2114
Langmuir 2009, 25, 2114-2120
Simple Method for Controlled Association of Colloidal-Particle Mixtures using pH-Dependent Hydrogen Bonding Pierre Starck† and William A. Ducker*,†,‡ Department of Chemical and Biomolecular Engineering, UniVersity of Melbourne, ParkVille, Victoria, Australia 3010, and Department of Chemical Engineering, Virginia Tech, Blacksburg, Virginia 24061 ReceiVed August 20, 2008. ReVised Manuscript ReceiVed NoVember 24, 2008 We describe a simple method for the controlled mixing of particles that could be used to produce materials with new properties. We demonstrate the procedure with sets of silica particles that have each been coated with one of two different organic thin films. One set of particles is functionalized with carboxylic acid groups and the other with ethylene oxide. Each of these sets of particles is stable in solution. When mixed, heteroaggregation can be induced reversibly and on demand simply by changing the pH. We provide evidence that control over aggregation is achieved by the ability to alter the number of hydrogen bonds between different types of particles and thus the strength of the attraction between different particles. We provide support for this mechanism by measuring the forces between a plate coated in a thin film of carboxylic groups and particles coated in ethylene oxide using colloid probe AFM. At pH 9, where we expect most of the acidic groups to be deprotonated, there is a strong repulsion between the particle and plate. However, at pH 3, the force is attractive, which we assign to the hydrogen bonding between the ether oxygen of the PEO and the hydrogen of the carboxylic acid group. Heteroflocculation occurred in the pH range of 3-4.5. At pH 5 and above, no flocculation was observed. Because the number of hydrogen bonds per surface area and therefore the strength of binding between dissimilar particles can be titrated through control of the pH, we can control the surface forces and avoid rapid coagulation that produces low density and disorganized particle arrangements. The control of interaction forces and therefore the approach of particles should allow the production of composite materials having mixture or product properties.
Introduction The combination of two types of particles (A and B) into a single material has the potential to produce new materials with properties that are not available in a single material. Van Suchtelen1 described two types of composite materials, sum and product composites. In the sum composite, the composite material has properties that are the sum of the two particle properties, and in the product composite, a new property is developed. An example of a product composite would be to combine piezoelectric and magnetorestrictive particles to produce a magnetoelectric effect. Such new materials are highly desirable, but their preparation depends on the ability to intimately mix the two types of particles. Two basic problems emerge when mixing particles. The first is that efficient coupling of the particles is required. In the above example, this would require strong mechanical coupling between the particles (i.e., close proximity or an incompressible matrix). The second is that it is also desirable to obtain a dense mass of particles so that the properties of air or vacuum or solvent are not included in the final properties (except in the cases where this is the aim, e.g., a low dielectric material). To aid in the development of composite materials, the aim of this article is to produce a method by which two different types of particles can be mixed together into a dense composite in which each type of particle is surrounded by the other type of particle, and the separation between particle centers is sufficiently small to produce a material that is homogeneous on the required length scale of the application. The obstacle to be overcome is that when an attempt to mix particles is made, the intermolecular interactions that are intrinsic to the particles often interfere with * To whom correspondence should be addressed. E-mail:
[email protected]. † University of Melbourne. ‡ Virginia Tech. (1) van Suchtelen, J. Philips Res. Rep. 1972, 27, 28–37.
both mixing and densification. For example, because van der Waals forces between particles depend on the product of polarizabilities, van der Waals forces tend to cause demixing of particles.2 The most obvious remedy to this problem is to immerse the particles in solvent (e.g., water) to reduce the van der Waals force and to seek conditions in which they have opposite charge. Then, the like particles will repel each other and particle A will be attracted to particle B. However, if the attractive force between the particles is large, then the particles will rapidly coagulate and form attractive networks that are trapped in conformations that are not at maximum density and difficult to mold.3-5 The aim here to obtain more subtle control over the interactions between the particles, preferably with a weaker and more tunable force, and to use water as the solvent so as to minimize the environmental impact. To this end, we note that the biological world is full of examples where (relatively weak) hydrogen bonds are used to create order that is reversible, and that hydrogen bonds have been used to control the order in particle mixtures. Of particular interest is an article by Starck and Vincent6 in which hydrogen bonding was used to coat particles A with particles B to control the rheology of suspensions. The method of achieving this coating was to produce latex particle A with hydrogen-bonding donors on the surface and latex particle B with hydrogen-bonding acceptors on the surface. The donors were the carboxylic acid groups on poly(acrylic acid) (PAA), and the acceptors were the ether groups on poly(ethylene oxide) (PEO), as shown in Figure 1. The hydrogen bonding between the two groups should lead to an (2) Israelachvili, J. Intermolecular and Surface Forces, 2nd ed.; Academic Press: San Diego, CA, 1991. (3) Velamakanni, B. H.; Chang, J. C.; Lang, F. F.; Pearson, D. S. Langmuir 1990, 6, 1323–1325. (4) Ducker, W. A.; Luther, E. P.; Clarke, D. R.; Lange, F. E. J. Am. Ceram. Soc. 1997, 80, 575–583. (5) Ducker, W. A.; Clarke, D. R. Colloids Surf., A 1994, 93, 275–292. (6) Starck, P.; Vincent, B. Langmuir 2006, 22, 5294–5300.
