Powder Technology, 62 (1990) 253 - 259 253

Flocculation of Oxides using Polyethylene Oxide

E. KOKSAL, R. RAMACHANDRAN*, P. SOMASUNDARAN** and C. MALTESH

Langmuir Center for Colloid. and lnterfacu, Henry Krumb School of Mines, Columbia University,New York, NY 10027 (U.S.A.)

(Received November 1,1989)

SUMMARY to H-bonding forces, physical hydrophobicinteractions and chemical mechanisms. Tadros[41 examined the adsorption and flocculationproperties of silica gel with polyvinyl alcoholand found that maximum flocculationoccurred at the point of zero charge of theoxide and also observed an increase in adsorp-tion of PV A upon heat treating the solid upto 700 °c. Tadros examined the mechanismof adsorption based on the degree of hydra-tion of the substrate. Evidently, other thanH-bonding, no generic mechanism could beutilized to explain polyethylene oxide adsorp-tion. The U.S. Bureau of Mines has found thatpolyethylene oxide is an excellent flocculantfor clays and red mud [5) and correlated theadsorption of the polymer with the electro-negativity index. Polyethylene oxide itselfhas been a widely used polymer in severalsystems and its characteristics have been thesubject of several reviews [6,7). The presentwork focuses on the adsorption of poly-ethylene oxide on clay and the constituentminerals of clay. It was the aim to study theadsorption/flocculation behavior of poly-ethylene oxide on a variety of oxide mineralssuch as silica gel, quartz, hematite, aluminaand kaolinite and utilize the behavior ofpolyethylene oxide on each mineral toexplain its overall flocculation characteristics.The oxides studied vary significantly in theirinterfacial properties such as point of zerocharge (pzc) and degree of hydration.

Flocculation of oxides using polyethyleneoxide was investigated. Polyethylene oxideadsorbed strongly on and flocculated sodiumkaolinite...H-bonding is considered to be thedriving force for polyethylene oxide adsorp-tion on kaolinite. The mechanism of adsorp-tion of polyethylene oxide on kaolinite wasexamined by studying the adsorption/flocculation characteristics of different oxideminerals using the same polymer. Poly-ethylene oxide flocculated silica gel but didnot adsorb on or flocculate hematite andalumina. Quartz was flocculated only belowpH 3. Results are examined in terms of thetype of surface groups and the degree ofhydration of the solids. It is proposed thatadsorption of polyethylene oxide on solidsrequires some optimum hydroxylation of thesurface with displaceable water species.

INTRODUCTION

Polyethylene oxide (PEO) has served as aneffective flocculant in several systems [1,2].However, the exact nature of the mechanismof polyethylene oxide adsorption remainsunexplained. Rubio [3] found that poly-ethylene oxide was an effective flocculanton substances recognized to be hydrophobicbut the polymer was ineffective on hydro-philic substrates such as rutile, quartz andcopper carbonate. Rubio attributed thespecificity of polyethylene oxide flocculation MATERIALS AND METHODS

Na-kaoliniteA homoionic sample of sodium kaolinite

was prepared from a well-crystallized sampleof Georgia kaolinite which was obtained from

.Present address: Union Carbide CorporationLimited, Technical Center, Bound Brook, New Jersey08805...To whom correspondence should be addressed.

0032-59101901$3.50 @ Elsevier Sequoia/Printed in The Netherlands

254

the clay repository at the University ofMissouri. The surface area of the sample wasdetermined by BET to be 9.4 m2/g. Themethodology used for clay preparation hasbeen discussed earlier [8].

The top half of the suspension was sucked outusing a suction device. About 20 cm3 of thesupernatant from the unsettled portion wascentrifuged for determination of residualpolymer concentration. Both 100-cm3 settledand unsettled portions were filtered sepa-rately using filter papers and the residue onthe filter paper was dried and weighed. Theper cent solid secttled was determined fromthe dry weight of the solid in the settled andunsettled portions.

