Physical metallurgy of aluminium alloys
Aluminium is essentially a soft and weak metal which has to be strengthened by alloying with suitable elements. The elements which are added to aluminium is appreciable quantities to increase its strength and improve other properties are surprisingly limited to only four, namely, magnesium, silicon, copper and zinc. These are added singly or in combination. It may be observed that these elements are situated close to aluminium in the periodic table. Magnesium and silicon are its close neighbours in the second period while copper and zinc are close neighbours of aluminium in the next period. Out of these four elements magnesium has a greater atomic diameter (3.1906 Ã...) than aluminium (2.857 Ã...), while silicon, copper and zinc have smaller atomic diameters: 2.345 Ã..., 2.551 Ã... and 2.659Ã... respectively. The differences in the atomic diameters of these elements and that of aluminium are within 15% and therefore alloying elements Mg, Si, Cu and Zn are favourably placed for forming substitutional solid solutions with aluminium. No element is known to have complete miscibility with aluminium in the solid state. Of all the elements zinc has the greatest soild solubility in aluminium with a maximum of 66.4 atomic per cent, while magnesium, copper and silicon have much lower solubilities i.e. 16.3, 2.48, and 1.59 atomic per cent respectively. These elements show a decrease in solubility with decreasing temperature. This decrease from appreciable concentrations at elevated temperatures to relatively low concentrations at room temperature is the fundamental characteristic that provides the basis for increasing substantially the hardness and strength of aluminium alloys by solution heat-treatment and subsequent precipitation ageing operations. This method of heat treatment of aluminium alloys was first discovered by a German scientist Alfred Wilm and subsequently elaborated by Merica Waltenberg, Wanga and Scott by studying the basic principles underlying precipitation hardening.
Magnesium, copper and zinc form compounds with aluminium which have dominant control on the hardening behaviour of these alloys. The compounds formed in case of magnesium and copper are Mg2Al3 and CuAl2 respectively. The compound formation is disputed in case of zinc but there are strong indications that a metastable phase of f.c.c. structure does form during the ageing of aluminium zinc alloys. When these four alloying elements are present in combination they may form binary and ternary compounds. When magnesium and silicon are present together in aluminium a stable compound of Mg2Si generally forms in addition to the parent Mg2Al3 and CuAl2 compounds. Similarly, MgZn2 and Mg2Zn11 may also form when the alloy is based on Al-Mg-Zn compositions. In the quaternary system Al-Cu-Mg-Zn isomorphous compounds CuMg4-Al6 and Mg3Zn3Al2 have been detected apart from Mg2Zn11 and Cu6Mg2Al5.These latter isomorphous compounds form continuous solid solutions that come into equilibrium with aluminium solid solution over a limited range of concentrations.
Rates of diffusion in the solid solutions are much slower than in the liquid solutions. Thus, depending upon the rates of solidification, equilibrium phase boundaries are displaced to varying degrees with the formation of non-equilibrium solid structures. These structures frequently include non-equilibrium constituents and cored solid solutions. The non-equilibrium contstituens include non-transformed or incompletely transformed intermetallic compounds. Indeed, equilibrium is not always desirable because preferred characteristics are often developed under non-equilibrium conditions.
Non-equilibrium structures also form, as a rule, rather an exception in quenched aged alloys. Though the solubility of the four alloying elements Mg, Si, Cu and Zn decreases with decrease of temperature, it is possible to retain them in solution on fast quenching. Due to differences in solubility at room and high temperatures the solute atoms try to precipitate out as compounds discussed above under suitable conditions of ageing. The precipitation of equilibrium phases in the aluminium alloys is generally not a straight forward step but involves a series of structural changes. The first stage of these changes has been identified in most of the alloys as rejection of solute atoms on preferred planes to form zones, the zones giving rise to metastable and stable phases. The zones and metastable phases maintain coherency with the aluminium lattice and as these grow in size the interface between zones or metastable phases and aluminium soild solution is appreciably strained and therefore hardened. The maximum strengthening effect depends on the number, size, shape and distribution of zones or a metastable phase which in turn are determined by the concentration of alloying elements and time and temperature of ageing the dependence of hardness on number of zones is shown in Fig. 2. The shapes of zones and metastable coherent phases are important because they determine the number of slip planes that can be obstructed by a given volume fraction of the precipitate. Their number increases in the order sphere-disc-rod. The shape of zones depends upon the relative diameter of solute and solvent atoms. Guinier was able to show by X-ray diffraction techniques that spherical zones exist in alloys of small atomsize differences, e.g. Al-Ag, Al-Zn, Al-Zn-Mg and precipitation occurs on {111} Al matrix planes. The zones form in the shape of discs or plates when there is a great difference in the atomic sizes as in Al-Cu. The preferred direction of precipitation of such zones is with {100} habit planes. The different models for zones are shown in Fig. 3. Needle or rod type zones form in Al-Mg-Si alloys of composition lying close to AlMg2Si. The zones are composed of layers of one row of silicon atoms bounded by two rows of magnesium atoms. This type of zones were named stringlets by Geisler and Hill and are 10-20 Ã... wide and 100Ã... long. Maximum strengthening is produced when the particles are about 100 Ã... dia. and interspaced at 100â€"150 Ã.... Marked softening occurs when the coherent phases become non-coherent or increase in size or acquire coarser distribution.
During room temperature ageing or during the early stages of ageing at intermediate temperatures of quenched aluminium alloys based on Mg, Si, Cu, Zn additions, zones are the first structures which are formed. The formation of zones proceeds with velocities 10 times the self-diffusion rate of solute atoms. To resolve this difference, vacancy-assisted diffusion has been invoked. It is supposed that the vacancies are created when the alloy is heated to solution temperature. These thermally generated excess vacancies are retained during quenching and help in the diffusion of solute atoms to nucleation sites which are additionally created by their condensation. The concentration of quenched-in vacancies depends upon the relative difference in the atomic sizes of solute and solvent atoms. When the solute diameter is appreciably smaller than that of the solvent e.g., Al-Cu, the vacancies are mobile and available for promoting the diffusion. The vacancies are immobilized and diffusion made slower if the solute diameter is larger than the solvent as in Al-Mg. This has obvious bearing on the extent of precipitation strengthening which is more in Al-Cu than in Al-Mg.
The concentration of quenched-in vacancies also depends upon the quenching speeds, specimen size, quenching bath temperature, solution temperatures adopted. The specimens quenched in air harden at a slower rate than the water quenched specimens. During air cooling the excess vacancies get annealed out and only fewer vacancies are left to assist diffusion. Similarly, as specimen size increases, hardening rate decreases appreciably for the same reasons. The massive thick specimens cool at slower rates than the thinner specimens and therefore are left with fewer vacancies. For the same reasons the higher the solution temperature and the greater the speed of quenching, the faster will be the rate of zone formation and the greater the speed of hardening. This in fact is the case. The effect of specimen size on quenchageing of Al-Cu alloys is shown in Fig. 4.
