J. Chem. SOC., Faraday Trans. 1, 1984, 80, 1491-1498 Adsorption of Goethite onto Quartz and Kaolinite B Y MARVIN c . GOLDBERG,* EUGENE R. WEINER~ AND PAUL M. BOYMELT U.S. Geological Survey, P.O. Box 25046 Mail Stop 424, Denver Federal Center, Lakewood, Colorado 80225, U.S.A. Received 27th July, 1983 The adsorption of colloidal goethite onto quartz and kaolinite substrates has been studied as a function of pH and NaCl concentration. Goethite adsorption was measured quantitatively by Fourier-transform infrared spectroscopy. The results indicate that adsorption onto both substrates is due primarily to coulombic forces; however, the pH dependence of adsorption is very different for the two substrates. This is explained by the fact that the surface charge on quartz is entirely pH-dependent, while kaolinite has surface faces which carry a permanent negative charge.Adsorption of goethite on to kaolinite increases markedly with increasing NaCl concentration, while adsorption onto quartz is relatively independent of NaCl concentration. This can be explained by the influence of NaCl concentration upon the development of surface charge on the substrates. A method is described for separating surface-bound goethite from free goethite. Sediments in natural waters can serve as sources or sinks of water-borne chemicals by virtue of adsorption and desorption processes occurring at their surfaces.l? The nature of sediment surfaces may change over time as the sediments acquire adsorbed organic and inorganic materials through interactions with dissolved and colloidal chemical species.The adsorbed species themselves then become important factors influencing further sorption processes and hence the chemical composition of the waters in which they reside. Among the most common sorbed species are the Fe1I1 hydrous oxides. Jenne2 has proposed that sorption to the hydrous oxides of iron and manganese is the principal control of heavy-metal mobility in soils and fresh waters. Crystalline iron(m) oxides such as goethite (a-FeOOH), lepidocrocite v-FeOOH) and haematite (Fe20,) are common constituents of the sediments, occurring as free particles and as sorbed layers on other particles such as clays and quartz.2v3 A gen- erally accepted mechanism for forming sorbed layers of crystalline Fe1I1 is one where ferric ion species [Fe3+, FeOH2+ and Fe(OH)t] and/or amorphous Fe(OH), are adsorbed onto the surface of a sedimentary mineral and then become converted into crystalline ferric layers over a period of time.Follett4 and F ~ r d h a m ~ - ~ have studied this process in detail for substrates of kaolinite, quartz and gibbsite. Follett found that amorphous Fe(OH), in colloidal suspension adsorbs specifically to all three substrates and that its adherence is unaffected by pH changes, prolonged washing, ultrasonic vibration or exposure to high concentrations of NHZ, Ca2+ or A13+. Fordham ex- amined the adsorption of ferric ion species from dilute acidic solutions onto kaolinite. He followed the change of ferric ion species adsorbed on the surface into amorphous Fe(OH), and its subsequent conversion into crystalline lepidocrocite.Although he observed no goethite or haematite after 15 weeks of ageing, these minerals, which are more stable thermodynamically than 1epidocrocitelO and are abundant in the natural environment, presumably could be formed eventually under appropriate conditions. t Permanent address: Department of Chemistry, University of Denver, Denver, Colorado 80208, U.S.A. $ Present address: Carborundum Co., Niagara Falls, New York 14302, U S A . 14911492 ADSORPTION OF GOETHITE ONTO QUARTZ AND KAOLINITE There seem to be no previously published reports concerning the direct adsorption of colloidal crystalline goethite onto sediments as a first step in developing crystalline coatings. The purpose of our study was to determine whether goethite coatings on sediments required precursors of ferric ion species or amorphous Fe(OH),, or whether they might also form by direct sorption of free crystalline goethite particles.The substrates studied were kaolinite and quartz. EXPERIMENTAL Goethite was prepared by the method of Atkinson et aZ.ll using an OH/Fe molar ratio of unity. Quartz, produced by the Ottawa Silica Co. as Flint Shot, was washed with 0.1 mol dm-3 HCl followed by 3 or 4 distilled-water washes to remove adsorbed ions. A well crystallized kaolinite was used as received from the Georgia Kaolinite Co. The point of zero charge (P.z.c.) for goethite was determined by the pH-drift method of Berube and de Bruyn12 to be at pH 7.2. Points of zero charge for quartz and kaolinite were taken from the literature.The P.Z.C. for quartz appears to be in the region of pH 2-313* l4 and that of kaolinite is reported to be somewhere between pH 4.5 and 8.2, depending on the ionic strength and method of meas~rement.