THE SOLUBlLlTY OF CALCltJM SULPHATE IN

SODlUM CHLORlDE AND SEA SALT SOLIXIONS

E. FURBY. E. GLUECKAUF AND L. A. MCDONALD i3crni~lr.v Division, Arnmic Energy ftesearch fitubfishment. Ifarwdl (f3t~hrnd~

(Rtrccivcd July 19. 1967)

t%c results of cxpcrimcnts on the soI?bility of cikium sulphatc in sodium chloride solutions and in simuhted sea water SOhtiOnS iit 1ctTlpCntUm Of Up IO Imc are giwn. They arc COmpatd with those uf simtiarexp&mtnts by other workers and found to be in wtisfitctorv agrcemcnt. A tabulation of the best moiaI whtbility products of C&501 arthydritc in sea s&t soluiions for the tcmpcmturc ran*= from 50 10 ,lNOC is giwn in the appcndir.

I_ INTRODUCFION

The solubility of calcium sulphate and its hydrates in sodium chloride solutions, and in sea water has been studied over a wide range of temperatures and concentra- tions (I-9).

The above mentioned investigations show that above 45C approximately, C&SO, (anhydrite) is the stable form_ and that below this temperature CaS0,2H20 (gypsum) is stable. Above S3C the mctastable iorm CaSO~{H20 (hemi-hydrate) is more stable than the dihydrate, though it is less stable than anhydrite. However, it appears &at nucleation of CaSO, (anhydrite) even above 45C does not take ptace spontaneously, and so. if a saturated solution of the dihydraie is heated above 45C, it does not precipitated the anhydrite. On the other hand, if the temperature is raised beyond 93C. and if its solubility product is exceeded, the metastable hemihydrate will readily precipitate. Once the solid hemihydrate has formed, a very slow change to the anhydrite phase takes plxe. (These transition temperatures vary slightly with the ionic strength of the solution.)

A knowledge of the solubilities is of considerable interest for the distillation of sea water, as c&&n sutphate precipitated at high temperatures forms an insoluble deposit on heat transfer surfaces, thereby affecting adverseiy the efficiency of the heat input sections.

in salt soiutioas of higher ionic strength, the calcium sulphate solubitity product is considerably increased. It is generally assumed that, as far as the solubility product of CaSO, is concerned, sea salt soh~tions behave like sodium chloride solutions of identical ionic strength. Careful study of the data recorded in the literature leads

264 Desalinrrlion, 4 (1968) 264-276

SoLUBlLlTY OF CI\LCIUM SUI.Pt!ATE 265

one to suspect the validity of this belief, since the solubility product is significantly higher for sea salt solutions in excess of I.5 molal. This is shown for equilibrations of the dihydrate at 25C (see Fig. 1) where 12) (sea salt) is compared with (7)

(sodium chloride). A similar ditrerence arises for anhydrite equilibrations at 125C

and 150C when comparing the data of (?,I (sea salt) with the measured data of (6) 1 (sodium chloride)*.

Fig. 1. The wlubility of calcium sulphatc dih>dnte in salt solutions at 15-C.

In the present progmmmc solutrifity experiments under otherwise identical condi- tions have hccn carried out with bcth sodium chloride solutions and sea salt solutions. in the experiments using naturat anhydrire as the equilibrating solid, the ap- proach to equilibrium by precipitation of excess calcium sutphatc was exceedingly slow, In order therefore to be czrtain of the equilibrium concentration, the fatter was obtained both by dissolution in unsaturated solutions, and by deposition of excess calcium sulphate from supersaturated solutions which however were unsaturated with respect to calcium sulphate dihydrate and hemihydrate.

2. EXPERIMENTAL

For the equilibrium of sea salt solutions multiples of a simplified composition were used:

N3Cl - 0.4861 mofes per kg water

MC& - 0.0363 moles per kg water

MgS@, - 0.01813 moles per kg water

This sofution contains all the major components of sea water except calcium suiphate

l Marshall and Slusher, in a recent publication of new results have also established this difference by direct experimentation. (O.R.N.L. Annual Progress Report December 3lsl, 1966). This *as not yet available when this paper was wxWen.