SUica gelSilica gel was purchased from Alfa Products

Inc. The surface area of silica gel was 300m2/g, particle size 70 p.m and pore volume1.6 cm3/g. The characteristics of silica gelwere provided by the manufacturer.

ms88% = .?'!'~~"'~

QuartzNatural Brazilian quartz prepared by wet

grinding and leaching was used. The surfacearea of quartz as measured by BET was 0.5m2/g.

ms + mus

SS% = solid settled per centms = weight of solid in 100 cm3 of settled

portionmus = weight of solid in 100 cm3 of unsettled

portionPolymer concentration was determined

using the method of Attia and Rubio [9].

HematiteSynthetic hematite was obtained from

Alfa Products Inc. The surface area ofhematite as measured by BET was 7.5 m2/g.

RESULTSAluminaAlumina of 0.3 /lm size was purchased

from Alfa Products. Effect of conditioning timeThe role of conditioning time in determin-

ing flocculation was investigated initially at aconcentration of 0.5 ppm of polyethyleneoxide. The polymer was added dropwise for55 s at a steady flow rate of 0.1 cm3fs. Con-ditioning time here refers to the time forwhich the suspension was further stirred afteraddition of the polymer. Figure 1 shows thatmaximum flocculation occurs when the

PolymerPolyox coagulant obtained from Union

Carbide Corporati9n was used as a flocculant.Polyethylene oxide used in the experimentshad an approximate molecular weight of5 million as determined by the manufacturer.Polyethylene glycol (PEG), molecular weight8000, was also used in some experiments.

All experiments were conducted in solutionmade up with triply distilled water. 100

~Adsorption and flocculation experiments

Five grams of the solid were equilibratedwith 195 cm3 of triply distilled water or solu-tion of appropriate ionic strength and pH in a250-cm3 glass beaker using a magnetic stirrerbar for about 2 h. The pH of the suspensionwas measured after equilibration and taken asthe final pH. The beaker was fitted withbaffle plates and the suspension stirred usinga propeller for 3 min. Five cubic centimetresof the polymer solution at the desired concen-tration was added dropwise while the suspen-sion was being stirred. As soon as all thepolymer was added, the stirring was stoppedand the suspension allowed to settle for 15 s.

v 80.E:70:!-0."eoat

~

PEO 5.106 MW/No-kOOIINte in WaterpH . 6.8Polymer Concentration' O.5j1pm

0 1 2 3 4 5 6 7 8 9 10Conditioning Time, min.

Fig. 1. Flocculation of sodium kaolinite as a functionof conditioning time.

255

stirring is stopped as soon as polymer additionis complete. Per cent solid settled droppedfrom 78% to 65% in 1 min and did not changesignificantly beyond that time period. It wasalso possible to visually observe the flocbreakage during conditioning after polymeraddition was complete. All flocculationresults reported in this work were obtainedat zero conditioning time after polymeraddition.

Polymer solutions and suspensions were pre-pared at the same ionic strength in theseexperiments. It is clear that maximum floc-culation occurs in triply distilled water bothin the presence and in the absence ofpolymer. Addition of salt led to reduction inflocculation and reached a constant value atabout 0.1 kmol/m3 NaCI. Figure 4 shows theeffect of pH on flocculation of kaolinite.Clearly, flocculation is enhanced at acidic pHconditions. The molecular weight of poly-ethylene oxide also had a significant role indetermining flocculation. Literature data haveshown the importance of molecular weight inflocculation of colloidal dispersions. In thisstudy, polyethylene glycol (MW 8000) andpolyethylene oxide (MW 5 million) were used.As shown in Fig. 5, the higher molecular

Effect of polymer concentrationFigure 2 shows per cent solid settled of

sodium kaolinite as a function of initialpolymer concentration. It is clear thatkaolinite is flocculated by polyethylene oxideeven at very low concentrations (1 ppm).Over the concentration range studied, norestabilization occurred. The effect of ionicstrength on flocculation is shown in Fig. 3.

100

~90 ~

~80

~ 10..'":! 60~~50

~~"~ "'"

""I ~~::::::::::~..PEO 5.10. MW INa-Kaolinlte In Wat...

