Flocculation of Oxides Using Polyethylene Oxide

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 hydrophobic interactions and chemical mechanisms.Tadros Flocculation of oxides using polyethylene [41 examined the adsorption and flocculation oxide was investigated. Polyethylene oxide properties of silica gel with polyvinyl alcohol adsorbedstrongly on and flocculated sodium and found that maximum flocculation kaolinite...H-bonding is considered to be the occurred at the point of zero charge of the driving force for polyethylene oxide adsorp- oxide and also observedan increasein adsorption on kaolinite. The mechanismof adsorption of PVA upon heat treating the solid up tion of polyethylene oxide on kaolinite was to 700 °c. Tadros examined the mechanism examined by studying the adsorption/ of adsorption basedon the degreeof hydra- flocculation characteristicsof different oxide tion of the substrate. Evidently, other than minerals using the same polymer. Poly- H-bonding, no generic mechanismcould be ethylene oxide flocculated silica gel but did utilized to explain polyethylene oxide adsorp- not adsorb on or flocculate hematite and tion. The U.S. Bureau of Mines has found that alumina. Quartz was flocculated only below polyethylene oxide is an excellent flocculant pH 3. Results are examined in terms of the for clays and red mud [5) and correlated the type of surface groups and the degree of adsorption of the polymer with the electro- hydration of the solids. It is proposed that negativity index. Polyethylene oxide itself adsorption of polyethylene oxide on solids has been a widely used polymer in several requires some optimum hydroxylation of the systemsand its characteristicshave been the surface with displaceablewater species. subject of severalreviews [6,7). The present work focuses on the adsorption of polyethylene oxide on clay and the constituent INTRODUCTION minerals of clay. It was the aim to study the adsorption/flocculation behavior of poly- Polyethylene oxide (PEO) has servedas an ethylene oxide on a variety of oxide minerals effective flocculant in severalsystems [1,2]. such as silica gel, quartz, hematite, alumina However, the exact nature of the mechanism and kaolinite and utilize the behavior of of polyethylene oxide adsorption remains polyethylene oxide on each mineral to unexplained. Rubio [3] found that poly- explain its overall flocculation characteristics. ethylene oxide was an effective flocculant The oxides studied vary significantly in their on substancesrecognized to be hydrophobic interfacial properties such as point of zero but the polymer was ineffective on hydro- charge(pzc) and degreeof hydration. philic substratessuch as rutile, quartz and copper carbonate. Rubio attributed the specificity of polyethylene oxide flocculation MATERIALS AND METHODS

.Present address: Union Carbide Corporation Na-kaolinite Limited, Technical Center, Bound Brook, New Jersey A homoionic sample of sodium kaolinite 08805. was preparedfrom a well-crystallized sample ..To whom correspondence should be addressed. of Georgiakaolinite which was obtained from

0032-59101901$3.50 @ Elsevier Sequoia/Printed in The Netherlands 254 the clay repository at the University of The top half of the suspensionwas sucked out Missouri. The surfacearea of the samplewas using a suction device. About 20 cm3 of the determined by BET to be 9.4 m2/g. The supernatant from the unsettled portion was methodology used for clay preparation has centrifuged for determination of residual beendiscussed earlier [8]. polymer concentration. Both 100-cm3settled and unsettled portions were filtered sepa- SUicagel rately using filter papers and the residue on Silica gel was purchasedfrom Alfa Products the filter paper was dried and weighed. The Inc. The surface area of silica gel was 300 per cent solid secttledwas determined from m2/g, particle size 70 p.m and pore volume the dry weight of the solid in the settled and 1.6 cm3/g. The characteristics of silica gel unsettled portions. were provided by the manufacturer. 88% = ms. ?'!'~~"'~ Quartz ms + mus Natural Brazilian quartz prepared by wet SS% = solid settled per cent grinding and leaching was used. The surface ms = weight of solid in 100 cm3 of settled areaof quartz as measuredby BET was 0.5 portion m2/g. mus = weight of solid in 100 cm3 of unsettled Hematite portion Synthetic hematite was obtained from Polymer concentration was determined Alfa Products Inc. The surface area of using the method of Attia and Rubio [9]. hematite as measuredby BET was 7.5 m2/g.

