Langmuir 2004, 20, 4093-4100
4093
An in Situ FTIR-ATR Study of Polyacrylate Adsorbed onto Hematite at High pH and High Ionic Strength Luke J. Kirwan,†,‡ Phillip D. Fawell,§ and Wilhelm van Bronswijk*,† A. J. Parker Cooperative Research Centre for Hydrometallurgy at the Department of Applied Chemistry, Curtin University of Technology, G.P.O. Box U1987, Perth, WA 6845, and CSIRO Minerals, P.O. Box 90, Bentley, WA 6982, Australia Received December 1, 2003. In Final Form: February 25, 2004 FTIR-ATR was used to examine in situ the interaction of polyacrylate and hematite at pH 13. Static light scattering and mobility measurements were used to assess solution polyacrylate dimensions and hematite surface charge, respectively. Polyacrylate adsorption occurred only with the addition of electrolyte (e.g., NaCl), and it was found that excess cations, up to approximately 1 M, facilitated adsorption, above which the effect was found to plateau. At pH 13 and at low ionic strength, adsorption of polyacrylate onto hematite is facilitated by cations in solution shielding both the negative acrylate functionality of the polymer and the negative hematite surface. The shielding of the hematite surface continues to increase with increasing salt concentration up to a measured 3 M. Similarly, the shielding of the polymer increased with electrolyte concentration up to approximately 1 M salt, beyond which no further increase in shielding was observed. At this concentration the polymer assumes a finite minimum size in solution that ultimately limits the amount adsorbed. The dimension of the polymer in solution was found to be independent of monovalent cation type. Thus, at high pH and high ionic strength adsorption is determined by the degree of hematite surface charge reduction. The cation-hematite surface interaction was found to be specific, with lithium leading to greater polyacrylate adsorption than sodium, which was followed by cesium. The stronger affinity of lithium for the hematite surface over sodium and cesium is indicative of the inverse lyotropic adsorption series and has been rationalized in the past by the “structure-making-structurebreaking” model. These results provide a useful insight into the likely adsorption mechanism for polyacrylate flocculants at high pH and ionic strength onto residues in the Bayer processing of bauxite.
1. Introduction The Bayer process for the extraction of alumina from bauxite generates large volumes of waste residue in alkaline liquors. This residue has a small particle size that contributes to its very poor settling behavior. This solid residue is removed by solid-liquid separation, which is typically achieved through flocculation in a series of gravity thickeners, with significant quantities of flocculant being consumed at each stage. The solid residue waste generally consists of silica, sodium aluminosilicates (desilication products, DSPs) and iron oxides (hematite, goethite), and is aptly named “red mud”. Flocculants used to remove the solid residue in the Bayer process are often polyacrylates or acrylate/acrylamide copolymers, and the hydroxamate functionality may also be introduced into the polymer structure. While there have been significant advances in the nature of flocculants used to treat bauxite residue, comparatively little fundamental research has been carried out to characterize the flocculant adsorption process. Refinery liquors are highly caustic, have a very high ionic strength, and contain a wide range of inorganic and organic impurities, many at much higher concentrations than those of added flocculants, which greatly complicates any experimental study. For this study, hematite was chosen as the most appropriate model substrate, as it is the dominant mineral found in Bayer residue solids in Australia. In this work we have sought * To whom correspondence should be addressed. Phone: +61 8 9266 7321. E-mail:
[email protected]. † Curtin University of Technology. ‡ Current address: Department of Inorganic, Analytical and Applied Chemistry, University of Geneva, 30 quai Ernest-Ansermet, Geneva 1211, Switzerland. § CSIRO Minerals.
to examine the fundamental aspects of how the acrylate functionality, which is dominant in Bayer-process flocculants, interacts with hematite in the adsorption process. Adsorption of polyacrylate onto hematite represents a special situation, where the surface charge of the substrate is expected to be negative due to the highly caustic environment, and where the polymer is 100% anionic; i.e., both the polyelectrolyte and substrate exhibit the same charge. This issue has been raised in studies of bauxite residue flocculation but has not been closely examined, with polymer adsorption generally being rationalized in terms of divalent cations providing a “bridge” between polymer and substrate. Calcium products are typically found in bauxite residue, and their presence has been considered necessary for flocculant adsorption at high caustic levels.1 In a comparable system Cosgrove et al.2 observed no adsorption for a negative poly(styrenesulfonate) onto a negatively charge polystyrene latex in zero added salt, but found adsorption increased with increasing salt. The increase in adsorption with ionic strength was attributed to a change in the adsorbed polymer’s conformation, with the development of loops made possible by the suppression of the intersegmental electrostatic repulsions associated with high ionic strength media; however, they did not attempt to explain the actual driving force for the attachment of the polymer to the substrate surface. Recent studies of flocculant adsorption on hematite have used settling and adsorption properties3-5 and atomic force (1) Connelly, L. J.; Owen, D. O.; Richardson, P. F. Light Met. 1986, 261. (2) Cosgrove, T.; Obey, T. M.; Vincent, B. J. Colloid Interface Sci. 1986, 111, 409. (3) Jones, F.; Farrow, J. B.; van Bronswijk, W. Colloids Surf., A 1998, 135, 183.
