospheres such as disso-ciated and burnt ammonia or charcoal gas are used in the heat-treatment of low-carbon and high-carbon steels, but from certain points of view a suitable atmosphere generated by the partial combustion of city gas is to be preferred. Both the ammonia derivatives are comparatively expensive, and even if the burnt ammonia is regenerated, thereby reducing operating costs, the capital expenditure of the gas plant may make it prohibitive for the smaller type of furnace installation. Charcoal-gas generators are not without their operating difficulties especially those associated with the quality of the charcoal, clinker formation and the removal of carbon dust from the generated gas. City gas has the advantage of being comparatively cheap and "on tap" in most if not all industrial centers. The use of raw city gas as a protective atmosphere in the heat-treatment of steels has been described,' but it presents certain difficulties because of its complex composition and its explosive nature when mixed with air. These factors, together with the tendency toward sooting at elevated temperatures, arising mainly from the hydrocarbon gases present, all militate against its use as a protective atmosphere in industrial furnaces. The following analysis, based on the analyses of city gas in different localities in Great Britain, indicates the wide limits existing

Jan 1, 1948

By A. Hultgren, B. Herrlander

The original aim of the present investigation was to study the mechanism of cracking on hot-deforming red-short steels. During the microscopical examination of hot-deformed soft steels attention was directed to various patterns of so-called veining in the ferrite, as related to variations in the deformation procedure and the heat-treatment of the steels. Later, a hot-short steel of higher carbon content was also studied. Historical Storey1 suggested in 1914 that veining developed during the gamma-alpha transformation from the growing together of several alpha nuclei within the same gamma grain. Rawdon and Berglund2 found that in soft steel, forged somewhat below A3, the ferrite showed profuse veining, while usually there was very little veining after forging at very high temperatures. They emphasized the similarity between the pattern of veining formed by deformation about 600°C and that of slip lines formed by deformation at room temperature. They also suggested a relation between the gamma-alpha transformation and veining. They became convinced from etching and from the slip-line pattern in cold-deformed material, that the orientation within any ferrite grain showing veining was uniform. In recrystallized alpha grains veining was absent. Veining did not appear materially to affect the properties of ferrite. Independent of veining, continuous networks, attributed to preexisting delta and gamma grain boundaries were sometimes found, the former only in cast steel. Those networks were usually connected with small inclusions. Ammermann and Kornfeld3 confirmed that recrystallized ferrite grains showed no veining. In soft steel annealed between A1 and A3 there was veining only in the ferrite grains formed during cooling. Electrolytic iron deformed at 880' was free from veining. Bannister and Jones' considered veining in ferrite to be a manifestation of microscopical and sub-microscopica~ inclusions. NorthCOtt,5,6 in studies on veining in steel and other metals and alloys, attributed veining to oxide inclusions. It could generally be removed by annealing in hydrogen, a suggestion earlier made by Rawdon and Berg1und. TrittOn (discussion in ref. 5) emphasized the importance of perfect polishing for bringing out veining by etching, and thought the veins or subboundaries were produced during the gamma-alpha transformation, when it was rapid, as a result of the volume change and contraction during cooling. Slight distortion would not allow projecting parts of a growing crystal to join up with atomic symmetry. Hanemann and co-workers showed that veining in ferrite consisted of ridges

Jan 1, 1948

By B. Herrlander, A. Hultgren

The original aim of the present investigation was to study the mechanism of cracking on hot-deforming red-short steels. During the microscopical examination of hot-deformed soft steels attention was directed to various patterns of so-called veining in the ferrite, as related to variations in the deformation procedure and the heat-treatment of the steels. Later, a hot-short steel of higher carbon content was also studied. Historical Storey1 suggested in 1914 that veining developed during the gamma-alpha transformation from the growing together of several alpha nuclei within the same gamma grain. Rawdon and Berglund2 found that in soft steel, forged somewhat below A3, the ferrite showed profuse veining, while usually there was very little veining after forging at very high temperatures. They emphasized the similarity between the pattern of veining formed by deformation about 600°C and that of slip lines formed by deformation at room temperature. They also suggested a relation between the gamma-alpha transformation and veining. They became convinced from etching and from the slip-line pattern in cold-deformed material, that the orientation within any ferrite grain showing veining was uniform. In recrystallized alpha grains veining was absent. Veining did not appear materially to affect the properties of ferrite. Independent of veining, continuous networks, attributed to preexisting delta and gamma grain boundaries were sometimes found, the former only in cast steel. Those networks were usually connected with small inclusions. Ammermann and Kornfeld3 confirmed that recrystallized ferrite grains showed no veining. In soft steel annealed between A1 and A3 there was veining only in the ferrite grains formed during cooling. Electrolytic iron deformed at 880' was free from veining. Bannister and Jones' considered veining in ferrite to be a manifestation of microscopical and sub-microscopica~ inclusions. NorthCOtt,5,6 in studies on veining in steel and other metals and alloys, attributed veining to oxide inclusions. It could generally be removed by annealing in hydrogen, a suggestion earlier made by Rawdon and Berg1und. TrittOn (discussion in ref. 5) emphasized the importance of perfect polishing for bringing out veining by etching, and thought the veins or subboundaries were produced during the gamma-alpha transformation, when it was rapid, as a result of the volume change and contraction during cooling. Slight distortion would not allow projecting parts of a growing crystal to join up with atomic symmetry. Hanemann and co-workers showed that veining in ferrite consisted of ridges

