Adsorption of Humic Acid to Mineral Particles. 1. Specific and

Apr 11, 1998 - Within natural soil systems a distinction can be made between a mobile and an immobile fraction. The mobile fraction consists of ..... ...
0 downloads 6 Views 116KB Size
2810

Langmuir 1998, 14, 2810-2819

Adsorption of Humic Acid to Mineral Particles. 1. Specific and Electrostatic Interactions A. W. P. Vermeer,† W. H. van Riemsdijk,‡ and L. K. Koopal*,† Department of Physical Chemistry & Colloid Science, Wageningen Agricultural University, Wageningen, PO Box 8038, 6700 EK Wageningen, The Netherlands, and Department of Soil Science & Plant Nutrition, Wageningen Agricultural University, Wageningen, PO Box 8005, 6700 EC Wageningen, The Netherlands Received June 12, 1997. In Final Form: September 15, 1997 The adsorption of humic acid to mineral particles can be characterized by specific and electrostatic interactions and by polydispersity effects. In this paper we focus on the adsorbed amount and discuss the importance of specific and electrostatic interactions. The adsorption of purified Aldrich humic acid onto hematite has been measured as a function of pH and salt concentration as well as in the presence of two cadmium concentrations. Besides the adsorbed amount, the thickness of the adsorbed layer has also been studied. The experimental results are discussed in relation to the polyelectrolyte behavior of the humic acid. At high pH and low salt concentration the adsorption is low and the humic acid molecules are adsorbed relatively flat on the surface. At low pH and high salt concentration a substantial fraction of the adsorbed humic acid is not in direct contact with the surface, which results in a relatively high adsorbed amount and a large layer thickness. PAHA adsorption at not too low pH leads to an overcompensation of the hematite charge and to the development of an electrostatic barrier against further adsorption. Substantial PAHA adsorption occurs at the point of zero charge of hematite; this shows that besides electrostatic interactions also specific interactions are important. Calculations based on the SCF theory for polyelectrolyte adsorption show that the trends with respect to the adsorbed amounts and layer thickness as a function of pH and salt concentration correspond well with the measured trends for PAHA. This suggest that at least partly the humic acid adsorption behavior is related to its polyelectrolyte character. Both experiments and theory indicate that a well-developed layer arises upon adsorption and that lateral interactions within this layer determine to a large extent the adsorption process.

Introduction Within natural soil systems a distinction can be made between a mobile and an immobile fraction. The mobile fraction consists of components that are soluble and transportable by the soil solution. The settled particles and the material bound to these particles belong to the immobile fraction. It is generally accepted that contaminants such as metal ions form complexes with the functional groups of the humic substances in the solution1-4 and that the adsorbed humic material plays an important role with contaminant binding to mineral particles.5-7 The adsorption of metal ions to both natural organic matter and mineral oxides will be altered when the humics are adsorbed on the minerals.8 To be able to understand the behavior of binding of metal ions to these complexes, the mechanisms of binding of humic acids to mineral particles have to be known. This paper concentrates on gaining insight into these mechanisms. * To whom correspondence should be addressed. Phone: (31) 317 482629. Fax: (31) 317 483777. E-mail: [email protected]. † Department of Physical Chemistry and Colloid Science. ‡ Department of Soil Science & Plant Nutrition. (1) Buffle, J. In Metal ions in biological systems; Sigel, H., Ed.; Marcel Dekker: New York, 1984; Vol. 18, p 165. (2) Ephraim, J. H.; Marinsky, J. A.; Cramer, S. J. Talanta 1989, 36, 437. (3) Sposito, G. CRC Crit. Rev. Environ. Control 1986, 16, 193. (4) Kinniburgh, D. G.; Van Riemsdijk, W. H.; Koopal, L. K.; Benedetti, M. F. In Adsorption of metals by geomedia: variables, mechanisms, and model applications; Jenne, E., Ed.; Academic Press: New York, 1997. (5) Tipping, E.; Griffith, J. R.; Hilton, J. Croat. Chem. Acta 1983, 56, 613. (6) Murphy, E. M.; Zachara, J. M.; Smith, S. C.; Philips, J. L. Sci. Total Environ. 1992, 117/118, 413. (7) Zhou, J. L.; Rowland, S. J.; Braven, J.; Mantoura, R. F. C.; Harland, B. J. Int. J. Environ. Anal. Chem. 1995, 58, 275. (8) Amirbahman, A.; Olson, T. M. Environ. Sci. Technol. 1993, 27, 2807.

