The Extraction of Potash and Other Constitutents From Sea Water Nittern

8/11/2019 The Extraction of Potash and Other Constitutents From Sea Water Nittern

1/10

9 6

T H E

J O U R N A L O F I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y

Vol. IO, No

I ORIGINAL PAPERS

THE EXTRACTION

OF

POTASH AND OTHER

CONSTITUENTS FROM SEA WATER

BITTERN

By

JOEL

H. HILDBBRAND

Received December 5 1917

COMPOSITION OF SEA WATER

The main constit uents of sea water, besides sodium

chloride, are magnesium sulfate, magnesium chloride

and potassium chloride, together with a small quantity

of magnesium bromide an d calcium salts. Dur ing

th e evaporation of the sea water t o secure sodium

chloride, the calcium present is almost completely

deposited as calcium sulfate,

so

that calcium salts are

practically absent from the mother liquor. By con-

sidering the various analyses of se a water we ma y calcu-

late the relative amounts of the solid salts that might

be obtained by evaporation. The salt works around

San Francisco Bay, with which we are primarily con-

cerned, produce something over

IOO,OOO

tons of sodium

chloride per annum; the amounts

of

the other salts

associated with thi s amoun t of so dium chloride would

be as follows:

TONS

Sodium chloride (NsCl) ......................... 100,000

Potassium chloride (RCl) ........................ 2,800

Magnesium chloride (MgClz.6HzO). ............... 27,300

Epsom

salts

(MgS04.7HzO).

.....................

16,000

Bromine (Br) extracted from the bromides.. ....... 240

A t

this time, when the country is suffering from an

acu te shorta ge of potassium salt s, th e am oun t of

potassium chloride indicated above is of considerable

importance. During the first half of

1917

the total

potash production of t he country, calculated on

the basis of

KzO,

was 14,000 ons, which amount was

but IO per cent of th e normal amo unt used before the

war. I t is evident th at the amount of potash that

could be extracted from th e bitterns of t he salt works

on San Francisco Bay alone would add about IO per

cent t o the countrys present annu al production of

potash. The amoun t of s alt actually produced in

this region is nearly 140 000 ons per annum, so t h a t

a liberal allowance for losses in working up the bittern

should leave still 3000 to ns of potassium chloride.

By utilizing the bitterns from other regions on the

Pacific Coast, notably San Diego, this amount would

be very greatly increased.

The other materials mentioned in the above table

also represent very considerable values, although they

have less relation to the present national emergency.

After th e removal of most of t he common salt in th e

salt ponds, the other salts would be contained in ap-

proximately IOO,OOO tons of b itte rn, having a volume

of approx ima tely

IOO,OOO

cubic yards.

The values represented by these materials, and their

imp ortance both as a nat ural resource of California

and in supplying the country with potash in t he present

acut e emergency, made th e stu dy of th is problem seem

a proper one to undertake at th is time.

of California.

1 This work has been supported by the Council of Defense of the Stat e

SCIENTIFIC BASIS OF METHODS FOR RECOVERY O F T

CONSTITUENTS OF BITTERN

We are ve ry for tuna te t o possess a vast fund

information upon the solubility relationships of t

various salts obtainable from sea water through t

classic work of van t Hoff and his co-workers. Th

work is described in great detail in ube r die Bildung

verh3ltnisse der ozeanischen Salzablagerungen (Le

zig Verlagsgesellschaft, I91 ). During the progr

of t he work two smaller volumes were published

1905

and

1909

by vant Hoff, entitled Zur Bildu

der ozeanischen Salzablagerungen (Braunschwe

View eg).

Inasmuch as very little of this work has been tran

lated into English, and in view

of

the difficulty

interpreting it in its formidable complexity, it see

desirable to give a general outline of it s nature .

Th e solubi lity of a single salt in its relation

changes in temp erature may be represented by simp

diagram s of the t ype fami liar to all trained chemis

4

s

In

I

0 20 4d 60

8 do

120 dG

/b

>rnperature deyrees

cen yrprode

F I G .

1

Fig. I are represented the solubility curves for t

ma in sa lt s with- which we have t o deal, v i z . sodi

chloride, potassium chloride, potassium sulfate, ma

nesium sulfate and magnesium chloride. I n this figu

solubility is expressed as the number

of

mols

of

a

hydrous salt per

1000

mols of water.

Of

cour

other u nits may be used, such as mols or grams of s

in a certain number of grams of wate r or of solu tio

or in a certain number

of

cubic centimeters in solutio

If we know th e solubil ity expressed in any of the

term s it is possible t o calculate

it

in any other term

the density

of

the solution being required where t

conversion is between a weight and volume bas

Th e laws of dilute solutions may fre quen tly be e

tended t o give a n approxi mate idea of th e behavior

con cent rated solutions. The solubility of a giv


2/10

Feb.,

1918 T H E J O C R N A L O F I N D U S T R I A L A N D E N G IN E E RI N G C H E M I S T R Y 97

salt is varied by th e introd uction in to the solution of

anot her salt. The effect of the second salt can be

predicted qualitatively by remembering tha t where the

salt s possess a common ion the solubility of each is

usually decreased by the presence of t he o ther.

If,

however, there is a strong tendency t o form a complex

salt t 5e solubility of one may be increased by th e

presence of the other.

Again, where there is no com-

mon ion th e solubilit y of one is increased by the pres-

ence of the other owing to t he i nteract ion of

the two

salts.

Ther e are various ways of r epresenting graphically

th e solubility relationships of salt pairs. The method

adop ted by v ant Hoff is t o represent the a mou nt of

each sa lt in the solution in ter ms of mols of a nhyd rou s

salt per

1000

mols of wate r, mea sured along tw o axes

at right angles to each other, as illustrated in Fig.

11.

Each curve here represents the composition of

a

solu-

tion satur ated with one component. The intersec-

tion s of th e cur ves represent t he composition of

a

solu-

tion saturated with both components.

A

point be-

tween these curves an d th e origin denotes the composi-

tio n of a n unsa tur ated sol ution. A point outside of

the cur ves would represent a mixture of a sat ura ted

A 7 2 .

,

\c

23

50 7s /OO

MO/Sa c /

p e r

/OOO mo/s HZ0

FIG.