Jan 1, 1948

By F. E. Harris

It has been said that carbon is "ubiquitous" with reference to iron alloys. Certainly at temperatures where carbon and iron form the solid solution, austenite, it may be readily added to, or removed from, steel surfaces. Differences in concentration potential are the sole requirements for carbon flow. To measure the rate of carbon flow, a proportionality factor, D, is arbitrarily expressed in units of (length)2 X (time)-'. If the term used for carbon concentration is weight Pet C, as is customary, then a factor must be found, changing this term to One of weight Per unit volume when the diffusion coeficient, D, is used in flow formulae. The determination of D values for carbon is complicated since D varies with the concentration itself. The Matano method has often been employed for systems involving a variant D. This method expresses all evaluation of unsteady flow whereby D is given a relationship with the numerical value of C-x slopes. The author contends that this solution is incorrect, except when D is invariant. The primary purpose of this paper is to support this contention. Here diffusion is controlled under two conditions of flow. The first part deals with the steady state; the second part considers unsteady flow under definite restrictions. Beginning with the fundamental concept, the development is continued until equations for C-x curves are obtained. The solution is then subjected to experimental data. It is realized that a great amount of reliable data is required for a rigorous proof. KO such research is claimed. Instead an attempt is made to set forth the physical significance of the diffusion mechanism, and particularly to express the factors which influence the flow of carbon in austenite. Of equal significance is the implication that gaseous media may be employed with reasonable accuracy in making these determinations. The Steady State Diffusion of Carbon in Au~tenite The fundamental experiment defining diffusion in solid metals may be readily performed for carbon in gamma iron.' For this experiment a thin plate of medium carbon steel is chosen. The surfaces of this plate may be taken as two parallel planes which may be considered infinite for points well in the Center of the planes. These two planes are held at different concentrations of carbon at a constant temperature. This temperature, 1700°F for example, is one where the single phase: carbon in solution with gamma iron, exists. When these conditions are held for a sufficient time the concentration of carbon of the different points of the solid ment down towards its steady state value. At points well removed from the ends, the concentration will remain the Same along planes parallel to the surfaces of the plate.

Jan 1, 1948

By F. E. Harris

It has been said that carbon is "ubiquitous" with reference to iron alloys. Certainly at temperatures where carbon and iron form the solid solution, austenite, it may be readily added to, or removed from, steel surfaces. Differences in concentration potential are the sole requirements for carbon flow. To measure the rate of carbon flow, a proportionality factor, D, is arbitrarily expressed in units of (length)2 X (time)-'. If the term used for carbon concentration is weight Pet C, as is customary, then a factor must be found, changing this term to One of weight Per unit volume when the diffusion coeficient, D, is used in flow formulae. The determination of D values for carbon is complicated since D varies with the concentration itself. The Matano method has often been employed for systems involving a variant D. This method expresses all evaluation of unsteady flow whereby D is given a relationship with the numerical value of C-x slopes. The author contends that this solution is incorrect, except when D is invariant. The primary purpose of this paper is to support this contention. Here diffusion is controlled under two conditions of flow. The first part deals with the steady state; the second part considers unsteady flow under definite restrictions. Beginning with the fundamental concept, the development is continued until equations for C-x curves are obtained. The solution is then subjected to experimental data. It is realized that a great amount of reliable data is required for a rigorous proof. KO such research is claimed. Instead an attempt is made to set forth the physical significance of the diffusion mechanism, and particularly to express the factors which influence the flow of carbon in austenite. Of equal significance is the implication that gaseous media may be employed with reasonable accuracy in making these determinations. The Steady State Diffusion of Carbon in Au~tenite The fundamental experiment defining diffusion in solid metals may be readily performed for carbon in gamma iron.' For this experiment a thin plate of medium carbon steel is chosen. The surfaces of this plate may be taken as two parallel planes which may be considered infinite for points well in the Center of the planes. These two planes are held at different concentrations of carbon at a constant temperature. This temperature, 1700°F for example, is one where the single phase: carbon in solution with gamma iron, exists. When these conditions are held for a sufficient time the concentration of carbon of the different points of the solid ment down towards its steady state value. At points well removed from the ends, the concentration will remain the Same along planes parallel to the surfaces of the plate.

