S O L Y B I L I T Y OF GYPSCJI I S AIQVEOUS S O L U T I O S S O F SODIUM CHLORIDE' BY F R X K K K. C A S I E R O S
T h e effect of one electrolyte in determining the solubility of another in water solutions has received considerable attention from investigators within the past few years. T h i s work has been directecl mainly to a verification of theories deduced from applications of the inass law to electrolytic dissociation, and quite satisfactory results have been obtained for \-cry dilute solutions. Excepting the early and now classical researches of Kopp,' E;arsteii,3 Rudorff,j and Eiige1,j but little satisfactory work is to be found described ill the literature on this subject when coiicentrations of any magnitude were involved or when the work was carried on through any great range of concentration. Moreover, the cases thus studied have been carefully selected so as to take advantage of some peculiar cliaracteristic of one or more of the coiiipoiieiits in the system under consideration and thus simplify the theoretical formulation of the factors involved. Such cases as one meets most frequently in everyday practice- two strong electrolytes in water, for exatiiple - have received but scant attention. Tiie reason for this is that snch cases n j),ioi-i could iiot be expected to give satisfactory results, so far as the theories under examination were concerned, it being a well-establislied fact that for such electrolytes the dissociation does iiot take place in accordance with the J ~ Z C I S S as we now foriiiulate it. For inany technical problems, studies in geoloqy, and similar investigations this niutual influen re of salts or electrolytes in solution is of very great importance. Qualitatively we already /mil,
'
Contribution from the Bureau of Soils, Y.S. Department of Xgriculture. I'ablishetl by perxnission of the Secretary of Agriculture. Liebig's Annalen, 34, 260 (rS4o). ' .\bhandl. der Berlin Akacl. (1841) . ' Pogg. Ann. 148,456, j j j ( 1 8 7 3 ) . Ivied. h n n . 2 5 , 626 ( I S S j ) . ' Comptes rendus, 102, I 13 (1886).
know something in this connection, but quantitative data have been almost entirely wanting up to the present.’ I n such studies we have to do with solutions often far from dilute, and we have 110 consisteiit theory to guide 11s ; furthermore, the particular salts involved in these problems have but seldoiii coininended theiiiselves to investigators. It is not often that me have to deal with but one pair of electrol? tes and their metathetical products i n the stody of such problems. but in tlie absence of satisfactory available data it was deemed wise to take a case of this kind for a preliiiiinary study. .4ccordingly, from time to time, as attention to other work of the laboratory permittc d, the solubility curves for several pairs of electroll. tes have been followed quantitatively. Some results are here recorded for the s! stein : TZ’ater, calcium sulphate, sodium chloride, and their reaction products. Solutions of sodium chloride of various concentrations were prepared. T h e sodium chloride used mas clieinically pure material obtained from Eimer and Amend and tested by anal yzing for clilorine with very satisfactory results. Into these sodinin chloride soiutions an excess of calciiiin sulphate i n the forni of g~psmii (CaS04.2HZO), finely divided, was introduced, and tlie solutions then allowed to stand, with frequent sliakings, from one to fourteen da!.s, according to the temperature a t which the work was being done, in order that equilibrium inight be reached. T h e gjpsuni used in the experiments described in this paper was from two sources. One sample was a very finely powdered gypsum obtained from Hot Springs, Fall River Count!., S.Dak., an analysis of which gave very satisfacI For a review of the earlier work see IV. IV. J. Nicol (Phil. Mag. 3 1 , 369 1 8 9 1 ) ) and J. E. Trevor (Phil. Xag. 32, 7 j (1891)). Nore recent investigations of interest in this connection are the ’‘ Solubility of Calcium Bicarbonate in Sodium Chloride Solutions.” Treadwell and Reuter (Zeit. anorg. Chem. 17, 192 (1SgS)); Solubility of Gypsum in Solutions of Hydrochloric Acid and Calj ;) “ Solubility of cium Chloride,” Lunge ( J o u r . SOC.Chern. Ind. 4, 31 ( 1 6 s ~ Gypsum in Solutions of Anirrionium Chloride, Magnesium Chloride, and Calcium Chloride,” Tilden a n d Shenstone (Proc. Roy. SOC. 38, 335 (rS8j)’) : “ Solubility of Calcium Sulphate in Several Acids,” Ostwald and Banthisch (Jour. prakt. Chem. [2] 29, j2 ( I S S ~ ) ) .