Small amount of impurities present may interact with vacancies to suppress the zone formation and therefore decrease room temperature hardening. Metals like Cd, Sn. In when present up to 0.1% in aluminium-copper base alloys completely suppress room temperature ageing but accelerate artificial ageing by a factor of 3 to 8. These impurities invariably have greater atomic diameter than aluminium. To lower the free energy of the alloy the impurity atoms associate with vacancies and thus bind them to their positions. There are therefore no free vacancies to take part in zone formation. This fact has been used in developing AlCu alloys doped with cadmium. Since such alloys will not age at room temperature the necessity of storing samples at low temperature to minimize room temperature hardening is eliminated for fabrication works. However, to avail of the full effect of cadmium the alloy must not contain high concentration of magnesium which precipitates out cadmium. In alloy containing magnesium such as Al-Mg, Al-Zn-Mg, silver has been reported to have the same effect as cadmium in magnesium free aluminium base alloys.
When zones have reached an equilibrium size after natural ageing for some time there is no further increase in hardness observed. For further hardening, the alloys require to be heated to elevated temperatures to form higher structures. These phases generally possess structures of true precipitates and grow in three dimensions when they remain coherent with the matrix they increase the strength. When coherency is lost and precipitates grow in size loss of strength takes place. These structural changes are described below for each alloy.
Aluminium-Magnesium Alloys
As the solid solution of Mg in Al is cooled, the solubility of Mg decreases from 18.9 atomic per cent at 450°C to 2.1 atomic per cent at 100°C and this is accompanied by the rejection of Al3Mg2 phase from the a-solid solution. In the quench-aged alloys formation of b is preceded by intermediate phase in the grains. The hardness curves against tempering time at a given temperature shows a broad peak to coincide with formation of . The peak falls off at longer times to reach values slightly lower than the initial hardness.
The heat-treatment of Al-Mg alloys produces no useful technological properties by structural hardening based upon precipitation or ageing. The small hardening observed in these alloys is due to dispersion of â€TM and which act as obstacles to dislocation propagation. In fact locked dislocations have been observed in the electron micrograph, in the immediate vicinity of Al3Mg2. The â€TM forms in the grain boundaries in 24 hrs at 150°C and then within the grains in 100 hrs. finally evolving into non-coherent after 50 to 100 days. The alloys however, are used due to their superior capacity to resist corrosion. The alloys can be worked up to 7% Mg. The 10% Mg alloys are used as castings only. To avail of the high corrosion resistance of 10% magnesium alloys attempts have been made to make this alloy workable by addition of other elements particularly misch metal etc. but without success.
Al-Si alloys
Silicon does not form any intermetallic compound with Al. It dissolves to the extent of 1.65% at 577°C to form solid solution which deposits particles of pure silicon below the solid solubility curve. Thus in commercial alloys containing up to 10% silicon there will be two kinds of silicon existing: the one resulting from the decomposition of the a-solid solutions the one produced by direct solidification from the eutectic melt. The two forms are crystallographically similar but differ in form and distribution. The eutectic silicon is generally coarse and mechanical properties are poor.
The eutectic structure is refined by a treatment known as modification. This consists in treating the molten alloys with metallic sodium or sodium fluoride. After modification the alloy usually contains 0.005 to 0.015% Na. The function of the modifying agent is not known, but it effects an astounding refinment of the eutectic, displaces the eutectic composition from 11.7 to 13.7 silicon and decreases the eutectic temperature. The effect is equivalent to super cooling the alloy and in fact quick cooling does tend to refine the structure. The modified 13% Si alloy has a U.T.S. of 13 tons/sq. in. and elongation 15% in contrast to unmodified alloy which has a U.T.S. of only 8 tons and elongation 5%
Al-Si alloys are not regarded as heat-treatable. Von Lanker has found that the properties of Al-10% Si alloys are significantly enchanced by quenching from 530°C. Quenching freezes in the vacancies and the precipitation of the silicon from the a-Al solid solution produces high toughness (twice that of untreated AL-10% Si)
Al-Si alloys have good resistance to marine corrosion, high degree of fluidity and low shrinkage and enable castings of intricate sections to be made dense and free from cracks.
laboratory investigation of carbon anode consumption in the electrolytic production of aluminium
Introduction
It is widely accepted that the primary product at the carbon anodes of aluminium reduction cells is carbon dioxide. Hence the theoretical lower limit of the anode consumption would be 0.25 g-atoms of carbon per Faraday or 0.1120g/amp-hr. In practice the consumption is much greater than this, usually in the range of 20 to 50% greater. There appears therefore to be considerable scope to reduce it, either by improving the carbon or by making changes in the cell design and operation conditions.
In spite of this the effect of such factors on the anode consumption does not seem to have been studied extensively in the laboratory. This is due presumably to the difficulties encountered in the operation of small, externally heated, cells, Drossbach used a graphite crucible lined with a sintered alumina tube, and with another such tube around the anode and extending below it into the electrolyte to collect the anode gas. In this laboratory the anode gas was collected by enclosing the cell in a gas-tight container of heat-resistant alloy, but this was unsatisfactory. Carbon dust accumulated on the surface of the electrolyte and allowed leakage of current between the anode and the graphite crucible-cathode. This was not successfully overcome by the use of insulating side linings of sintered alumina or boron nitride. The anode gas could have changed in composition after emerging from the electrolyte by reaction of CO2 with the carbon dust and the tops of the anode and crucible. Some finely divided carbon was also found in the cooler parts of the container indicating that some disproportionation of CO had occurred.
In 1956, Vajna described a simpler arrangement in which the anode consumption was determined by weight loss and related to the quantity of electricity passed. The anode specimens were completely immersed in the electrolyte to prevent air oxidation, and the graphite crucible was unlined and provided with a loose fitting cover. Such an arrangement differing in detail from that of Vajna was tested in this laboratory and found to give reproducible results. It should be useful for evaluating different anode carbons. It has been used to determine the effect of current density, electrolyte temperature, carbon baking temperature, and certain changes in electrolyte composition on the consumption of some carbons. The results may be interpreted with respect to the mechanism of the consumption in execess of that corresponding to formation of CO2 at 100% current efficiency.
MATERIALS
Anode Carbon
No.1
A prebaked anode block from commercial production.
Nos. 2, 3, and 4
Prebaked-type specimens prepared in the laboratory from petroleum coke aggregate. They were pressed in a 77-mm diam, heated mold at 420 kg/sq cm (5970 psi) and baked over 48 hr to a final temperature of 1000°C. For the investigation of the effect of baking temperature two specimens of No. 3 were rebaked in a nitrogen atmosphere at 1150 and 1300°C, respectively.
