March, 1957
NOTES
and to extrapolate the absorbances back to the values they presumably would have been before any decomposition occurred. It is felt that the uncertainty in this procedure is well under one per cent. Solutions of pH e ual to 6 and 14, containing the triazine virtually completdy converted to the molecular and anion forms, respectively, were prepared and measured in a similar manner. The absorptivity of the molecular form is about 3.3 X lo8 liter mole-' cm.-1 and the corresponding value for the anion is about 2.7 X 10' at 390 mp. It was found that the addition of small amounts of sodium sulfite to the solutions before adding the triazine resulted in improved stability, robably due to the reduction of oxygen, and t8hatthis dignot significantly affect the observed values of K . The absor tion s ectra (approximate, because of instability) of tffe two Lrms are shown in Fig. 1.
Acknowledgments.--It is a pleasure to acknowledge that this research wm supported through the partial use of grants from the Research Corporation and from E. I. du Pont de Nemours and Co. We are indebted to Mr. Herbert Richardson for the preparation of the triazine.
0.G
0.2
350 370 390 Wave length, mp. Fig. 1.-Absorption spectra of the two forms of 3hydroxyl - 1,3 - diphenyltriazine: concentration about 2.5-lo-' M . Curve 1, pH 6,all triazine as molecular form; curve 2,pH 14,all triazine present as anion form.
Results Table I gives the values of K found a t various pH values. The quotient was calculated by means of the formula K = (H+)(A - A&)/(Ab - A ) where A is the observed absorbance of the solution containing both forms, A , is the absorbance of the solution containing the triazine completely in the molecular form, and A b is the absorbance of the solution containing the triazine completely in the anion form, all solutions having the same total concentration,
379
THE SOLUBILITY OF SODIUM HYDRIDE IN SODIUM BY D. D. WILLIAMS,J. A. GRANDA N D R. R. MILLER Chsmislru Diuieion, Naval Rsaearch Laboratory, Waahinglon. D. C. Recaivcd August SO, 1986
As a continuation of the study of impurities in alkali metals' the solubility of sodium hydride in sodium has been determined. This system has implications bearing upon corrosion data interpretation from the standpoint of both direct reaction and apparent oxide content correlation. This impurity in a sodium system can come from two principal sources, hydrogen-sodium reaction (impure cover gas), and as a product of the reaction between sodium and sodium hydr~xide.~J Previous work at NRL2 established two facts pertinent to this study: that sodium hydride resulted from the Na-NaOH reaction a t temperatures above 325", and that the hydride so produced was stable when in solution in an excess of either primary reactant. Stability in NaOH has been shown by Gilbert,4 and mutual solubility of sodium-aodium hydride has been reported by Banus, et al.' The stability in solution results in a lowering of the dissociation pressure of the sodium hydride. Thus, if a sodium system is heated, under other than high vacuum conditions, in order to rid it of any contained h y d r o ~ i d e , ~sodium .~ hydride is introduced as an impurity. A sample drawn from such a system for oxygen analysis, without accounting for the contained hydride would result in erroneous values. The solubility of sodium hydride in sodium was, of necessity, determined by an indirect method, based upon the lowering of the dissociation pressure2s4 of this normally unstable material.2b7.8 Thus, a t a given temperature, a mixture of sodium and sodium hydride will exhibit a true dissociation pressure only so long as the condensed portion of the system consists of two phases: saturated sodium and solid sodium hydride. During stepwise temperature increases the first incremental increase in temperature which failed to exhibit a corresponding increase in pressure indicated the temperature of saturation as the lower of the two temperatures.
TABLE I Experimental DISSOCIATION QUOTIENTO F ~-HYDROXY-~,~-DIPHENYIATRIThree techniques, with different apparatus, were used to AZINE establish the curve for the solubility of sodium hydride in PH K x 1019 PK sodium. Low temperature (240-300") runs were made by 11.30 2.63 11.58 (1) D. D. Williams, J. A. Grand and R. R. Miller, in preparation. 11.40 2.79,2.51 11.55,11.60 (2) D. D. William8, NRL Memorandum Report f33, 1952. 11.45 2.80,2.89 11.55,11.54 (3) A. Klemenc and E. Rvetlik, 2 anorg. Chem.. 269, 153 (1958). (4) H. N. Gilbert, U. 5. Patent 2,377,876 (June 1045). 11.50 2.94,2.85 11.53,11.55 (5) M . D. Banus, J. J. McSharry and E. A. Sullivan, J . A m . Chem. 11.55 2.83 11.55 Soc., 7'7, 2007 (1955). 11 .49,11.52 11.60 3 .23,3.05 (6) J. D. Noden and K. Q. Bagley, Culcheth Laboratories, Tech.