(
/’
tory results for the formula CaS04.2H20. T h e material was powdered so fiue1)- that it displaj ed a decided tendency to“ cake,” and when used the solutions had to be shaken frequently. T h i s ((caking” did not resemble the setting of plaster of Paris, but the substance behaved more like a plastic clay. T h e other gypsum sample was obtained bj. pulverizing in an agate mortar some very fine c r j stals of gypsum kitidlj- furnished by Nr. IYirt Tassin, of the Vnited States Kational 1Inseuiii. I t ma) be said at once that curves were plotted from data obtained with both these materials and with calcium sulphate obtained by precipitation from mixing solutions of calcium chloride and sodium sulphate, all of which agreed most satisfactorily. T h e material first described possessed one advantage in that its very finely divided coiiditioii enabled it to go into solution more rapidly and greatly shorteiied the time required to obtain final eqnilibrium i n the system. Preiimiiiary experillietits indicated the advisability of using rather large volumes of solution. .4boiit 300 cc was fotind to be a very convenient amount. I n the experiments at higher temperatures tlie portions of solutions required for analysis were drawn off with a pipette as quickly as possible, care being taken to avoid drawing off any solid gypsum. In some cases considerable difficulty n-as experienced in this respect, but it is not believed that an! of the results here presented are open to criticism on this account. M’hen working a t tlie lower temperatures, a 100cc portion was filtered through a Schleicher und Schiill ” folded filter, discarding the first runnings. This 100cc portion was then diluted to 500 cc, and aliquot parts taken for the analytical det:rininations. T h e chlorine was determitied bv titrating with a carefully standardized silver nitrate ( N , IoAIgNO3’) solution, using potassium chromate as indicator. T h e sulphates were estimated gravimetically as barium sulphate, in the usual iiianiier. I n some cases the calcium was also determined, but as nothing especial was thus gained, this last procedure did not seem to be advisable. For a constant temperature bath a large washtub filled with ((
559
Sohbilz'fy qf Gj@siiiit
water was used. ITith such a large volnme of water the temperaiure changed quite slowly, and by passing in a stream of water or steam, as the case might be, no serious difficulty was experienced in keeping a temperature constant within a degree or a degree and a half Celsius. T o attempt greater accuracy in this clirection was not jiistifable in view of the results obtained, and the advantage deri\-ed in being able to work with large volumes of solution in a bath of this size far outweighed any disadvantage arising from a variation in the temperature of the magnitude described. Determinations made at 15" C T h e first series of determinations obtained were made a t 15' C. They are given in Table I., in which the first coluinn contains the amounts of sodium chloride in the solution, assuming all the sodium and chlorine to be thus combined, and the second column containing the correspot:ding amounts of calciuin sulphate found to be soluble. A much more complete set of determinations was then made at 2 3 " .
TABLE I. ____
NaCl and CaSO, in water at 15" C _ _
~
-
~
Grams XaCl p t r liter
Grams CaSO, per liter
0.6
2.3
1.1
2. j
j .1
3.1 3.7
10.6
31.1 51.4 159.9
4.8 5.6 7.4
Determinations made at 23" C T h e data given in Table 11. are the results of experiments made with more than ordinary care. T h e calcium sulphate used was prepared by pulverizing selected gypsum crystals. Other series made at this temperature with calcium sulphate from various soiirces showed most satisfactory agreements, and it is not deemed worth while to present them here. T h e results
given in this table are presented graphically in Fig.
Fig
I.