I1

solution with one or both solid salts, depending upon

its position.

O n evapora tion of a n unsa tur ated solu-

tion the relative amounts of the two salts would re-

main the same until the solution becomes saturated,

so

that, for example, a solution having t he composition

represented b y the point

a

in Fig.

I1

would, on evapora-

tion, change in composition as represented b y th e mo-

tion along the line ab. As soon as the curve A B is

reached, representing in t his case th e composition of

a solution sat urat ed with potassium chloride, solid

potassium chloride will separate and the solution must

bec0m.e relatively richer in sodium chloride,

so

tha t a s

the evaporation proceeds from

b

the solution will

change in composition along the solubility curve to-

wards

B.

Similarly, an unsaturated solution having

the composition represented by c would, on evapora-

tion, change in composition as represented by move-

ment along the line

cd .

At

d

sodium chloride would

begin to crystallize, whereupon the solution would

become richer in potassium chloride, its composition

changing along the line

dB.

It is evident that the

final result in th e eva porati on of an y solution of thes e

two salts would be a saturated solution having the

composition represented by

B

changing into a mixture

of t he t wo solid salts.

The effect of tem pera tur e may be indicated on a

third axis at right angles to the others, giving

a

solid

figure, as represented in perspective in Fig. 111.

Where a double salt may be formed, the solubility

relationships at a given temperature would be repre-

sented by a diagram such as that in Fig. IV. This

diagram represents the solubility at 30 of m ixtures of

sodium sulfate decahydr ate, a nd magnesium sulfate

heptahydrate, which form the double salt, astra-

kan ite , NazMg(S04)2.4HzO. Th e middle por tio n of

the curve seen in this figure represents the composition

of solutions sat ura ted with astr akan ite. Solid astr a-

kanite, which contains equivalent quantities of the

two salts, has a composition lying upon a line bisecting

th e angle between th e tw o axes. The composition

of

the solid salt is represented by a point on this line a t

E

expressing th e numbe r of mols per

1000

mols

of

water in the solid salt . The composition of solid

sodium sulfate, N a2S04. 10Hz0, which lies along t he

line O A is at a distance from th e origin corresponding

to its water content at

F .

Similarly, solid magnesium

sulfate has the composition

represented by the point

FIG.I11

G. When an unsaturated solution containing these

salts is evaporated, its composition will, as in the

previous case, move along a line away from the origin

unti l one of t he curves representing the compositi on

of t he sa turat ed solution

is

reached, when the solution

will change in composition along this line in the direc-

tion aw ay fro m the line representing the composition

of the solid which is separating. Th us a solution having

the composition represented by

a

would, on evapora-

tion, change in composition along the line

ab

when,

on furth er evaporation, sodium sulfate would separate,

and finally, at B both sodium sulfate and the double

salt would separate, the solution remaining consta nt

in composition until it had all disappeared. Similarly

an unsat urat ed solution of composition represented b y

G would change in composition in the direction

c d B

the solids separating being first pure astrakanite and

the n a mixture of astr akan ite and sodium sulfate.

The point B represents, therefore, the end-point of

Crystallization for soluti ons which contain more sodium

sulfate tha n magnesium sulfate.

Fig. V represe nts the sol ubili ty of mixtures of

magnesium chloride an d potassium chloride, from which


3/10

98

T H E J O U R N A L

O F

I N D U S T R I A L A N D E N G I NE E R IN G C H E M I S T R Y

Vol.

I O , N

MO/S YgS

0 per

OOO mo/s H 0

FIG. V

it is possible to crystallize the double salt carnallite,

KMgCls.6H20. Unlike the previous instance, how-

ever, the line

OE

somewhere upon which lies the point

representing the composition of solid carnallite, does

not intersect the curve

BC

which expresses the compo-

sition of solutions sat ura ted with carnallite. This

fact makes the pa th of crystallization, during th e

evaporation

of

solutions of th ese two salts , somewhat

different from that considered above. A solution

having the composition represented by

a

will, on

evaporation, change in composition till

b

has been

reached, whereupon potassiu m chloride begins to

crystallize out, and the solution, becoming richer in

magnesium chloride, will move along bC. When the

solution has reached the composition represented by

C,

carnallite will begin t o sepa rate, bu t since carnallite

contains more potassium chloride than does the sat-

urated solution at C, it is evident that while carnallite

crystallizes, the solution will tend to move along the

line CB instead of remaining at C. The phase rule,

however, requires that while both potassium chloride

and carnallite are present, the solution must remain

constant in composition at C. Therefore, inste ad

of

the liquid phase disappearing a t this point, as was th e

case in th e former salt pair, i t is one of th e solid phases,

potass ium chloride, which will now disa ppear, being

changed over into carnallite.

It

is no t unti l all of th e

potassium chloride has been

so

changed that the solu-

tion can move from C t o B . B will thus represent an

end-point of crystallization, while C will not. It is

evident, therefore, t ha t in o rder t o prepare crystals of

carnallite it is necessary t o use a solution containing

more th an th e equivalent amo unt of magnesium chlo-

ride, the relative amounts of the two salts being such

that, on evaporation, the line BC will be intersected

slightly above

C.

Similar co nsiderations show

US

t ha t

on tre ating solid carnallite with water, instea d of dis-

solving as such, it would tend to change into solid

potassium chloride and a solution whose composition

is th at represented by C .

It

is obvious, therefore, that

I

/67

/

/

I

f l o h Nper

1000

mo/s

H2

fiG. v

it is not difficult to obtain potassium chloride fr

carnallite, a point

of

importance in the treatment

salt bitterns, as will be discussed later. After

remova l of t he potas sium chloride the solution can

evaporated, carnallite separating, while the compo

tion of th e solution changes frofn C t o

B.

This c

nallite can be treated with water, leaving solid pot

sium chloride, etc.

Solutions containing magnesium and potassi

chlorides and sulfates are in equilibrium with so

phases

at 2 5 O

according to the da ta in Table

I

and

TABLE

System, KCl-MgClaKnSOrMgSO4, a t

25

COMPOSITION

OF

SOLUTI

SOLID HASES

Mols of constituents

per 1000 mols HzO

RzCln MgClz

MgSO4

Kz

A

KCI

.................................