Jan 1, 1948

By M. F. Hawkes, R. F. Mehl

The rate of isothermal transformation of austenite to pearlite depends upon the rate of nucleation, N, and the rate of growth, G, of pearlite in austenite. Values of N are given in terms of the number of nuclei forming in unit time per unit grain boundary area of unreacted austenite; values of C are given in terms of the linear rate of radial growth. Variations in the isothermal rate with temperature result from variation of N and G with temperature. Depth of hardening of a steel depends fundamentally upon values of N and G; changes in carbon and alloy content effect changes in N and G and thus affect depth of hardening. Quantitative studies are available on N and G for the formation of pearlite in plain carbon eutectoid steels.3 Systematic studies in this field are necessary if the phenomena are to be well enough known to furnish a proper basis for full rationalization and for theory. Accordingly, the studies are continuing, both with respect to the effect of alloying elements upon the values of N and G in the formation of pearlite and to the values of N and G for the formation of ferrite in hypoeutectoid steels. The present paper is the first relating to the effect of alloying elements upon the values of N and G for pearlite. It offers special interest since cobalt is known to have an anomalous effect upon the rate of formation of pearlite, accelerating it, whereas all other elements retard it.4 It will be shown that cobalt increases both N and G for pearlite and that this effect is inherent in the Fe-Co-C system and not dependent upon factors relating to austenite heterogeneity nor upon any recognizable adventitious factor. Inasmuch as the rate of diffusion of carbon in austenite and the interlamellar spacing of pearlite are variables related to G and possibly to N (in a manner not yet certain), measurements of the effect of cobalt upon them are also reported here. In the course of the work, measurements were made on the effect of cobalt upon the marten-site temperature and these are given. COBALT AS AN ALLOYING ELEMENT IN STEEL Properties of Cobalt Cobalt resembles iron more closely than does any other chemical element. Its atomic number is 27, hence it is the element immediately following iron in the periodic system. The two differ in atomic structure only in the fact that cobalt contains seven electrons instead of six in the third, or transition, level of .the third shell. The next shell, which is the

Jan 1, 1948

By M. F. Hawkes, R. F. Mehl

The rate of isothermal transformation of austenite to pearlite depends upon the rate of nucleation, N, and the rate of growth, G, of pearlite in austenite. Values of N are given in terms of the number of nuclei forming in unit time per unit grain boundary area of unreacted austenite; values of C are given in terms of the linear rate of radial growth. Variations in the isothermal rate with temperature result from variation of N and G with temperature. Depth of hardening of a steel depends fundamentally upon values of N and G; changes in carbon and alloy content effect changes in N and G and thus affect depth of hardening. Quantitative studies are available on N and G for the formation of pearlite in plain carbon eutectoid steels.3 Systematic studies in this field are necessary if the phenomena are to be well enough known to furnish a proper basis for full rationalization and for theory. Accordingly, the studies are continuing, both with respect to the effect of alloying elements upon the values of N and G in the formation of pearlite and to the values of N and G for the formation of ferrite in hypoeutectoid steels. The present paper is the first relating to the effect of alloying elements upon the values of N and G for pearlite. It offers special interest since cobalt is known to have an anomalous effect upon the rate of formation of pearlite, accelerating it, whereas all other elements retard it.4 It will be shown that cobalt increases both N and G for pearlite and that this effect is inherent in the Fe-Co-C system and not dependent upon factors relating to austenite heterogeneity nor upon any recognizable adventitious factor. Inasmuch as the rate of diffusion of carbon in austenite and the interlamellar spacing of pearlite are variables related to G and possibly to N (in a manner not yet certain), measurements of the effect of cobalt upon them are also reported here. In the course of the work, measurements were made on the effect of cobalt upon the marten-site temperature and these are given. COBALT AS AN ALLOYING ELEMENT IN STEEL Properties of Cobalt Cobalt resembles iron more closely than does any other chemical element. Its atomic number is 27, hence it is the element immediately following iron in the periodic system. The two differ in atomic structure only in the fact that cobalt contains seven electrons instead of six in the third, or transition, level of .the third shell. The next shell, which is the

Jan 1, 1948

By K. A. Grange, W. S. Holt, E. T. Tkac

Quantitative knowledge of the time clement involved in austenite transformation in a particular steel provides a sound basis for understanding and planning heat-treatment. Such knowledge is conveniently obtained by measurement of austenite transformation as it occurs at each of a series of temperature levels. The results of these measurements usually are summarized in the form of the now familiar isothermal transformation diagram (I-T diagram, TTT diagram, or S-curve). Although a considerable number of such diagrams already have appeared in the literature, there remain many compositions of commercial importance for which no I-T diagram is available. The diagram for many such steels may be predicted from published data with sufficient accuracy for most practical purposes, but for others no satisfactory prediction can be made because the effect on austenite transformation of one or more of the important alloying elements present is not well enough known. Such a steel is Nitralloy, Type 13s-Modified, which, containing about 1.25 per cent aluminum, differs from any whose I-T diagram has been published. This paper presents results of a study of the isothermal transformation behavior of a sample from a commercial heat of this aluminum-chromium-molybdenum steel, together with pertinent supplementary data including the equilibrium transformation temperatures (Aeaand Ae,), the temperature range of martensite formation, and the end-quench harden-ability. These data are directly applicable to the hcat-treatment of this type of steel and are of interest in revealing some of the fundamental effects of aluminum as an alloying element in steel. Material All specimens were prepared from a 1 1/4 i-in. diam bar forged from a 5 by 5-in. billet from a commercial heat made in an electric arc furnace. The composition of the bar was; C, 0.41 per cent: AIn, 0.57: P, 0.016: S, 0.005: Si, 0.24: Ni, 0.17: Cr, I.j7: Mo, 0.36: All 1.26. The forged bar was normalized from r750°F (95s°C) and then tempered for one hour at rzoo°F (650°C), this tempering treatment having been given to facilitate the machining of specimens. Equilibrium Transformation Temperatures In hypoeutectoid steel the minimum temperature at which austenite is completely stable, Aep, and the maximum temperature at which austenite is completely unstable, Ael, are of great significance in heat-treatment. Since these temperatures are dependent upon com-