T h e or-
I
dinates represent reacting weights gypsum per liter of solution, and the abscissas the corresponding reacting weights of sodiuin
TABLE 11. S a c 1 and CaSO, in water at ---.--__
Granis KaC1 per liter _ _ __ . _
~
I
_ 1-
0.99
4.95 10.40 30.19 49.17 75.58 129.50
197.20 229.70 306.40 315.5 5l
Reacting weights KaC1 per liter ~ _ 0.017 1
1
I
,
1 1
0.0852 0.1790 0.5200
0.847 I 1.3018 2.2314 3.3965 3.9568 5.2785 5.4345
I
Grams CaSO, per liter
23'
C
~_ Reacting weights CaSO, per liter
_
-
Grams gypsuni per liter -
~~
2.37 3 02
0.017 6 0.0223
2.Y9 3.82
3.51 4.97 5.94
0.0262
4.48 6.31
0.0368
6.74 7.50 7.25 7 -03 j.68 5.37
1
0.0439 0.0499 0.05 jj 0.0536 0 Oj20
0.0420
7 .j 1 8.53 9.42 9.17 8.88 7.19
6.79 chloride. For convenience in reference, the corresponding values in grains per liter and for anhydrous calcium sulphate are also given i n the table. By referring to Fig. I it will be observed that the solubility of the calcium sulphate or gypsum reaches a maximum in solu0.0397
The solution i n this case w-as in contact xvith both gypsum and sodiuni chloride in the solid phase.
tioiis containing about 135-140 grains of sodium chloride per liter, upward of 7.5 grams calcium sulphate - equivalent to 9 . ~ grams gypsum -being dissoll-ed at this concentration, against approximately z grams calciuin iulphate, cqiii\ d e n t to 2 . 7 grams gypsum soluble in pure water. In ryiew of the generally accepted ideas 011 tlie mutual effect of electrolytes in solution, the existence of this ~iiaxiiiiiim solubility point was entirely unexpected and surprisi1ig.I T h e data given are calculated with reference to a coiistant voluiiie of solution. I t seemed possible that the relatively large ainouiits of salts present in the more coiiceiitrated qolutions had replaced so much of tlie volume of the solvent in this constant voliimeof solution that the existence of this iiiaxiinum solubility was onl?. apparent atid not real. Experiments to test this point nil1 be described presently. I t was deemed advisable to determiiie esperiinentally whether or not the composition of the solid phase, containing both sodinin chloride and gypsiiin, would affect the co~iiposition of the solution in contact with it. Therefore two mixtures coiltaining. respectively, 45 grams of sodiuin chloride with I grain of gjpsiim and 45 grams of sodium chloride with I O graiiis of gypsuin, were each brought into contact with IOO cc of distilled water. T h e mixtures n-ere shaken continiiously for four da) s, then placed in the constant temperature bath and allowed to settle. analysis of the supernatant solutions gave identical results for both cases. As there was no reason n $rio?-i to expect a different rrsuit with this pair of electrolytes 110 further experiments on this point were deemed necessar? , Determinations made a t higher temperatures I n Table 111. the data obtained at higher temperature^ are given. T h e results are not preiented in graphic form liecause the curves would lie so close together, when plotted on all!. convenient scale, as to fail to bring out more cleat-1)-than the tables do any points especially worth noting. -
~~
’ It had been observed, SOC.38, 335 (188j).
however, by Tilden and Shenstone. Proc Roy
TABLEIIX. -
XaC1 and CaSO, i n water at various temperatures
___ 300
-
~~
c
Grams Grams SaCl CaSO, per liter per liter
52’
Grams NaCl per liter
C
700
Grams , Grams CaSO, SaCl per liter per liter
c
szo
Grams Grams CaSO, NaCl per liter per liter ~ __
c Granis CaSC), per liter
_____-__--
2.2
0.0
2 ‘27
34
1.0
2.18
4.9
5.0
2.65
58 7.4
10.1
29.5
3.30 4.65
7.6
48.8
z.j.+
74.9
6.23
128.7 195.1
7.00
-
7.jI