For the adsorption of humic acid on positively charged minerals an increase in adsorption is observed with decreasing pH and increasing salt concentration.9-12 In general the adsorption is due to both specific and Coulombic interactions.6,11-13 Parfitt et al.14 gave spectroscopic evidence for specific interactions via a “ligand exchange” between a surface hydroxyl or bound water molecule and the oxygen of the carboxylic group of the humic acid. Tipping12 and Murphy et al.6 reported humic acid adsorption isotherms at a given pH and salt concentration, described the adsorption by a ligand exchange mechanism, and modeled the adsorption using the Langmuir equation. Differences in adsorption were ascribed to differences in the composition of the humic substances, and it was found that humic acid adsorption is higher than fulvic acid adsorption. Gu et al.13 studied the reversibility of the adsorption of natural organic matter (NOM) to iron oxide particles and reported a strong hysteresis between adsorption and desorption. A ligand-exchange mechanism was proposed, and the isotherms were described using a modified Langmuir equation in which a hysteresis parameter was incorporated. Summers and Roberts9 have stressed the importance of the electrostatic interactions and the role of both pH and the indifferent electrolyte concentration on humic acid adsorption. (9) Summers, R. S.; Roberts, P. V. J. Colloid Interface Sci. 1988, 122, 367. (10) Zhou, J. L.; Rowland, S. J.; Fauzi, R.; Mantoura, R. F. C.; Braven, J. Water Res. 1994, 28, 571. (11) Davis, J. A. Geochim. Cosmochim. Acta 1982, 46, 2381. (12) Tipping, E. Geochim. Cosmochim. Acta 1981, 45, 191. (13) Gu, B.; Schmitt, J.; Chen, Z.; Liang, L.; McCarthy, J. F. Environ. Sci. Technol. 1994, 28, 38. (14) Parfitt, R. L.; Fraser, A. R.; Farmer, V. C. J. Soil Sci. 1977, 28, 289.

S0743-7463(97)00624-0 CCC: $15.00 © 1998 American Chemical Society Published on Web 04/11/1998

Adsorption of Humic Acid to Mineral Particles

Divalent cations have a relatively strong effect on the adsorption of humic acids onto mineral particles.12,15,16 Tipping12 reported slightly higher values for the adsorption in the presence of Ca2+ and Mg2+ and postulated that the cations and the positive surface groups compete for the anionic groups in the humic acid molecules. The competition causes fewer contacts between the humic acid molecules and the surface, and this may lead to an increased adsorption. Furthermore, due to the formation of metal ion-humic acid complexes the lateral electrostatic repulsion between the adsorbed humic acid molecules is decreased, and this also may contribute to an increase in adsorption. Engebertson and Wandruszka16 also found an increase in humic acid adsorption in the presence of divalent metal ions. According to these authors the increase is due to rearrangements within the humic acid that lead to a more compressed structure of the complex. In most of these studies the macromolecular character of the humic acid is largely neglected. This is certainly the case when Langmuir type models are used to describe the behavior.6,12,13 However, Summers and Roberts9 have taken an opposite view and considered humic substances as a polydisperse mixture. They studied the effects of the polydispersity of the humic substances on the adsorption isotherms and reported that previously developed concepts for well-defined synthetic polymers17-20 were also applicable to humic acid adsorption. The macromolecular approach of humic substances is in agreement with those of several others21-24 that describe the behavior of humic acids in solution as flexible polyelectrolytes. Chen and Schnitzer22 mentioned that fulvic and humic acids behave like flexible polyelectrolytes, not exclusively composed of condensed rings but as chains with numerous linkages about which relatively free rotation occurs. Ghosh and Schnitzer23 showed that, under the conditions that normally prevail in natural soils, humic acids can be described as flexible linear colloids. Cameron et al.21 visualized the humic acid molecules in solution as a series of charged, occasionally branched hydrated strands and concluded that the mean distribution of molecular mass is spherical and Gaussian about the center. Branching results in an increased coil density within the molecule, giving rise to more compact coils than those for a linear molecule of equivalent weight. By accepting the view that humic acids can be considered as a kind of flexible polyelectrolyte, their adsorption can also be compared with general polyelectrolyte adsorption results. A brief review of polyelectrolyte adsorption has been given by Cohen Stuart et al.25 In general, chain type polyelectrolytes behave similarly to neutral polymers at high ionic strength and low degrees of dissociation. However, at low salt concentration and strong dissociation the polyelectrolyte nature is very dominant. A further (15) McKnight, D. M.; Wershaw, R. L.; Bencala, K. E.; Zellweger, G. W.; Feder, G. L. Sci. Total Environ. 1992, 117/118, 485. (16) Engebertson, R. R.; Wandruszka, R. V. Environ. Sci. Technol. 1994, 28, 1934. (17) Cohen Stuart, M. A.; Fleer, G. J.; Bijsterbosch, B. H. J. Colloid Interface Sci. 1982, 90, 310. (18) Koopal, L. K. J. Colloid Interface Sci. 1981, 83, 116. (19) Cohen Stuart, M. A.; Scheutjens, J. M. H. M.; Fleer, G. J. J. Polym. Sci., Polym. Phys. Ed. 1980, 18, 559. (20) Hlady, V.; Lyklema, J.; Fleer, G. J. J. Colloid Interface Sci. 1982, 87, 395. (21) Cameron, R. S.; Thornton, B. K.; Swift, R. S.; Posner, A. M. J. Soil Sci. 1972, 23, 394. (22) Chen, Y.; Schnitzer, M. Soil Sci. Soc. Am. J. 1976, 40, 866. (23) Ghosh, K.; Schnitzer, M. Soil Sci. 1980, 129, 266. (24) Cornel, P. K.; Summers, R. S.; Roberts, P. V. J. Colloid Interface Sci. 1986, 110, 149. (25) Cohen Stuart, M. A.; Cosgrove, T.; Vincent, B. Adv. Colloid Interface Sci. 1986, 24, 143.

