424 MILNER AND TOWNEND : THE DETERMINATION OF ALUMINIUM [Vol. $6 The Determination of Alumiiiium in Copper-Base Alloys BY G. W. C. MILNER* AND J. TOWNEND A method for the determination of aluininium in brasses by the selective precipitation of aluminium as its benzoate complex has been studied and modified to enable the accurate determination of this element in all types of copper-base alloys. The alloy is dissolved in nitric acid, tin is removed by filtration and then the pH of the solution j s adjusted to the region of pH 4. Copper and iron are then reduced by boiling with hydroxylamine hydro- chloride and the aluminium is selectively precipitated by the addition of ammonium benzoate, After filtration, the aluminium benzoate is dissolved in a hot ammoniacal tartrate solution, re-precipitated as aluminium oxinate and the determination completed volumetrically .The range of the method is from 0.1 per cent. to approximately 12 per cent. of aluminium. THE older methods for the determination of aluminium in copper-base alloys involve the following time-consuming processes : removal of the copper by precipitation as sulphide or by electrolysis, precipitation of aluminium and iron hydroxides together, re-solution of the precipitate in hydrochloric acid after filtration and. then selective precipitation of aluminium hydroxide by means of sodium thiosu1phate.l s 2 Later methods fall into two main groups according to whether they do or do not involve a preliminary separation of the aluminium from the other alloying constituents by electrolysis with a mercury cathode.This electrolytic procedure gives virtually complete deposition of copper, zinc, iron and nickel, but only partial deposition of mangane~e,~ and since this element interferes in the subsequent method for determining aluminium, an additional step is necessary to separate it from the aluminium. The results of the determination of aluminium by this type of procedure are very reliz~ble,~ but in the analysis of many samples in an inspection laboratory the elqctrolytic procedure can delay the determinations; it also yields large quantities of highly contaminated me:rcury that must subsequently be purified ; for this reason the A.S.T.M. method5 incorporates a mercury-cathode electrolysis after the copper has been deposited electrolytically on a platinum cathode.In the group of methods that do not use mercury-cathode electrolysis is that of Edwards,6 who recommends an 8-hydroxyquinoline procedure for the determination of aluminium in aluminium bronzes that makes use of potassium cyanide to suppress the interference of copper, iron and nickel, and ferrocyanide to remove zinc and manganese as insoluble precipitates. This method, however, is not applicable to brasses. A method proposed by Bayley7 for the determination of aluminium in.brasses makes use of ammonium benzoate as a direct selective precipitant for aluminium after the reduction of copper and iron with hydroxylamine hydro- chloride. As there is a real need of a purely chemical procedure that is both rapid and accurate and that is applicable to the determination of aluminium in many types of copper-base alloys, the benzoate precipitation procedure was studied to find if it would meet this need.* Present address : Chemistry Division, Atomic Energy Research Establishment, Harwell, Berks.July, 19511 I N COPPER-BASE ALLOYS 425 EXPERIMENTAL According to Smales,* the pH range for the precipitation of aluminium benzoate is from 3.5 to 5, with an optimum pH of about 4. In Bayley’s method the pH is adjusted to this region by the addition of diluted ammonium hydroxide (1 + 1) to a nitric acid solution of 0.5 g of the brass to give the first permanent precipitate of cupric hydroxide. This precipitate is then just cleared by the careful addition of diluted hydrochloric acid (1 + 4). The solution is brought to the boil and 20ml of a 2 per cent.solution of hydroxylamine hydrochloride containing 2 per cent. of ammonium hydroxide and 10 per cent. of ammonium chloride is added to reduce the copper completely. The aluminium is then precipitated by the