Desaiina~ion, 4 (1948) 264-276

266 E. FURBS, E. GLUECKAWF AND L. A. MCDONALD

(0.0103 molal). This simplified mixture is based on the sea water composition given by Sverdrup, Johnson and Fleming (10) by replacing minor components; K+ was replaced by Na+, and Br- and HCO; by Cl-. Identical equilibrations were then carried out with sold&m chforide solutions and with these sea salt solutions of comparabIe ionic strength.

Tlte solid seeds for 25C equilibrations were gypsum (C&O,. 2H,O) and. for the higher temperatures two types of anhydrite designated as A and B, Type A was freshly pr-ecipitated gypsum converted to hem&hydrate by heating in an oven at approximately 130C and finally converted to anhydrite by boiling in 3N sodium chloride solution for several days. Type B was natural anhydrite supplied by Gregory and Bottley of London and ground to pass 180 mesh (c 85 p dia.) The fines were removed partly by elutriation and partly by extensive boiling in sodium chloride solutions. X-ray powder diffraction analyses gave identical patterns for each type of material; both corresponding to the dead-burnt insoluble variety of anhydrite (p-C&O,).

The calcium sulphate solubility products of the two materials were not signiticantly different but equ~ib~um was achieved more quickly with the Type A material, which had a much greater surface area. fn general equilibration experiments were ailowed to proceed for 2 week.

3. ANALY!%

In the absence of magnesium+ a 5 ml aliquot of the calcium solution, dilutrd with water to approximately 50 ml, was butTered to pH 10 by the addition of 2 ml of an ammonium chloride/ammonia buffer and titrated with O.OlM E.D.T.A. using solochrome black (e&chrome black) as iadicator. Where magnesium in significant amounts was present, as in the sea salt solutions, a modification of the above procedure was necessary. The solochrome black indicator was replaced by calcon and the E.D.T.A. titration was carried out at a pH of 12.3 obtained by the addition of diethyl- amine (5 ml per 100 ml of solution). The latter reagent precipitates the magnesium quantitatively as the hydroxide after which the solution may be titrated to the magenta/ blue end point very similar to the solochrome change. There is in this case some little co-precipitation of calcium with the magnesium, and some adsorption of the indicator on the hydroxide precipitate but the degree of reproducibility did not appear to be affected to a marked extent.

Other analyses, in particular the SO; - and Cl- contents, were carried out when required using standard procedures.

.

The results of the equitibrations are given in Tables I-V and these data are compared with those from other sources in Figs. 1,2 and 3. The data from Tanaka, Nakamura and Hara (2) have been utilised on the assumption that the ratio of excess sulphate (So,- - - Ca+ +) to chloride in rheit sea water had the same value as that given in

lksalination, 4 (1968) 264-276

SOLUBILITY OF CALCIUSI SULPHATE 257

(II/, i.e. 0.0322 - this is very similar to the value obtained from (10) which gives 0.0328.

towc LT=EhoI* 0s f3L1 SCLT~OY GVOLk1

Fig. 2. Ihe solubility of mhydritc in sait solution at 7SC.

Fig. 3. The solubility of anhydrite in salt solution at IWC.

TABLE I THE x)LIJBtLITY OF CALCIUM WLPHATE DIHYDRATE

lr* SDDNM CHLORIDE SoLUTtOFis AT 25c

2.62 2.92

4.96 5.12

K, = solubility product or mass product.