By referring to the tables it will be observed that the curve for 30’ C \vould lie above that for 23’ for the rnost part, but the curves for 50’ and the higher temperatures, respectively, would fall lower and lower on tlie chart. ,It first sight this might be attributed to the fact that gypsum slion-s a maximuin solubilitjr in pure water at about 37. j ” C, as may be seen by referring to llarignac’s results.’ But that this fact alone is not sufficient to account for the results noted seeins probable when it is observed that the solubility of the gypsum in the sodium chloride solutions, even at go’,inay be increased to nearlj- 3. j times its solubility in pure water, whereas tlie solubility of the calcium sulphate in pnre Ivater is 37.5”, the ~naxiiiiiitnpoint, is about 1.119 times its soliibility at o‘, and abont 1.243 times its solubility at 99’, as calculated froin \Iarigiac’s data. T h a t a cansal relation exists between tlie facts observed is to be expected, but what it mal- be is not obvious at present. X s far as was observed, the curves detennined at teinperatures above 30’ C were asymptotic, and did not show the existence of maximum” points. T h a t such points may exist is of coiirse possible, but if so they innst lie near the points indicating ( &
the extreme solubility of sodium chloride in the system. T h e labor involved in attempting to locate them did not appear to be justified a t the time the curves were determined, aiid the purposes for which this inyestigation was undertaken has not since indicated the desirability of a more esliaustive esainination for the 111. Determinations made at 2 6 O C I n this series the deterniiiiatioris were made 011 solutions of the volutne of a small pycnometer or specific gravity bottle, arid in each case the w t u a l weight of the solutions was deternnined. Having deterinined the weights of the salts present. the \T eights of water actually preseiit as solvent were then calculated bj. 511btracting tlie weights of the salts present from tlie total weights of the solution. T h e ~ o l u m eof the pycnometer a t 26" \vas carefully determined by boiling it for some time in distilled water, coolitig to 26') weighing it filled, carefull) d r ~ i i i g aiid , weigliiiig it empty. T h e data obtained fol1ow-s : Grami
lyeight of pycnometer filled lyeight of pycnonieter empty
I 3Lj.53f5~
-small. I n this case about eight days were required for resolution to be complete ; in the two other cases about sixteen aiid twentytwo days, respectively, were required, but in all three cases the re-solution was complete in the course of time. I t thus appeared that the formation of a different hydrate above SO' was a gratuitous assumption ; the more probable explanation of the facts observed is that the rate of solution of gypsiini crystals is very slow, eLTen when large amounts of sodium chloride are prcsen t .I Since the experiments just described were made, a paper by \-ai] 't Hoff and Armstrong" has been published, from which it appears that the transition point for the two solid hydrates, CaS04.2H20and CaS04.>$H20,is 101.45' when in contact with a solution of sodium chloride of considerable concentration. With certain other substances however, this transition teiiiperature is much lower. For example, these investigators found it convenient to prepare the hj.drate, CaS04.$H20, by keeping g j psuiii, CaSO4.2HzO,in contact with concentrated nitric acid a t 40' for some time. It might be supposed that this transition takes place at this lower temperature on account of the strong affinity of the nitric acid for water. T h e results of Vater,3 rather tend to confirm this view. H e found that from saturated solutions of sodium chloride at all ordinary temperatures calcium sulphate separated in the form of gj.psum, CaS04.2H20. T h i s Since this work was done, Ostnald (Zeit. phys. Chem. 34, 4 9 j (1900') ) has suggested t h a t a substance may have different solubilities. depending upon the size of the particles of the solid in contact with its solution, aiid Hulett (Zeit. phys. Chem. 37, 355 (1901)) has rerified the fact for gypsum. I t is believed. however, that it has been amply demonstrated that i n these gypsum solutions final equilibrium conditions will be the same, and it is the rate of solution which has been the main disturbing factor hitherto i n their study. Sitzuiigsber. Akad. XViss. Berlin, 28, j59-576 ( 1900). 'j Sitzungsber. Akad. Wss. Berlin, 28, 269-294 ( 1 9 ~ 0 ) . See also Zunino. Gazz. chini. Ital. 30, I., 333 ( I F ) .