. . . . .

B MgCh.6Hz0

............................ 108

..

C M g S 0 ~ 7 H z 0

............................. 55

D

E

F

H

I a)

K

L

M

N

P

Q

R

..

5 5 ii:5

1 105

104

73

. . . . .

. . . .

42

..

25 21

9

55

8

62

4 5

70

2

99

..

..

..

14

15

58 5

22

..

11

16

15

13.5

12

(a)

The composition

of

the solution a t th is point is given by diff

figures in vant Hoffs earlier and late r books, The lat ter are dou b

incorrect,

as

the former agree with those

of

H.

S.

van Klooster,

J .

P

Chem.,

21

(1917), 513.

represented b y van t Hoff along four axes, as in F

V I , each pair of axes representing solutions contain

a common ion. The boundary lines correspond

solutions saturated with the two constituents rep

sented by the enclosing axes. Where salt pairs c

taining no common ion are present it is impossible

represent th e composition by a p oint in th e plan e

this figure.

A

mixture of equ iva lent quan titie s

potassium sulfate and magnesium chloride would e

dently lie a t t he origin 0 and would be indistingu

able from pure water by its position in th e plane.

order to make this distinction it is necessary t o in


4/10


5/10

IO0

T H E J O U R N A L O F I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y

Vol.

IO, N

of water remo ved by

s.

We may then represent what

has become of th e original solution during eva poratio n

b y means of the following equa tion :

IOOOHZO

8MgSO4 + IoMgClz + ZKzClz = ~ ( I O O O H ~ O

16MgS04

+

55MgC12

+

9KzClz) + ~o&Mg(S04)~.6HzO)

~(MgS04.7HzO)

+

SHZO

By equat ing coefficients of the various substan ces

present, i t is possible to set u p the following equatio ns:

Coefficients of H10: 1000 = xooop + 6p

+

7 r +

s

Coefficients

of

Mg:

IS =

71p +

p

+ r

Coefficients of Clz: 12 = 6 4 p

Coefficients

of

Kz z =

99

+ s

The solution of these equations gives the following

values

:

p = 0.188;

q =

0.31; Y = 4.38; s

=

779

These values of

q Y

and s represent the amounts

of the respective substances which have sepa rate d

by evaporating the original solution and p represents

the amount of solution left.

If, instead of tak ing the

TABLE 2

System, NaCl-KCl-MgClrMgSO4-NazSO4, t 25'

MOLS

PER

1000 MOLSHzO

NazCls KzCh MgClz-MgSO4 NazSOc

aturation with NaCl an d

A

B

C

D

E

F

G

H

I

L

M

N

P

Q

R

S

T

U

V

W

X

Y

z

MgClz.6HnO. ....................

KC1. ...........................

NazSOd

.........................

MgClz 6HzO Ca rna llite ..

.........

KCI.

darnaliite..

................

KC1: Glaserite..

NazSOc Glaserite..

. .

.........

NazSO4' Astrakani te ..............

MgSOa.'7HaO Astrakanite..

.......

MgS04.7HeO: MgS0~6Hz0.......

MgS04.6Hz0, Kieserite. ..........

Kieserite, MgClz.6HzO. . . . . . . . . . . .

KC1, Glaserite, Schiinite. . . . . . . . . .

KCl, Schonite, Leonite. . . . . . . . . . .

KC1, Leonite, Kainite. ...........

KCl, Kainite, Carnallite.. ........

Carnallite, Kaipite, Kieserite. .....

NaB04. Glaserite. Astrakanite..

...

Glaserite Astrakanite Schonite.

Leonite hstrakanite bchonite..

...

Leon ite' Astrakanit: MgSOc 7Hz0

Leonite' Kainite MgbOd 7Hzb..

...

MgSOd5HzO Kdinite MgSO4 7Hz0

MgS04.6HzO' Kainit; Ries erke.. .

Carnallite, MgCh.BH;O, Kieserit e.

No.

FIELD FORMUGA

1 ALZD

2. BFMNPQE

3 CGSH

4 DZRQE

5 FMTSG

6 SHIVUT

7 VIJXW

8 JXYK

9 KYRZL

10 TUNM

11 NUVWP

12 PWXYRQ

55.5

...

5 1

2

5.5

44 20

44 10.5

46 ...

26 ...

2.5

44.5 15:;

1

:i

4 . . .

2.5 . . .

1 ...

23 14

19.5 14.5

9.5 9.5

2.5 6

0 . 5 1

26 8

27.5 10.5

22 10.5

10.5 7.5

9

7.5

3.5 4

1.5 2

0 0.5

. . .

103

...

ii1i:5

70.5

...

. . .

...

7

67.5

79

102

21.5

25.5

47

68

85.5

16

16.5

23

42

45

65.5

77

100 ,

. . .

. . .

. . .

...

. . .

...

...

1 i : i

34

12

9.5

5

14

14.5

14.5

5

8

1i :5

19

19

19.5

13

10

5

...

12.5

...

:5

14.5

3

...

...

. . .

...

...

...

...

22

...

...

...

...

;.4Ha0

MINERALOGICAL

DESIGNATION

Bischofite

Sylvite

Thenardite

Carnallite

Glaserite

Astrakanite

Epsom salts

Not found

Kieserite

Schonite

Leonite

1

Kainite

am oun t of t he original solution represented 5 y IOOO

mols of water, a different am oun t is take n, propor-

tionate am ounts of th e solids and water are obtained

from the solution on evaporation to the same point.

When we come t o consider the ev aporati on of sea water,

we have in addition to the above components large

amounts of sodium salts. Since during evaporat ion

sodium chloride is always present, it is possible to

represent saturated solutions such as are obtained on

evaporating sea water by solid models similar t o t he

one considered above. By introdu cing sodium chlo-

ride as anothe r comp onent no new degrees of fre edom

are introduced, provided it is stipulated that solid

sodium chloride shall always be present. Van't Hoff

and his co-workers have determined the solubility

relationships at z 5 O and 83 . Fig. VI11 represents th e

results for 2 5

O

contained in Table z ; results for

are found in Table 3 and Fig. IX . The amoun

sodium chloride present is not consi,dered in the p

jection? but is counted in th e to tal number of dissol

mols which would be represented in a solid mod

Sodium sulfate may be expressed in term s of t he o t

salts present, since NazS04

=

Na2Clz

+

MgSOa

MgC12, or, = NazClz

+

KzS04- KzC12. T hu s po

C, Table 2 is represented in Fig. VI11 by coun

I Z / ~ divisions to the left of the origin and 12l / Z

visions along th e axis. Its position in a sp

model would be 63l/2 divisions vertic ally above

point so obtained.

