77Powder Technology, 54 (1988) 77 - 83

Charge Effects in the Adsorption of Polyacrylamides on Sodium Kaolinite andIts Flocculation

G. ATESOK*, P. SOMASUNDARAN and L. J. MORGAN

School of Engineering and Applied Science, Columbia Univeraity, New York, NY 10027 (U.S.A.

(Received January 22, 1987; in revised form September 15,1987)

SUMMARY

Adsorption of relatively high molecularweight non-ionic and anionic po lyacry lam ideson sodium kaolinite has been studied simulta-neously with various flocculation responses(per cent solids settled, settling rate andsupernatant clarity) under controlled condi-tions of ionic strength, temperature and agita-tion. Flocculation properties are correlatedwith polymer adsorption and particle zetapotential, and mechanisms of flocculation areexamined as a function of the charge of thepolymer and clay.

ionic strength not only affects polymer-solvent interactions in the bulk and at thesurface, but also adsorbs competitively on thesolid surface [7,9]. pH also shows complexeffects on the adsorption of polymers,affecting both functional groups on polyacryl-amide and charging properties of clay. Sepa-rate charging mechanisms for clay edges andfaces can even result in opposite charges.

While polymer adsorption clearly governsflocculation, the effect it has depends on theresponse monitored [8]. Per cent solidssettled and settling rate can show quite differ-ent trends from supernatant clarity. Theseflocculation properties reflect the differentinteractions that can occur, in a clay-polymersystem [10]. For instance, settling rate is ameasure of the size and density of the flocsand floc networks and finally of floc compres-sibility. Supernatant clarity reflects the sizedistribution of flocs and the ability of thepolymer to attach to fine particles. A systemcould have a high settling rate and yet yield aturbid solution of suspended fines. In order tofully elucidate flocculation, it is important tosimultaneously measure adsorption densityand flocculation response and to monitorseveral responses.

The study examining flocculation responsesof sodium kaolinite to the adsorption of non-ionic and anionic polyacrylamides isexpanded in this work in order to ascertainany correlation between polymer adsorption,flocculation and zeta potential. Various floc-culation responses (solids settled, settling rate,supernatant turbidity) to reagentizing time,polymer charge and to pH are monitored.

INTRODUCTION

Polymer adsorption has been widely inves-tigated primarily because of its importance inthe understanding of the flocculation mecha-nisms utilized in various industrial applica-tions such as beneficiation of mineral fines[1,2], waste water treatment [3] or paper-making [4]. Among the most frequently usedflocculants, polyacrylamides have been exten-sively studied, but their flocculation behavioris not fully understood due to the complexityof the adsorption phenomena influenced by anumber of polymer, mineral and solutionproperties.

Polymer type and molecular weight as wellas concentration and dosage influence adsorp-tion and impact on flocculation response[1,2,5 - 7]. Solution properties such as tem-perature, pH and ionic strength also play animportant role in the adsorption [1,2,7,8].Interactions governing adsorption depend onthe nature of the adsorbent as well as thepolymer. For instance, salt added to control EXPERIMENTAL

*Present address: Department of Mineral Dressing,Mining Faculty of Istanbul Technical University,Tesvikiye, Istanbul (Turkey).

MaterialsA well-crystallized sample of Georgia kaol-

inite was obtained from the clay repository

@Elsevier Sequoia/Printed in The Netherlands0032-5910/88/$3.50

78

~~

at the University of Missouri. The surface areaof this sample, determined by the nitrogenadsorption technique (BET), is 9.94 cm2/g.Particle size distribution of this sample hasbeen determined by a Leeds and NorthropMicrotrac Particle Size Analyzer to be 80%under 5,um and 100% under 22 ,um. A homo-ionic sample of this kaolinite was preparedusing a standardized technique describedearlier [11, 12].

The I~-labeled non-ionic polyacrylamide(PAM, 5.2 X 106 molecular weight) and 33%hydrolyzed polyacrylamide (HPAM, 6.6 X 105molecular weight) used here were synthesizedand characterized by American CyanamidCompany.

Fiscber-certified HCl and NaOH were usedto adjust the pH of the suspension. AmendDrug and Chemical Company reagent grade,99.96% pure NaCI was used to adjust theionic strength. All solutions were made intriply distilled water.