Jan 1, 1948

By W. S. Holt, K. A. Grange, E. T. Tkac

Quantitative knowledge of the time clement involved in austenite transformation in a particular steel provides a sound basis for understanding and planning heat-treatment. Such knowledge is conveniently obtained by measurement of austenite transformation as it occurs at each of a series of temperature levels. The results of these measurements usually are summarized in the form of the now familiar isothermal transformation diagram (I-T diagram, TTT diagram, or S-curve). Although a considerable number of such diagrams already have appeared in the literature, there remain many compositions of commercial importance for which no I-T diagram is available. The diagram for many such steels may be predicted from published data with sufficient accuracy for most practical purposes, but for others no satisfactory prediction can be made because the effect on austenite transformation of one or more of the important alloying elements present is not well enough known. Such a steel is Nitralloy, Type 13s-Modified, which, containing about 1.25 per cent aluminum, differs from any whose I-T diagram has been published. This paper presents results of a study of the isothermal transformation behavior of a sample from a commercial heat of this aluminum-chromium-molybdenum steel, together with pertinent supplementary data including the equilibrium transformation temperatures (Aeaand Ae,), the temperature range of martensite formation, and the end-quench harden-ability. These data are directly applicable to the hcat-treatment of this type of steel and are of interest in revealing some of the fundamental effects of aluminum as an alloying element in steel. Material All specimens were prepared from a 1 1/4 i-in. diam bar forged from a 5 by 5-in. billet from a commercial heat made in an electric arc furnace. The composition of the bar was; C, 0.41 per cent: AIn, 0.57: P, 0.016: S, 0.005: Si, 0.24: Ni, 0.17: Cr, I.j7: Mo, 0.36: All 1.26. The forged bar was normalized from r750°F (95s°C) and then tempered for one hour at rzoo°F (650°C), this tempering treatment having been given to facilitate the machining of specimens. Equilibrium Transformation Temperatures In hypoeutectoid steel the minimum temperature at which austenite is completely stable, Aep, and the maximum temperature at which austenite is completely unstable, Ael, are of great significance in heat-treatment. Since these temperatures are dependent upon com-

Jan 1, 1948

By Gerald Edmunds

The liquidus line on the phase diagram for temperature versus composition of a binary alloy system, representing the boundary between the homogeneous melt and the heterogeneous melt plus solid, besides showing the minimum casting temperature, is useful in determining eutectic compositions and the presence of phase changes such as peritectic transformations and intermetallic compound formation. The liquidus often is determined by the cooling curves of thermal analysis, but supercooling frequently occurs, so that the temperature values are low. An alternative method, the one employed here, consists of determining the composition of samples of the melts from heterogeneous alloys that are at equilibrium. With meticulous attention to fundamentals, this method is capable of yielding values limited in precision only by the temperature control and measurement and analyses. Apparatus By starting, on the one hand, with a supersaturated solution and allowing the excess solute to precipitate, and on the other hand with an undersaturated solution and allowing more of the solute to be dissolved, both processes being carried out at the same temperature, equilibrium may be said to have been approached from both directions; that is, from above and from below. If the analyses show the solutions to contain the same fraction of solute when equilibrium has been approached from above and from below, one is assured that equilibrium has been reached in both determinations. In order to carry out such determinations, it is necessary to have equipment for obtaining and maintaining the proper temperatures; for holding the melt, measuring its temperature, and stirring, sampling and analyzing it. The present work required temperatures from 420° to 900°C. The equipment consisted of: A furnace, which was a vertical alundum tube, 13/4-in. diameter inside and 18 in. long, wound with Nichrome resistance wire, surrounded by a 3-in. i.d. alundum tube wound for 5 in. of length near each end, for compensation of end losses. This muffle was contained in an iron shell with Transite end plates, the intervening space being packed with Sil-O-Cel. The outside dimensions of the furnace were 12 in. in diameter by 181/2 in. in height. Heating current was supplied by three independent variable-voltage transformers (Variac), one connected to each end compensator winding and manually controlled, and one connected to the main winding through a series resistance of 5 ohms. This resistance was short-circuited for

