Investigation of Interparticle Forces in Natural Waters: Effects of

River Humic Acid (SHA), has been investigated using an atomic force microscope (AFM). Two particle surfaces were used, alumina and a hydrous iron oxid...
1 downloads 0 Views 151KB Size
Download PDF
Environ. Sci. Technol. 2004, 38, 4791-4796

Investigation of Interparticle Forces in Natural Waters: Effects of Adsorbed Humic Acids on Iron Oxide and Alumina Surface Properties SYLVIA SANDER, LUKE M. MOSLEY, AND KEITH A. HUNTER* Department of Chemistry, University of Otago, P.O. Box 56, Dunedin, New Zealand

The nature of interparticle forces acting on colloid particle surfaces with adsorbed surface films of the internationally used humic acid standard material, Suwannee River Humic Acid (SHA), has been investigated using an atomic force microscope (AFM). Two particle surfaces were used, alumina and a hydrous iron oxide film coated onto silica particles. Adsorbed SHA dominated the interactive forces for both surface types when present. At low ionic strength and pH >4, the force curves were dominated by electrostatic repulsion of the electrical double layers, with the extent of repulsion decreasing as electrolyte (NaCl) concentration increased, scaling with the Debye length (κ-1) of the electrolyte according to classical theory. At pH ∼4, electrostatic forces were largely absent, indicating almost complete protonation of carboxylic acid (-COOH) functional groups on the adsorbed SHA. Under these conditions and also at high electrolyte concentration ([NaCl] > 0.1 M), the absence of electrostatic forces allowed observation of repulsion forces arising from steric interaction of adsorbed SHA as the oxide surfaces approached closely to each other (separation < 10 nm). This steric barrier shrank as electrolyte concentration increased, implying tighter coiling of the adsorbed SHA molecules. In addition, adhesive bridging between surfaces was observed only in the presence of SHA films, implying a strong energy barrier to spontaneous detachment of the surfaces from each other once joined. This adhesion was especially strong in the presence of Ca2+ which appears to bridge SHA layers on each surface. Overall, our results show that SHA is a good model for the NOM adsorbed on colloids.

Introduction In natural waters, colloidal particles are characterized by extreme complexity and diversity. However, their surface properties appear to be rendered quite uniform by ubiquitous surface films of natural organic matter (NOM) that envelop the particles (1, 2). Thus, almost all particles, regardless of inherent mineral type, are found to be negatively charged because of the ionization of acidic functional groups such as carboxyl -COOH on the NOM dominating the surface chemistry (3). The NOM coating on natural particles is highly heterogeneous in size and chemical properties (4). Usually * Corresponding author phone +64-3-479-7917; fax: +64-3-4797906; e-mail: [email protected]. 10.1021/es049602z CCC: $27.50 Published on Web 08/17/2004

 2004 American Chemical Society

such NOM is characterized on an operational basis with common groupings being humic and fulvic acids, and these have different size characteristics when derived from different sources (5). These molecules have been shown to contain various functional groups and structures (e.g., phenolic groups, quinone structures, nitrogen and oxygen as bridge units, and carboxylic acid groups), the exact nature of which is often difficult to determine. The characteristics of the organic matter also change with changing solution properties (e.g., pH, salinity, ionic composition). In estuaries, a large decrease in the magnitude of the negative charge on suspended particles is observed with the initial salinity increase from freshwater values (6), and the aggregation of colloidal material commonly occurs in this salinity region (6-10). The decrease in negative charge is caused by the seawater cations screening the charge of negative functional groups on the organic coatings (8). This may also lead to humic acid molecules changing their configuration from a more extended form at low ionic strength to a more coiled form at higher ionic strengths (11). The development of colloid probes for the atomic force microscope (AFM) has enabled the measurement of very small interparticle forces (12), and since that time the AFM has been used widely to examine interactions between synthetic organic molecules adsorbed to surfaces (13-16). The advantage of AFM techniques is that they provide fundamental information on the forces existing between surfaces in solution and that the solution composition can be readily changed to elucidate the properties of the surface or an adsorbed material. AFM has been used to investigate the structure of humic acids sorbed on various mineral surfaces (17, 18). Recently, we reported the first use of AFM to measure the force interactions between iron oxide surfaces coated with NOM of both seawater and freshwater origin (19). The interparticle forces were dominated by electrostatic repulsion at low ionic strengths while at higher ionic strengths steric repulsive forces, arising from confinement of organic molecules between the surfaces, were found. Large adhesive forces were also observed under certain solution conditions (high ionic strength, low pH), and these would prevent particle separation. However these measurements were conducted only for an iron oxide coated surfaces coated with NOM of localized origin and a synthetic polyelectrolyte. In the present study, an AFM was used to directly measure the forces between iron oxide and alumina surfaces coated with Suwannee River Humic Acid (SHA). This is an internationally available aquatic humic acid (International Humic Substances Society, IHSS) that is widely used as a reference material for the study of freshwater NOM. These measurements were carried out at various solution pH values and ionic strengths. The aim of our study was to determine if SHA is a suitable model substance for simulating the interparticle forces applicable to river water and estuarine colloids.

