Phase Relations in the System Sodium Oxide-Uranium Trioxide-Water

Phase Relations in the System Sodium Oxide-Uranium Trioxide-Water at 50 and 75°1. John E. Ricci, and Frank J. Loprest. J. Am. Chem. Soc. , 1955, 77 (...
0 downloads 0 Views 1MB Size
Download PDF
April 20, 1955

PHASE RELATIONS IN

THE

SYSTEM SODIUM OXIDE-URANIUM TRIOXIDE-WATER 2119

Acknowledgment.-The authors are indebted to Dr. S . Triebwasser for his helpful discussions and suggestions, to Mr. M. Berkenblit for preparation

[CONTRIBUTION FROM THE

of charges and chemical analyses, and to Mr. K. Asai for design and construction of the variac drive. NEWYORK,N. Y.

DEPARTMENT OF CHEMISTRY,

N E W YORK UNIVERSITY]

Phase Relations in the System Sodium Oxide-Uranium Trioxide-Water at 50 and 7501 BY JOHN E. RICCIAND FRANK J. LOPREST RECEIVED OCTOBER 4, 1954 Equilibrium phase relations in the system sodium oxide-uranium trioxide-water at 50 and 75’ have been investigated by phase rule methods. In addition to the terminal solids of NaOH.HZ0 (I) and UOs.2HzO (VI), there are a t 50’ four intervening solid combinations of the two oxides, all incongruently soluble, the concentration of uranium trioxide in the liquids throughout being of the order of 10-20 milligrams per liter. For solutions almost saturated with sodium hydroxide, containing!rom 42.8 t o 45.2% NazO (the solubility of NaOH.HzO), the saturating solid is a pink hydrated compound, solid 11, with a hlgh ratio of sodium oxide, probably 5 or 6 NazO per UO,. The next solid phase, solid 111, for liquid concentrations extending down to 0.0106% NazO, is a solid solution of considerable range in composition and ranging in color from bright orange to plain yellow. If it is anhydrous its upper sodium oxide limit is a t 4 3 N a ~ O . l l U O s ,while if it is slightly hydrated it may just include the formula NasUO,.H20; its lower limit is very nearly Naz0.3U0,. The familiar diuranate formula, Na2U207, is therefore simply a point in the continuous range of this principal solid solution. This phase is followed, for liquid conce?trations down to 0.00058% NazO, by the compound Na20.6UO3, solid IV, which shows little, if any, tendency t o take up adjacent solids in solid solution. There follows finally another solid solution, V, ranging approximately from Na20.12UOs t o N a ~ 0 . 1 8 U O ~and , saturation with both this solid and UOa.2H20 (VI) occurs a t 0.00012% NazO in the liquid. The 75’ isotherm was studied only for liquid concentrations below 30% NazO, showing in this region the same sequence, nature and compositions of solid phases as at 50’. The NazO concentrations of the three isothermally invariant liquids covered, each saturated with two solids, are lower than a t 50”.

This is a report of the phase rule investigation of Bersohn and Brady2bnot only noted this important two isotherms of the system sodium oxide-uranium effect of washing in changing the composition of trioxide-water. The information in the literature the solid before analysis but even suggested that regarding the composition of various “uranates” the original solid might have been itself variable, in of sodium and their equilibrium relations is some- the nature of a solid solution. The results of the what fragmentary and conflicting. Reference present investigation corroborate this surmise. works in inorganic chemistry mention both the With regard to the corresponding compound “normal” or mono-uranate, such as Na~U04,from “(NH&U207,)’ which had also been reported in high temperature reactions, and a variety of “poly- the early literature,S we may note that Carston uranates,” particularly the diuranates, both of and Norton6 failed in their attempt to prepare a sodium and of potassium, presumably known as solid of exactly such composition. definite compounds. The actual information] howStill other formulas have been reported for preever, is meager, especially with regard to the ques- cipitates obtained in connection with the study of tion of equilibrium relations in aqueous sys- the pH titration curves for the reaction of sodium tems. hydroxide with uranyl salt solutions. The ques“Na2UZO7”has been reported as formed by high tionable nature of these results is apparent when i t temperature reactions,2 and from the ignition of is noted that the first precipitate in such a neutraliprecipitates obtained from uranyl salt solutions zation is probably hydrated uranium trioxide, and treated with excess of sodium hydroxide. The con- that i t would be very unlikely that equilibrium siderable variation in the final composition of such of solid phases would subsequently hold during the precipitated products has been attributed to a va- continued neutralization. On the basis of the charriety of effects. Jolibois and Bossuet,a observing acteristics of the pH titration curve, although withthe sodium oxide content to vary from 1 or 2 up to out direct analysis, Britton’ inferred that the first S%, considered the precipitate to be merely hy- precipitate was ‘(uranyl hydroxide.” Similarly, in drated uranium trioxide with adsorbed sodium hy- a study of the action of calcium hydroxide on uranyl droxide. Metzger and Heidelberger4reasoned that nitrate solution, the first precipitate was found to the original precipitate was probably actually contain hardly any calcium.8 Guiter has reported NazUz07 but that it suffered decomposition in the compound 2Naz0.7U03as being formed in the washing to bring the composition down to the titration of uranyl nitrate a t room t e m p e r a t ~ r e , ~ observed values of SNaz0.20U03 and 2Naz0.5U03. and the compounds Naz0.4U03 and Na20NJ03 from uranyl acetate solutions at room temperature (1) This work was supported by the United States Atomic Energy Commission under Contract No. AT.(30-1)-1256 with New York and a t the boiling point, respectively.1° In the University. Part of the work described was done by F. J. Loprest in titration of uranyl nitrate, Wamser, Belle, Bernpartial fulfilment 01 the requirements for the degree of Doctor of Philosophy. (2)(a) C. Zimmermann, A n n . , 218, 285 (1882); (b) R. Bersohn and E. L Brady, U.S.A.E.C.literature, Mon. C-109,March, 1946. (3) P . Jolibois and R . Bossuet, C o m p l . t e n d . , 174, 1625 (1922). (4) F. J. Metzger and M. Heidelberger, THISJOURNAL, 81, 1040 (1909). This paper contains references to earlier work on these precipitates.

(5) E. Peligot, A n n . , 4S, 255 (1842). (6) A. I. Carston and T . H. Norton, A m . C h e m . J . , 10, 219 (1888). (7) H. T . S. Britton, J . C h e m . SOC.,127, 2148 (1925). (8) C . Tanford, R. L. Tichenor and H. A. Young, AECD-2653. Feb.,

1945.

(9) H. Guiter, Bull. m c . c h i m . France, 403 (1946). (10) H. Guiter, ibid., 275 (1947).