Jan 1, 1944

By Gerald Edmunds

The liquidus line on the phase diagram for temperature versus composition of a binary alloy system, representing the boundary between the homogeneous melt and the heterogeneous melt plus solid, besides showing the minimum casting temperature, is useful in determining eutectic compositions and the presence of phase changes such as peritectic transformations and intermetallic compound formation. The liquidus often is determined by the cooling curves of thermal analysis, but supercooling frequently occurs, so that the temperature values are low. An alternative method, the one employed here, consists of determining the composition of samples of the melts from heterogeneous alloys that are at equilibrium. With meticulous attention to fundamentals, this method is capable of yielding values limited in precision only by the temperature control and measurement and analyses. Apparatus By starting, on the one hand, with a supersaturated solution and allowing the excess solute to precipitate, and on the other hand with an undersaturated solution and allowing more of the solute to be dissolved, both processes being carried out at the same temperature, equilibrium may be said to have been approached from both directions; that is, from above and from below. If the analyses show the solutions to contain the same fraction of solute when equilibrium has been approached from above and from below, one is assured that equilibrium has been reached in both determinations. In order to carry out such determinations, it is necessary to have equipment for obtaining and maintaining the proper temperatures; for holding the melt, measuring its temperature, and stirring, sampling and analyzing it. The present work required temperatures from 420° to 900°C. The equipment consisted of: A furnace, which was a vertical alundum tube, 13/4-in. diameter inside and 18 in. long, wound with Nichrome resistance wire, surrounded by a 3-in. i.d. alundum tube wound for 5 in. of length near each end, for compensation of end losses. This muffle was contained in an iron shell with Transite end plates, the intervening space being packed with Sil-O-Cel. The outside dimensions of the furnace were 12 in. in diameter by 181/2 in. in height. Heating current was supplied by three independent variable-voltage transformers (Variac), one connected to each end compensator winding and manually controlled, and one connected to the main winding through a series resistance of 5 ohms. This resistance was short-circuited for

Jan 1, 1944

By E. A. Anderson, P. W. Ramsey, E. J. Boyle

Only a few zinc-rich alloys are in commercial use today, and these depend largely upon solid solution effects for their desirable properties. Typical examples are rolled alloys containing about 1 Per cent of copper, with Or without additions of elements such as magnesium, and die-casting alloys containing about 4 Per cent of aluminum with small amounts of magnesium and sometimes copper." A number of elements are capable of forming alloys with zinc, Which do not enter into solid solution except in limited degree. The Present Paper considers One such system, zinc-titanium, in which alloys having potentially useful properties in the rolled form were found. Zinc-titanium System The zinc-rich portion of the zinc-titanium diagram, as determined by standard methods in this laboratory,3 is shown in Fig. I. The portion of the diagram of significance in the present work extends from 0 to about 0.5 per cent titanium. Gebhardt4 places the eutectic at 0.45 per cent, but the evidence in favor 01 the composition given in Fig. I is considerable. The solid solubility of titanium in zinc is very low indeed and possibly places these alloys in a somewhat different category from the commercial alloys previously mentioned. Alloy Preparation All of the alloys used in this work were prepared with titanium metal of 98.9 per cent purity. In most cases high-purity zinc (99.99 per cent) was used but sufficient work was done later to show that zinc of about high-grade composition would yield approximately equal results. Best practice in alloy manufacture seemed to be to prepare a zinc-titanium hardener by adding about 4 per cent of titanium metal to molten zinc and holding under a boric acid cover, with frequent stirrings, at 750°C, until the titanium had dissolved. The desired titanium content could readily be obtained in the final alloy by adding the calculated amount 01 this hardener. The alloys proved to be quite stable in composition, unless overheated, and, with reasonable temperature control, required no cover or flux to prevent loss of titanium content by oxidatioll, Hot-rollig Tests Three alloys containing, by analysis, 0.05, 0.12, and 0.23 per cent titanium were selected for preliminary study. Slab cross sections a a marked grain refinement* (Fig. 2) in the two alloys of highest titanium content. Cast unalloyed zinc has a columnar grain structure similar to that shown for the alloy of lowest titanium content but without the scatter of small equiaxed grains at the center.