Methods Force-distance relationships were measured with a Nanoscope III AFM (Digital Instruments) by methods described previously (12). Briefly, this involves gluing (EPON 1004 heat glue) a spherical particle (2-40 µm in radius, R) of a suitable material onto an AFM cantilever tip with the aid of an optical microscope and X-Y-Z translation stage. Cantilevers (Digital Instruments) were standard V-shaped silicon nitride and had a spring constant of 0.58 N m-1 measured according to ref 20. The AFM was then used to measure the force (F) between this colloidal probe and a flat plate made of the same material. VOL. 38, NO. 18, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4791

The measurements were made with both the cantilever and the plate immersed in an electrolyte solution, allowing examination of the effects of solution composition and pH on the force-separation properties. According to the Derjaguin approximation (21), the intensive property F/R between a sphere and a flat plate is twice the magnitude of the corresponding force exerted between two identical spheres, allowing a direct correspondence between the experimental measurements and the actual forces between pairs of colloid particles. Force-separation curves may be measured either during approach of the colloid particle on the tip toward the flat surface or during separation from the flat surface. Unless otherwise stated, the results presented below are in the form of approach curves. The region of zero separation was defined as the point when the movement of the tip was equal to the movement of the flat plate surface. The normalized force curve data was fitted, where applicable, to DLVO theory in which the force is modeled as the sum of a repulsive electrostatic double-layer force and an attractive van der Waals force. The electrostatic force was calculated from the Poisson-Boltzmann equation by means of an exact numerical solution (22) using a constant potential and Debye length as boundary parameter (23). This allowed fitting for separations greater than 5 nm. The van der Waals force (FVDW) was calculated using the equation

FVDW A ) R 6D2

(1)

where R is the sphere radius, D is the separation distance between the surfaces, and A is the Hamaker constant. The Hamaker constant A ) 1.35 × 10-19 J for the iron oxide coated silica was taken from measurements in our previous work (19). There appear to be no other published values of Hamaker constants for hydrated iron oxides. For alumina, we calculated a value of 2.5 × 10-20 J from separate force measurements made in NaNO3 electrolyte, which is similar to a reported value for Al2O3 of 6.0 × 10-20 J (24). It is considered that, at present, Hamaker constants can be determined only within a factor of about 7 at best (24). Only the absolute value of the surface potential (ψ) can be measured by AFM, so here we report only absolute values. The sign of ψ can usually be deduced from a knowledge of the surface properties (e.g., the pH at which the surface has a zero net charge, pH(PZC)). Two different surfaces, silica coated with iron oxide and alumina, were examined as models for the underlying minerals in natural water particles. The silica spheres had a diameter of 4-5 µm and were used with a flat quartz plate. Both silica surfaces were coated with iron oxide as described previously (19). Basically this involved raising the pH of a 10-5 M FeCl3 solution in the AFM cell to pH 7 using small additions of a NaOH solution. The RMS roughness of the plate surface after formation of the iron oxide coating measured by AFM using an oxide sharpened tip was approximately 7 ( 3 nm, and XPS measurements suggested that the iron oxide coating was not completely uniform. The alumina spheres had a diameter of 25-35 µm and were obtained from R. S. A. Le Rubis (France). These are fused and recrystallized alumina spheres (99.995% Al2O3) and represent a mix of R-, δ-, and η-Al2O3 phases. Once attached to the cantilevers, the spheres were UV irradiated for at least 45 min and allowed to cool before assembly in the AFM cell. The alumina spheres were used with a sapphire plate (13 mm diameter) obtained from Harrick Scientific Company and had no specific orientation. RMS roughness measurements made by AFM using the polished sapphire surface and an alumina sphere attached to the tip gave approximately 5 ( 3 nm. Furthermore, scanning electron micrographs showed that the surface of the alumina sphere in contact 4792

