Role of Kaolin in Anion Sorption and Exchange1

DALE H. SiELiNG2

'-T~»HE process or mechanism by which soluble phos-1 phates are "fixed" in the soil and the factors in-

fluencing the process have been the subjects of manyinvestigations. It has been rather widely accepted thatthe components of the clay fraction and the iron andaluminum of the soil solution are chiefly responsiblefor "fixation" in acid soils. Ravikovitch (13, I4)3

concluded that phosphate was sorbed as a result ofequivalent ionic exchange with other anions of thesoil complex and could be released from the sorbedcondition by anionic exchange. Scarseth (15) pro--posed that phosphate ions replaced hydroxyl ionsfrom the exposed alumina groups of the alumina-silicate complex of electrodialized bentonite at pH6.0. Murphy (11) observed that ballmilled kaolin wasvery effective in sorbing phosphate from the solutionscontaining H2PO4- ions and that sorption increasedprogressively as the pH of the solution decreased toa maximum at pH 3.0 and at pH's below 3.0 decom-position of the mineral prevented further fixation. Bymeasuring the water formed by the interaction ofphosphates with ballmilled kaolin and halloysite andfrom X-ray data obtained from these phosphatedminerals, Stout (17) concluded that "fixation" con-stituted a simple ionic exchange of phosphate forhydroxyl ions in the crystal lattice of these minerals.He also concluded that the process of exchange wasreversible. Subsequent investigations by Black (i)and by Kelly and Midgley (5) largely confirmed thefindings of Stout; however, each of these investiga-tors emphasized the importance of free aluminum andiron oxides in the "fixation" process. Coleman (2)attributed the sorption of phosphate by both kaoliniticand montmorillonitic clays to the free hydrous oxidesof iron and aluminum, which had been incompletelyremoved in the isolation of these minerals, by aprocess in which one OH ion was replaced by oneH^PCV ion. That the hydrated iron and aluminumoxides can act effectively in the sorption of phosphatehad been established earlier by Lichtenwalner, Flen-ner, and Gordon (8), by Heck (4), and more recent-ly by Kelley and Midgley (5).

If anionic exchange is largely responsible for phos-phate fixation and if the mineral kaolin can be con-sidered as an anion exchange complex, then certainfundamental concepts concerning phosphate fixationand exchange in soils might be established by study-ing the factors influencing this process using kaolinas a representative soil mineral. The advantages arethat this mineral may be obtained in a relatively"pure" state and that the procedure is analogous tothat which has been used rather widely for estab-lishing facts concerning base exchange reactions withsuch materials as bentonites, kaolinites, etc.

If anionic exchange is a property of kaolin then alyotropic series for anions, similar to that which hasbeen fairly well established for cations, should exist.Such a lyotropic series would show the relative affini-ty of the various anions for kaolin and other soil col-loids. Kurtz, DeTurk, and Bray (7) reported avast difference in the replacing abilities of the variousanions for sorbed phosphate in Illinois soils. Thefluoride ion was the only ion that quantitatively re-placed the phosphate. Dean and Rubins (3), in avery recent report, show that citrate and hydroxylions removed all of the sorbed phosphate from eightsoils having a rather wide range of phosphate sorp-tion capacities; however, fluoride removed the sorbedphosphate from all soils except those having a veryhigh capacity for phosphate sorption. Arsenate andtartrate were less effective than fluoride, and acetatewas the least effective of any of the ions tested.These investigators report that there is always morephosphate than arsenate retained by soils as ex-changeable anions when comparable methods of satu-ration are employed.

Arsenates and phosphates are so much alike chemi-cally it was thought that they would react in a verysimilar manner in anionic sorption and exchange.Data concerning the relative reactivities of these-anions in the exchange reaction with soil colloidswould throw much light on the agronomic practicesto be employed under conditions where large amountsof arsenates are used as insecticides and herbicides.If arsenates are fixed in a manner similar to phos-phate and w,ith about the same energy of fixation,then their addition to the soil would result in a libera-tion of fixed phosphate with subsequent temporaryadvantage for the growing crop; however, an accu-mulation of fixed arsenate in the soil as a result ofcontinual dusting and spraying programs, might leadto a very toxic condition whenever soluble phosphateswere added as fertilizing materials with the resultingrelease of the fixed arsenates into the soil solution.The replacing ability of any ion for another in ionicexchange has been pointed out by Mattson (10) asdepending upon their relative activities or concentra-tions and their affinities or energies of combinationfor the sorbing complex. To establish an ion's relativereplacing ability, it would be necessary to contact awide variation of concentrations of the ion in solutionwith a constant quantity of sorption complex contain-ing a definite amount of the ion to be replaced andthen to determine the relative concentration of thetwo ions in the equilibrium Solution. This has beenthe method most commonly employed for determiningthe relative replacing abilities of the yarious cationsin base exchange reactions.

'Contribution No. 6n, Massachusetts Agricultural Experiment Station, Amherst, Mass."Research Professor of Soil Chemistry.3Pigures in parenthesis refer to "Literature Cited", p. 169.