hydrate also separated from solutions coiitaitiiiig as much as 3 percent of magnesium chloride ; hiit froni satiirated solutioiis of magnesiniii chloride gypsum, CaS04.2H20,separated below 40°, and the hydrate, CaSC)+.>I H 2 0 , above 40'. 3lagiiesiuiii chloride has a strong tendency to crystallize under such conditions as tlie liexah~-drate ; that is to say, shows a strong affiiiitj. for water. From these facts it would seem tliat g!.psiiiii is not a stable hydrate above 40" iii the presence of other substances which are strongly deliquesceiit, but tliat in the presence of a neutral substaiice like sodium chloride, which does not foriii a hydrate either in the solid or liquid pliase, it is stable u p to a temperature of 101.45' C. I t will be recalled that tlie iiiaxiiiiiini solubility of .calciiiiii sulphate in water is a t about 40' C (more esactl). 37. jc), and tlie idea snggests itself that perhaps this is tlie true transition point ; the snpposed solubility curve for gypsuin froiii 0" to 100' being in realit!. t\vo solubilitj. carves- oiie for the dihydrate and the other for t h e hemili>.drate-iiitersecting a t 37.5", both being so flat the!. rnii into each other as though but one curve. Rut this view is negatived by the fact that the transition does not take place in contact wit11 socliuin chloride solution until a teinperature of 101.45' ; and, as will presently be showi, it does not apparently take place a t all in a perfectly indifferent medium a t any teiiiperature below that sufficient to drive out all the water from tlie stibsiance. Further, the transition of gypsum to tlie heiiiihjxlrate wlieii in contact with pure water above 40' has not !,et beeii observed. T h e statements of vaii 't Hoff and .\rnistroiig and of I'ater were confinned by experinieiits in which soiiie cleavage pieces of a very fine g\.psiiiii crystal were p.it into the various solutions a t the teinperatures stated. b'lieii trailsforination to the hemili>.drate, CaS04.>< H 2 0 , took place, tlie surface of the gj'psiiiii sooii became veri- white aiid opaque and the surface rouglieiEd, but after standing soiiie tiiiie, if dropped into water, the opacit!. disappeared and the substance gradually became perfectly clear atid traiisparent again. T h i s traiisformatioii froin oiie hydrate
to another conld be very beau ti full^. and readil!. follon ed when the cleavage piece was in contact with sodium chloride solutions, for the changes were theii fairly rapid. I t seemed north while to 5ee if the transition temperature would be the same when the 11) drate was in contact with other presumably neutral substances. Therefore a clear transparelit cleavage piece of g ! psuin crystal was dropped into a test-titbe containing melted paraffin and tlie tern per at 11re gradual 1!. raised by i inin ersi ng i 11 a water-bath to which sodiuiii chloride was added from time to time to raise the boiling teinperature. S o transition to the hemihydrate appeared when the temperature reached 101.50' C. T h e teinperature was then raised to 104' and kept constant for about six hours. KOformation of tlie hemihydrate could be observed, and the temperature was then gradually raised to I 15' witliout any change in the gypsnm being apparent. T h e test-tube containing the paraffin and gypsum was then withdrawn from the bath and quickl!. heated to about 145' C over a Bunsen lamp. ;It this temperature the cr)stal rapidly lost water, cruinbled to pieces, and the residue appeared to be the anhydrous salt, CaS04, although a more thorongh analytical examination would he desirable before making this statenleiit positively. From these facts it is apparent that the equilibrium between the 1 1 drates ~ of calciiiin sulpliate needs further stiidj-. But the work was stopped a t this point, as further investigation in this direction \vas scarcely pertinent to the subject of this paper. Solubility of gypsum in pure water T h e solubility of calciuin sulphate in pure water has been frequently determined, but the published resnlts are ver!. coiiflicting. 'I'o account for the discrepancies most astonishing statements as to the supersaturation and under saturation, etc., are recorded. The real cause of the difficulty seenis to 1iai.e been the very slow rate of solution, the time necessar!. for the obtaining of final eqiiilibriuin het\veen the solid solute and solution being relatively very long. For the sake of a coinparisoii, some of the principal determinations a t temperatures near that a t wliicli we worked are presented in Table 1'1.
TABLE 1-1. Solubility of gypsum in pure water
1
Parts 'water re'quired to j dissolve I part Temper- calcium ature sulphate ~
~
Authorities
~
,
Parts water required to dissolve I part gypsum
?vIarignac' Poggiale' Church' Cozza' Droeze' Goldammer' Kohlrausch and Rose' e From data of this paper T h e first column gives the authority, with reference ; the second column, the temperature at which the solubility was determined ; the third column, the parts of water required to dissolve one part of calcium sulphate, calculated on the basisof the anhydrous salt ; and the fourth column, the corresponding calcolation on the basis of the dihydrate. Droeze seems to have given inuch attention to this solubility determination and, besides making experimental investigations himself, has given a critical review of the work of his predecessors i n this direction. H e objects to Poggiale's figures, that they were probably obtained by working with a supersattirated solution. T h a t this could have been possible seeins absurd, and
' A n n . Chim. Phys.
( 5 ) I , 274 (1574).