addition of 10ml of a 10 per cent. solution of ammonium benzoate, filtered on a paper-pulp pad after digestion for 5 minutes and washed with 1 per cent. ammonium benzoate solution. The aluminium benzoate precipitate is dissolved in a hot ammoniacal tartrate solution, the aluminium re-precipitated as its oxinate and finally determined volumetrically in the usual way. This method applied to the determination of aluminium in a number of standard and synthetic brasses gave good results for amounts of aluminium less than 1.0 per cent., but for amounts in the region of 2 per cent.the results were from 5 to 10 per cent. below the amount present. On occasions it was observed that the aluminium precipitated as the hydroxide instead of as the benzoate; when this occurred the determinations were rendered invalid because of the insolubility of aluminium hydroxide in the. ammoniacal tartrate solution. It was considered desirable to apply a more effective buffering action in the requisite pH region to prevent any aluminium precipitating as hydroxide. The above method was therefore modified by the addition of 50 ml of a pH 4-2 buffer solution (4.1 g of sodium acetate and 35 ml of N hydrochloric acid, diluted to 250 ml) after the removal of the cupric hydroxide precipitate with dilute hydrochloric acid; the method was then applied to a number of brass samples.Although all the precipitates then had the granular form of the benzoate, the results for the higher amounts of aluminium were still low. The aluminium benzoate precipitate was allowed to digest hot for longer times of up to 30 minutes before filtration. The precipitate was also washed with a solution of ammonium benzoate adjusted to pH 4 by the addition of acetic acid. But the results for samples containing 2 per cent. of aluminium were still 3 to 5 per cent. below the amount present. In the penultimate operation the aluminium oxinate is dissolved from a fairly large paper-pulp pad with hydrochloric acid, and the low results could conceivably be caused by the incomplete extraction of the oxinate from the paper pulp.However, it was found that complete extraction was always attained on applying the same technique to known amounts of aluminium oxinate filtered on similar sized paper-pulp pads. The difficulty was eventually overcome by increasing the volume of 10 per cent. ammonium benzoate solution used to precipitate the aluminium from 10 to 20 ml, and with this modification results for four different synthetic brasses each containing 2-00 per cent. of aluminium were 2-01, 1-99, 1.99 and 2-00 per cent. It appears, therefore, that the low results for aluminium by Bayley’s method were due either to the incomplete precipitation of the benzoate or to some aluminium being precipitated as the hydroxide instead of the benzoate, or to a combination of both.This method, incorporating the use of the pH 4.2 buffer and 20ml of 10 per cent. ammonium benzoate precipitant, was next applied to the determination of aluminium in aluminium bronzes containing approximately 10 per cent. of aluminium, starting with samples weighing 100mg so as to give suitable titres in the volumetric finish. In every experiment the aluminium was precipitated as the hydroxide instead of as the benzoate, so making the determination completely useless. As in these samples there was only about 90mg of copper compared with approximately 300mg in brass samples, it was considered essential to reduce the amount of the hydroxylamine reagent; on using 6 ml of the hydroxyl- amine reagent on a synthetic bronze containing 10 per cent. of aluminium, 9.10 per cent.was found. These results indicated that the method developed for brasses is applicable only to types of alloys containing a fixed proportion of copper; before it can be satisfactorily applied to alloys containing differing amounts of copper, the amount of hydroxylamine reagent to be added must be accurately related to the amount of copper present. As the aim of this investigation was to develop a single method suitable for the determination of aluminium in all types of copper-base alloys, it