Alumina RESULTS Alumina of 0.3 /lm size was purchased from Alfa Products. Effect of conditioning time The role of conditioning time in determin- Polymer ing flocculation was investigatedinitially at a Polyox coagulant obtained from Union concentration of 0.5 ppm of polyethylene Carbide Corporati9n was used as a flocculant. oxide. The polymer was added dropwise for Polyethylene oxide used in the experiments 55 s at a steady flow rate of 0.1 cm3fs.Con- had an approximate molecular weight of ditioning time here refers to the time for 5 million as determined by the manufacturer. which the suspensionwas further stirred after Polyethylene glycol (PEG), molecular weight addition of the polymer. Figure 1 showsthat 8000, was also used in some experiments. maximum flocculation occurs when the All experiments were conducted in solution made up with triply distilled water. 100

~ Adsorption and flocculation experiments Five grams of the solid were equilibrated v. 80 E with 195 cm3 of triply distilled water or solu- :70 tion of appropriate ionic strength and pH in a :! 250-cm3glass beaker using a magnetic stirrer ."eo-0 at bar for about 2 h. The pH of the suspension ~ was measuredafter equilibration and taken as PEO 5.106 MW/No-kOOIINte in Water the final pH. The beaker was fitted with pH .6.8 baffle plates and the suspensionstirred using Polymer Concentration' O.5j1pm a propeller for 3 min. Five cubic centimetres of the polymer solution at the desired concentration was added dropwise while the suspen- 0 1 2 3 4 5 6 7 8 9 10 sion was being stirred. As soon as all the Conditioning Time, min. polymer was added, the stirring was stopped Fig. 1. Flocculation of sodium kaolinite as a function and the suspensionallowed to settle for 15 s. of conditioning time. 255 stirring is stopped as soon as polymer addition Polymer solutions and suspensionswere pre- is complete. Per cent solid settled dropped pared at the same ionic strength in these from 78% to 65% in 1 min and did not change experiments. It is clear that maximum floc- significantly beyond that time period. It was culation occurs in triply distilled water both also possible to visually observe the floc in the presence and in the absence of breakageduring conditioning after polymer polymer. Addition of salt led to reduction in addition was complete. All flocculation flocculation and reached a constant value at results reported in this work were obtained about 0.1 kmol/m3 NaCI. Figure 4 showsthe at zero conditioning time after polymer effect of pH on flocculation of kaolinite. addition. Clearly, flocculation is enhancedat acidic pH conditions. The molecular weight of poly- Effect of polymer concentration ethylene oxide also had a significant role in Figure 2 shows per cent solid settled of determining flocculation. Literature data have sodium kaolinite as a function of initial shown the importance of molecular weight in polymer concentration. It is clear that flocculation of colloidal dispersions. In this kaolinite is flocculated by polyethylene oxide study, polyethylene glycol (MW 8000) and even at very low concentrations (1 ppm). polyethylene oxide (MW 5 million) were used. Over the concentration range studied, no As shown in Fig. 5, the higher molecular restabilization occurred. The effect of ionic 100 strength on flocculation is shown in Fig. 3. ~ 90 ~ 80 ~ ~ 10 '".. ~"~ ~ :! 60 '" " ~ ""I ~~::::::::::~.. ~50 PEO 5.10. MWINa-Kaolinlte In Wat... Pafrmer Concentratian . 0 0 0.5,pm

, I I I . I I I I I 234557.9101112 pH Fig. 4. Effect of pH on the flocculation of sodium - kaolinite. ..

~ 80 .! ; 70 go ~ go080 ~ 50

40 IonicStre",th(MI pH (Noturol) 0 -2 6.89 3.10 4.95 0.5 4.71 1 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 to Ionic Strength, kmole/m3 NoCI Fig. 3. Effect of ionic strength on the flocculation of sodium kaolinite. 256

I:~ 1 eo~ # 10 40 PEO5.10. _'&".'.a .. Wato, 0" . ..8 (Nat.all

0 2 4 6 . 10 12 14 .. 16 20 22 25 1.."101 POI,.. COftCM'ro'i..., pp" Fig. 9. Flocculation of alumina as a function of polymer concentration. weight polyethylene oxide is a much better flocculant than polyethylene glycol. This is probably due to the bigger molecule being more amenableto the formation of tails and loops that are favorable for bridging flocculation to occur. In order to identify the key solid properties responsible for polyethylene oxide adsorption, flocculation of a seriesof oxide minerals of varying hydration and surface charge characteristicswere studied. Figures 6 - 9 show the flocculation of silica, quartz, hematite and alumina along with adsorption densitiesas a function of polymer concentration. Figures 10 - 12 show the flocculation of these oxide minerals as a function of pH. Silica gel was strongly flocculated by poly- DISCUSSION ethylene oxide at natural pH (around 6.9) but the polymer had little effect on the settling The kinetics of flocculation monitored by behavior of alumina and hematite over a wide stirring the suspensionafter addition of the pH range. Quartz in contrast to silica gel was polymer show the importance of hydro- flocculated only at pH 2.5. dynamics in determining flocculation (Fig. 1). 257

The role of polymer concentration and 100 ionic strength (Figs. 2 and 3) indicates that . over the concentration range studied restabilization does not occur. Most interestingly, 80 .~~~---~ ~ 'V. increasingionic strength led to a decreasein :70 U). flocculation. The general effect of adding an