Such a model may be construc

in a way similar to t ha t previously described.

Th e composition of se a water which has been ev

orated until it is saturated with sodium chloride i

follows, expressed in mols of each constituent:

TABLE3

Syst em, NaCl-KC1-MgClrMgSOcNazSO4, a t 83O

MOLS ER

1000 MOLSHzO

Sat ura tion with NaCl and NazClr KsCla MgClz MgSO4 Na

0

A MgClz.6Hz0

....................

B

KC1

...........................

C NarSO4 ........................

D MgClz.6Hz0, Carnallite. . . . . . . . . .

E

KCl Carnallite

. . . . . . . . . . . . . . . . .

F

KCl: Glaserite.. . . . . . . . . . . . . . . . .

G NazSOa, Glaserite

. . . . . . . . . . . . . . .

H NanSO4, Vanthoff ite

. . . . . . . . . . . .

I Vanthoffite Loeweite

K Loeweite Kieserite . . . . . . . . . . . . . .

L Kieserite: MgCla.6HzO., .........

P KCl Glaserite Langbeinite..

....

Q KCl: Carnallit;, Kieserite. ....... 2

R KCl Langbeinite Kieserite.. . . . . 11

S Glaieerite, NaaSOs: Vanthoffite..

.

43

V Loeweite, Glaserite Vanthoffite.. 34. 5

W

Loeweite, Glaserite,'Langbeinite.

,

30

Y Loeweite, Kieserite, Lang beinitq ., 16

Z Carnallite , MgCla.6Hz0, Kieserite

59

1

39

56.5

1

1.5

39.5

43.5

51

35

12.5

1

29.5

1

37

. .

2

10

39

21

..

..

3;:s

12

15

22.5

26.5

24.5

10.5

2

. . .

121

...

117

92

. . .

. . .

4.5

22

61.5

120

13

86.5

76

. . .

8.5

12

42

116

. . .

. . .

. .

8

. .

..

4

11

1615 .

12.5

5.5 .

1 .

10 .

5

.

5

17.5

.

16.5

14

.

7.5 5

1

MINERALOGICA

DESIGNATIONo. FIELD FoRmuLA

ALZD

BFPRQE

CGSH

DZQE

FPWVSG

NSVI

XVWYK

RYRQZL

WPRY

Bishcofite

Sylvite

Thenarditearnallite

Glaserite

Vanthoffite

Loeweite

Langbeiniteieserite

1oooHz0, 47NazC12, 1.03KZCl2, 7.36MgC12, 3.57MgS

By the use of a solid model i t is possible t o deter m

th at this solution, on further evaporation, would int

sect the surface at the point

a

where Epsom s

would begin to separate. Th e composition of

solution at this point

is

approximately 1000H

I INazC12, 6K2Cl2, zoM gS0 4, 41 M g CL Fur th er ev

oration would lead to the boundary between t

field and the kainit e field W X after which these t

salts would separate together.

It

is possible t o cal

late as before the am ount s of each substance remov

from the solution when the latter has the composit

indicated, say, by X . Suppose that

10 000

g. of

original solution are used. Th e number of grams cor

sponding to th e n umber of mols of each substance

the original solution is 24,790, so t h a t 1 0 0 0 0 g. of so

tion would contain, instea d of th e previous num ber

mols of ea ch const itu ent , only 0.404 of these qu an

ties, namely , 404Hz0, IgNa2 Cl2, 0.4zK2Cl2, 3Mg

and 1.44MgS04. On evapora tion this solution wo

yield mols of th e solution satu rate d a t X ontain

its constituents in the proportions indicated in Table

together with qMgS04.7Hz0 + rNazClp + sKC


6/10

MgS01.3H20 + tH2O. We can, therefore, write the

following equation:

404Hz0

+

rgNazClz

+

0.42KaClz 4 3MgClz + I . LFM~SO~

p(1oooH~0+ 3.5 Na ~Cl ~. 4K~C12

+

65.5MgC12

+

13MgS04)

+ pMgSOa.7Hz0

+

rNazClz+ sKCl.MgS04.3HzO + tHz0

By equating corresponding coefficients and solving

th e resulting equation s, we o btain t he following values:

p =

0.0458;

q =

0.37;

r =

18.9;

s

= 0.47; t

=

354

tions it will be possible to determine what will take

place.

I t is evident from the position of point

a

in the

diagram for

2 5 O

t ha t only a small amount of Epso m

salts will have been crystallized by evaporation

of

the

mother liquor from sea water before kainite will begin

to separate.

I t

is true that kainite shows

a

great

tendency to supersaturation, and unless suitable nuclei

ISGHOFITE

L

Feb.,

1918

T H E J O U R N A L O F I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y 101

and by the aid of geometric and algebraic considera- separation of th e potassium salts from the magnesium

Keso

FIG.

VI11

Hence we conclude that 354 mols of water have been

evap orated, and 0.37 mol of Epso m salt s, 18.9 mols

of NazClz an d 0.47 mol of kainite are in the solid

portion. Similar calculations may be made to de-

term ine what will happ en during all sor ts of changes.

For example, instead of removing water, a cert ain salt

may bt: added to a solution saturated with other salts,

F I G .

I x

are present this field might not be present, which will

allow th e evaporati on and separati on of Epsom salt s

to continue somewhat further until the potassium

chloride field is reached. The condit ions obta inin g

in solar evapo ration are, however, very favorable to

the crystallization

of

such

a

subst ance because of t he

presence of many impuritie s. I n order to get a good


7/10

I 0 2 T H E J O U R NA L O F I N D U S T R I A L

salts in the bitter n of sea water it is not desirable to

carry on the evaporation so as to separate more than

a small amou nt of Epso m salts a t 2.5 . O n comparing

Figs. VI11 and I X it will be seen tha t a t the higher

temperatures it is possible to continue the evaporation

much farthe r before a ny salt containing potassium will

crystallize from the hot solution.