Nos. 5, 6, 7, 8 and 9
Soderberg-type specimens prepared in the laboratory from petroleum coke aggregate. Numbers 5, 6, 8, and 9 were baked over 48 hr to final temperature of 1000°C. Number 7 was baked over 36 hr to 800°C, then two specimens each were redbaked for 0.5 hr in a nitrogen atomsphere at 940, 970, 1020 and 1050°C, respectively.
No. 10
Commereial 51 mm electrographite rods with an apparent density of 1.66 g/cu cm.
No.11
Lumps of commercial low-ash coke produced from beneficiated coal; ash content 1.4% silicon 0.30% iron 0.14% sulfur 0.53% carbon (by combustion) 97.8%, hydrogen 0.15% real density (pyenometric, kerosene) 1.97 g/cu em. The lumps were friable relative to baked carbon bodies and quite macroporous.
Electrolyte Materials
The principal material was crushed, cryolite-base, industrial electrolyte containing 6% calcium fluoride and 9% alumina. Synthetic cryolite containing 2 to 4% alumina was also used, particularly for the melts with no calcium fluoride. Industrial aluminium fluoride containing 90% total fluoride, reagent grade sodium fluoride, ceramic grade calcium fluoride, technical grade sodium chloride, and metal grade alumina were used as additives.
APPARaTUS
The arrangement of the cell is shown in Figure1. The electrolyte was in an electrographite crucible of 200 mm ID, which was inside an open container of 80-14-6 Ni-Cr-Fe alloy. This was inside a furnace with a vertical-strip electrical heating element of the same alloy. The top of the crucible was partially protected from air oxidation by a cast alumina ring. On this rested one 32 mm layer of dense alumina brick, and two 64 mm layers of fireclay insulating brick. The anode was put through a central hole 60 to 80 mm sq, which allowed enough circulation of air to burn the carbon dust which would otherwise collect on the electrolyte.
The anodes were machined to approximately 38 mm diam and 30 mm high, with a threaded hole in one end into which the anode rod was screwed. The rod was of copper, 16 mm diam and 600 to 700 mm long. The upper part was copper pipe and the lower 300 mm solid copper rod. It was held in place by two clamps, and could be moved vertically very easily to keep the anode immersion at the desired level. The lower part was usually protected by a layer of frozen electrolyte and was not appreciably attacked.
The crucibles were machined from 254 mm graphite cylinders. Usually they were replaced after 4 to 6 days because the sidewall above the electrolyte had become thin or even burnt through by air oxidation. There were some premature failures from electrolyte penetrating through the graphite, freezing it to the bottom of the alloy container on cooling, and cracking the graphite on reheating. Such failures were reduced by using instead of a flat bottom a thicker, hemispherical bottom as shown in Figure 1.
The current was supplied by a three-phase selenium rectifier, and the total quantity of electricity for each run was measured with a copper coulometer. The electrical connection to the crucible was made through the alloy container.
The furnace temperature was controlled from a thermocouple installed through the side of the furnace with its hot junction between the heating element and the crucible container. It was measured with another Pt/10% PtRh thermocouple below the container. The temperatures reported are these measured temperatures minus the difference relative to the electrolyte temperature. For 30 amp the correction was 8 to 9°C.
PROCEDURE
General
The anode was weighed, screwed to the anode rod, and immersed in the molten electrolyte. In a few minutes the electrolyte which froze on the anode melted, and the rectifier voltage was adjusted to maintain the current close to the desired value, usually 30 amp. The anode immersion was checked frequently with a small metal rod and maintained in the range of 2-5 mm. This prevented exposure of the top of the anode to air and allowed a protective layer of frozen electrolyte to be maintained on the copper rod. At the end of the run, usually after 6 hr, the anode was lifted, allowed to cool in air, unscrewed from the copper rod, and weighed. It was pulverized to minus 65 mesh (0.2 mm) and the amount of electrolyte in it determined by combustion at 750°C, correcting for the ash content of the original carbon. The consumption was expressed as a percentage of that corresponding to formation of CO2 at 100% current efficiency (0.1120 g/amp-hr). Except as indicated it was not corrected for the impurities in the carbon nor for the loss by air oxidation which occurs on removal of the anode from the cell. For three of the carbons tested the mean loss by air oxidation was 0.46 g, standard deviation 0.11 g, equivalent to 2.3±0.6% consumption for 6 hr runs at 30 amp.
In the early part of this work the electrolysis was carried out at 20 amp for 6hr. This was increased to 30 amp for 6 hr to consume a greater thickness of carbon and increase the weight loss, and to have a current density closer to industrial practice. The current was nearly equally distributed over the side and bottom of the anode and, with 30 amp for 6 hr, a layer 4 to 5 mm thick was consumed. Ignoring the bit of current which flowed from the top of the anode, the current density for 30 amp was o.64 amp/sq cm initially, and 0.9 to 1.1 after 6 hr.
The carbon consumed of 4 to 5 mm thickness was deemed ample to give a representative weight loss, since the surface roughness after electrolysis at 30 amp was such that no particles protruded by more than 2 mm and only a few protruded 1 mm. This was confirmed by tests which showed no trend of consumption with electrolysis time over the range of 0.5 to 6 hr.
For a new crucible the electrolyte was made up by weight to give a molten electrolyte depth of about 75 mm. When the depth dropped below 60 mm, usually after the third 6-hr-run, more of the original composition was added. In addition small corrective additions were sometimes required to maintain the desired composition.
To melt the electrolyte quickly before start up of a test and to compensate for the chilling caused by insertion of the cold anode, the control temperature was raised 15 to 20°C. Then the temperature was reduced by resetting the controller, and within a half hour after the start of electrolysis the desired temperature was achieved. Except in the investigation of the effect of electrolyte temperature, the desired temperature was that which gave a protective coating of frozen electrolyte 2 to 3 mm thick on the anode rod. For 300 amp this was 20 to 25°C above the primary freezing point. Small changes in freezing point, and thus in electrolyte composition, could often be detected by a change in the amount of frozen electrolyte on the anode rod.
Unless otherwise indicated the electrolyte had an NaF/AIF3 weight ratio of 1.3 to 1.4, with 6% calcium fluoride and 5 to 9% alumina. For 30 amp the temperature which gave the desired coating of frozen electrolyte on the anode rod was 970 to 980°C. It decreased 3 to 5°C in the course of 5 or 6 tests with the same crucible. This is attributed to a small increase in alumina content which would lower the freezing point.
The above procedure applies particularly to electrolysis at 30 amp with normal electrolyte compositions and temperatures at which some frozen electrolyte protects the anode rod. Some special modifications are described below.