Pafrmer Concentratian

. 00 0.5 ,pm

, I I I . I I I I I234557.9101112

pH

Fig. 4. Effect of pH on the flocculation of sodiumkaolinite.

-..

~ 80.!; 70go~080go

~50

40

IonicStre",th(MI pH (Noturol)0 -2 6.893.10 4.950.5 4.711 4.94

I I I I I I I . I

0 0;1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 toIonic Strength, kmole/m3 NoCI

Fig. 3. Effect of ionic strength on the flocculation ofsodium kaolinite.

256

*

I:~1 eo~# 10

40 PEO 5.10. _'&".'.a .. Wato,0" . ..8 (Nat.all

0 2 4 6 . 10 12 14 .. 16 20 22 251.."101 POI,.. COftCM'ro'i..., pp"

Fig. 9. Flocculation of alumina as a function ofpolymer concentration.

weight polyethylene oxide is a much betterflocculant than polyethylene glycol. This isprobably due to the bigger molecule beingmore amenable to the formation of tails andloops that are favorable for bridging floccula-tion to occur.

In order to identify the key solid proper-ties responsible for polyethylene oxideadsorption, flocculation of a series of oxideminerals of varying hydration and surfacecharge characteristics were studied. Figures6 - 9 show the flocculation of silica, quartz,hematite and alumina along with adsorptiondensities as a function of polymer concentra-tion. Figures 10 - 12 show the flocculationof these oxide minerals as a function of pH.Silica gel was strongly flocculated by poly-ethylene oxide at natural pH (around 6.9) butthe polymer had little effect on the settlingbehavior of alumina and hematite over a widepH range. Quartz in contrast to silica gel wasflocculated only at pH 2.5.

DISCUSSION

The kinetics of flocculation monitored bystirring the suspension after addition of thepolymer show the importance of hydro-dynamics in determining flocculation (Fig. 1).

257

100.~. ~~---~ ~80

'V.:70.U)~600.~50

PEO ,.10. _/Ollarf. in Water

PoIJmer Cancenrratlan-0- 0.5 ppm-- 040

30

2 3 4 5 6 7 8 9 10 11 12pH of Solution

Fig. 11. Effect of pH on the flocculation of quartz.

100 PEO 5.108 MW/Alumlna in Water

Polymer Concentration0 0A 25ppm.

~ 70

;: 60~:;50..

#40

~

30

, I I I I I I I I I '.1. 3 4 5 6 7 8 9 10 11 11.

pH of Solutio..

Fir. 12. Effect of pH on the flocculation of alumina.

The role of polymer concentration andionic strength (Figs. 2 and 3) indicates thatover the concentration range studied restabili-zation does not occur. Most interestingly,increasing ionic strength led to a decrease inflocculation. The general effect of adding anelectrolyte to a dispersion is to lower itsstability by compressing the electrical doublelayer. The double layer compression in con-junction with bridging by polymer normallyleads to an increase in flocculation, unlikewhat is seen in the clay/polyethylene oxidesystem in this study. This is possibly due tothe complex crystal and electrical doublelayer structure of clay. Depending on solutionpH, clay crystals expose two different sur-faces, a negatively charged layer surface anda positive or negatively charged edge surface.Consequently, due to oppositely chargeddouble layers, the edge of one clay particlecould be associated with the surface ofanother resulting in mutual coagulation ofclay [11 - 13]. In the absence of an indiffer-ent electrolyte, both double layers are suffi-ciently well developed for edge/face internalmutual coagulation to occur, causing rela-tively high solid per cent settled in thepresence and absence of polymer. With in-creasing electrolyte concentration, bothdouble layers are compressed and theireffective charge will be reduced, leading to areduction in edge/face flocculation.