~~’ l6 The uncertainties in P.Z.C. values for quartz and kaolinite are discussed further in the Results and Discussion section. A typical experiment consisted of stirring 0.1 g of kaolinite or quartz (particle-size range between 10 and 20 pm) with a distilled-water suspension of goethite (0.2 g per 250 cm3, particle- size range < 1 pm). The pH was adjusted with either 0.01 mol dm-3 HCI or 0.01 mol dm-3 NaOH. NaCl was added to adjust the ionic strength. After stirring for 19 h the final pH was recorded and the solids allowed to settle. During stirring, the pH change was always toward neutrality and the variation was < 1 pH unit.The final pH after stirring was taken as the pH at which sorption occurred. Sorption of goethite onto the substrate produced larger composite particles which were visually distinguishable from the untreated substrate. Electron micrographs of goethite-sorbed particles showed goethite platelets adhering to substrate particles. Goethite-sorbed substrate was separated from free goethite by exploiting their different settling-out times after stirring ceased. Goethite was formed as a distilled-water suspension with an average particle size < 1 pm. A portion of the suspension was tested for the presence of amorphous ferric hydroxide by treatment with acid ammonium oxalate.lS The amount of amorphous hydroxide was < 1 % .Quartz and kaolinite were sieved to obtain a particle-size distribution between 10 and 20 pm. Free goethite remained in suspension after 2 days, whereas the quartz and kaolinite particles settled out in CQ. 30 min (fall distance 10 cm). When a run ended and stirring stopped, the larger goethite-sorbed substrate particles settled out in ca. 15 min. The unsorbed goethite remaining in suspension was removed by siphoning off the suspension. In experiments where varying the solution ionic strength might have caused flocculation of goethite, the salt concentrations were adjusted in the goethite suspensions before adding the substrate material. These suspensions were stirred for 19 h to complete the flocculation process and all the flocculated goethite was allowed to settle out.Further flocculation did not occur. The goethte remaining in suspension was siphoned off and used in subsequent experiments. At the end of the sorption experiment the separated goethite-sorbed substrate particles were repeatedly washed with distilled water adjusted to an ionic strength matching that of the experiment, in order to remove any non-bound goethite. Goethite adsorption to kaolinite occurred at all pH values between 2 and 7. Wash solutions between pH 2 and 10 had no effect on the adherence of goethite to kaolinite, regardless of the pH at which the adsorption occurred. Goethite adsorption onto quartz was detected only in suspensions between pH 5 and 8.5. Wash solutions with pH values outside this range removed goethite from the substrate.After washing, the goethite-sorbed substrate particles were air-dried and examined by infrared spectroscopy to determine the quantity of goethite adsorbed. Infrared spectra were obtained from 1.0 and 0.1 % KBr pellets on a Digilab FTS-20B IR Fourier-transform spectrometer, interfaced with a Nova 2 computer. Computer analysis of the recorded spectra produced relative measurements of the amount of goethite adsorbed onto quartz and kaoliniteM. C. GOLDBERG, E. R. WINER AND P. M. BOYMEL 1493 I I I I I I I I 0 3600 3200 2800 2400 2000 I600 1200 800 wavenum berlcm -' 0 Fig. 1. Infrared spectra from the goethite-kaolinite system: (a) standard spectrum of pure goethite, from one of a series of standardized samples, (b) pure kaolinite, (c) goethite-sorbed kaolinite and ( d ) spectrum of surface-bound goethite resulting from subtraction of (b) from (c).under different conditions of pH