161

l62 SOIL SCIENCE SOCIETY PROCEEDINGS 1946

If it is established that arsenates are sorbed bysoils in equal quantities and in a -similar manner asphosphates, and will quantitatively replace fixed phos-phates, then the phosphate fixation capacity of a soiland its fixed phosphate content might be estimated byusing arsenate as the replacing ion. The advantagesof such a procedure are quite evident when one con-siders that arsenate may be determined by direct titra-tion by taking advantage of the quantitative reduc-tion of arsenate to arsenite in a strongly acid solutionand that appreciable quantities of arsenate are not

.found in a great majority of soils. A further ad-vantage is that the phosphate liberated from the fixedcondition could also be estimated in the presence ofarsenate.

The 'purpose of this research has been to establishthe relative positions of phosphate and arsenate inthe lyotropic series for anion exchange with soil col-loids and to obtain from these data necessary infor-.mation for formulating a method to determine theanion exchange capacity of soils. This research -wasintended to further clarify the role of kaolin in anionexchange and sorption reactions.

. EXPERIMENTALINFLUENCE OF GRINDING ON ANION SORPTIVE

CAPACITY OF KAOLIN

A sample of acid-washed kaolin4 was selected asthe "pure" material to be used in.all of the testsreported. Other-samples of kaolin and halloysite weresubjected to many of the same reactions and gavevery similar or identical results, indicating that thedata obtained are representative for the minerals ofthe kaolin type. This acid-washed kaolin containedless than 0.2% of material soluble in 10% hydro-chloric acid.

To determine phosphate sorption capacity of thekaolin, duplicate 5-gram. samples were shaken for 18hours with 50 ml of i.o N sodium phosphate solutionhaving a pH value of 3.0, the resulting suspensioncentrifuged, and the phosphate remaining in thesupernatant' solution determined. Duplicate sampleswere also treated with a i.o N sodium arsenate solu-tion at pH 3.0 and the same experimental procedureswere employed-. Neither phosphate nor arsenate wassorbed by the kaolin under the conditions of theseexperiments.

- Murphy (n) showed that ballmilling increasedthe sorptive capacity of kaolin for phosphate and itwas later indicated by Stout (17) that this processdid not alter the crystal structure of the kaolin. Thekaolin selected for this study, and shown to have noactivity for anion sorption, was ballmilled continu-ously for 30 days. Samples were removed at definiteintervals to determine the progress of the activationof this kaolin. Table I shows the sorptive capacity ofthe kaolin for arsenate and phosphate after it hadbeen ballmilled for 10, 20, and 30 days. These datawere obtained' by the same method employed for de-termining the sorptive capacity of the original kaolin.

TABLE i.—Influence of grinding on the sorptive capacityof kaolin.,

MM

e.* AsO4 sorbed per gram of kaolin. .e. PO4 sorbed per gram of kaolin . . . .

Time of ballmilling

10.days

3-263-50

20days

3-423-77

'30days

3-764.40

'*Milliequivalent. In this report this term refers to one-third the milli-molar weight of the tribasic acids HaPO* and H3AsO< and is equivalent to10.34 milligrams of P and 24.97 milligrams of As.

These data show that kaolin may be made activefor anion sorption by ballmilling and that the activityincreases as the ballmilling is continued. A sample ofthe acid-washed kaolin "micronized" by the Micro-nizing Processing Co., Moorestown, New Jersey, hadno sorptive capacity for anions when tested in thesame manner. The use of the term ballmilled kaolinin the remainder of this report refers to the acid-washed kaolin which had been ballmilled for 30 days.

EFFECT OF TIME ON SORPTION OF PHOSPHATE .AND ARSENATE BY KAOLIN

It has been shown by numerous investigators thatphosphate sorption increases with time of contactwith the sorbing material (i, 2, 7). It seemed advis-able, therefore, to establish^an arbitrary time of re-action, to be used in future investigations, whichwould give values very nearly those obtained overlonger periods of time. For this purpose a series ofI-gram samples .of ballmilled kaolin was placed in 15ml calibrated conical centrifuge tubes and 10 ml ofsodium phosphate solution, containing 51.3 m. e. ofphosphate and having a pH of 3.1, were added to eachtube. The tubes were then shaken end over end, andat regular time intervals two tubes of the series were-removed from the shaker for analysis. The analyticalprocedure consisted of centrifuging the suspensions;determining the pH of the supernatant solution withthe Beckman pH meter, employing the glass elec-trode; washing the residues three,times with 10 mlportions of water by resuspending and'centrifuging;and combining the solutions and 'washings of eachsample in a volumetric flask from which appropriatealiquots could be taken for the phosphate determina-tion by the method of Sherman (16) after silica hadbeen removed by perchloric acid dehydration. Theamount of phosphate sorbed was determined by sub-tracting the amount found in the solution from theamount originally added to the sample. That thismethod would give the same results as analyzing theresidues for phosphate, was shown when the residuesof one- series were analyzed and gave identical resultsfor both methods. Since the use of the solutions en-tails fewer analytical operations, this technic was usedin all subsequent experiments.