These results give an average value of K = 2.85 (10.14)X 1O-I2.
Note #80,1954. (7) F. 0 . Keyes, J . Am. Chem. Soc., 94, 779 (1912). (8) A. Herold. Compl. rend., 228, 686 (1049).
NOTES
380
Vol. 61
expoain sodium to an excess of sodium hydride and sodium Both of the above methods suffer from th? presence of oxide, dtering, and: ( 1 ) analyzing the filtrate b amalgama- sodium oxide in the samples, and they are limited to a low tion and titration, and (2) analyzing the filtrate l y controlled temperature range because of accelerated attack upon the dissociation in a calibrated volume. Higher temperaturea glass systems at hi her temperatures. Furthermore, both (300-445') were investigated by both absorption and methods tend to yidd low results by reason of the temperadesorption of hydrogen in an ap aratus which provided for ture changes involved in the filtering and coolin steps. simultaneous recording of P-V-8 data. As the saturated sodium cools, solid sodium hydrife sepal The low temperature apparatus consisted of two 50-ml. rates and dissociates to some indeterminate extent dependPyrex flasks joined by an inverted U shaped piece of tubing. ing upon run temperature and cooling rate. Thus, some One Aask was equipped with an inlet port and was used hydride is irretrievably lost. The bulk of the hydride soluaa a still for introducing pure sodium into the second flask. bility data, therefore, was gathered by the indirect method of This latter flask was fitted with an inlet port, thermocouple hydrogen evolution and absorption. well, vacuum manifold outlet, and a sintered glaas filter The a paratus for absorption and desorption studies conbelow which was a thin-walled ampule of about 3-ml. ca- sisted o?a metal saturation pot attached b means of a metal pacity. Sodium hydride (98%) waa placed in the satura- ball joint to a Pyrex system which incruded a 1,260-cc. tion chamber and metallic sodium in the distilling pot. calibrated volume, a mercury manometer a n d l a vacuym The system was purged by evacuation and the sodium was manifold connection. A stopcock between the saturation distilledinto thesaturation chamber. Heat waa applied to the ot and the calibrated volume allowed for the isolation of the Na-NaH mixture and the filter until a desired temperature Bot from the cold part of the system The volume of the was reached. The melt was allowed to soak at this tempera- heated saturation pot was about 1 % of'the calibrated volume ture for five minutes. (Attack on the glass system, by the which wae at ambient temperature. sodium, also saturated the melt with sodium oxide.). NiFor desor tion, known wei ht,s of sodium metal and sotrogen pressure was then admitted, forcin the saturated dium hydri& were added to &e purged and nitrogen-filled metal into the sample bulb, which was &en sealed off. saturation After evacuation and isolation from This sam le waa then anal zed in the apparatus described the vacuumchamber. manifold, the temperature of the eaturation in the $RL modification K of the Pepkowitz and Judd chamber was raised with the sto cock between the pot and method for oxy en in sodium. The total apparent oxide the volume closed. h e n the desired temperaalkalinity thus h e r m i n e d was corrected for that due to turecalibrated reached, the sto cock was opened slowly and excese oxide saturation,l the remainder being alkalinity due to so- solid WM sodium hydride alrowed to dissociate, the hydro en dium hydride. The presence of traces of sodium hydride passing into the calibrated volume. When no further [ywae confumed during analysis by the evolution of gas drogen evolution was apparent, the pressure and tem erabubbles when the amalgamation residue was hydrolyzed. ture of the system were checked against the known &BOFor some runs, this apparatus was modified by replacing ciation pressure of sodium hydride.TJ If the pressure on the sample am ule with a calibrated volume receiver and a the eyetem waa lesa than that specified by these references, manometer. h t e r saturation and filtration, the analysis then the temperature of the saturation ohamber was slow1 for hydride WM made by heating the aample under vacuum. lowered the P-T relation for the eystem agreed w d The tem erature and preeeure of evolved hydro en were referenceuntil values. If, on the other hand th? fist check renoted. !he amount of sodium hydride was tbus c%culable vealed that the P-T relation checked the literature values, Bample weight was determined by reaction with alcohol and then the temperature was raised until an incremental intitration, creme did not result in a corresponding increase in the pressure. These points of conformity to sodium hydride diesociation date were taken as the eaturation temperatures for I the varioue runs. The antici ated range could be redictad, end waa shifted by varykg the origind charge TO CONVERT TO WT HYDROGEN dium hydride. 