Studies on the Mechanism of Ionic Exchange in Colloidal Aluminum

STUDIES ON THE MECHANISM OF IONIC EXCHANGE I N COLLOIDAL ALUMINUM SILICATES*

Downloaded by GEORGETOWN UNIV on August 29, 2015 | http://pubs.acs.org Publication Date: January 1, 1931 | doi: 10.1021/j150338a011

BY HANS JENNY

Introduction I n 1819 the Italian chemist Gazzari46 made the interesting observation that clay decolorizes liquid manure and retains soluble substances which in turn can be released to growing plants. Gazzari’s work is perhaps the first study on exchange adsorption or ionic exchange which a t present holds such a prominent place in general colloid chemistry. However, it was really the classical work of Thomas Ways2 on “The power of soils to absorb manure” which, in 1850, stirred the agricultural chemists and divided them into two hostile camps in regard to the nature of these processes. So revolutionary were Way’s discoveries from the viewpoint of soil fertility that he himself came to the conclusion: “It cannot be regarded in any other light than as a direct impress of the wonderful hand of providence.” Way was the originator of the following instructive experiment: If a solution of KCI percolates through a column of soil, differential adsorption takes place. Only the positive part of the salt (K) is retained, while the negative (Cl) remains in the soil solution. I n the place of the adsorbed potassium another element, mostly Ca, appears in the liquid phase. This replacement is “wonderful” indeed, since the valuable element K, indispensable for plant growth, is retained by the soil and prevented from leaching at the cost of the common element Ca. For 80 years this experiment has remained the corner stone of all investigations on base exchange in soils. It has been confirmed again and again in various modifications. There is no general agreement, however, as to the mechanism of the phenomenon. Way was inclined to believe that the process was a chemical one, although he was aware of certain abnormal, non-chemical characteristics such as speed of reaction, influence of temperature, existence of lyotropic series and others. Liebig,34not knowing the law of mass action, vehemently opposed Way and advanced a physical explanation of adsorption. Since that time the disputation about the chemical or physical nature of base exchange in soils has not come to rest. The situation is quite similar to the one existing in the chemistry of proteins and in the explanation of adsorption in general colloid chemistry. I n studying the voluminous literature on ionic exchange one gains the impression that apparently the diversity of opinion is partly due to psychological prestige reasons, since the experimental material presented is in most cases not adequate to substantiate the conclusions drawn. Most of the workers picked up a random sample of soil and conducted experiments without

* Contribution from the Department of Soils, Missouri Agricultural Experiment Station, Columbia, Mo. Journal Series No. 334.


2218

HANS JENNY

knowing exactly what they were working with. Natural soils are too complicated and represent systems too variable to yield precise information about the mechanism of ionic exchange. Fortunately there are two promising ways to avoid this uncertainty of the nature of the adsorbent. One is the study of ionic exchange with artificial clays (zeolites, permutits); the other is the isolation of the colloidal fraction from soil and its subsequent purification by methods which do not affect the nature of the colloid. A systematic study with such artificial and purified colloids in which both the nature of the colloidal complex and the properties of the participating ions are considered should reveal a qualitative and quantitative insight into the mechanism of ionic exchange. Such was the aim of the investigation herewith reported.

Practzcal Szgnzjicance of Ionic Exchange. Aside frorn theoretical interest, ionic exchange is of far-reaching importance in soil studies and in many industrial problems. The fixation of fertilizers added to the soil is still the classical problem of applied ionic exchange. The ions adsorbed in soils can be easily utilized by plants; in the process of the liberation of nutrients by living organisms ionic exchange undoubtedly plays a very significant r81e. It is being realized more and more that there is a relation between ionic exchange and soil formation. The first step in the weathering of aluminum silicates is an ionic exchange in which the hydrogen ions of water and carbonic acid replace the cations on the surface of the crystal lattice of the minerals. The differential leaching of K, Na, Ca, Mg, from soils is easily understood on the basis of the behavior of cations in the exchange adsorption.*: GedroizZo has even gone so far as to propose a scheme of soil classification which is based on the nature of the adsorbed ions on the colloidal soil complex. The great problem of the reclamation of alkali soils is now more fully understood. The important investigations of Gedroiz,20 Sigm0nd,~7K e l l e ~ Burgess12 ,~~ and others have shown that ultimately it is a base exchange reaction on a very huge scale. Physical, as well as chemical, properties of soils are closely associated with exchange adsorption. Wiegner59 and his pupils have been able to show that coagulation, dispersion, viscosity, hydration and structure of colloidal clays are greatly affected by the nature of the adsorbed ions. These observations have been confirmed by Baver' and others. A direct consequence of base exchange studies with clays has been the invention of water softening36 by the zeolite or permutit process, in which Ca-ions of the water are exchanged against Na-ions of the aluminum silicate. I n sugar factories the principle of base exchange has become a vital factor in the development of refining methods, There are yet many possibilities for the application of base exchange in industry, such as in the tanning of leather, stabilization of emulsions and suspensions in general, manufacturing of milk products and others.

IONIC EXCHANGE I N COLLOIDAL ALUMINUM SILICATES

2219


The Nature of the Colloidal Complex. Colloidal clays found in natural soils exhibit crystalline properties. On the basis of Debye-Scherrer diagrams the presence of montmorillonite, beidellite, halloysite and bauxite in soils has been definitely established.22*28Fortunately, the recent work of Bragglo and Paulinga permits a deeper insight into the nature of these colloidal complexes. According to these investigators silicates are made up of ions which act as impenetrable spheres with characteristic diameters. On account of their abundance and large size the oxygen ions (r = 1.32A) play an important r6le in the strycture of aluminum silicates. The small but highly charged Si++++(r = o.3gA) and Al+++(r = 0.57A) ions are surrounded by four or six oxygen ions forming tetrahedrons or octahedrons which are considered the “building stones” of aluminum silicates. The remaining electric valence forces of the 0-- ions are saturated by cations (K+, Ca++) which are packed in the interstices of the oxygen polyhedrons. The crystal is held together by the electrostatic forces existing between cations and anions. The packing of the building stones can be close or open. St. Naray-Szabo*’ expresses the density of the packing quantitatively by the L‘volume”occupied by one oxygen ion. From the viewpoint of ionic exchange it is very significant to note that mainly such crystallized aluminum silicates as have an open packing are distinctly capable of base exchange and adsorption. I n the following selected list (Table I) the correlation between density of packing and ionic exchange capacity is indicated:

TABLE I Packing of Oxygen Ions and Intensity of Ionic Exchange Mineral

Corundum Quartz Muscovite Nosean Ultramarine

Volume per oxygen ion ka Ionic exchange 14 .o Very little or none

Formula

A1203 Si02

18.7

K(A12,Mg3)(A1SiaOlo)(OH)2 19.2 NaeAle(SiOl)6.zNa2S04 19

11

J?