Operation at Different Current Densities
In the determination of the effect of current density, currents of 15 to 90 amp were used with the same size of anode. Electrolysis was carried on for a total of 180 amp hr, thus for 12 hr at 15 amp, 6 hr at 30 amp, etc. This was done so that the surface areas of the specimens at the conclusion of electrolysis would all be nearly the same. The effective anode area was taken to be that of the side and bottom of the specimen. With passage of 180 amp-hr this area was reduced by 33 to 40%. The specimen dimensions were measured before and after each test and the effective areas and corresponding current densities calculated. The mean current density per test (arithmetic mean) ranged from 0.4 to 2.6 amp/ sq cm for currents from 15 to 90 amp.
Operation at Different Temperatures
In the measurement of the effect of temperature on anode consumption no frozen electrolyte formed on the copper rod at the higher temperatures. In only a few cases a thin layer mainly of alumina protected the copper. It was possible to avoid both air oxidation of the anode and corrosion of the copper rod by very carefully controlling the anode immersion at 2 to 3 mm.
An attempt was made to use greater immersion combined with water cooling of the copper rod 100 to 200 mm from the anode to maintain the frozen layer on the rod. This was abandoned since the increase in consumption with increasing electrolyte temperature was much smaller with water cooling than without. The explanation is that the temperature of the anode surface was reduced by water cooling. Thus a 50°C rise in electrolyte temperature, combined with the required water cooling, resulted in a calculated increases in the temperature of the anode surface of only 17°C.
Operation at Different Electrolyte Compositions
For the effect of Na F/AlF3 ratio two pairs of ratios were selected, each with a given primary freezing point (weight ratios of 1.0 Vs 2.4 and 1.4 vs 1.65). Hence the comparison could be made at the same temperature with the desired temperature interval above the liquidus temperature and the same amount of frozen electrolyte on the copper rod.
During the first measurements with electrolyte of ratio 1.0, a small portion solidified on the bottom of the crucible under the anode. A sample of the solid material removed from the melt showed a ratio of 1.4 and an alumina content of 1.5%, suggesting that cryolite had crystallized. This difficulty was avoided by careful control of the electrolyte temperature at 942±1°C, and by small additions of aluminium fluoride or Sodium fluoride (changing the ratio by 0.03 or less) to maintain the desired amount of frozen electrolyte on the copper rod.
After the first measurements with electrolyte of ratio 2.4 the anode had a conical shape. Thus in one case the diameter near the upper, threaded end was 34 mm, while at the lower end it was only 24 mm indicating that most of the current had flowed from the lower end of the anode. This was associated with a tendency for a thin layer of electrolyte to freeze on the upper part of the anode. The difficulty was avoided by careful control of the amount of frozen electrolyte on the copper rod by means of small additions of sodium fluoride, the temperature being held constant at 942±1°C.
For the effect of alumina content two pairs of compositions were selected, with calcium fluoride contents of 0 and 10%, respectively. The primary freezing point, and also the electrolyte temperature, varied with alumina content, so the combined effect of a change in alumina content and temperature was measured. The effect of alumina alone was calculated using the previously determined correlation between anode consumption and electrolyte temperature.
The effects of both calcium fluoride and sodium chloride were tested at levels of 0 and 10% with corresponding differences in electrolyte temperature. The effect of these additives at constant temperature was calculated (or estimated) from the observed results, as in the case of alumina content.
In the above procedures for alumina, calcium fluroide, and sodium chloride content, the changes in electrolyte temperature were advantageous since they permitted maintenance of the desired layer of frozen electrolyte on the copper rod. In addition the observed results (without adjustment to constant temperature) are pertinent, since there would be corresponding changes in electrolyte temperature in industrial practice.
alumina extraction from a pennsylvania diaspore clay by an ammonium sulfate process
Introduction
Although the major portion of the aluminium produced today comes from alumina processed by the Bayer process, certain reasons warrant investigation of new source materials and new methods for processing these materials. Some of these reasons are as follows.
1. Aluminium is the most abundant metallic element in the earthâ€TMs crust; therefore, it occurs in most rocks and minerals.
2. Bauxite, which is formed by lateritization, occurs mainly in hot humid climates; therefore, certain industrial nations are without a domestic supply.
3. Because certain nations lack a domestic supply of bauxite, there is the added cost of transportation and the fear of losing this supply during political upheavals and war.
4. The constantly increasing demand for aluminium may lead to a scarcity of bauxite in the future.
5. Certain nonbauxite deposits may have an extremely high alumina content and yet be unsuitable for use in other industries.
The investigation discussed herein was prompted by the last of these reasons. It has been known for many years that a high-alumina diaspore lay occurs in Pennsylvania. The Pennsylvania diaspore clay occurs in three areas in central Pennsylvania as shown in Figure 1. The clay in these deposits is of three general types. One portion containing high alumina and less than 3% ferric oxide is suitable for use in the refractory industry. Another portion with 40 to 70% alumina and 5 to 25% ferric oxide content would be a potential feed for aluminium extraction and has used in this investigation. The third type is a high-iron flint clay containing 35 to 39% alumina and 3 to 25% ferric oxide.
Of all the aluminium-bearing rocks and minerals, the nonbauxite type most often investigated as a source of alumina has been clay. Due to their high-alumina content (20 to 70%) and widespread occurrence, lays have been investigated as a possible source of aluminium since the beginning of the aluminium industry in 1854. Clays, unlike bauxites have a relatively high silica content which limits the use of the Bayer process for extracting their alumina. Generally other processes, either acid or alkaline in nature, are considered, each having certain advantages and limitations. In alkaline processes, Silica is a serious contaminant and usually the temperatures required for optimum alumina recoveries are higher than in acid processes. On the other hand, iron removal, which is a problem in the acid processes, gives little trouble in alkaline processes. Acid processes require specially constructed equipment.
During World War II, clays from the Pennsylvania high-alumina diaspore deposit near Clearfield were investigated for alumina extraction by the lime-soda method. The results were technically encouraging, but the costs made the method noncompetitive. The high temperatures required (1000-1250°C) and the need for desilicating the leach liquors greatly added to the costs.
The purpose of this investigation was to determine if the alumina in the Pennsylvania diaspore clay could be extracted by the ammonium sulfate process. In this process, clay and ammonium sulfate are mixed and fused to convert the alumina to a water-soluble double salt of ammonium sulfate and aluminium sulfate (ammonium alum). The alum thus formed is leached with a hot solution of dilute sulfuric acid and the alum is crystallized from the clarified liquor causing an initial purification. After further purification, alumina is precipitated by adding the alum crystals to an ammonium hydroxide solution made by passing the exhaust gases from the roasting reaction through water.
The acid ammonium sulfate process was accepted as offering a possibility for future commercial application for a number of reasons. The temperatures required for extraction were lower than in the alkaline processes. Iron, which was a serious contaminant, could be more easily removed than silica which interfered in the alkaline processes. Ammonium sulfate, which was recovered and recycled, is abundant at a moderately low cost as a coke furnace by- product.