and salt concentration. Because the particle-size distribution can influence the i.r. spectral pattern," care was taken to ensure that all calibration and experimental samples had the same particle-size distribution. Because the quartz spectrum was especially sensitive to particle-size distribution, we restricted the quartz particle-size range to < 20pm. Fig. 1 and 2 illustrate the spectral subtraction procedure used to determine the amount of goethite adsorbed to the substrate. Spectra were obtained from pure substrate, goethite-sorbed substrate and standardized samples of pure goethite. The pure-substrate spectrum was computer-subtracted from the spectrum of goethite-sorbed substrate, leaving a sample goethite spectrum.A standard-goethite spectrum was then compared by subtraction with the sample- goethite spectrum to obtain a quantitative measure of the amount of goethite absorbed. The characteristics of the pure-goethite spectrum [fig. l(a)], a broad band centred at 3 160 cm-l [v(OH) from hydroxy groups hydrogen-bonded to adsorbed water molecule^^^] and1494 ADSORPTION OF GOETHITE ONTO QUARTZ AND KAOLINITE 4000 3600 3200 2800 2400 2000 1600 1200 800 400 Fig. 2. Infrared spectra from the goethite-quartz system : (a) pure quartz, (b) goethite-sorbed quartz, ( c ) spectrum of surface-bound goethite resulting from subtraction of (a) from (b). Spectrum (c) is compared by subtraction with a pure-goethite standard spectrum, such as fig.1 (a), to obtain a quantitative measure of surface-bound goethite. wavenum berlcm-' sharp peaks at 900, 800 and 640 cm-I, make goethite easy to recognize in the presence of other spectral features. The absence of the v(0H) bands at 3660 and 3486cmp1, which are characteristic of dry goethite,'*. l9 indicates that the goethite in the measured samples was covered with adsorbed water. With a kaolinite substrate the sharp kaolinite peaks between 3600 and 3750 cm-' [fig. 1 (b)] can serve as a convenient internal reference for quantitative subtraction of kaolinite from goethite-sorbed kaolinite [fig. 1 ( c ) ] . After subtracting kaolinite, the height of the remaining goethite band at 3160 cm-' [fig l(d)], which is relatively free from overlap by kaolinite spectral features but is close in frequency to the kaolinite reference peaks, was compared with a standard-goethite spectra to measure the amount of goethite adsorbed.Reproducibility was satisfactory only if the goethite band closest to the kaolinite reference peaks was used for the measurement, possibly because the influence of particle-size distribution upon band shape, mentioned earlier, is wavelength-dependent. The subtraction procedure is more difficult with a quartz substrate. The broad quartz absorption band around 3460 cm-l overlaps the goethite band at 3 160 cm-l, and structure in the quartz spectrum below 900 cm-I interferes with the goethite bands in that region (see fig. 2). The most satisfactory spectral subtraction procedure proved to be one where the quartz band at 3460 cm-l was subtracted to leave no trace on the side of the goethite band at 3 160 cm-l.The height of the remaining goethite band at 3160 cm-l [fig. 2(c)] was then compared with standard-goethite spectra, as in the goethite-kaolinite system.I 50 60 70 80 1495 Fig. 3. Adsorption isotherm of colloidal goethite onto quartz. There is no measurable dependence on NaCl concentration. RESULTS AND DISCUSSION GOETHITE-QUARTZ The adsorption isotherm of goethite particles onto quartz is shown in fig. 3. Adsorption was observed only between pH 5 and 8.5, with a maximum at pH 6.7. Goethite adsorption onto quartz is expected at pH values between their respective P.Z.C. values of pH 7.2 (goethite) and 2 or 3 (quartz), where goethite and quartz have surface charges of opposite sign. However, the negative surface charge density on quartz develops slowly at first, making the P.Z.C.difficult to determine a~curate1y.l~ It remains quite small (< 10 pC cm-2) until ca. pH 5 or 6, above which the charge development is ac~e1erated.l~ Our spectroscopic measurement technique detected no adsorbed goethite until this pH region was reached. Increasing the ionic strength by adding NaCl had no detectable influence at any pH value. Although increasing the NaCl concentration is known to increase the quartz surface charge density, the