2120

JOHN

E. RICCIAND FRANK J. LOPREST

sohn and Williamson1’ reported the formation of 3Na20.7U03 and Naz0.7U03. These investigators made X-ray diffraction measurements on mixtures of the solids, and they further reported that the two compounds exist not only as anhydrous entities but also, in each case, as hydrates with both 12 and 16 moles of water per “formula.” At the same time, however, they noted the effect of washing in changing the final composition. The only previous phase rule study involving alkali iLuranates’’seems to be that of Flatt and Hess12on a portion of the quaternary system uranyl nitrate-potassium hydroxide-water, a t room temperature and a t the boiling point. The few measurements made indicated, but not very precisely, the formation of solids of the compositions 2Kz0. 5UOa and Kz0.7U03. The relations in very high concentrations of alkali, or up to saturation with either sodium hydroxide or potassium hydroxide, were not studied in any of the previous work. It seemed desirable therefore to attempt to determine the actual equilibrium phase relations in the system sodium oxide-uranium trioxide-water, especially since the information so obtained might be expected to prove useful in connection with certain uranium recovery processes. The various difficulties to be expected in such a study were already known. The solids involved are not well crystallized when formed from aqueous reactions, and they are incongruently soluble, so that they are decomposed by washing, separated only imperfectly from mother liquor, and not obtainable in a pure state for direct analysis. In addition, the solubilities involved are very low, and equilibrium with respect to solid phases is expected to be approached with extreme slowness, especially in low alkali concentrations. If an equilibrium state could be approached in a fairly satisfactory manner, however, it seemed possible that the compositions of the solids might be determined by the indirect methods of isothermal phase diagrams. The general plan then was to prepare synthetic “complexes” or mixtures of known quantities of the three components, to bring these to equilibrium by stirring a t constant temperature, to analyze the liquid phase, and from the two compositions involved for each experiment to extrapolate the tieline to the composition of the saturated solid or mixture of solids.13 Materials and Analysis.-The uranium trioxide was supplied, in two ten-pound batches, by the United States Atomic Energy Commission. The uranium content was determined both b y ignition to Us08 a t 1000” and by the standard procedure14 of reduction with zinc in a Jones reductor, treatment with air to adjust the oxidation state t o I V , and titration with ceric sulfate or potassium permanganate, either potentiometrically or, more conveniently, with fcrroin as indicator. These solutions were standardized both with iron and with pure black oxide, U30,.lb The two samples of the yellow trioxide contained some water (about 1 and 5%, respectively); also, since the titration results (11) C. A. Wamser, J. Belle, E. Bernsohn and B. Williamson, THIS J O U R N A L , 7 4 , 1020 (1952). (12) R. F l a t t a n d W. Hess, Helv. Chim. Acta, 21, 1506 (1938). (13) A. E. Hill and J. E. Ricci, THIS J O D R N A L , 63,4305 (1931). (14) C. W. Sill and H. E. Peterson, Anal. Chenz., 24, 1175 (1952). (15) Supplied by Dr. C . J. Roddeu, New Brunswick Laboratory, U.S.A.E.C.,New Brunswick, New Jersey.

VOl. 77

were higher than those by ignition in both cases, the discrepancy indicated the probable presence of nitrate, amounting to about 2% and 1%in terms of uranyl nitrate, respectively. The presence and the amount of nitrate were then verified by colorimetric determination, performed with a Beckman model B spectrophotometer, involving the extent of decoloration of a constant amount of indigo carmine,16 by comparison with known standards containing nitrate. The material was therefore purified, in 50-g. batches, by boiling in a large quantity of water, followed by washing, filtration and drying. The product contained less than 0.1% of nitrate, determined by a spot test with brucine, as described by Feig1.I’ Many experiments were run with this purified uranium trioxide, but all of them were eventually discarded, as it became evident that the solid used failed t o come to equilibrium in the aqueous system in any reasonable time. All the experiments reported were run with a more active form of uranium trioxide prepared from the peroxide essentially according to a procedure described by Kraus.18 The original uranium trioxide was dissolved in nitric acid and treated at go”, with 10% Hi02 solution added dropwise and very slowly, with vigorous stirring, t o precipitate the peroxide. Too rapid addition of hydrogen peroxide gives a non-settling precipitate which cannot be washed by decantation or filtration. After thorough washing on a biichner funnel, the peroxide, prepared in 50-g. batches, was finally converted t o the trioxide (dark orange in color as compared t o the original yellow material) by ignition, with occasional The product, grinding, for about ten hours a t 250-300’. partially hydrated Uo3, contained only negligible amounts of peroxide, of UIV and of nitrate (the latter not greater than 0.05 t o 0.1%), and its uo3 content, assayed as U308, varied, in the various batches, between 94.7 and 98.7’%. I n solutions of very low uranium content the uranium was determined by a spectrophotometric method involving dibenzoylmethane, recently described by Yoe, Will and Black.lg The absorption measurements were made both with the model B and with the model DU Beckman instruments. The sodium oxide for the ternary complexes was introduced by weight as carbon dioxide-free sodium hydroxide solutions of known concentrations. The sodium oxide content of the final saturated solutions of the ternary system was determined either by titration or by flame photometry, depending on the size of sample and its concentration. Titrations were made with various standard sulfuric acid solutions (from 0.1 to 1 N) by means of weight burets, with methyl red as indicator, carbon dioxide being boiled out at the end-point. The Beckman model B spectrophotometer with flame attachment was used for the flame photometry measurements, which were based on calibration curves prepared with various standard sodium chloride solutions. I n a few cases the sodium content was determined gravimetrically by weighing as NaCl after evaporation with aq. HCl, this being possible because the uranium oxide content was negligible. In the analysis of aliquots of wet residues the sodium was determined either with the flame photometer or gravimetrically as the triple sodium-zinc-uranyl acetate; in these wet residue analyses the uranium content waq determined by titration. Experimental Procedure.-For the majority of the experiments, covering the range below about 3070 NaaO in the liquid, complexes, ranging from 150 to 300 g. in total weight, mere prepared, in long-necked round-bottom Pyrex flasks, from known weights of assayed GOa, KaOH solution and water. Each flask was clamped in position, almost totally immersed, in a large constant temperature water-bath, and i t was fitted with the stainless steel two-bladed stirrer of an individual stirring motor, run at about 1000 r.p.m. The stirring shaft was passed through a close fitting hole in layers of “parafilm” and filter paper a t 50°, and of filter paper and aluminum foil a t 7 5 ’ , closing the mouth of the flask, and sealed around its rim with cellophane tape. This arrange(16) C. J . Rodden. editor-in-chief, “Analytical Chemistry of t h e hlanhattan Projcct,” llcGraw-€Iill Book Co., Inc , h-ew York, N. Y , 18.51, p 215 (17) F Feigl, “Qualitativp Analysis hy Spot Tests,” 2 n d ed.. Xordemann Pub1 Co., Inc , X-ew York, N. Y . , 1939, p. 212. (IS) C. A. Kraus, U.S.A.E.C. literature, A-360, Oct., 1942. (19) J . H. Yoe, F. Will and R . A . Black, Anal. Chcm., 25, 1200 (1953).

April 20, 1955

PIIASE RELATIONS IN

THE

212 1 SYSTEM SODIUM OXIDE-URANIUMTRIOXIDE-WATER

ment was found t o be convenient for continuous vigorous agitation, carried on in some cases for a total of 16 days, and it did not result in too much evaporation. Evaporation caused a change in the calculated composition of the total complex being stirred, and for the purpose of the necessary correction t o be applied the extent of evaporation was determined by recording the loss of weight of the flask and contents (including the stirrer). On the basis of several experiments followed in this way the average rate of evaporation a t 50” was found to be 0.3% by weight per day for solutions containing not more than -20% NazO. For higher concentrations the change was negligible and at very high concentrations, near saturation with sodium hydroxide, there was actually a gain in weight. At 75’, the evaporation for dilute solutions was -0.06% per hour, falling off somewhat for the highest concentrations. Absorption of carbon dioxide by the alkaline solutions was found either negligible or entirely absent, on the basis of differential titration with two indicators. For analysis of the solution after a certain period of stirring, the flask was first allowed to stand a t room temperature for the solid to settle, if it would settle a t all. Samples of the liquid were then taken either by centrifuging or by suction filtration through ultrafine sintered glass biichner-type funnels. For solutions of very low concentration the sodium oxide content of the first portions filtered was usually too low, and it was necessary t o establish, by many tests on a variety of samples, a suitable procedure ensuring t h t f a!ter the discarding of a certain number of “mashing” or priming” portions, a sample would be obtained representing the final constant value for the solution being analyzed. In many cases successively filtered portions were analyzed until constant, regardless of the pre-established routine. The first portion, moreover, often was turbid, but the retention of solid improved after a distinct layer of it had accumulated on the filtering disc. After sampling and analysis the complex was returned t o the thermostat for further stirring, and the procedure was repeated in order t o verify the attainment of equilibrium. For the experiments a t 50” dealing with liquid concentrations below 0.05% NapO, from three t o eight such successive analyses were generally necessary. In some cases the material remaining from a completed experiment was then used as the start of a new total complex, after addition of further small quantities of the components. In addition t o the desirability of economizing in material, this procedure made it possible to start with the solid phase or phases already close t o the new equilibrium state being investigated. Since the greatest difficulty in the low alkali region seemed t o be the slowness with which the mixtures came t o equilibrium, several variations were tried in the order of mixing, in stirring of the uranium trioxide with water before addition of the sodium hydroxide solution, etc., but with no observable improvement in any case. While the experiments a t 75” were carried only t o -30% h’azO in the liquid, those a t 50” were extended t o saturation with NaOH.HZ0. In this high alkali region preliminary experiments were made with “open compexes” as described above, but on a smaller scale, for the sake of the very effective high speed stirring thus possible. I t was found, however, that the solutions took up water from the air so irregularly that the total complex composition a t the time of analysis of the liquid phase became too uncertain. Since these experiments did indicate that equilibrium in this region was being attained much more rapidly than in the dilute region, the final experiments for the high concentration region were made in closed tubes without internal stirring, and simply rotated on a wheel in the water-bath for periods up t o 10 days. Re-analysis showed that the stirring time used had been sufficient. The liquid phase in these experiments, very close t o or a t saturation with sodium hydroxide, could be neither filtered nor centrifuged, because the temperature could not be allowed t o fall. The mixtures were simply allowed t o settle in the water-bath, and the clear liquid was rapidly withdrawn through a hot pipet for analysis.