Related Literature
Many early investigators attempted to produce a pure alumina by fusion with ammonium sulfate. As early as 1909, a U.S. patent was issued to Rinman in which he claimed that a water soluble alum could be produced by heating kaolin, feldspar, or bauxite with ammonium sulfate at 250-400°C. The alum was said to crystallize in a high degree of purity without filtering. In 1917 a Swedish patent was issued to Hultman which claimed to produce a crystalline alum by fusing clay and acid potassium or ammonium sulfate. A U.S. patent was issued to Whittington in 1925. This patent described a process in which bauxite or clay was heated with ammonium sulfate to temperatures of 525-560°C until ammonia ceased to be given off. The aluminium sulfate was water soluble while the iron was said to convert to the insoluble oxide.
A semicommercial process, patented in the United States in 1924, using ammonium sulfate was developed in Germany by Buchner. The first stage of this operation consisted of heating solid ammonium sulfate in a suitable vessel so that ammonia was driven off converting the sulfate to bisulfate. The ammonia was collected and used elsewhere in the process. Clay was treated with the fused bisulfate and a little water in a digester at temperatures around 200°C resulting in a slurry of ammonium aluminium sulfate, ammonium iron sulfate, and insolube material. An excess of hot saturated solution of ammonium sulfate was added to this slurry which was then filtered. On cooling, the ammonium alum crystallized in a state of very high purity, while the iron salts remained in solution in the excess ammonium sulfate. Aluminium hydroxide was precipitated by adding the alum crystals to an ammonia solution containing two or three times the theoretical quantity of ammonia. The ammonia and ammonium sulfate were used cyclically.
In a later work, Seyfried used a similar process for processing clays form the area around Castle Rock, Washington. Raw clay was roasted in a rotary kiln at 750°C and crushed. The calcined clay was then leached with molten ammonium bisulfate which was manufactured at the pilot plant. The temperature of the leach was 106°C or the boiling point of the slurry. Since ferrous iron did not interfere with alum crystallization, ammonium sulfite was added to the pregnant liquor before cooling to crystallize the alum. Crystallization recovered about 77% of the alumina present in the clay and approximately 8% remained in solution. Alumina was recovered by redissolving the alum crystals in a 50% aqueous ammonia solution. The ammonia and ammonium sulfate were again used cyclically.
St. Clair, working with a clay from the western United States, preferred to roast the clay-ammonium sulfate mixture. The mixture was baked at temperatures between 360 and 400°C for 2 hr. Ammonia losses were said to be negligible in this temperature range, but at higher temperatures the ammonium sulfate decomposed. It was also found that particle size was important, especially for hard dense clays. The roasted product was leached at 70°C with water having a slight excess of ammonium sulfate, or in dilute sulfuric acid solutions to keep the resulting alum from hydrolyzing. The alum solutions formed were ideal for crystallization, since a saturated solution of ammonium alum contained 126 mg of Al2O3 per 1000 g of water at 90°C and only 16.5 g of Al2O3 per 1000 g of water at 20°C. Aluminium hydroxide was precipitated with aqueous ammonia produced from the gases evolved during baking. The ammonia and ammonium sulfate were used cyclically.
Raw material
According to Foose, the high-alumina refractory clays of central Pennsylvania are all under-clays associated with carboniferous coal beds. The nodular or diaspore-bearing clays are limited to the Mercer horizon of the Pottsville series. The clays occur in lenticular masses which have a wide range of thickness.
X-ray diffraction showed that the alumina is present as the minerals diaspore (b-alumina monohydrate), kaolinite, and boehmite (a-alumina monohydrate), as shown in Figure 2. Boehmite generally occurs in small amounts but some samples have been reported which analyze up to 15% boehmite. The alumina content of these clays generally ranges from 40 to 70%. The particular samples used in this investigation were collected near Lock Haven, Pennsylvania in the Haneyville district. Chemical analysis of this sample shown the following values (in %): SiO2, 13.59; Al2O3; 56.13; Fe2O3, 9.47; TiO2, 3.12; H2O (105°C), 0.16; and loss on ignition 16.40.
The mineral proportions were estimated by use of x-ray diffraction to be about 75% diaspore, 20% kaolinite, and 5% boehmite.
The minerals diaspore and boehmite pose a special problem since on heating to dehydration they convert to insoluble species. According to Pask and Davies, the dehydration reaction for both these minerals takes place at about 550°C. Schwiersch, however, found that dehydration actually began at 340°C. On dehydration, diaspore converts to corundum while boehmite converts first to an amorphous species and then to g-alumina. All of these species are relatively insoluble in sulfuric acid.
Kaolinite which is the only true clay mineral found in the clay deposit is a hydrated aluminium silicate with the formula Al2O3.2SiO2.2H2O. Pask and Davies found that the dehydration of kaolinite takes place at 600°C. Some workers feel that on dehydration kaolinite takes place at 600°C. Some workers feel that on dehydration kaolinite converts to a mixture of amorphous alumina and silica, while others feel that an amorphous compound known as metakaolinite is formed. At any rate the material formed by the dehydration of kaolinite is soluble in sulfuric acid as shown by Pask and Davies.
The characteristics of the minerals contained in the clay dictated the method used for extracting the alumina. Since diaspore and boehmite and their dehydration products are insoluble, the method proposed by Buchner in which the raw or calcined clay was extracted with molten ammonium bisulfate could not be used. In the present investigation, the clay was roasted with solid ammonium sulfate silmilar to the method used by St. Clair.
Procedure
The raw clay was crushed and ground to minus 60 mesh. The ground clay was then mixed with solid ammonium sulfate in glass containers agitated by mixing rolls. After mixing, a small amount of water was added to the clay-ammonium sulfate mixture to insure binding. and the mixture was pelletized. Pelletizing was conducted with a Carver laboratory press using a stainless steel mold. The pellets were 3 cm in diamter and ranged between 4 and 4.5 cm in height which gave a pellet weighing between 45 and 50 g. The pellets were dried at 110°C for 12 hr and weighed before roasting. Using the total weight of the pellet, the precentage of alumina in the raw material, and the ammonium sulfate alumina mole ratio in the pellet, the alumina and ammonium sulfate content of each pellet was calculated.
After the pellets were prepared, they were placed in hard-baked ceramic crucibles and roasted in a muffle furnace. The furnace was brought to temperature and the pellet inserted. Temperatures were measured with a Chromel-Alumel thermocouple inserted in the roasting zone of the furnace. In certain test in which a controlled heating rate was used, the pellets were placed in the cold furnace and the temperature was increased at a controlled rate. Duplicate samples were tested. When roasting was conducted in a controlled atomosphere, the pellet was crushed and the three-to-ten mesh particles were roasted. These were placed in a ceramic boat and were roasted in a tube furnace with gas flowing through the tube.