A second series of i-gram samples of ballmilledkaolin>was"treated in a similar manner, except that10 ml of sodium arsenate solution containing 38.8

•"Obtained from J. T. Baker Chemical Co., Rahway, N. J., and designated1 as lot No. 73115.

SIBLING : KAOLIN IN ANION SORPTION AND EXCHANGE

TABLE 2.—Effect of time on sorption of phosphate and arsenate by kaolin.

i63

< Description

M.e. PO4 sorbed per gram kaolin from 10 ml soln. containing 51.6 m.e. PO4. . . .

M.e. AsO4 sorbed per gram kaolin from 10 ml soln. containing 38.8 m.e. AsO4. . .

Apparent volume of arsenated kaolin (ml) . . . . . . . . ' . . . . . . . . . . . . . . . . . . . . . . . .

Time of reaction

I hr.

5-54-3°1.82.73.601.8

1 8 hrs.

°6.65.002.08.24-653-1

72 hrs.

IO.O5-071.8

10.65.002.8

144 hrs.

13.05.08i-7

n-35-H2.7

240 hrs.

14.35-121-7

11.75-153-2

m. e. of arsenate at pH 3.0, were used in place of thesodium phosphate solution. The same general technicwas employed as has been described for the phosphateseries except that the arsenate determination wasmade by the> titration method of Kolthoff and Fur-man (6) without the removal of silica which does notinterfere with this determination. This method wasmodified slightly to use a 20% solution of KI ratherthan the prescribed dry 'salt which was found toproduce a heavy, difficultly, soluble precipitate whichinterfered with the accuracy of the subsequent titra-tion.

The amounts of phosphate and arsenate sorbed byballmilled kaolin as influenced by the time, of contactof the kaolin with the solutions are recorded in Table2, along with the observed pH's of the equilibriumsolutions and the apparent volumes of the phosphatedand arsenated kaolins.

When the effect of time on the sorption of arsenateor phosphate by kaolin, as shown in Table 2, is illus-trated graphically in Fig. i, it will be observed that avalue approaching maximum sorption is not reacheduntil the time exceeds 144 hours. It should not beinferred that maximum sorption is reached at anyfixed time since Kurtz et al. (7) observed that phos-phate was still sorbed in large quantities by Illinoissoils after 45 days' contact; however, the curves inFig. i indicate that the rate of sorption of theseanions by ballmilled kaolin as a function of time be-comes small and nearly constant for both phosphateand arsenate after 144 hours of shaking. The observedchanges in the pH's of the equilibrium solutions to afairly constant value within the first 72 hours indicatethat the exchange of arsenate or phosphate ions forhydroxyl, as suggested by several investigators, israther rapid and that a large proportion of the anionssorbed after this time may result from molecularsorption rather than ionic exchange. The markedincrease in the apparent volume of the arsenated col-loid as compared to the nearly constant volume of thephosphated colloid suggests a sorption of solution orsolvent attended by swelling. This observed swellingwas considered when the analytical technic used in thearsenate systems was devised since it was observedthat the volume of the supernatant solutions variedconsiderably and therefore aliquots of the solutionwould not represent a definite fraction of the un-sorbed arsenate.

INFLUENCE OF CONCENTRATION AND REACTIONON SORPTION OF PHOSPHATE AND ARSENATE

BY KAOLIN

It has been reported by various investigators thatthe sorption of phosphate by kaolin increases pro-gressively as the acidity of the solution increases frompH 7.0 to 3.0 (i, i i) . In most sorption reactions,except those involving true ionic exchange, the sorp-tion of the solute from solution depends upon theconcentration of the solution if all other conditionsare kept constant. Further, the percentage sorptionis greater in the more dilute solutions.

To determine the influence of solution concentra-tion on phosphate and arsenate sorption by kaolin, aseries of I-gram-samples of ballmilled kaolin wasshaken for 216 hours with 10 ml of sodium phosphatesolutions at pH 3.6 and containing from 5.1 to 72.1m. e. of phosphate. A parallel series was shaken with

o

«cOS.

edLJ*•«*QLJODQC

Oz,<*LJ

PO+ SORBED

TIME IN HOURSFIG. i.—Effect of time of sorption of phosphate

and arsenate by kaolin.

164 SOIL SCIENCE SOCIETY PROCEEDINGS 1946

TABLE 3-—Influence of solution concentration on sorption of phosphate and arsenate by kaolin.