10.1021/la8027258 CCC: $40.75 2009 American Chemical Society Published on Web 01/15/2009
Controlled Association of Colloidal-Particle Mixtures
Langmuir, Vol. 25, No. 4, 2009 2115
Figure 1. Hydrogen bonding between carboxylic acid and ether groups.
Figure 3. Surface chemistry of silane-modified particles and AFM substrates.
Figure 2. Schematic of the mixing procedure.
attractive force between the two different particles and was observed to lead to control of the rheology.6 The pH-controllable hydrogen bonding between carboxylic acid groups on poly(methacrylic acid) and PEO has also been used to make selectively dissociable thin films.7 The basic scheme for the proposed procedure is shown in Figure 2. The aim is to have a flocculated mixture of two different particles that is well mixed. This requires that the force be controllable: particles A must attract particles B only under certain conditions. For example, PEO-coated particles attract COOHcoated particles at low pH. The attraction causes PEO-coated particles to be coated in COOH particles. Alternately, the particles can be mixed at high pH and then “cross-linked” by lowering the pH. If the force is made only weakly attractive (i.e., only a few hydrogen bonds per particle, pH 4.5-5) then the particle mixture will still be workable and can also be made dense. The first aim of this article is to demonstrate that we can obtain good control of the properties of particle mixtures by coating the particles with thin organic films that lead to controllable forces between the particles. We define good control as the ability to cause heterocoagulation on demand, for the attraction to be reversible, and for the strength of the attraction to be easily tunable (e.g., by pH, salt, temperature changes). Following the design of Starck and Vincent, we coat one particle type with an ether and the other with a carboxylic acid, and we use the same combination of PAA and PEO. The second aim of the article is to demonstrate that ethers form attractive interactions with carboxylic acids and thereby to validate the mechanism. We will test this with the atomic force microscope (AFM) colloid probe technique.8,9 Previous AFM force measurements between (7) Sukhishvili, S. A.; Granick, S. J. Am. Chem. Soc. 2000, 122, 9550–9551. (8) Ducker, W. A.; Senden, T. J.; Pashley, R. M. Nature 1991, 353, 239–241. (9) Ducker, W. A.; Senden, T. J.; Pashley, R. M. Langmuir 1992, 8, 1831– 1836.
different functional groups were reviewed separately by Butt10 and Noy.11 Previous measurements investigated symmetrical interactions for -NH2, -CH3, and -COOH functionalized solids in water9,12-14 and -OH, -COOH, and -CH3 terminated samples and tips in different liquids9,15,16 The most relevant work for the current study is that of Hammond and co-workers, who concluded that, between ethylene oxide groups and carboxylic groups of SAM surfaces, there is only a very weak adhesion force (2-5% of that for other hydrogen-bonding interactions) at pH 2.5 and no adhesion at pH 4.5 and 7.12 We examined silica particles because of the ready availability of silica substrates for surface chemical experiments. Silica has surface silanol groups that dissociate in aqueous solution to impart a negative charge and therefore double-layer stabilization to the particles.17 The point of zero charge is usually in the range pH 2-4. Thus, silica particles are usually stable when the pH is above about 3-4, and the stability decreases with increasing salt concentration,18 but there are notable exceptions to this behavior.19 We coated the silica particles using silane coupling because silanes can be coupled to any material that exhibits surface hydroxyls, which is a common surface group in water. In fact, multivalent silanes can also be used to coat materials without surface hydroxyls.20 Schematics for the chemistries of the coated particles are shown in Figure 3. A polymeric carboxylic acid was used so as to achieve better interparticle adhesion by exploiting the greater conformation freedom of the polymer over an inflexible short tether to the solid. In this study, we used a variety of materials that all have silica-like interfaces. For the AFM measurements, we used glass particles and silicon wafers. The silicon wafers have a native layer of silicon oxide. For particle stability and adhesion measurements, we used colloidal silica particles. Throughout (10) Butt, H.-J. Surf. Sci. Rep. 2005, 59, 1–152. (11) Noy, A. Surf. Interface Anal. 2006, 38, 1429–1441. (12) Jiang, X.; Ortiz, C.; Hammond, P. Langmuir 2002, 18, 1131–1143. (13) Kidoaki, S.; Matsuda, T. Langmuir 1999, 15, 7639–7646. (14) Frisbie, C. D.; Rozsnyai, L. F.; Noy, A.; Wrighton, M. S.; Lieber, C. M. Science 1994, 265, 2071–2074. (15) Papastavrou, G.; Akari, S. Colloids Surf., A 2000, 164, 175–181. (16) Warszynski, P.; Papastavrou, G.; Wantke, K. D.; Mohwald, H. Colloids Surf., A 2003, 214, 61–75. (17) Verwey, E. J. W.; Overbeek, J. T. G. Theory of the Stability of Lyophobic Colloids; Elsevier: New York, 1948. (18) Kobayashi, M.; Skarba, M.; Calletto, P.; Cakara, D.; Borkovec, M. J. Colloid Interface Sci. 2005, 292, 139–147. (19) Allen, L. H.; Matijevic, E. J. Colloid Interface Sci. 1969, 31, 287–296. (20) Schwartz, D. K.; Steinberg, S.; Israelachvili, J.; Zasadzinski, J. A. N. Phys. ReV. Lett. 1992, 69, 3354–3357.