19

23. I 23 . I

Little Distinct Pronounced Very pronounced

Pronounced ionic exchange is observed mainly in the natural and artificial zeolites which have such big interstices that large ions such as K+and Ba++, as well as HzO molecules can move freely in the interior of the crystals. Crystals with close packing exchange only a t their outer surfaces. Their ability to adsorb ions depends, therefore, on the size of the colloidal particles. Thus, Kelley28was able to increase the exchange capacity of bentonite clays by mere grinding in a ball mill. On the other hand, Wiegner66 demonstrated that in case of permutit gels (related to ultramarine in the above table) the degree of dispersion is without material effect upon ionic exchange. The importance of crystal structure in relation to ionic exchange is illustrated convincingly in Figs. I to 3 . A1203 and Si02 simply do not permit ions to enter the crystals on account of the architectural arrangement of the oxygen ions. I n the case of ultramarine the pattern of alternating aluminum and silicon tetrahedra

HANS JENNY

2220


furnishes large cavities consisting of a cubical space with an edge of 3 . 7 A and six crosswise arranged prolongations in which there is superfluous space for “wandering” constituents.23 On the basis of these modern pictures of aluminum silicates ionic exchange can be much better visualized than has been hitherto possible. The electronegatively charged framework of 0-- and possibly OH- ions acts as the attractive wall (inner layer of the Helmholtz double

FIG.I Structure of aluminum oxide(AllOj). Close packed system (Bragg’O).

FIG.3 Structure of ultramarine (Jaeger*$). The large circles represent oxygen ions. Notethe bigcavityin the interior of the cell containing exchangeable Na ions.

FIG.2 iirrangement of oxvgen ions in an open packed system (St. Naray-S~abo~’).

layer) for the adsorption of cations. The structural arrangement of the 0-ions remains undisturbed as long as the positive and negative electrical charges are balanced. The forces of attraction are mainly electrostatic in nature (except perhaps of €and I-are ) greater the closer the cations can approach the oxygen ions. I t is to be expected, therefore, that ions of different size and electric charge act differently both in their adsorption and release. The search for such relationships between ionic exchange and properties of the ions was one of the main objects of this investigation.

Presentation of Experimental Material * I n these investigations ionic exchange studies were made on both artificial and natural coIIoidal aIuminum silicates. Artificial aluminum silicates: Ordinary permutits furnished by J. D. Riedel in Berlin and Folin-permutit obtained from Arthur H. Thomas Company were converted into pure

* Credit for part of

the experimental work is due to E. R. Shade and W. H. Allison.


IONIC EXCHANGE IN COLLOIDAL ALUMINUM SILICATES

8 h

.. E s R m 2221

0 N

W W W W

? ? ? ?

n o a m

HANS JENNY

2222

.a x"

O

?

M

O

N

O

fr:': r: N H

u ) W N


rd

8 x

w violm'o

h b W b G .

h G . W

9'0 H

I

r; & AH W N

b

N

0

* v i W

h

h

0

. . . . .

C A H h h N L C h W

0 G.G.



0

0

H

N

N

N b

o o o o '

? 70 P m o N H

5

W

L 3

1

1 0 CI N

o'

2223

?? ? ? ?

- 0

o'o 0

d.N N H

CI

HANS JENNY


2224

Li-Na-K-"4-permutits by leaching with chlorides for several months as previously described.24 H-permutit was prepared by electrodialysis in a Bradfield three compartment ce1L6 Natural a l u m i n u m silicates: Adopting the method of Bradfield,4 soil collected from the B-horizon of the Putnam silt loam was churned and aft,er standing for several days was run through a supercentrifuge in order to isolate the colloidal particles. According to the data of Baver' these are of the dimensions of 100-150 mp (#). Electrodialysis in Bradfield cells converted the clays, which originally contained various exchangeable cations, into pure hydrogen systems. Pure basic clays were then obtained by adding the proper amounts of LiOH, NaOH, KOH, or ",OH to the H-clays. Corresponding samples of bentonite clays from Wyoming were prepared in the same way. The amount of base added was equal to the saturat>ioncapacity of the clays which was determined by tit,ration curves according to Bradfield' and by replacement of H+ by KCl. Both methods gave very similar results. The exchange experiments were conducted as follows: stock solutions s; clay sols ( 3 - 5 9 / ) were transferred into volumetric flasks (250, 500 cc.), and redistilled, carbon dioxide-free water was added nearly up to the graduation mark. These colloid-water systems were kept for a week in a constant t ~ m perature room in order to permit the colloidal particles to reach the hydrat'ion

TABLE VI11 System: NH4-bentonite monovalent cations. r.770 g. colloid containing 1.980 milliequivalents NH,; volume 500 cc., time of reaction IO days, room temperature; pH of sol = 8.56, y% = NH4 exchanged in percent,age of total NH4. a = milliequivalents salt added.

+

NH4-bentonite +LiCl pH of 801 Y% 8,

69.75 81.37

3.56 8.89

NH4-bentonite +NaC1 __ pH of Y% a sol

-

70.25

80.86

5.04 8.15 12.59 8 . 0 5

NH4-bentonite +KC1 pH of sol Y% a

78.34 84.90

5.02

12.55

8.14 8.20

TABLEIX System: H-permutit monoualent cations. 2 . 5 0 0 g. permutit containing 1.743 milliequivalents exchangeable H+; volume 1 2 5 cc., time of reaction one week, temperature 26'C. yyo = H+ exchanged in percentage of total exchangeable H. a = inilliequivalents of salt added. H-permutit + LiCl H-permutit 4- NaCl H-permutit + KC1

+

y%

1.13 1.41 2.54

3.11

a

DH of supernatant liquid

2.135 5 . 0 0 5,338 4.91 10,677 4.88 21.354 4 . 8 6

Y% 2.69 3.20 4.52

6.51

a

pH of supernatant liquid

4.75 4.70 I O . O ~ O4 . 6 4 20.140 4 . 5 2 2.014 5.035

Y%

a

7.92

2.008

11.01

5.020

15.55

10.040

20.88

20.080

pH of supernatant liquid

-

-



.^

64 h v ) v ) v ) v )

N

m m m m m

. . . . .

m h d h

2225

HANS JENNY

2226

-E0

$..


x

m O

O N

O o

N h

N

W

O i

0 N

9 ? t ?