A t

this tempera-

tu re and in the presence of th e magnesium chloride

which exerts a dehydrat ing effect, inst ead of E pso m

salts cryst allizing, kieserite, MgS04.HZ0, is obtain ed.

The fact th at t he solubility of magnesium sulfate tends

to decrease at higher temperatures, while the solu-

bilities of po tassium chloride and magnesium chlor ide,

and hence, carnallite, increase, causes the kieserite

field at 83' to become large at the expense of t he

fields of potassiu m chloride a nd carnallite.

I t

is evi-

den t, therefore, th at most of t he sulfate present in the

solution could be removed as kieserite by evaporating

the bittern at higher temperatures until the carnallite

boundary is approached. During this evaporation,

the solids which separate would be sodium chloride and

kieserite. By removing these from the hot solution

they could be obtained relatively uncontaminated

with potassium.

If,

now, the m other liquor from these

crystals is cooled, the growth of th e carnallite field

as lower temperatures are reached indicates that this

salt would separate as the solution cools, while the

mother liquor from th e carnallite would consist largely

of a solu tion of magnesium chloride. These considera-

tions seem to indicate th e possibility of a satisfa ctory

process for the sep aration of th e bitter n int o three main

constituents: magnesium sulfate, carnallite, and a

solution of magnesium chloride. The re would remain

the necessity, first, of separating magnesium sulfate

from the sodium chloride accompanying it, second,

of treati ng th e carnallite for the recovery of potassium

chloride, according to the principles discussed earlier,

and , third , th e evaporatio n and cooling of the magne-

A N D E N G I N E ER I N G C H E M I S T R Y

Vol.

IO, N

increase is caused by encountering the boundary

the carnallite field shown in Fig. IX , and th e sub

quent separation

of

carnallite.

The crystals which

deposited from the solution after this point is reac

contain a considerable am oun t of potassium in

for m of carnalli te. A calcula tion of t he amo unt

water which should be removed in order to reach

carnallite boundary a t

83 O

gave a figure correspond

very closely with that indicated by the above cur

The original mother liquor at

2 o

s saturated b

with sodium chloride and magnesium sulfate, but si

the solubility of sod ium chloride does not mate ria

change with the temperat ure, whereas th at of mag

sium sulfate does increase during th e first par t of

evaporation, the solution is satura ted with sodi

chloride but not with magnesium sulfate, hence

first crystals t o separa te consist largely of so di

chloride, which was found to be the case with the

bo th of th e microscope an d of a chemical analy

It

is possible, therefore, t o remove a n additiona l amo

of sodium chloride from th e magnesium sulf ate

filtering the hot solution by the aid of the centrif

during th e early stages of the evapo ration. T

procedure simplifies the fu rther purifica tion of

*IJ6

/L 1.3.

I I

I

I

11

IO

2 0

so

IO SO

FIQ.X-WATER EVAPORATED

N

PER

CENT

OF

BITTERN

AKE

0

jza

sium chloride liquor to obtain MgCl2.6H20.

, magnesium sulfate which separates as evaporat

Table 4 gives the results of th e analy sis

roceeds.

E V A P O R A T I O N E X P E R I M E N T S

The process outlined above, on the basis of t he equ i-

librium diagrams, was first tested on a laboratory

scale by evaporatin g weighed qua ntiti es of bi ttern .

I n one set of experiments the evapo ration was carried

on at the boiling point of the solution . Crops of crys-

tals were removed from the solution from time to time

by centrifuging the liquid through a muslin bag. The

density of th e solution was read b y the aid of a hydrom-

ete r made of pyrex glass, th e smal l coefficient of ex-

pansion of which made its readings nearly co rrect in

spite of changes of temp erature.

Th e. boiling point

was read with a thermometer graduated to one-tenth

of a degree.

The a moun t of water, when each reading

of densi ty and boiling point was made, was determined

by weighing the vessel containing th e hot solution.

It will be seen from the results, plotted in Fig. X,

that the density and boiling point rise gradually until

water has been removed amounting to about 36 per

cen t of the weight of the original bit tern. The dens ity

and boiling point from here on increase more rapidly

with the furthe r removal of water. This more rapid

;he crystals remo vedf rom t he solution by th e aid of

centrifuge at the stages of evapora tion indicated

Fig.

X.

TABLE

Crop from

cooled

mother Crop

liquor from

fina

Crop 1 Crop 2 Crop 3 Crop

4

from 4 evaporatio

MgSO4.... 15.1 24.6 20.1 24.1 1.3 3.5

MgCla.. , . . 12.0 12.3 21.8 16.9 39.2 38.2

KC1. . . .

.

4.8

NaCl

...

. 20.3 20.3 13.3 17.0 1.3

HzO

. . . . . .

52.8 36.4

40 5

24.9 54.7 5 7 : 5

6.3

4.4

17.1 3.6

1 1

The centrifuge employed was not very efficie

and some cooling took place during the process,

that the respective crops are contaminated with

salts that should remain in the mother liquor.

is eviden t from t he results of this analysis, togeth

with th e course of the density a nd boiling-point curv

that the process contemplated furnishes the desir

separation. I n a second experiment, based upon t

results of t he first, the solution was evaporated un

the density ha d reached a value of 1.35 at the boili

poin t of th e solution,

121 .

The crystals separati

up to this point were removed and the mother liqu


8/10

Feb.,

1918

T E E J O UR N AL O F I N D U S T R I A L

allowed to cool. Th e crystals separating on cooling

should be carnallite, and it will be seen from the anal-

ysis of these cry stal s in Tab le 5 that their composition

approx imates closely t o th at of carnallite. The mother

liquor from th e carnallit e consists principally of a

solution of magnesium ch loride, as is confirmed by its

analysis. The potassium content of the first two frac-

tions may be attrib uted t o the cooling in the centrifuge

inevitable

3n

working on such a small scale.

A s

is to

be expected, the proportion of Mg S04 .H2 0 to NaC1

is greater in the second crop of crystals th an in th e first.