Ibid. ( 3 ) 8, 469 (1843). Jsb. Chem. p. 192 (1S67). Gazz. Chim. Ital. p. 135 (1873). Ber. chern. Ges. Berlin, IO, 330 (1877). Chem. Centrbl., p. 70s (18SS). Zeit. phys. Chem. 12, 241 (1893). Calculated from the published data. Under the circumstances the agreement with Marignac's results is remarkable.
Solzibiliiy of Gypsziiiz
573
it is much more probable that Poggiale did some very bad analytical work, for his restilts are plieno~nenallyhigh. Again, Droeze objects to Church’s results, that they were obtained froin unsaturated solutions. T h e reasons for this conclusion are vague, btit we are inclined to hold the concluslon itself to be correct, for Droezc’s results agree fairly well with those of Marignac, Goldarnmer, and Kolilrausch and Rose, while the fignres of Poggiale, Church, and Coaza show wide variations. T h e solubility, as determined from tlie data presented in this paper, is somewhat greater than is indicated by the figures of tlie majority of the investigators cited, but not quite as great as found by Droeze. No special care was observed in obtaining this determination, for at the time the experiment was made i t was regarded only as an incident in tlie gathering of data for the solubility curve and not as a point under special investigation. Kevertlieless, it is believed that tlie value as gi\Ten here is approximately correct, for it is not easy to see in what respect the details of the experimental determination could be bettered. There can be little or no doubt that final equilibrium between the solid salt and the solution had been reached. T h e analytical operations involved were carefully performed and tlie calculations were based on the actual weights of the solvent, as well as the solute found to be present.
Theoretical Discussion T h e effect of one electrolyte upon another in aqueous solution can be accounted for by the electrolytic dissociation theory, at least qualitati\-el~-. Attempts have not been wanting to forninlate a quantitative expression for the phenomena observed. Nernst’ and Noyes,* with his pupils, in particular have attacked the problem with marked SLiccesS by applying tlie mass Zaw of Guldberg and Waage to the particular types of equilibria I
Zeit. phys. Cheni. 4, 372 (1889) ; Theoretical Chemistry (Palmer), p.
406. Zeit. phys. Cheni. 6, 241 (isgo); g, 603 (18923 ; 16, 125 ( 1894) ; 26, 152 (1898) ; 27, 267 (1898) : Jour. Am. Chem. S O C . 19,930 (1898) : 20, 194 742, 751 ( 1 8 9 8 ) ; 21, 2 1 7 , 5 1 1 (1899).
574
/.i*lZJZb k:
~‘lZiJl~J~OJ2
between electrolytes and their resu!ting ions which may be presented. For the case where a pair of binary electrolytes reacts mitli the formation of another pair of binary electrolj tes and four ions tlie theory involves the solutions of nine simultaneous eqiiationr containing nine variables. For particnlar cases eight of these variables are determined by various means, and in consequence there is obtained one equation in oiie unknown qaaiitity, but it is of a high order ant1 involves a very large iiuiiiber of terms. In all the cases so far studied they were so selected that certain restrictions were imposed, arising froin the nature of the particular cases themselves, which greatly facilitated their study. Certain special assumptions could he made by which the algebraic expression of the relation betn eeii the quantities iiivoll ed would be verj iiiuch simplified. In the case which is the subject of this paper no such assumptioiis appear jnqtified. In the first place, the concentrations in;.olved are far from what may fairly be considered dilute. -111 the substances involved in the metathesis CaSO, - SaCl CaC1, 4 NalSO, are S ~ I - O J electrol>,te\ Z~ to which the Ostwald dilution law does not appl! - that is to sa!, to which the mass law is m t applicable as a statement of the equilibria between the quantities of dissociated ions and iiiidissociated salt through any range of concentrations. T h e ionization constants, which appear in the equations to which reference lias just been made, can not be obtained, therefore, in any way wliicli we can at present see. In the ca5e of the calcium chloride and tlie sodium sulphate supposed to be in the solution there did not appear to be any practicable n a y to decide iii how far tlie dissociation would take place as for a di-ionic electrolyte, or how far as for a tri-ionic electrolyte, with a consequeiit tiiodificatioii of the theoretical statements. These considerations have finall) caused the abandonment of any attempt to make a ccinpaiison betneen theoretically comFuted results and those actiiall! fonnd, and, in the present state of oiir knowledge, the results can be presented only as empirical determinations in the trust that in time the) may find