N C Q m N W N LOW

.

.

.

.

W r r ) d O H

N

- L O

z w

0

0

r r ) h d

9 9 9 " N

L

O

E

0

N

0

worn

,

.

W

.

W

.

W L O 1 0 N C

l

H

H

h N

1 1 1 1


2227

m 0 d-0,

? ? O'm m r r ) w

N

d-w o m

???" N

.-*

N

N

N

Q

d

e

I

roo

0 0

- 0

0 0

N ? ? ?

e


I

o w a N ? h N ? 0 0

h w w w w

J

m o

0 0 0

N

d-

a% w

N

m m o

N " 9 ? ?

I

MI

2

m N

.a

N

B

E

0 N N o\W r')dvirorr)

o w

0 0 0 - 0

09?

? ? t ? 1

h w w w w

? - ? ? N 1

w \ o w w w

roo

0 0 0

d

mm w

N ? ? ? ? N

m o

n m

N

N 9'9" I o m m ION row w w n

N

O

$1

n

rom

N

0'3

N m m N

0 0

N

.am%

N

*row

- 0

b o

O

N

d N

n

h w m

N

row w w

ci

H

'9 ? ? ? \ o w w m

roo

0 0

N ? ? ?

w

m

1'9

h h h w

0

. . . . .

H

d

N

I

m o 0 0 0 N ? ? ? ?

N

N

N ? ? ?

. h . d . N. u.

d

N

m o

w ' 6 w w w

N

0

w w w m

? ? ? h N

h

N

"?1?

w h w w w rorow

""

'owww

0 m

d-

0

d-m?

I

N ? ? t ?

V 3 0 N N O

? ? ?

N

N

II

* m o oH

HANS JENNY


2228

equilibrium. This precaution is necessary since it had been found that by diluting concentrated clay sols the hydration of the colloidal particles may become so strong as to bring about apparent negative adsorption. After equilibrium had been reached, neutral salt solutions (0.1-1.0 normal) were added and the flasks filled with water up to the graduation mark. After a week’s standing a t constant temperature (within i I O C ) during which time the flasks were shaken daily, the colloidal particles were separated from the liquid by centrifuging or ultrafiltration. The amount of ions exchanged was either found by direct determination (in release of ions) or by difference (in adsorption of ions). NHa was determined by distillation (occasionally also Nessleration) ; K was analyzed by the chloroplatinic method; and H was determined by titration using cresolphthalein as an indicator, Since the supernat,ant liquid of the H-systems treat.ed with chlorides always cont,ained Al+++, titration of H+ had to be made in the presence of AICl,. Ext,ensive titration with mixtures of HCl, AlC13 and LiCl, NaC1, KCl showed that the endpoint is reached at pH 9.2 which is in good agreement with the data reported by Britton.” All experiments reported were entirely repeated a t least twice.

TABLE XVI System: Li, N a , K-bentonite HCl (old sols) 0.950 g. colloid cont,aining 0.8075 milliequivalents exchangeable bases, volume 2 5 0 cc., time of reaction one week, temperature 28OC. y% = bases exchanged in per cent, as determined by HCl adsorption. a = milliequivalents of HCl added.

+

Li-bentonite

Y% 43.85 59.83 70.58

+ HC1

Na-bentonite

Y% 42.65 60.40 6 9 . IO

a

0.404 0.808 I ,615

+ HC1 a

0.404 0.808 I ,615

K-bentonite

Y% 39.96 50.36 59.41

+ HCI a

0.404

0.808 I ,615

TABLE XVII Release of adsorbed LiJ N a , and K , ions against HCl (Young sols) Fifty cc. of electrodialyxed H-bentonite (1.11 g. colloid) and 30.82 cc. of H-Putnam clay (1.65 g. colloid) were each diluted with 180 cc. HzO. After 24 hours 0.936 milliequivalents of LiOH, NaOH, KOH, respectively, were added and the resulting salt clays were shaken daily for a week; 0.936 milliequivalents of HC1 were then added and all systems were brought to a volume of 2 5 0 cc. After 9 days the sols were centrifuged and the HCl remaining in the clear supernatant liquid was determined by titration. The following results were obtained: exchanged (symmetry value) Li-bentonite HC1 = 77.90% JJ JJ ,J Na-bentonite HC1 = 76.41% 11 J f Jf K-bentonite HC1 = 74.12% ,, It ” Li-Putnam clay HCl = 84.48% ,, 17 ” Na-Putnam clay HC1 = 8 6 . IO% 1, ,l ” K-Putnam clay HCl = 80.90%

+

+ +

+ + +

IONIC EXCHANQE IN COLLOIDAL ALUMINUM SILICATES

2229

TABLEXVIII Release of adsorbed Li, Na, K from permutits by HC1 I

g. permutit in IOO cc. volume; time of reaction one week; room temperature Millie uivalents of exclangeable Milliequivalents bases in I g. HCl added permutit


Nature of permutit

Li-permutit I Na-permutit I K-permutit I Na-permutit I1 K-permutit I1

K)

x)

JO

70

Percent exchanged (Symmetry value)

1.78

1.78

82 .os%

2.34

2.34

77.40%

I .64

I .64

73.20%

2.72

2.72

2.72

2.72

89.98% 8s .45%

90

110

uo

WJ

M.€.Sail in equilih

FIQ.4 Adsorption of cations by H-Putnam clay. ( I gr clay plus increasing amounts of NaC1, KCl.)