Langmuir, Vol. 14, No. 10, 1998 2811

comparison of the adsorption of polymers and polyelectrolytes has been given by Lyklema,26 who also discusses the most important results derived from polymer and polyelectrolyte adsorption theories. Vermeer et al.27 have used the self-consistent field (SCF) theory for polyelectrolyte adsorption to show that the adsorbed amount and the conformation of weak acid polyelectrolytes, adsorbed on a surface with a pH-dependent charge, are influenced by the electrostatic interactions between the different segments of the polyelectrolyte and by the electrostatic and specific (non-Coulombic) interactions between the polyelectrolyte and the surface. For such a system the charges on both the surface and the adsorbed polyelectrolyte are determined by the pH and salt concentration in the local environment of the chargeable groups. At low pH, where the polymer is only weakly negative or even neutral and the surface is strongly positive, the adsorbed polyelectrolyte molecules protrude relatively far into the solution with loops and tails, and a high adsorbed amount results. At high pH where the polymer is highly negative and the surface is almost uncharged, the adsorption values are relatively low, and a flat conformation with a relatively high fraction of trains (segments in contact with the surface) determines the adsorption. With increasing pH the positive surface charge can become overcompensated by the negative charge of the adsorbed polyelectrolyte and an electrostatic barrier arises that minimizes further adsorption. Comparable effects were calculated by Dahlgren and Leermakers.28 The overcompensation of the surface charge by adsorbed polyelectrolytes and the influence of an arising electrostatic barrier on the adsorption isotherm were measured by Meadows et al.,29 who also studied the conformation of the adsorbed molecules by ESR measurements, and De Laat and VandenHeuvel.30 In view of the importance of the electrostatic interactions, as revealed by both the experiments on humic acid and the theoretical results, we will study the adsorption of purified humic acid onto iron oxide as a function of pH and salt concentration. Emphasis will be given to the adsorbed amount and the adsorbed layer thickness. Some attention will be paid to the influence of the presence of metal ions on the adsorption. In the second part of the paper the experimental results will be compared with the results calculated with the self-consistent field theory for the adsorption of (weak) polyelectrolytes31-33 on a variably charged surface.27 The aim of this comparison is to look for similarities in behavior and to gain a better understanding of the adsorption process. In the next paper the effect of the polydispersity of the humic material on the kinetics of adsorption, the appearance of the isotherm, and the adsorption hysteresis will be studied in more detail. Materials and Methods All experiments were performed at room temperature (21 °C) and with KNO3 as indifferent electrolyte. Chemicals used were p.a. quality, and pure water was obtained by percolating it (26) Lyklema, J. Fundamentals of Interface and Colloid Science: Solid-Liquid Interfaces; Academic Press: London, 1995. (27) Vermeer, A. W. P.; Leermakers, F. A. M.; Koopal, L. K. Langmuir 1997, 13, 4413. (28) Dahlgren, A. G.; Leermakers, F. A. M. Langmuir 1995, 11, 2996. (29) Meadows, J.; Williams, P. A.; Garvey, M. J.; Harrop, R. A.; Phillips, G. O. Colloids Surf. 1988, 32, 275. (30) De Laat, A. W. M.; Van den Heuvel, G. L. T. Colloids Surf., A 1995, 98, 53. (31) Bo¨hmer, M. R.; Evers, O. A.; Scheutjens, J. M. H. M. Macromolecules 1990, 23, 2288. (32) Evers, O. A.; Fleer, G. J.; Scheutjens, J. M. H. M.; Lyklema, J. J. Colloid Interface Sci. 1986, 111, 446. (33) Israe¨ls, R.; Leermakers, F. A. M.; Fleer, G. J. Macromolecules 1994, 27, 3087.