Desalination. 4 (l%S) 264-216

268 E. FURBY, E. GLUECKAUF AND L. A. MCDONALD

TABLE II

THE SOLUBILlTY OF CALCIUM SULPHATE DIHYDRATE IB

su SALT soLInloss N-FOLD CONCENTRATED A-r 25%

N TOffli

ionic SZrel&?fh moial

K. x loi moiaf

1.0 0.79 16.1 2.74 z-g 30.4 3.5 4.0 2:90

34.5 37.7

7.0 5.12 32.7

,

TABLE 111

-WE SOLUBIUTY OF CALCIUM WLPHATE (XATUluL AsHYDRtTF) IS SODfUM CHLORIDE SOLUTIOSS AT 75C

r&al N&l sohrlion mofai CaSO4 conrent mofal iofric slrengzh molol u, Y 101 InofaI

0305 0.0246 0.606 6.05

1 .o:! 0.0314 I.14 1.23 o.ou9 1.37 * f:S 2.08 0.0386 2.23 14.4 3.17 0.0368 3.32 13.5 4.z 0.0342 4.46 11.7

TABLE IV THE SOLUBILITY OF CALCSUH SULPHATE

SOLUBILITY OF CALCIUM SULPIIATE 269

TABLE V

T-HE 5OLL;RILIN OF CALCWJM SlXPi1ATE fPREPARED ANIIYDRITE) IN SODlUbl CIILORIDE AS0 SEA SALT SOLUTIOSS AT lw-c

N&l CasoI rutat K. x IW St-u salr CamJ Tozaf K, x 10 solurion content iO& nrolof :: tv content ionic mulai fnolal .nwlai strengrh v $02 smen.qrh

1. 102 fnolai tnuiai nmlai __ .-__... -. . ..-_ __ _. ._... .- ._ _._.

0.666 2.17 0.753 4.70 1.01 1.43 0.733 4.67 1.66 2.95 I .9R S.:O 2.825 1.43 1.98.5 9.:ti 2.40 3.07 2.52 9.40 3.61 1.35 zss IO.86 2.76 2.91 2.88 8.45 4.08 1.29 2.88 11.53 S.QZ 2.63 5.13 6.91 7.02 0.88 5.10 12.6

5. DISCL%SION

Fig. I shows that at 2sC there is a good agreement between data obtained in the present work and those of Shternina (7) and Marshall and Stusher (9,) for sodium chloride sotutions, and between the present work and the data of Tanaka, Nakamura and Hara f2f for sea-salt solutions. It also shows that sodium chloride and sea salt solutions of the same ionic strength definitely give different solubility products

KS -= ([Ca] [SO; -]I in solutions of concentration greater than I.5 molai. This difference increases with the ionic strength. It might be suspected that the presence in sea-water of excess sulphnte reduces the activity of the calcium and that the presence of magnesium reduces the activity of the sulphate in solution thus giving higher solubility mass products for calcium sulphste.

In the case of magnesium chloride this effect is shown clear@ from the calcium sulphate solubility data by Shternina (7) in magnesium chtoride soiutions. Comparing calcium sutphate solubiiity products at the same ionic strength, the solubility products in magnesium chloride solution were shown to he substantiaiiy higher than those in sodium chloride, the difference increasing with ionic strength. Those in sea salt SO~U- tions take up a position between the two others.

At 75C and IWC this difference between sodium chloride and sea salt solutions is even more strongly marked than at room temperature (Figs. 2 and 3. reqxctively). These figures give the solubility of anhydrite in solutions of sodium chloride and sea salt by plotting molal solubility product against the molal ionic strength (3 ~mq

270 E. FURBY, E. GLUECKAUF AND L. A. MCUOSALD

6. CONCLUS;OF;S

In conclusion, the data obtained from sea salt solutions at 25C and 100C in the present work have confirmed results estabfished by Tanaka, Nakamura and Hara. The muIts of the tatter authors at higher temperatures can be shown to folfow

closely the pattern which might be predicted from a consideration of the behaviour of activity coefficients in electrolyte solutions, and can therefore be accepted with all the more confidence. The results of the present work show also that the soiubiiity product of calcium sulphate is higher in sea water solutions than in sodium chloride solutions of cquivalcnt ionic strength when the latter is in excess of 1.5 molal.

The assistance of Mr. 0. gratcfuIIy acknowledged.