10.1021/la036248u CCC: $27.50 © 2004 American Chemical Society Published on Web 04/09/2004
4094
Langmuir, Vol. 20, No. 10, 2004
microscopy6 to obtain indirect information on the process, but do not provide any indication of the molecular origins of the surface interactions. Infrared spectroscopic techniques such as diffuse reflectance Fourier transform infrared (DRIFT)7-10 and polarized grazing angle (PGA)11 have been a popular choice in examination of the adsorption mode of carboxylates onto mineral surfaces because of the molecular vibrations associated with the carboxylate functionality, offering considerable potential for adsorbed species characterization. This has been demonstrated by Deacon and Phillips,12 who utilized transmission infrared spectroscopic data to determine the relationship between the carbon-oxygen stretching frequencies of carboxylato complexes and the type of carboxylate coordination. By comparing the separation between the antisymmetric and symmetric stretching frequencies of the carboxylate ion (∆COO-) bound to transition metals to that of the sodium salt, they were able to propose a set of rules for identifying the bonding mechanism in complex ions. Polymer adsorption onto mineral particles is more complex and is typically characterized by “trains” of adsorbed polymer with “loops” and “tails” of unadsorbed polymer that extend into the solution. Adsorbed segments may represent only a fraction of the total polymer chain length, and polymer tails and loops that were unadsorbed while in solution may collapse onto the surface during any washing and drying, distorting the true binding behavior. Thus, DRIFT studies of polyacrylate adsorbed on hematite, such as those by Jones et al.,7 may not be indicative of the in situ adsorbed species, as dry ex situ samples are required. FTIR-ATR spectroscopy offers the possibility of in situ examinations of the solid-solution interface and has been demonstrated in a previous paper to successfully characterize the mode of adsorption of poly(acrylic acid) onto hematite at low pH.13 This was achieved by coating the nonabsorbing infrared reflecting element (IRE) with a thin uniform film of hematite, and while internal reflection occurs at the hematite-IRE interface, radiation penetrates into the sample layer and is attenuated, allowing an absorption spectrum to be obtained. The advantage of this strategy is that information about the solid-liquid interface is obtained without any sample treatment that may change the surface characteristics. In that study it was found that the adsorbed segments of the polymer were able to be distinguished from the unadsorbed loops and tails, and that poly(acrylic acid) was chemisorbed to the hematite surface via a bidentate chelate complexation directly to a surface ferric ion. There has been little work reported on the in situ interaction of carboxylates with mineral surfaces at high pH. Specht and Frimmel14 studied the in situ adsorption of malonic and succinic acids on kaolinite over a range of (4) Jones, F.; Farrow, J. B.; van Bronswijk, W. Colloids Surf., A 1998, 142, 65. (5) Farrow, J.; Jones, F.; van Bronswijk, W. 5th International Alumina Quality Workshop, Bunbury, Australia, March 21-26, 1999; Alumina Quality Workshop Inc.: Collie, Western Australia; p 466. (6) Bremmell, K. E.; Scales, P. J.; Healy, T. W. 5th International Alumina Quality Workshop, Bunbury, Australia, March 21-26, 1999; Alumina Quality Workshop Inc.: Collie, Western Australia; p 489. (7) Jones, F.; Farrow, J. B.; van Bronswijk, W. Langmuir 1998c, 14, 6512. (8) Lee, D. H.; Condrate Snr., R. A.; Reed, J. S. J. Mater. Sci. 1996, 31, 471. (9) Parker, R. W.; Frost, R. L. Clays Clay Miner. 1996, 44, 32. (10) Gong, W. Q.; Parentich, A.; Little, L. H.; Warren, L. J. Colloids Surf. 1991, 60, 325. (11) Allara, D. L.; Nuzzo, R. G. Langmuir 1985, 1, 52. (12) Deacon, G. B.; Phillips, R. J. Coord. Chem. Res. 1980, 33, 227. (13) Kirwan, L. J.; van Bronswijk, W.; Fawell, P. D. Langmuir 2003, 19, 5802.