The effect of pH on flocculation as shownin Fig. 4 is consistent with this explanation.The flocculation was enhanced at acidic pH,but in the presence and in the absence ofpolymer. In the absence of polymer, highsolid settled per cent in acidic pH can resultfrom internal mutual coagulation. At acidicpH, the flat surfaces still possess a negativecharge, whereas the edge surfaces carry apositive charge, leading to mutual coagula-tion. However, at alkaline pH, the positivecharge is reversed and electrostatic repulsionprevents coagulation. The effect of polymerin determining flocculation as a function ofpH suggests that conformational aspects maybe playing a role in enhancing flocculation atacidic pH. As polymer concentrations usedin the study were very low, it was not possibleto obtain quantitative information on adsorp-tion using the analytical technique of Attiaand Rubio [9]. Flocculation results, however,suggest that bridging is more effective at~

Prolonged agitation after complete additionof polymer resulted in floc breakage (asobserved visually) and thus led to a decreasein per cent solid settled. This is consistentwith the observations of Stanley and Scheiner[2] who reported that clay formed strongflocs when polyethylene oxide was used as aflocculant but with time, the flocs disinte-grated if the supernatant was not removed.They attribute the floc breakage to the weakinteraction between the polymer and thefilling water. The hydration water of the clayexists as two different types [10], the hydra-tion shell that is tightly bound and the fillingwater that is loosely bound to the claysurface. The initial interaction between theclay and the polymer is probably a waterbridge involving the filling water moleculeswhich is a weak interaction that is broken upupon rigorous stirring.

258

number of negative sites on silica gel are lessthan on quartz throughout the entire pHrange, permitting polyethylene oxide adsorp-tion on silica gel and flocculation under allpH conditions.

The necessity for neutral or positive siteson the solid surface cannot explain theinertness of polyethylene oxide for aluminaand hematite throughout the entire pH rangestudied. In addition to H-bonding, entropychanges also contribute to adsorption ofpolymers. The release of water molecules, theloss of water from the polymer and theentropy of dilution of bulk phase are impor-tant favorable factors for polymer adsorption.Of all the above factors, release of the watermolecules from the solid surface upon adsorp-tion is perhaps the most significant and isexpected to govern the overall effect ofentropy. The replacement of water moleculesduring the adsorption process is related to theconcentration of the hydroxyl groups on thesurface, since these species provide the sitesfor the initiation of multilayer adsorption ofwater molecules [14]. For example, apartially hydrated surface was found to bethe best for H-bonding with polymer, whereasa completely hydrated surface (depselycovered with OH groups) preferred water topolymer [1,15,16]. The concentration ofsurface hydroxyl groups on oxides has beendetermined by several investigators [14, 17].The Table shows the hydration of oxides interms of the number of OH groups per100 A 2. It is clear that alumina and hematiteare strongly hydrated in contrast to silica gel.The absence of adsorption of polyethyleneoxide on hematite and alumina is thereforeattributed to the inability of the polymerto displace the strongly H-bonded watermolecules associated with the surfaces.

acidic pH. One reason for this is the attractionbetween the negative charge on the etheroxygen of the polymer being able to attachto the positive sites of clay at acidic pH. Thus,at acidic pH, there are more sites for thepolymer to anchor on clay.

Solid per cent settled in the presence ofpolymer increased at all pH values, suggestingpolymer adsorption over the entire pH range.(Depletion flocculation at these polymerconcentrations is not expected in thissystem.)

The results on the adsorption and floccula-tion behavior of the oxides silica gel, quartz,hematite and alumina (Figs. 6 to 12) revealthe importance of the solid surface propertiesthat influence polyethylene oxide adsorptionon minerals. Silica gel was strongly floc-culated at its natural pH, whereas quartz wasflocculated only at pH 2.5. Most interestingly,polyethylene oxide did not adsorb or floc-culate hematite and alumina over the entirepH range.