-

Results and Discussion Relations at 5O0.-The principal data for the experiments a t 50” are listed in Table I. Columns A and B give the composition of the total complex as calculated from the weights of the components

TABLE I

THESYSTEMNaZO-UOa-H20

AT 50’ I n weight percentage. B C D % NazO % NazO % NazO Hr.

A

%i:Oa

No.

In complex

complex

Liquid

...

0

5 7

8 9 10 11 12 13 14

3.89 5.40 3.91 5.47 5.30 3.86 5,41

15 16 17 18 19 20 21 22 23

5.46 5.46 3.88 5.44 5.43 5.38 5.39 5.41 5.46

45.20

43.78 43.59 43.33 43.23 42.83 42.86 42.42

36 37

38 39 40 41 42 43 44 45 46 47 48 49 50

51 52 53 54

55 56 57 58 59 60 61 62 63 64 06

4

a2

45.15 45.25 45.26 45.16 45.1.5 45.18

95 54.4 43.3 38 35.2 33.7

6 2 18 36 18 18

a,i a,i

44.28 44.26 43.88 43.82 43.62 43.33 43.15

36.1 36.4 34.7 37.0 33.8 35.5 34.3

144 240 144 240 96 144 240

+ solid I1 + solid I11

42.25 42.12 42.26 41.80 41.55 41.38 41.17 40.78 40.87

Liquid 5.66 5.39 5.46 5.44 12.94 5.40 5.48 9.73 5.41 12.79 13.10 9,772 8.889 3.173 8.403 7.256 6.861 5.233 5.460 5.410 5.508 1.817 17.058 5.0916 4.7788 0.7055 2.925 0.7635 2,954 2,964 1.324 1.9214 2.960 5.011 5.044 5.070 1.348 1.357 0.2326 1.373 0.2524 8.164

Notes on

r ~ n g method

+ solid I1 (hydrated compound)

Liquid

24 25 26 27 28 29 30 31 32 33 34 35

sfir-

+ solid I + solid I1

48 45.61 45.08 44.1 43.67 43.47

Liquid

in

“solid” + solidliquid I (NaOH.Hz0)

Liquid 0.3 1.74 5.12 9.4 9.65 9.87

In

43.07 42.94 42.91 42.98 42.75 42.77 42.66 42.42 42.54

33.0 32.9 31.4 27.2 26.5 23.7 20.8 17.4 17.2

21G 240 144 240 216 216 240 240 240

+ solid 111 (solid solution)

40.40 39.43 38.37 37.42 34.37 36.35 35.45 33.91 34.47 26.26 18.59 17.874 15.533 14.774 9.532 7.630 6.353 3.657 2.549 1.971 1,320 0.778 2.196 0,8992 .7459 ,2659 ,4666 .2558 .4501 ,4297 ,2540 ,3115 ,4158 .5775 .5497 ,6213 ,1498 ,1342 ,03404 .1276 ,03221 .e680

42.26 41.18 40.09 39.10 38.31 37.95 37.00 36.66 35.94 28.85 19.98 18.79 16.11 14.93 9.458 7.408 6.037 3.255 2.062 1.462 0.770 ,592 ,420 .381 ,257 ,1943 ,184 ,1791 ,161 ,145 ,124 ,123 ,1208 ,0958 ,0713 ,0544 ,0304 ,0196 ,0176 ,0151 ,0146 ,014

14.00 12.90 12.54 11.89 11.50 11.82 12.15 11.91 12.04 11.06 10.59 10.36 10.3 10.6 10.3 10.22 10.23 10.19 10.09 9.95 9.79 9.88 9.84 9.60 9.60 9.57 9.00 9.46 9.07 8.89 9.04 9.04 9.19 8.86 8.73 8.48 8.16 7.80 6.66 7.59 6.59 7.19

360 168 168 168 3 168 168 2 168 5 3 2 5 5 4

5 4 4 4 4 13 26 115 24 40 62 17 23 17 17 16 22 21 31 35 24 7 7 65

7 65 36

a,i a,i a,i n,i

21 22

JOHN

E. RICCIAND FRANK J. LOPREST Liquid

TABLE I (Concluded) A %,UOI

No.

in complex

B % NaaO in complex

3.165 0.1872 1.394 0.4999 0.2543 3.479

71 72 73 74 75 70 77

0.4608 ,4994 .2239 ,2304 3.221 o.2100 ,4076 0.9896 3.221 0.2318 0 A183 3.253

78 79

xn 81 82 83

0.2315 ,0232 ,0996 ,0403 ,02875 ,2286 ,0381 .03.53 ,0222 ,0238 ,1702 ,0212 ,0323 ,01374 .1590 ,0212 ,0358 .I300

Liquid

87 88 89

on 91 92 93 94 95 90

97 98 99 100 101 102 in3 104 105 100 I 07 108 109 iin Ill 112 113

114 113 116 117 118 I19

Liquid 120 121 122 123 121 125 126 127 137 128 129 130 136 137 131 136 137

liquid

0.01603 ,01507 ,08935 ,00333 ,01297 ,01193 .01110

0.09856 ,4481 ,5019

,00255 ,00870 ,00950

,06880

(3) (3)

0.0119 ,0104 ,0091 ,0083 ,0127 .0080 ,0111 ,0085 ,0109 ,0123 0113 ,0111 ,0102 ,0092 ,0127 ,0109 ,0115

,0128

0.0108 ,0094 0092 ,0086 ,0081 ,0080 ,0076 ,0075 ,0073 ,0065 ,0063 ,0057 ,0056

,0054 ,0045 ,0044 ,0039 ,0033 ,0033

,0030 ,0028 I0022 ,0022 ,0022 ,0021 .0020 ,0016 ,0011 ,0010 ,0009 .on08 ,00050 ,00045 ,00045 ,00045 .00040

0.00043 ,00044 ,00077 ,00086 ,00045 ,00071 ,00059 ,00068 .00081 ,00062 ,00053

.00048 ,00069

,00064 ,09514

%,NazO in “solid”

6.50 6.46 6.11 6.03 5.99 5.98 5.50 5.12 4.88 4.79 4 71 4.69 4.54 4.42 4.38 4.30 3.82 3.69

Hr. stirrmg

56 120 72 15 04 72 64

3.82 3.85 3 8 3.6 3.85 3.55 3.9 3.4 3.53 3.67 3.7 3.6 3.51 3.4 3.72 3.5 3.4 3.53 3.3 3.52 3.3 3.41 3.41 3.4 3.52 3.38 3.47 2.99 3.4 3.1 3.15 3.15 3.22 3.28 3.22 3.46

15

87 87 127 140

68 44

135 87 114 93

125 114 67 91 64 117 44

136 136

84 91 64 106 64 89 64 64 98 43 52 43 22 22 66 41 22 22 73 240 240 260 300 123 114 130 100

00200

,00043 ,00049 ,00044

2.94 2.81 2.62 2.48 2.42 2.14 2.01 2.01 1.92 1.81 1.79 1.77 1.77 1.75 1.62 1.59 1.59

123 123 159 136 160 19 16 76 68 21 100 200 30 22 21 32 32

,4957

b,d.c,h b,h b,h b,g b,R,h b,h b.h b,h b,d,h b,d,c,h b,h b,h b,h b,h a,h b,h b,h b,h b,h b,h b,h b,h b,h b,h b,d,c,h b,h b,fh b,h b,h b,h a,g

b,h b,f,h b,h b,f,h b,h

b,d,e,g,h

+ solid V (solid solution)