Kirwan et al.
pH. For the adsorption of malonic acid at pH 7.4, only the symmetric and antisymmetric stretching vibrations of the carboxylate ion were present, at positions similar to those in the unadsorbed solution species spectra. The authors proposed, in agreement with the rules outlined by Deacon and Phillips,12 a bidentate bridging complexation. It is also possible that the ionized carboxylate group of malonic acid was electrostatically attached to the kaolinite surface, as this interaction would exhibit the same stretching frequencies as those shown in the study. Steric considerations would suggest only one carboxylate is bound, with the other carboxylate unbound and deprotonated, exhibiting the same stretching frequencies associated with the bound carboxylate group. Succinic acid was found to adsorb onto kaolinite in a manner significantly different from that of malonic acid at high solution pH. At pH 6.5 and 11.5, the adsorbed species spectra developed near equal intensities for frequencies characteristic of the carboxylate ion and the carbonyl functionality. The symmetric and antisymmetric stretching vibrations of the carboxylate ion (1552 and 1396 cm-1, respectively) were at positions similar to those in the unadsorbed solution species spectra, while the carbonyl stretch (1650 cm-1) remained unshifted from that exhibited by adsorbed succinic acid at low pH. This would indicate that monodentate complexation, which was dominant at low pH, persists going from low to high pH and that one of the carboxylate groups is nonbinding and deprotonated, and hence exhibits the characteristic stretching frequencies of the carboxylate ion. Using a similar FTIR-ATR approach, Hind and Bhargava15-17 have demonstrated that the technique can be applied to high caustic, high ionic strength environments. They examined the adsorption of different chain length quaternary ammonium compounds adsorbed on solid sodium oxalate and gibbsite in synthetic Bayer liquors. However, the peak positions in the spectra of their quaternary ammonium compounds were insensitive to the adsorption state, and therefore did not lend themselves to mechanistic interpretation. FTIR-ATR can distinguish some carboxylate binding modes, but it cannot characterize substrate surface charge nor polymer size, both of which are relevant to the adsorption process. Hence, mobility measurements, to indicate the charge at the surface of shear, and multiangle laser light scattering (MALLS), for polymer size, have been used to complement this study. While MALLS is a versatile technique for the characterization of polymer solutions, as it is able to measure over an extensive range of solution pH and ionic strength, conventional electrophoretic measurements of particle mobility are restricted to relatively benign pH conditions and very low electrolyte concentration. In this study, electrophoretic and hence mobility measurements have been obtained utilizing the phase analysis light scattering principle.18 The technique enables measurements to be made in more forcing pH conditions (pH range 4-14) and in high electrolyte concentration (up to 3 M in this study). A limitation of this technique is the likely reaction of chloride ions with the palladium (14) Specht, C. H.; Frimmel, F. H. Phys. Chem. Chem. Phys. 2001, 3, 5444. (15) Hind, A. R.; Bhargava, S. K. Light Met. 2000, 65. (16) Hind, A. R.; Bhargava, S. K.; Grocott, S. C. Langmuir 1997, 13, 3483. (17) Hind, A. R.; Bhargava, S. K.; Grocott, S. C. Langmuir 1997, 13, 6255. (18) Tscharnuter, W.; McNeil-Watson, F.; Fairhurst, D. In Particle Size Distribution (III): Assessment and Characterisation; Provder T., Ed.; American Chemical Society: Washington, DC, 1998; pp 327-340.
Polyacrylate Adsorption on Fe2O3
Langmuir, Vol. 20, No. 10, 2004 4095
Table 1. Adsorbates Used in This Study adsorbate
manufacturer
mol wt
code
pimelic acid poly(acrylic acid) sodium polyacrylate
Aldrich Aldrich Ciba Specialty Chemicals
160 ∼450000 ∼13000000
PAA450K PAA13M
electrode during the electrophoretic mobility measurements, and hence, nitrate salts, which do not interfere, were chosen as electrolytes for these measurements. 2. Experimental Section 2.1. Flocculant Solutions. Details of adsorbates used are given in Table 1. Concentrated stock solutions (1000-10000 ppm) were prepared by the addition of powdered adsorbate to an appropriate aqueous solution followed by mixing on a stirrer table for several days to ensure optimal dispersion. The solution was made up to the required pH, and then stirred for a further day before the final pH was recorded. Dilutions were made from the stock solutions as required. In examining cation effects, alkalimetal polyacrylate solutions were prepared from powdered poly(acrylic acid) dispersed in water, with the pH adjusted using LiOH, NaOH, or CsOH, and the ionic strength increased by using the corresponding alkali-metal chloride. 