2812 Langmuir, Vol. 14, No. 10, 1998 through a mixed-bed ion-exchange column, an active carbon column, and a microfilter. Hematite Particles. The two types of hematite (R-Fe2O3) suspensions used are indicated with the letters B 34 and P,35 respectively, according to their method of preparation. The suspensions are well aged and washed with HCl and dialyzed against purified water till a constant resistance. The hematite B particles resemble parallelograms with an included angle of 60° and a mean size of 50 nm and a BET surface of 43 m2 g-1. The hematite P particles are highly monodisperse35 and nearly spherical with diameters of 86, 212, 402, and 570 nm. The BET surfaces of the three largest P samples were respectively 6.8, 4.3, and 2.6 m2 g-1. The porosity of the samples is negligible. Humic Acid. A purified Aldrich humic acid (Aldrich-Chemie; code: H1,675-2) denoted as PAHA was used. For the purification36,37 10 g of humic acid was added to a 1 l aqueous solution containing 5 mL of concentrated HF and 5 mL of concentrated HCl. The solution was stirred for 8 h and afterward filtered over a Whatman Cellulose filter, grade 2, to remove silica and other soluble minerals. The humic acid residue was washed several times with 1 M HCl, neutralized, dissolved in a NaOH solution of pH 9 for 24 h, and filtered. The suspension was brought to pH 2 with 1 M HCl, stirred for 24 h, and centrifuged using a Beckman JS-7.5 centrifuge at 7500 rpm for 1 h. With these steps, the fulvic acid, humic acid, and humin fractions are separated on the basis of their solubility at different pH values. The humic acid precipitate was then dialyzed against slowly flowing purified water until the resistance values before and after dialysis were equal (about 1 week). Finally the sample was shaken with an acid Dowex 50W-X8 resin for 2 weeks to remove all trace metals. The humic acid obtained in this way is in its proton form. The PAHA sample is freeze-dried and stored in a glass container. Before use the freeze-dried PAHA was resuspended overnight in a KOH solution with pH approximately 10, to a concentration of 2 g L-1, to ensure complete dissolution of the sample. Other PAHA solutions were made from this stock solution. PAHA is ash free, and its elemental composition is as follows: C, 55.8%; O, 38.9%; H, 4.6%; N, 0.6%. The concentration of trace metals is below the detection limit of ICP measurements. PAHA has a mean molecular weight of 23 000 Daltons,38 and the sample is of amphiphilic nature. Proton Adsorption Measurements. Titrations were performed using the automated Wallingford titration system.39 The concentrations of the hematite suspensions were on the order of 20 g L-1, and those of the PAHA solutions were on the order of 2 g L-1. For the hematite a polypropylene basket was used to prevent adsorption to the glass vessel and the suspension was stirred continuously to prevent settling of the suspension. After removal of possible carbon dioxide at pH 3 for about 2 h, successive acid and base titrations were performed at different KNO3 concentrations. The KNO3 concentration was adjusted at pH 3, and after addition an equilibration time of 30 min was allowed. After each addition of titrant, the rate of drift was measured over a 2-min interval after an initial delay of 20 s to allow adequate mixing. The electrode readings were accepted when the drift was less than 0.5 mV min-1. A maximum reading time of 20 min was set for two successive additions. To obtain a good distribution of data points over the pH range studied, the doses of HNO3 and KOH were such that a change of about 5 mV occurred for each addition. The proton consumption of the samples is obtained by (34) Breeuwsma, A.; Lyklema, J. Discuss. Faraday Soc. 1971, 52, 324. (35) Penners, N. H. G.; Koopal, L. K. Colloids Surf. 1986, 19, 337. Penners N. H. G.; Koopal, L. K.; Lyklema, J. Colloids Surf. 1986, 21, 457. (36) Aiken, G. R.; McKnight, D. M.; Wershaw, R. L.; MacCarthy, P. Humic Substances in Soil, Sediment, and Water; Wiley-Interscience: New York, 1985. (37) Thurman, E. M.; Malcolm, R. L. Environ. Sci. Technol. 1981, 15, 463. (38) Vermeer, A. W. P. Interactions between humic acid and hematite and their effects on metal ion speciation. Ph.D. Thesis, Wageningen Agricultural Universtity, Wageningen, The Netherlands, 1996. (39) Kinniburgh, D. G.; Milne, C. J.; Venema, P. Soil Sci. Soc. Am. J. 1995, 59, 417.

Vermeer et al. subtraction of a theoretical blank. The volume of titrant required for a pH change of an equivalent volume of the blank electrolyte solution was calculated from the actual proton concentration and the activity coefficient. The latter was derived from the calculated ionic strength using the improved Davies equation.40 PAHA Adsorption Measurements. All adsorption measurements were carried out in polyalymere tubes, each containing about 0.5 m2 of hematite and varying initial concentrations of PAHA (5-500 mg L-1). Hematite was added by weighing an amount of stock suspension in the tubes. PAHA was added using a volumetric pipet. The H, KNO3, and Cd(NO3)2 concentrations in the tubes were adjusted by adding the required amount of stock solutions with Metrohm 665 Dosimat burets. After all components were added, the pH was checked and adjusted to the desired pH within 0.1 unit. The suspension was shaken, head over head, for about 18 h. Then the pH was readjusted, the suspensions were shaken again for about 1 h, and the final pH was measured. From a separate adsorption experiment it was found that the PAHA concentration became constant after 14 h. After centrifugation for 30 min at 7500 rpm using a Beckman JA-21 or JA-20 centrifuge, the humic acid content of the supernatant was determined with a Hitachi U-3210 spectrophotometer at 254 nm. Control experiments without the oxide showed that no loss of humic acid occurred during centrifugation. The absorbance at 254 nm for a given humic acid concentration is pH- and salt-dependent, and calibration lines were made for each required pH and salt level. Particle Size and Layer Thickness Measurements. The dynamic light scattering (DLS) set up was built from standard components: an ALV-125 laser light spectrometer/goniometer, an ALV-5000 digital correlator, an ALV-800 transputerboard, and a Spectro Physics 35-mW HeNe laser with a wavelength of 632.8 nm. The scattering angle θ was 90° for all the performed experiments. A detailed description of the experimental setup and theory has been described by Lyklema.41 DLS is a modern technique to obtain the average radius of particles dispersed in a solution. The method uses the fluctuations of the scattered light intensity due to Brownian motion of the dispersed particles, which in turn is related to the diffusion coefficient of the (hydrated) particles. From the intensity fluctuations a timedependent autocorrelation function can be obtained. The diffusion coefficient (D) is obtained by fitting a curve, using the second cumulant method, through the autocorrelation function. From the diffusion coefficient the radius of the particles is calculated with the Stokes-Einstein relation. The required viscosity of the PAHA solutions was measured with an automatic viscometer (Viscosometric MS type 53 000). The radii of bare hematite P particles were measured at low pH (between pH 2 and 4) and without the addition of any salt (addition of indifferent electrolyte or increasing the pH partly aggregated the hematite particles). The radii of the PAHA molecules were measured at different pH values and salt concentrations at a PAHA concentration of 50 mg L-1. Prior to the measurements the PAHA solutions were filtered using a 300-nm Millipore filter. For the coated particles the following procedure was used. A small amount of hematite P was added, while stirring strongly, to a filtered PAHA solution adjusted to the required pH and salt concentration. Final and initial PAHA concentrations were about 50 mg L-1, due to the low hematite concentration. After the mixture was stirred strongly for 14 h, it was filtered through a 600-nm Millipore filter and the hydrodynamic radius was measured. This special procedure is required to avoid particle aggregation during the sample preparation.42 The hydrodynamic radius of both the bare and the coated particles was calculated from several hundreds of measurements; experiments with a high rms error were withdrawn. The (40) Davies, C. W. Ion interactions; Butterworth: London, 1962; p 34. (41) Lyklema, J. Fundamentals of Interface and Colloid Science: Fundamentals; Academic Press: London, 1991. (42) Liang, L.; Morgan, J. J. Chemical modelling of aqueous systems II; ACS Symposium Series 416; American Chemical Society: Washington, DC, 1990; p 293.