After roasting, the charge was air cooled. crushed and ground with a mortar and pestle. A small aliquot of the roasted product was taken for x-ray analysis. The remainder was stored in sealed glass containers prior to extraction. Extraction of the roasted product was conducted in a 500-ml three-neck distillation flask.
After extraction, the slurry was filtered while hot through a Buchner funnel using Schleicher and Schuell Blue Ribbon filter paper. The residue was washed serveral times with hot water. The filtrate was diluted to 500 ml and kept for alumina and sulfate analysis. Alumina was analyzed by an acid-base-back-titration method using a centrifuge; sulfate was analyzed by the gravimetric barium sulfate method. The solids were kept for x-ray analysis.
the recovery of alumina from its ores by a sulfuric acid process
INTRoDUCTION
Over the past 100 years, a voluminous literature has accumulated on acid processes for the recovery of alumina from its ores, yet the alkaline (Bayer) process has been and is still being used to produce practically all the alumina required for reduction to metal. It is therefore interesting to examine the reasons governing the great amount of attention given to acid processes, particularly over the past decade.
The Bayer process usually operates with high-grade bauxites containing 50 to 60% alumina and preferably not more than 5% reactive silica. Countries such as Canada, Great Britain, and the United States which lack indigenous deposits of suitable bauxite, must therefore import supplies and add the freight charges to the initial cost of raw material. Obviously, considerable economic advantage would be gained if it were possible to use a domestic ore and to site the alumina plant in close proximity to the ore deposit. For strategic reasons irrespective of cost, it is also desirable to use local ores if possible. Hence, many processes have been devised to modify low-alumina, high-silica ores in such a way as to provide a cheap and suitable starting material for the normal Bayer process. Some of these have been successfully applied on the industrial scale, despite the fact that it is theoretically simpler to beneficiate high-silica ores by acid treatment. In closely related fields, e.g., the manufacture of aluminium sulfate for the industry, it is obviously preferable to use an acid process, working with cheap domestic ores rather than pure alumina hydrate available at considerably greater cost from the Bayer process. For this reason alone, much research into sulfuric acid processes has been instituted in countries such as the United States and Germany. Once solid or liquid alum has been made by a satisfactory process, a logical extension is to produce a more valuable material, viz., alumina of suitable purity for reduction to metal.
From the above remarks, it must be evident that extensive application of acid processes in the aluminium industry is desirable and potentially feasibleâ€"since such application has not yet occurred, it is necessary to examine the disadvantages which have delayed the acceptance of most acid processes by the industry. Naturally, the primary disadvantage is high production cost and the factors leading to this can be summarized as follows, on the basis of comments by Pearson, Edwards et al.. and Tilley et al.
w The extraction efficiency is generally not as high as in the Bayer process.
w Complexities are introduced by the elaborate recycling operations which are required to conserve reagents.
w Technical difficulties occur in the calcination stage. particularly where various alums represent the starting material and high temperatures are needed to remove the last traces of sulfure trioxide from the alumina.
w Whereas the removal of silica (and titanium) rarely presents difficulty in most acid processes, the separation of iron from aluminium is a major problem.
w Acid solution are highly corrosive and the materials of construction required for any acid process are more expensive than the mild steel generally employed in the Bayer process.
It is proposed to discuss these five factors briefly in turn. with particular reference to the C.S.I.R.O. process, and to amplify and support the comments at a later stage in the paper.
1. High extraction efficiency is relatively unimportant if the raw material is easy to mine, readily available, and therefore cheap compared with imported bauxite. Apart from this, the overall extraction obtained by many acid processes from high-silica ores is undoubtedly greater than can be obtained by the Bayer process. In the C.S.I.R.O. process, no changes in operating conditions are needed to obtain high recoveries from ores containing alumina monohydrate, whereas higher temperatures are required for this purpose in the Bayer process.
2. Elaborate recycling may be necessary in acid processes where ammonium or potassium alums are crystallized. The same is true for processes using nitric or hydrochloric acids. However, with sulfuric acid processes, recycling is not expensive and some loss of regeneration efficiency can be tolerated without serious effect on the cost structure. The C.S.I.R.O. Process is only slightly more complex than the Bayer Process and can in fact be broken down into a very similar sequence of unit operations.
3. It is true that the calcination of alums presents technical difficulties since the water content of crystalline aluminium sulfate is 56% and that of potash alum 45.7%. This problem does not arise to the same extent in the C.S.I.R.O. process, since the intermediate product, a basic aluminium sulfate, contains only 20% water and shows no tendency to dissolve in its water of hydration. Furthermore, this product being basic contains less than half the amount of sulfate per mole of alumina compared with ordinary alum and the recovery of sulfurous gases at the calcination stage is thus correspondingly less critical in terms of sulfur wastage. It is possible that high temperatures or long retention times may be needed to reduce the sulfur content of the alumina to a desired minimum but laboratory tests (q.v.) have shown that a satisfactory product can be obtained at temperatures not in excesss of those used for the formation of -alumina in the Bayer process.
4. It is generally conceded that iron contamination represents a major problem in most acid processes. In the past, the problem has been overcome either by efficient removal of ferric iron from the aluminium solutions or by converting the iron to the ferrous state prior to crystallization or hydrolysis of the desired aluminium compound. (Many elaborate devices, including solvent extraction and precipitation of ferrocyanide, have been suggested for the removal of iron but most of these appear to be unnecessarily costly.) In the C.S.I.R.O. process both of the fundamental approaches are used, i.e. much of the iron is precipitated at around 130°C and pH 3-3.5 during the modification stage and the remainder is reduced with sulfur dioxide prior to hydrolysis. As a result, the product of hydrolysis (basic aluminum sulfate) contains as little as 0.002-0.004% iron. No buildup of iron in the circuit occurs as a consequence of this procedure, since hydrolyzed ferric oxide is discarded with the digestion residue.
5. Corrosion is undoubtedly a major problem in acid processes, even at calcination stage where acid gases are evolved at high temperatures. Stainless steel is a suitable material of construction for the digestion stage where oxidizing conditions can be maintained. With the C.S.I.R.O. process, attention has been directed particularly to the hydrolysis stage, which is conducted at high temperatures and necessarily under nonoxidizing conditions to keep iron in the ferrous state. Moreover, at this stage, some free sulfur dioxide is likely to be present and must be taken into account in selecting suitable materials of construction. In the past, lead, graphite, and acid brick have been used separately or in combination under such conditions, but it is hoped that modern developments will provide superior materials which can perhaps be clad onto mild steel (e.g., Ti-0.3% Pd alloy). The problem of cost with the autoclaves is less serious than that with auxiliary equipment (valves, pipelines, etc.), which together with settling and filtration make up approximately half of the equipment costs. The lower cost for autoclaves is achieved by brief retention times which give adequate throughput with relatively small-scale equipment.