.4956

,00805 ,00805

,4243 ,4244

,00587 ,00587

Liquid 136 137 139 (1) 138 139 140

+ s(Aid I\‘ + solid V

,5150 ,5070 3.292 0.09789 .504i ,5138 ,5132 3.3247 (4)

(2) (2)

In

+ solid IV (Na204UO8)

0.4995 0.03045 3.2543 ,1391 n . n 9 8 8 ~ . o i3n5 ,1012 ,0123 ,4725 . 0269 1.4380 ,0607 0.1008 ,0116 0 . 1016 ,0109 3.214 ,1243 0.4997 ,0254 ,1021 .moz ,1017 .now 1.4416 ,0579 0.1022 ,0090 1.437 ,0599 n.ioo9 ,00805 ,1010 ,0074 .496R ,0215 1010 ,0068 1.4453 .0551 0 1027 . OO(i3 1.4473 ,0532 1,4463 ,0533 0,0986 0057 0.4997 0203 1.441 ,0525 0 4962 ,0194 3 3004 1031 0.1002 . on45 ,1002 004 1 ,0171 5010 ,5396 ,01799 ,5277 ,01797 3.4651 ,1181 0.5010 .Olio8 0 4516 ,01056

84 85 80

% .Nm0

132 133 136 (1) 137 (1) 139 (2) 134 135

D

+ solid I11 + solid IV

Liquid 66 67 68 69 70

c

Vol. 77 0.00030 ,00028 ,00021 .00020 ,00018 ,00018 ,00016

1.54 1.58 1.43 1,43 1.22 1.32 1.33

I6 16 20 20 GO

22 22

b,f,h b,d,h b,d,c.h b,d,c,h b,d,e,g,h b,d,h b,f,h

+ solid 17 + solid VI (U03.2FT20)

0.4807

0.00544

0.5154

0.00474

0.00012 .00012 ,00014 ,00013 . 00008 .00011

1 23-0 1 .23-0

1.22 1 09 1.03-0 0.89

24 24 22 84 00

40

b,d,c,g b,d,e,g b.d,e,h b.h b,d,c,R b,h

a Na2O in liquid determined by titration. * Na2O i n liquid determined by flame photometry. NazO in liquid UO3 predetermined as NaCl by evaporation with HCl. stirred with water, for one hour t o a day, before addition of NaOH solution. e NaOH solution added in successive small increments. f UOa stirred for one day with NaOH solution and 50 ml. of water before addition of further i50 ml. of water. Liquid phase isolated by centrifuging. * Liquid phase isolated by ultra-fine filtration. Liquid phase sampled by pipet after settling. i Complexes rotated in closed tubes (clear liquid obtained by settling); all others internally motor-stirred, open t o air.

taken a t the start of the run. Column C gives the sodium oxide concentration of the liquid phase. The value listed is that which is considered to be the most probable value representing equilibrium for the mixture. It is averaged from determinations made in a series of consecutive samplings upon repeated stirring, each sample being analyzed in duplicate in almost every case. The sixth column lists the total time of stirring to which the liquid concentration listed in column C corresponds. The actual individual average analyses, for each stirring time, are listed separately in Table IA.?O Examination of this table shows that in most of the experiments the attainment of equilibrium was definite. In some cases the final values were not quite constant, but it was nevertheless clear that the system was not far from equilibrium. It will be seen later, moreover, that the effect of small uncertainties in the final value of yo NasO in the liquid upon the calculation of the composition of the solid phase, at least for experiments with liquids with very low sodium oxide concentration, is very small. Although the uranium content of the liquids was also determined in many cases, the values found are not included in Table I. The values were crratic and not reproducible, and the most that can be said is probably simply that the solubility of uranium trioxide in these solutions is roughly of the order of 10 mg. per liter. The difficulties may have been partly in analysis, partly in imperfect separation of solid suspensions from the liquid to be analyzed, and partly in the possible effect of traces of carbon dioxide absorbed during the stirring; but the principal difficulty was probably the lack of attainment of true solubility equilibrium. There seems to be no question that equilibrium was (20) This table, and t h e similar Table I I I A mentioned later, for the d a t a a t 75O, have been deposited as Document number 4426 with the AD1 Auxiliary Publications Project, Photoduplication Service, Library of Congress, Washington 25, D. C. A copy may be secured b y citing t h e Document number a n d b y remitting $1.25 for photoprints, or $1.25 for 35 mm. microfilm in advance by check or money order payable to: Chief, Photoduplication Service, Library of Congress.

April 20, 1955

PHASERELATIONS IN

THE

SYSTEM SODIUM OXIDE-URANIUM TRIOXIDE-WATER 2123

attained in the experiments with respect to the which a wet residue point could be fixed. For either formation of the correct solid phases as function of method the only analytical determination necessary the sodium oxide concentration of the liquid. This for fixing the liquid composition was that of SOis the process followed by successive analysis of the dium oxide since the uranium trioxide content was liquid for its sodium oxide content. But the at- negligible. But whereas a synthetic total complex tainment of saturation or solubility equilibrium of can usually be prepared with almost any desired the solid phases with the liquid solution, which degree of accuracy by direct weighing of the compowould be followed through the uranium trioxide nents, the wet residue has to be analyzed for both content of the liquid, may well be a very much slower sodium and uranium, and it furthermore entails process, in view of the very low solubilities (ura- errors of handling that are difficult even to estinium trioxide concentration in the liquid) involved. mate. Unless, for example, the whole wet residue sample is dissolved and aliquoted for analysis, Calculation of Composition of Solid.-Direct analysis of solid phases, satisfactorily free of mother the portions used for the two different determinaliquor, was obviously impossible in this system. tions do not pertain to the same total composition. Centrifuging, and centrifuging followed by suction But after dissolving in acid, the sodium oxide filtration, gave nothing better than a smooth wet content cannot be determined by titration, but paste; and continued washing could not be used only gravimetrically or by flame photometry, and because of the incongruent solubility relations of the the precision thus possible is not sufficient for the distinctions in question. solid phases. The use of the total complex brings the second What was done, therefore, was to calculate the tie-lines fixed for each experiment by the composi- point of the tie-line close to the liquid composition, tions of the total complex and of the liquid phase, and it was, in addition, deliberately brought very and to consider the possible solid phase composi- close (which simply means that very little uranium tions on the basis of the behavior and disposition of oxide was used in the complexes), in the low alkali the tie-lines in the various regions of the system. region because of the slowness with which the solid The directions of the tie-lines were determined by phases came to their equilibrium compositions even algebraic extrapolation to the anhydrous base of the with the violent stirring used, and in the very high composition triangle of the system. The percent- alkali region because of the bulkiness of the solid age (D)of sodium oxide on the base of the triangle, phase formed and the difficulty of obtaining suffito which each tie-line extrapolates, is calculated cient clear liquid for analysis through settling. The from the data of Table I by the formula approach to equilibrium was slower the lower the sodium oxide concentration of the liquid, and most ( B C) D = B + (100 - A - B ) of the experiments dealt with solutions containing (A B) - (C L9 (1) less than 0.02% NazO. Actually, however, the in which U,the percentage of UOS in the liquid, may accuracy of the extrapolation is higher in this rebe neglected without error. The possible presence gion than in the region of higher concentrations, of nitrate to the extent of one or two parts per where more solid was used. In the higher range, thousand of the uranium trioxide used is also negli- the quantity ( B - C) of equation 1 is the difference gible in its effect upon this extrapolation. The of two large quantities, and a very high accuracy is quantities A and B used for the calculation, morerequired in the titrations for both B and over, were not in every case those actually listed therefore C. In the low alkali region, on the other hand, C is in the table. They were first corrected if necessary, for the evaporation occurring during the stir- so small compared to B , except in a few cases in ring of the complex up to the time of the analysis which extremely little solid was used, that the error of the final solution taken as representing equilib- of the extrapolation hardly depends on the relative rium. [This correction simply involved dividing B accuracy of the determination of the sodium oxide or ( A B ) by the factor (1 - t r ) , in which t = time content of the liquid, which was done by flame phoand r = weight fraction lost by evaporation per tometry. In this case the principal uncertainty is unit time; r = 0.3% per day a t 50°, and -1.4% per simply that resulting from the extreme slowness of equilibration of the solid phases. These relations day a t 7 5 O . I Theoretically, the accuracy of such extrapolation may easily be verified by substitution of the numerifor the solid composition depends on having the sec- cal values from the table. The number of deciond term of equation 1 small compared to the term mals recorded for D (% NafO on the NaZO-UO3 base B, and on the precision of the differences repre- of the composition triangle, or 70Nan0 in the solid sented by numerator and denominator of the frac- if this is assumed to be anhydrous), in Table I, is tion; geometrically, then, it depends on the dis- based on consideration of the uncertainties in the tance, in the usual triangular plot of the ternary analysis of the liquid, in the agreement of successive isotherm, between the two points fixing the tie- analyses (Table IA), in the time of stirring, and line involved, namely, the liquid composition and in the rate of loss of water by evaporation. Except the complex composition. For this reason it would for experiments in very high alkali concentrations, seem still better to use the analysis of a wet the scattering of the experimental points seen in residue to fix the second point. Many experi- Figs. 2 and 5, therefore, is not the result of analytical ments were made, in fact, by the wet residue difficultybut of incomplete equilibrium, or of a (parmethod,21 but they were all discarded. What tially) false equilibrium which persisted despite the would seem to be gained by the method was overbal- long and vigorous stirring and despite the use of anced by the much lower analytical accuracy with very small amounts of solid. (21) F.A H. Schreitirmakers. Z. phrsik. Chem.. 11. 75 (18'33). In a few special experiments of Table I, those