Adsorption of Humic Acid to Mineral Particles

Figure 1. Surface charge of hematite B as a function of pH in the presence of three concentrations of KNO3: (9) 10-3 M; ([) 10-2 M; (2) 10-1 M.

Figure 2. Proton adsorption isotherms of PAHA as a function of pH in the presence of three concentrations of KNO3: (9) 10-3 M; ([) 10-2 M; (2) 10-1 M. standard deviation for the data in such a series was less than 5% for all presented radii. The layer thickness was calculated by subtracting the radius of the bare particles from that of the covered particles.

Results Proton Adsorption Isotherms of Hematite and PAHA. The surface charge density/pH curves of hematite B for three salt concentrations are given in Figure 1. The point of zero charge (pzc) as determined by pH-STAT titrations upon addition of KNO3 equals 8.9. Both the pzc and charge density curves are in good agreement with previous results.34 The proton-binding isotherms of PAHA at different salt concentrations are shown in Figure 2. The relative positions of the three surface charge/pH curves have been fixed by pH-STAT titrations. The curve at the salt concentration 0.001 M was fixed relative to a reference point (Q0). The value of Q0 () -0.8 mequiv g-1) is obtained by measuring the consumption of 0.1 M HNO3 by PAHA using a back-titration.43 The thus obtained value of Q0 is based on the assumption that at pH ) 1 the degree of dissociation is negligible and represents the acid groups that dissociate between pH 1 and the pH at Q0. The relatively smooth charge density curves and the salt concentration effects on the charge density at a given pH correspond well with results obtained for other humics.4,44 Adsorption of Humic Acid on Hematite. Typical adsorption isotherms of PAHA on hematite B at 0.1 and (43) Boehm, H. P.; Diehl, E. Z. Electrochem. 1962, 66, 642. (44) De Wit, J. C. M.; Van Riemsdijk, W. H.; Nederlof, M. M.; Kinniburgh, D. G.; Koopal, L. K. Anal. Chim. Acta 1990, 232, 189.