2116 Langmuir, Vol. 25, No. 4, 2009
the discussion of results, we assume that differences in behavior between these interfaces, once coated in silane, are negligible.
Experimental Section Materials. All solutions were prepared in water purified by an Easypure UV system (Barnstead) with charcoal, deionizing, UV light, and 0.2-µm filter stages. The water has a specific resistivity of 18.2 MΩ cm and a surface tension of 72 mJ m-2. Sodium chloride (NaCl, 99.99%) supplied by Aldrich was baked overnight in an oven at 500 °C to remove organic contamination. Toluene, ethanol, and 2-propanol were obtained from BDH. Toluene was initially dried on 4 Å molecular sieves (Ajax), distilled over small pieces of sodium, and then stored on 4 Å molecular sieves before use. 3-Aminopropyltriethoxysilane (APTES, 99%) was obtained from Sigma-Aldrich and 2-[methoxy-(polyethyleneoxy)6-9-propyl]-trimethoxysilane (SiPEG, MW of 460-590, purity >90%) was purchased from Gelest. Both silanes were used as received. PAA (MW ≈ 8000), 2-morpholinoethanesulfonic acid (MES) buffer, and 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) were obtained from Sigma Aldrich. Grafting of APTES to Silicon Wafers. A silicon wafer was washed twice in water and sonicated in ethanol/water (1:1) for 5 min, dried with nitrogen, and then cleaned and oxidized with “piranha” solution (30% H2O2 and concentrated H2SO4, 1:3 vol; caution: strong oxidizing agent) for 30 min. The wafer was then washed twice with water and once with ethanol, blown dry with nitrogen, and heated in an oven at 120 °C for 2 h. The wafer was immediately immersed into the APTES silane solution. The grafting was performed in a sealed beaker containing the silane solution, APTES, 0.5 vol %, in toluene for 18 h at room temperature. The surface was washed and sonicated in toluene, ethanol, and then water. Grafting of PAA to APTES-Functionalized Wafers. The APTES-silicon wafer was immersed in a 3 mg mL-1 PAA solution at pH 5.5 and 0.5 mM NaCl solution for 30 min. The PAA/APTES layer was then chemically cross-linked with a solution of EDC (20 mg mL-1) in MES buffer (0.05 M, pH 5.5) to form amide bonds between amines and acid groups in the film.21 The reaction mixture was shaken for 1 h, and then the wafer was rinsed several times with water. Preparation of PEG-Coated AFM Colloidal Probe. A silica sphere (r ≈ 10 µm) was glued to a silicon nitride cantilever using epoxy glue. The sphere on cantilever was then cleaned under a UV lamp for 30 min and then put directly in a Si-PEG solution, 0.5% (v/v), in toluene for 18 h at room temperature. The sphere-cantilever was then washed carefully in toluene, ethanol, and water. Grafting of PEG and APTES to Silica Particles. The silica micropowder used in this study (Fuso Chemical Co. Ltd.) consisted of nearly monodisperse spherical particles of diameter 0.30 µm (range: 0.27-0.34 µm), with a density of 2200 kg/m3, a maximum moisture of 0.5%, and a surface area of 12.15 ( 0.05 m2/g. The average hydrodynamic diameter of the silica particles was determined using dynamic light scattering (Malvern Instruments, HPPS). Silica particles were modified with either APTES or Si-PEG in a solvent-phase reaction following a method described by Jaroniec et al.22 In this reaction, 5-g silica particles were dispersed in 400 mL of dry toluene in a round-bottom flask under nitrogen. The silane (0.8 mL) was added to the solution, and the mixture was refluxed for 18 h. The sample was then cooled to ambient temperature, filtered on Buchner funnel, washed with dry toluene, heated overnight in a vacuum oven at 100 °C, and finally stored in a desiccator before further use. Grafting of PAA to APTES-Functionalized Particles. Silica particles modified with APTES were then redispersed in water at 2% vol and mixed with a PAA solution (10 mg mL-1) at pH 5.5 for an hour, centrifuged, and then cross-linked with a solution of EDC (20 mg mL-1) in MES buffer (0.05 M, pH 5.5) to form amide bonds as explained above. (21) Lahiri, J.; Isaacs, L.; Tien, J.; Whitesides, G. M. Anal. Chem. 1999, 71, 777–790. (22) Jaroniec, C. P.; Kruk, M.; Jaroniec, M.; Sayari, A. J. Phys. Chem. B 1998, 102, 5503–5510.