Phosphate sorption

SampleNo.

i2345

. 6789

10ii1213

M.e. PO4added

per gramkaolin

5-110.320.625-730.936.038.641.243-746.351-561.872.1

M.e. P04sorbed

per gramkaolin

2-95-88-39-7

11.411.612. 112.013-013-013.814.214-5

pHofequilibrium° solution

6.506.506.10 •5.855-705-555-455-395-385-305.154-954-87

Apparentvolume of

phosphatedkaolin, ml

1.40 'i-55i. 601.701.65

- .1.701.701.701.70 ,1.701-75i. 801.90

Arsenate sorption

SampleNo.

i23456789

10ii1213

M.e. AsO4added

per gramkaolin

6.613-3

' 2O.O26.633-336.639-746.650.053-356.659-966.6

M.e. AsO4sorbed

per gram' kaolin

3-05-57-78.9

IO.Iii. 813-614.216.116.818.319-52O.2

pHofequilibrium

solution ,

5-87 '5.605-525-355.105.025.104.864.864.854-854.854.62

Apparentvolume ofarsenatedkaolin, ml .

i-71-71-7i-71-73-43-53-73-9

- 4.04-25-o4-5

sodium arsenate solutions containing 6.6 to 66.6 m. e.of arsenate at pH 3.0. After shaking, the amounts ofthe sorbed arsenate and sorbed phosphate were de-termined by the technics previously described. Them. e. of arsenate and phosphate sorbed by the ball-milled kaolin from solutions of different concentra-tions, the pH's of equilibrium solutions, and theapparent volumes of the arsenated and phosphatedkaolins' are recorded in Table 3.

These data for phosphate sorption are presentedgraphically in Fig. 2 by the curve designated pHvariable and show that phosphate is sorbed by theballmilled kaolin in a regular manner and forms a

ME P04 ADDED PER GRAM OF KAOLINFIG. 2. — Influence of solution concentration and reaction

on the sorption of phosphate by kaolin.

typical sorption curve. The increase in pH of theequilibrium solution indicates an exchange of phos-phate ions for hydroxyl ions with the increase beingsmaller for the more concentrated solutions becauseof the greater buffer capacity of these solutions. Thereis a slight increase in the apparent volume of thephosphated colloids as the amount of the sorbed phos-phate increases indicating swelling of the particles asa result of phosphation or the imbibition of the sol-vent or solution.

These data for arsenate sorption are shown graphi-cally in Fig. 3 by the curve designated pH variable

>

102.0 30 » 7 0 80ME AsO* ADDED PER CRAM OF KAOUN

FIG. 3.—Influence of solution concentration and reactionon the sorption of arsenate by kaolin.

SIBLING: KAOLIN IN ANION SORPTION AND EXCHANGE 165

and pH constant at 4.86. The sorption curve dis-closes that in the more dilute solutions the sorptionof arsenate by ballmilled kaolin follows the same gen-eral pattern as was shown for phosphate; however,an entirely different phenomenon takes place in themore concentrated solutions. Whenever the concen-tration of the arsenate exceeds 33 m. e. per gram ofkaolin in 10 ml of solution there is a marked increasein the volume of the arsenated colloid. The volumeincreases progressively as the concentration of thesolution increases and is attended by an almost con-stant pH of the equilibrium solution. This indicatesvery strongly that secondary reactions involving sorp-tion of molecules predominate at the higher concen-trations after the initial exchange reactions have beencompleted.

The sorption of phosphate and arsenate by ball-milled kaolin from the solutions of different concen-trations but identical initial pH values, results inequilibrium solutions having a very wide differencein reactions; therefore, an attempt was made to regu-late the reaction of the solutions so that the equilibri-um reactions would be nearly constant regardless ofthe concentration of the solution. To achieve thiscondition one.series was prepared to contain 3 ml ofglacial acetic acid in each 10 ml of solution of phos-phate or arsenate used for each sample of the ball-milled kaolin. These solutions had an initial pH valueof 2.6 regardless of the concentration of the phos-phate or arsenate. Another series was prepared con-taining sodium arsenate or sodium phosphate at pH7.0 and from the data presented in Table 4 there wasno appreciable shift in the pH's of these solutionsupon contact with the kaolin. This confirms the ob-servations of many other investigators (i, 5, il).The sorption and analytical details are identical tothose used in the preceding experiments. The dataconcerning the sorption of phosphate and arsenatefrom solutions of varying concentrations but havingidentical equilibrium pH's are shown in Table 4.

Data for phosphate are shown in Fig. 2 and thosefor arsenate are graphically represented in Fig. 3.The graphical representations show clearly that thesorption of both arsenate and phosphate increasemarkedly as the acidity of the solutions is increasedfrom pH 7.0 to 3.3, and that the sorptions in bothinstances increase as the concentration of the solute,in a given volume of solution, increases. These dataare generally what one would predict for sorption ofthese ions by hydrous alumina or indicate that finelydivided kaolin exhibits sorption properties very simi-lar to those' of hydrous alumina. Weiser, Milligan,and Purcell (19), Marion and Thomas (9), andMattson (10) observed that aluminum "hydroxide"began to precipitate at pH 3.0 when alkali hydroxideswere added to aluminum chloride solutions and thatthe precipitation was complete at pH 7.8 to 8.0. Whenthe sulfate was used, the precipitation was complete,at a much lower pH of the equilibrium solution andthe precipitate, contained sulfate in measurable quan-tities. With phosphate, Mattson showed that the pre-cipitate was isoelectric at pH 4.0 to 6.45 and con-tained phosphate; the isoelectric pH and the phos-phate content depended upon the concentration ofthe phosphate — the higher the phosphate concen-tration the lower the isoelectric pH and the morenearly the ratio of phosphate/aluminum approachedunity.