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 18, 2004

with the sapphire surface was very smooth on a 10 nm scale. The plate surface was cleaned in concentrated nitric acid, rinsed with Milli-Q water (MQ), and polished with Al2O3 (0.015 µm Aldrich) on a polishing pad for a minimum of 5 min until the surface became hydrophilic. At this stage, the surface was rinsed again with MQ and assembled in the AFM stage. Care was taken to ensure that the iron oxide and alumina surfaces were not exposed to organic and inorganic (particular silica) contaminants. Contamination of iron oxide and alumina due to coating with silica causes the point of zero charge to decrease. Therefore all solutions used were kept in acid-cleaned Teflon or fluorinated plasticware. The pH of all solutions injected into the cell were also kept between pH 4 and pH 10 so that no significant dissolution of the silica or metal oxide layers, or the quartz AFM cell, would be expected. The NaCl, NaOH, and CaCl2 solutions were prepared from AR-grade chemicals that had been baked to remove organic contamination at 450 °C for at least 12 h. The pH of the salt solutions was adjusted with NaOH or quartz distilled HNO3. Following the characterization of the Fe and alumina surfaces, adsorption of the humic acids was carried out using methods based on our earlier research (19). SHA was obtained from IHSS in freeze-dried form, and fresh 100 mg/L solutions were prepared prior to each experiment. SHA has been found to have a molecular weight of about 2.2 kDa with a relatively high polydispersity (Mw/Mn ) 1.6) (25). All solutions were filtered through precombusted (500 °C, >12 h) 0.7 µm glass fiber (Whatman GF/F) filters. Prior to injection into the AFM cell, the SHA solution was adjusted to pH 4.5 with NaOH. About 20 mL (ca. 200 times the AFM cell volume) of the SHA solution was slowly injected into the AFM cell at a rate of ca. 1 mL every 2 min. The solution was left in the AFM cell for at least 16 h to allow for the adsorption of organic material and equilibration with the surfaces. Following absorption of the humic acids, the excess organic material was rinsed with several 1 mL aliquots of deionized water. For each experiment, we made at least 30 force curve measurements, both with different particles on the tip and in different regions of the plate surface for a given sphere and surface combination. It is not practical to calculate an “average” curve or errors in the measured force from replicate curves because of small variations in the zero force between measurements. However, almost all replicates gave forceseparation curves that were very similar in magnitude and general shape. From these, we selected a typical example for presentation in this paper. Following each experiment, the particle on the cantilever was examined and its radius measured using transmission electron microscopy (Cambridge Stereoscan).

Results and Discussion Prior to the adsorption of humic acids, the properties of the iron oxide and alumina surfaces were examined by AFM. The forces between iron oxide and alumina surfaces in 0.001 M NaCl solutions having varying solution pH are shown in Figure 1, panels a and b, respectively. For the Fe surface at pH 6.1, the forces between the iron-coated surfaces were monotonically attractive and of short range (0-15 nm). This suggest that the net charge on the particles is very small and that attractive van der Waals forces dominate. This region of low charge is within the range of previously reported values for the point of zero charge (PZC) of iron oxide, as already reported (19). A repulsive force was measured at higher (pH 9.0) or lower (pH 4.0) pH as the surface hydroxyl groups are protonated and deprotonated, respectively. The best fit values for the Debye length (κ-1) obtained using DLVO theory were 11 (pH 4.0), and 10.3 nm (pH 9.0), which compare well with the theoretical value for this electrolyte of 9.9 nm.