TABLECOMPOSTTIO TI ON OF

MATERIAL BTAINED T VARIOUS

STAGES

N

PER CENT

Removed :t Removed Theoretical Final

b. p. 121 , from for mother

B. p. 116'

d.

1.35 cooled liquor carnallite liquor

K . . . . . . . . . . .

3 . 1

5 . 7

10.0 14 .1 0 .4

C1. . . . . . . . . . . 3.5 19 .1 36 .6 38 .4 23 .4

SO4

. . . . . . . . . .

16.9

32.2 trace 0 . 0 2 . 7

Mg . . . . . . . . . .

3 . 9 9 . 2

7 .6

8 . 7 7 . 9

The curve given in Fig. XI was obtained by an

evaporation in which no crystals were removed, thus

avoiding the inevitable losses occurring through at-

tempts to remove crystals from the hot solution.

The break in the boiling-point curve in Fig. XI is at

a higher temperature than th at in Fig. X. This is

doub tless due to the use of different samp les of bi tte rn

in the two experiments, so that the carnallite field is

encountered a t different points in the two cases. I t

may be noticed that the break is more pronounced in

th e case where it occurs at the lower tempe ratur e which

is jus t what would be expected on th e basis of t he solu-

bility diagram in Fig. IX . Th e composition of the two

samples of bitte rn used in th e above experiments is

given in Table 6. The sodium content is not given.

TABLG

Bittern used in gettin curves

in Fig. X in Jig. XI

K

.......................

1.48 1.76

C1

.......................

15.82 18.22

so4 ...................... 5.81 3 .88

Mg ...................... 5 . 3 8 6 . 3 2

TABLG

Bittern used in gettin curves

in Fig.

X

in Jig. XI

K

.......................

1.48 1.76

C1

.......................

15.82 18.22

so4 ...................... 5.81 3 .88

Mg ...................... 5 . 3 8 6 . 3 2

O U T L I N E

O F P R O P O S E D P R O C E S S

I . E V A P O R A T I O N O F THE

BITTERN-The .bitterns

from various sources will vary somewhat depending

on th e temperature of the liquid in the last salt pond,

and whether or not any Epsom salts are allowed to

separate. There is, in fact, no reason apparent why

a crop of Epsom sa lts should not be rem oved b y cooling,

either artificially or by storage till winter, before the

subseq uent process of se paration is applied. Th e

process of solar evaporatio n should n ot, ho wever, be

carried far enough to cause any potassium salts to

crystallize, as it is probably not desirable to separate

the potassium content into two portions. Th e varia-

tions in th e composition of th e bitte rn caused by an y

of

the above factors would not cause any serious diffi-

culty, as during the later evaporation the separation

of NaCl and MgS04.H20,kieserite, would take place

in such proportion as to make the resulting liquid

converge towards a fairly uniform composition.

I t

is more important, under present conditions, to

recover all

of

the potassium salts, and hence to pre-

vent their contaminating the NaCl and kieserite frac-

tion, th an i t is to recover all of the Epsom salts, or

t o

A N D E N G I N E ER I N G C H E M I S T R Y I 0 3

obtain pure magnesium chloride from the final liquor.

Such contamina tion would result, if t he evap oration

were continued as far as the carnallite boundar y,

for some cooling during the separatio n of th e kieserite

from the mother liquor is inevitable, and if t he solu-

tion is saturated with carnallite before this separation

begins, some of it will crystallize along with the kieser-

ite. On th e other han d, if the evapo ration is not con-

tinu ed so far, a little of t he su lfate will remain in th e

solution, and will probably pass thro ugh t he succeeding

operations and come down with the magnesium chlo-

ride at the final stage of the process. Since very pur e

magnesium chloride will probably not be desired, the

presence of th is sulfate can do no har m.

Inste ad, therefore, of c ontinuin g the ev aporation

as far as the break in th e boiling-point curves, as in

Figs. X and XI , it will doubtless be better t o evapora te

till the boiling point is about 120 C. This will re-

sult in the recovery of p ractical ly all of th e carnallit e

and still allow leeway for variations in the bitterns

used.

FIG.

XI-WATER EVAPORATEDN PER CENT

OF

BITTERN TAKEN

Th e best typ e of evapora tor for this operation will

doubt less be of t he film type, where a given par t of

the liquid is not boiling for a very long time. Ther e

is a tendency for magnesium chloride to hydrolyze,

giving magnesium hydroxide and hydrochloric acid,

which escapes with the steam. If the liquid is evap -

orat ed in a ke ttle it is boiling for such a len gth of t ime

th at a considerable a mou nt of magnesium hydroxide

is formed. If, on the o ther ha nd, the liquid is allowed

to flow over a heated surface, the evaporation takin g

place ve ry quickly, there is little time for this hydrolysis

to take place. This liquid may then be kept in a

settling ta nk without f urth er loss of hydrochloric

acid, even near the boiling temperature, provided

actu al boiling does no t ta ke place.

I n principle, the n, the process indicated is a s follows:

Evaporate until the boiling point

of

the liquid

is

raised to about 120 C. and the density is approxi-

mately

1.35.

The liquid running off from the evap-

orator should be caught in a steam-jacketed tank


9/10

104

T H E

J O U R N A L

O F

I N D U S T R I A L A N D E N G I N E E R I N G

CHEMISTRY

Vol. IO, N

where i t is allowed t o se ttle. The clear liquor is al-

lowed to run

off

to a cooling tank, in which the car-

nallite will sepa rate . Th e sludge of NaCl and kieserite

is

run in to centrifuges, previously heated, where i t is

separated from the adhering mother liquor, which is

run into th e cooling tank mentioned above. The

separatio n of th e Na Cl from the material remaining

in the centrifuges and th e recovery of Epsom salt s will

be discussed later.

containing the carnallite may be cooled by the fresh

bittern going t o the evaporator, in order to utilize t he

heat conte nt of the latter . After it has been thor-

oughly cooled, the carnallite which has separated is

removed and freed from

its

mother liquor by centri-

fuging. Th e recovery of the potassium chloride from

this carnallite will be discussed later.