+

+

+

2124

JOHN

E. Rxccr A N D FRANK J. LOPREST

numbered 136(1-3), 137(1-4) and 139(1-2), the solid composition was calculated a little differently. The final liquid and solid compositions for these experiments are listed in Table I, in order to have the results together for comparison, but the data used in calculation of the solid composition are given separately in Table 11. These mixtures were based on the material remaining from a completed experiment, such as number 136 itself. Taking into account material removed for analysis from the original total complex, such as No. 136, a new total composition was prepared by addition of further small amounts of the components. Since the bulk of the original NazO and UOS was always in the solid phase, none of which, in most cases, was removed in the sampling of the liquid, the next total composition, such as No. 136(1), is known quite accurately with respect to the actual weights of UO3 and NazO in it, but the total weight of the complex is known only approximately. When solid phase was also removed in the preceding sampling, the amount removed was also taken into account. The data here therefore comprise the actual weights of Uos( = a ) and of NazO ( = b) in the total mixture, the (approximate) total weight of the mixture ( = c), and the percentage of NazO in the equilibrium liquid ( = d ) . Then the percentage of NaaO, on the anhydrous basis, in the equilibrium solid (= e) is calculated as e = 100 X

b

a

- d(c

-a

- b)/100

+ b - d ( c - a - b)/100

(2)

in which it is assumed, again, that the UOS content of the liquid is negligible, and, with altogether negligible, error, that the weight of the liquid phase is equal to (c - a - b ) . DATAFOR

No.

136 136(1) 136(2) 136(3) 137 137(1) 137(2) 137(3) 137(4)

139 139(1) 139(2)

THE

TABLE I1 SPECIAL EXPERIMENTS IN TABLE I (50”)

W t . of UOI in mixture, a

2.2314 2.2314 2.2314 2.2314 2.2310 2.2310 2.2310 2 2310 2 2310 2.6442 2 6342 2.6442

Total W t . of NazO wt. of in mixture, mixture, b C

0.02818 .03293 .03754 ,04224 .02818 ,03292 .03754 .04170 ,04658 ,02834 ,03309 ,03316

313 283 323 293 311 265 320 290 345 200 280 26 5

%

% NazO

NazO in solid,

0.00012 .00021 ,00049 ,00069 .00012 .00020 ,00044 ,00064 ,00081 .00008 .00013 .00018

1.23 1.43 1.59 1.77 1.23 1.43 1.59 1.75 1.92 1.05 1.22 1.22

in liquid, d

C

If a group of tie-lines radiates, practically, from the water apex of the triangle-as is obviously the case for all those involving liquids of very low alkali concentration-it is clear that they can give no information regarding the water content of the solid phase. ,4t the same time, however, these tielines give the actual ratio of NazO to u03 in the solid regardless of its water content. On the other hand, if a group of tie-lines does not radiate from the water apex, they can fix even the water content of the solid if they have a common intersection in the triangle; but in the absence of such an inter-

VOl. 77

section the value of D obtained by extrapolation does not represent the percentage of Na20 in the solid unless this is actually anhydrous. The corresponding values of C and D listed in Table I are plotted in Figs. 1 and 2 . In order to show the results for the low concentration region on one diagram the vertical coordinate in Fig. 2 represents the logarithm of C (yoNazO in the liquid) rather than the concentration itself. The simplest interpretation of the high concentration results seems to be that suggested by the lines drawn through the points in Fig. 1 ; and our interpretation of the low concentration region (Fig. 2 ) is shown separately, for clarity, in Fig. 3 . A horizontal represents an isothermally invariant solution saturated with two solid phases, and a vertical represents a range of solution concentrations saturated with a single solid of fixed composition and hence a solid “compound.” Along other curves the solid composition varies with the liquid composition, representing, therefore, saturation with a solid solution of a certain range. There are evidently then six solid phases, which will be distinguished as I for pure NaOH.Hz0 to VI for pure U03.2H20, with four intervening solid phases, 11-V, containing both NazO and UOS. -41though the scale cannot bring out the variations in the liquid compositions near the water apex, the complete isotherm is shown in the usual triangular fashion in Fig. 4. The relations are also shown schematically, without regard to scale, in the diagram of Fig. 4a. The most surprising of these solids is solid 11,saturating liquids of composition on the roughly vertical section c-d of Fig. 1. Figure 4 shows that this solid contains more than two moles of NaPO per mole of UOS. If point a, in Fig. 4a, represents the solubility of NaOH.H20in pure water a t 50°, the first solubility curve, a-b, for saturation with this solid in presence of U03 in the liquid, is too short for detection, and the horizontal b-c, in Fig. 1, represents the solution saturated with both NaOH.HzO and solid 11, point b,c in Fig. 4(a). The concentration of this liquid is experimentally indistinguishable from the solubility of NaOHeH20 alone, or 45.20% NazO, point a. The next horizontal section, d-e, represents liquid, with 42.78% Na20 as its average composition, saturated with both solid I1 and solid 111, which is clearly a solid solution covering a long range of coniposition. Between these invariant liquids or for concentrations between 45.2 and 42.S% Na20, the liquid is in equilibrium with a single solid (11),the composition of which must lie on the group of tielines (Fig. 4) joining the short solution curve c-d (Fig. 4(a)) and the base of the diagram a t the narrow region labeled “11.” These nine tie-lines, taken from points 6-14 of Table I, are not parallel to each other ; nor do they have a common intersection in the triangle. They intersect irregularly; but this is the result of the low precision in the extrapolated value of D for this region, because of the inevitable uncertainties in the handling and analysis of the practically saturated sodium hydroxide solutions involved in determining both the total complex composition and the liquid composition. For

PHASE RELATIONS IN

April 20, 1955

THE

SYSTEM SODIUM OXIDE-URANIUM TRIOXIDE-WATER 2125

L

G M

$39:

-4 I

1 60

l

l

l

l

8 6 4 2 0 % NaOl on anhydrous base of triangle ( = D). Fig. 3.-System NazO-UOa-HzO a t 50 ’: phase relations 011 basis of results in Fig. 2. 10

1

I

I

50

l

l

40

I

1

I

20

30

10

% NazO on anhydrous base of triangle ( =D). Fig. 1.-System NazO-UOa-HzO a t 50 O: tie-line relations for the high alkali concentration region.