Hydrogen bonding has often been proposedto be the driving force for the adsorption ofnonionic polymers on oxides. In the case ofpolyethylene oxide, the hydrogen bondoccurs between the ether oxygen of thepolymer and the surface hydroxyls. Atalkaline pH values, the negative MO- site isnot favorable for H-bonding. The behaviorof silica gel and quartz can be explained usingthe H-bonding mechanism. The zeta potentialmeasurement of silica gel and quartz is shownin Fig. 13. At any given pH level above 3,quartz is more negative than silica gel. Quartzbeing a stronger proton donor has morenegative sites, which are not favorable forH-bonding with the ether oxygen. At very lowpH, quartz is flocculated by polyethyleneoxide due to an increase in SiOH sites and adecrease in the negative sites (H3SiO4-). Thezeta potential results also suggest that the

TABLEConcentration of surface hydroxyl groups

Number of OH/IOO A2Oxide

Silica (amorphous)Titania (anatase)Titania (rutile)a-Hematitea-Aluminar-Alumina

4.2 .5.14.9 . 6.2

2.7 - 8.5

4.6 - 9.1

1510

Fig. 13. Zeta potential of quartz and silica gel.

259

REFERENCESCONCLUSIONSThe flocculation behavior of several oxides

- kaolinite, alumina, hematite, silica gel andquartz - with polyethylene oxide was investi-gated. The flocculation of clay was stronglyaffected by system variables such as condi-tioning time, ionic strength and pH. Increas-ing the conditioning time after polymeraddition led to floc breakage. Flocculationof kaolinite was enhanced in the absence ofindifferent electrolyte due to electrostaticattraction between the oppositely chargedclay surfaces, leading to mutual coagulation.

The adsorption of polyethylene oxide onoxide minerals was governed primarily by theH-bonding characteristics and the degree ofhydration of the substrate. Silica gel wasstrongly flocculated by polyethylene oxidewith the silanol sites serving as the principaladsorption sites. Quartz was not flocculatedby the polymer above pH 3 due to thepredominance of negative species on thesurface. Alumina and hematite were inert tothe polymer due to their strong hydrationand the polymer's inability to displace thebound water.

1 J. Rubio and J. A. Kitchener, J. Coli. Inter(.Sci., 57 (1976) 132.

2 D. A. Stanley.and B. J. Scheiner, Colloids andSurfaces, 14 (1985) 151.

3 J. Rubio, Colloids and Surfaces, 3 (1981) 79.4 Th. F. Tadros, in T. F. Tadros (ed.), The Effect of

Polymer on Dispersion Propertie" AcademicPress, London, 1982, pp. 1 - 38.

5 A. G. Smelley, B. J. Scheiner and J. R. Zatko,Bureau of Mines Investigations 8498, 1980, U.S.Department of the Interior.

6 F. E. Bailey and J. V. Koleske, Poiy(EthyleneOxide), Academic Press, New York, 1976.

7 POL VOX Water-8oluble Resins Are Unique,Technical Bulletin F-44029, Union CarbideCorporation.

8 A. F. Hollander, P. Somasundaran and C. C.Gryte, J. App. Poly. Sci., 26 (1981) 2123.

9 Y. A. Attia and J. Rubio, Br; Poly. J., 7 (1975)135.

10 R. Prost, in S. W. Bailey (ed.), Proc. Int. Cl4yConference, Mexico City, 1975, Applied Publish-ing, Wilmette, IL, p. 351.

11 R. K. Schofield and H. R. Samson, Disc. FaradaySoc., 18 (1954) 135.

12 A. Kahn, J. Coli. Inter!. Sci., 13 (1958) 51.13 H. yon alphen, J. Coli. Inter!. Sci., 19 (1964)

313.14 A. C. Zettlemoyer and E. McCafferty, Cooatica

Chemica Acta, 45 (1973) 173.15 O. Grist and J. A. Kitchener, Trans. Faraday Soc.,

61 (1965) 1026.16 T. F. Tadros, J. Coil. Inter!. Sci., 64 (1978) 36.17 Adsorption of Inorganics at Solid-Liquid Inter-

face" M. A. Anderson and A. J. Rubin (eds.),Ann Arbor Science Publishers, Ann Arbor,Michigan, 1981.

ACKNOWLEDGMENTThe authors wish to thank the National

Science Foundation (NSF-CPE-83-18163) forfinancial support of this work.