Langmuir, Vol. 14, No. 10, 1998 2813

0.01 M KNO3 are shown in parts a and b of Figure 3, respectively. It is observed that the humic acid adsorption increases substantially with decreasing pH. The adsorption isotherms show an initial high-affinity character, followed by a weakly increasing isotherm at elevated PAHA concentrations. These features are commonly observed for the adsorption of humic substances onto mineral particles.10-13,15,45 At 0.1 M KNO3 the high-affinity character seems somewhat stronger than that at 0.01 M salt. The higher initial affinity at 0.1 M than at 0.01 M KNO3 indicates that even at this low PAHA concentration overcompensation of the surface charge occurs. At pH 6 and 9 a slightly higher adsorption is observed, whereas at pH 4 the adsorption is substantially increased with increasing ionic strength. The fact that at pH 9, which is very close to the pzc of hematite, a substantial PAHA adsorption is observed indicates that next to Coulombic interactions also specific attractions are present. Following Van der Steeg,46 the salt effect on the adsorption may be classified as screening enhanced, indicating that the nonelectrostatic attraction between humic acid groups and the surface sites dominates the attraction. Salt screens the electrostatic interactions but is most efficient in screening the mutual repulsion between the adsorbed humic acid molecules, and hence the adsorption is enhanced with increasing salt concentration. Adsorption/desorption experiments showed that the adsorption was reversible for pH changes larger than two pH units. For smaller pH changes, the adsorption was not completely reversible within the time scale of the experiment (18 h). As was mentioned above, surface charge compensation may be an important factor in the adsorption process. As a first approximation, to study the degree of charge compensation, one can compare the initial charges associated with the surface and the adsorbed PAHA to obtain an impression of the overall charge of the complex for a given amount of adsorbed humic acid. Such calculations show that at pH 4 the net total initial charge of the oxide particles plus that of the adsorbed humic acid remains positive. At pH 6 this overall charge is slightly negative, and at pH 9 the total initial charge becomes significantly negative. Qualitatively this result is in agreement with electrophoretic measurements by Tipping and Cooke47 and Davis and Gloor,48 who showed that for pH > 3 the positive iron oxide particles become negative due to humic acid adsorption, even at very low coverage. Due to overcompensation of the charge of the hematite particles, a negative potential profile will develop around the particles. This causes an electrostatic repulsion between the free PAHA molecules and the hematite/PAHA complex that will strongly inhibit further adsorption of the negatively charged humics, minimize the extent of charge overcompensation and lead to a very slow adsorption process.30 We will return to the kinetics of the PAHA adsorption in the following paper.49 Considering the charge compensation in more detail, the charge adjustment due to the interaction of both the surface and the adsorbed humic acid has to be taken into account. In general, this type of charge adjustment upon the adsorption of a weak polyacid on a variable charged surface has been shown theoretically by Vermeer et al.27 (45) Amal, R.; Raper, J. A.; Waite, T. D. J. Colloid Interface Sci. 1992, 151, 244. (46) Van der Steeg, H. G. M.; Cohen Stuart, M. A.; De Keizer, A.; Bijsterbosch, B. H. Langmuir 1992, 8, 2538. (47) Tipping, E.; Cooke, D. Geochim. Cosmochim. Acta 1982, 46, 75. (48) Davis, J. A.; Gloor, R. Environ. Sci. Technol. 1981, 15, 1223. (49) Vermeer, A. W. P.; Koopal, L. K. Langmuir, accepted.

2814 Langmuir, Vol. 14, No. 10, 1998

Vermeer et al.

Figure 3. Effect of pH on the adsorption of PAHA onto hematite B (a) at 0.1 M KNO3 and (b) at 0.01 M KNO3: (9) pH ) 4; ([) pH ) 6; (2) pH ) 9.

When there is a significant charge difference between a surface and an adsorbed weak polyelectrolyte, the component with the highest absolute value of the charge is able to induce charges on the other component. Translating this result to our hematite humic acid system would mean that at pH 4 charges will be mainly induced on the humic acid, whereas at pH 9 the hematite charge will be mainly increased. At the intermediate pH values an additional charge will be induced on both the iron oxide surface and the adsorbed humic acid. Due to mutual compensation of the induced charges, the effect on the adsorption will be relatively small at pH 6. Assuming near charge compensation to occur implies that a relation exists between the adsorbed amount and the overall (initial plus induced) surface charge. Taking pH 6 as a reference state, it is then possible to predict the adsorption at pH 4 from the initial surface charge density of hematite (the PAHA charge adjusts) and to estimate the induced charge on hematite at pH 9 from the adsorbed amount of PAHA (the hematite charge adjusts). The ratio of the charge densities of pure hematite at pH 4 and 6 ranges from 1.9 to 1.7 (depending on the KNO3 concentration). This is in reasonable agreement with the adsorption ratios that range from 1.4 to 1.8. The ratio of the adsorbed amounts of PAHA at pH 9 and 6 is about 0.60-0.45. This might indicate that at pH 9 the charge density of hematite, induced by PAHA adsorption, would be about 30-60 mC m-2, depending on the salt and PAHA concentration. This value cannot be checked, but its magnitude is not unreasonable when compared with the charge adjustment observed in theoretical calculations.27 To investigate the influence of the particle size of the monodisperse hematite P particles on the humic acid adsorption, measurements were made on a series of hematite P particles at pH 6 and 0.01 M KNO3. All experiments were performed with an equal surface area per tube. The results are shown in Figure 4 (solid symbols). The adsorption isotherms on the different hematite P particles are within experimental error the same. The differences between the radii of the hematite particles are not large enough to introduce conformational differences within the adsorbed layer. For the sake of comparison the isotherm of the much smaller, parallelogram-like hematite B particles is also shown in Figure 4 (open symbols). At a PAHA concentration of about 50 mg L-1 the adsorption onto hematite B is about half that of the hematite P particles. At first sight this difference seems very unlikely, because at the small hematite B particles a higher or at least equal adsorbed amount would be expected. However, if it is taken into account that a

Figure 4. Adsorption isotherms of PAHA onto different hematite particles at pH ) 6 and 0.01 M KNO3: (9) hematite P, 212 nm (diameter); ([) hematite P, 403 nm; (2) hematite P, 570 nm; (O) hematite B (equivalent diameter of approximately 50 nm). (×) Adjusted adsorption isotherm of PAHA onto hematite B particles (see text).