Mathematical Formulation of the Process of Ionic Exchange The magnitude of ionic exchange is among other things a function of the amount of the salt added. Most investigators have been working with moderate concentrations only, omitting very dilute or highly concentrated systems. The neglect of including wide concentration ranges has been in many instances the direct cause of fruitless disputations as to the chemical or physical nature of the exchange process. There is no question that in all systems studied ionic exchange reaches a distinct maximum at which further additions of neutral salts remain ineffective in releasing adsorbed ions (See Fig. 4). The maximum (flat branch of the curve), however, does not necessarily correspond t o the total exchange capacity of the colloid. Removal of released ions and repeated additions of electrolyte will in many cases liberate additional amounts of adsorbed ions until the absolute exchange maximum is reached. For permutits the maximum exchange capacity lies between 400-500 milliequivalents of cations per IOO g. colloid, depending upon the preparation of the material. In many cases the cation-aluminum ratio in permutits is one. Natural clays contain considerable inert material and have much lower exchange capacities. The values found by the author are 85-100 milliequivalents of base for bentonites and 55-70 for Putnam clays as determined by

HANS JENNY


2230

continuous leaching with neutral salts, or potentiometric titration of H-systems with &(OH)* up to pH 7. I t is always necessary to state the methods used in determining the total ionic exchange capacity of a colloid, since not all methods give concordant results: There is considerable controversy among soil scientists as to the [[correct” method of measuring the exchange capacity. A considerable number of equations have been used to describe the functional relationships in ionic exchange reactions. Some of these are entirely empirical; others are supported by more or less convincing theoretical reasoning (Ganssen, Wolf, Freundlich, Rothmund and Kornfeld, Schmidt, Arrhenius, Wiegner, Vageler). The following three widely used formulae will be considered : Gunssen’s” On the basis of the law of mass action Ganssen derived the following formula:

K =

where K m n g x

X?

(m.n - x) (g - x) = equilibrium constant, = exchange complex (permutit) in g, = total amount of exchangeable bases in the exchange complex, = total amount of salt in solution, = amount adsorbed.

Wiegner’s e q ~ u t t o n s . ~Wiegner, ~J~ who approached base exchange from the colloid-chemical viewpoint, found that Freundlich’s parabolic isotherm was able to describe satisfactorily the exchange reactions: x/m = k . cIfP where x/m = adsorbed amount per g. of adsorbing substance, c = concentration of added salt after equilibrium has been reached, k,p = constants. Later WiegnerZ4modified the equation in order to express the fact that ionic exchange is independent of dilution:

where a = concentration of salt added a t the beginning of the reaction. Vageler’s Recently Vageler has suggested a new equation which expresses the observation that under higher concentrations the ionic exchange reaches a maximum: x.s y=x f C y = amount adsorbed per g. adsorbing substance, x = amount of salt in equivalents added per g. adsorbent, S = maximum exchange capacity (saturation capacity), C = half value, that is the conc. of x at which 50% of S is exchanged.

* H. W. Kerr’szg“new” theory and equation of base exchange are identical with the work of Ganasen, published twenty years ago.



2231

This hyperbolic equation is identical in form with the well-known adsorption isotherm of Langmuir. The latter corresponds in it way to the law of mass action of surface reactions (Pauli).40 In the present study all three types of equations have been used to describe the experimental data, but no equation has proved entirely satisfactory. It depends on how wide a range in concentration is investigated and what ions are included. I n dilute to moderate concentrations of ions with low atomic weights the Freundlich-Wiegner equation is very valuable, but a t higher concentrations, especially for larger ions, (including H) the formula fails. Here the Vageler-Langmuir equation proves to be more satisfactory. No equation is able to characterize, for instance, the finer differences existing between closely related ions, such as Na and Li in the case of exchange with H-clay. An equation which embraces the behavior of several cations a t various concentrations would have to contain more parameters than just one or two. This, of course, would render the physical interpretation of the constants difficult. In order to numerically characterize the individual behavior of various ions a single magnitude would be extremely desirable. For this purpose a single point on the exchange curve must be used, since no satisfactory formula having but one parameter exists. The numerical magnitude selected in this study is called “symmetry value.” It expresses the intensity of ionic exchange in percentage when the number of ions added to the system is made equal to the total number of exchangeable ions on the colloidal complex. For example, IOO grams of colloidal clay containing 60 milliequivalents of exchangeable NHa are mixed with a solution of NaCl having 60 milliequivalents of Na. Ionic exchange takes place and the Na ions replace 20 milliequivalents of NHa ions. The symmetry value is then 20/60. IOO = 33.33%. It is of interest to note that this symmetry value is in many cases approximately identical with the magnitude of the constant k in the Freundlich-Wiegner adsorption isotherm. It might be worth while in this connection to point out a principal difficulty in the formulation of equations for base exchange reactions. Investigators who approach ionic exchange reactions on colloidal clays from the classical chemical standpoint often express the concentration of exchangeable cations as equivalents per total volume of the sol, as for instance, “one milliequivalent of hydrogen ions in 1000 cc. clay-suspension” (pH 3). I n formulating equations for such “solutions” the exchangeable ions are treated as if they were uniformly distributed in the entire volume of the sol. But such is certainly not the case. The cations do not move at liberty in the solution. They are assembled on the surface of the suspended colloidal particles and are much closer together than the common concentration expression (pH 3 in the above case) indicates. I n other words, the exchange reactions on the micelle occur in concentrated rather than in dilute solutions even if only one single colloidal particle were suspended in a liter of water. Attempts to measure the order of magnitude of ion concentrations on permutit surfaces lead to values > I - 2 normal. Under such circumstances the laws of classical chem-

2232

HANS JENNY


istry which apply only to dilute solutions cannot be expected t o yield good quantitative results. The “constants” in the equations may not be constants. The formulae will have to be modified. Inter-ionic effects and polarization phenomena must be taken into consideration and individual ionic properties such as charge, radius and electric field strength will certainly play an important r81e. It will be shown in the following discussion that a mere application of the laws of dilute solutions can embrace but a small part of the multitude of reactions encountered in ionic exchange. The Avidity of the Colloidal Complex as a Factor in Ionic Exchange If natural clays and artificial aluminum silicates are subjected t o electrodialysis they take up exchangeable hydrogen ions. The H-systems are usually called H-clay, H-permutit, etc. The pH of such systems depends upon the concentration of the solid phase present as shown by many investigators, notably by Bradfield.6 Wiegnerj* and Pallmann3shave recently investigated in great detail this phenomenon and distinguish in such systems two types of hydrogen ions, “free ions,” which are in true solution (very small in number), and “swarm ions,” which are on the surface of the colloidal particles moving and sedimenting with them. From a theoretical standpoint it is of great interest to note that the adsorbed H-ions also affect the Hz-electrode as well as inverl sucrose. Upon addition of a neutral salt ionic exchange occurs, the cation of the sslt taking the place of the hydrogen ion on the colloid. The H-ions enter the intermicellar liquid and can be determined (by titration) after separation of the solid from the liquid phase by centrifuging or ultrafiltration. Using Zsigmondy’s notation, the reaction can be written as follows:

I K+ I K+ I H+

3HfC1-

+ (n-3) Kf C1-

The nature of the inner layer (negative sign) is not indicated in the above equation. It is assumed that it consists mainly of negatively charged 0-and OH- ions. Colloidal clays show great differences in the amount of hydrogen ions released. This is well illustrated in Fig. 5 , which shows the amount of H+ released from five grams of H-colloid upon addition of various amounts of KCl. The number of hydrogen ions freed depends, of course, upon the number of exchangeable H-ions adsorbed per unit colloid (capacity factor) and also u p o n the strength wzth whzch they are tied u p to the negative places on the collozdal mzcelle (intensity factor). The affinity between the positive outer (Hf) and negative inner layer (O--, OH-, in Helmholtz’s double layer) need not be constant for all colloidal aluminum silicates, since it obviously depends on the lattice arrangement of the negative ions in the colloidal complex. The ease of replacement of hydrogen ions which shall be called avzdzty ( T r ~ o g ) ~ * varies considerably in different soils as pointed out by P a ~ k e r . ~ Q


2233


z

2

4

6

8

IO

12 14

16

18 M E . KCI In equlbb

FIQ.5 Release of adsorbed hydrogen ions by KC1 from various colloidal aluminum silicates.

I n Table XIX the magnitude of avidity for three different aluminum silicates is expressed in various numerical terms. Irrespective of the formula

TABLEX I X Avidity of Various Aluminum Silicates in the Exchange of H+ against KC1 Numerical characterization

Constant k (Freundlich’s isotherm) Symmetry value Constant C (VagelerLangmuir equation) Apparent dissociation constant p K Increase in avidity

H-Permutit

H-Putnam clay

H-Bentonite

5.94

12.05

8.0%

I 5 .O%

15.22 4%

23.31

10.74

6.87

7.28

5.97

5.67

18

>

or mode of presentation used to characterize the intensity of the ionic exchange, it is clearly seen that all values increase from left to right. H-permutit has the smallest avidity and H-bentonite the greatest. According t o Vageler’s C-value 23.31 milliequivalents of KC1 must be added to replace 50 per cent of the adsorbed H+ from permutit, while only 6.87 milliequivalents are necessary to replace 50 per cent in the case oE H-bentonite. If one adds the same amounts of K+ ions as there are exchangeable H+ ions on the colloids, only 8 per cent of the H+ ions are replaced from the permutit, while 18.4 per cent are released from the bentonite (symmetry value). I n other words, the H-ions are held more tightly o n permutits than o n bentonites. P u t n a m clay occupies a n intermediate position. In the same table figures are given for the “apparent dissociation constant.” It has been calculated according to Bradfield,8 who accepts the

2234

HANS J E N S E

viewpoints of Loew, Ramann, Truog and others, and who considers clays as colloidal acids. H e characterizes the nature of the mineral soil acids with the apparent dissociation constants based on the following equation :


pH

=

pK

[salt] + log [acid]

Experimentally pK is determined by adding just enough base to the H-clay to neutralize one-half of the exchangeable H+ ions. The resulting pH value corresponds to pK. Saturally the dissociation const,ant and the avidity index must stand in close relation to each other since both are expressions for the affinity between the H-ion and the clay “anion.” It is evident from Table X I X that on the basis of pK, H-permutit is a weaker acid t’han either H-Putnam clay or H-bentonite.

TABLE XX Effect of Base and Time of Reaction on the Magnitude of the Apparent Dissociation Constant of Colloidal “Clay Acids” (Based on pH measurement,s of sols) “normality” = o.0028 N Base used for neutralization

LiOII

Time of reaction

4 days

46 days

NaOH

4 days

46 days

KOH

* H-Permutit

Permutit

x

*

10-7

0.141 0,355 0.224

Putnam clay

x

Io-:

Bentonite clay

x

Io-’

14.46

8.91 28.19

j.62

12.60

j .02

0.692

II.7j

30.91

4 days

0.52j

10.72

46 days

I ,203

30.20

21.38 70.80

was ground in a mortar until it formed a relatively stable suspension.

Cnfortunately the dissociation constant of colloidal clay acids is not of such great value as the dissociation constant of true weak acids. The trouble is that pK does not prove to be a constant; it varies enormously with the experimental methods used as seen in Table XX. From data of Baverl it can be calculated that the dissociation constant for a 0.016“normal” clay acid suspension varies 1900 per cent according to the base applied (LiOH, It is also certain that pK varies with dilution, and furthermore it has a very great time factor. For these reasons the term “avidity index” instead of “dissociation constant” is used in order to avoid an impression of accuracy and consistency that does not exist. The qualitative similarity between the avidity index and the dissociation constant leads to important suggestions regarding the release and adsorption of ions by colloidal aluminum silicates of various avidities, which demonstrates the heuristic value of the clay acid-viewpoint in ionic exchange.


223s

a) Release of H-ions. Under comparable conditions weak clay acids release less H+ than strong* clay acids. This statement follows directly from the empirical definition of avidity. It has been tested by adding different salts to the H-colloids. Table X X I shows the relationship expressed in terms of symmetry values.