CHLORIDE-The mothe r liquor from th e carnall ite con-

tains a very little potassiu m, a little sulfate, a consider-

able amo unt of colloidal organic matte r, the brom ine

cont ent of th e sea water, an d a large amou nt of mag-

nesium chloride. Th e liquid must be evapor ated fur-

ther in order to recover MgC12.6Hz0. During this

evaporation, however, the temperature rises con-

siderably , unless vac uum evapo ration is employed,

charring the organic matter, and strongly darkening

the magnesium chloride which separates

on

cooling.

To destroy this organic matter, therefore, as well as

to recover the bromine, preliminary treatment with

chlorine is desirable. Th e details of this tre atm ent

are now the sub ject of investigation in this labora-

tory. We can only say at the present time th at there

seems to be good prospec t of success.

Th e disposal of th e large quanti ties of m agnes ium

chloride that would be obtained from these bitterns

presents a n economic problem. The possible outlets

seem to be as follows: magnesium oxychloride c ement,

magnesium oxide and hydrochloric acid, and metallic

magnesium. Th e use of magnesium oxychloride ce-

ments might be greatly increased by skilful advertising,

hydrochloric acid might be substituted for sulfuric

acid, for certain purposes, and t here seems to bq good

reason t o anticipat e a large production of magnesium

in the future.

2 .

TH E

R E C O V E R Y O F

THE

CARNALLITE-The l iquor

3 . THE

RECOVERY

O F B R O M I N E A N D MAGNESIUM

4.

THE SEPARATION O F SODIUM CHLORIDE AND

SULFATE-The separation O f the sodium

AGNESIUM

chloride an d the kieserite obtained in the first pa rt of

th e process is complicated b y the possibility of fo rming

astrakanit e, Na2Mg (S04). 4H20, at ordinary t em-

pera tures an d of loeweite, Na4Mg,(S04)4.sH2O1 or

vanthoffite, N S ~ ~ M ~ ( S O ~ ) ~ ,t higher temperatures.

In order to put the separation of the magnesium from

the sodium salts on an exact basis

it

is desirable to

have a knowledge of th e solubility rela tionships of th e

chlorides and sulfates of these two metals. I t has

been found possible, by using data given by vant

Hoff, Seidelll a nd Roozeboom,2 to construct th e equi-

librium diagram for all but two points which are un-

1Am Chem.

J .

27

1902), 52 ; see

also

Schreinemakers and Baat, 2.

2. hysik . Chem. 2 1888). 518.

phys ik . Chem.

67

1909), 5 3 3 .

important for the present purpose. The dat a

are given in Table

7

and are represented graphic

in Fig. XII , giving a diagram similar to that in

V I , where potassium chloride is considered instead

sodium chloride. Th e two undeter mined points h

been add ed more or less at rand om, for the sak

completing the fields, and are denoted by in terroga

marks on th e figure.

Now the material obtained from the first stage

our process contains MgS04 and NaCl in nearly equ

len t amoun ts, and hence, if dissolved in water, w

be represented by a point lying nearly vertically ab

the origin, at a distance increasing as the solutio

evaporated.

It

might, therefore, cut the surface

the solid model in the astrakanite face, which wo

prevent th e separation of the sodium from the m

nesium.

A

little magnesium chloride, however, wo

if added , raise the solution away from the astra ka

field, so that we would have only NaCl and MgS

7Hz 0 to deal with.

Th e solubilities of these two salts ar e affected

differently by t he tem perature th at we may anticip

their separatio n by first cooling, removing Epsom sa

then evaporating partly at higher temperatures,

moving sodium chloride, th en cooling again, etc. T

portion of the equilibrium diagram th at can be c

structed for 8 O from v an% Hoffs data shows th a

that temperature loeweite and vanthoffite intr

themselves between t he magnesium sulfate an d sodi

chloride fields, even when a considerable amount

magnesium chloride is added, so that i t may not

advisable t o evaporate the solution for the remo

of sodium chloride at too high a temperatu re. T

great tenden cy of these double salt s towards su p

saturation might allow the evaporation to proce

without their formation.


10/10

Feb., 1918

T H E J O U R NA L O F I N D U S T R I A L

TABLE

System,

NaCl-MgC1r-MgSO4-NatSO4,

at

2 5 O

Solid phases

NazClz...............................

NazS04.1OH~0..

.....................

MgSO4.7HzO..

........................

hIgClz.6HzO..........................

MgCla.6HzO; N azClz.. . . . . . . . . . . . . . . . . .

NazC1z. NazS04. ......................

NazSO;. NazS04.10HzO(*).. . . . . . . . . . . . .

NazSO4.1OHeO;NaaM (S04)2.4Hzo

.....

MgS04 7Hz0 NazMgfSOa)z.4Hz0.:.

MgSO4.6HzO MgSOa.Hz0.

. . . . . . . . . . . . .

MgS04.6HzO: MgCIa.6HzO.

. . . . . . . . . . . . . . . . . .

NazSO4; NazSO4.1OHzO; NaaMg(S04) ~.4H zO.

NazClz. MgS04 7Hz0. NazM g(S 04) ~.4 H~ 0...

NazClaf MgSO4.HzO; MgCh.6ZO.. .....

MgSO1:7Hz01 MgS04.6HzO.

. . . . . . . . . . . .

NazCb; NmSb4; NazMg(S04)~.4HnO.

. . . . . . .

46

NazClz MgS04:7HzO: MgSO4 6H z0. .

....

NazCIz M gS04.6H z0: MgSO4HzO.

......

26

. . . . 5 5 . 5

. . . . . .

. . . . . .

. . . .

....

i:

.... 51

.... 30

. . . . . .

. . . . . .

. . . . . .

. . . .

.... 4

2.5

. . . . 1

*) From the experimental work of Professor W

still in progress in this laboratory, this point may be

. . . . .

. . . . .

3 5 : 5

. . . . .

1i:5

. . . 58

. .

108

. . . .

103

. .

. . . . .

27

. . .

35

35

. . .

48 26

73

15

Not determined

104

14

. .

Nitdetermined

7 34

67.5 12

, .

71 9 5

102

5 . .

16.5 3

C. Blasdale, which is

considerably in error.

In order to have the desired data it is very important

th at the equilibria here involved should be determined

or temperatu res both lower an d higher th an

2 5 O .

Work on the solubilities at o o s now in progress in

th is laborato ry and will be published as soon as possi-

ble.

