414 JACKSON: A NEW METHOD FOR THE DETERMINATXON OF SODIUM [VOl. 75 A New Method for the Determination of Sodium in Calcined Alumina and Aluminium Hydrates BY H. JACKSON SYNOPSIS-A method is described for the determination of sodium in alumina and aluminium hydrates. It consists in heating the sample with hydrochloric acid in a sealed tube at 200" C. Only partial dissolution of alumina is necessary for complete solution of the sodium; complete dissolution of alumina can, however, be effected under suitable conditions. Aluminium hydrates are dissolved completely. Sodium is determined gravimetrically, after removal of excess acid, by precipitation as the triple acetate. Separation of aluminium salts is un- necessary. The method is simple and the blank low. THE standard method for the determination of small amounts of sodium in calcined alumina (a-alumina) by the J.Lawrence Smith method is tedious and needs a skilled analyst to carry it out. The method involves sintering with calcium carbonate and ammonium chloride, extraction of calcium and sodium chlorides with water, and separation of the calcium as oxalate before the final determination of sodium by one of the usual procedures. The main fault of the method apart from the time required, lies in the high blank obtained, which is a disadvantage when the sample has a low sodium content. Wichers, Schlecht and Gordon1 used a technique involving treatment with hydrochloric acid at high temperatures in sealed tubes for the dissolution of a number of refractory oxides, ceramic materials and minerals.The method described below, p. 417, is based on this technique. Complete solution of the sample is unnecessary because the sodium appears to be present entirely on the surface of the crystal. Complete solution of a-alumina in hydrochloric acid can be readily attained, however, under suitable conditions. In preliminary work it was observed that when a-alumina was shaken with water, an alkaline solution with a pH of about 10 was obtained. Extraction of sodium from alumina with water in a Soxhlet apparatus was incomplete, however, even after 24 hours. Extraction with diluted hydrochloric acid, 1 + 1, on the other hand, gave a solution containing all the sodium together with most of the iron and a small amount of aluminium after boiling for 24 hours, The treatment was carried out in an all-silica apparatus consisting of a flask fitted with a stirrer and reflux condenser.Sodium was determined in the filtrate, after removal of iron and aluminium with ammonia, by precipitating with zinc uranyl acetate reagent. As this method was rather lengthy, and required a considerable expenditure on apparatus t o carry it out as a batch process, experiments were carried out on the extraction at tempera- tures above 100" C. in sealed glass tubes. It was found that alumina calcined at 1200" C. (or-alumina) could be completely dissolved in a 25 per cent. excess of hydrochloric acid by heating at 250" C. for 5 hours. Hence aAugust, 19501 IN CALCINED ALUMINA AND ALUMINIUM HYDRATES 41 5 simple method is available whereby the minor impurities present in alumina, e.g., sodium, calcium, iron, sulphur and phosphorus may readily be determined without introducing a high blank from reagents.Aluminium mono- and tri-hydrates, which are only slowly dissolved by hydrochloric acid at normal boiling-point, are also readily dissolved under pressure by heating with excess of hydrochloric acid for 1 to 2 hours at 200" C. The time required to carry out the determination was shortened by precipitating the sodium directly with zinc uranyl acetate, without separation of aluminium, which did not interfere. Monax tubing of 4.5 mm. bore and 1.5 to 2-0 mm. wall thickness was used throughout the work since this glass was easy to seal by means of an air-gas flame and no contamination of the solution by soda from this glass was observed. Any other borosilicate glass with low soda contefit, such as W1* or Pyrex, is suitable, but Pyrex requires a gas-oxygen flame for sealing off.Taking the tensile strength of glass as approximately 10,000 