JOHN E. RICCIAND JOHN A.
3618
SKARULIS
Vol. 7 3
the lower range (10 to -50") is so slow, however, that all that could be reasonably read from the results is a curve attributable to a form stable above -50-60". This form, crystallizing in small octahedra, persists metastably evidently down to lo", and its solubility curve, with some points approached from undersaturation, some from supersaturation, and a few from both directions, was determined from 10 to 9.5". For each point the solid phase was examined microscopically and found to be always the same. The values, listed in Table IV, fall on a smooth curve with a minimum solubility near 85". These values represent measurements agreeing on repeated analysis with continued stirring at each temperature. The density of the solution a t 24.95', with 45.33% LiIo3, was 1.587. Many widely scattered and unreproducible "solubilities" lower than the values on this curve were observed a t temperatures below 5 5 O , and one higher than the curve a t 60" but none
below the curve a t temperatures above 50". It is probable, therefore, that the form involved is stable above and unstable below -50-60". Most attention was giver1 to 25", where a longconstant value of 43.86y0 was obtained with several different samples of starting material. This is lower than the value in Table IV, and therefore must pertain to a form stable (relatively) a t 25". Whether it is the most stable form a t 25", however, cannot be said, although its solubility does agree fairly well with the value ( ~ 4 4 . 0 7 extrapolated ~) from the solubility curve of LiIOs in the system LiIOs-HIO3-H20 as presented in Table 111 and Fig. 2. It also agreed in crystalline appearance with the LiIOs solid phase obtained in the ternary system, in the form of long hexagonal rods, as already mentioned. It is probably the form studied crystallographically by Zachariasen and Barta. 22 Since this solid is definitely anhydrous according to Fig. 2, we may infer that the higher temperature form involved in Table IV must also be anhydrous. TABLE IV For comparison, we note that the solubility SOLUBILITY OF ONE FORM OF LiI08 (OCTAHEDRAL CRYSTALS) values in the literature are few, a t scattered U, undersaturation; S , supersaturation; m , metastable temperatures, and sometimes with no information Temp Solubility Approach concerning the purity of the salt and the attainOC. wt. ?& LiIOa from ment of equilibrium. LUhden~ann~~ reported 9.93 47.19 ( m ) u 42.18% a t 10'; H e y d ~ e i l e r gave ~ ~ values of 23.5 20.24 45.86 (m) S and 38.3%, for two forms, at 18"; G r i i n e i ~ e n ~ ~ 24.95 45.33 (m) U&S reported 38% a t 18"; Mylius and Funkz6reported 29.94 44.89 (m) U 44.6% a t 18". Ira addition a hydrate of the salt 34 * 95 44.45 ( m ) U was also reported (LiIOa.Hz0, at -60") by Ditte,lG 40.00 44.12 (m) U although it has not again been mentioned. 45.00 43.84 (m) U h S (22) W.H.Zachariasen and F. A . Barta, Phys. Rev., 36, 1693 (1930); 50.06 43.51 (m) S 37, 1326 (1931). 55.1 43.35 (?) U I
60 2 65 3 75.5 85 5 99 1
43 43 42 42 42
10 00
U U
82 76 85
S U
u
[CONTRIBUTION FROM
(23) R. Liihdemann, 2. physik. Chon., BSQ, 133 (1935). (24) A. Heydweiler, Ann. Physik, ST, 741 (1912). (25) E. Griineisen, Wmensch. Abh. PhyS.-Teihn. Rcichsansl., 4, 246 (1905). (26) F. Mylius and R . Funk, Bcr.. 30, 1716 (1897).
NEW YORK,N. Y.