magnitude of the increase is sma1114 and any influence on goethite adsorption was below our limit of sensitivity. Washing goethite-coated quartz removed goethite if the wash solution pH was below 5 or above 8.5.The behaviour of the goethite-quartz system is consistent with a non-specific adsorption mechanism and indicates that purely electrostatic attractions between pH-dependent charge sites are responsible for goethite adsorption onto quartz. GOETHITE-KAOLINITE Adsorption isotherms of goethite onto kaolinite are shown in fig. 4. The dependence upon NaCl concentration and pH is very different from the goethite-quartz system and suggests a different surface-adsorption mechanism. For pure kaolinite the1496 0 l 0 - n M k x 008- E 0 .- W .- u L Y N 0.06 5 m \ u .4 3 004- 0 M 002 ADSORPTION OF GOETHITE ONTO QUARTZ AND KAOLINITE - - 0 12 nQ 2.0 3.0 4.0 5.0 ' 60 7.0 PH Fig. 4. Adsorption isotherm of colloidal goethite onto kaolinite, for several NaCl concentrations: 0, 0.001; A, 0.01 and 0, 0.1 mol dm-3.dependence of surface charge upon pH is more complicated than for quartz. Kaolinite plates are believed to carry a permanent small negative charge on their basal planes, owing to isomorphous substitution, and a pH-dependent charge on the plate edges that has a P.Z.C. between pH 4.5 and 8.2, depending on the ionic strength and the measuring technique.14*15 The edge charge is positive at pH values below the P.Z.C. owing to adsorption of hydronium ions to the edges, and is negative above the P.z.c., owing to adsorption of hydroxyl ions. The overall negative charge density is small at low pH, owing to occupation of negative charge sites by H+, but it increases rapidly as the pH is increased14 and appears to go through a maximum around pH 10.5.15 The adsorption of cations, such as Fe3+, FeOH2+ and Fe(OH)i, onto kaolinite will occur only on the basal faces in very acidic solutions but will extend to the plate edges at pH values above the P.Z.C.of the kaolinite edge charge. Colloidal particles such as amorphous Fe(OH), will also be absorbed on the negative kaolinite basal faces at pH values below the P.Z.C. for the colloid, where the colloid particles carry a net positive charge. In the pH region above the P.Z.C. values for both the colloid and kaolinite, adsorption of colloid is negligible because both particles carry net negative charges. The surface charge on goethite is positive at pH values below its P.Z.C.of 7.2, and adsorption onto kaolinite is expected in this region. We were unable to obtain consistent measurements above pH 7. The data suggested that coagulation of the goethite colloid might have been the reason, causing larger particles of pure goethite to settle and be collected along with goethite-sorbed substrate particles. Fig. 4 shows that adsorption onto kaolinite occurs at pH 2 and increases with pH for low NaCl concentrations. It seems likely that the concentration of positive geothite particles is greatest adjacent to the negative basal planes of kaolinite and small nearM. C. GOLDBERG, E. R. WEINER AND P. M. BOYMEL 1497 the positively charged plate edges, resulting in preferential adsorption onto the basal planes.* As the pH increases, the net negative charge density also increases as the amount of surface-bound H+ decreases and more goethite is adsorbed.However, the fact that goethite, which was adsorbed at higher pH values, is not removed when the pH is lowered suggests a bonding mechanism more complex than in the goethite-quartz system. We observed no changes in the goethite spectrum that might be attributed to the adsorption process, although small changes in the broad goethite band would be hard to detect. Nor did we find any new bands from goethite-surface bonds. Although these results are inconclusive, they at least give no support to a specific bonding mechanism. Follett4 reported similar adsorption behaviour for the adsorption of amorphous Fe(OH), onto kaolinite. He proposed an adsorption mechanism whereby positively charged colloid particles are attracted to the permanently negative surface of kaolinite, increasing the net positive charge on the kaolinite.Changes in pH can increase or decrease the additional charge density carried by the colloid, but have no influence on the portion of the colloid neutralized by being in