Talanta. 1968. Vol. 15. pp. 1259 to 1265. Pergamon Press. Printed in Northern Ireland

DETERMINATION OF THE PRECISE COMPOSITION OF THE TRIPLE SODIUM URANYL ACETATES OF MAGNES- IUM, ZINC AND NICKEL, AND SOME OBSERVATIONS ON THE USE OF THE FIRST TWO COMPOUNDS FOR

THE DETERMINATION OF SODIUM

R. G. MONK Atomic Weapons Research Establishment, U.K.A.E.A., Aldermaston, Berkshire, U.K.

(Received 27 May 1968. Accepted 28 May 1968)

Sunnna~-Three triple sodiumuranyl acetates NaM(UOr,)8(CHJCOO)p* nH,O in which the bivalent metals M were magnesium, zinc and nickel, have been precipitated and the air-dried compounds analysed for uranium by a highly precise method. Despite contrary claims it has been established that the compounds are precise hexahydrates the maximum deviation of n from 6 in any one analysis being 0.1. The precipitation of sodium as the magnesium or zinc compound gives results which depend on the excess of the reagents, and positive errors can be obtained. It is also concluded that, contrary to the usual belief, the magnesium compound is rather less soluble than the zinc one and is therefore somewhat more sensitive for sodium determination.

A VERY large number of papers have been published on the determination of sodium by precipitation of compounds of the type NaM(UO&(CH,COO),*nH,O, where M is a bivalent metal, and the method in various forms-gravimetric, titrimetric and spectrophotometric-has found application to a wide range of materials but generally in circumstances requiring no better precision than about 1%. The work described in this paper was part of an investigation into the suitability of the triple acetate method for more precise sodium determinations. This work was abandoned because it was found that the fraction of sodium precipitated depended on the excess of reagents used and, for some unexplained reason, positive errors could be obtained. Certain definite conclusions could be drawn, however, some of which are at variance with opinions expressed by other workers. The most important aspect of the investigation forms the major subject of this paper and concerns the precise composition of three of the compounds of interest -those in which the bivalent metals were magnesium, zinc and nickel.

In view of the length of time that the method has been established it seems extraordinary that final agreement on the composition of these compounds has not been reached. While there is no doubt of the empirical formula NaM(UO&,(CH,COO),~nH,O, varying values have been quoted for n in the com- pounds where M is Zn and Mg, the most widely favoured for determining sodium. The magnesium compound, dried for 30 min at llO”, was assigned 9 molecules of water by Streng,l but Caley and Foulk,2 on the basis of an analysis for UO,, Mg, Na and CHaCOO, decided that the water content corresponded to 6.5 molecules.

Barber and Kolthoff3 assigned 6 molecules of water to the zinc compound and their careful investigation cannot easily be faulted. They washed the precipitate with por- tions of 95 % alcohol (saturated with the compound), then with ether and dried it by suction. They showed that the product was stable when exposed for several days to

1259

1260 R. G. MONK

atmospheres dried by calcium chloride or concentrated sulphuric acid. It was de- hydrated completely in 2 hr at a pressure of 40 mbar and a temperature of 90-loo”, and a further 2 hr gave practically no further loss in weight. At a temperature of 110” slow decomposition of the zinc salt proceeded beyond loss of water but the magnesium compound was found to be more stable and lost only 0.3% of its original air-dried weight in 20 hr at 110”.

The accuracies of the determinations of sodium in known quantities of pure sodium compounds in both of the investigations quoted above, indicated that the compounds being weighed were stable under the specified conditions and that the water contents were consistent with the formulae claimed within about O-5 molecule.

Feldstein and Ward4 showed that the air-dried nickel compound also contained a number of molecules of water close to 6 but stated that their analysis was not suffi- ciently accurate to decide between 6 and 6.5.

Despite the evidence indicating a composition approximating to the hexahydrate the original Streng formula for the magnesium compound with 9HsO is still widely quoted in text-books, particularly general reference works on inorganic chemistry. Without any published experimental evidence the zinc salt is also quoted as containing 9H,O in at least two books.5v6 One of these gives the zinc compound with 9HsO on one page and 6& H,O on another but quotes a gravimetric factor for the latter which is actually correct for 6Hz0.

Duval and DuvaP claim that the triple acetates with zinc and magnesium are mixtures of hydrates and contain between 6 and 9 molecules of water, the water content increasing with decreasing temperature of precipitation. They give pyrolysis curves which show constant weight up to 75” in the case of the zinc salt and up to 91” in the case of magnesium. They state that the initial horizontal portion of each curve is the region of mixed hydrates, a statement which is contrary to the phase rule. In fact the horizontal affords good evidence for the existence of only a single hydrate over that particular temperature range.