relation exists between the adsorbed amount and the surface charges, as postulated above, the differences in PAHA adsorption on hematite P and hematite B can be explained by the differences in surface charge density. At pH ) 6 the ratio of the charge densities on hematite B and P is 1.8. This is indeed close to the observed adsorption ratios. Layer Thickness of the Adsorbed Humic Acid. The layer thickness of PAHA adsorbed at hematite P (86 nm) in 0.01 M KNO3 increases from 37 to 55 nm upon decreasing the pH; see Table 1. The layer thicknesses were measured at overall PAHA concentrations of about 50 mg L-1, to prevent aggregation of the hematite particles. Comparable experiments with silica particles in the presence of 50 mg L-1 PAHA at a pH where no adsorption occurred did not show any effect of PAHA. Thus the presence of PAHA in solution is not affecting the measurements. The measured thicknesses indicate that the adsorbed molecules protrude far into the solution. Experiments with hematite P(570 nm) particles gave comparable layer thicknesses, indicating that the increased radius was not caused by aggregation of the hematite particles. For the sake of comparison in Table 1 also the adsorption levels of PAHA at 50 mg L-1 on hematite B are shown together with the diameters of dissolved PAHA molecules at different pH values. A relevant procedure to compare these properties is by taking the ratios of the measured values, using again pH ) 6 as a reference. Comparison

Adsorption of Humic Acid to Mineral Particles

Langmuir, Vol. 14, No. 10, 1998 2815

Table 1. Experimental Data at 0.01 M KNO3 from Adsorption (onto Hematite B) and Dynamic Light Scattering (Using Hematite P) Experiments adsorbed amount of PAHA

thickness of PAHA adsorbed layer

diameter of PAHA in solution

pH

Γmax (mg m-2)

ratio

δPAHA (nm)

ratio

dPAHA (nm)

ratio

4 6 9

1.2 0.8 0.4

1.5 1.0 0.5

55 40 (37a)

1.35 1.00 0.91

130 138 143

0.94 1.00 1.04

a To overcome aggregation of the hematite particles, hematite was added to the PAHA solution at pH 6. The pH was increased to pH 9 after 0.5 h.

Figure 5. Adsorption isotherms of PAHA onto hematite in the presence of (a) 0.1 M KNO3 and 10-4 M Cd(NO3)2 and (b) 0.01 M KNO3 and 10-4 M Cd(NO3)2: (9) pH ) 4; ([) pH ) 6; (2) pH ) 9; (+) pH ) 4 and 10-3 M Cd2+.

of these ratios reveals two important aspects. (1) The hydrodynamic layer thickness of the adsorbed PAHA and the adsorbed amount both decrease with increasing pH, but the hydrodynamic layer thickness decreases less than the adsorbed amount. (2) The hydrodynamic diameter of the humic acid molecules in solution increases slightly with increasing pH, whereas the hydrodynamic layer thickness decreases substantially. Comparison of the radius of the free PAHA molecules with the layer thickness of the adsorbed PAHA, at a given pH value, shows that the PAHA molecules are flattened upon adsorption. The degree of flattening increases with increasing pH. At high pH values a significant part of the adsorbed molecule is in close contact with the surface. The hydrodynamic layer thickness is largely determined by “tails” of the adsorbed molecules that protrude into the solution.50,51 Thus the relatively low decrease in hydrodynamic layer thickness with increasing pH, compared to the strong decrease in adsorption, can be explained by the fact that some parts of the PAHA molecules protrude relatively far into the solution at all pH values. The pH dependence of both the adsorbed amount and the hydrodynamic layer thickness of PAHA compares well with results reported by Wang and Audebert52 for the adsorption of a cationic polyelectrolyte on silica. Influence of Cadmium Adsorption. The effect of the presence of 10-4 M cadmium on the PAHA adsorption isotherms is shown in Figure 5 for three pH values and two salt concentrations. Several observations can be made when the isotherms are compared with the isotherms of Figure 3. Especially at pH 9 the adsorption of humic acid is increased when cadmium ions are present. The strong effect at pH 9 can be understood by the more profound

cadmium adsorption to the humic acid at high pH53,54 and by the relatively strong cadmium adsorption on hematite. Cadmium adsorption to the humic acid diminishes the electrostatic repulsion in the PAHA layer, and due to cadmium adsorption on the surface the electrostatic attraction between PAHA and the surface increases. It is also observed that the shape of the isotherms is changed compared to that of the isotherms without the cadmium; in the presence of cadmium ions the adsorption is higher at low PAHA concentrations. A slight maximum exists in the adsorption isotherms at pH 6 and 9. Results reported by Tipping12 and McKnight et al.15 showed comparable effects of cadmium on the humic acid adsorption isotherms and their adsorption plateaus. The reason for this behavior is that in the presence of a constant amount of cadmium the Cd binding to PAHA is relatively large at low PAHA concentrations. The effect on the lateral electrostatic repulsion is therefore stronger at low than high PAHA concentrations. A similar maximum was also reported for the adsorption of synthetic polyelectrolytes in the presence of divalent ions on kaolinite55 and anatase.56 At pH 4 and 0.01 M KNO3 experiments with a constant amount of hematite and PAHA and a total cadmium concentration of 10-3 M showed that the PAHA fraction removed from the solution upon centrifugation increased strongly as compared to the situation at 10-4 M cadmium. At this point the difference between adsorption and aggregation has to be emphasized. At pH 4, 0.01 M KNO3, and 10-3 M Cd2+, the isotherm shown in Figure 5b reflects both PAHA adsorption and aggregation. From dynamic light scattering experiments it was found that PAHA aggregated strongly at pH 4 and such a high cadmium

(50) Scheutjens, J. M. H. M.; Fleer, G. J.; Cohen Stuart, M. A. Colloids Surf. 1986, 21, 285. (51) Fleer, G. J.; Cohen Stuart, M. A.; Scheutjens, J. M. H. M.; Cosgrove, T.; Vincent, B. Polymers at Interfaces; Chapman & Hall: London, 1993. (52) Wang, T. K.; Audebert, R. J. Colloid Interface Sci. 1988, 121, 32.

