July, 1938
1637
CALCIUM S U L F A T E - S O b I U M S U L F A T E - W A T E R
means of equation (8) and are given in Table IV. The corresponding values of A(o(peXptl. - c p ~ )have also been calculated and are included in the same table (111). For some of the salts A p has been plotted against m in Fig. 6. Cerium, praseodymium and neodymium chlorides were omitted From the plot because the values of Ap for these salts were so close to lanthanum chloride which has been included. The values of Ap representing the deviation from Scatchard’s equation are not proportionately large when compared to those given by him for the higher valence sulfates measured by Robinson and Jones.l4 I n fact, the actual values of A p found in this investigation are less than those for the sulfates. (14) Robinson and Jones, THIS JOURNAL, 68, 959 (1936).
[CONTRIBUTION FROM THE CHEMICAL
Ternary Systems. XXIV.
I am indebted to Professor Scatchard and Dr. Wood for the use df their data for the osmotic coefficients of potassium and sodium chlorides prior to publication and for several suggestions regarding the use of equation (8). For these courtesies I wish to express thanks.
Summary 1. The osmotic coefficients of aluminum, scandium, yttrium, lanthanum, cerium, praseodymium and neodymium chlorides have been determined in aqueous solution a t 25’ by the isotonic method of Robinson and Sinc1air.l 2. The activity coefficients of these salts have been computed from the osmotic coefficients. 3. The data obtained in both 1 and 2 have been compared with theory. DURHAM,
N. H.
LABORATORIES OF NEWYORK
RECEIVED APRIL30, 1938
UNIVERSITY ]
Calcium Sulfate, Sodium Sulfate and Water
BY ARTHUR E. HILLAND
JOHN
H. WILLS‘
Early study of this system by Fritsche2 dis- double salts with gypsum and mirabilite (Glauclosed the existence of two double salts, glauberite ber’s salt) as 30.2’ for the transition involving the (CaSOk.Na2S04) and a salt of the composition labile double salt and 29’ for that involving glauCaS04-2Na2S04.2H2O1which has been variously berite. The present research contains a repetition known as syngenite, from a mistaken reading of of the 25’ isotherm and additional isotherms a t its composition, and as the labile double salt, from 35, 50 and ’is’, including in addition t o the its metastable relationship to other double salts. previously known double salts, anhydrite a t the Isotherms of the system have been studied a t 22’ higher temperatures and a new double salt, by Cameron and Seidell13and a t 25’ by Cameron sodium pentacalcium sulfate; these data permit and Breazeale14in both cases without inclusion of the construction of a fuller polytherm than was double salts. Barre5 compiled the concentrations previously possible. of isothermally invariant solutions saturated with In carrying out work in this system in which two salts, glauberite and each of the components there occur so many degrees of metastability at of the system and the labile salt with each of the any given temperature, much use has been made components, giving an outlined polytherm of the of van’t Hoff)s*generalization that the retardasystem from room temperature to 100’. D’Ans tion of salts in coming to equilibrium with their and Schreiner6 discussed the two double salts solutions is related to the mean valence, calcuand added a point of saturation by glauberite and lated by dividing the total valence representing gypsum at 60’. Van’t Hoff7 determined the in- all the ions in the formula by the total number of variant points of saturation by each of these such ions, giving each water molecule in a hydrate an arbitrary mean valence of 4/3 ; this rule guided (1) This paper is an abridgment of the thesis presented by Mr. Wills in partial fulfilment of the requirements for the degree of us in this research, as in other related r e s e a r c h e ~ , ~ Ph.D. at New York University, June, 1935. in deciding upon appropriate time of contact for (2) Fritsche, J . prakt. Chem., ‘?a, 291 (1857). Glauberite had been prepared previously only in the dry way. (3) Cameron and Seidell, J . Phys. Chem., 6 , 643 (1901). (4) Cameron and Breazeale, ibid., 8, 335 (1904). ( 5 ) Barre, Ann. chinz. phys., [SI 34, 145 (1911). (6) D’Ans and Schreiner, e. anorg. Chem., 63, 129 (1909). (7) Vnn’t Hoff, 2. physik. Chem., 46, 257 (1903).
(8) Van’t Hoff, “Zur Bildung der ozeanischen Salzablagerungen,” F. Vieweg und Sobn, Braunschweig, 1905; Vol. I, p. 32; “01. 11, p. 17. (9) Hill, THISJOURNAL, 66, 1071 (1934); Hill and Yanick, ibid., 67, 045 (1935).
164s
, ~ K T H U K E.