2.2. Hematite Colloid Preparation and Deposition. Colloidal hematite was prepared from dilute FeCl3 and HCl solutions. Details of this preparation and characterization have been described previously.13 The colloid was found to have a very narrow distribution (d50 ) 0.096 µm) and to be pure hematite. A thin layer was formed on the zinc selenide crystal of the ATR accessory by carefully dispensing 30 µL of the colloid to a diameter of 5 mm and allowing the slurry to dry. 2.3. Spectral Measurements. FTIR-ATR spectra were obtained at a resolution of 4 cm-1 using a Bruker IFS 66 instrument, a Harrick “Seagull” variable-angle ATR accessory and flowcell, and an MCT (mercury cadmium telluride) detector. The number of scans accumulated depended on the intervals between acquiring spectra; 56 and 128 scans were accumulated for intervals of 30 and 60 s, respectively, while 256 scans were accumulated for intervals greater than 60 s. The incident angle of the ATR accessory was set at 45°. The experimental concept is shown schematically in Figure 1. The background measurement for all spectra was that of the clean dry ZnSe crystal. A number of backgrounds were recorded to simplify the removal of water vapor and carbon dioxide peaks from spectra that were recorded at various intervals over a number of days. After the hematite film was deposited, solutions were pumped through the cell at a constant flow rate (1.0 ( 0.02 mL min-1) using a peristaltic pump, with spectra recorded at regular intervals. Equilibrium was first established using the requisite pH reference solution, and its spectrum recorded. The procedure was then repeated with a solution containing adsorbate. The adsorbed species spectra as functions of time were obtained by subtraction of the reference spectrum from the sample spectrum. Peaks due to water vapor and carbon dioxide were removed by a scaled subtraction of the difference spectrum of two selected background spectra. This minimizes the problems associated with subtracting a single background spectrum, or using the hematite-zero adsorbate solution as the spectral acquisition background, when the relative amount of water vapor and carbon dioxide varies significantly with time. Concentrated flocculant solution species spectra were also acquired with the flowcell in the absence of the hematite colloid layer. Where peak area rather than peak intensity of spectral bands is examined, it was determined by a valley-valley peak area calculation using Bruker Opus version 2.0 software. 2.4. Surface Charge Measurements. Electrophoretic measurements were obtained at the University of New South Wales using a Brookhaven ZetaPlus instrument. Measurements were recorded at 23 °C using an applied field of 3 V and a frequency of 2 Hz. A total of three subsamples were analyzed for each sample, and the error was taken as the standard deviation. Each subsample was subject to three analytical runs by the instrument at five cycles per run.
Figure 1. Schematic representation of the FTIR-ATR experimental arrangement as used in this study, showing hematite deposited onto the ATR crystal and polymer adsorbed from a dilute solution (applied at a constant flow rate).
Figure 2. Spectrum of concentrated polyacrylate (PAA450K, 10000 ppm) in solution at pH 13. The hematite colloid was diluted to a concentration of 30 ppm, with the pH adjusted using LiOH, NaOH, or CsOH solutions and the ionic strength using the corresponding alkali-metal nitrate salt (chloride salts could not be used due to reactions with the palladium electrode during measurements). Prior to measurement, samples were vigorously shaken and then placed in an ultrasonic bath for approximately 1 min to aid particle dispersion. 2.5. MALLS. Static light scattering was carried out with a DAWN DSP photometer (Wyatt Technology) fitted with a K5 flowcell. A 1 wt % dextran solution (0.2 µm filtered) was used for normalization. PAA450K (200 ppm) and PAA13M (40 ppm) were prepared in water (0.1 µm filtered) with the pH adjusted using LiOH, NaOH, or CsOH and ionic strength using the corresponding alkali-metal chloride. Solutions were introduced into the flowcell at 10 mL h-1 through a 5 µm filter, except for PAA13M at pH 9, 11, and 12, where no filter was used due to the viscosity of the solutions. Scattered light was collected at 15 collimated detectors at fixed angles around the flowcell. Of 40-60 scans collected over approximately 30 min, approximately 10% of the lowest, most reproducible scans were kept to construct the Debye plot and subsequently determine the radius of gyration. The collection of data and determination of the radius of gyration was conducted through the use of DAWN 3.30 software, applying the Debye formulation with a fifth-order polynomial fit used in all cases.