These data indicated the necessity of determiningdefinitely whether the activity of the ballmilled kaolinto sorb anions was due to the exposed basoid groupsof the crystal lattice or to the free alumina that had.been sheared from the alumino-silicate complex bygrinding.

INCREASE IN SORPTIVE CAPACITY OF KAOLIN

BY ACTION OF ALKALI AND HEAT

Strong bases are widely used to decompose alumino-silicates in quantitative analysis and are also em-

TABLE 4.—The influence of solution concentration on arsenate and phosphate sorption by kaolin from solutions having• constant equilibrium reactions.

Phosphate sorption

SampleNo.

i2

' 3456789

10u12

M.e. POiadded

per gramkaolin

10.3 ,20.731.04i-351-662.072-34-69.2

13.818.423-0

M.e. PO4sorbedper gram

~ kaolin

IO.O13-013-814.414.915-616.6

I.O2.O2-73-44-0

pHofequilibrium

solution

3-33-33-33-33-33-33-3

. 7-7-7-7-7-

Apparentvolume of

phosphatedkaolin, ml

1.61.8i-9i-91-92.O2.O1-5i-4i-51-51.6

i Arsenate sorption

SampleNo.

i23456789

. 10ii12

13H

M.e. AsO4added

per gram •kaolin

7-915-723-631-539-447-255-16.5

13-119.626.232.739-245-8

M.e. AsO4sorbed

per gramkaolin

6.7I I.O14.916.419.02I.O22-9

O-30.50.91-42.22-43-1

pHofequilibrium

solution

3-33-33-33-33-33-33-37-07-07-07.07.07.07.0

Apparentvolume ofarsenatedkaolin, ml

1-72-34-34-54-54-75-31-31.4i-51.62.O2.O2.O

i66 SOIL SCIENCE SOCIETY PROCEEDINGS 1946

. ployed as dispersing agents for soil colloids. If thetreatment of kaolin with a strong base would increaseits sorption capacity for anions, this effect could beattributed either to higher dispersal of the mineral,with subsequent exposure of anion -sorbing activegroups, or to the decomposition of the kaolin into itsconstituents — alumina and silica.

For these tests i-gram samples of the acid-washedkaolin, which had not been activated by ballmilling,were treated with sodium hydroxide solutions ofknown strength in the following ways: (a) Shakenend over end for 15 hours; (b) heated to boiling,cooled, and allowed to stand for 15 hours; (c) heatedat 75° C with a reflux condenser'for 15 hours; and(d) heated to dry ness in an oven at 135° C for 15hours. To the resulting alkaline mixtures were addedenough arsenic acid to adjust the pH to 3.0 andenough' sodium arsenate at pH 3.0 to -give the de-sired concentration of -arsenate. The sorption technicand the analytical procedures were identical to thosepreviously described. Sorption of arsenate, attendedby swelling of the kaolin, was observed only with thedry heating at 135° C and the wet heating at 75° C.The sorption by the 'material which had been dryheated was many times that shown by the materialprepared by wet heating. Heating of the kaolin at135° and 700° C for 18 hours, without the additionof the alkali, did not activate the kaolin for anionsorption.

Other minerals, talc, alumina, silica, and electro-dialized bentonite, were heated with sodium hydrox-ide at 135° C for 18 hours and their arsenate sorp-tion abilities deterrnined. Of these minerals only.alumina sorbed arsenate although it had no capacityfor arsenate sorption previous to the alkali treatment.

That the quantity of sodium hydroxide, employedin the activation of kaolin influences its capacity tosorb arsenate is shown in Table 5.

These data show that there is an increase,of anionsorption by the NaOH activated kaolin as the quan-tity of alkali increases. It seems logical, therefore, toassume that the kaolin had been decomposed by theNaOH to give sodium aluminate and sodium silicate,both of which were precipitated as hydrous oxidesupon addition of the arsenate solution at pH 3.0, andthat the arsenate was sorbed by the hydrous aluminaor formed an insoluble basic aluminum arsenate.

TABLE 5.—The arsenate sorption of kaolin as influenced by theamount of sodium hydroxide used in its activation by heating

at i35°Cfor 18 hours.

SampleNo.

i2 •

' 34

Weightof

kaolin,grams

I.O1.01.0I.O

M.e.NaOHadded

per gram

1-523-056.10

12.20 .