MethodsFlocculation procedureThe details of the flocculation experiments

have been described elsewhere [8]. Importantpoints include, 2 hours pre-conditioning ofclay with salt solution before polymer isadded. The final solid/liquid ratio is 0.025.For per cent solids settled, 100 cm3 of suspen-sion is suctioned out after 30 s settling andthe two portions are f"lltered, dried andweighed. Settling rate is determined in a flat-bottomed graduated cylinder and turbidity ismeasured in the cylinder after 10 min ofsettling. A Zeta Meter is used to measure theelectrophoretic mobility of clay particlesfrom which the zeta potential is obtained.

shown in Figs. 1 and 2 for PAM and HP AMat natural pH. Adsorption shows typicalpolymer isotherm behavior for both poly-mers, reaching plateau adsorption values of8 mg/g for the non-ionic and only 1.7 mg/gfor the anionic polymer. Whereas the plateauis not attained until 200 mg/kg for the non-ionic, it is reached at considerably lowerdosage for the anionic polymer, 50 mg/kg.Clearly, charge-charge repulsion plays agoverning role restricting adsorption of thenegatively charged polymer. It is interesting,however, that the initial isotherm slope issteeper for the anionic than non-ionicpolymer, suggesting stronger adsorption. .Typically, PAM adsorption is thought tooccur via hydrogen bonding; here, there isevidence that, for the hydrolyzed polymer,the charged functional group contributes tothe adsorption step, probably at cation sites,until repulsive forces become predominant.

R~ULTS AND DISCUSSION

Monitoring the effect of reagentizing timeon per cent solids settled and on polymeradsorption establishes that polymer adsorp-tion (from a 25 ppm solution) is finishedwithin 2 min, plateauing at 1 mg/g for bothPAM and UP AM. Solids settled reaches itsstable value after 10 min conditioning..Thus,10 min conditioning with the polymer is usedfor the rest of the experiments.

Effect of polymer doseThe comparison of adsorption density and

flocculation as reflected in settled solids is

79

"~I-' ""\_7;..c'_""'~i'- ; 5.2"" _/Na-

1CoeIiIIU.pM : 4.8-5.0IlL : 0.025T : 24*2"CI : 3 '10-2 .-v..s

. NoCI~ C"".Speed: 1200rp8I. r ~ Condo Tim. : 10mln..'..

...

: 1.0...II.0.8

...'C.c.o.sz:;l-t; 0.4. '

\..~ ,

i{-'O..~ ~ , ,

~

Gal

100 ~z

80~I-

&ojM. . . . . . . . . 140 z 20 !

0 0 #0 2 4 . . 10

AOSORPTION ~NSIT~ ""I,. . . . . . . . . . .0 0.2 OA oa Q8 1.0

SURFAC£ COVERAGE

Fig. 3. Flocculation reeponses of sodium kaolinite asa function of lurface coverage by PAM.

significance. This response can be seen inFigs. 5 and 6 for PAM and HP AM, respec-tively. The non-ionic polymer clarifies thesuspension rapidly and reaches a high finalclarity but only at low doses. Even thoughadditional polymer enhances settling, theclarity suffers drastically, suggesting that asignificant amount of the fine fraction of theslurry lags behind the settling mas andremains suspended. It is important to notethat at the intennediate doses, even though

. The effects of the two polymers on the sus-pension stability are quite opposite; while thenon-ionic polymer causes flocculation of thekaolinite at low dosage, the anionic polymerproduces only dispersion, lower per centsolids settled, under the tested conditions inspite of adsorption.

Since per cent solids settled cannot be con-sidered to be an absolute measure of floccula-tion of these types of suspensions, two otherproperties of the system, settling rate andsupernatant transmittance have been mea-sured. These are plotted in Figs. 3 and 4 as afunction of adsorption density as well as sur-face coverage. Settling rate is calculated fromthe constant settling region, interface heightUB. time. Settling experiments could be doneonly up to concentrations of 30 mg/kg addi-tion for 33% hydrolyzed polyacrylamide sincethe slurry/supernatant interface is not visibleat higher concentrations; complete dispersionof the sample results with the anionicpolymer above 30 mg/kg polymer dosage.

It can be seen from Fig. 3 that both thesettling rate and the supernatant transmit-tance are found to go through a maximum asa function of adsorption of the non-ionicpolymer. The maxima in solids settled,settling rate and transmittance all occur atquite low adsorption density, 1 Ing/g, corre-sponding to only 10% surface coverage. Thissuggests a bridging mechanism betweenparticle-polymer/particle rather than betweenparticle-polymer/polymer-particle. As theadsorption density increases, it beginshampering flocculation, as shown by thedecrease in responses measured.

In contrast to this, Fig. 4 shows that thehydrolyzed polyacrylamide yielded a maxi-mum solely in the case of settling rate. Itproduced only an increase in turbidity uponthe addition of anionic polymer. ComparingFig. 2 and 4 reveals that while adsorptiondoes not enhance per cent solids settled, atvery low dosage it can increase the settlingrate but at the same time supernatant clarityworsens dramatically. In other words, whatlittle adsorption contributes to flocculationonly increases the rate of settling and finesare in fact stabilized by adsorbed polymer.The overall effect is negative.