lb. per sq. in. and the bursting pressure (A) in lb. per sq. in. as A = (2W/D) 10,000, where W is the wall thickness and D the internal diameter, the calculated bursting pressure of tubing of the above dimensions is over 6000 lb. per sq. in. If this is halved for safety the bursting pressure will still be over 3000 lb. per sq. in. The pressure developed by hydrochloric acid (32 weight per cent., sp.gr. 1.16) calculated from the formula given (Zoc. cit.) is 743 lb. per sq. in. and 1635 lb. per sq. in. at 200" and 250" C. respectively so that the safety margin is high, especially when working at the lower temperature.No accidents occurred from bursting tubes throughout this work, although well over 100 determinations have been carried out. As a precaution, however, each tube may be wrapped in asbestos paper and placed in a steel shell made from a piece of %-inch gas piping of suitable length with a loose-fitting screwed cap at each end. EXPERIMENTAL SECTION I-THE DETERMINATION OF SODIUM IN CALCINED ALUMINA- The method described on p. 417 was applied to several samples of alumina from normal to finely ground samples. The sodium was also determined by the J. Lawrence Smith method. Treatment of reaction ides-Surface impurities were removed from the tubes by heating in a chromic - sulphuric acid mixture for half an hour, followed by boiling in nitric acid for half an hour after an intermediate rinse in distilled water.The tubes were then washed in distilled water and dried. Tubes 14 inches long (35.5 cm.) with a bore of 4.5 mm. gave a total volume of 5.65 ml. Thus a sample weight of 0.4 g. alumina with a 10 per cent. excess of hydrochloric acid, sp.gr. 1.16, over the stoichiometric amount, could be taken without having the tube more than two-thirds full. Dissolution of alumina at 250" C.-Preliminary tests were carried out at 250" C. in a thermostatically controlled air-oven on 0.4 g. calcined alumina and 2.6 ml. hydrochloric acid (10 per cent. excess) for various times. Each tube was sealed off, after thickening the wall, with a well-rounded end and, after cooling, the contents of the tube were mixed by a vigorous shaking.The tube was then placed horizontally in the oven with the alumina distributed along its length. Even after 6 hours, complete dissolution of the alumina was not reached. Since the reaction is of the first order, increased acid should increase the rate. This was found to be so and by using a 25 per cent. excess of acid over the stoichiometric amount (total 3.0 ml.), 0-4 g. alumina was completely dissolved in 5 hours at 250" C. After treatment, the tubes were removed from the oven and placed vertically to cool after shaking the contents out of the top of the tube. As there is no pressure increase due to the reaction, the tubes were opened by making a scratch about half an inch from the end, which was then broken off.The contents of the tube were washed out by inversion over one limb of a U-tube of 2 to 3 mm. diameter drawn down to a jet and reaching to the end of the reaction tube. Any undissolved alumina was removed directly by filtration under reduced pressure through a pulp pad contained in a sintered glass funnel of porosity 1 (see Fig. 1). Aluminium was removed from the solution as chloride2 and sodium determined gravi- metrically in the filtrate by precipitation with zinc uranyl acetate in the usual manner. High blanks were obtained, amounting to as much as 16 rng. (weight of precipitate), compared to a weight of 120 mg. for the sample. The other wider limb of the U-tube was connected to a wash-bottle. * A tungsten-sealing glass manufactured by the General Electric Company, Wembley.416 JACKSON: A NEW METHOD FOR THE DETERMINATION OF SODIUM [VOl.75 Extraction of sodium at ZOO" C.