5% Uo3. It seems likely therefore that this solid is hydrated, and probably rather highly hydrated. If the compound were anhydrous, the extrapolations of the tie-lines to the base of the triangle would indicate, with an average of 35.2% NazO, a 5 :2 mole ratio of NazO to U03, for which the theoretical percentage of NazO is 35.14. But since it is almost certainly hydrated, the NazO proportion must be still higher. For the mole ratios 3 :1 to 6 : 1 the following formulas all fall on the bundle of tie1 I lines : 3Naz0.U03.2Hz0 (or Na2U04.4NaOH), 4NazO.UO3.6.5Hz0, 5NazO~U03.11Hz0, and 2 -11 6Naz0.U03.15Hz0. The agreement for either of the last two formulas is quite good, but only the last, which may be written as NazU04.10(NaOH. HzO), contains sufficient water for the hydration of all the excess of NaOH over “Na2U04.” Extrapoz lated to the theoretical value of 30,8270UOafor this formula, the nine tie-lines there give an average -3 value of 39.9% NazO (average deviation, 0.5) compared to the theoretical figure of 40.07%. The intersection of the tie-lines with the line joining NaOH.HzO and Na2UO4 thus occurs almost %.&:;:;-; -4 I I I I I I I 1 6 ? ....... exactly (on the average) a t the mole ratio 10 : 1. Double compounds of “salts” with NaOH, or 1 2 1 1 1 0 9 8 7 6 5 4 3 2 1 0 compounds with excess of NazO relative to the % NalO on anhydrous base of triangle ( = D). “acidic” oxide, are rather rare. Among those reFig. 2.-System NazO-UOa-HzO a t 50 tie-line relations ported are 4Na20.A1203.10Hz0 ( = 8NaOH.AltOr for the low alkali concentration region. 6HzO),2Na20.BzO3.Hz0 ( = NaOH.NaBOZ), NaOH. 3Na&s04.HzO, NaOH.Na3AsO~~10Hz0 and NaOH, an uncertainty of only one part per thousand in each Na2C4H40g3Hz0 (tartrate). 2 2 The classical examof these two points fixing the tie-line the uncer- ple is perhaps “tetrasodium chromate,’’ Na4CrOb. tainty of the extrapolated value of D a t the base of 13HzOz3( = 2NaOH.Na2Cr04.12HzO). the triangle is about f1% NazO for a complex conIn contrast to solid 11, the next one, solid 111, a solid solution, ranges in color from bright orange a t taining -5% yo,. This solid I1 is bright pink in color, and along the its high NaaO end (point e) to plain yellow a t the section b-c of Fig. 1, with white solid NaOH.HzO low NasO end (point f ) ; and while also extremely also present, the color varied accordingly as a mix- finely divided i t settles more readily than does the ture of the pink with white. The pink solid is ex(22) All cited in A. Seidell, “Solubilities of Inorganic and Metal tremely finely divided, and of great apparent bulki- Organic Compounds,” D. Van Nostrand, N . Y.,Third Edition, 1940 ness. Although it settles to leave a clear superna- (Vol. I, PP. 1146, 1150), and Supplement t o Third Edition, 1953 ( p p . tant liquid, in some cases only l or 2 ml. of such 415, 429). (23) Ref. 22, Third Edition, p. 1255. For the phase diagram of the liquid was visible even after many days of standing system NazO-CrOa-HZO a t 30°, see F. A. H. Schreinemakers, 2 . p h y s i k . in a test-tube for a 50-g. complex containing only Chem., 55, 71 (1906).

I

1

2 0

- I

L,

I

I

O:

2126

JOHN

VOl. 77

E. RICCIAND FRANK J. LOPREST

(The value 6.5y0NazO for point f is estimated not from the logarithmic plot of Fig. 2 but from a direct plot of C against D on an appropriate scale for this region.) Since there is a gap in solid composition between compound I1 and the edge of solid solution 111,these results definitely rule out the existence, as a distinct equilibrium phase in the aqueous system at this temperature, of an anhydrous “normal” or mono-uranate, “Na2U04’’ (theoretical 7,NazO = 17.82; point u on Fig. 4),although such a formula, as already suggested, may be iiivolved as part of compound 11. If the solid solution I11 is anhydrous, then, it covers a t this temperature the NazO: UOa mole ratios from -S:11 (theoretical, l3.6achNazO) to -1 : 3 (theoretical, ti.7475 NazO), so that the “diuranate,” “NasUzO~”(with 9.787, NazO), is merely a point in this range of variable solid. If the solid solution is hydrated, however, it would cover a greater Pig. 4.-System Na20-UOa-HzO: tie-line relations a t 50 ’; range of NazO: UOa ratios in its composition. The u = “Na2U04,” v = Na2U04.H20or 2NaOH.U03. solid, however, may well be anhydrous. If i t is hydrated its high NazO edge would be some point on the limiting tie-line running to 13.3% NazO on the base of Fig. 4, and its edge might possibly be the formula “ N a 2 U 0 ~ ~ H 2 0although ,” the point representing this formula, with 4.92% HzO (point v on Fig. 4), lies just outside and to the left of the limiting tie-line for the solid solution, which intersects the KaLU04-Hz0 line a t 0.07, Hz0. If the water content of the solid solution is greater than 5-67, then its composition, according to its limiting tie-line, extends to Nan().UO3 ratios greater than unity. The lower limit of -1 : 3 for the mole ratio included in the solid solution (point f ) is independent of the water content assumed for it. It seems likely, all things considered, either that this solid solution is anhydrous or that it has only a simll water content. Although the precision along some parts of the LI:, curve e-f is not high, it is nevertheless clear that the ----4 solid solution is continuous between the limits e and I1 I11 IVV Fig. 4(a).-Schematic diagram of the phase relations a t 50”. f , with no sufficient evidence of any break, or invariance in the liquid coniposition, to suggest a discontinuity in the solid solution. The middle part of compound 11. Along the section d-e of Fig. 1, the the curve is very steep (Figs, 1, 2 ) , so that over a mixture of these solids showed the expected varia- considerable range of sodium oxide concentrations tion of color from distinct pink to distinct orange in the liquid the solid phase composition is almost and an immediate improvement in settling as the constant, near -lo.@’, NazO; but this does not solid solution became part of the mixture. The correspond to any very simple mole ratio of the change between the solids is reversible. The com- oxides. In terms of Fig. 4, this means a tendency pound I1 is very incongruently soluble, being sta- for the tie-lines in this region to converge a t a poitit ble only in contact with highly concentrated alkali. on the base of the diagram. This possibility of a Dilution with water rapidly changes the color from practically invariant composition a t some point in pink to orange, indicating hydrolytic decomposition the solid solution range vanishes, however, if the of the compound to the solid solution 111; the reverse change, by addition of concentrated NaOH solid solution is hydrated, for then its composition to the yellow or orange solid, is much slower but lies somewhere above the base of the triangle. (Figure 4 shows only a few of the actual tieeasily observable. Solid 111, and only a The long curve labelled e-f (Figs. 1, 2 , 3) repre- lines for the region Liquid single token tie-line for each of the subsequent resents the major part of the phase diagram (Fig. 4). Solid IV and Liquid Solid v.) Here the liquid, with alkali concentrations ranging gions, Liquid The solid solution I11 is the sole saturating phase from 42.8% NazO (point e) down to 0.01% NazO (point f ) , is saturated with a single continuous solid for liquid compositions ranging from point e to point phase of variable composition. If this solid is an- f, where the curve (Fig. 2 ) again becomes horizonhydrous its composition (values of D in Table I , or tal, indicating the appearaice of an additioiial solid the region “111” a t the base of Fig. 4) runs from phase, solid IV. The hori7ontal f-g (Fig. 3) there~-l:3.,?7~ NazO (point e) to -6.5% NazO (point f ) . fore represents a third isothermally invariant soluI