was desirable to reconstruct the method completely. Other possible sources of error were next investigated.426 MILNER .4ND TOWNEND THE DETERMINATIOK OF -4LUMINIUM [Vol. 76 As the reduction of copper by hydroxylamine hydrochloride proceeds, the solution becomes increasingly acid, so that, even on starting with an aqueous solution of copper sulphate, the acidity of the solution eventually increases sufficiently to prevent further reduction and the copper is incompletely reduced.In Bayley’s method the ammonia present in the hydroxylamine reagent neutralises this developed acidity and so allows the complete reduction of the copper. But for a method of general application it seemed better to use a pure solution of hydroxylamine hydrochloride and a pH 4 buffer of such capacity that a t least 500mg of copper could be reduced without a change of pH. A series of tests was performed on 500-mg amounts of copper, the amounts of sodium acetate and hydrochloric acid used in the pH 4.2 buffer being increased by the same factor to give buffers of approxi- mately the same pH, but of increasing buffering action. The results of the tests are shown in Table I.TA4BLE 1 CHOICE OF BUFFER SOLUTION FOR COPPER REDUCTION Ratio of sodium acetate (g) to N hydrochloric acid (ml) Observations 0.82/7 Copper incompletely reduced 2/17 Copper incompletely reduced 3125.5 Copper incompletely reduced 4-1/35.5 5/42-7 7-5/64 Copper completely reduced, but pH reduced below 4 Copper completely reduced, but benzoate precipitate gelatinous Copper completely reduced and benzoate precipitate filterable It was further found that 15ml of 5 per cent. hydroxylamine hydrochloride solution was just sufficient to reduce 500 mg of copper. With this quantity of hydroxylamine hydro- chloride and 7.5 g of anhydrous sodium acetate dissolved in 64 ml of N hydrochloric acid as buffer, the method was applied to synthetic alloys with the results shown in Table 11.TABLE :“I RESULTS ON SYNTHETIC ALLOYS WITH THE CHOSEN BUFFER Composition of alloy h f v Aluminium Copper, Aluminium, Iron, found, mg mg mg mg 500 10.00 nil 10.02 100 10-00 nil 10.04 100 10.00 10 10.02 Although a good result was obtained with the alloy containing iron, a slight brown turbidity remained after the copper reduction, arid the oxinate precipitate was discoloured. This behaviour was possibly due to the formation of a basic acetate of iron; the effect was avoided by adding the hydroxylamine hydrochloride to the cold solution of the alloy and then heating to boiling. The final method is given below. METHO:D The following method is applicable to coppers, brasses and aluminium bronzes for The range of the method is samples that contain the equivalent of 0.5 g or less of copper.from 0.1 per cent. to approximately 12 per cent. of aluminium. REAGENTS- All reagents should be of the highest purity and distilled water must be used throughout. Nitric acid, 0.5 per cen,t.-Dilute 0.5 ml of concentrated nitric acid (sp.gr. 1.42) to 100 ml Hydrochloric acid, diluted (1 + 4)-Dilute 20 ml of concentrated hydrochloric acid Bufer soZution-Dissolve 75g of powdered anhydrous sodium acetate in 640ml of Hydroxylamine hydrochloride, 6 per cent. solution-Dissolve 5 g of solid reagent in 100 ml with water. (sp.gr. 1.16) to- 100 ml with water. hydrochloric acid. of water.July, 19511 I N COPPER-BASE ALLOYS 427 Ammonium benzoate, 10 per cent.solution-Dissolve 10 g of solid reagent in 100 ml of water with warming. Aminonium benzoate wash solution, 1 per cent.-Prepare by dilution from the 10 per cent. solution. AmmoniacaZ tartrate solution-Dissolve 25 g of tartaric acid in water, add 120ml of ammonium hydroxide (sp.gr. 0.880) and 5 g of potassium cyanide, and dilute to 1 litre with water. 