~60 electrolyte to a dispersion is to lower its .0 PEO ,.10. _/Ollarf. in Water stability by compressingthe electrical double ~50 PoIJmer Cancenrratlan layer. The double layer compression in con- -0- 0.5 ppm 40 -- 0 junction with bridging by polymer normally leads to an increase in flocculation, unlike 30 what is seen in the clay/polyethylene oxide

2 3 4 5 6 7 8 9 10 11 12 system in this study. This is possibly due to pH of Solution the complex crystal and electrical double Fig. 11. Effect of pH on the flocculation of quartz. layer structure of clay. Depending on solution pH, clay crystals expose two different surfaces,a negatively charged layer surface and

100 PEO 5.108 MW/Alumlna in Water a positive or negatively charged edge surface. Consequently, due to oppositely charged Polymer Concentration 0 0 double layers, the edge of one clay particle . A 25ppm could be associated with the surface of another resulting in mutual coagulation of ~ 70 clay [11 13]. In the absenceof an indiffer- ; - : 60 ent electrolyte, both double layers are suffi- ~ ciently well developed for edge/faceinternal :;50 .. mutual coagulation to occur, causing rela- #40 tively high solid per cent settled in the 30 presenceand absenceof polymer. With in-

, I I I I I I I I I '. creasing electrolyte concentration, both 1. 3 4 5 6 7 8 9 10 11 11. double layers are compressed and their pH of Solutio.. effective chargewill be reduced, leading to a Fir. 12. Effect of pH on the flocculation of alumina. reduction in edge/faceflocculation. The effect of pH on flocculation as shown in Fig. 4 is consistent with this explanation. Prolonged agitation after complete addition The flocculation was enhancedat acidic pH, of polymer resulted in floc breakage (as but in the presenceand in the absenceof observedvisually) and thus led to a decrease polymer. In the absenceof polymer, high in per cent solid settled. This is consistent solid settled per cent in acidic pH can result with the observationsof Stanley and Scheiner from internal mutual coagulation. At acidic [2] who reported that clay formed strong pH, the flat surfaces still possessa negative flocs when polyethylene oxide was used as a charge, whereas the edge surfaces carry a flocculant but with time, the flocs disinte- positive charge, leading to mutual coagula- grated if the supernatant was not removed. tion. However, at alkaline pH, the positive They attribute the floc breakageto the weak chargeis reversedand electrostatic repulsion interaction between the polymer and the prevents coagulation. The effect of polymer filling water. The hydration water of the clay in determining flocculation as a function of exists as two different types [10], the hydra- pH suggeststhat conformational aspectsmay tion shell that is tightly bound and the filling be playing a role in enhancing flocculation at water that is loosely bound to the clay acidic pH. As polymer concentrations used surface. The initial interaction between the in the study were very low, it was not possible clay and the polymer is probably a water to obtain quantitative information on adsorp- bridge involving the filling water molecules tion using the analytical technique of Attia which is a weak interaction that is broken up and Rubio [9]. Flocculation results, however, upon rigorous stirring. suggestthat bridging is more effective at~

~ 258 acidic pH. One reasonfor this is the attraction number of negativesites on silica gel are less between the negative charge on the ether than on quartz throughout the entire pH oxygen of the polymer being able to attach range, permitting polyethylene oxide adsorp- to the positive sites of clay at acidic pH. Thus, tion on silica gel and flocculation under all at acidic pH, there are more sites for the pH conditions. polymer to anchor on clay. The necessity for neutral or positive sites Solid per cent settled in the presenceof on the solid surface cannot explain the polymer increasedat all pH values,suggesting inertness of polyethylene oxide for alumina polymer adsorption over the entire pH range. and hematite throughout the entire pH range (Depletion flocculation at these polymer studied. In addition to H-bonding, entropy concentrations is not expected in this changesalso contribute to adsorption of system.) polymers. The releaseof water molecules,the The results on the adsorption and floccula- loss of water from the polymer and the tion behavior of the oxides silica gel, quartz, entropy of dilution of bulk phaseare impor- hematite and alumina (Figs. 6 to 12) reveal tant favorable factors for polymer adsorption. the importance of the solid surface properties Of all the above factors, releaseof the water that influence polyethylene oxide adsorption moleculesfrom the solid surfaceupon adsorp- on minerals. Silica gel was strongly floc- tion is perhaps the most significant and is culated at its natural pH, whereasquartz was expected to govern the overall effect of flocculated only at pH 2.5. Most interestingly, entropy. The replacement of water molecules polyethylene oxide did not adsorb or floc- during the adsorption processis related to the culate hematite and alumina over the entire concentration of the hydroxyl groups on the pH range. surface, since these speciesprovide the sites Hydrogen bonding hasoften been proposed for the initiation of multilayer adsorption of to be the driving force for the adsorption of water molecules [14]. For example, a nonionic polymers on oxides. In the caseof partially hydrated surface was found to be polyethylene oxide, the hydrogen bond the best for H-bonding with polymer, whereas occurs between the ether oxygen of the a completely hydrated surface (depsely polymer and the surface hydroxyls. At coveredwith OH groups) preferred water to alkaline pH values, the negative MO- site is polymer [1,15,16]. The concentration of not favorable for H-bonding. The behavior surface hydroxyl groups on oxides has been of silica gel and quartz can be explained using determined by severalinvestigators [14, 17]. the H-bonding mechanism.The zeta potential The Table shows the hydration of oxides in measurementof silica gel and quartz is shown terms of the number of OH groups per in Fig. 13. At any given pH level above 3, 100 A 2. It is clear that alumina and hematite quartz is more negativethan silica gel. Quartz are strongly hydrated in contrast to silica gel. being a stronger proton donor has more The absenceof adsorption of polyethylene negative sites, which are not favorable for oxide on hematite and alumina is therefore H-bonding with the ether oxygen. At very low attributed to the inability of the polymer pH, quartz is flocculated by polyethylene to displace the strongly H-bonded water oxide due to an increasein SiOH sites and a moleculesassociated with the surfaces. decreasein the negativesites (H3SiO4-). The zeta potential results also suggestthat the TABLE Concentration of surface hydroxyl groups