4.0 MULTIPLY BY 0.0417 The hot eaturation pot was isolated, and the hydrogen in the calibrated volume removed by evacuatlon. The hyI drogen e uivalent to the remaining soluble hydride was then recover88 and metlsured by repeated evolution into, and evacuation from, the calibrated volume. Evolution and measurement were always made with appro date isolation of the hot and cold parts of the system so !tat the volume measured for removal was always at ambient temperature. This hydrogen removal could be done at saturation temperature because aodium hydride in solution still exhibits a dissociation pressure ( a few millimeters a t 400") which is sufficient to effect complete recovery, although the equilibrium pressure becomes lower as the solution becomes more dilute. A second means of dissociating the soluble h dride is to cycle the melt through a temperature range o 300350' with intervening evacuations. Both methods were used and were equally effective. The isolation stopcock between the saturFtion pot ?nd the calibrated volume was closed during the Initial heating period EO as to minimize the amount of hydride required to satisfy both metal saturation and equilibrium system pressure. Since the apparatus was not designed to operate at pressures greater than one atmosphere, solubility determina1 .o tions by this method were made only to 425'. In the absor tion studies, the same apparatus was used. It was chrtrgecf after purging, with sodium metal and hydrogen gas. The temperature of the esturation pot was raised to the various desired temperatures and the hydrogen pressure dro was recorded until no further absorption was indicated. &his method was much slower than the desorp0 tion study, but did allow for higher temperatures to be used. 200 300 400 500 In early runs, the reaction pot was constructed of nickel, but diffusion a t temperatures above 400" necessitated the subT ("(2.). stitution of a 304 stainless steel tube. Fig. 1.--The solubility of sodium hydride in metallic The data from all runs are shown graphically in Fig. 1. sodium. -_The agreement of the data from the different types of runs lends credence to the indirect approach to the s o l u b W 10) D. U . Williams and R. R. Miller, And. Chem., 28, 1865 (1951).
1
Yo
OFSO-
P
March, 1957
NOTES
problem. A brief discussion of the deviations may be in order, however. It may be seen that the absorption data tend to run somewhat hipher than the desorption data. The reason for this is the limited formation of undissolved hydride. The initial H Zpressure during the absorption runs was between 600 and 650 mm. Thus, at temperatures below 4000, the melt absorbed sufficient Hg to saturate itself and, in addition, enough to coat the surface with a thin layer of sodium hydride. Since the system was static, this layer prevented further hydride formation,lO and accounts for the slightly hi her solubility figures re orted. %he wide deviation of t i e point at 265' is explained by the fact that, at this temperature, the melt probably was not saturated with sodium oxide, and subsequent correction for anticipated saturation by this material resulted in a low hydride value. The 240' figure was not corrected for oxide, since no attack was in evidence and no alkalinity was found. This solubility study was limited by the va or pressure of sodium metal and the dissociation pressure ofsodium hydride. Both become significant to these data at 425'. The desorption apparatus would not allow for operation at pressures In excess of one atmosphere, and excemve distillation of sodium vitiated higher temperature results in the ahsorption runs. Data were collected a t 500', but were not mfficiently reliable to be included in this report. There were indications, however, that the system was approaching a limiting solubility of approximately 5 wt. % NaH (under about 1 atmosphere HZ pressure). Hence, extrapolation of the data herein reported is not recommended. The results of this study demonstrate a possible source of error in present sodium analysis for oxygen by the amalgamation method. The error with respect to sodium hydride, a8 well as the likelihood that it will be present in most sodium systems, is particularly significant. The error will not necessarily be predictable in a sample drawn directly from saturated sodium unless precautions are taken to prevent dissociation of hydride during sample handling and cooline. Excessive local heating during amal amation in the oxide analysis method can also result in hyfride loss.