The matter has been further complicated by the use of mixed aqueous-alcoholic solutions of reagents for determining sodium as the sodium magnesium uranyl acetate. Having used an aqueous-ethanolic reagent solution, Kahane8 claimed that the pre- cipitate contained 8 molecules of water but Caley and Rogers9 showed that it con- tained about 1.5 molecules of ethanol and 4.5 molecules of water. Williams and HainslO used an aqueous isopropanol reagent solution and claimed that the magnesium triple salt so produced contained 7 molecules of water and 0.3 molecules of isopropanol.

Because of these uncertainties in the compositions of the triple acetates it was decided to carry out careful analyses of the three compounds in which the bivalent element was zinc, magnesium and nickel, to determine the number of water molecules with certainty. In such a case it is believed that the best method is to determine that constituent for which the most accurate analytical method is available, preferably using a method not requiring elaborate elemental separations. In the present instance the determination of uranium by a highly precise titrimetric procedure based on the oxidation of uranium(IV) to uranium(Vl) was the obvious choice. The method used was based on a combination of well-tried standard methods and involved reduction of uranium(VI) to uranium(IV) in 1M sulphuric acid by amalgamated zinc in a Jones reductor, aerial oxidation of any uranium(III) to uranium(IV), addition of excess of

Precise composition of the triple sodium uranyl acetates 1261

iron(m), and titration of the iron@) produced. Cerium(IV) sulphate was used as oxidant with ferroin as indicator, and was standardized against sodium oxalate. As it was required to establish the number of molecules of water in the compounds to

within 0.05 or better, an error of not more than 0.05% was aimed at. Accordingly weight titration was used and the final step in both the standardization of cerium(IV) sulphate and the analysis of the triple acetate precipitates was back-titration with 0*02M iron(

As it had been shown that the fraction of sodium precipitated depended on the excess of reagent added, analyses were carried out on precipitates corresponding to more and less than 100 % recovery to determine whether any of this effect was attribu- table to variations in the composition of the precipitate.

EXPERIMENTAL

Precipitation of Triple Ace6ates

Reagents

Precipitating solutions. Uranyl acetate solutions containing also zinc, magnesium or nickel acetate were prepared from analytical grade reagents and were made 1M with respect to acetic acid. They were saturated with the triple salt at room temperature and filtered immediately before use.

Standard sodium solutions. Solutions containing 3-6 mg of sodium per ml were prepared by weighing out analytical grade sodium chloride that had been ignited in a platinum crucible at 600-700”, dissolving it in water and diluting accurately to 250 ml.

Procedure

All precipitations were carried out at room temperature in stoppered glass test-tubes which were agitated for 30 min by rotation at about 20 rpm in a motor-driven shaker. Then 5-ml portions of standard sodium solution were delivered from a calibrated pipette to the tube and, if required, the solution was evaporated to dryness and the residue dissolved in the appropriate volume of water for the experiment. The measured volume of precipitating solution was added and, after shaking as described above, filtration was carried out on a weighed porosity 3 sintered glass crucible, and the precipitate was washed with a few ml of the precipitant and then with several portions of 95% alcohol saturated with the triple salt, a total of 25-30 ml being used. The precipitate was washed three times with a few ml of ether and dried by suction. The crucible and contents were weighed after standing for 20-30 min in the balance case, which contained no desiccant. The method was essentially that used by Barber and Kolthoff’ and it was found that replicate determinations carried out under the same conditions generally agreed within 0.142 %.

Reagents

Determination of Uranium Contents of Precipitates

Standard cerium(ZV) sulphate, O-2N in 1Msulphuric acid. This solution was prepared as described by Wilson and Wilson,” 110 g of ammonium hexanitratocerate (prepared by recrystallizing 99 % pure material from dilute nitric acid) and 60 ml of concentrated sulphuric acid being heated to re move nitric acid and the residue dissolved and diluted to 1 litre with IM sulphuric acid.

Zron(ZZZ), 0*4M. Prepared by dissolving 192 g of analytical grade ferric alum in 1M sulphuric acid and diluting to 1 litre with the same acid.

Standardization of cerium(ZV) sulphate

The solution was standardized against approximately O-5 g portions of analytical grade sodium oxalate which had been recrystallized from water and dried at 120”. The method was essentially that of Walden, Hammett and Chapman, I* the cerium(IV) sulphate being delivered from a weight burette, with the quantities adjusted so that the oxalate was in slight excess. The solution was heated to about 50”, more cerium(IV) sulphate added until the solution was just yellow, the solution cooled to room temperature, 0.1 ml of 0.025M ferroin added, and the solution back-titrated with 002M iron(H) sulphate of which the titre against the cerium(IV) solution had previously been determined. The relative standard deviation of the mean of 3 titrations was 0.01%.