3. Results and Discussion Ideally the commercially used PAA13M flocculant would have been used in this study; however, it is only available in its sodium salt form. Hence, a lower molecular weight poly(acrylic acid), PAA450K, which is available in its acid form, was used to study the effects of pH, ionic strength, and cation type, and PAA13M was used only to confirm the changes in molecular size at pH > 9 and in alkaline NaCl solutions. 3.1. Adsorption Mode at pH 13. Figure 2 shows the solution spectrum for PAA450K at pH 13, measured with a clean ZnSe crystal, highlighting the symmetric (1408 cm-1) and antisymmetric (1562 cm-1) stretching of the
4096
Langmuir, Vol. 20, No. 10, 2004
Figure 3. Adsorption of polyacrylate (PAA450K, 50 ppm) onto hematite at pH 13 in 1 M NaCl.
carboxylate functionality. The peak at approximately 1680 cm-1 and the asymmetry of the 1562 cm-1 peak are likely artifacts of spectrum subtraction, a consequence of less water at the hematite-solution interface, and thus a less intense 1640 cm-1 water band, in the polyacrylate solution spectrum. This is evident in Figure 4, where the 1680 cm-1 peak increases as more polyacrylate adsorbs with time. It should also be noted that the polymer concentration of 10000 ppm used in Figure 2 is several orders of magnitude greater than the values used in the adsorption studies. At the low concentrations used for the adsorption studies, the contribution to the measured spectra from the polymer in solution is insignificant. Attempts were made to achieve polymer adsorption on a hematite-coated ZnSe crystal by contacting it with a continuous flow of a 50 ppm PAA450K solution at pH 13, a concentration well below the solution detection limit. No polymer peaks could be detected even after more than 1 h of continuous contact, suggesting that no polymer adsorption was observed. This result was unaffected by variations in the solution flow rate or doubling of the polymer solution concentration. Adsorption from 50 ppm PAA450K at pH 13 was only achieved when there was an excess of sodium ions present. The pH 13 hydroxide concentration alone represents 0.1 M sodium ions, and an additional 1 M sodium (as NaCl) was required to achieve the adsorption spectra shown in Figure 3. The peak absorbance of approximately 0.013 at 1562 cm-1 for the hematite-adsorbed species spectra (Figure 3) was very similar to the absorbance of approximately 0.012 of the equivalent peak in the concentrated (10000 ppm) unadsorbed polymer species spectra (Figure 2). This highlights that the polymer was both adsorbed and concentrated onto the hematite film, and that the solution cation plays an essential role in the adsorption of the polymer’s carboxylate functionality onto hematite under caustic conditions. Most importantly, it also shows that adsorption can be achieved without the involvement of calcium19 or the need to postulate divalent cation bridging. The ∆COO-salt for the sodium salt of poly(acrylic acid) in solution was almost identical to the ∆COO-adsorbed at pH 13 with 1 M NaCl (154 and 153 cm-1, respectively). Use of the rules outlined by Deacon and Phillips12 would indicate a bidentate bridging structure for adsorption. However, such an assignment does not take into account the involvement of sodium ions, which cannot be adequately described under the Deacon and Phillips rules, as they only outline modes of coordination. Therefore, since there was no observed adsorption of negatively charged (19) Chen, H. T.; Ravishankar, S. A.; Farinato, R. S. Int. J. Miner. Process. 2003, 72, 75.
Kirwan et al.
polyacrylate to negatively charged hematite at pH 13, with no additional ionic strength, this would indicate that there is an electrostatic repulsive barrier that is too large to be overcome and facilitate adsorption. With increasing ionic strength, adsorption is seen to occur, as depicted in Figure 3. The lack of any significant spectral shifts while also taking into account the role of the sodium ions suggests that this adsorption is likely not chemisorption but rather a van der Waals interaction, with the increasing ionic strength screening the negative hematite and polymer charges, and lowering the electrostatic repulsive barrier, allowing adsorption to occur. Thus, the adsorption process could be most simply likened to the coagulation of two negatively charged particles (polymer and hematite) by the addition of electrolyte. In their DRIFT study of the adsorption of polyacrylate onto hematite at pH 13, Jones et al.7 obtained spectral evidence of asymmetric bidentate bridging complexation of the carboxylate directly to surface iron atoms. The disparity in the results can be attributed to the ex situ nature of the DRIFT technique, with the consequent collapse of the polymer loops and tails onto the hematite surface distorting the infrared derived information. The proposed van der Waals interaction at high pH found in this study is different from the mode of adsorption observed in situ for poly(acrylic acid) on hematite at pH 2,13 where the polymer chemisorbed to the hematite surface through a bidentate chelate complexation. That result was reinforced by studies of pimelic acid (HOOC(CH2)5-COOH), which structurally mimics a small segment of poly(acrylic acid), and was found to have the same mode of adsorption at pH 2 as poly(acrylic acid). Therefore, it may be expected that the pimelic acid molecule should also adsorb under similar caustic conditions as used in this study; however, no adsorption was observed at pH 13, regardless of the amount of electrolyte added. The attractive van der Waals interaction would be much smaller in magnitude due to its