RECEIVED NOVEMBER 29, 1950
DEPARTMENT OF CHEMISTRY, NEWYORK UNIVERSITY ]
Some Aqueous Salt Systems Involving Fluosilicates BY JOHN E. RICCIAND JOHN A. SKARULIS The solubility relations in several aqueous systems, some ternary and one quaternary, involving fluosilicates are reported for 25'. The isotherm of the system KsSiFs-KBr-HnO shows the saturating solid KlSiFs to be anhydrous. The three ternary systems involving the salt pairs ( NH4)&iiFpMgSiF6,SrCl&rSiFs, and (NH4)tSiFarSiFs were studied with 0.5% aqueous HoSiFs as solvent, to prevent hydrolytic precipitations. The first two are simple, with (NHI)~W~, MgSiF143H20, SrClt.6H~0and SrSiFe2HeO as sole solid phases. The third pair forms a congruently soluble double salt the formula of which seems to be (NH4)tSiF&SrSiF@. In connection with these systems the solubilities of (NH4)BSiF6and SrSiF~2Hs0in presence of H&iFs, up to -30%, were also determined. The three systems containing the pairs NH&-NH4CI,NH4CI(NH&SiF6, and NH4F-(NH4)2SiF& were studied with pure water as solvent. The first two are simple, the only solids being the anhydrous salts. The third involves the already known incongruently soluble double salt NHnF-(NH4)2SiFs. The 25" isotherm of the quaternary system NH4F-NHrCI-( NH&3iF&-HtOhas two solutions of threefold saturation. One is a transition point in isothermal evaporation, with the phase reaction (NH4)aiFs liquid F? NH&l NHIF.(NH&~F~ H t G , and the other is the congruent drying-up point for the solids NH4F NH&I NHIF.(NH&~IF~.
+ + +
The literature contains little information on solubility equilibria of the fluosilicates. Solubilities of a number of fluosilicates, some of them of uncertain dependability, are cited in Mellor's "Treatise,"l in Seidell's "Solubilities,"2 and in a compila(1) J. W. Mellor, "A Comprehensive Treatise on Inorganic and Theoretical Chemistry," Longmans, Green & Co , New York, N. Y . , 1025, Vol. VI, pp. 944-958. (2) A. Seidell, "Solubilities of Inorganic and Organic Substances," D. Van Nostrand Co., New York, N , X.'1940, Vol. I, pp. 810, 970,
+
+
tion by Carter of values known to 1930.3 Carter also reported some further measurements, in particular the solubilities of the sodium, potassium and No double salts of barium salts from 0 to -80'. the fluosilicates are mentioned other than the compound NHdF. (NH&SiFO. This was first prepared 1098. A few further individual solubilties have since been reported in the literature. 13) R. H. Carter, Ind. Eng. Chrm., U, 886 (1980).
Aug., 1951
SOMEAQUEOUSSALTSYSTEMS INVOLVING FLUOSILICATES
by M a r i g n a ~ ,and ~ its crystal structure has been studied by Hoard and Williams.s The only study of aqueous salt systems involving fluosilicates seems to be that of the system Na2SiF6-NaC1-H&l reported by Anosov and Chirkov,6 in which the solid phases are the separate, single anhydrous salts a t the temperature 15'. Some of the difficulties involved in such studies are the hydrolytic instability of many of the salts with precipitation of silica or metallic fluoride or both, the corrosive action of the solutions on glass ware, and the lack of rapid and accurate analytical methods for fluosilicates, fluorides and silica in the presence of each other. The systems here reported, dealing principally with the fluosilicates of ammonium, strontium and magnesium, axe presented as a possible start for the more systematic investigation of the aqueous solubility relationships of the fluosilicates among themselves and with other related salts. Ammonium fluosilicate is a stable compound easily prepared and purified by recrystallization from water. It was planned to investigate a number of systems a t 25' with it as a component in conjunction with some of the more stable fluosilicates as well as with other salts, with the aim of determining any double salt formation. Fluosilicates other than those of the alkali metals suffer hydrolytic precipitation in pure water solution, because of the reaction Sips' -I- 4Hz0 Si(OH)4 4H+ 6F-.' Exploratory experiments on the alkaline earth fluosilicates showed that such precipitation could be prevented by the use of 0.5% fluosilicic acid as solvent in the case of both strontium and magnesium fluosilicates a t room temperature. Since the over-all phase diagram with such a solvent would hardly differ from that with pure water, the ternary systems of each of these salts with ammonium fluosilicate were investigated as a starting point. The salts (NH4)2SiF6 and SrSiF6 were found to form a congruently soluble anhydrous double salt ("4)$iF6.6SrSiF~ as one of the solid phases a t 25'. In connection with this isotherm the simpler system SrC12-SrSiF6-Hz0 was also studied for the purpose of confirming the dihydrate of SrSiFB as the equilibrium solid phase for strontium fluosilicate a t 25'. The barium salt of fluosilicic acid is only slightly ~ o l u b l ewhile , ~ the calcium salt was found to hydrolyze to such an extent as to require too high a concentration of acid to prevent precipitation of calcium fluoride and silica. Ternary systems involving these salts were therefore not attempted. In order to add to the scanty information about the stability relations of the double salt NH4F. the ternary system NHJ?-(NH4)zSiFs-HzO was studied, together with the corresponding system with NH4Cl in place of NH4F, in
-
+
+
(4) J. C. G. de Marignac, Annalcs de Mines, [ 5 ] 15, 224 (1859). (5) J. L. Hoard and M. B. Williams, THIS JOURNAL, 64, 633 (1942); also, H. Baker, J . Chcm. Soc., S I , 762 (1879). (6) V. Y. Anosov and S. K. Chirkov, J . Agplicd Chcm., U.S . S. R . , 6, 224 (1938).