HILLA N I ) J O H N IT. WILLS
\?Ol.
the different salts, for which periods as short as oiie or two minutes may sometimes not be exceeded without chaiige, while in other cases weeks or months are necessary. The entire list 01' salts iiow kliowii in this sysceni, .i.rith the meat1 valeiiee for each. is as follows
($0
gypsum in sodium sulfate solution in the form of a network of fine spicules so knitted together that a weight of approximately 2% of the salt will hold the mass in suspension so rigidly that there is no moveinelit upon inversion of the vessel. This condition obviously retards the attainment ui' equilibrium and may leave a considerable pordrlle 1 orniiii 4 tion of gypsum unchanged, which is probably the h~lrdblhtC nn:so, lU€I*U 1 i.3 (1-nnanied) Na2S04.Xi20 ) 1 .;:j) cause of Cameron, Bell and Robinson's error. rhenardrte NasSO1 1 .3'> This network retains a large proportion of mother Labile salt 2Na,S04 CaSO4.2H10 1 4.j liquor even upon centrifuging a t 1800 r. p. m., so Gypsum CaS042H20 I Xi that a direct analysis such as D'Ans and Schreiner Glauberitc Na&Oc Cas04 100 made is not always possible. We are able, howPcntasulfate NazS04.5CaSOp3H10 1 O i (Hemihydrate) (Cas04 0 5H20) ! I 71; ever, to confirm their findings, by the use of both a'i~~l~ycirite CaSOl 2 til methods ; algebraic extrapolation of the tie lines Of this series, the second and the eighth, for the labile salt in Table I gave (assuming the sodium sulfate heptahydrate and calcium sulfate water content of the salt correct) agreement with hemihydrate, did not appear spontaneously in the original formula of Fritsche with an average any oi our experiments, aiid were omitted from deviation of 0 757, for the six cases there tabuconsideration ; the others have been considered lated, with a maximum deviation of 1.30%, in in the appropriate isotherms Special attentioil experiments in which the amounts of solid phase was paid to the labile salt, coiiceriiing which a were usually less than 15% of the weight of the sample was also prepared by washing number of points have apparently remained uti- system. out the mother liquor with alcohol-water mixsettled in the eighty years since its discovery arid tures, which was fouiid to contain 29.75% Cas04 of course special at tentioii was likewise giveii to (theory 2S.ti3), 61.5iTo Nag904 (theory 62.27), the new peritasalt after it had appeared 111 oiic aiid 8.68% HzO (theory 7.90), in agreement with oi the isotherms. the formula CaS04.2NazSO4.2H-fa0. The Labile Salt, Sodium Hemicalcium SulWith respect to the metastability of the salt, fate.-The compositiori of this salt as proposed by FAtsche and its condition as metastable with our findings are in agreement with those of Barre respect to glauberite as held by van't Hoff were and of D'Ans and Schreiner that the salt can be disputed by Cameron, Bell and Hobinson,lo who held unchanged for long periods of time in contact believed the formula to be L'CaS04.3Na2SO4 with its solutions a t lower temperatures (25' (anhydrous) on the basis of indirect determination e. g.), and that only a t 75' and above does the of composition in a sodium chloride solution, and change toward glauberite begin a t once. The also believed that their dilatometric experinients action of the salt when treated at 25' with pure showed it to be stable. Experimeiital answer to water instead of a sodium sulfate solution with this view appears in the €our isotherms presented which it is in equilibrium (3Ck34Oj00)is noteworthy; in this paper, which shoiv by solubility curves the decomposition cannot be better described than that the labile salt is metastable with respect to as explosive, both because of the astonishing speed glauberite between 2.5 and 7.5L, iii accord with a t which the solid disappears, and also because van't HOE'Sconclusitrn bawd upon the transition o f the unexpected concentration attained. A teniperatures. I'ht matter o f the composition of sample was treated with cold water and immedithe salt was gone into by D'Aiis aiid Schreiner,b ately filtered from a small residue of calcium sulwho were able t u obtain samples sufficiently freed fate, the solution was analyzed and found to confrom mother liquor, by use of a hydraulic press, tain i.Ol(,L calcium sulfate. At 25' the solubility to permit ahalyses, which agreed well with thc o f gypsum is 0.Z0S70 and that of anhydritei1 is formula CaS01*2Na&304-2Hz0.I t is worth riot- O.E.LC/, ; the solubility of hemihydrate from the ing that each of the methods which led to the con- curve of ChasseventL2is about 0.8%,, all figures flicting results is accompanied by unusual diffi- heiiig considerably less than the value found. ctilties i n the case o f this salt, which forms from 11) HdJ, T H I S J O U R h A L , 69, 2242 ( 1 s d i t i'
I
( i1-x
11, 40'1 f J W 7 >
12) (.ha\irre.it
Ann
ihzin
phvs
[10] 6, 244 (1926)
July, 1938
CALCIUM SULFATESODIUM SULFATE-WATER