It

is planned , also, to include potassium salts

in this work, so that a diagram for o o similar to t ha t

in Fig. VI11 can be constructed, and which might sug-

gest a modification of th e first tr ea tmen t of th e bittern .

CARNALLITE-The recovery of po tass ium chloride

from the carnallite was discussed earlier in connec-

tion with Fig. V. We m ay ask whether h ot or cold

water should be used for this purpose.

A

reference to

th e tables shows th at th e proportion of p otassium

chloride to magnesium chloride in th e solution a t equi-

librium with potassium chloride and carnallite is much

less at 2 5 th an at 83. This makes it obvious th at

a much smaller prop ortion of potassium chloride goes

into solution at the lower temperature . The composi-

tio n of t he solution in equilibrium with potassium

chloride and carnallite at

2 5 O

is as follows:

1oooH2O

+

5.5K2C12

+

72.5MgC12. Fro m thi s i t is possible t o

calculate the a mou nt of water t o be used in extracting

the magnesium chloride from the carnallite at this

temperature.

If I

mol of carnal lite is used we can

write the following equation:

5

T H E

RECOVERY OF P O T A S S I U M CHLORIDE F R O M

XMgCla.6H20 xHzO = yKsCi2 ~ ( I o o o H ~ O 5.jKzClz

+

72 ~MgC12)

From this we find

=

7.8;

y = 0 .425 ; z =

0.0138.

That is , I mol, or

2 7 7 . 5

g. of carnalli te, require s 7 . 8

mols,

or 140.4 . of wate r, or, th e weight of wate r re-

quired is approximately half the weight

of

the car-

nallite. At th e lower temperat ures th at would natu r-

ally be used somewh at more water would be required,

but relatively less KC1 would be dissolved.

The liquor used in extracting the carnallite may

then be partly evaporated and cooled, whereupon

ano the r crop of carn allite crystals will be obta ined .

To obtain the maximu m amou nt of carnallite but no

magnesium chloride the solution should be evaporated

t o such an extent th at on cooling with separatio n of

carnallite its composition will correspond to point

B

in Fig. V, which is 1oooHz0

+

105MgCL

+

K2C12.

Th e amo unt of evapora tion necessary is calculated

from th e following equatio n:

XOOOH~O72.5MgC12 + 5.5KzC12

xHzO +

yKMgC13.6H20+

~ I O Q O H ~ OIogMgClz + K2C12)

A N D E N G IN E E RI N G C H E M I S T R Y

This gives x =

340;

= 9.8 ;

z

=

0.6;

hence 2 5 7 2 0 g.

of solution should lose 6120 g. of water, giving 2 7 2 0 g.

of carnallite on cooling; or,

I

ton of solution should

lose 0.238 ton of water , and deposit 0.107 ton of car-

nallite.

It

will probably not pay, here or in stage

2

of the

process, t o recover the small amo unt of potassium

chloride remaining in the mother liquor from the car-

nallite,

Condensed summary of th e above process:

Evaporate bittern till boiling point becomes about 120 , and density

hot)

Separate solid and liquid while hot (settling tank and centrifuge).

A .

Solid. NaCl and MgSO4.HzO. Dissolve out NaCl wit h cold

water (containing some MgClz?) dissolve residue

in

hot water

and cool with ice machine, getting MgS04.7HaO.

1.35.

B . Liquid. Cool.

leaving

I.

Solid Carnallite. Extract wi th minimum amount cold water,

1. Solid KCl.

2. Solution. Evaporate partly, cool.

a. Solid carnallite, add to

I.

b. Solution of MgClz, add to 11.

Evaporate, cool, recover solid MgClz.6HnO.

11. Solution, mainly MgClz. Bleach with Clz and remove Brn.

The above process is being tested in this laboratory

on a semi-commercial scale under th e d irection of

Professor Merle Randall, and will be described in a

later publication.

I t

may be mentioned, however,

th at an excellent separatio n of a ctua l bitte rn has been

obtained into one lot of mater ial consisting of kieserite

an d sodium chloride, another consisting of carnallite

of a high degree of pu rit y an d whiteness, a nd a mother

liquor consisting of magnesium chloride solution con-

taining bu t very sm all amou nts of su lfate an d of potas-

sium. For example, using

I

5 lbs. of b itter n, and evap o-

ratin g till the boiling poin t was

I I ~

he thre e fractions

of materia l obtai ned had t he following composition:

NaC1, Kieserite

Carnallite Fraction Mother

Fraction Found Theoretical

Liquor

K

. . . . . . . . . . . . . .

1.6

11

.o

14.1 Trace

c1

. . . . . . . . . . . . .

21.1 37.4 38.4 23.4

so

. . . . . . . . . . . . . 22.2 0 .4 0 . 0 1.25

Mg

. . . . . . . . . . . . .

10.0 8 .1 8 .7

...

The writer wishes, in conclusion, to express gre at

apprec iation for the cordial cooperation of t he Oliver

Salt Company, which has given information and has

furnished samples of material and bi ttern .

Generous credit should be given to Messrs.

A. H.

Foster, W .

D.

Coughlan, C.arl Iddings and

W.

D.

Ram age for much of th e experimen tal work herein

described, and to Mr. Iddings for drawing the illus-

trati ons . Professor W. C. Bray has given considerable

time to th e final criticism of th e man uscript an d the

checking of the figures necessitated by t he absence of

the auth or from Berkeley due t o his acceptance of

a

commission in th e army .

Since concluding the abo ve work there ha s appeare d in

C h e m i c a l A b s t r a c t s

Vol. 11

(1917))

2719 ,

a brief outline

of a process by T. Nishimura, J . C h e m . Ind. T o k y o Vol.

20 (1917), 587, for extracting potassium from bittern

which, apart from certain serious errors in translation,

seems to be fundamentally similar to that herein de-

scribed, and which we may welcome, therefore, as

addit ion al evidence of th e feasibility of working up

these bitterns instead of allowing the m to be largely

wasted, as a t present.

BERKELEY,AL.

The Extraction of Potash and Other Constitutents From Sea Water Nittern

Documents

Transcript of The Extraction of Potash and Other Constitutents From Sea Water Nittern