Jan 1, 1944

By P. W. Ramsey, E. J. Boyle, E. A. Anderson

Only a few zinc-rich alloys are in commercial use today, and these depend largely upon solid solution effects for their desirable properties. Typical examples are rolled alloys containing about 1 Per cent of copper, with Or without additions of elements such as magnesium, and die-casting alloys containing about 4 Per cent of aluminum with small amounts of magnesium and sometimes copper." A number of elements are capable of forming alloys with zinc, Which do not enter into solid solution except in limited degree. The Present Paper considers One such system, zinc-titanium, in which alloys having potentially useful properties in the rolled form were found. Zinc-titanium System The zinc-rich portion of the zinc-titanium diagram, as determined by standard methods in this laboratory,3 is shown in Fig. I. The portion of the diagram of significance in the present work extends from 0 to about 0.5 per cent titanium. Gebhardt4 places the eutectic at 0.45 per cent, but the evidence in favor 01 the composition given in Fig. I is considerable. The solid solubility of titanium in zinc is very low indeed and possibly places these alloys in a somewhat different category from the commercial alloys previously mentioned. Alloy Preparation All of the alloys used in this work were prepared with titanium metal of 98.9 per cent purity. In most cases high-purity zinc (99.99 per cent) was used but sufficient work was done later to show that zinc of about high-grade composition would yield approximately equal results. Best practice in alloy manufacture seemed to be to prepare a zinc-titanium hardener by adding about 4 per cent of titanium metal to molten zinc and holding under a boric acid cover, with frequent stirrings, at 750°C, until the titanium had dissolved. The desired titanium content could readily be obtained in the final alloy by adding the calculated amount 01 this hardener. The alloys proved to be quite stable in composition, unless overheated, and, with reasonable temperature control, required no cover or flux to prevent loss of titanium content by oxidatioll, Hot-rollig Tests Three alloys containing, by analysis, 0.05, 0.12, and 0.23 per cent titanium were selected for preliminary study. Slab cross sections a a marked grain refinement* (Fig. 2) in the two alloys of highest titanium content. Cast unalloyed zinc has a columnar grain structure similar to that shown for the alloy of lowest titanium content but without the scatter of small equiaxed grains at the center.

Jan 1, 1944

By B. Blumenthal

THE present investigation was planned as a survey of the lead-rich portion of the ternary lead-antimony-silver system by thermal analysis. Since, however, a rigorous application of the conventional thermal analytical method is time consuming, an attempt was made to speed up the investigation by reducing the time required for a single thermal analysis. This was done by increasing the rate of cooling to 12° to 18°C. per minute, since such a rate does not permit even an approach to equilibrium conditions, the melt was stirred vigorously by a stirrer operating at the high rate of 665 r.p.m. This proved to be effective in obtaining cooling curves with distinct points of arrest and of excellent reproducibility, which satisfactorily compare with those obtained at a rate of 1" to 2°C. per min. with or without slow stirring (Fig. I). By this method the lead-rich end of the lead-alltimony system and a series of ternary alloys in sections parallel to the lead-antimony side of the ternary system were investigated and it was observed that the two branches of the liquidus curve of the binary alloys did not intersect at the euteotic temperature. An equivalent effect mas observed in the ternary system. The eutectic temperature of the binary lead-antimony system was found to be 252.0°C., which confirms the recent in- vestigation by Pellini and Rhines,1 who determined the eutectic temperature to be 2 51.5°C. The observation that the curves of primarY crystallization of the lead-antimony system do not intersect at the eutectic temperature was previously made by Dean.2 Dean explains the phenomeilon by assuming the existence of a compound Sb4pb. He felt that supercooling was not the cause of the phenomena because he was unable to influence them by inoculation. The X-ray investigation by Solomon and Morris-Jones,3 the metallographic investigations by Stead,4 Charpy,5 Gontermann,6 and Dean, Zickrick and Nix,7 and the numerous investigations of physical properties, definitely disprove the occurrence of any compound. The work of Dean was not confirmed by Broniewski and Sliwoski,8 whose experiments did not show any irregularity and who report a eutectic composition of 13 per cent antimony and a eutectic temperature Of 2500 to 252°C. at a rate Of coo1ing of about 6°C. Per min. Abell Redlich and Adler9 then report a eutectic composition of 12.1 per cent upon reviewing the previous publications and adding a few of their own cooling curves. In 1934 Quadrat and Jiriste10 found from careful analyses of the bottom portion of slowly cooled hypereutectic melts that the eutectic only 11 .4 to 11.5 per cent antimany. (Previous analyses by Stead4 .in 1891 gave 13 per cent.) Weaver,ll however, reported in 1935 that cooling cunres indi-

Jan 1, 1944

By B. Blumenthal

THE present investigation was planned as a survey of the lead-rich portion of the ternary lead-antimony-silver system by thermal analysis. Since, however, a rigorous application of the conventional thermal analytical method is time consuming, an attempt was made to speed up the investigation by reducing the time required for a single thermal analysis. This was done by increasing the rate of cooling to 12° to 18°C. per minute, since such a rate does not permit even an approach to equilibrium conditions, the melt was stirred vigorously by a stirrer operating at the high rate of 665 r.p.m. This proved to be effective in obtaining cooling curves with distinct points of arrest and of excellent reproducibility, which satisfactorily compare with those obtained at a rate of 1" to 2°C. per min. with or without slow stirring (Fig. I). By this method the lead-rich end of the lead-alltimony system and a series of ternary alloys in sections parallel to the lead-antimony side of the ternary system were investigated and it was observed that the two branches of the liquidus curve of the binary alloys did not intersect at the euteotic temperature. An equivalent effect mas observed in the ternary system. The eutectic temperature of the binary lead-antimony system was found to be 252.0°C., which confirms the recent in- vestigation by Pellini and Rhines,1 who determined the eutectic temperature to be 2 51.5°C. The observation that the curves of primarY crystallization of the lead-antimony system do not intersect at the eutectic temperature was previously made by Dean.2 Dean explains the phenomeilon by assuming the existence of a compound Sb4pb. He felt that supercooling was not the cause of the phenomena because he was unable to influence them by inoculation. The X-ray investigation by Solomon and Morris-Jones,3 the metallographic investigations by Stead,4 Charpy,5 Gontermann,6 and Dean, Zickrick and Nix,7 and the numerous investigations of physical properties, definitely disprove the occurrence of any compound. The work of Dean was not confirmed by Broniewski and Sliwoski,8 whose experiments did not show any irregularity and who report a eutectic composition of 13 per cent antimony and a eutectic temperature Of 2500 to 252°C. at a rate Of coo1ing of about 6°C. Per min. Abell Redlich and Adler9 then report a eutectic composition of 12.1 per cent upon reviewing the previous publications and adding a few of their own cooling curves. In 1934 Quadrat and Jiriste10 found from careful analyses of the bottom portion of slowly cooled hypereutectic melts that the eutectic only 11 .4 to 11.5 per cent antimany. (Previous analyses by Stead4 .in 1891 gave 13 per cent.) Weaver,ll however, reported in 1935 that cooling cunres indi-