8-Hydroxyquinoline, 2 per cent. solution-Dissolve 5 g of solid reagent in 15 ml of glacial acetic acid and dilute to 250ml with water. Potassium bromate, 0.1 N solution-Dissolve 2-7840 g of potassium bromate and 15 g of potassium bromide in water and dilute to 1 litre. Sodium thiosulphate, approximately 0.1 N solution-Dissolve 25 g of sodium thiosulphate crystals in water and make up to 1 litre.Potassium iodide, 15 per cent. solution-Dissolve 15 g of solid reagent in 100 ml of water. Starch, 0.5 per cent. solution-Make 0-5 g of soluble starch into a paste with a few milli- litres of water and add to 100ml of boiling water. PROCEDURE- as follows- Cool and filter if necessary. Cool. Weigh an amount of sample according to the percentage of aluminium in the sample For aluminium contents less than 2.0 per cent., weigh out 0.5 g. For aluminium contents from 2.0 to 4.0 per cent., weigh out 0.25 g. For aluminium contents greater than 4-0 per cent., weigh out 0.1 g. Transfer the weighed sample to a 400-ml conical beaker, dissolve it in 5 ml of concentrated nitric acid (sp.gr. 1-20), add approximately 10 ml of water and boil to remove nitrous fumes.If tin is present in the sample, evaporate the solution to a paste, re-dissolve the salts in about 15ml of 0.5 per cent. nitric acid and filter through a small, tight, paper-pulp pad. Wash well with approximately 50ml of hot 0.5 per cent. nitric acid and collect the filtrate and washings in a 400-ml conical beaker. If tin is absent, dilute the sample solution, after removing the nitrous fumes, to approximately 75ml with water. Carefully add diluted ammonium hydroxide (1 + 1) to give the first permanent pre- cipitate and then add diluted hydrochloric acid (1 + 4) dropwise to just clear the precipitate. Add 70 ml of buffer solution and 15 ml of hydroxylamine hydrochloride solution, and heat to boiling. Boil for one minute, then remove from the hot-plate and add 20ml of 10 per cent.ammonium benzoate solution in one rapid addition. Stand the solution by the side of the hot-plate for about 15 minutes to allow the aluminium benzoate precipitate to settle and the supernatant liquid to become perfectly clear. Filter through a fairly large paper-pulp pad of medium compactness* and wash the precipitate and precipitation beaker well with hot ammonium benzoate wash solution. Remove surplus liquid from the pad by suction and carefully transfer the pad and precipitate to the precipitation beaker. Wash the funnel with 50ml of hot ammoniacal tartrate solution and then with hot water, and collect the washings in the precipitation beaker. Dilute to about 150 ml with water and digest at 80" to 90" C for a few minutes to dissolve the precipitate.Then precipitate the aluminium oxinate by addition of 20 ml of 8-hydroxy- quinoline solution with constant shaking. Replace the beaker on the hot-plate and maintain the solution at 80" to 90" C until the supernatant liquid becomes quite clear. Filter through a small loosely-packed paper-pulp pad and wash the precipitate well with cold 5 per cent. ammonium hydroxide solution and finally wash once with cold water. Add 40 ml of hot (80" C) concentrated hydrochloric acid (sp.gr. 1-16) to the precipitation beaker to dissolve traces of aluminium oxinate, and then dissolve the precipitate from the pad into the bottle with this acid. Dilute the final volume of solution in the bottle to 200ml with water and cool. Add a few drops of methyl red indicator solution and run in the bromate - bromide solution with constant shaking until the colour changes from orange to yellow.Add 4 ml of bromate - bromide solution in * It is important not to make this pad too large, otherwise a t the filtration stage of the aluminium oxinate the amount of pulp is too great to facilitate efficient washing of the precipitate. Thin and loosely packed pads, however, often allow the aluminium benzoate to pass through; the ideal size can only be determined satisfactorily by experience with the method. Transfer the filter funnel to the neck of a 500-ml glass-stoppered bottle. Wash the beaker and pad well with hot water.428 MILNER AND TOWNEND: THE DETERMINATION OF ALUMINIUM [Vol. 76 excess and then 10ml of potassium iodide solution. Mix well and titrate with sodium thiosulphate solution until the brown-coloured precipitate just clears.Then add about 5ml of starch solution and continue the titration until one drop of thiosulphate turns the solution bright yellow. STANDARDISATION OF THE SODIUM