(7) (a) I. G. Ryss and N. P. Bakina, Compl. urnd. (Doklady) Acad. Sci. U. R. S. S., ( N . S.),8, 21 (1936), cited in C. A , , SO, 7058 (1936). 1.2 X 10-1' for this reaction a t 20°. (b) I. G . Ryss, J . Phys. give K Chcm., U.S.S.R., 41, 197 (1947), cited in C.A . , 41, 6112 (1947), gives R 6.4 X 10-fl e t 11'.
-
3619
which, however, no double salt appeared at 25'. With the additional simple system NHa-NH4CIHzO these ternarv svstems were then used in the study of the qiatlmary system NH4F-NH4Cl(NHJ2SiFgH20at 25'. . In .addition, the system K2SiF6-KBr-H20 was also investigated in order to determine if precipitated gelatinous potassium fluosilicate was hydrated. When a solution of a potassium salt is added t o fluosilicic acid, a slightly opalescent, almost transparent, gelatinous precipitate forms.* When filtered, it is still gelatinous but dries to a fine white powder. Upon resuspension of this amorphous powder in water the material regains its former appearance. It seemed possible that the gelatinous material was an unstable hydrate, but the phase diagram with KBr shows it to be anhydrous. Matedds.-Five general methods are available for the preparation of the fluosilicates? (a) reaction of gaseous SiF, with a solid metallic fluoride; (b) solution of the metallic fluoride together with silicic acid in hydrofluoric acid and evaporation; (c) digestion of BaSiFs with a solution of the metallic sulfate, filtration of the BaSOI, and evaporation of the liltrate; (d) neutralization of fluosilicic acid with either the carbonate or the hydroxide, and evaporation; (e) precipitation of the fluosilicate by the addition of alcohol to a solution containing fluosilicic acid and the metallic chloride. The first two of these methods were not used in the present work. Although the other three methods were tried in the preparation of strontium and magnesium fluosilicates, the last was found t o be the most convenient. This method avoids the evaporation of solution, a process requiring chemically resistant non-glass apparatus if a pure product is t o be obtained. Moreover, the use of the carbonate t o neutralize the fluosilicic acid is a slow process involving persistent foaming, while the use of the hydroxide causes precipitation of silica before complete neutralization. Ammonium Fluosilicate.-This salt was prepared by the treatment of 30% fluosilicic acid (Baker and Adamson) containing ,0.3 to 0.4% hydrofluoric acid, with 28% aqueous ammonia. The hydrofluoric acid was added to prevent precipitation of gelatinous silica which otherwise occurs before the calculated amount of ammonia is added (such precipitation was observed by Berzelius, ref. 8, p. 192). The neutralization was carried out in a special pot cooled in a n icewater mixture, the pot being a one-liter cylindrical container made of a plastic material chemically resistant to hydrofluoric acid. During the dropwise addition of the ammonia, the solution was stirred by means of a motor-driven, waxed glass stirrer. The small crop of ammonium fluosilicate crystals forming after sufficient cooling was collected by suction on a filter paper in a biichner funnel, and the mother liquor was immediately returned to the plastic pot and evaporated by heating until again saturated. Cooling t o room temperature then gave a second crop of crystals. The mother liquor from this was combined with that from another neutralization, evaporated, and cooled, yielding a third crop. This cycle of processes was repeated several times before the liquid was discarded. The solid ammonium fluosilicate was recrystallized once from water by the cooling of a saturated solution a t 50" t o room temperature. A final yield of 80-90% was possible in preparation of a pound of the salt. The final crystals were collected on a biichner funnel, washed with a small amount of water, and given a preliminary air drying. Then they were placed in a large platinum dish, heated for several hours in a n oven a t go", pulverized, and finally heated at 80' for 24 hours. The salt thus prepared gave a clear solution in water, and was 99.9% pure on the basis of ammonia, determined by distillation into standard H&o4 followed by NaOH titration with methyl red as indicator. Determination of fluosilicate by the method described later under the system "IF(NH4)&iFcH20 gave 79.6% SiF6 against a theoretical value of 79.75%.. Magnesrum Fluosilkatc.-About 90 g. of C.P. MgCL. 6H10 was added to 500 d.of 95% alcohol together with the (8) J. J. Berzelius, Ann. Phys. Chcm., 1, 188 (1824). (9) Summarized from Meltor's "Treatise," ref. 1.