The slope of the transmittance-time curveis a measure of the speed with which the sus-pension clears and is a parameter of practical

80

with the anionic polymer is dispersed with acloudy supernatant. At higher polymer do~,the non-ionic polymer continues to adsorb inincreasing amounts, enhancing the settling butnot the transmittance. In this concentrationrange, additional adsorption of the hydro-,lyzed polyacrylamide is prevented, presum-ably due to electrostatic repulsion from theadsorbed layer of anionic polymer. In the caseof the non-ionic polymer, electrostatic forcesare unimportant and solvation and hydrogenbqnding forces predominate. Similar to resultson solids settled, these properties show thatmaximum flocculation is obtained at frac-tional saturation coverages which are muchsmaller than the 0.5 which h~ been proposedin the past [13]. Apparently, particles do notflocculate in this case by particle-polymer topolymer-particle bridging. In fact, this inter-action is repulsive. More importantly, theadsorbed anionic polymer molecules require alarger bare share of the particle surface. Thisis consistent with the idea that unchargedpolymers tend to form coils in solution andcharged polymers tend to be extended. Thus,when adsorbing, the former would require lessarea than the latter.

Effect of molecular weightThe effect of molecular weight may be seen

by comparing these results with those usinglower molecular weight polymers in an earlierphase of the work. Not surprisingly, increasedmolecular weight has no effect on non-ionicPAM adsorption density (on a weight basis) orsettling rate. However, the smaller polymerresulted in plateau transmittance at 80%clarity, while the longer polymer yielded amaximum. Or, some steric stabilization iscaused by the longer adsorbed loops in thelatter case. Contrastingly, even a smallincrease in molecular weight of the anionicpolymer impacts on flocculation response.Adsorption density is reduced, probably dueto the increased anionicity per molecule, lead-ing to earlier charge-charge repulsion whilesettling rate is not affected. Lower molecularweight HP AM shows a maximum in transmit'.tance at very low adsorption density, but atthe higher molecular weight, only dispersionis observed. Again, the longer chain could leadto a thicker adsorbed layer which leads to anapparent higher charge density in theadsorbed layer and this in turn to repulsion.

0 2 4 6 8 10 12 14 16SETTLING TIME, .-;IIU'.'

Fig. 6. Supernatant transmittance for HPAM as afunction of settling time.

18

the suspension clarifies more rapidly than inthe absence of the polymer, ultimate clarity ispoor. In contrast to the non-ionic polymer,the ionic polymer makes the slurry super-natant more turbid at all tested dosages. Forboth polymers, the adsorbed layer can lead tosteric stabilization of fine particles.

At low polymer doses, the sample with thenon-ionic polymer is in a flocculated statewith a clear supernatant, whereas the one

81

Effect of pHThe type of surface site is as important as

the functional group of the polymer inadsorption. For sodium kaolinite, the chargecharacteristics of the surface will depend onthe presence of electrolytes and particularlyon the pH of the solution. Figure 7 shows thezeta potential of kaolinite in the absence andpresence of the anionic polymer. The fJgureshows that the isoelectric point of this sampleis 4.7 and does not change in the presence ofHP AM. At low pH, the polymer reduces thezeta potential, masking surface charge. Athigh pH, within the scatter, the anionicpolymer does not significantly affect themeasured interfacial potential.

Results obtainoo for floccuJation andadsorption for the same sample are given as afunction of pH for the anionic polyacryl-amide in Fig. 8 and for the non-ionic in Fig. 9.Adsorption density of the anionic polymercan be seen from Fig. 8 to decrease withincrease in pH, which is consistent with thelack of effect on zeta potential at higher pH.At low pH, where an effect on zeta potentialwas observed, some flocculation is obtained.This is due to bridging between the positivelycharged kaolinite and the adsorbed negativepolymer molecules. The polymer causes nofloccuJation above pH 3 in spite of the factthat there is significant adsorption at pH 5.Near pH 5, the kaolinite particles report a

zero charge and the adsorbed anionic polymerdoes not contribute toward any bridging. Ifanything, there is already sufficient polymeron the surface to cause repulsion betweenadsorbed layers. Above pH 7. where themineral is highly negatively charged, adsorp-tion is 'insignificant and there is no floccula-tion.

Adsorption density is not affected by pHin the case of PAM even though a more con-centrated solution leads to higher adsorption(Fig. 9); hydrogen bonding and solvationmust be the predominant interactions. Inter-estingly, with the non-ionic polymer,increased flocculation is obtained withincrease in pH but only at low initial concen-tration even though the adsorption densityremains essentially unchanged as a function of

82

pH. Differences in conformation of adsorbedpolymer, suggested by observed bulkier flocs,could be contributing to this effect. Thereasons are not clear at present, but theresults suggest a role in bridging played by thenegative surface sites and adsorbed polymerlayer as long as the layer is not so thick thatit masks the surface leading to stabilization.At high polymer dosage, both the adsorptionand flocculation remain approximately thesame with increase in pH until pH 11, wherethe polymer has lost its flocculating powerdue to hydrolysis of the functional groups.