ACKNOWLEDCt3WNT

Flint. who carried out the X-ray diffraction analyses. is

I. E_ P. PATNIXX ASV A. H. WHIZ-E. J. Amer. Chem. Sot., 51119?9) 360. 2. Y. TAXMA, K, NAYAMURA AKD R. HARA. Kogyn Kagaku Fashi. B(1931) 779. 3. 7. Touruur. T. KLWWAR~; A&V R. HARA. Kogyo Kagaku Fd.whi, 34 (1933) 1651. .a. F. G. STANDWORD AND H. F. BJORK. A&. Ciwvn. Ser., 27 (1960) I IS. 5. F. G. StANotfom ASD J. R. SISEK. Chem. .&I!. progr.. g (1961) 58. 6. w. t. MARsHALL, R. SLUWER ASD E. V. Josy J. Chcm. and Eqq. Dam. 9(z) (1%) 187. 7, E. B. SHTERNINA, Dokkdy .&ad. Xa& SSSR, 60 (19-W 247. x. Y/. F. LAIGELWR, D. H. CALDWELL AF;D W. 8. LA~RESCE. Ind Eqq. Chum., 32 (1950) 126. 9. s). t_. .%~ARsuALL &SD R. SLUSWER, 1. P/p. chum.. 70 (1966) #fs.

IO. H. V. !%ERDRUP, M. W_ Jofrxsos AND R. H. FLEMISG, 77re Uceuns, Prentice and Half, Inc., MY* (I946).

Il. Y. MATUNO, 1. KUCANEMMU AND R. HARA, Clwm. Sue., lirpan. Ind. Eng. Section, $I (1941) 142.

fL N. F~JERR~W, Z. .E&trrochcm., U (i918) 32t. 13. E GLUECKAUF, The Structure of ETecrrofy~ic Solurionr (Ed. W. T. Hamer) J&n Wily &

Sons Inc.. (1959) p. 97.

Desalinutiott. 4 ( 1968) 264-276

SOLUBfLlTY OF CALClVbf SULPHATE 271

APPENDiX

c&o., SOLt'ltLtTl~~ OF TANAKA, !%RSMUfU &Xl3 ihRA

These experimental observations not only agreed satisfactorify with the authors own observations at 25C and 1OOC. but showed also an excellent consistency when data on anhydrite, hemihydratc and dihydrate were compared with each other.

The activity sotubility products can be expressed in the form

where 7 is the mean molal activity cocfticicnt of CaSO, in the saline solution,

(I, is the water activity of the saline solution II is the hydration number of the solid (O,& and 2)

Then, if orte compares the moEA soiubility products mcJ - ~~ts04 - tr,: for different hytfratcs at the same ccmpetaturc and for the same saline solution, the activity CO- efticicnt j of the CaSOa in the solution is identical for these, so that constant ratios of this product ought to be oht;&uzd. This constancy is obtained (within ;I few per cent) where the obssr\ed datit permit this comparison; at SW and 75 for anhydrite and dihydrstc, and at 100 and 125C for anhydrite and hemihpdrate. (An exception to this agrccmcnt vvcrc the data zit the very bighcst concentrations (chlorinity 14SO],o) H here for unexplained reasons this value ditkred conside~bl~ for anhydrite and dihydratc from that found for salinities of from 16 to l~~~~*. The water activity was computed on the basis of - log tt,,. =: x stat +Q, ,,jSS which leads to the formula :

- Jogo, .:- i 0.010;s I i O.WOSS f2 (W

In order to earrj- out these comparisons, the soiubility data of Tim&a, Naknmura and Hara, originally given in wt ~~C$X3+ as function of the chlorinity given in wtO! rno of Cl, [Cl] had to be converted to the modality scak (Irr,mofs/kg H20) using for the weight ratio of total salt in sea water to the weight of chlorine a value of

R, = 1.818

The molality of excess sulphate (,A) was based on the molar ratio in sea Water

R, = SO;- =a = 0.0323 (A31

A nominal ionic strength I was also defined on molal basis:

272 E. FURBY, E. GLUECIAUF AND L. A. MCDONALD

As the data observed by Tanaka, Nakamura and Hara were rather widely spaced as regards the salinity of their solution (16.29, 49.64, 97.24 and 118.06 wt o/00 chlo- rinity). it was decided to rc-compute the table of solubititics, smoothing any small deviation by means of the activity product equation, particularly as no indication was given by the Japanese authors of how they carried out their intcrpofations. Bearing in mind that the activity coefhcients in Eq. (Al) at high concentration arc usually better represented by an equation of the tyxz

1% . =z - Az,z,f. + Bf GW

thrn b:r the conventional extended Debye Huckel equation (II. 12) WC intended to use for the smoothing operation a plot of

Q = logEi, i trlo~.~,,. - Rrl,t If3 (AN

against I.

where SA, -.= 1.98 n,,aos DT- had the va!ues given in Table AI.

In the cz~sc of the dihydrate this plot gives paraltel almost straight lines for all tem- peraturcs.

Thus:

log INcz * mso< .+Zlogn., -=Q,O +8.4,1 3-0.326f -t-O.166 x lo-I

Yalucs of Q~arc given in Table Al.

I 0 25 50 75 100 15 150 57s1 200

.-W i .927 I.990 2.046 2.128 2,224 2.316 2.428 2.5,% 2.700 Q&r=0 4.522 J-785 5,092 5.430 5.785 6.142 6.485 f?&n = 1 (85 ;) 4.48 4.68 4.97 5.30 Q&n = 2 4.386 4.392 4.458 4.62

But for the anhydrite, it was found that slightly curved plots were obtained. On the other hand a plot of Q not against I, but against the chlorinity in weight OfoO reduces the curvature and gave effectively straight lines of constant gradients, very suitabfe for interpolation. The mass products of the observed anhydride could therefore be accurately represented by an equation

I logmc. * m,, = Qp- 1.379 x 10 [CI] -!- 8AJ I3 i A (A7)

Desahation, 4 (1968) 264-276

SOLUBILITY OF CALCIUM SULPHATE 273

where A is a small correction. The significance of Eq. A7 is merely that a not strictly linear term in I of Eq. 5 for log y could be repiaeed by a linear term in [Cl], the coefficient of which moreover happened to be independent of temperature.

Fig. Al shows plots of Q versus [C!] using the experimental results of Tanaka, Nakamura and Hara. The values of Qrextrapoiated from this plot for [Ci] =0 are shown in Fig. AZ together with the values of log K,, of Marshall, Slusher and Jones, uhich though obtained from NaCt solutions should lead to a similar extrapolated

65

60

-002.D2SC

50

nPrnrOl?lTE EATA (.I\)

X HEhWiYCIRATE 04TA (HI

@ DIHYORATE DATAID) NAKAUURA

0 20 40 60 60 100 I20 I40 I60

CHLORINITY I..

Fig. Al. Plot of Q (from EZq. A63 against chlorinity, showing straight lines of constant gradient.

iksaharion, 4 (1968) 264-276

274 E. FURBY, E. GLUECKAUF AND L. A. MCDOSALD

-65 0

..I I I, ! I * * : ; , , 50. I00 150

TEYPEAANRE Y

Fig. XL Comparison of solubiiity produc?s at inlinitc dilution, Q of Ref. (2) and Log K. of Ref. (6).

values-as indcvtd they do for values below 1 WC-for the anhydrite. hemihydratc and dihydrate data.

The re-computed data are given in Table Ail for log KS, and in Fig. A3 for the molal solubilities of Ca for the whole range of temperatures and chiorinitics.

For the range of the temperatures from 50% to 150C the original Japanese in- terpolations (after scale conversion) agree quite well with the smoothed data of Table Aff and deviations of 4% or more are rare and mostly confined to chlorinities above Iof * Joa. But deviations are very pronounced over the whole extrapolated range of 175 and 2OOC, where as a rule, they exceed 10%.

Our thanks are due to Mr. A. Gardner for computing tables and functions.