TABLE XXI Release of H+ from Electrodialyzed Colloidal Aluminum Silicates (Symmetry values) Substance added

H-Permutit

6.78% -

KNOs KCl NaCl LiCl

% % 1.0 %

8.0 2.7

H-Putnam clay

H-Bentonite

7 .SI%

13.78% 19.92% 18.4 % 14.6 % 1 8 . 0 yo

11.30% 13.0 % 6.2 % 6.7 %

It is evident that for all of the five different cations and two different anions used, more H-ions are replaced from the bentonite (stronger acid) than from the permutit (weaker acid). This shows that the differences in affinity existing between the various aluminum silicates and H-ions are independent of the salts used in the exchange experiments in this investigation. On the basis of stoichiometrical replacement the corollary follows that a “strong”* clay acid H-system adsorbs more cations than a weak clay acid. This statement is of particular importance from the viewpoint of fertilizer applications to soils of humid regions which possess considerable amounts of exchangeable H-ions in their colloidal complexes. b) Adsorption of H-ions. Under comparable conditions weak clay acids adsorb more H-ions than “strong” clay acids. In other words, if HCl is added to basic clays, K-clays for instance, the aluminum silicate of lowest avidity will adsorb the most H-ions. I n Table X X I I the effect is shown quantitatively on the basis of symmetry values. TABLEX X I I Adsorption of H+ from HC1 by Various Aluminum Silicates (Symmetry values) (From Tables 111, IV, VII, XIII, XVI, XVIII) System

Percentage of H adsorbed

NH4-permutit NHrPutnam clay NH4-bentonite clay K-permutit K-Putnam clay K-bentonite clay

* “Strong”

in a relative meaning, as contrasted with “weak.”

83.53% 74.82% 60.90%

8s .45%

69. ss% 50.36%


2236

HANS JENNY

In every case permutit takes up the most H-ions and bentonite the least. If HCl is replaced by weak acids such as acetic acid, the effect still holds for Putnam clay and bentonite but is less conspicuous for permutits. If the ionic exchange is stoichiometric the corollary follows that adsorbed monovalent cations are more easily replaced (by Hf) f r o m a weak clay acid t h a n f r o m a strong one. From the clay acid standpoint the entire problem of ionic exchange corresponds to the phenomenon of distribution of a base between two acids. If both acids are weak and their dissociation constants known, ionic exchange might be predicted quantitatively, as has been attempted by Bradfield. Since in this study only neutral salts of strong acids are brought into contact with clays the application of the above principle is excluded. S u m m a r y of Discussion of Avidity of Soil Colloids. I j Various colloidal aluminum silicates have different affinities for hydrogen ions. Bentonite has a high avidity, permutit a low one. Putnam clay takes an intermediate position. 2 ) From the clay-acid viewpoint, the avidity index can be related to the apparent dissociation constant of the clay acids. Putnam clay is a weaker acid than bentonite, while permutit is weaker than Putnam clay.

3) The following relations existing between ionic exchange and strength of clay acids have been suggested:

a) The weaker the clay acid the less H-ions will be released upon addition of a neutral salt. b) The weaker the clay acid the less cations will be adsorbed by H-systems. c) The weaker the clay acid t'he more H-ions will be adsorbed by basic systems. dj The weaker the clay acid the easier are adsorbed cations replaced by H-ions. Lyotropic Series and Exchange-Adsorption The adsorption of monovalent cations by "4and H-colloids exhibits characteristic differences in the three types of aluminum silicates investigated, The various silicates will be discussed separately.

a) Permutit systems (artificial zeolites). Fig. 6 and Tables 11, 111and IX, show that the intensity of ionic exchange is different for all cations studied. The individual behavior of t'he various ions is regularly manifested throughout the range of concentration used and can be expressed by various numerical values as indicated in Table XXIII.



L_..__-..

I

. 10

,

20

"

'

40

30

'

'

50

"

Bo

2237

--- Li

'

70 M.E. Salt in quilib.

FIQ.6 Adsorption of cations by "4-permutit. (From Table 11)

TABLEXXIII Lyotropic Series as characterized by Various Numerical Values (NHd-permutit monovabnt chlorides, equilibrium concentrations)

+

Salt added

Symmetry value

Wiegner-Freundlich Vageler-Langmuir isotherm isotherm

(s = 4.15)

K

(I/P=O.57)

s

C

I/P

0,392 0.296 0.226 0.226 0.296

15.21 29.46 45.07 56.75 57.98

2.30 3.45 3.88 4.07 4.31

61.56 14.53 6.46 2.09 1.44

0.029 0.209 0,715 1.947 2.421

k

LiCl NaCl KC1 RbCl CsCl

16.5% 32.5% 48:0%

54.5% 56.5%

Ganssen equation

FIG.7 Adsorption of cations from chlorides by H-permutit. (From Table IX)

HANS JENNY


2238

The magnitudes of all constants in the various equations clearly express a systematic trend in regard to the cation used. Whether the formula is based on strictly chemical principles (law of mass action) or whether it designates physical adsorption, the lyotropic series is always manifested. The above type of comparison is valid for all systems studied. For the sake of simplicity and uniformity the symmetry value will be mainly used for the quantitative characterization of the exchange intensity of the various cations. The lyotropic series is independent of concentration and is noticeable in pure K’H4-systems, in pure H-systems (Fig 7 ) or in mixtures of NHa-H-systems, as illustrated in Table XXIT’.

TABLEXXIV Adsorption of hlonovalent Cations by Various Permutit Systems (Symmetry Values) Li

Na

K

Rb

NH4-permutit (Table 11)

%

%

%

%

%

16.j

32.5

48 o

54 5

56.5

% -

NH4-H-permutit (Table 111) pH range

16 o

33.5

44.0

-

-

83 6

-

-

System

H-permut it (Table IX) pH range

7

0-7.1 I

O

4 9-5 o

6 9-7.1 6 8-7 5

H

Cs

2

4-64

2 7

8.0

-

-

-

4.75

4.60

-

-

-

Although H-permutit adsorbs fewer cations than NH4-permutit, the lyotropic series is noticeable in every case, both under alkaline and acid conditions. For all permutit systems the lyotropic series assumes the following form in regard to adsorption: Li

< Na < K < Rb < Cs < H

Li is least adsorbed and H is best adsorbed. The differences are very great, amounting to 240 per cent between Li and Cs in NHpperrnutit and 700 per cent between Li and K in H-permutit. b) Putnam clay-systems. Tables IV, V, VI and X and Figs. 8, 9 and I O give information as to the existence of the lyotropic series in the natural soil colloid, Putnam clay. Again pure NH4-and H- and combined NH4-H-clay systems were studied. A glance a t the various graphs reveals at once that the lyotropic series is clearly manifested, but at the same time one may notice a significant difference in the case of Li and Na. Both curves are closely associated. I n dilute concentrations Li is slightly better adsorbed than Na, while a t higher concentration it is somewhat less attracted. Table XXV expresses the adsorption intensity on the basis of symmetry values.