.SoluhiZify of
[email protected]~~N
575
tise in lielpiiig to develop a inore cotnprelieiisi\.e and satisfactor!. theory for soliitions. There does not appear to be an!. rational esplaiiation for the appearance of tlie maxiintiin point on the solubilit!, curve. One would naturally expect the solubility of the gypslim to continually increase as the concentration with respect to sodiutn chloride increases. Noyes' has brought forward some evidence to shon- thyt as the coiicentratioii increases with respect to clilorides of the alkali earth metals, there is a formation of COIII-
-
plex ions, for example, CaC13 or CaCl+. T h e e\.idence he obtained was rather against tlie assuinption that the sulpliates of tlie alkali metals would dissociate i n any but a iiorinal wa!., as triioiiic electrolytes ; but lie worked with quite dilute soliltions ( u p to a tenth molar), and it does not follow that complex ions would not be found in solutions of the concentrations considered in this paper. I t is difficult to see how the forination of- such complex ions could affect the gypsum in contact with tlie solution, otherwise than to increase its solnhilit). ; and tlie fact that tlie solubilit!. actually decreases beyond the ~naxiiniiriipoint is, as far as it goes, rather to be considered as evidence against the formation of the complex ions. T h e same thing iiiiglit be said as to the suggestion that all electrolytes are considerabl~.11jTIroljzed a t high coiiceiitrations. I t seems more probable that tlie plieiioinei~oiiis connected with the condensation of the solvent and the effect of such condeiisation upon the solution teiisioiis of the various electrolytes, aiid it is eariiestl!. to be desired that some one wit11 time and opportniiities undertake an investigation of this subject. For assistance in carrying out the experiiiieiital and ana1J.t. ical details of the work described in this paper, t!ie autlior wishes to record his sincere ackiiowledgnients to Messrs. Frank I). Gardner and Atherton Seidell.
Summary Some of tlie points brought oiit in the foregoing pages to which attention is especially directed, may be stated as follows : ~-
' Phys. Re\. .. 1 2 , I j ( r g o i ) ; Joiir. Am. Chein. phys. Clieiii. 36, 61
(1901I.
S O ~ 23, . 37 (1901) ; Zeit.
5 76
Frank K. Crr meJron
( I ) Below 37.5" C the solubility curves for g>-psnm in aqueous solutions of sodium chloride show a inaxiniuii1 point. Above this temperature the existence of maximum points is doubtful. ( 2 ) At 23' C this maxiniuni solubility takes place in a solution containing from 13j to 140 grains sodium chloride per liter. (3) T h e solubility of gypsum at 23" in a solution containing 1 3 j grains of sodium chloride per liter is about 9.3 grains per liter. T h i s is equivalent to about 7.j grams of anhydrous calcium sulphate per liter. (4) T h e solubility of gj-psum in solutions containing less than 140 grams of sodium chloride per liter is very little affected by change of temperature. (5) Alaxitnum points in the solubility curye still persist, even if the solubilities be calculated on the basis of the weight of solvent present rather than the volume of the solution. (6) condensation of the waier as solvent takes place when gypsum and sodium chloride are brbught into solution. (7) T h e rate of solution of gypsum is very slow, and a t ordinary temperatures the time required for equilibrium to be bronght about is often great. This fact accounts for the apparent insolubility of gypsum crystals in that volume of solution from which they may have been originally obtained. (8) T h e transition temperature at which the dihydrate and hetnihydrate of calcium sulphate are in equilibrium appears to be dependent upon the nature of the substance or substances in the solution ~ i t which h they are i n contact. (9) At 26" the solubility of gypsuiii is about I part in 372 parts of pure water. ( I O ) T h e specific gravity of a saturated solution of gypsum in pure water is 1.0026 at 26' C, and 1.0031 a t 31'. (11) T h e application of the mass law, as formulated by Guldberg and Il'aage, to the phenomena of electrolytic dissociation fails to account for the observations here recorded. (12) T h e existence of a maximum point on the solubility curve is not satisfactorily accounted for by the hypothesis of the existence of complex ions in the solution. I t is probably connected with the condensation of the solvent in aqueous solutions of electrolytes.