contact with permanent negative sites on the kaolinite surface. Thus, pH changes do not release colloid particles back into suspension. Our observations of goethite adsorption onto kaolinite in acidic suspensions are consistent with Follett’s mechanism. Fig. 4 also shows that adsorption of goethite increases at any pH < 7 if the NaCl concentration is increased, with saturation occurring around 0.10 mol dmP3 NaC1. This behaviour is expected because the negative surface charge density of kaolinite increases rapidly with increasing NaCl concentration,14 and this should increase goethite adsorption.Another possible mechanism, which also may play a role, is easier adsorption of goethite onto Na+-exchanged kaolinite surfaces than onto H+-exchanged surfaces. Na+ is an easily exchangeable cation, while H+ is much more difficult to displace.20 At low pH and low Na+ concentrations, the kaolinite negative surface sites are occupied by tightly bound H+. Increasing the pH or increasing the Na+ concentration allows Na+ to compete more effectively for surface positions. Eventually, at high enough pH and/or Na+ concentration, most of the H+ is displaced from the negative surface sites. The fact that the extent of goethite adsorption follows the displacement of H+ by Na+ could indicate that goethite absorbs more readily onto a kaolinite surface exchanged with Na+ than onto one exchanged with H+.A similar increased affinity for adsorption of Cu2+ onto Na+-exchanged kaolinite surfaces has been observed by McBride.21 CONCLUSIONS Colloidal goethite can adsorb onto the surface of both quartz and kaolinite. The pH dependence of adsorption onto quartz indicates a purely coulombic mechanism governed by the pH dependence of the goethite and quartz surface charges. The pH dependence of goethite adsorption onto kaolinite indicates a different binding mechanism. A possible mechanism is based on the fact that in acid solutions the kaolinite has permanent negative sites. Adsorption of positively charged goethite onto these sites results in a binding force, owing to effective charge neutralization in the region where the particles are in contact, which is not pH-dependent.The influence of NaCl concentration on goethite adsorption was found to follow closely its influence on the substrate surface charge concentration, both adsorption and charge concen- tration increasing with NaCl concentration. The data also allow the possibility that goethite displaces Na+ more readily than H+ from a kaolinite surface.1498 ADSORPTION OF GOETHITE ONTO QUARTZ AND KAOLINITE J. D. Hem, U.S. Geological Survey Water Supply Paper 1473 (U.S. Government Printing Office, Washington, D.C., 1970). E. A. Jenne, in Molybdenum in the Environment, ed. W. R. Chappell and K. K. Peterson (Marcel Dekker, New York, 1977), vol. 2, pp. 425-553. U. Schwertmann and R. M. Taylor, in Minerals in Soil Environments, ed. J. B. Dixon and S. B. Weed (Soil Science Society of America, Madison, Wisconsin, 1977), pp. 145-176. E. A. C. Follett, J. Soil Sci., 1965, 16, 333. A. W. Fordham, J. Soil Res., 1969, 7, 185. A. W. Fordham, Aust. J. Soil Res., 1969, 7, 199. A. W. Fordham, Aust. J. Soil Res., 1970, 8, 107. A. W. Fordham, Aust. J. Soil Res., 1973, 11, 185. A. W. Fordham, Ausr. J. Soil Res., 1973, 11, 197. lo W. L. Lindsay, Chemical Equilibria in Soils (Wiley, New York, 1979), p. 133. l 1 R. J. Atkinson, A. M. Posner and J. P. Quirk, J. Inorg. Nucl. Chem., 1968, 30, 2371. l2 Y. G. Berube and P. L. DeBruyn, J. Colloid Interface Sci., 1968, 27, 314. l3 R. 0. James and T. W. Healy, J. Colloid Interface Sci., 1972, 40, 65. l4 A. C. Riese, Ph.D. Thesis (Colorado School of Mines, Golden, Colorado). lE, J. B. Dixon, in Minerals in Soil Environments, ed. J. B. Dixon and S. B. Weed (Soil Science Society of America; Madison, Wisconsin, 1977), pp. 385-388. E. R. Landa and R. G. Gast, Clays Clay Miner., 1973, 21, 121. l7 J. Hlavay and J. Inczedy, Acta Chem. Acad. Sci. Hung., 1979, 102, 11. IR J. D. Russell, R. L. Parfitt, A. R. Fraser and V. C. Farmer, Nature (London), 1974, 248, 220. 2o R. N. Yong and B. P. Warkentin, Soil Properties and Behaviour (Elsevier, New York, 1975). '' M. B. McBride, Clays Clay Miner., 1978, 26, 101. R. L. Parfitt, J. D. Russell, and V. C. Farmer, J. Chem. Soc., Faraday Trans. I , 1976, 72, 1082. (PAPER 3/1306)