1262 R. G. MONK

Procedure The weighed precipitate (l-1.8 g) was dissolved in 50 ml of 1M sulphuric acid and reduced to

uranium(W) by passage through a Jones reductor as described by Kolthoff and Lingane.” After aerial oxidation of uranium(II1) to uranium(W) 20 ml of 0.4Miron(III) solution were added followed by 0.1 ml of 0*025&f ferroin indicator, and the solution was titrated with the standard cerium(IV) solution, again delivered from a weight burette. The small excess of cerium(IV) was back-titrated with @02M iron(I1) sulphate. A blank was also carried out and deducted from the observed titration.

RESULTS AND DISCUSSION

Table I shows the percentage of UO, found in each precipitate analysed, together with the recovery of the sodium added and the initial and final concentrations of UOZs+ and M”. The concentrations in the mother liquor were calculated from the known volumes of the sodium and reagent solutions, the weight of precipitate, and the concentrations of the reagent solutions.

Three different magnesium uranyl acetate solutions were used, the first being that of Caley and Foulk2 in which the uranyl concentration is low and which therefore requires a large volume of reagent per determination.

Two zinc uranyl acetate reagents were used, each containing the same concentra- tion of zinc as Barber and Kolthoff’s reagent. The uranyl acetate concentration of one solution was l-10 times that of Barber and Kolthoff and that of the other O-76 times.

One solution only of the nickel uranyl acetate reagent was used, that of Feldstein and Ward.4

The standard deviation of a single uranium determination corresponded to 0.03 molecules of water and the maximum deviation of the uranium content of any one precipitate from the theoretical for a hexahydrate corresponded to less than 0.07 molecules of water. The deviations from the theoretical hexahydrate values of the mean values for the three compounds were 0.01, O-02, and 0.04 for the magnesium, zinc and nickel compounds respectively. It is clear, therefore, that the air-dried sodium uranyl acetates with magnesium, zinc and nickel are precise hexahydrates and that statements that have been made to the contrary are incorrect. The results for the uranium content of sodium copper uranyl acetate obtained by Caley and Rogers9 also agreed very closely with the theoretical value for the hexahydrate and it is evident that NaM(U02),(CH,COO),~6H20, where M is a bivalent metal, constitutes as well- defined a class of compounds as the alums. It is difficult to account for the 6.5H,O found by Caley and Foulk2 for the magnesium compound but their uranium figures appear to have been definitely low.

The variation in the fraction of sodium precipitated is clearly not due to any variation in composition of the precipitate and must be connected only with the excess of uranyl ion, bivalent metal and acetate in the mother liquor. When the standard technique of shaking for 30 min with small volumes of precipitant that was at least 0*15M in uranyl acetate was used, it was found that the apparent fraction of sodium precipitated increased with increasing excess of reagent from below 100% to a maxi- mum of 100.6-100.8%. High sodium recoveries were also obtained in some experi- ments by Barber and Kolthoff;3 Kolthoff and Lingane13 obtained similar results, their gravimetric results being confirmed by titrimetric determinations of uranium. This phenomenon is difficult to explain. In the present work the precipitating solutions were shaken with excess of triple acetate and filtered just before use so that super- saturation could not have occurred. The only likely explanation appears to be the introduction of sodium from the glassware used, as it was observed that the triple

30

30

0.14

5 0.

142

8 G P

1264 R. G. MONK

acetate deposited slowly from the filtered reagent on standing in glass bottles. Also, in the experiment in which 150 ml of the dilute reagent were used and precipitation carried on overnight in glass, a result of 101.4% was obtained. If this explanation is correct then results below about 100.7 % obtained by the standard technique represent incomplete precipitation of the total sodium present. It would therefore appear that this system needs study with sodium-free apparatus-plastic or silica-and that excess of reagent should be such that results of at least 100.7 % are obtained in glass. Future work could, with advantage, include determinations of the solubilities of the triple acetates in precipitating solutions and mother liquors, a tracer technique with sodium-22 or -24 being used.

The relative sensitivities of the zinc and magnesium reagents are of interest. It is usually considered14 that the zinc reagent is the more sensitive of the two but that the magnesium reagent has the advantage of being less susceptible to interference by