IONIC EXCHANGE I N COLLOIDAL ALUMINUM SILICATES I


t

FIQ.8 Adsorption of cations b NH4-H-Putnam clay. (From T d l e IV)

looeo

I

K

.-

10 2

4

6

8

10

I2 I4

16 18 20 M E . Sol? in rquillb

FIQ.9 Adsorption of cations by NHrPutnam clay. (From Table V) 60 50

2

4

6

0

10

IC

14

I6 I8 20 ME.Sdlr in rquilib

FIG.I O Adsorption of cations by H-Putnam clay. (From Table X)

2239

HANS JENNY

2240

TABLE XXV


Adsorptinn of Monovalent Cations by Various Putnam Clay Colloids (Symmetry Values) System

Li

Na

K

"4-Putnam clay (Table V) pH of sol at equilibrium

31.5% 7.24

32 .a%

63 ' 7%

7.04

8 .oo

KH4-H-Putnam clay (Table IV) pH of sol at equilibrium

37.5%

35.7%

54.0%

5.26

5.23

5.05

NH1-H-surface soil (Table VI) pH a t equilibrium

33.2%

35.8%

55.0%

5.31

5.22

4.88

H-Putnam clay (Table X) 6.7% pH at equilibrium: sol 3.91 Intermicellar liquid 4.05

6.2%

13 .o% 3.72 3.67

H-Sharkey clay (Table XII)

10.4%

9 .o%

3.93 3.93

15.2%

For all five systems the symmetry values express quantitatively the adsorpt'ion similarity of Li and Na. In natural clays and soils the lyotropic series is also independent of the nature of t,he adsorbed ion, concentration of salt added and acidity or alkalinity of the ayst,em. In fact, every cation brings the sol to a pH value characteristic for the added ion. The order of the ions in the lyotropic series must be written as follows: Li=Ka Nepermutit >K-permutit, etc. Experiments previously cited show that this series actually exists. 2) Divalent systems: For exchange with divalent cations the ionic replacement in permutits was also found to be stoichiometrical. For every divalent cation adsorbed, two monovalent cations are liberated. From the viewpoint of space relations a highly interesting situation arises. Although the entire

* Crystal lattice radii.

HANS JENNY


2256

system remains electrically balanced, yet the place of one of the two cations remains unoccupied, since the divalent cation certainly is not split up into two parts. For instance, the volume of two Na-ions is 7.88 A3, while the volume of one Ca-ion, which replaces two Na-ions, is only 4.99 A3. The “empty” space can thus be filled with water molecules. This mechanism leads to the conclusion that permutits with divalent cations are more hydrated than permutits with monovalent cations, provided the ingoing and outgoing ions are of somewhat similar size, such as, for example, Li and Mg, Na and Ca, K and Ba. From analogy with monovalent syst’ems it, would be expected t,hat within t,he divalent lyotropic series, the hydration decreases from Mg-permutit to Ba-permutit, according to the size of the divalent ions. 3 ) T h e H-system: Extending the ideas so far developed to H-systems, the conclusion follows at’ once that these should contain the most water of hydration. The H-ion is the smallest ion of all, it occupies very little space in the cavity which, therefore, can be filled with HzO molecules nearly to the maximum of its wat,er holding capacity. For instance, if a Na-permutit is converted into a H-permutit, the “cavity-water” should be increased at about the rate of one water molecule per Na-ion replaced. Data indicate, however, that the number of water molecules in H-systems is considerably greater than that provided by mere space relationships. The author believes that the additional water molecules represent water of constitution formed during the process of replacement. Remembering that aluminum silicates consist of a framework of oxygen, and possibly hydroxyl ions, it can be easily ima,gined that replacement, of ca?ions by H-ions creates ”potential water molecules” which are 1iberate:l upon ignition of the new compound. Fig. 2 1 illustrates schematically this formatic.1.of crystal lattice water molecules as a consequence of replacement of cations by H-ions.

1

OH - H

OH- K I

crystal

1

I I

OH - H

bH- K crystal

OH- K

OH - H

OH- K

OH - H

I

l OH- and O= ions on the surface of the crystal, as a part of the latter.

+ 4 KCl

1 H20 - molecules belonging to the crystal.

FIG,2 1 . Formation of crystal lattice water by adsorption of HA.

Before the exchange the analyses of the water-free colloidal particle would yield xSiOz, yAlzOI and zKzO, while after exchange H2O would be found in place of KzO. The hydration of the H-system is thus considered to consist of at least two main parts. First, the ordinary water of hydration (water films around the



2257

colloidal particles and cavity water) and second, water of constitution. This water of constitution also explains why in many cases the H-systems exhibit a lower exchange capacity than the original zeolite complex, and furthermore it may account in part for the well known observation that ignition destroys the ionic exchange properties of the colloid. Summarizing, it can be said that size consideration in ionic exchange can plausibly explain the observed relations between water adsorption of permutit gels and the nature of adsorbed ions (even including the abnormal H-system). For natural soil colloids (clay, bentonite-sols) complications may occur on account of differences in crystal structures and variations of size of colloidal particles due to ionic exchange and its consequent changes in the electrokinetic potential. I n such systems the effect of the hydration of the ions in the outer surface and the polarization of water molecules due to the surface oxygen ions may overshadow to some extent the rBle played by the cavities.

General Summary Ionic exchange (replacement of cations) has been studied in various artificial and natural aluminum silicates such as permutit, zeolite, Putnam clay and bentonite. Electrodialyzed colloidal aluminum silicates contain exchangeable 2) hydrogen ions and behave in many respects as colloidal acids. The strength of these acidoids varies as follows: permutit < Putnam clay < bentonite Permutit is the weakest acidoid. 3 ) I n all exchange reactions very pronounced lyotropic series can be observed. The position of the various elements within the series varies according to the nature of the aluminum silicate. For the adsorption (intake) of monovalent cations the following relationships were found: Permutit: Li < Na < K < R b < Cs < H Putnam clay: Li = Na < K < H Bentonite clay: Na 4 Li 4 K < H The release or outgo of adsorbed cations takes place in the reverse order of the intake. 4) The explanation of the individual behavior of the cations in the exchange process is based on Coulomb’s law which expresses the magnitude of attraction existing between negative oxygen ions of the crystal lattice and adsorbed cations. Accepting the ideas of Wiegner, stress is laid on the radii of hydrated ions. It appears that these “hydrodynamic” radii vary according to the nature of the colloidal system. 5 ) The strong adsorption and difficult release of the hydrogen ion is easily understood by considering the strong chemical bonds existing between the adsorbed hydrogen ion and the oxygen and hydroxyl ions of the rigid crystal frame. 6 ) Ionic exchange affects the hydration of the colloidal particle. On the basis of space relationships it is suggested that clays with adsorbed divalent I)

2258

HANS JENNY

cations should contain more water than clays having monovalent cations. H-systems are necessarily most hydrated.