3620
JOHN
E. RICCIAND JOHN A. SKARULIS
Vol. 73
minimum amount of water needed for complete solution. to settle and the solution decanted. The precipitate was The filtered solution was added dropwise with stirring to 250 washed thoroughly by decantation and then again repeatml. of 30% H&iFe in the plastic pot. The final alcohol con- edly, with water, on filter pa er in a biichner funnel. It and dried at 100". centration was about 50% by volume. Well formed white was finally washed with 95'$alcohol crystals of MgSiFv6HzO settled out leaving a clear super- By conversion t o K&Oc with sulfuric acid a purity of 99.7% natant liquid. After decantation the crystals were collected was calculated, while direct titration of SiFo' with 0.1 N by suction on filter paper on a biichner funnel. They were NaOH and phenolphthalein as indicator, in a hot solution then redissolved in a minimum amount of 5% H&iFs and re- of the salt, gave 100.0%. precipitated by the very slow, dropwise addition of an equal Ammonium Fluoride.-For the preparation of complexes volume of 95% alcohol, with stirring. The crystals were of definite composition this salt was used in the form of an again filtered and the process repeated twice more. During analyzed solution of C.P. material (J. T. Baker), which was the final precipitation the stirring was done manually with first adjusted to the methyl red end-point by addition of a a hard rubber spatula rather than with the waxed stirrer. small amount of H F . The neutralized solutioq, preserved After a final washing with 95% alcohol containing a little in a waxed bottle, contained 21.54% ammonium, deterHzSiFs, the crystals were placed in an evacuated desiccator mined by distillation and titration, and 22.70% fluoride, over CaCl, for removal of alcohol and surface moisture. determined as PbClF.11 These values give 1:1.001 for A solution of the salt thus prepared gave no turbidity the ratio NHd:F, corresponding very closely to pure NH4F. Because of The original solid salt, dried for several weeks over solid when tested with AgNOa in presence of "0,. the insolubility of MgF2 in water the turbidity of a solution KOH in an evacuated desiccator, was used when exact total of magnesium fluosilicate in pure water WLESused as a cri- compositions were not required; the KOW was intended to terion of purity as well as of the extent of hydrolysis. The take up COZfrom decomposition of ammonium carbonate dry, well formed, free-flowing crystals gave only a very probably accounting for the original h l i n i t y . slight turbidity when dissolved in water at room temperaOther MatarbZs.-C.P. NH&!1 was used without further ture. Upon standing this turbidity increased a bit. But purification. C.P. KBr was recrystallized from water, it was found that with 0.3% HiSiFe a s solvent no turbidity ground after preliminary drying at 105", and heated finally developed with standing overnight. Ten grams of the a t 150". Strontium chloride was used as recrystallized C.P. crystals was dissolved in 40 ml. of each of a series of fluo- SrCls-GHgO. silicic acid solutions with concentrations varying from 0.1 Procedure and Preliminary Tests.-Mixtures of the comt o 1.0%. The solutions were allowed to stand in beakers ponents were brought t o equilibrium in closed tubes rotated coated lightly with wax. The decreasing turbidity with in a large water-bath maintained a t 25 f 0.02". The soluincreasing acid concentration was easily observed, and it was bility tubes and bottles were carefully waxed with material decided to work with 0.5% acid as solvent to provide a obtained by melting down the fluosilicic acid bottles. The margin of safety, inasmuch as the 0.3% solution remained stoppers were of rubber. For analysis the saturated soluclear for the 24 hours of the test. tions were sampled by means of waxed pipets fitted with No attempt was made t o prepare a salt having the exact filter paper tips. composition of the hexahydrate. Instead, the hydrated In all cases the samples were rotated for a t least two or salt was thoroughly mixed and analyzed for magnesium three weeks, after which time several of the complexes in when complexes of known composition were to be made up. each system studied were reanalyzed on consecutive days to The ratio of Mg t o SiFs was 1:1.005, with, originally, a con- verify equilibrium. In each system equilibrium was aptent of 63.5% MgSiFs