M.e.AsO4

addedper gram

50-0 .50.050.050.0 •

M.e.AsO4

sorbed ' .per gram

7-55I3-32I7-5I19.84

Apparentvolumeof arse-nated

colloid,ml

2-53-84-44-2

REDUCTION OF SORPTIVE CAPACITY OF ANIONACTIVE KAOLIN BY ACID EXTRACTION

If the active constituent of ballmilled kaolin andof alkali-activated kaolin is hydrous aluinina, then itshould be possible to extract this active constituentwith mineral acids. To find whether the active con-stituent could be extracted with mineral acids, sixo.4-gram samples of ballmilled kaolin and three 0.4-gram samples of alkali-activated kaolin were ex-tracted with three io-ml portions of 10% HQ. Theresidues were washed free of chloride ions with waterand the combined acid extracts and washings of eachsample were used for silica and alumina determina-tions. The acid extraction removed 0/0778 gram ofAl2Os and 0.0033 gram of SiO2 from each sampleof the ballmilled kaolin and 0.0791 gram of A12O3and 0.0794 gram of SiO2 from each sample of thealkali-activated-kaolin. Three acid extracted residuesof the ballmilled kaolin were tested for arsenate sorp-tion and three were tested for phosphate sorption.Neither of these anions was sorbed by the acid de-activated kaolin, although 9.1 m. e. of arsenate and12.i m. e. of phosphate were sorbed by i-gram sam-ples of the unextracted kaolin. Three samples of thealkali-activated kaolin which had been extracted withacid were tested for arsenate sorption. There wasnone; however, the alkali-activated kaolin which hadnot been extracted sorbed 9.4 m. e. of arsenate per•gram. Thus 10% HCI was effective in removing theactive constituent from both the ballmilled kaolin andthe alkali activated kaolin and in both cases theamounts of alumina removed were practically iden-tical while the amounts of silica were quite .different.This indicates that the activity of the samples wasassociated with the extractable alumina content.

That strong mineral acids are capable of decom-posing kaolin is well known since this procedure isused in the commercial production of aluminum salts.It is therefore conceivable that the 10% HQ couldhave decomposed the activated kaolin and thus re-moved the anion sorption groups from the crystallattice rather than simply soluting the free alumina.To determine whether the action was one of solutionrather than of decomposition, the experiment wasrepeated using o.i N HQ for the extractant-on oneseries of samples arid o.i M tartaric acid for another.It was believed that neither of these reagents 'could beconsidered active enough to decompose the kaolin,regardless of particle size, but that either or both ofthe reagents would be effective in dissolving hydrousalumina. Each sample was extracted until the ex-tractant had the same pH before and after contactwith the kaolin indicating that all basic substanceshad been removed from the samples. Six extractionsof 10 ml each were required for each sample withboth reagents and were followed by three extractionswith hot water to remove the last traces of the acids.The 'extracts were used for alumina determinationsand it was found that the o.i N HCI removed exactlythe same quantity of alumina (0.1945 gram per gramof kaolin) as did the 10% HQ while the o.i M tar-taric acid removed only a little less (0.1812 gram).

SIBLING: KAOLIN IN ANION SORPTION AND EXCHANGE

The samples extracted with both o.i N HC1 and o.iM tartaric acid were inactive for sorption of bothphosphate and arsenate. 'These data indicate that theactive constituent of ballmilled kaolin and of alkali-activated kaolin is free hydrous alumina and not thealumina of the crystal lattice of the kaolin.

ANION SORPTION BY ALUMINA

If the activity of ballmilled kaolin for anion sorp-tion is due to the free hydrous alumina, then a fresh-ly precipitated hydrous alumina should show similarsorption activity for phosphate and arsenate underthe same experimental conditions. Samples of freshlyprecipitated "aluminum hydroxide" were prepared byprecipitating the aluminum frqm an aluminum sulfatesolution with ammonia at pH 7.0. This precipitatecontained 0.1945 gram of alumina and was washedseveral times with ammonium nitrate and water toremove the excess sulfate. Another series of sampleswas prepared from aluminum chloride and the pre-cipitates were dried and ignited at 700° C to produceanhydrous alumina. Sorption experiments with ar-senate and phosphate were conducted with both thecompletely hydrated "aluminum hydroxide" and theanhydrous alumina. Fig. 4 shows graphically thesorption of phosphate by hydrated alumina and an-hydrous alumina from solutions of increasing con-centrations of phosphate at an initial pH of 3.0. Fig.5 consists of curves showing the sorption of arsenateby anhydrous and hydrated alumina from solutionscontaining increasing concentrations of arsenate at

/ME P% ADDED PER SAMPLEFIG. 4.;—Influence qf hydration of alumina on the sorption

of phosphate.

Mt As04 ADDED PER SAMPLEFIG. 5.—Influence of hydration of alumina on the sorption

of arsenate.

pH 3.0. The dotted curves in each figure representthe sorption of the anion by ballmilled kaolin contain-ing an equivalent quantity of alumina.