FIGURE 1. pH dependence of the normalized force (F/R) as a function of surface separation for (a) iron oxide-coated silica and (b) alumina surfaces in 0.001 M NaCl solution. The solid lines show forces calculated from DLVO theory as explained in the text. The best fits for the surface potential (ψ) and Debye length (K-1) were as follows: iron oxide: pH 4.0, |ψ| ) 45 mV, K-1 ) 11 nm; pH 6.1, not calculated; pH 9.0, |ψ| )45 mV, K-1 ) 10.3 nm; alumina: pH 4.4, |ψ| ) 23 mV, K1 ) 7 nm; pH 6.2, |ψ| ) 17 mV, K1 ) 8.5 nm. These results compare well with the calculated Debye length of 9.9 nm for this electrolyte. Similarly for the alumina particle and sapphire plate surfaces, a low repulsive force was observed at pH 9.3 (Figure 1b), which is near previously reported PZC values for alumina surfaces (26, 27). As the pH was lowered a repulsive force of greater magnitude developed on the alumina surface, which is consistent with the deprotonation of hydroxyl groups on the alumina. Because the alumina and sapphire surfaces may not be identical materials, we also measured the PZC of the alumina particle and the sapphire plate surface separately in experiments in which the other surface was silica (i.e., alumina particle + quartz plate or silica particle + sapphire plate). This is possible because the silica has such a low PZC (pH < 4) (26). The results indicated a PZC for the alumina particle in the pH range of 8.6-10.2, as reported by Veeramasuneni et al. (26) for R-alumina. For sapphire, a PZC of pH ∼9.3 was found, also consistent with ref 26. We were not able to work above pH 10 because of the possibility of dissolution of silica from the surface of the AFM cell. As with the iron oxide surface, the Debye lengths calculated by DLVO fitting agreed quite well with the theoretical value of 9.9 nm for this electrolyte: 7 nm (pH 4.4) and 8.5 nm (pH 6.2). Next, a series of NaCl solutions of different concentration at pH ∼6 were introduced into the AFM cell (Figure 2a,b). For the iron oxide surface there was no significant doublelayer repulsive force observed at this pH in 0.001 M NaCl (Figure 1a), with the result that there was very little variation of long-range (>10 nm) forces in different NaCl concentrations (Figure 2a), attractive forces being observed in all cases. By contrast, alumina exhibited a significant double-layer force in 0.001 M NaCl at pH ∼6 (Figure 1b). As expected, this repulsive force reduced significantly as the NaCl concentra-

FIGURE 2. Effect of NaCl electrolyte concentration (at pH ∼6) on the normalized force (F/R) as a function of surface separation for (a) iron oxide and (b) alumina surfaces. Solid lines for alumina surfaces in 0.001, 0.01, and 0.1 M NaCl are calculated from DLVO theory. Fitted DLVO parameters were 0.001 M NaCl: |ψ| ) 17 mV, K1 ) 8.5 nm; 0.01 M: |ψ| ) 15 mV, K1 ) 3 nm; 0.1 M: |ψ| ) 9 mV, K1 ) 1 nm. The fitted Debye lengths compare well with theoretical values for these electrolytes of 9.9 nm (0.001 M), 3.0 nm (0.01 M), and 0.95 nm (0.1 M), respectively. tion increased. The range of this force scales with the theoretical Debye length, which is consistent with an electrostatic double-layer force being present. In 1 M NaCl (and to a lesser extent in 0.7 M NaCl), the electrostatic force is sufficiently screened so that the attractive van der Waals forces become evident at separations below about 15 nm (Figure 2b). The behavior of both oxide surfaces with respect to electrolyte concentration and pH observed in Figures 1 and 2 is classical and can be understood in terms of the balance between van der Waals attraction and electrical double-layer repulsion and how the latter is moderated by electrolyte composition (double-layer screening) and pH (surface ionization of the oxides). However, once the iron oxide and alumina surfaces had been exposed to SHA solution overnight, very different force interactions were observed compared to the clean surfaces. Figure 3a,b shows the force-separation behavior at different pH values in 0.001 M NaCl for both iron oxide and alumina surfaces after adsorption of SHA (corresponding to Figure 1a,b without SHA). For both oxide surfaces, the force is much lower at pH ∼4 than at higher pH, indicating that the PZCs are now similar and much lower that when SHA is absent. This is consistent with the adsorption of SHA having carboxylic acid functional groups (e.g., -COOH, pKa ∼ 4). At higher pH, uniform repulsive forces are observed, consistent with ionization of such functional groups in the adsorbed SHA. However, there are differences between the two oxide surfaces. In the absence of SHA, the repulsive forces for iron oxide are about twice as great as those for alumina, such a difference probably resulting from different charge densities on the two oxides. However, with SHA added, there is a large increase (4-6×) in the magnitude of the repulsive forces at VOL. 38, NO. 18, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4793