THE C.S.I.R.O. PROCESS
The process has already been described in general terms elsewhere and the purpose here is to explain in greater detail the reasons behind the choice of specific operating conditions and to provide some supporting data. Instead of describing each stage of operation separately (apart from the synopsis), an account will be given in such a way as to amplify the earlier comments on the problems associated with the use of an acid process, using the following headings: synopsis of process, experimental procedures, extraction efficiency, control of impurities, recycling operations, calcination, liquid-solid separations and costing. The reason for this approach is that the process is fully cyclic and changes in the results for one stage affect the entire scheme of treatment.
It should be stressed that although the process is potentially adaptable to continuous operation, all the results reported below refer to batch experiments on laboratory scale.
Synopsis of Process
The C.S.I.R.O. process makes use of temperatures well above 100°C to obtain good extraction of alumina and a high yield of intermediate product at the hydrolysis stage.
The intermediate product is a basic aluminium sulfate (referred to henceforth as BAS) closely related to the mineral alunite with respect both to its chemical behavior and x-ray diffraction pattern. In fact, BAS can be regarded as hydrogen alunite with the oxide formula expressed as 3Al2O3.4SO3.9H2O. The theoretical composition compares favorably with the figures in parentheses, which refer to the average composition for BAS calculated from several hundred laboratory preparations: Al2O3, 38.8% (39.5), SO3. 40.6% (40.5); and H2O 20.6% (20.5).
Generally speaking, the operating conditions depend to a large extent on the chemical behaviour of BAS, which is remarkably stable and is the only product of hydrolysis found under process conditions in the temperature range of 130°-350°C. Particularly at higher temperature. e.g., above 200°C, BAS can even be formed from acid solutions, i.e., those in which the SO3:Al2O3 ratio is greater than 2.35, which represents the ratio for normal aluminium sulfate. A detailed account of the hydrolysis reaction will be given elsewhere, since we are only concerned here with the special conditions under which a high yield of BAS is obtained for process purposes; viz., temperatures of 200°C or more and the use of solutions with a relatively low SO3:Al2O3 ratio.
For subsequent discussion, a typical but not necessarily final flow sheet is provided in a simplified form as Figure 1. The essential features can be outlined as follows:
Digestion at 180°C is used to recover further alumina from modification residues, e.g. point2.
Modification at 130°C is used to render the acid digestion liquors slightly basic by dissolving alumina from the fresh ore. A convenient index for acidity or basicity of the liquors is given by the SO3:Al2O3 ratio by weight, neglecting other metallic elements present in small amount. A basic liquor having a ratio around 1.90 is desired if high yields of BAS are wanted at the hydrolysis stage.
The modification residue is returned to the digestion stage, thereby providing, in effect, a two-stage countercurrent procedure.
Reduction of the modified liquors is necessary for the conversion of iron to the ferrous state, preferably by using sulfur dioxide. A wide choice of temperatures is available for this reaction, and the ultimate selection would be a compromise between heat balance and reduction efficiency.
The reaction involved requires 0.398 g SO2 (approximately 0.2 g of sulfur) per gram Fe2O3 in solution, viz.,
2H2O + SO2 + Fe2(SO4)3 2FeSO4 + 2H2SO4
Provided that the theoretical quantity of sulfur dioxide is used, the increase in acidity only affects the SO3:Al2O3 ratio by an amount of less than 0.02, assuming that modified liquor contains between 2 and 3 g Fe2O3 per liter. Care must be taken to use operating conditions which militate against the formation of dithionates, since these compounds consume extra sulfur (later released during hydrolysis) and slow down the rate of reduction of ferric iron.
Hydrolysis at temperatures around 200°C is used to precipitate BAS in pure condition and high yield. Some compromise is needed between cost of equipment and temperature of operation, but the preference is for the higher temperatures in order to obtain a rapid reaction. Under these circumstances, of course, equilibrium conditions are not attained, but the yield is still satisfactory.
A major improvement is achieved by recirculating some reactive alumina from the calcination stage, since this material reduces the overall So3:Al2O3 ratio during hydrolysis. Once again, a compromise is necessary because recirculated alumina is largely converted to BAS (80% conversion assumed in Fig. I) and this places an increased load on the calciner. The amount selected in Figure 1 gives a net hydrolysis yield calciner. The amount selected in Figure 1 gives a net hydrolysis yield of 60% from the entering reduced and modified liquor.
Calcination is undertaken in the temperature range of 900-1300°C depending on the type of alumina required and the amount of residual sulfur which can be tolerated in it.
Acid regeneration is effected by absorbing the sulfurous gases from calcination in the acid filtrate recovered from the hydrolysis stage. This step produces an acid liquor suitable for reuse in the digestion stage.
Experimental Procedures
Much of the work to be described has been done in type 316 stainless steel reactors holding 500-550 ml of pulp. Apart from minor modifications these are similar to the reactor described by Warren. Provision is made for agitation and aeration, using Snyder-type turbine-aerators with an overpressure of at least 60 psi oxygen to protect the stainless steel.
For the hydrolysis stage, a stainless steel body has also been used but with platinum, gold or glass liners to eliminate the corrosion which would otherwise occur. This provision is necessary because of the nonoxidizing conditions which must prevail during the precipitation of BAS. The hydrolysis autoclaves have a working volume of 200 ml, and agitation is provided by rotation of the autoclave on an axis inclined at 30° from the vertical. Tests have shown that more violent modes of agitation do not have a major effect on hydrolysis yield.
Both types of autoclave are heated externally by a gas flame, which ensures a more rapid heatup than was previously obtained with an electric winding applied to the outer shell. The flame is adjusted for the different operations so that an approximately uniform time of 15-18 min is required to reach temperature. Since data in the sequel are reported on the basis of time at temperature, any possible effects of this heatup period must be borne in mind.
Both to freeze high temperature equilibria and to simulate flashdown conditions, all reported results are based on quench-cooling, in which the hot autoclave is placed in a bath of cold water at the completion of a run. By this procedure, the liquor in the autoclave is brought to within 10° of room temperature within 2 min. Sampling of solutions during the course of an experiment has been attempted, but difficulties have arisen from the anaerobic conditions necessarily prevailing in the sampler tube and by blockage with crystals or undissolved ore particles.
Ore samples for testing have generally been given a medium grind in a laboratory disintegrator, a typical product being 20% 100-200 mesh BAS and 40% minus 200 mesh. For laboratory purposes, it is most convenient to crush in this way, although it is appreciated that a plant feed would be produced by wet grinding with circuit liquor, probably in a ball or rod mill.