Jan 1, 1944

By R. S. Dean

The commercial production of electrolytic manganese on a small scale commenced in 1939. The writer made a short report on the progress of production and utilization in Mining and Metallurgy for January 1941. Progress during the last two years, naturally, has been more rapid. In June 1940, Congressman (now Senator) Scrugham presented a proposal to Congress for an appropriation authorizing erection of a pilot plant for " the production of metallic manganese by electrolytic or other means." As a result of this, a pilot plant, having a capacity of a few hundred pounds per day was built at Boulder City, Nev. Subsequent appropriations permitted expansion of the plant to about a ton per day. Much has been learned in this plant about electrolytic manganese, and the product even in its small way has found war uses of considerable importance. Meanwhile, the Knoxville plant of the Electro Manganese Corporation, the sole commercial producer, was expanded to about 4 tons per day to supply needed material for the war program. Other proposals for commercial plants have not been favorably received by the War Production Board. The Bureau reported on one such proposal by the American Alloys and Chemical Corporation. Practice at BouldeR City The present practice, as carried out at Boulder City, which is being described in more detail in a current paper by the Bureau staff at Boulder City, starts with a manganese dioxide ore containing about 20 per cent manganese. Currently this ore comes from the Three Kids deposit near Las Vegas. The steps for the production of electrolytic manganese, in general, are as follows: 1. The manganese in the ore is reduced to MnO. 2. The MnO is dissolved from the ore in spent electrolyte, which contains about 38 to 47 grams per liter of free sulphuric acid, 135 grams per liter of ammonium sulphate, and 10 to 12 grams per liter of Mn as sulphate. In the leaching step the manganese is built back up to 32 to 36 grams per liter of Mn as sulphate; the pH is adjusted to neutral by means of gaseous ammonia and passed through thickeners. 3. The electrolyte is purified by adding H2S, which precipitates the heavy metals. After filtering, ferrous sulphate is added to the solution and oxidized with air and the solution is filtered and clarified on a precoat filter. 4. The purified catholyte is electrolyzed in a diaphragm cell using stainless-steel cathodes and lead-silver anodes. The current density is about 45 amp. per sq. ft. Current efficiencies of 60 to 65 per cent are regularly obtained.

Jan 1, 1944

By R. S. Dean

The commercial production of electrolytic manganese on a small scale commenced in 1939. The writer made a short report on the progress of production and utilization in Mining and Metallurgy for January 1941. Progress during the last two years, naturally, has been more rapid. In June 1940, Congressman (now Senator) Scrugham presented a proposal to Congress for an appropriation authorizing erection of a pilot plant for " the production of metallic manganese by electrolytic or other means." As a result of this, a pilot plant, having a capacity of a few hundred pounds per day was built at Boulder City, Nev. Subsequent appropriations permitted expansion of the plant to about a ton per day. Much has been learned in this plant about electrolytic manganese, and the product even in its small way has found war uses of considerable importance. Meanwhile, the Knoxville plant of the Electro Manganese Corporation, the sole commercial producer, was expanded to about 4 tons per day to supply needed material for the war program. Other proposals for commercial plants have not been favorably received by the War Production Board. The Bureau reported on one such proposal by the American Alloys and Chemical Corporation. Practice at BouldeR City The present practice, as carried out at Boulder City, which is being described in more detail in a current paper by the Bureau staff at Boulder City, starts with a manganese dioxide ore containing about 20 per cent manganese. Currently this ore comes from the Three Kids deposit near Las Vegas. The steps for the production of electrolytic manganese, in general, are as follows: 1. The manganese in the ore is reduced to MnO. 2. The MnO is dissolved from the ore in spent electrolyte, which contains about 38 to 47 grams per liter of free sulphuric acid, 135 grams per liter of ammonium sulphate, and 10 to 12 grams per liter of Mn as sulphate. In the leaching step the manganese is built back up to 32 to 36 grams per liter of Mn as sulphate; the pH is adjusted to neutral by means of gaseous ammonia and passed through thickeners. 3. The electrolyte is purified by adding H2S, which precipitates the heavy metals. After filtering, ferrous sulphate is added to the solution and oxidized with air and the solution is filtered and clarified on a precoat filter. 4. The purified catholyte is electrolyzed in a diaphragm cell using stainless-steel cathodes and lead-silver anodes. The current density is about 45 amp. per sq. ft. Current efficiencies of 60 to 65 per cent are regularly obtained.