FIGURE 3. pH dependence of the normalized force (F/R) in 0.001 M NaCl as a function of surface separation after adsorption of SHA: (a) iron oxide and (b) alumina surfaces. Solid lines are calculated from DLVO theory as described in the caption to Figure 1. Fitted DLVO parameters were (a) iron oxide: pH 4.0, |ψ| ) 37 mV, K1 ) 18 nm; pH 6.1, |ψ| ) 100 mV, K1 ) 19 nm; pH 9.2, |ψ| ) 88 mV, K1 ) 18 nm; (b) alumina: pH 6.2, |ψ| ) 32 mV, K1 ) 10.5 nm; pH 9.3, |ψ| ) 29 mV, K1 ) 10 nm. The fitted Debye lengths agree well with the theoretical value of 9.9 nm for this electrolyte in the case of the alumina surface but are higher for the iron oxide surface.

FIGURE 4. Effect of NaCl electrolyte concentration (at pH ∼6) on the normalized force (F/R) as a function of surface separation after adsorption of SHA: (a) iron oxide and (b) alumina surfaces. Solid lines are calculated from DLVO theory. Fitted DLVO parameters were (a) iron oxide: 0.0001 M NaCl, |ψ| ) 125 mV, K1 ) 37 nm; 0.001 M, |ψ| ) 100 mV, K1 ) 18 nm; (b) alumina: 0.001 M NaCl, |ψ| ) 32 mV, K1 ) 11 nm; 0.01 M, |ψ| ) 19 mV, K1 ) 3 nm; 0.1, 0.7, and 1 M, |ψ| ) 9 mV, K1 ) 1 nm. The fitted Debye lengths compare well with theoretical values for these electrolytes of 30 nm (0.0001 M), 9.9 nm (0.001 M), 3.0 nm (0.01 M), and 0.95 nm (0.1 M) respectively, except for iron oxide in 0.001 M NaCl.

pH > 4 (cf. Figures 3a and 1a), whereas for alumina the forces are of similar magnitude to the SHA-free case (cf. Figures 3b and 1b). This suggests that the adsorption density of SHA on iron oxide may be greater than that on alumina or that the conformation of adsorbed molecules is such that a greater proportion of ionized groups is exposed to the solution in the case of iron oxide. It may also be related to the different sizes of the particles used (4-5 µm for iron oxide-silica, 25-35 µm for alumina). In support of this, the surface potentials calculated by DLVO fitting for both surfaces show much higher potentials in the case of SHA adsorbed to iron oxide than to alumina at both pH 6 and 9, which would also be consistent with a greater adsorption density of acidic functional groups on the iron oxide surface. The fact that the inherent surface properties of the iron and aluminum oxides are completely masked in the presence of the adsorbed SHA shows clearly that SHA is very strongly adsorbed by the surface and is retained when the SHA solution is replaced in the AFM cell by the various electrolyte solutions. This is quite different from a classical adsorption equilibrium in which the adsorbed SHA remains in equilibrium with SHA in solution and indicates that, in our case, the material is essentially irreversibly adsorbed. Humic acids are a complex mixture of different molecules, some of which may have greater affinity for adsorption than others depending on pH, the identity of the mineral surface, the average MW, polydispersity and source of the organic matter sample and the ionic strength (28). For example, it has been recently shown that the adsorption of low MW humic substances, which have a greater density of acidic functional groups, is