These data in Figs. 4 and 5 show that phosphateand arsenate are sorbed in somewhat greater amountsby hydrated alumina than by an equivalent amountof ballmilled kaolin; however, the sorption of theseanions by dehydrated alumina is very much less thanthat by the kaolin. This indicates that the activeconstituent of ballmilled kaolin is a partially hydratedalumina similar to y —A1OOH which was identifiedby Weiser et al. (19) as being present in freshly pre-pared precipitates of alumina formed in solution hav-ing a pH range of 5 to 8. These investigators reportedthat this hydrated alumina was converted to a crystal-line substance having the formula Al2O3-SO3-i.5H2Owhen aged in sulfate solutions at lower pH valuesand that the X-ray patterns were entirely differentfrom those of the unsulfated hydrate. The formationof an analogous crystalline phosphated .aluminaseems likely and would account for the X-ray dataof Stout (17) which he obtained from the materialproduced by treating ballmilled halloysite with phos-phate solutions at pH 3.0 and which he interpretedas being evidence of the formation of a phosphatedhalloysite. That phosphate ions may enter the aluminacomplex to replace other anions was shown by Matt-son (10). This investigator concluded that the re-placement of anions within the complex by otheranions depended upon the relative activity of theions involved; therefore, at very low pH valuesH2PC>4~ and H2AsO4~ ions would be expected toreplace the less active hydroxyls. From a purely

i68 SOIL SCIENCE SOCIETY PROCEEDINGS 1946

stochiometric standpoint, the formation of the basicphosphate from y —A1OOH by replacing the hy-droxyl with the chemically equivalent H2PO4- wouldlead to a compound having a P2O5/A12O3 ratio ofunity which was suggested by Mattson (10) as beingthe ultimate of phosphation. Marion and Thomas (9)have recently presented evidence that certain organicanions could occupy the coordinated valences of thealuminum ion and thus prevent its precipitation by^hydroxyl ions — some of these ions such as citrate'and tartrate were so strongly coordinating that theycould not be replaced by hydroxyls even at high pHvalues. That citrate ions are very effective in replac-ing fixed phosphate was shown by Dean and Rubins .(3) and indicates that the effectiveness of organicmatter in preventing phosphate fixation in the soil isrelated to occupancy of the coordinated valences ofthe alumina by organic anions that are not replacedby the active phosphate ions.

RELATIVE REPLACING ABILITIES OF PHOSPHATEAND ARSENATE IONS

That ballmilled kaolin is a system containing freehydrous alumina in a state of hydration somewhatanalogous to that found in, mineral soils has beenindicated by the investigations reported above. Thismaterial is of ideal physical character for studyingthe relative replacing abilities of various anions inthe anionic exchange-reactions with hydrous^ alumina.

Arsenate and phosphate are sorbed by alumina ofthe ballmilled kaolin in approximately equivalentquantities under the same experimental conditionsuntil the concentration of the solution exceeds 25m. e. in 10 ml of solution. Above this critical concen-tration, the sorption of arsenate exceeds that of phos-phate because of the secondary reactions.

To determine the relative replacing activities ofthese anions arsenate solutions were used to replacesorbed phosphate from the alumina of ballmilledkaolin and phosphate solutions were used to replacesorbed arsenate.

A phosphated alumina was prepared by shaking 50grams of ballmilled kaolin for 7 days with 500 ml ofsodium phosphate containing 5 m. e. phosphate perml and having pH of 3.0. The pH of the suspensionafter phosphation was 4.8. The suspension was cen-trifuged and washed six times with hot water. Thephosphated colloid was dried at room temperatureand the sorbed phosphate was determined by releas-ing it with a sodium hydroxide solution. After 45months' storage, the amount of phosphate which

could be liberated by NaOH was the same—11.3m. e. per gram of colloid.

An arsenated alumina was prepared in an analo-gous manner using sodium arsenate solution at pH3.0. The resulting equilibrium suspension showed apH value of 4.8 and the air-dried colloid contained10.2 m.' e. of replaceable arsenate per gram of kaolinafter 45 months' storage.

A duplicate series of phosphated samples equiva-lent to 0.1945 gram of A12O3 was treated with 10 mlof sodium arsenate solutions adjusted to pH 4.8 andcontaining from 5 to 60 m. e. of arsenate per" sample.These suspensions were shaken end over end for 9-days and then subjected to the same analytical'technicas was previously described. The amounts of phos-phate liberated and arsenate sorbed are presented inTable 6, along with the pH of the equilibrium solu-tions and the apparent volume of the residual colloid.

Data in Table 6 show that the affinity of phosphate, for the alumina is very much greater than' that of

arsenate. The arsenate was very ineffective in replac-ing the sorbed phosphate even in the higher concen-trations of solution. The amount of arsenate sorbedwas greater than the amount of phosphate releasedand may be accounted for by the replacing of hy-droxyls that had not previously been replaced by thephosphate ions, since there was an increase in the pHin all cases.

A duplicate series of arsenated alumina samplesequivalent to 0.1945 gram of AUOs were treated in alike manner with sodium phosphate solutions at pH4.8. The solutions contained from 5.5 to 88.5 m. e. ofphosphate in 10 ml. The same general technic wasused for the replacement and analysis as was em-ployed in the preceding experiment. The amounts ofarsenate released and of phosphate sorbed by eachsample are shown in Table 7, together with the pH'sof the suspensions arid apparent volume of the resi-dual colloid.