(53) Saar, R. A.; Weber, J. H. Can. J. Chem. 1979, 57, 1263. (54) Milne, C. J.; Kinniburgh, D. G.; De Wit, J. C. M.; Van Riemsdijk, W. H.; Koopal, L. K. J. Colloid Interface Sci. 1995, 175, 448. (55) Ja¨rnstrom, L.; Stenius, P. Colloids Surf. 1990, 50, 47. (56) Bo¨hmer, M. R.; El Attar Sofi, Y.; Foissy, A. J. Colloid Interface Sci. 1994, 164, 126.

2816 Langmuir, Vol. 14, No. 10, 1998

Vermeer et al.

concentration, whereas the radius of the PAHA molecules decreased upon addition of small amounts of cadmium ( 4.5, rather than the available surface area seems to be the factor that determines the adsorbed amount. The importance of such an electrostatic barrier for the adsorption of synthetic polyelectrolytes has also been mentioned by other authors.30,62 Final Remarks and Conclusions The adsorption isotherms of PAHA onto hematite particles show an increased adsorption with decreasing pH. Increasing the salt concentration and adding cadmium nitrate also increase the adsorption, but the effects are much smaller than the pH effect. The hydrodynamic radius, measured by dynamic light scattering, shows a decreasing thickness with increasing pH. A relatively small decrease of the hydrodynamic layer thickness is observed as compared to the decrease in the adsorbed amount. The adsorbed amount and the conformation of the adsorbed layer are only slowly altered by changes of the conditions (pH and salt and cadmium concentration). These effects show that time effects may be important even after waiting times of 10-20 h. In the next paper49 these kinetic aspects will be discussed in somewhat more detail. The same trends are also observed with the calculated results for the adsorption of linear polyelectrolytes on a surface with a variable charge. The general conclusion of this study is that qualitatively (62) Bain, D. R.; Cafe, M. C.; Robb, I. D.; Williams, P. A. J. Colloid Interface Sci. 1982, 88, 467.

Langmuir, Vol. 14, No. 10, 1998 2819

the adsorption of naturally occurring humic substances onto mineral oxide particles can be understood as a function of pH and salt concentration using an advanced polyelectrolyte adsorption theory. Adsorption of the PAHA molecules onto mineral surfaces occurs due to electrostatic attraction between the hematite surface and the PAHA molecules combined with a specific (non-Coulombic) attraction. The adsorbed amount is determined by two opposite forces; charge compensation and specific attraction enhance the adsorption, whereas lateral electrostatic repulsion and loss of entropy inhibit the adsorption. A clear resemblance is observed between the behavior of PAHA and that of simple linear polyelectrolytes near a surface with a variable charge. At high pH and low salt concentration the humic acid molecules are adsorbed relatively flat on the surface, but some parts of the molecule that are not in the direct vicinity of the surface protrude relatively far into the solution due to lateral repulsion effects. At low pH and high salt concentration the adsorbed layer can be described by a large fraction of adsorbed segments in the vicinity of the surface but not in direct contact with the surface. Due to the high fraction of these segments, the adsorbed amount per unit area is relatively high. The relatively small decrease of the hydrodynamic layer thickness with increasing pH as compared to that of the adsorbed amount points to the fact that the length of the humic acid parts that protrude into the solution decreases only slightly with increasing pH. For pH values above 5-6 the surface charge is overcompensated by the charge associated with the adsorbed humic acid. Due to the surplus of negatively charged humic acid groups, a negative potential is developed around the initially positively charged hematite particles. Once this electrostatic barrier is developed, further adsorption is inhibited and a PAHA depletion zone arises in the solution. It may be concluded that the adsorbed humics can be described as a dynamic layer, protruding into the solution. Slow rearrangements occur due to changes in the environmental conditions, and the fraction of humic acid groups protruding into the solution can be adjusted. It is clear that the calculations based on the SCF theory help us understand the adsorption mechanisms of the experimental hematite/PAHA system. The distribution of the humic material near the interface and the electrostatic humic-humic and humic-oxide interactions have a large influence on the adsorption behavior. In our opinion the trends in adsorption with such a complicated system can be much better compared with those obtained with the SCF theory for polyelectrolyte adsorption than with Langmuir type isotherm equations, as has been done in the past.6,12,13 Acknowledgment. This work was partially funded by the European Union under Contract No. STEPCT900031. Joanne Klein Wolterink is kindly acknowledged for carrying out a series of preliminary calculations that have paved the road for the present ones. LA970624R