much lower molecular weight, and hence much lower polarizability, which is a major contributor to the van der Waals interaction. The reason it does not adsorb may also indicate that there are additional non-DVLO repulsive forces. For example, the highly structured water layer at the hematite surface, as discussed later, may present a steric barrier, which the low molecular weight molecule cannot overcome, while the high molecular weight polymer can. As the adsorbed species spectra at pH 13 in the presence of 1 M NaCl were found to be indistinguishable from the unadsorbed solution spectrum at pH 13, it was not possible to discriminate between the adsorbed and unadsorbed segments of the polymer. Such discrimination is possible for the adsorption of poly(acrylic acid) onto hematite at pH 2, for which the adsorbed carboxylate peaks exhibit a clear shift of position.13 3.2. Effect of Electrolyte (NaCl) at pH 13. Due to the irreversible nature of polymer adsorption, a separate hematite film had to be cast for each adsorption experiment. Adsorption of polyacrylate onto different hematite films cast under identical solution conditions was found to be reproducible (1560 cm-1 peak absorption (5%), implying that each cast film gave a consistent surface area for polymer adsorption, and that valid comparisons between films could be made. To define the effect of electrolyte concentration on polyacrylate adsorption onto hematite at pH 13, the spectra obtained from 50 ppm polymer solutions containing between 0.1 and 3 M NaCl were acquired as a function of time. Figure 4 shows the results for the adsorption of
Polyacrylate Adsorption on Fe2O3
Langmuir, Vol. 20, No. 10, 2004 4097
Figure 4. Adsorption of 50 ppm polyacrylate (PAA450K) onto hematite at pH 13, as a function of NaCl concentration.
Figure 5. Radius of gyration (rg) of PAA450K (200 ppm) as a function of solution pH.
PAA450K, as monitored by the area of the antisymmetric stretching peak of the carboxylate ion (1560 cm-1). The results in Figure 4 show that the adsorption of polyacrylate at pH 13 was enhanced with increasing electrolyte up to a NaCl concentration of approximately 1 M. Above this ionic strength there appears to be little or no increase in the amount of polymer adsorbed. The observed effect of electrolyte is similar to the findings of Cosgrove et al.2 for the contact between a negative poly(styrenesulfonate) and a negatively charged polystyrene latex, with adsorption only achieved through the addition of salt. They did not specify an adsorption mechanism and attributed the increase in adsorption with ionic strength to a change in the polymer adsorbed conformation only (the development of loops made possible by the suppression in the intersegmental electrostatic repulsions associated with high ionic strength media). However, it is necessary to consider not only the effect that ionic strength has on the polymer’s solution properties, but also how it influences the effective surface charge of the substrate. 3.3. Effect of Electrolyte (NaCl) at pH 13 on Polyacrylate. In salt-free solutions, sodium polyacrylates of low molecular weight are known to assume a rigid rod form, while at salt concentrations above 0.02 M the polymer’s conformation condenses to a random coil.20 Rodlike behavior may be observed at low ionic strengths (90%).26 It was postulated that a positively charged cation-dioxane complex preferentially adsorbed to the silica surface to such a significant extent as to change the potential at the shear plane and hence reverse the particle mobility. The charge reversal was not found for Li+ and Na+ ions, and it was argued that they are more strongly hydrated and hence less likely to coordinate with the dioxane. It should be noted that, for the silica case, depending on the dielectric constant of the nonaqueous solvent, it could potentially be an ion-ion correlation effect. An example of where monovalent ions have caused charge reversal in aqueous systems was reported by Rowlands et al.,27 who examined the surface properties of aluminum hydroxide in high ionic strength environments using electroacoustic measurements. In 0.5 M NaCl they found that the iep shifted from 9.1 to approximately 11, while in 3 M NaCl the measured surface potential remained positive. The authors suggested that, under such forcing conditions, while the measured trends in the ζ potential were significant, the normal picture of the diffuse double layer was no longer valid and the absolute value of the measured ζ potentials was no longer accurate. They concluded that, in alkaline, high ionic strength solution, sodium ions adsorb so close to the particle surface that the net charge at the surface where liquid flow begins to develop is positive. Similarly, the results in the present study show that, under caustic conditions, sodium ions appear to have an affinity for the hematite surface, effectively shielding and even reversing the negative charge of the hematite surface. The origin of this preferential adsorption of sodium ions onto the hematite surface is discussed in detail in the following section. The affinity of sodium for the hematite surface must also be considered in the polyacrylate adsorption mechanism. The critical added NaCl concentration to give effective charge reversal at pH 13 was in the range 0.40.5 M. At pH 13, in the absence of added NaCl, there are insufficient sodium ions to neutralize the hematite surface, (25) Kjellander, R. Ber. Bunsen-Ges. Phys. Chem. 1996, 100, 894. (26) Kosmulski, M.; Matijevic, E. Langmuir 1991, 7, 2066. (27) Rowlands, W.; O’Brien, R.; Hunter, R.; Patrick, V. J. Colloid Interface Sci. 1997, 188, 325.