on the basis of the magnesium deter- proached from both directions, Le., some complexes from mination (precipitation with oxine as described later under undersaturation, others from supersaturation. In the latter the system (NH&SiFe-MgSiFgHzO). The theoretical cases the mixtures were first shaken at temperatures not value is 60.63% for MgSiFsBHsO. On standing in air, exceeding 35", because of the unstable nature of the mecovered by a watch glass, the water content rose, later tallic fluosilicates. analyses giving 62.1 and then 61.5% MgSiFs. As observed Specific analytical procedures are described under the inby others,'* moreover, dehydration at 105" caused consid- dividual systems. erable decomposition. The heated product gave a very As already mentioned, 0.5% H&iFs was used as solvent turbid solution, apparently precipitating MgFz. in systems involving magnesium and strontium fluosilicates, StrMltium Fluasilicate.-The procedure was the same as to prevent hydrolytic precipitation. That slow hydrolysis that for the magnesium salt, with recrystallized SrCla. of SrSiFe did not occur in the O.5y0acid was ascertained by 6 H z 0 as the starting material. Tests again showed that the constancy of the density of the saturated solution with 0.5% H&iFa was more than enough to prevent turbidity in time. The measurements were made directly in the solua solution of 5 g. of the salt (SrSiFs2HzO) in 40 ml. Anal- bility tube, without removing it from the bath, by means of ysis of the hydrated salt gave 13.0070 loss of weight (as a Westphal balance constructed from an analytical balance H2O) a t 105", 47.59% residue on ignition (as SrFz), a value set over the bath. The specific gravities observed varied of 1:1.003 for the ratio Sr:SiFs (strontium being determined from 1.14798 by +3, -3, -3, f2, 0 in the last place after as the sulfate by the method described under the system 1, 2, 3, 10, 15 days of stirring. The solid was then filtered SrCl&rSiFcH,O), and a negative test for chloride. Theo- and analyzed, and the ratio Sr :SiFe was found t o be 1:1.002 retical values for SrSiFs6H10,are 13.56% H30 and 47.28*% indicating no appreciable hydrolysis. SrFz. In contrast t o MgSiF~6Ha0,this hydrate loses its In addition parts of the isotherms of the systems (NH& water at 105" without significant decomposition. The S i F r H & F r H z O and SrSiFE-HzSiFgHzO were also studied anhydrous salt so obtained was 99.7% pure by strontium in order t o determine that the solid phases involved would determination, while the determination of both strontium not be affected by the 0.5% fluosilicic acid, or that the presand fluosilicate gave 1:0.99 for the ratio sr:siFE. In addi- ence of the acid would cause no unusual behavior in the tion, the slight turbidity of its aqueous solution cleared up ternary phase relations t o be investigated. in 0.5% HzSiFS. The analyzed hydrated and anhydrous materials were both used for the making up of complexes The measurements for the system (NH&SiF6-of definite composition. H&FC--HzO are listed in Table I, in terms of Calcium Fluosilicate.-The same method was again used, with calcium nitrate as starting material. A gelatinous pre- weight percentage. For the analysis, both of satucipitate, difficult t o filter, formed if a large excess of HpSiFe rated solutions and of wet residues, (NH*)&iF6was was not present. When the well formed crystalline salt was determined as ammonia after distillation with altreated with water a very turbid solution was obtained to- kali, while HzSiF6was titrated directly with NaOH gether with a gelatinous precipitate not easily dissolved by with phenolphthalein as indicator. This was done addition of H&iFe. If this is the behavior of pure CaSiFe in ice-cold solution saturated with KNOs to precipi(see also ref. 10, p. 369). the aqueous solubility reported in tate KaiFe and prevent reaction of SiFa- with the the literaturea has littie meaning. Potassium Fluosilicatc.--Sice this salt is only slightly alkali. To check the method, 10-d.aliquots of soluble the use of atcohof is unnecessary. A dilute solution -15% H&iFs were pipetted into each of six beakers of C.P. KCl was added to a large volume of dilute (-5%) (11) Procedure as In W. W.Scott, "Standard Methods of Chemical HaiFa. The transparent potassium fluosilkate wm allowed Analysis," D.Van Nostrand, Co., Inc , New York, N. Y., 1939, Vol. I,
(10) Pr.