-In an attempt to reduce the value of the blank, further experiments were carried out at a lower temperature (ZOO" C.) and sodium determinations made on the resulting solutions, although complete solution of alumina was not attained. The results are given in Table I and show that complete solution of the alumina is unnecessary for a perfect extraction of the sodium. Moreover the blank value is reduced to a much more reasonable level (2 to 4mg. weight of precipitate). This indicates that the attack on the glass at 200" C. and with a shorter time of extraction is much less than the attack at 250°C. Examination of the tubes after treatment failed to show any surface attack; several of the tubes were used for a number of determinations.Fig. 1. U-tube for washing reaction tubes As shown in Section 11, p. 419, aluminium was found to have no effect on the precipitation of the triple sodium salt. Consequently, no attempt was made to separate it in these tests. The time saved was considerable. Care had to be taken, however, during the final stages of evaporation. Loss by spirting was prevented by finishing the evaporation on a low temperature hot-plate, working at 120" to 130" C., controlled by a Sunvic Energy Regulator; with this low temperature there was no danger of hydrolysis of the aluminium chloride to hydroxide, and excess acid was readily removed. TABLE I EXTRACTION OF SODIUM (CALCULATED TO Na,O) FROM ALUMINA WITH HYDROCHLORIC ACID (25 PER CENT.EXCESS) (SAMPLE ~.406) 0.4 g. alumina, 3-0 ml. hydrochloric acid Time of Alumina Temperature treatment dissolved, Na,O found, Yo % 200" c. Y9 99 250" C. 1 hour 12.33 2 n 19-95 4 99 31.45 5 9) 100 0.506 0-507 0-602 0-504 As a result of these experiments the sample weight was increased to 1 g . using the same amount of hydrochloric acid, thereby increasing the precision of the analysis, especially where the sodium content of the sample was low. Experiments using the larger sample weight gave results in close agreement with those obtained for a 0.4g. sample.August, 19501 I N CALCINED ALUMINA AND ALUMINIUM HYDRATES 417 Variations in time and temperature of treatment-Further experiments were carried out t o determine the limits of the treatment, in which the time was reduced to 30 minutes and the temperature to 160" C.The results are summarised in Table 11. TABLE I1 EFFECT OF VARIATION IN TIME AND TEMPERATURE ON THE EXTRACTION 1 g. alumina, 3.0 ml. hydrochloric acid OF SODIUM FROM ALUMINA Na,O (mean of four results at r 1 200" C. for Sample No. Time 160"C., 17OoC., 180"C., 190°C., 200°C., 1 hour), Na,O found at a temperature of A % % % % % % o;* } 0.313 0y-o } 0.210 R103 1 hour 0.267 0.292 0.30'7 0.308 30 min. 0.306 R104 1 hour 0.181 0.190 0.202 0.210 30 min. 0-206 R391 1 hour 0.263 - R392 1 3) 0.326 R406 1 5) 0.242 - - 0.283 - 0.348 - 0-506 - - - - - - Table I1 shows that below 190" C. the sodium is incompletely extracted in 1 hour and that a time of 30 minutes is just sufficient at 200" C.It is however recommended that a temperature of 200" C. for 1 hour should be used to ensure complete extraction of all samples, At temperatures above 200" C., the blank increased with an increase of temperature. Since the pressure increasesrapidlyabove 200" C., it is not advisable to go much above this temperature in the interests of safety. Blank determinations varied from 0.7 to 4.9mg. with an average of just over 2 mg. (weights of precipitates), and are very small compared with the weight of precipitate from a sample. METHOD I. THE DETERMINATION OF SODIUM IN ALUMINA REAGENTS- Diluted acetic acid-(1 + 1). Et her-dry . Hydrochloric acid-sp.gr. 1.16 (32 per cent. by weight). Zinc uranyl acetate yeage&-- Solution A-77 g.of uranyl acetate, U0,(C2H302),.2H20, 13.3 ml. of glacial acetic acid and 410ml. of water. Solution B-231 g. of zinc acetate, Zn(C2H,0,),.3H,0, 6.6 ml. of glacial acetic acid and 262ml. of water. Heat solutions A and B to 70" C., mix, stir until clear an' allow to stand for 24 hours. Filter immediately before use. Alcoholic wash solution-Alcohol (95 per cent ., saturated with sodium zinc uranyl acetate, NaZn(U0,)3(C,H,0,)9.6H,0). Prepare by precipitating 0-1 g. of sodium chloride, dissolved in 5 ml. of water, with 50 ml. of zinc uranyl acetate reagent as described below under procedure. Filter the precipitate on a small sintered glass crucible (porosity 3) and wash thoroughly with 95 per cent. ethyl alcohol. Transfer