The data in Table 7 show also that phosphate has agreater affinity for alumina than does arsenate. Thatmuch more phosphate was sorbed than there wasarsenate released may have resulted from the sorp-

. tion of molecules or anions without equivalent ex-change as was postulated for the excessive sorptionof arsenate by the alumina of ballmilled, kaolin fromthe more concentrated'solutions. It will be noted thatswelling of the arsenated colloid upon contact withthe phosphate solutions was similar also to that ofthe more concentrated solutions of arsenate withalumina. The lower pH value of the equilibrium solu-

TABLE 6.—The replacing of sorbed phosphate from hydrous alumina by sodium arsenate at pH 4.8 and by sodium hydroxide. '

SampleNo.

i23456

M.e. AsO4added per

sample

5-010.02O.O40.060.0

(Extracted with

Symmetrycone.

0.440.891-773-545-30

u .6 m.e. NaOH i

M.e. AsO4sorbed per

sample

o.470.650.84I.IO1-35

a 10 ml solution)

M.e.'PO4released per

sample

0.090.250.52o.751.22

11.29

pHofequilibrium

solution

5-75-65-55-25-0".4

Apparent volumeof equilibrium

colloid, ml

1.61.61.61-71.8!.»>

SIBLING: KAOLIN IN ANION SORPTION AND EXCHANGE 169

TABLE 7.—The replacing of sorbed arsenate from hydrous alumina by sodium phosphate at pH 4.8 and by sodium hydroxide.

SampleNo.

I23456

' M.e. PO4added per

sample

5-5II.O22.144.288.5

(Extracted with

Symmetrycone.

0.54i. 082.174-348.68

1 1. 6 m.e. NaOH ii

M.e. POisorbed per

sample

3-64-88.2

IO.210.8

i 10 ml solution)

M.e. AsO4released per

sample

2.93-96.3 -8.18-5

10.2

pHofequilibrium

solution

3-94.24i4.64.6

U-5

• Apparent volumeof equilibrium

colloid, ml

2.22.73-52-32.32.1

tions may result from this additional sorption as aresult of selective anion sorption rather than mole-cular sorption as was previously postulated. That thesorption of different anions by hydrous alumina willvary considerably was pointed out by Weiser (18)who explained -that neutralization of the charge orthe exchange is an equivalent reaction but that theamount of any ion sorbed by the neutralized particlewould depend upon the nature and concentration ofthe electrolyte. It is obvious from the data just pre-sented that the use of arsenate for determining theaiiion exchange capacity of soils is not to be recom-mended even though the amount of arsenate sorbedby a given soil might be very nearly the same as theamount of phosphate under the same conditions. Thefact that arsenate will not effectively replace fixed orsorbed phosphate would cause a considerable error indetermining the anion exchange capacity of thosesoils which already contained a large quantity offixed phosphate. Piper's" (12) method, which is basedupon saturating the sorption complex with phosphatefollowed by release of the fixed phosphate by sodiumhydroxide, seems to present a favorable method sincethe practical agronomists are largely interested in thebehavior of phosphate and the measure of the soil'sability to fix this ion.

That soils containing large amounts of fixed resi-dual arsenates might release toxic concentrations ofthis element as a result of phosphate fertilizationseems very probable.

SUMMARY AND CONCLUSIONS

Kaolin may be activated for anionic sorption byballmilling or by heating with alkali. The longer thetime of ballmilling or the greater the amount of alkaliused, the higher will be the sorption capacity of theresulting material; however, the rate of increase ofactivation becomes less as these activating processes.are increased.

The active constituent of ballmilled and alkali acti-vated kaolin was removed by extracting the materialswith .0.1 N HC1, o.i M tartaric acid, or 10% HC1.This active constituent was believed to be a hydrousalumina such as y —A1OOH which is known to bepresent in freshly precipitated alumina and to exhibitproperties similar to the ballmilled kaolin.

The alumina of ballmilled kaolin sorbed phosphateand arsenate in practically equivalent amounts fromthe more dilute solutions of these ions; however,from the more concentrated solutions the sorption of

arsenate far exceeds that of phosphate. The amountof either anion sorbed was dependent upon the re-action of the equilibrium solution and the initial con-centration of the solution. The lower the pH withinthe range of pH 3.0 to pH 7.0, the greater the sorp-tion and the higher the concentration at any fixed pH,the more of the anion was sorbed per unit of alumina.

Freshly precipitated hydrous alumina sorbed phos-phate and arsenate in greater quantities than anequivalent quantity of alumina contained in ball-milled kaolin. Freshly prepared anhydrous aluminawas much less active in anion sorption than thealumina of ballmilled .kaolin. Aged commercial"aluminum hydroxide" had no anion sorption activityunless it had been activated by alkali and heat.

Phosphates were effective in replacing sorbed ar-senate from the alumina of ballmilled kaolin, but thereaction was not an equivalent one. Arsenates re-placed only a very small percentage of the sorbedphosphates from the alumina of the ballmilled kaolineven when the concentration of the arsenate wasfive times that of the sorbed phosphate. Therefore,the use of arsenate as an analytical reagent for meas-uring the anion sorption of soils or for determiningthe exchangeable phosphorus in soils is not recom-mended.

170 SOIL SCIENCE SOCIETY PROCEEDINGS 1946