mechanism were the correct one. there should be a marked hydrogen isotope effeci and a significant decrease in reaction rate in D O 4 In the hope of deciding between these mechanisms, rates in H 2 0 and I920 solutions were determined and compared at the same 02,Pu(II1) and free SOU- concentrations. Only a moderate hydrogen isotope effect was observed. The solubility of oxygen in DzO was determined so that a correction for the concentration difference in the two solvents at the same partial pressure of oxygen could be applied. A known volume of water was saturated with oxygen, transferred to a vacuum line and distilled into a large bulb chilled with Dry Ice. The liberated gas was determined manometrically. For a partial pressure of O2of 1 atm. the solubility was found to be 1.28 X M in HzO and 1.41 X M in DaO. Three determinations in each solvent were made, the mean deviation was about 2.4%. No measurements have been made in salt solutions, but it is probably safe to assume the same solubility ratio and that 02 is about 10% more soluble in DnO solutions than in HzO solutions. The acid dissociation quotient of DSO4- in DaO solutions was needed for the calculation of the free Sod- concentrations in those solutions. It WM determined by making optical measurements on the Ce(II1)SOr system. Absorbancies were determined for a series of Ce(II1) solutions which were 0.012 M in HClO, and with Na2SOd ranging from 0 to 0.05 M . In these solutions, the acid concentration was low enough so that an approximate value for the acid dissociation quotient was sufficient for the determination of free sulfate concentration as a function of absorbancy. Using the same Ce(II1) concentration, absorbancies were determined for a series of solutions all 0.05 M in NaaS04, but with HC104 concentrations ranging up to 0.11 M . Free Sod- concentrations were estimated from the absorbancy vs. SO,- function and the acid dissociation quotients were calculated. The results of such determinations in both H20and DZO solutions of unit ionic strength are given in Table I. The fact that the values obtained in the
(IO) D. T. Hurd, "Chemistry of Hydrides," John Wiley and Sons, New York, N. Y., 1952, p. 31.
THE EFFECT OP DzO ON THE RATE OF THE REACTION BETWEEN OXYGEN AND Pu(II1)' 1 %P.~B. BAKERA N D T. W. NEWTON University of California, Los Alamos Scientific Luborolory, Loa Alamos, New Meaico Received Auouat $7, 1966
It has hecii shown recently that the reaction IEtwecn I'u(II1) and oxygen in aqueous sulfate solutions is in termolecular: first order in the oxygen concentration and second order in the Pu(II1) concentration.2 One possible mechanism would involve an activated complex in which two Pu(II1) sulfate complexes were linked by means of an oxygen bridge; this structure would be similar to that postulated for one of the Pu(IV)-H202 complexes.a Another possible mechanism would involve the transfer of a hydrogen atom from the hydration shell of each of two Pu(1II)SO~-complexes to the oxygen rnolec~le.~If a hydrogen atom transfer (1) This work was done under the auspices of the U. S. Atomic Energy Commission. (2) T. W . Newton and F. R. Baker, J. Phwa. CAem., 60, 1417 (1956). (3) R. E. Conniok and W. H. McVey. paper 4.12 in "The Transuranium Elenients." editcd by G. T.Seabora. ,J. J. Katz and W. M. Manning, "National Nuclcar Energy Seriea," Division IV. Vol. 1 4 8 , McGraw-Hill Hook Co., Inc.. New York. N . Y.. 19.19. (4) This was suggested to us by Professor Norman Davidson, ROC also R. E. Huffman and N. Davidson, J. Am. Chem. SOC.,78, 4Y3U
(1956).
381
TABL I ~ DETERMINATIONS OF THE ACID DISSOCIATION QUOTIENTOF THE BIRULFATE IONIN SODIUM PERCHLORATE SOLUTIONS OF UNITIONIC STRENUTH AT 25' Fraction DaO
NasS04, No. of M aoln.
0.05
0.892
.05
5 4
0.876
.05
5
0
" 3 0 4
range.
M
0.05 t o 0 . 1 1 .05 to .I1 . 0 3 t o -11
Acid dissociation quotient
0.091f0.006 .052f .004 .055f ,003
H2O solutions are in good agreement with those obtained potentiometrically6indicates that the optical method for determining the acid dissociation quotients is valid. Assuming that the effect of D20 is linear in its concentration, the data lead to the estimate that the acid dissociation quotient of DSOd- in pure DzO solutions a t 25" and unit ionic 0.003. The ratio of acid disstrength is 0.051 sociation quotients found here is in substantial agreement with the ratios which have been re-
*
(5) E. Eiohler and 8. W. Rabideau, ibid., 77, 6501 (1955).