TABLB II.--coMpARISON OF ZINC AND MAGNESIUM URANYL ACETATE SOLUTIONS AT SIMILAR CONCENTIUTIONS FOR PRECIPITATING SODIUM

Molarity of reagent in mother liquor

Uranium Zinc

Apparent Molarity of reagent in Apparent recovery of

sodium added, mother liquor recovery of

sodium added, % Uranium Magnesium %

0.085 1.01 97.7 0.085 1.04 100.0 0.101 1.04 99.2 0.092 I.05 100.1 0.087 1.20 99.5 0.078 1.23 99.8 0099 1.23 99.7 0.099 1.21 loo*5

potassium. Such comparisons are generally misleading as they are between the Barber and Kolthoff3 zinc reagent which is about 0.22M with respect to uranium and the Caley and Foulk2 magnesium reagent with a uranium concentration of about O-1 1M. Any differences therefore relate to the effect of uranium concentration rather than to the particular bivalent acetate used. Comparison should be made on the basis of equal uranium concentrations but it should be noted that even under these conditions Feldstein and Ward4 claimed the zinc reagent to be the more sensitive in their qualita- tive tests. The present work does not support this view but indicates that the difference, while not very great, is somewhat in favour of the magnesium, i.e., the solubility of the triple magnesium salt in a given reagent is less than that of the zinc salt in a reagent of corresponding composition. This is shown by the results in Table II.

These experiments were not specifically designed for this comparison, so the reagent concentrations in the mother liquors from the zinc experiments do not correspond exactly to those in the magnesium experiments. Nevertheless the greater precipitation of the magnesium compound under similar conditions is quite clear. It is of interest to note that the zinc triple acetate is also more soluble in 95% alcohol than the magnesium one. Barber and Kolthoff3 found O-5 mg/ml for the solubility of the former and Caley and Foulk2 0.2 mg/ml for the latter.

Zusammenfassung-Drei temlire Natriumuranylacetate NaM(UO& (CH,CGO), - nH,O, in denen die zweiwertigen Metalle M Magnesium, Zink und Nickel waren, wurden gefallt und die an der Luft getrock- neten Verbindungen mit einer sehr genauen Methode auf Uran analysiert. Entgegen abweichenden Angaben wurde festgestellt, dat3

Precise composition of the triple sodium many1 acetates

die Verbindungen genaue Hexahydrate sind; die hiichste Abweichung des n-Wertes von 6 in allen Analysen betrug 0,l. Die Fgllung von Natrium als Magnesium- oder Zinkverbindung gibt Ergebnisse, die von dem Reagentieniiberschuu abhiingen; positive Fehler sind moglich. Es wird such geschlossen, daB im Gegensatz zur herrschenden Meinung die Magnesiumverbindung betrachtlich unl&.licher als die Zinkverbindung und daher bei der Natriumbestimmung etwas emp- findlicher ist.

R&r&-On a precipite trois acetates triples uranyl-sodium NaM- (UO,),(CH,COG),* nH,O, dam lesquels les mttaux bivalents M sont le magnesium, le zinc et le nickel, et analyd les composes s&ches a l’air pour l’uranium par une methode hautement precise. Malgre des declarations contraires on a dtabli que les composes sont des hexa- hydrates p&is, la deviation maximale de n de 6 n’atteignant dans aucune analyse 0,l. La precipitation du sodium a l’ttat de compose du magnbium ou du zinc donne des resultam dependant de l’ex& des reactifs et l’on peut obtenir des erreurs positives. On conclut aussi que, contrairement a la conviction habituelle, le compose magnesien est quelque peu moins soluble que le zincique et est par consequent un peu plus sensible pour le dosage du sodium.

REFERENCES

1265

1. A. Streng,Z. Anal. Chem., 1884,23,115. 2. E. R. Caley and C. W. Foulk, J. Am. Chem. Sot., 1929,51,1664. 3. H. H. Barber and I. M. Kolthoff, ibid., 1928, 50,1925. 4. P. Feldstein and A. M. Ward, Analyst, 1931,56,245. 5. Handbook of Chemistry and Physics, 48th Ed., pp. B227 and B271. The Chemical Rubber

Company, Cleveland, Ohio, 1967. 6. F. J. Welcher, Ed., Standard Methods of Chemical Analysis, 6th Ed., Vol. 2, Part A, p. 429.

Van Nostrand, Princeton. 7. T. Duval and C. Duval, Anal. Chim. Acta, 1948,2,97. 8. E. Kahane, Bull. Sot. Chim. France, 1930,41,382. 9. E. R. Caley and L. B. Rogers, Ind. Eng. Chem., Anal. Ed., 1943,15,32.

10. D. Williams and G. S. Haines, ibid., 1944, 16,157. 11. C. L. Wilson and D. W. Wilson, Comprehensive Analytical Chemistry, Vol. lB, p. 245, Method 3.

Elsevier, Amsterdam, 1960. 12. G. H. Walden, L. P. Hammett and R. P. Chapman, J. Am. Chem. Sot., 1933,55,2649. 13. I. M. Kolthoff and J. J. Lingane, ibid., 1933,55, 1871. 14. I. M. Kolthoff and P. J. Elving, Treatise on Analytical Chemistry, Part II, Vol. 1, Section A,

p. 365. Interscience, New York, 1961.