Polyacrylate Adsorption on Fe2O3
Langmuir, Vol. 20, No. 10, 2004 4099
Figure 9. Adsorption of PAA450K (50 ppm) at pH 13 (LiOH, NaOH, or CsOH) and 0.2 M added electrolyte (LiCl, NaCl, or CsCl) as monitored by the area of the 1560 cm-1 peak.
Figure 10. Mobility of the hematite colloid (30 ppm) at pH 12 measured as a function of electrolyte (LiNO3 or NaNO3) concentration.
leading to a negative surface charge that inhibits the adsorption of polyacrylate. This is despite the concentration of sodium ions being sufficient to shield the negative charge of the carboxylate groups on the polymer significantly. With further increasing ionic strength, the hematite surface charge undergoes a charge reversal. It is possible that the adsorption of the polyacrylate under these conditions also has an additional electrostatic component; however, this is likely to be very short range and weak, due to the high ionic strength and the associated screening of the polymer charge. Therefore, added electrolyte changes the effective surface charge, which is the critical factor in facilitating adsorption, whereas the size of the polymer in solution, also affected by the added electrolyte, more strongly influences the extent of adsorption. 3.5. Effect of Cation Type on Adsorption. Figure 8 shows that a critical sodium ion concentration (0.4-0.5 M) exists to neutralize the hematite surface charge, and from Figure 4 it can be seen that this concentration corresponds to a substantial enhancement in polyacrylate adsorption. That being the case, it could be expected that altering the cation may influence the adsorption. The impacts of lithium, sodium, and cesium cations were therefore compared. PAA450K was used as the adsorbate because it is a poly(acrylic acid), as the sodium polyacrylate PAA13M could not be readily converted to its acid form without sodium ion contamination. The application of PAA450K to the hematite surface at pH 13 in the presence of a 1 M concentration of each added salt showed that there was no discernible difference in the adsorption densities achieved. However, when the additional electrolyte concentration was decreased from 1 to 0.2 M salt concentration, it was quite apparent that there was a significant effect attributable to the cation type, with the presence of lithium ions giving the greatest adsorption density and cesium the least (Figure 9). The impact of lithium cations on adsorption is a reflection of either greater shielding of the negative surface charge of the hematite relative to sodium or cesium cations or a stronger effect on the polymer’s solution dimensions. On the basis of the magnitude of the enhancement in adsorption, the former appears more likely. The effect that the cation type had on the ability to shield the negative moieties of the polymer was examined by measuring the radius of gyration (rg) of PAA450K at pH 13 (0.2 M salt) by MALLS. It was found that there was insignificant variation in rg for the different cations (150155 ( 15 nm), in agreement with Boisvert et al.,23 who found that the interaction of monovalent cations and polyacrylate is nonspecific and purely electrostatic. This suggests the impact of the solution cation identity is predominantly on the hematite surface charge. The mobility of hematite is most negative at approximately pH 12 (Figure 7), indicating that the charge
at the surface of shear of the hematite is also most negative at this solution condition. This condition is thus the most appropriate for evaluating the effect that the cation type has on the ability to shield the negative charge of the hematite surface. The results of electrophoretic mobility measurements carried out on dilute colloidal hematite samples that had been made up to pH 12 using either sodium hydroxide or lithium hydroxide and the corresponding electrolyte NaNO3 or LiNO3 are given in Figure 10. For both the Li+ and Na+ systems at pH 12, the measured mobility was negative in the absence of added electrolyte, and an increasing electrolyte concentration made the mobility less negative. At approximately 0.50 M LiNO3 and 0.65 M NaNO3, the mobility became zero, and beyond these electrolyte concentrations, the mobility was positive and increased in magnitude. Furthermore, LiNO3 reduced the negative surface charge of hematite to a greater extent than NaNO3. The cations clearly have a high affinity for the hematite surface, to the point that at very high salt concentrations the surface becomes positively charged. The positive mobility at high cation concentrations once again indicates a preferential adsorption of the cation to the hematite surface, with the shear plane positioned beyond these closely adsorbed cations. As the lithium cation reduces the negative surface charge of the hematite to a greater extent, it would appear to have a stronger interaction with the hematite surface than the sodium cation. This is contrary to what is classically accepted for substrates such as mercury, silver iodide, and silica, whereby increasing attractive interactions between the surface and counterion are usually associated with increasing size of the counterion. However, the result found above is not unprecedented, with the same adsorption sequence (Li+ > Na+ > K+ > Cs+, inverse lyotropic adsorption sequence) being observed by Johnson et al.28 (on alumina) and Berube and de Bruyn29 (on rutile). From observations of coagulation experiments, Amhamdi et al.30 suggested that hematite also followed the inverse lyotropic adsorption sequence. Berube and de Bruyn29 described the inverse lyotropic adsorption sequence by the model of an ionic double layer, which has become more commonly known as the “structure-making-structure-breaking” model. When a surface has a strong affinity for water, then the water molecules associated with the surface exhibit a high degree of order and structure, and are classified as “structure makers”. The authors concluded that when a surface exhibits a strong affinity for water, then strong specific adsorption (28) Johnson, S.; Scales, P.; Healy, T. Langmuir 1999, 15, 2836. (29) Berube, Y. G.; de Bruyn, P. L. J. Colloid Interface Sci. 1968, 28, 92. (30) Amhamdi, H.; Dumont, F.; Buess-Herman, C. Colloids Surf., A 1997, 125, 1.