This seems to be evidence that the mechanism by which the salt decomposes on treatment with water cannot be a simple change into diwolved sodium sulfate and solid calcium sulfate in one of the three forms noted, since a solution cannot supersaturate itself from a given solid phase; either a fourth solid calcium sulfate is formed (an amorphous body, perhaps, which may be the condition of the oft-quoted “soluble anhydrite”) or else the primary change is into dissolved material which only later precipitates. Investigation of this phenomenon of decomposition is being continued. Microscopic identification of the labile hemisalt is comparatively easy, by two of its properties : first, the unusual slimness of the long needles, which run quite unifornily between 1 and 2 p in breadth, whatever be the length, and by their refractive index y = 1.510 (=t0.003), which is sufficiently different from that of glauberite as given by Larsen13 (7 = 1.536) as well as that of the pentasalt given later in this paper. The refractive indices in the other directions could not be determined accurately enough. Sodium Pentacalcium Sulfate.-The new salt, resembling the pentacalcium sulfates found in the potash series and the ammonia series except as to extent of hydration, was first found a t 75’ when solutions of about 14% sodium sulfate content were treated with gypsum; in general, it may be prepared easily a t temperatures of 60 to 75” in solutions of 10 to 14y0sodium sulfate upon addition af gypsum, or more rapidly by introduction of g y p s m and glauberite in proportions corresponding to the reaction 4(CaS04.2Hz0) CaS04.Na&304---f 5CaS04.NazS04.3Hz0 5HzO. The reaction is completed in four days or less; if the concentration of the solution is such that the salt is metastable only with respect to anhydrite (see Fig. 4), subsequent change to that solid is very slow and need hardly be taken into consideration, but operation in the conceritration where the pentasalt is metastable with respect to glauberite will result in fairly rapid and eventually complete change into that salt. Samples of the pentasalt may be washed free of mother liquor without appreciable decomposition, and are found upon analysis to be in close agreement with the formula given; a sample gave CaS04 77.59% (theory 77.63), NazS04 16.21 (theory 16.20),
+
+
(18) Lilrsen, “The Microscopic Determination of the Nonopaque Minerals,”U 5 Govt Printing Office, Washington, D C , 1 9 3 4 , ~156. .
1649
H2O 6.20 (theory 6.17). This water content, which was confirmed by a number of analyses giving 6.2% of water or a few tenths highei, corresponds to a tiihydrate, wherea the corresponding potassium and ammonium salts have been found by all investigators to be monohydrates. The salt crystallizes in rather fine needles of an average diameter of about 7 p , which under the microscope are seen to be prisms (or pinacoid faces) with pyramidal ends. They show oblique extinction; ZAc is 11’ ( = ~ = 2 )which , is quite close to that found for the corresponding potassium and ammonium saltg by Gabriel;14 the refractive indices found were CY = 1.5537 (*0.003), y = 1.567 (*0.003). Experimental Procedure The materials used were precipitated gypsum and analyzed anhydrous sodium sulfate. Where salts were expected to form by reaction of these components, they were weighed accurately with weighed quantities of water, to give a known original complex as ofie of the points t o be used in extrapolations for the composition of the solid phases. The materials were rotated in stoppered bottles in thermostats for appropriate times. For analysis, solution was filtered off and a part evaporated t o constant weight, first a t 100’ and finally for a day a t 200’; in the remainder af the solution, the calcium was precipitated as oxalate and weighed as oxide. It was found that the precipitated calcium oxalate occluded sodium sulfate in weighable amount, whereas occlusion of potassium sulfate did not occur in our previous the most effective method of elimination was found to be the volatilization of the sodium salt from the bottom of the platinum crucible to the lid, which occurred a t temperatures reached by a Meker burner or blast lamp, and subsequent volatilization from the lid by heating it in the open; failure to do this would leave the results one per cent. or more too high in calcium. The 25’ Isotherm.-The results a t 25’ are given in Table I and shown in Fig. 1. The solubility curve for gypsum passes through a minimum a t a concentration of about 1.5y0of sodium sulfate, and through a maximum at about 2U%. The figures of Cameron and Breazeale4 are distinctly higher than found by us; the isothermally invariant point for gypsum and mirahilite (14) GabrieI, THISJOURNAL, 6 1 , 686 (1935).
AKTHUK13. HILI, AND J o m
1tj50 TABLE
J
CaS04-Na2S04-HI0AT 25'
_-
-
0.573 1 529 3.109 5 964 10.80 14.83 17.89 19 24 24 21
__?$I
x1
- 0 209 0.595 ,148 ,139 l.C,Ol 3.200 ,144 6.251 ,161 10.93 .181 15.23 . I94 ,197 18.09 20.15 .198 .19i 21.75 t
3 773 3 7x9 2 5til 3.803 1.17Cl 3.304 1.093 3 803 1 276
25.87
,188
0 325
2'7.98 29. 78
.180 174
-__
33,85
. 155
21. 70 21.72 3 8 . !11
--
___
___
_ _-
23.29
1 138
__
.
~
-~__ . ._.
3G 37
0 358
33.31 32.84 31.74 31 60 30 69 31 01
700 426 427 693 317 744
,120 ,049
32.61 29.31 26. ti0 34.13
,065 ,113 . 175 I055
32.72 32,49 31.41 31.0