-

+

+

+

April 20, 1923

PHASE RELATIONS IN THE SYSTEM SODIUM OXIDE-URANIUM TRIOXIDE-~VATER 2127

tion, saturated with the limit f of solid solution I11 successive small increments of very dilute NaOH and with solid IV. The concentration of sodium solution. The original suspensions had the sodium oxide in this solution, 0.0106~0,is averaged from content of the water used, namely, ~ 0 . 0 0 0 0 5 % points 66-85 of Table I, with an average deviation NazO. Then through seven successive additions of 0.0012. The range of solid composition (mixture of NaOH solution in experiment 13G, for example, of two solids) in equilibrium with this liquid is from the analysis gave repeatedly 0.0001270 NazO (in each case also constant with time), up to and in6 5 % NazO (point f ) to 3.5% Na10 (point g). Since the section g-h (Figs. 2 , 3) is fairly vertical, cluding the total composition for the mixture rethe solid IV which saturates these liquids (between corded in Table I1 as No. 136. A t this point, 0.0106 and 0.00058% NazO) appears to be one of then, the solid, with 1.23% NazO, was a mixture fixed composition rather than a solid solution. of U03.2Hz0 and a solid solution of the composiPoints 82-119 of Table I may be attributed to this tion of point j. With the next addition of sodium section, and the average solid composition given by oxide, to give the total composition of complex these points is 3.49% NazO, with an average devia- No. 13G(l) of Table 11, the equilibrium concention of 0.18 and with extremes of 3.9 and 3.0. The tration of sodium oxide in the liquid began to rise, composition is therefore very close to the formula and the percentage of sodium oxide in the solid also Naz0.6U03, with a theoretical sodium oxide per- rose, so that we were now dealing with curl-e centage of 3.49; the theoretical values for the 1:5 i-j of Fig. 3. Further additions of sodium oxide, and the 1: 7 formulas are 4.16 and 3.00, respectively. with equilibration and analysis, gave the subseThere is a slight trend in the solid compositions, quent points of the series, 136(2) and 186(3). Simiwith higher values near point g and lower values lar procedures and observations are involved in near point h, but in view of the scattering of the the series numbered 137 and 139. Relations at 7Y.-A portion only of the 75" isopoints, the solid may be taken to be either a pure, definite compound or a solid solution of short range therm of the system was investigated, for the purpose of verifying the relations in the low alkali on either side of the 1: 6 compound. As may be seen from Fig. 2 , the data for the hori- region, with the expectation that heterogeneous zontal section h-i of Fig. 3 are rather poor a t this equilibrium might be attained more rapidly a t the temperature, the attainment of equilibrium in this higher temperature. The higher alkali region, inregion being persistently slow. The simplest inter- volving the solid phase 11, the pink compound, was pretation, however, is that this is a horizontal sec- not attempted a t 75", and whether this solid pertion representing a fourth isothermally invariant sists as an equilibrium phase a t 75" is not known. For the part of the i s o isotherm studied, the data solution, now saturated with solid IV and with the limiting composition of another solid solution phase, and results are listed in Table I11 and plotted in a t point i. The 24 points 113-137(2) of Table I Fig. 5 . The information given in Table I11 is simigive an average of 0.00058~0NazO as the concen- lar to that in Table I. The liquids for experiments tration of this liquid, with an average deviation of 1-6 were sampled by centrifuging and titrated for 0.00014, while the solid composition (mixture) sodium oxide content. All others were sampled ranges from 3.5 to -1.7 or 1.8y0NazO (point i). by ultra-fine filtration and analyzed for sodium by The next section of the curve of Fig. 3, or i-j, repre- flame photometry. The actual successive analyses sents saturation with a solid solution, solid V, con- made in testing for equilibrium are listed separately taining from 1.7 or 1.8 to 1.1 or 1.2% NazO, or in Table III-l.20 The procedure followed for coverina the Na20: UOX mole ratios between -1 : 12 the special experiments 17(1), 18(1), 23(1), 22(1), and -?:18. The interpretation of the section h-i as a horizontal for +q saturation with solids IV and V, although somewhat questionable on the basis of the 50" data alone, is .: strongly corroborated by the much clearer relations observed a t 75" 2 0 (Fit. 5 ) . I/ ' ginally, the horizontal j-k represents the last isothermally invariant -1 solution, that for saturation with 2 solid V (point j) and U03.2H20 (solid .= --2 (VI)). The last six points of Table I give 0.00012% NazO as the concen- 2 tration of this solution, with an aver- 8 -3 age deviation of 0.000013. (There 4 must also be a curve "k-1," of course, -4 for saturation with U03.2H20 alone, but this, like that for NaOH.H20, -5 -I was not detectable; Fig. 4(a).) The 12 10 8 tj 4 0 actual constancy of the liquid coni% XazO on anhydrous base of triangle ( = D). position in this region was observed by starting with suspensions of UO:j. Fig. 5 -System Na90-UO3-H20 at 7 5 " : tic-line relatior~sfor thc low alkali 2 H z 0 in pure water and then adding region. 1

-

+

'5

c

212b

E. I ~ I C C. ~I N DF R . ~ NJ.KLOPREST

JOHN

24(1,2) and 25(1,2,3) was the same as that already

explained for the corresponding at - experiments -

.io"

TABLE 111 T H E S Y S T E M S a 2 0 - ~ 0 3 - H 2 0 AT i'*itni>r,sition of

'-1

75

weight percentage.)

(Jii

complex R

n

C (;

\

1,;

r; S a n O

UO,

27,204 18.744 3,962 0.6434 ,2762 ,1908 1. ~ R O X ,1582 11974

1

9,7X.i

12..558 5,322 1.9609 1.9439 1.9897

:i

1 )

(i

i

x

IO 11 12 1.i

in liquid

4 3 3 43 23 76

70 179

29.31 20.12 5.682 0.4.56 ,0914 ,0268 ,0083 ,00714

?lap0

in "sdid"

10.8 10.6 10. 6 9.57 8.81 7.69 7.02 6 . G!I

+ solid I11 + solid I\'

Liquid 1 ,9778 1 ,9891 1 .9X.%

11

5.: NaiO

+ solid I11 (solid soln.)

Liquid 2

Ilr. stirrini:

0 . 1381

. 181G5 K255 ,10191 ,08967 ,08083

45 104 65 63 70 04

0 0038

.00826 ,00358

6 . 37 t i . 07

,00412

5.79 4.69 4.20 3.74

10

Liquid solid I V (Sa?O.GUOa) 2.0150 0.07450 48 0.00214 2.0107 .07201 90 ,00023

3.49 3.45

17 18 19 17(1) 20 21 19(1) 23( 1) 24(2) 2,5(3) 22

Liquid solid IY solid \1,9817 0.06916 40 (I. 000105 2.0224 ,06823 42 ,000113 88 ,000136 1.9605 ,05613 41 000126 2.0001 ,04832 63 .000178 2.0076 ,04814 42 .000168 41 .000113 66 .000153 85 .000140 ti7 ,000168 1 ,9741 ,03061 70 ,001)l:$l

,

L9W8 1 96;il 1.9771

11

00443 ,00358

+

1 .i

+

-

+ solid 1- (solid soln.) 18 .0001u2 + solid V + solid \.I (UOs,2H20)

3.37 3.26 2.78 2.37 2.35 2.34 1.94 1.75 1.74 1 55 1.52

Liquid 21(1)

Liquid 25(2) 23 22(1) 24 2 q 1) 23

I)ATA

2.0279

0.02049

1.9860

,01851

11 68 90 44

1.9724

.00751(i

11

.48

FOR THE

a

17(1)

3.2G6 2.8133 3.0362 2.636 2.9519 2.958 2.9519 4.988 2.951'3

19(1)

23(1) 24(2) 25(3) 24(1) 23(2) "(1)

2,3(1)

1.16 1.00 0.94 92 C r

. i i

:3 8

TABLE ITI N TABLE111 (75') SPECIAL EXPERIMEXTS

Wt. of UOa in mixture S O .