greater than that of higher MW substances on positively charged surfaces such as hematite because of a greater contribution from ligand exchange and/or electrostatic attractions rather than hydrophobic interactions (28). This selective adsorption of the most surface-active fraction most likely accounts for the adsorbed NOM found on particles in the aquatic environment (2, 29, 30). The absence of electrostatic double-layer forces at pH ∼4 in Figure 3a,b allows other forces to be observed. For SHA on iron oxide at pH 4.0 (Figure 3a), a short-range repulsive force that is jagged in nature is seen commencing at 10 nm separation. Evidence for similar, but less pronounced, discontinuities in the force curve for the two higher pH cases is also seen around this separation distance. These localized repulsive forces arise from steric repulsion caused by the work required to confine the polymer to a smaller volume on the surface (31) and have been observed in the presence of synthetic polymers (15, 31) and NOM films in seawater and freshwater (19). The jaggedness is likely a result of the AFM tip pushing through the organic layers adsorbed on the surface. The short-range barrier provided by steric repulsion is an important mechanism for stabilizing the particles. In the case of alumina, the steric repulsion barrier is less evident because of the dominance of the van der Waals attractive force. Figure 4a,b shows the force-separation behavior at different ionic strength (NaCl solutions at constant pH) for both surfaces after adsorption of SHA (corresponding to Figure 2a,b without SHA). The trends are similar for both surfaces, and as also seen in Figure 3a,b, the forces for the

4794

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 18, 2004

iron oxide case are much larger in magnitude than those observed in the absence of SHA (cf. Figures 2a and 4a) whereas for alumina the general magnitude is similar with and without adsorbed SHA. As with Figure 3a,b, these results also demonstrate the essentially irreversible nature of SHA adsorption. As expected the long-range electrostatic repulsion barrier reduces with increasing ionic strength because of screening of the double layer by electrolyte. In natural water systems, this decrease in the repulsive barrier by electrolyte is expected to lead to aggregation of the colloids (19). At low ionic strengths (10-3-10-4 M NaCl), the long-range repulsive force begins at a separation distance of 70-100 nm from the surface, increasing exponentially toward the surface until a separation of 5-10 nm. At this point a change in the nature of the repulsive force is seen, particularly in the case of iron oxide. The force curve becomes steeper and more jagged in nature, indicative of steric repulsion as the adsorbed SHA layers on the particle and the plate encounter each other. We note also that as the ionic strength increases to 0.1 and 0.7 M NaCl, the steric repulsion region shrinks closer to the surface, indicating some collapse of the adsorbed SHA layer. This is consistent with our previous observations for iron oxide surfaces coated with NOM obtained from river and seawater (19). Humic acids and related polyelectrolytes have many groups or segments that can potentially be adsorbed to a surface. However, only some of the functional groups bind to the surface, and electrostatic repulsion occurs between the segments that extend into solutions leading to a relatively flat conformation (32). SHA and other types of NOM have highly negatively charged functionalities that would repel each other in low ionic strength solutions (3). This is also consistent with measurements of the hydrodynamic radius of Aldrich humic acid in solution (11) that show a monotonic decrease from approximately 10 to 1 nm with ionic strength increasing from 0.001 to 1 M NaCl. This decrease was interpreted as a continuous change in the conformation from the charged, uncoiled macromolecule at low ionic strength to a fully coiled molecule by about 0.5 M NaCl (11). Increasing the pH also increased the hydrodynamic radius slightly, and this effect was more significant at low ionic strength (11). The results in the present study also appear to indicate that the SHA molecules coil increasingly as the ionic strength increases. Thus at higher ionic strengths where the double-layer thickness is negligible, the repulsive force is primarily of steric origin and is similar to that observed at pH ∼4 in Figure 3a,b where the SHA is largely un-ionized. Similar force profiles have been reported for interactions between surfaces coated with synthetic polymers (13, 15). When certain macromolecules such as adhesives are present on a surface, the force-separation curves usually show hysteresis effects (33) in which the force measured as the colloid particle approaches the flat plate surface (approach curve) is different from that in which the colloid particle is moving away from the plate (separation curve). The separation curve often exhibits abrupt discontinuities because of unwinding or detachment from the surface of the macromolecule (34, 35). Figure 5a,b shows approach and separation curves for alumina surfaces with and without adsorbed SHA in 0.001 M NaCl electrolyte. In the absence of SHA (Figure 5a), the approach and separation curves are identical, indicating that as the colloid particle moves relative to the plate surface, the forces are in equilibrium on the time scale of the approach (1-2 in-out cycles per second). In this simple case, the dominant force is electrostatic repulsion of the double layers which is able to change rapidly through the diffusion of small ions (Na+, H+, Cl-, OH-) in the double layer. In the presence of adsorbed SHA (Figure 5b), there is pronounced hysteresis with much stronger attractive forces close to the surface on separation. This suggests that the

FIGURE 5. Normalized force (F/R) as a function of surface separation for both approach and separation for alumina surfaces in 0.001 M NaCl (a) pH 6.2 with no SHA adsorbed and (b) pH 4.2 with SHA adsorbed.