Extraction Efficiency
The usual variables affecting extraction efficiency operate in the C.S.I.R.O. process, viz., nature of ore, particle size pulp density (and hence liquor concentaration), temperature, time at temperature, and excess acidity. These factors will be discussed in order by providing some illustrative examples, but it must be realized that generalizations are impossible. Each ore has its own characteristics and any process, whether acid or alkaline, must be tailored to cater for the idiosyncrasies of the source material. Moreover, it is common practice to test ores with pure acid or alkali, whereas the reagent in plant operation is a recycled liquor containing alumina as well as various impurities. As a final complication, the C.S.I.R.O. process involves a two-stage extraction of alumina, so that the residue discarded from the digestion stage is the product of a reaction between recycled acid liquor and modification-stage residue. It is fortunate, for testing purposes, that the latter residue behaves very much like bauxite (or original source material) at the digestion stage, and many of the results are reported below on this basis. Similarly, it has been found legitimate in laboratory work to use “synthetic recycle liquors†(made from alum with an appropriate addition of acid) since no significant difference in results has been observed to date when circuit impurities (especially iron) have been incorporated.
initial softening in some aluminium base precipitation hardening alloys
It has been reported by previous workers that some extent of softening is observed before setting in of the usual hardening process when ageing is carried out on the Al-Cu and Al-Mg precipitation hardening alloys. The possible reason for the initial softening has been suggested as relief of thermal strain. No experimental evidence in support of this postulate has been reported so far.
The present work was undertaken to make a systematic study of initial softening in certain Al-Cu and Al-Mg alloys. It was proposed to study the phenomenon of intial softening as a function of solute concentration, quenching medium and temperature of ageing. Hardness measurements were carried out to follow the process of softening and relief of thermal strain was studied by analysing X-ray line profile.
Experimental procedure
Preparation of Alloys
Binary Al-Cu and Al-Mg alloys were prepared from super purity aluminium (99.9%) and high purity copper and magnesium. All melting was carried out in graphite crucibles placed in electrical resistance furnaces. Required quantity of magnesium wrapped in aluminium foil was added to molten aluminium. Loss of magnesium in the form of oxide was substantially reduced by keeping the metal dipped in the molten aluminium till all of it got melted. The alloys were chill-cast in mild steel moulds. Hexachloroethane was used as degassant.
The cast alloys were forged and then annealed at 350°C for three days to ensure removal of microin-homogeneity and cast structure. Annealing was followed by machining out disc shaped specimens 20mm dia. Ã- 8mm thick. On one side of the specimens, numbers were punched for identification and the other side was polished so that subsequent to heat-treatment very little polishing was required for taking hardness values, thus reducing the handling of heat-treated speciments to a minimum.
The nominal compositions of six binary Al-Cu alloys were from Al-2% Cu to Al-4.5% Cu at intervals of 0.5% Cu. The actual composition varied from the nominal within the limits of ± 0.05%.
The nominal composition of binary Al-Mg alloys were Al-6% Mg, Al-8% Mg and Al-10% Mg. The actual composition varied from the nominal within the limits of ± 0.1%.
Heat Treatment
Initial solution treatment was carried out in muffle furnaces for at least 48 hours at 520°C for all the Al-Cu alloys except Al-4.5% Cu which was solution treated at 530ºC. Al-Mg alloys were solution treated at 450ºC. The solution treated specimens were quenched in (1) water at (20 ± 1)°C or (2) brine water at 0ºC. Microscopic examination of as-quenched specimens revealed the complete dissolution of the second phase into the parent phase.
The solution treated, quenched specimens were aged at 110°C, 130°C, 150°C, 170°C, 190°C and 210°C for the binary Al-Cu alloys and 200°C, 250°C and 300°C for the Al-Mg alloys. Within 5 minutes of quenching, the specimens were put for ageing treatment. It has been shown by previous workers that heat treatment does not result in loss of either copper or magnesium.
Hardness Measurements
Hardness values were determined on Vickers Hardness tester with 5 kg load; 4 specimens were taken out at the end of each ageing period and 3 hardness values were determined on each specimen. Average of a set of 12 readings was taken for determining each hardness value on ageing.
X-ray Diffraction Studies
For the purpose of X-ray studies the specimens were mounted on perspex sheet after polishing them. A blank run with perspex sheet was carried out to ascertain that no peaks appear due to perspex. The target used in the X-ray tube was iron. Manganese filter was used to cut off Kb radiation. However, as no crystal monochromatizer was used, the radiation of Kµ consisted of doublet.
YPC-50 type X-ray diffractometer unit was used. The counter was run at the rate of ½° per minute. Bragg angle vs intensity graph was plotted with automatic stripchart potentiometer. The graph was run for 2° on either side of the peak position to determine the background. The half width (b) of the diffracted line was evaluated by dividing the area under the curve by the peak height above the background.
Results
Figures 1 and 2 show the quenched hardness values of binary Al-Cu and Al-Mg alloys with (a) water at 20°C and (b) brine water at 0°C as quenching media.
In Figures 3 to 5 are plotted the relationships between the extent of softening and solute concentration in binary Al-Cu and Al-Mg alloys at various ageing temperatures. Figures 4 and 5 compare the extent of softening in Al-Mg alloys with the two types of quenching media. In case of binary Al-Cu alloys the extent of softening with water quench was very small, especially with lower copper contents (Cu less than 3.5%) and hence have not been plotted for comparison.
Figure 6 shows the temperature/time-to-minimum hardness relationship for the binary Al-Cu alloys.
Table I summarizes the results obtained on the X-ray diffractometer.
Discussion
Quenched Hardness
From Figures 1 and 2, it is evident that quenched hardness increases with increase in solute concentration (viz. Cu or Mg as the case may be). The hardness of quenched alloy is attributable to:
w Solid solution hardening.
w Lattice distortion due to supersaturation.
w Lattice distortion due to quenching.
In comparing the quenched hardness of alloys quenched under similar conditions, it is only supersaturation that alters the hardness value. This explains the linear increase in as-quenched hardness with increase in (a) % Cu and (b)% Mg. The quenched hardness, with brine water at 0.C as the quenching medium, is higher than with water at 20ºC as the quenching medium. This is true for every alloy of Al-Cu and Al-Mg and the difference in the as quenched hardness values increases with increase in solute concentration.
The extra hardness with brine water quench is attributable to (1) higher lattice distortion due to greater severity of quenching and (2) creation of dislocation loops due to possible collapse of vacancy clusters4 created due to large number of trapped thermal vacancies.
The greater effectiveness of rate of quenching to alter the as-quenched hardness, with increase in solute concentration may be attributed to the following reason.
When the solute concentration is higher, larger number of solute atoms are available in the matrix and they are helpful in setting up lattice distortion during the process of quenching.