Oxide Number of OH/IOO A2

Silica (amorphous) 4.2 .5.1 Titania (anatase) 4.9 . 6.2 Titania (rutile) 2.7 - 8.5 a-Hematite 4.6 - 9.1 a-Alumina 15 r-Alumina 10 Fig. 13. Zeta potential of quartz and silica gel. 259

CONCLUSIONS REFERENCES The flocculation behavior of severaloxides 1 J. Rubio and J. A. Kitchener, J. Coli. Inter(. - kaolinite, alumina, hematite, silica gel and Sci., 57 (1976) 132. quartz- with polyethylene oxide was investi- 2 D. A. Stanley.and B. J. Scheiner, Colloids and gated. The flocculation of clay was strongly Surfaces, 14 (1985) 151. affected by system variables such as condi- 3 J. Rubio, Colloids and Surfaces, 3 (1981) 79. 4 Th. F. Tadros, in T. F. Tadros (ed.), The Effect of tioning time, ionic strength and pH. Increas- Polymer on Dispersion Propertie" Academic ing the conditioning time after polymer Press,London, 1982, pp. 1 - 38. addition led to floc breakage.Flocculation 5 A. G. Smelley, B. J. Scheiner and J. R. Zatko, of kaolinite was enhancedin the absenceof Bureau of Mines Investigations 8498, 1980, U.S. Department of the Interior. indifferent electrolyte due to electrostatic 6 F. E. Bailey and J. V. Koleske, Poiy(Ethylene attraction between the oppositely charged Oxide), Academic Press,New York, 1976. clay surfaces,leading to mutual coagulation. 7 POL VOX Water-8oluble Resins Are Unique, The adsorption of polyethylene oxide on Technical Bulletin F-44029, Union Carbide oxide minerals was governed primarily by the Corporation. 8 A. F. Hollander, P. Somasundaran and C. C. H-bonding characteristics and the degree of Gryte, J. App. Poly. Sci., 26 (1981) 2123. hydration of the substrate. Silica gel was 9 Y. A. Attia and J. Rubio, Br; Poly. J., 7 (1975) strongly flocculated by polyethylene oxide 135. with the silanol sites servingas the principal 10 R. Prost, in S. W. Bailey (ed.), Proc. Int. Cl4y adsorption sites. Quartz was not flocculated Conference, Mexico City, 1975, Applied Publish- ing, Wilmette, IL, p. 351. by the polymer above pH 3 due to the 11 R. K. Schofield and H. R. Samson, Disc. Faraday predominance of negative specieson the Soc., 18 (1954) 135. surface. Alumina and hematite were inert to 12 A. Kahn, J. Coli. Inter!. Sci., 13 (1958) 51. the polymer due to their strong hydration 13 H. yon alphen, J. Coli. Inter!. Sci., 19 (1964) and the polymer's inability to displace the 313. 14 A. C. Zettlemoyer and E. McCafferty, Cooatica bound water. Chemica Acta, 45 (1973) 173. 15 O. Grist and J. A. Kitchener, Trans. Faraday Soc., 61 (1965) 1026. ACKNOWLEDGMENT 16 T. F. Tadros, J. Coil. Inter!. Sci., 64 (1978) 36. The authors wish to thank the National 17 Adsorption of Inorganics at Solid-Liquid Inter- face" M. A. Anderson and A. J. Rubin (eds.), ScienceFoundation (NSF-CPE-83-18163)for Ann Arbor Science Publishers, Ann Arbor, financial support of this work. Michigan, 1981.