)natural pH do not reflect the effect ofadsorbed anionic polymer which would havebeen expected to yield a more negative poten-tial. However, this is a difficult region tomake measurements and the effect may havebeen missed. Adsorbed polymer does reducezeta potential at low pH, which is the regionwhere flocculation is observed.

Low doses of non-ionic PAM yield a floc.culated state with a clear supernatant. Athigher adsorption, settling is enhanced butclarity worsens, which could be due to stabil-ization of the fine fraction by the adsorbedpolymer. Although the adsorption densitydoes not change, flocculation is enhanced athigh pH. This is attributed to the formationof large bulky flocs in the basic pH region,which may also entrap IMlditional materialwhile settling. Conformational changes inadsorbed polymer layers are not well under-stood at present. There seems also to be somerole played by negative surface sites in thebridging by the non-ionic polymer as seen bycontinued adsorption at high negative inter-facial potential.

\

CONCLUSIONS

Adsorption, flocculation and zeta potentialresults suggest that kaolinite flocculation bypolyacrylamides arises from a bridging mecha-nism in which two particles are linked by anadsorbed polymer molecule, or particle-polymer/particle interaction. As the adsorp-tion density increases, it begins hamperingflocculation. Once the adsorption is highenough, adsorbed polymer sees adsorbedpolymer and repulsion or steric stabilizationresults. Particle-polymer/polymer-particleinteraction leads to dispersion, as evidencedby high turbidity and slow settling in cases inwhich higher adsorption is measured.

In addition to hydrogen bonding forces,electrostatic forces playa role in the adsorp-tion of polyacrylamides. At natural pH, non.ionic PAM adsorption is governed primarilyby hydrogen bonding and solvation inter-actions, while clearly electrostatic forces areimportant for charged polymer. Very lowdoes the non-ionic; however, the initial iso-pH «5), but with even a slight increase ineither of these, dispersion results. The anionic.polymer adsorbs much less on kaolinite thanthose the non-ionic; however, the initial iso-therm slope is higher than that of the non-ionic. This suggests that although the netinterfacial potential is zero, individual positivesurface sites have a higher driving force foradsorption than adsorption by hydrogenbonding in the case of PAM. The negativecharge quickly becomes detrimental as posi-tive surface sites are saturated and/or chargencharge repulsion becomes predominant. Evena small increase in molecular weight of thecharged polymer suppresses adsorption andleads to dispersion. Zeta potential results at

ACKNOWLEDGEMENTS

The authors gratefully acknowledge thesupport of the Multiphase and the InterfacialPhenomena Program of the National ScienceFoundation.

REFERENCES

1 P. Somasundaran, in P. Somasundaran (ed.), FineParticle Proceaing, AIME, New York, vol. 2,1980, p. 947.

2 P. Somasundaran, Consolidation and Dewateringof Fine Particle.. USBM, 1982, pp. 407 - 437.

3 P. A. Rey and R. G. Varsanlk, in J. E. Glasa (ed.),Adl1anc" in Chem;.try ,erie" vol. 213, ACSPubliaher, 1986, pp. 113 - 143.

4 R. H. Pelton, J. CoiL and Interface Sci., III(1986) 475.

5 W. F. WaDes, J. CoIL And Interface Sci.. 27(1968)797.

6 A. F. HoD.nder, M.S. The,;', Henry KrumbSchool or Mines, Columbia University (1979).

7 A. F. Hollander, P. Bomuundaran and C. C.Gryte, J. App. PolymerSci., 26 (1981) 2123.

8 P. Somasundaran, Y. H. Chia and R. Gorelik, inE. D. Goddard and B. Vincent (eds.), PolymerAd80rption and Dilperaion Stability, ACS Sympo-sium Series, 1984, p. 393.

9 P. J. Dodson and P. Somuundaran, J. CoIL andInterface Sci., 97 (1984) 481.

83

10 Donald H. Napper, Polymeric Stabilization ofCoUoidal Dilperaiona, Academic Presa, NYC,1983.

11 A. F. Hollander, P. Somaaundaran and C. C.Gryte, in P. H. Tewari (ed.), AdlOrption from

Solution, Plenum Press, New York, 1981, p. 143.12 H. Van alphen, An Introduction to Cloy Colloid

Chemistry, Wiley Science, New York,I977.13 K. V. La Mer and T. W. Healy, J. PhY8. Chem., 67

(1963) 2417.