References L. D. Baver: Missouri Agr. Exp. Sta. Research Bull., 129 (1929). 2 M. Born: Z. Physik, 1, 221-249 (1920). 3 M. Born: “Atomtheorie des festen Zustandes” (1923). R . Bradfield: Missouri Agri. Exp. Sta. Research Bull., 60 (1923). 6R. Bradfield: J. Phys. Chem., 28, 170-175 (1924). E R. Bradfield: Ind. Eng. Chem., 20, 79 (1928). R. Bradfield: Proc. 1st Int. Congr. Soil Sc., 4, 858-868 (1928). 8 R. Bradfield: J. Phys. Chem., 35, 360-373 (1931). * W.L. Bragg and J. West: Proc. Roy. Soc., 114A,450-473 (1927). W. L. Bragg: Trans. Faraday Soc., 25,291-314 (1929). 11 H. T. S. Britton: “Hydrogen Ions” ( I 29). 12 P. S. Burgess and W. T. McGeorge: lrizona Agr. Exp. Sta. Tech. Bull., 15, 359-399 (1927). 13 P.Debye: “Polar Molecules” (1929). 14 C. Drucker: Z. Elektrochemie, 20, 79-81 (1914). P. Drude and W. Ernst: Z. physik. Chem., 15, 79-85 (1894). 18 K . Fajans: Naturwissenschaften, 9, 1921. 17 K . Fajans: 2. Elektrochemie, 34,502-518 (1928). K . Fajans: Science, 72,99-107 (1930). R. Ganssen (Gans): Centr. Mineral. Geol., 1913,699-712, 728-741. *a K . K. Gedroiz: Kolloidchem. Beihefte, 29, 149-260 (1929). 21V. M. Goldschmidt: Trans. Faraday Soc., 25, 320-345 (1929). 22 S. B. Hendricks and W. H. Fry: Soil Sei., 29, 457-479 (1930). 23F.>I. Jaeger: Trans. Faraday Soc., 25, 320-345 (1929). 24 H . Jenny: Kolloidchem. Beihefte, 23, 428-472 (1927). 26 H. Jenny: Missouri Agri. Exp. Sta. Research Bull., 162 (1931). 28 H.Kappen: “Bodenaziditat” (1929). 2’ W.P. Kelley: J. 4m. Soc. Agr., 22, 977-985 (1830). , W.P. Kelley, W. H. Dore, and S. M. Brown: oil Sei., 31,25-45 (1931). ** H. W. Kerr: J. Am. Soc. Agr., 20, 309-355 (1928). 30 F. Kohlrausch and W. Hallwachs: Z. physik. Chem., 12, 538-539 (1893). 31 W. Kossel: Ann. Physik, 49, 229-362 (1916). 32 I. Langmuir: J. Am. Chem. Soc., 38, 2 2 2 1 (1916). 3 3 J. E. Lennard-Jones and B. M. Dent: Trans. Faraday Soc., 24, 92-108 (1928). 34 J. Liebig: Ann., 94,3 3 384 1855). 35 A. R. Martin: Dept. &Ind!Res. Tech. Paper, 1, London (1929). 36 I. S . Mukherjee: Trans. Faraday Soc., 16, 103-115 (1921). 57 St. Naray-Szabo: Z. physik. Chem., 1930,356-377. 38 H. Pallmann: Kolloidchem. Beihefte, 30,334-405 (19 0) 39 F. W. Parker: Proc. Int. Congr. Soil Sc., 2, 164-173 t1928). 4 0 U‘. Pauli u. E. Valko: “Elektrochemie der Kolloide” (1929). 41L. Pauling: J. Am. Chem. Soc., 51, 1010-1026 (1929). (*L. Pauling: Science, 71, 545 (1930). 43 L. Pauling: J. Am. Chem. Soc., 53, 1367-1400 (1931). 44 J. A. Prins: Naturwissenschaften, 19, 435-442 (1931). 46 H. Remy: 2. physik. Chem., 89,467-488, 529-569 (191j ) . 46 F. Sestini: Landw. Versuchsst., 16,409, 411 (1873). 47 A. A. J. de Sigmond: Special publication, California Bgr. Exp. Sta. (1927); Soil Sc., 21, 4 5 475 (1926). 48 M. $;ne]: 2. PA. Diing. Bodenk., 9A, 121-135 (1927). 4 9 E. Truog: Ind. Eng. Chem., 8 341-3 J (1916). 60 P. Vageler and J. Woltersdod: 2. Pi: Dung. Bodenk., A 15, 329-342 (1930); A 16, 184-204 (1930). 51 E. W. Washburn: Jahrb. Radioakt. Elektronik, 5, 493-552 (1908); 6,69-125 (1909). 52 J. T. Way: J. Roy. A r. Soc England, 11, 313-379 (1850). 5 3 T . J. Webb: J. Am. Cfem. Soo., 48, 2589-2603 (1926). 54 E. T. Wherry, C. S.Ross, P. F. Kerr: ‘Colloid Symposium Annual, 1930, 191-193. 55G. Wiegner: J. Landw., 60, 1 1 1 - 1 5 0 , 197-222 (1912). 58 G. Wiegner: Kolloid-Z., 36 (Erganzungsband), 341-369 (1925). 5’ G. Wiegner and H. Jenny: Kolloid-Z., 42,268-272 (1927). ‘8 G. Wjegner: Kolloid-Z., 51, 9 60 1930). b9 G. Wiegner: J. Roc. Chem. &d., $0, 55-62, 65-71, 103-112 (1931). B o I. Zoch: Chem. E r d e , 1, 219-269 (1915).


1

’

Studies on the Mechanism of Ionic Exchange in Colloidal Aluminum

Recommend Documents