Stolba, Silxbsr. Whm. Gel. Wirs., 290 (1877),
p. 405.
Aug., 1951
SOMEAQUEOUSSALTSYSTEMSINVOLVING FLUOSILICATES
3821
containing various amounts (0.0 to 5.0 g.) of as previously described, required 25.46, 25.43 and (NH4)zSiFs. The volumes were adjusted to 100 25.45 ml. in the presence, in addition, of 1 , 3 and 5 g. ml., finely divided KNOI was added with stirring SrSiFs, respectively. Finally, the direct (hot) tiuntil in slight excess, and the mixtures were cooled tration of total fluosilicate in presence of strontium in an ice-water bath for an hour, with frequent thor- was veriiied by analysis of samples of strontium ough stirring. They were then titrated to the fluosilicate both in this manner and by the method methyl red end-point with 1 N NaOH, phenol- described later for the system NHJ?-(NH4)2SiFephthalein was added, and the titration was contin- HzO, involving precipitation of KzSiFgand hot titraued to the first pink color. Another aliquot of the tion of the filtered precipitate with NaOH. The fluosilicic acid was treated in the same manner but two procedures gave values differing by only 3/ in platinum for the purpose of standardization. In 1000. cold titration with phenolphthalein as indicator TABLEI1 the titer was 22.90 ml. of 1N NaOH, while continuation of the titration hot required an additional PARTIALSYSTEMSrSiFe( =A)-H2SiF6( =B)-HsO AT 25"; SOLIDPHASESrSiFe.2H20 45.79 ml. The second titration corresponds to the Original complex Saturated solution reaction SiFe' f 40H- 3 Si(0H)r 4- 6 F- and % A %JB % A %'oB should require twice the volume for the cold titra(14.9)" 0.00 tion. Blanks run on the KNOs and indicators were 14.58 0.42 found to be negli ible. The six test aliquots re19.89 3.95 11.39 4.52 quired 22.79 ml. ?with average deviation of 0.02) 21.52 7.84 8.22 9.53 with methyl red, and 22.91 f 0.04 with phenol22.83 12.03 5.10 15.49 phthalein as indicator. Phenolphthalein was there23.63 16.25 2.45 21.87 fore chosen as the indicator. TABLEI PARTIALSYSTEM(NH&SiFa( =A)-H&iFe( = B)-HzO 25"; SOLIDPHASE,( NH,)SiFe Wet residue %A % B
AT
26.45 19.44 1.44 27.82 (48.68 13.37Ib 1.18 29.93 a Estimated by extrapolation to 0% HzSiFe. b Wet residue.