4100
Langmuir, Vol. 20, No. 10, 2004
is going to be favored by ions that will not disrupt the structural order of the water in the surface region. Hence, small, heavily hydrated cations such as Li+ and Na+ will preferentially adsorb over larger, less hydrated cations such as K+ and Cs+. Dumont et al.31 expanded on this model, suggesting that the strength of the water-solid interaction is indicative of the degree of ordering of water molecules on the surface for a given oxide and showed that the structure-making or -breaking properties of the substrate could be predicted from the oxide’s pzc and heat of immersion. In their review, hematite (along with other metal oxides such as rutile and alumina) was classified as a structure maker due to its high pzc (8.2) and high heat of immersion (532 mJ m-2). 4. Conclusion As an in situ technique, FTIR-ATR avoids the inevitable binding changes associated with ex situ sample preparation, and its application to polyacrylate adsorption on hematite has provided evidence of how such polymers adsorb at high pH and in high ionic strength solutions. At pH 13 polyacrylate does not adsorb unless there is an excess of cations present, such as sodium, indicating that these ions play a crucial role in the adsorption of the carboxylate functionality onto hematite under caustic conditions. The spectrum of the adsorbed species at pH 13 with added electrolyte (NaCl) was found to be indistinguishable from the unadsorbed solution spectrum under the same conditions, and hence, it was not possible to discriminate between the adsorbed and unadsorbed segments of the polymer. The spectral band positions are consistent with a van der Waals interaction, with the adsorption of the negatively charged polyacrylate to the negatively charged hematite surface facilitated by the presence of sodium ions. Although they could also be taken to indicate a bidentate bridging structure for adsorption, this is unlikely as that does not take into account the role of the sodium ions. The use of monovalent sodium ions to facilitate the adsorption highlights that divalent ions are not necessarily (31) Dumont, F.; Van Tan, D.; Watillon, A. J. Colloid Interface Sci. 1976, 55, 678.
Kirwan et al.
needed to facilitate adsorption between negatively charged polymers and negatively charged surfaces. In particular, there is no need to postulate any bridging mechanism by divalent cations such as calcium, as has been proposed in other studies of bauxite residue flocculation. Adsorption of polyacrylate onto hematite at pH 13 was found to increase with increasing electrolyte concentration (NaCl) up to a concentration of approximately 1 M, where the effect plateaus. This is attributed to a decrease in polyacrylate size and charge, which leads to a potentially greater adsorption density and greater affinity for the negative hematite surface, and a reduction in the hematite negative surface charge. At pH 13, Na+ ions preferentially adsorb to the negative hematite surface, changing the effective surface charge. This is the critical factor in facilitating adsorption. At appreciably high electrolyte concentration (>0.5 M) there is charge reversal of the hematite surface charge from negative to positive. The size and apparent charge of the polymer in solution were also affected by added electrolyte (up to 1 M), having a most significant effect on the amount of polymer adsorbed. At high solution pH, the type of cation does not have a significant effect on the solution properties of polyacrylate but does have a large impact on the surface charge of the hematite, which becomes positive with increasing cation concentration and does so more rapidly for smaller more highly hydrated countercations. Hence, polyacrylate adsorption increased as a function of cation type in the series Cs+ < Na+ < Li+, corresponding to the relative affinity of the cation for the hematite surface, which is indicative of the inverse lyotropic adsorption series, and is a structure-making-structure breaking model. Acknowledgment. This research has been supported by the Australian Government’s Cooperative Research Centre program, through the A.J. Parker Cooperative Research Centre for Hydrometallurgy. This support is gratefully acknowledged. L.J.K. is grateful for the support of an Australian Research Council Postgraduate Award. LA036248U