THIOSULPHATE- With a pipette, place 25.0 ml of the bromate - bromide solution in a glass-stoppered bottle and add 10 ml of potassium iodide solution, 125 ml of water and 40 ml of concentrated hydrochloric acid (sp.gr. 1.16). Mix well and then titrate with the thiosulphate solution, using starch as indicator, to the disappearance of the blue colour. If x ml of thiosulphate are needed, the factor (f) of the thiosulphate is 25/x. If b = the volume in millilitres of the bromate - bromide solution added, CALCULATION OF RESULTS- t = the volume in millilitres of the thiosulphate solution added f = the factor of the thiosulphate solution, and then x 0.022475 b - - t x f the percentage of aluminium = (weight of sample in g) RESULTS Excellent recoveries of known amounts of aluminium added to pure copper have already been reported above for this method.The possible interfering effects of the alloying elements usually present in copper-base alloys were next :;tudied by adding known amounts of pure aluminium to a typical alloy brass and then determining the total aluminium by the recom- mended procedure. The brass used had a percentage composition of: Cu, 55.66; Sn, 1.30; Ni, 2.85; Fe, 0.97; Mn, 0.17; As, 0.20; Al, 0.48; Zn, remainder; volumes of a standard aluminium solution were added to increase the aluminium percentage by 0.5 and 1.0 per cent.respectively. The recoveries of aluminium under these conditions are shown in Table 111. TABLE 1111 RECOVERY OF ALUMINIUM ADDED TO A STANDARD ALLOY Aluminium Aluminium present, recovered, % % Sample only . . .. .. 0.48 0.48 0.48 0.48 Sample + 0.50% of aluminium . . .. .. 0.98 0.99 0.98 0-985 Sample + 1.0% of aluminium . . .. .. 1.48 1.49 1-48 1-49 As the results of these experiments suggested that interference from the usual alloying constituents of copper-base alloys was negligible, the procedure was applied to the determina- tion of aluminium in different types of alloys with the results shown in Table IV. TABLE :[V DETERMINATION OF ALUMINIUM IN ALLOYS OF KNOWN COMPOSITION Aluminium Composition, yo by recom- A r -, mended Type of alloy Cu Pb Fe Mn Sn Ni Zn A1 method, % Aluminium bronzes- (a) Highly alloyed .. 79.58 - 4.51 1.03 0.005 5.11 0.005 9.72* 9.70 (b) Nickel absent . . 84.90 - 3.15 1.92 0.14 - 0.80 9.17’8 9.23 (c) Low manganese . . 83.17 - 3.33 0.17 0.005 3.93 0.05 9-32* 9.33 Byasses- Rem. 0.35t 0.35 (c) Propeller metal . . 55.75 0.06 0.87 0.23 1.17 0.005 Rem. 0.2lt 0.19 (a) 60/40 type .. . . 56.08 0.92 2.01 0.28 - - (b) B.C.S. “B” . . . . 58.8 0.78 0.91 1.03 1.75 1.01 33.9 1-62? 1-58 (d) Aluminium brass . . 77.86 0-02 0.02 - 0.02 0.03 20.0 1.99t 1-96 * By a mercury cathode electrolysis method.4 t By gravimetric analysis.1JJuly, 19511 IN COPPER-BASE ALLOYS 429 This procedure is fairly rapid, as a single determination takes only about two hours. Moreover, the results obtained for aluminium in different types of copper-base alloys have been found to be reliable and in good agreement with the results obtained by classical methods of analysis. The procedure is, therefore, ideal for the inspection analysis of copper-base alloys. The Admiralty has granted permission for this paper to be published. REFERENCES 1. The British Aluminium Co., Ltd., “The Chemical Analysis of Aluminium and its Alloys,” 2. Scott, W. W., and Furman, N. H., “Standard Methods of Chemical Analysis,’’ Fifth Edition, 3. Lundell, G. E. F., and Hoffman, J. I., “Outlines of Methods of Chemical Analysis,” Chapman 4. Unpublished experiments of Bragg Laboratory. 5. American Society for Testing Materials, “A.S.T.M. Methods of Chemical Analysis of Metals,” 6. Edwards, W. T., Analyst, 1948, 73, 556. 7. Bayley, W. J., Lecture to the Birmingham and Midlands Section of the Society of Chemical 8. Smales, A. A., Analyst, 1947, 72, 14. NAVAL ORDNANCE INSPECTION DEPARTMENT Publication No. 405, London, 1949, p. 165. Volume 1, Technical Press Ltd., London, 1939, p. 55. and Hall, Ltd., London, 1945, p. 95. 1946, p. 206. Industry, November 16th, 1949; Chem. and Ind., January 14th. 1950, No. 2, 34. BRAGG LABORATORY JANSON STREET, SHEFFIELD, 9 November, 1950