the precipitate to a bottle containing 1 litre of ethyl alcohol, 95 per cent., shake, and allow to stand for 24 hours.Filter the solution immediately before use. Exposure to low temperatures during cold weather should be avoided. PROCEDURE- Introduce 1 g. of alumina into a dry glass tube 8 mm. external x 4 to 5 mm. internal diameter and 14 inches long of Monax, W1 or similar borosilicate, low soda, glass. (If a greater volum-e is required, the tube should be longer rather than wider. The contents should not occupy more than two-thirds of the volume.) Add 3.0 ml. of hydrochloric acid and seal off the tube as close to the end as possible, making sure that the end is thickened and Store the zinc uranyl acetate reagent and the alcoholic wash solution at 20" C.418 JACKSON: A NEW METHOD FOR THE DETERMINATION OF SODIUM [Vol.75 well-rounded and that no pinhole remains. Shake the tube vigorously to mix the contents and place it horizontally in an air-oven maintained at a temperature of 200" C. t 5" for 1 hour, with the alumina distributed along the length of the tube. Then remove the tube, stand it vertically and allow to cool to room temperature. Cut off the tip and wash it and the contents of the tube on to a medium-tight pulp pad contained in a sintered glass funnel of porosity 1. Draw the filtrate, under suction, into a 150-ml. silica beaker, using a Witt's apparatus, and wash the pad well with water. Evaporate the filtrate to dryness, carrying out the final stage on a hot-plate at 120" to 130" C. Stir until the residue has dissolved, add a further 1 ml. of water if the aluminium chloride does not dissolve com- pletely, and place the beaker in a thermostat at 20" C.f 1". Precipitate the sodium with 10ml. of zinc uranyl acetate reagent and allow to stand for 30 minutes with occasional stirring. Filter off the precipitate through a sintered glass crucible, porosity 3, and wash with five 2-ml. portions of zinc uranyl acetate reagent followed by five 1-ml. portions of alcoholic wash solution. Finally, wash with 5 ml. of ether, dry at 100" C. for 30 minutes, cool and weigh. Wash the crucible, under suction, with hot water followed by alcohol and ether, and again dry and weigh. The difference between the two weights is sodium zinc uranyl zcetate. Carry out a blank determination on 3-0 ml. of hydrochloric acid. The factor for conversion to Na,O is 0-02015.A selection of results obtained by the procedure described above, together with figures obtained by the Lawrence Smith method, are shown in Table 111. The standard deviation estimated from sixteen results on five samples of alumina with sodium contents from 0.2 to 0.5 per cent. as Na,O was *0.006 per cent. TABLE I11 DETERMINATION OF SODIUM IN ALUMINA BY EXTRACTING WITH Add 1.0 ml. of water and 5 drops of acetic acid (1 + 1). Sample No. R103 R104 R391 R392 R406* HYDROCHLORIC ACID FOR 1 HOUR AT 200°C. Na,O, % L f 7 Lawrence Smith Sample weight, Found Mean method g. 0-4 0.300 7 0.313 0.4 0.5 1.0 0.317 0.4 0.206 1 0.319 0.227 - 0.265 0.348 0.4 0-219 0.5 0-206 1.0 0.209 1.0 0-274 1-0 0-349 - 0.4 0.505 - 0.460 * See also Table I. SECTION II-THE DETERMINATION OF SODIUM IN ALUMINIUM MONO- AKD TRI-HYDRATES Similar experiments to those in Section I were carried out on samples of aluminium mono- and tri-hydrates, using sample weights of 0.5 g.and 0.6 g. respectively with 3.0 ml. of hydrochloric acid (25 per cent. excess), and heating at a temperature of 200" C. Complete solution was attained after 1 to 18 hours heating for the tri-hydrate, and 1+ to 2 hours for the mono-hydrate. It was necessary, however, to take the tubes out of the oven after half an hour and shake vigorously to disperse the cake that formed along the glass, and which otherwise was difficult to dissolve. THE DETERMINATION OF SODIUM IK THE: PRESENCE OF ALUMINIUM CHLORIDE A series of tests was carried out on synthetic solutions containing different amounts of standard sodium solution, together with aluminium chloride solution, prepared from super purity aluminium, equivalent to that obtained from 0.6 g.of aluminium tri-hydrate. The