FIGURE 6. Normalized force (F/R) as a function of surface separation for both approach and separation for alumina surfaces with adsorbed SHA at pH ∼6 in 0.01 M CaCl2 solution (symbols) and 0.01 M NaCl solution (dashed and dotted lines). large adsorbed SHA molecules are compressed during approach of the colloid particle but cannot relax to equilibrium easily when the sphere is withdrawn (15). The detachment of the particle in steps over a large range (4 nm) implies that the SHA bridges between the two surfaces. The jumps indicate unwinding or detachment of sections of the polymeric material. Similar hysteresis has also been observed in studies involving synthetic polyelectrolytes (15, 36, 37) and both marine and freshwater NOM (19). Figure 6 shows approach and separation curves for SHA adsorbed on alumina in 0.01 M CaCl2 at pH 6.2 as compared with equivalent curves in 0.01 M NaCl electrolyte. The approach curve shows a decrease in the repulsive force compared to equivalent NaCl solutions (dashed line in Figure 6), which is consistent with the Schulze-Hardy rule that the double-layer repulsion is reduced with increasing valency of the counterion (38) as well as higher ionic strength. However, the separation curve with Ca2+ present exhibits a very pronounced hysteresis indicative of bridging attraction VOL. 38, NO. 18, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4795

between the surfaces. For Na+ as cation there is almost no corresponding hysteresis (dotted line in Figure 6). This implies that, when a divalent ion such as Ca2+ is present, bridging between the functional groups of organic matter adsorbed on each surface can take place (13, 39). In summary, we have shown that the surface force properties of iron hydroxide and aluminum oxide colloids, after adsorption of SHA, are very similar to each other in respect of the effects of pH and ionic composition. This confirms earlier work on the effects of adsorbed NOM on various mineral particles (30, 40, 41). Furthermore, we have shown that the behavior of adsorbed SHA with respect to colloid particle interactions is very similar to that exhibited by marine and freshwater NOM (19). Thus we suggest that SHA provides a suitable and conveniently available model material for the further study of colloid properties in natural waters, especially by AFM techniques. These have considerable potential for better understanding of particle interactions in natural waters and also in water treatment systems.

Acknowledgments We thank William Ducker, Kate McGrath, and Urszula Nowastaska for their help with this research. We are grateful to the Marsden Fund of the Royal Society of New Zealand for financial assistance.

Literature Cited (1) Loeb, G. I.; Neihof, R. A. In Applied Chemistry at Protein Interfaces; Baier, R. E., Ed.; American Chemical Society: Washington, DC, 1975; p 319. (2) Hunter, K. A.; Liss, P. S. Nature 1979, 282, 823. (3) Hunter, K. A. Limnol. Oceanogr. 1980, 25, 807. (4) Santschi, P. H.; Lenhart, J. J.; Honeyman, B. D. Mar. Chem. 1997, 58, 99. (5) Perminova, I. V.; Frimmel, F. H.; Kudryavtsev, A. V.; Kulikova, N. A.; Abbt-Braun, G.; Hesse, S.; Petrosyan, V. S. Environ. Sci. Technol. 1983, 37, 2477. (6) Hunter, K. A. Geochim. Cosmochim. Acta 1983, 47, 467. (7) Boyle, E.; Collier, R.; Dengler, A. T.; Edmond, J. M.; Ng, A. C.; Stallard, R. F. Geochim. Cosmochim. Acta 1974, 41, 1719. (8) Eckert, J. M.; Sholkovitz, E. R. Geochim. Cosmochim. Acta 1976, 40, 847. (9) Sholkovitz, E. R.; Boyle, E. A.; Price, N. B. Earth Planet. Sci. Lett. 1978, 40, 130. (10) Powell, R. T. Mar. Chem. 1996, 55, 165. (11) Cornel, P. K.; Summers, R. S.; Roberts, P. V. J. Colloid Interface Sci. 1986, 110, 149. (12) Ducker, W. A.; Senden, T. J.; Pashley, R. M. Nature 1991, 353, 239. (13) Dahlgren, M. A. G.; Waltermo, A.; Blomberg, E.; Claesson, P. M.; Sjostrom, L.; Akesson, T.; Jonsson, B. J. Phys. Chem. 1993, 97, 11769.