Saturated solution % A %B
The tie-lines fixed by compositions of solutions and original complexes indicated, both graphically and by algebraic e ~ t r a p o l a t i o n ,that ~ ~ the solid 71.44 2.58 6.67 phase is SrSiFs.2H20 throughout. The average 73.29 4.48 13.09 extrapolation error is -0.5 ( h O . l ) ~ oH2SiFaa t the 72.80 6.04 17.65 line representing the theoretical percentage of 60.97 10.3 22.57 SrSiFe in the dihydrate. The lowest concentration 60.93 11.9 27.39 of H2SiFsin the saturated solution was 0.42%, calGraphical and algebraic consideration of the tie- culated from the observed weight percentage of lines fixed by the compositions of saturated solu- SrSiF6 and the fact that the complex was prepared tions and wet residueP shows that the solid phase from 0.5% HzSiF6 and solid SrSiF6.2H20. The is anhydrous (NH4)2SiFs from pure water up to a solubility of SrSiF6.2HzO in pure water, by extrapoconcentration of -27% HzSiF6. The algebraic ex- lation to O%H2SiF6, is then -14.9% in terms of trapolation gives an average error of +0.6( j=0.2)% SrSiFo. The only value in the literature, for comHzSiF6 a t 100% (NH4)2SiFa. For the solubility of parison, is that of 3.22% a t 150.14 System K&3iF~-KBr-H20.-The saturated (NH~)~S~FF, a t 25O, here observed as 18.75%, Jatlov and Pinaevskaya12areport two values, 18.75 and solutions were analyzed for total solid by evaporation and drying at 105' (in platinum), and for KBr 18.32%. For the partial isotherm of the system SrSiF6- by the Mohr method. The calculation of the perby difference has, of course, little HzSiFe-HnO the measurements are listed in Table centage of 11. The complexes were prepared from SrSiF6. numerical significance, since the quantity found was 2Hz0, analyzed (-30%) HzSiFe, and water. For of the order of magnitude of the uncertainties inthe analysis of the saturated solution, H2SiFa was volved, The precise value of the low concentration determined as described in the preceding system and of KzSiFe in solution, however, does not affect the total fluosilicate was determined on a separate indirect determination of the composition of the sample, usually smaller, by direct titration of the solid phase, which was the purpose in studying this solution with NaOH, first neutralizing the acid at isotherm. Tests showed the Mohr titration of broroom temperature, then heating, and completing mide in presence of KzSiF6, which gives an acidic the titration of the fluosilicate ion. The SrSiFo was reaction, to be satisfactory. The measurements are listed in Table 111. calculated by difference. This analytical scheme was checked in various ways. Solutions with Graphical and algebraic extrapolation of the tieknown percentages of (3.16 and 3.74) in lines shows that the solid potassium fluosilicate is presence of fluosilicic acid, gave 3.17 and 3.77y0, anhydrous. The average deviation from pure respectively. Aliquots of a solution of HzSiFa KzSiF6, b extrapolation, is f0.27 ( *O.26)y0 KBr KaiF6. The solubility of K&iFs, requiring 25.42 (average) ml. of 1 N NaOH for cold a t titration in presence of KNOs and phenolphthalein -0.15%, was determined only roughly. Carter' reported 0.177 g./lOO ml. solution at 2 5 O , while in(12) F. A. H.Schreinemakcra, 2. phys. Chcm., 11, 76 (1893). 18.75 16.31 13.75 12.29 10.16 8.54
0.00
O lOd
(1%) V. 9. Jntlov and E.N.Pinaevskaya, J . Gdn. Cham., V ,S. 9. R., 18, 269 (1946).
(13) A. E.Hill and J. E.R i d , THIIJOURNAL, 64, 4306 (1981,. (14) R. Frcoceoiua, 2. 4 ~ 4 1 .Chrm., 49, 146 (16190).
3622
JOHN
E. RICCIAND
JOHN
A. SKARULIS
Vol. 73
terpolation of Ryss's values at 20 and 40" (ref. 7b) gives -0.134% a t 25".15