Dcsaha!ion, 4 (1968) 264-276

IA

RLE

A

ll

WM

. W

LUIiIL

l-I-

Y

WX

NX

T

Or

C&

h

~S

llSl)

RI1

1?

IN

SFA

S

ALT

S

OLC

;TIO

HS

_. _

..__

.__.

._

_ _.

_.

__... ._ 10

20

30

40

50

60

70

80

90

? lo

o

0

2

11

0

a

8 2,

I20

13

0

&

h ;;

14

0

Q

I50

8

0.9

23

7.5

I.882

13

.1

2.8

71

1

7.6

I91 I

21

.2

4,9

87

2

3.7

6.1

06

2

5.4

7.2

72

2

6. I

8,4

88

2

6.2

9.7

56

2

5.9

ll.08

35

.5

12

.47

2

4.7

13

.92

2

13

.7

l&43

22

.5

17

.03

2

1.2

18

.70

2

0.0

75

1

00

I 1

0

_ _ .

. . . _

___ .

. -,

. .

_._ ,_.

.._

4.4

4

2.4

3

8.0

4

4.6

2

11

.2

6.6

8

13

.8

8.5

0

IS.9

1

0.0

0

17

.3

II.?

18

.2

12

.0

18

.5

l2.S

18

.7

12

.H

18

.7

13

,o

18

.5

13

.2

lt1.2

1

x3

17

.8

13

.4

17

.3

13

.5

16

.7

13

.6

1.8

5

1.4

1

1 .o

i(

O.H

2

3.6

0

2.8

0

2.1

%

1.7

1

5.2

7

4.1

6

3.3

0

20

62

60

78

5

.41

4

33

3

.49

8,0

5

b.4

8

52

5

42

7

9-0

6

7.3

6

G.O

L 4

.95

9.8

0

H.0

3

6.6

2

5.5

4

IO,2

9

X.5

0

7.0

9

5 .W

lUe65

8.8

5

7.4

6

6.3

8

IO,8

7

9.0

9

7.7

2

6.6

8

1 I,1

9.3

7

8.0

7

IAM

Il.3

9.7

2

HA

4

7.4

4

11

,s

10

.05

8

.60

7

.83

I l-8

1

0.4

6

9.2

9

8.3

0

11

.9

IO.7

4

9.6

1

8.6

6

12

0

I30

.-.

140"

..

_ . . .

ISO

. .

160 ._

0,GJ

0.4

87

I ,3

4

I .0

6

2.0

8

I .6

8

2,8

I ?,

30

3.4

8

2.8

8

4.0

7

3.4

1

4.ii

3

.00

5.1

0

4.3

3

5 4

5

4.7

5

5.7

8

5.1

2

6.1

6

5.5

0

65

7

5.8

8

6.9

4

6.2

9

7.4

0

6.7

4

7,8

3

7.2

4

17

0

18

0

20

0

___

_. ._

. .._-

_ -.

_ .

_....

0,3

77

0

,29

6

0.&

40

0.0

7G

1.3

6

I,11

1 S

H

I.56

2.3

9

20

1

2.u

2

.44

3.3

0

2.8

7

3.7

2

3.2

7

4.1

3

3.6

8

4.5

3

4.0

7

4.0

4

,44

5.2

6

4.8

0

5,'l

O 5

.2h

6.1

4

5.7

4

6.7

2

6.3

6

0.1

87

8

0.4

52

5

0.7

70

5

1.1

2

8

I.48

z 1

.86

5

z

2,2

G

z

2.0

6

E

3.0

5

z f

3.4

4

z

3x

4.2

0

4,6

6

4.7

0

4.7

5

276 SOLUBlLlTY OF CALCWhf SULPHATE

- 2.5

- 2.0

: -15 ! f

P , 1 , * I f ! 0 50 100

tdS 60

wt %, Cl

Fig. A3. Mohf sohbilitics of Ca in sea *mer solutions ~~ccn!rated to a given wt 7; of cklotinc, according to new ircitmtnt of data of Ref. (2) (me=: mols Cas01 per 1000 g of wter).

Desahhtion, 4 ( 196S) 264-276