4796

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 18, 2004

(14) O’Shea, S. J.; Welland, M. E.; Rayment, T. Langmuir 1993, 9, 1826. (15) Biggs, S.; Healy, T. W. J. Chem. Soc., Faraday Trans. 1994, 90, 3415. (16) Rojas, O. J.; Claesson, P. M.; Muller, D.; Neuman, R. D. J. Colloid Interface Sci. 1998, 205, 77. (17) Maurice, P., A.; Namjesnik-Dejanovic, K. Environ. Sci. Technol. 1999, 33, 1538. (18) Namjesnik-Dejanovic, K.; Maurice, P., A. Geochim. Cosmochim. Acta 2000, 65, 1047. (19) Mosley, L.; Hunter, K.; Ducker, W. A. C. Environ. Sci. Technol. 2003, 37, 3303. (20) Senden, T. J.; Ducker, W. A. Langmuir 1994, 10, 1003. (21) Israelachvili, J. N. Intermolecular and Surface Forces; Academic Press: London, 1992. (22) Verwey, E. J. W.; Overbeek, T. G. Theory of the Stability of Lyphobic Colloids; Elsevier: Amsterdam, 1948. (23) Meagher, L.; Franks, G. V.; Gee, M. L.; Scales, P. J. Colloids Surf. A 1999, 146, 123. (24) Ackler, H.; French, R.; Chiang, Y.-M. J. Colloid Interface Sci. 1996, 179, 460. (25) Chin, Y.-P.; Aitken, G.; O’Loughlin, E. Environ. Sci. Technol. 1994, 28, 1853. (26) Veeramasuneni, S.; Yalamanchili, M. R.; Miller, J. D. J. Colloid Interface Sci. 1996, 184, 594. (27) Franks, G. V.; Meagher, L. Colloids Surf. A 2003, 214, 99. (28) Hur, J.; Schlautman, M., A. J. Colloid Interface Sci. 2003, 264, 312. (29) Hunter, K. A. In Marine Particles: Analysis and Characterization; Hurd, D. C., Spencer, D. W., Eds.; American Geophysical Union: Washington, DC, 1991; p 259. (30) Hunter, K. A.; Liss, P. S. Limnol. Oceanogr. 1982, 27, 322. (31) Gregory, J. In The Scientific Basis of Flocculation; Ives, K. J., Ed.; Sijthoff & Noordhoff: Amsterdam, 1978; p 101. (32) Van der Schee, H. A.; Lyklema, J. J. Phys. Chem. 1984, 88, 6661. (33) Callow, J.; Crawford, S.; Higgins, M.; Mulvaney, P.; Wetherbee, R. Planta 2000, 211, 641. (34) Clausen-Schaumann, H.; Seitz, M.; Krautbauer, R.; Gaub, H. E. Curr. Opin. Chem. Biol. 2000, 4, 524. (35) Mitsui, K.; Nakajima, K.; Arakawa, H.; Hara, M.; Ikai, A. Biochem. Biophys. Res. Commun. 2000, 272, 55. (36) Bremmell, K. E.; Jameson, G. J.; Biggs, S. Colloids Surf. A 1998, 139, 199. (37) Kjellin, U. R. M.; Claesson, P. M.; Audebert, R. J. Colloid Interface Sci. 1997, 190, 476. (38) Lyklema, J. In The Scientific Basis of Flocculation; Ives, K. J., Ed.; Sijthoff & Noordhoff: Amsterdam, 1978; p 3. (39) Hunter, R. Introduction to Modern Colloid Science; Oxford University Press: Oxford, 1993. (40) Neihof, R. A.; Loeb, G. I. J. Mar. Res. 1974, 32, 5. (41) Loeb, G., Neihof, R. A. J. Mar. Res. 1977, 35, 283.

Received for review March 14, 2004. Revised manuscript received June 27, 2004. Accepted July 7, 2004. ES049602Z