0.000061 .no0056 .000058 ,000068 .000041 ,000055

I .?I1

Total Wt. of Sa?O w t . of in mixture mixture b r

Yc S a x 0

in liquid d

0.0794G ,05583 .05425 ,04673 ,04656

115 101 145 129 114

0.0001 26 ,000113 ,000153 .000140 ,000168

,03927

139

.oonio3

,03178

121 1.57 139

.000001 on0068

,04748

,02301

,000041

'.; N a r O in solid

e

2,317 1 ,94 1.75 1.74 1.55 1 . : ~ 1.16 0.94 0 77

1.01. 77

(Table 11), and the data are given separately in Table IV. With respect to the sequence and nature of solid phases and invariant solutions, the results for this part of the system a t 75" are seen to corroborate the conclusions drawn from the more widely scattering results a t 30"; and the solids will therefore be identified by the same numerals as for the 50" isotherm. Solid 111, saturating the liquids along the curve e-f, is again a solid solution, with its h i t f still at -6.57, XazO. Solid IV is again clearly very nearly the compound Naz0.6UOs (theoretical, :3.49% NazO), and solid V is close to 1.3% Na20 or -16 UOa per SazO. The concentration in the liquid a t each of the three invariants studied is lower than a t 50". For the liquid f-g saturated with solids I11 and IV, points 9-16 of Table TI1 give an average value of 0.00:38f': NazO, with 0.00034 as average deviation. The eleven points 17-22 give, for liquid h-i, saturated with solids 117 and V, an average of 0.00014(;, NasO with average deviation of 0.000020. For the last invariant, j-k, saturated with solids V and VI, the last six points of Table 111 give 0 . 0 0 0 0 S i ' ~ NapO (average deviation, 0.000006), which is just barely over the observed reading for the distilled water used. That the terminal solid V I is CO:l.L)H20was determined by direct analysis. Solid rotated with water for two months at 25" was air-dried to constant weight and then assayed as U,O, a t 1000", giving S9.06 and 8S.99yo UOa. Solid stirred violently with water for i 2 hours at 7 3 " , and dried rapidly by washing with alcohol arid ether, gave, by similar assay, 8S.61 and SS.CC;S;I, UOs. The theoretical value for U03.2Hz0 is SS.Sl% UOa. The various solid phases from 111 t o VI differed little in apparent physical properties. The color of the principal solid solution, solid 111, varied from yellow to bright orange with increasing sodium oxide content. \Vith less than about Of-; NazO in the solid it settled rather. poorly from the mother liquor, which then contained less than 0.1'; NazO. In the higher concentration range the solid settled m-ell enough to permit sampling of the liquid by celltrifuging. The remaining three solids, IV, IT, VI, were all yellow; solids V and 17 settled fairly well but solid IV, the compound Naz0.6UOa, seemed to be the most finely divided, settling extremely slowly. Long centrifuging and even weeks of standing failed to give a clear liquid, which was obtainable only by ultra-fine filtration. The presence of this solid could almost be established visually, in suspensions of mixed solids, by this characteristic. The solids in experiments 53, %(I), and 25(2) a t i 5 ' , for esaniple, settled fairly well, but when the total conipositiori was changed to 25(3), to bring it from the invariant j-k, for solids V and VI, to the invariant h-i for solids I V and V, the change was visible, in agreement with the result of analysis. Similar observations were made, in reverse, in going from experiment 22 to 22(1), also at i.5". The final numerical relations for the solid phases and isothermally in\-ariant solutioiis for both teinpcratures may he suininarizctl as follows

April 20, 1935

SEPARATION OF EUROPIUM FROM SAMARIUM BY ELECTROLYSIS

2129

ple "mono-uranate," or Na20.UO3, a t least as an equilibrium phase in the aqueous system, despite the formation of the unexpected solid I1 with a much higher NazO:UOa ratio. As for the familiar "diuranate" formula, NazUzOi, or Naz0.2U03, we see that in these isotherms a t least, this is merely a point in the continuous range of compositions comprising the principal solid solution 111, although it may be that such a compound becomes a distinct phase in itself a t much higher temperature. In some ignitions of wet mixtures of NazO and excess 500 760 UOa, made for the purpose of testing the result for 45.2 ( 5 8 .5)z4 L + I possible use in analysis of wet residues, it was obL + I + I I 45.2 served that the weight of the residue obtained a t 1000° could be accounted for best (but roughly) in terms of a mixture of NazUz07and U308. On the basis of the present results, then, all the At the higher temperature the lower limit of the formulas of sodium uranates previously reported solid solution I11 appears to be about the same as a t (references 1, 4,9, 10, 11) through the direct or in50". Solid IV remains close to the formula NazO. direct analysis of precipitates obtained from aque6U03 in composition, with very little, if any, tend- ous solutions, would have to be explained as repreency to take up adjacent solids in solid solution. senting portions of a solid solution, mixtures of The composition spread of solid V is possibly be- equilibrium solids, or decomposed (hydrolyzed) coming narrower with increasing temperature. residues of these. Thus the reported NazO:U03 The fact that the sodium oxide concentrations of mole ratios 1:2 , 9 : 20, 3 :7 and 2 :5 would represent the invariant liquids involving the solids 111-VI are regions of the solid solution I11 ; the 'icompounds" lower a t 75" than a t 50" suggests that the actual with ratios 2 :7 and 1:4, if they were equilibrium solubilities of the various solids are decreasing with solids, would be mixtures of the solids I11 and IV; rising temperature. A solid of approximately the and the "compounds" 1:7 and 1: 8 probably reprecomposition Naz0.16U03 is formed a t a sodium hy- sent solid IV, or Na20-GU03, partly decomposed droxide concentration of -4 x M a t 50' and by washing into a mixture of solids IV and V. From -2 X AI a t 75"; and the solid Naz0.6U03 the behavior encountered in the present direct study requires NaOH a t -2 X M a t 50" but only of the system, and from the phase equilibria established, it may be said that it is highly improbable 4.5 x 10-5 M a t 750 Although the solid phases found here to be the that equilibrium could be expected to hold in ordiequilibrium solids in the aqueous system are not nary precipitations such as those used in attemptnecessarily the same as those to be expected in the ing to determine the compositions of the solids by anhydrous system Na20-U03 a t high temperatures, direct analysis or those occurring during the titrathe results do seem to rule out the existence of a sim- tion of uranyl salt solutions with alkali. Solids I, NaOH.H20 a t 50'; KaOH a t 75" 11, hydrated compouni with probably 5 r: 6 moles of NalO per UO3, a t 50 ; not studied a t 75 111, solid solution wit! BazO:U03 mole ratioso from -8: 11 t o -1 : 3 a t 50 ; same lower limit a t 75 IV, compound, probably anhydrous, Na20,6UOa, a t 50" and 75'; possibly with slight solid solution on either side of this formula LT,solid solution with XazO :UOs Tole ratios from -1 : 12 to -1 : 18 a t 50'; -1 : 16 a t 75 VI, UOa.2H20, from 25 to 7 5 " . Invariant liquids (yoNazO)

NEW YORK,N. Y.

(24) Seidell (ref 22), Vol I , p 128-1; by interpolation.

[CONTRIBUTION FROM THE LOS

ALAMOSSCIENTIFIC LABORATORY]

The Separation of Europium from Samarium by Electrolysis1 BY E. I. ONSTOTT RECEIVED OCTOBER28, 1954 Europium is eficiently separated from samarium by electrolysis of the europium citrate complex ion at a lithium amalgam cathode. Samarium oxide containing 1.6% europium oxide is quantitatively freed of the europium with one electrolysis. Under optimum conditions, 99.9% of the europium is removed into the mercury phase, while less than 1Oyoof the samarium is removed. Parameters affecting the electrolysis are discussed.

The separation of europium from samarium is not difficult, since europium is easily reduced to the divalent state.* However, an efficient method of separating a small amount of europium from a relatively large amount of samarium has not previously been reported. Marsh3 has used sodium amalgam to separate europium from samarium, but his method depends

on the selective oxidation of samarium after extraction of both samarium and europium. Several extractions are necessary in order to obtain pure samarium. The ion-exchange separation of europium from samarium is possible, but not p r a ~ t i c a l . ~ McCoy5 previously has shown that the mercury cathode separation of europium from samarium is practical. Actually the electrode employed by Mc-

(1) Work done under the auspices of t h e Atomic Energy Commission. (2) H. S RlcCoy, THISJ O U R N A L , 5 7 , 173H (1935); 59, 1131 ( 1 9 3 7 ) ; I,, F. Yntema, ibid , 62, 2782 (1930). (3) J. R. hI;rrsh, J . C/ieii%.S u c . , 398 (1972); 531 (I!Jl3).

(4) B. H. Ketelle and G . E. Boyd, THISJOURNAL. 69, 2800 (1947); F. H.Spedding, E I Fulmer, J. E Powell. T.A . Butler and I. S. Yaffee, ibid., 7 2 , 4860 (1950); S. W. Mayer a n d E. C. Freiling, ibid., T 6 , 5647 (1953) ( 5 ) H. N. RfcCoy, ibid., 63, 3432 (l!)&l).