solutions were evaporated to dryness, the final stage being carried out on a low temperature hot-plate working at 120" to 130" C. as before.August, 1950 J I N CALCINED ALUMINA AND ALUMINIUM HYDRATES 419 The normal treatment of the residue prior to precipitation is to dissolve it in 1 ml. of water and a few drops of acetic acid. It was found that this was insufficient in the presence of so much aluminium chloride as a viscous solution was obtained and the salt was frequently co-precipitated with the sodium triple salt, which made the solution difficult to filter.The amount of water was increased to (a) 2 ml. and (b) 3 ml., and the sodium precipitated with 20ml. of reagent. This was wasteful of reagent, and so synthetic solutions with 2ml. of water and only 10 ml. of reagent were tested. The results are given in Table IV, and show that satisfactory recoveries are obtained by this latter method. Filtration and washing were carried out in the normal way and the determinations completed gravimetrically. TABLE IV DETERMINATION OF SODIUM IN SYNTHETIC SOLUTIONS CONTAINING ALUMINUM CHLORIDE Aluminium chloride added equivalent to 0.6 g. of Al(OH), Volume of solution before precipitation, ml . 2 2 2 2 3 3 Volume of reagent added, ml. 20 20 10 10 20 20 Na,O on 0.6 g. Found, Taken, 0-134 0.133 0.270 0.267 0.134 0.133 0-269 0.267 0.134 0-133 0-268 0.267 A I \ % % 11.THE DETERMINATION OF SODIUM IN ALUMIKIUM HYDRATES REAGENTS- PROCEDURE- Introduce 0.6 g. of tri-hydrate or 0-5 g. of mono-hydrate into a reaction tube, add 3.0 ml. of hydrochloric acid and seal off the end of the tube. Mix the contents of the tube by vigorous shaking and place horizontally in an air-oven at 200" C. f 5" with the hydrate spread along the length of the tube. Allow the reaction to proceed for 1 to 2 hours until the solid has dissolved completely. This is facilitated by removing the tube from the oven after Q to 2 hour, and shaking vigorously to re-distribute the solid which tends to cake on the walls of the tube. After solution of the sample, remove the tube from the oven and stand it vertically until cold. Cut off the end of the tube and wash it and the contents of the tube into a 150-ml. silica beaker.Evaporate the solution to dryness, carry out the last stages on a hot-plate at 120" to 130" C. to effect removal of excess hydrochloric acid without decomposition of the aluminium chloride or loss by spirting. Dissolve the residue in 2.0 ml. of water and 5 drops of acetic acid (1 + 1). Precipitate the sodium with 10 ml. of zinc uranyl acetate reagent and complete the determination as in I., p. 418. Carry out a blank determination on 3-Oml. of hydrochloric acid. A series of eighteen determinations was carried out on three samples of tri-hydrate, each containing approximately 0.3 per cent. of Na,O, by the recommended procedure. The standard deviation was 50.006 per cent. The results are shown in Table V. As for Method I. TABLE V DETERMINATIONS OF SODIUM IN ALUMINIUM TRI-HYDRATE Al(OH),, 0.6 g. Hydrochloric acid, 3.0 ml. 1 to 1Q hours at 200" C. Sample Na,O found, yo Mean Na,O, A No. f v Y O S4890 0.324 0.323 0.307 0.313 0.313 0.317 0-316 S6002 0-290 0.296 0.275 0-280 0.289 0.293 0.287 S6003 0.286 0.290 0.281 0.279 0.286 0.288 0.286420 SHORT: THE DETERMLNATION OF SMALL [Vol. 75 TREATMENT OF RESIDUES All solutions containing uranium salts should be collected and the uranium recovered The author wishes to thank the British Aluminium Company for permission to publish by one of the published method^.^^^!^ this work. 1. Wichers, E., Schlecht, W. G., and Gordon, G. L., J . Res. Nat. Bur. Stand., 1944, 33, 451470. 2. Hillebrand and Lundell, “Applied Inorganic Analysis,” Wiley & Sons, New York, 1929, p. 392. 3. Rodden, C. J., And. Chew., 1949, 21, 331. 4. Rune Hedin, “Colorimetric Methods for Rapid Analysis of Silicate Materials,’’ p. 84 (Swedish 5. Clark, F., Analyst, 1949, 74, 411. REFERENCES Cement and Concrete Res. Inst. at the Royal Inst. of Tech., Stockholm, 1947). RESEARCH LABORATORIES CHALFONT PARK BRITISH ALUMINIUM Co., LTD. GERRARDS CROSS, BUCKS. Jamavy, 1960