which the factor 0.00825 represents (MgSiF6. 6HzO/MgSiF6)/200. The average percentage of in the solution is therefore close to -0.3. TABLE 111 The solubility of the magnesium salt in absence of SYSTEM KzSiFs(A)-KBr(B)-H20 AT 25" the ammonium salt is 23.5470 MgsiF6 with -0.3Y0 Complex Solution Solid %B a t HzsiF6 in solution. Roughly, by analogy with % A %B % A %B phase 100% A SrSiF6 in Table 11, one may estimate a solubility of ... 0.00 0.15 0.00 A -23.7% in pure water. This is to be compared 10.01 5.90 .02 6.58 A -0.22 with 24.1%, interpolated from the measurements by 9.99 7.13 .02 7.92 A .00 Jatlov and PinaevskayaI6 a t 20 and 40'; these au10.04 9.92 .01 11.03 A -I- .68 thors also reported the solubility a t 20' in presence 9.76 15.50 .09 17.14 A 3- .20 of up to 46% H2SiF6. 9.98 20.33 -07 22.56 A C .OT With the small percentage of disregarded, 9.87 23.82 .19 26.30 A .73 the percentages listed in Table IV may be plotted .18 9.72 26.87 .09 28.61 A in the usual fashion on a triangular diagram with 9.03 30.28 .05 33.26 A .09 the corners representing ( N H ~ ) z S ~MgSiF6 F~, and 6.52 34.07 .07 36.42 A .22 "H20." The plot shows clearly that the solid 8.11 36.88 .09 40.05 A .56 phases are pure (NH&SiF6 and MgSiFe.6H20. For 10.06 39.48 .14 40.47 A +B precise algebraic verification of these exact solid 0.00 .00 40.51 B phases, the three components may be taken strictly System (NH&SiFe-MgSiF6-0. 5% HzSiF6 at as (NH4)2SiFe, MgSiF6.6HzO and 0.5% H2SiF6. 25 '.-For the analysis of the saturated solution, Algebraic extrapolation of the tie-lines then gives the ammonium salt was determined by ammonia errorsof +0.09, 4-0.29, +0.06 and -0.05% (NH& distillation, and the magnesium salt in a separate siF6 a t 100% MgSiF6.6H20,and an average error of >'fgSiF6.6Hz0at 100% ("413sample by the quinolate method for magnesium. -0.03 ( * O . l l ) % For this purpose the sample was weighed in a small SiF6. System SrCl~-SrSiF6-0.5% HzSiFe at 25'.platinum dish covered with a watch glass. A platinum cover was then substituted for the glass one The saturated solutions were analyzed for total and sulfuric acid was added carefully to the sample. strontium and for chloride. The preparation of The solution was evaporated to fumes of SO1 to samples for strontium determination was the same drive off H F and SiF4, cooled, and transferred to a as that already described for magnesium, up Pyrex beaker. It was again evaporated to fumes of through their evaporation practically t o dryness in SO%fumed strongly, cooled, diluted with water and beakers to expel excess H2S04. After addition of just neutralized with ammonia. After filtration to 100 ml. of water and 100 ml. of 9570 alcohol, the remove suspended material, the solution was ali- precipitate of SrS04was allowed to settle overnight. quoted if necessary and the magnesium was precipi- It was filtered through asbestos in an ignited gooch crucible, being washed and transferred with 50% altated and weighed as the quinoline dihydrate. The results are listed in Table IV. Since the cohol containing a little HzS04. The strontium sulcomplexes were made up from (NH&SiFs, MgSiF6. fate was finally washed with small portions of 95% 6H20and 0.5% H2SiF6, the actual weight percent- alcohol, dried at l l O o , and ignited to constant Chloride was determined by the age of HzSiF6, both in the complexes and in the weight at -800". saturated solutions, is not uniform; %H&%F6 = Volhard method, the excess of AgN03 being ti0.5 - %(NH4)2SiF6/200 - 70 MgsiF6(0.00825), in trated with KCNS in suitable aliquots after filtration of AgC1. Tests were made showing that SrSiFe TABLE IV did not interfere with the chloride determination.
+ + + + +
1..
SYSTEM ( NH&S~FB-M~S~F~-O.~%HZS~F~ AT 25 A = (NH&SiFa, B = MgSiFs, W = H 2 0 Complex
%A
25.33 24.77 24.95 25.64 a: : : {
17.98 17.07 10.94 9.46 6.80 3.82 1.90 0.00
Solution %A %B
% B
6.96 9.64 10.70 12.07 15.08 5.55" 18.73 20.35 25.72 25.04 25.87 25.81 30.35
}
TABLE V SYSTEM SrC&SrCIF&.50/oHzSiFa AT 25" A = SrClz, B = SrSiF6, W = H,O
Solid phase
15.85 14.77 14.36 13.61 12.48
7.85 10.91 12.23 14.01 17.32
A A A
11.62 11.39 11.35 10.45 7.51 4.14 2.37 0.00
20.19 20.85 20.96 21.28 22.08 22.85 23.06 23.54
A A B.6W A B6W B.6W B.6W B.6W B.6W B.6W
Complex % A %B
...
A
A
+ +
(15) There is consjderable disagreement a t other temperatures also. While Carter (ref. 3) gives 0.152 g./lOO ml. a t 16O, A. A. Wassilieff and N. N. Martianoff, 2. ~ d Ckcnz., . 103, 103 (1935). give 0.115 at 1 7 O : ~ n