Jan 1, 1944

By F. Keller, J. D. Edwards

The anodic treatment of aluminum presents problems of scientific as well as of commercial interest.l-3 Of particular interest is the fact that, during the anodic oxidation process, the oxide continues to form at the metal-oxide interface under any oxide previously formed. This has led to speculation as to the mechanism involved in the formation of the relatively thick oxide coatings that are used commercially for decorative or protective purposes. Furthermore, the ability of certain types of oxide coatings to adsorb dyes and other substances has stimulated research to determine the actual structure of these adsorptive coatings. It has been found that anodic oxide coatings on aluminum are composed essentially of aluminum oxide. They are formed by the action of oxygen ions penetrating to the metal surface during the electrolytic oxidation treatment. These coatings can be formed in a number of different electrolytes, as for example, those which contain sulphuric acid, chromic acid, oxalic acid, or boric acid. These are the commonest and most useful electrolytes employed commercially. Except for certain specific conditions of formation, the coatings in general have been found to be amorphous alumina, as far as can be determined by X-ray or electron-diffraction methods. Anodic oxidation processes can be arranged in three rather general classes if they are grouped in relation to the solvent action of the electrolyte on the coating.4 In the first class, the electrolyte has little or no solvent action on the coating that is formed. In general, coatings produced under such circumstances are nonporous and non-adsorptive. In the second class, the electrolyte exerts an appreciable solvent action on the coating. These coatings are porous and adsorptive. Finally, for the third class, the electrolyte tends to dissolve the coating about as rapidly as it is formed. This action produces electrolytic brightening or anodic polishing of the aluminum surface and at most leaves only a very thin film of oxide. The nonporous and nonadsorptive type of oxide film is represented by the coatings that are formed when solutions of boric acid are employed as the electrolyte. It is especially significant that these impervious films are formed in electrolytes that exert little or no solvent action on the coating. Where a boric acid electrolyte is used, the coating tends to form rapidly, with the result that the flow of current is soon reduced to substantially zero. This indicates that the growth of the coating has stopped. As a rule, the thickness of the coating is roughly proportional to the voltage employed for formation and the coatings of this type are exceedingly thin. In an electrolyte of this type, the coating has a high resistance when the aluminum is made anode; consequently, any current that may flow (leakage current) is very

Jan 1, 1944

By F. Keller, J. D. Edwards

The anodic treatment of aluminum presents problems of scientific as well as of commercial interest.l-3 Of particular interest is the fact that, during the anodic oxidation process, the oxide continues to form at the metal-oxide interface under any oxide previously formed. This has led to speculation as to the mechanism involved in the formation of the relatively thick oxide coatings that are used commercially for decorative or protective purposes. Furthermore, the ability of certain types of oxide coatings to adsorb dyes and other substances has stimulated research to determine the actual structure of these adsorptive coatings. It has been found that anodic oxide coatings on aluminum are composed essentially of aluminum oxide. They are formed by the action of oxygen ions penetrating to the metal surface during the electrolytic oxidation treatment. These coatings can be formed in a number of different electrolytes, as for example, those which contain sulphuric acid, chromic acid, oxalic acid, or boric acid. These are the commonest and most useful electrolytes employed commercially. Except for certain specific conditions of formation, the coatings in general have been found to be amorphous alumina, as far as can be determined by X-ray or electron-diffraction methods. Anodic oxidation processes can be arranged in three rather general classes if they are grouped in relation to the solvent action of the electrolyte on the coating.4 In the first class, the electrolyte has little or no solvent action on the coating that is formed. In general, coatings produced under such circumstances are nonporous and non-adsorptive. In the second class, the electrolyte exerts an appreciable solvent action on the coating. These coatings are porous and adsorptive. Finally, for the third class, the electrolyte tends to dissolve the coating about as rapidly as it is formed. This action produces electrolytic brightening or anodic polishing of the aluminum surface and at most leaves only a very thin film of oxide. The nonporous and nonadsorptive type of oxide film is represented by the coatings that are formed when solutions of boric acid are employed as the electrolyte. It is especially significant that these impervious films are formed in electrolytes that exert little or no solvent action on the coating. Where a boric acid electrolyte is used, the coating tends to form rapidly, with the result that the flow of current is soon reduced to substantially zero. This indicates that the growth of the coating has stopped. As a rule, the thickness of the coating is roughly proportional to the voltage employed for formation and the coatings of this type are exceedingly thin. In an electrolyte of this type, the coating has a high resistance when the aluminum is made anode; consequently, any current that may flow (leakage current) is very

Jan 1, 1944