Starck and Ducker Mixtures of Particles. Mixtures of functionalized silica particles (PAA and PEO) were prepared from equal masses of 2% vol solution, each in 0.01 M NaCl, and shaken for 5 min. The samples were left to equilibrate for at least 30 min before pictures were captured using a digital camera (Olympus). For the experiments performed at 80 °C, the samples were left to equilibrate for 30 min in the bath. The photographs were taken immediately after removal from the water bath. Colloidal Probe AFM Force Measurements. An Asylum Research MFP-3D AFM was used to measure the forces between PEO-functionalized 20-µm-diameter silica particles glued to a silicon nitride tip and a PAA-functionalized silicon wafer in aqueous salt solutions, 0.1 M NaCl, and two different pH values (3 and 9). The basic method is described by Ducker et al.8,9 The AFM measures the change in endslope of a cantilever while the cantilever is translated normal to the solid-liquid interface. The translation is measured using a linear variable differential transducer. We convert the endslope to the deflection as described previously. The separation is the sum of the deflection and the translation. The AFM probes (NP, Veeco, silicon nitride, square pyramid tip, nominal stiffness of 0.12 N/m) were cleaned under ultraviolet light before use. Each spring constant was measured using the thermal method before the sphere was glued onto the cantilever and was found to be in the range 0.10-0.12 N/m. Fluid exchange made use of a custom designed open fluid cell, consisting of an Asylum cantilever mount modified to hold two lengths of PEEK tubing (Valco Instruments). This enabled the exchange of fluids within a liquid capillary formed between the surface and the cantilever mount. Tapping-Mode AFM Imaging of the PAA Surface in the Presence of PEO Particles. To measure the interaction between particles and plates, we used tapping-mode AFM to image the surface of the PAA plate after contact with a solution of PEO particles (1% vol) at 0.01 M NaCl and at pH 9 and 3. The plate was then rinsed with a salt water solution at the same pH as the particles. The measurements were also made in salt water at either pH 9 or 3. The drive frequency of the cantilever was set at 10 kHz, and the drive amplitude was 200 µV. During imaging, the set point ratio (A/A0) was usually above or at 0.9 where A0 is the free amplitude of the cantilever and A is the set point amplitude used during imaging. X-ray Photoelectron Spectroscopy. XPS analysis was performed using an AXIS-HSi spectrometer (Kratos Analytical, Inc.) with a monochromated Al KR source, a hemispherical analyzer operating in the fixed analyzer transmission mode, and the standard aperture (1 mm × 0.5 mm). The total pressure in the main vacuum chamber during analysis was of the order of 10-8 mbar. Spectra were recorded at an emission angle of 60° with respect to the surface normal, which corresponds to a depth of penetration of about 5 nm. All elements were identified from survey spectra acquired at a pass energy of 320 eV. To obtain more detailed information about chemical structure, high-resolution spectra were recorded from individual peaks at 40 eV pass energy, yielding a typical peak width of 1.0-1.1 eV. These data were quantified using a minimization algorithm to determine the contributions from specific functional groups. The atomic concentrations of the detected elements were calculated using integral peak intensities and the sensitivity factors supplied by the manufacturer. Peak assignments are based on the measured binding energy values charge-corrected with respect to the aliphatic hydrocarbon peak at 285.0 eV.
Results Characterization of Modified Silicon Wafers. The efficacy of procedures for modifying the plates and particles was examined by preparing films on a silicon wafer (with a native layer of silicon dioxide) and measuring the surface composition using XPS. The XPS spectra of silane-modified silicon wafers are shown in Figure 4. Peak assignments were made using the NIST XPS database.23 The adsorption of the amino silane is demonstrated by the N1s peak at 400 eV. The conversion of some of these groups to amides by reaction with PAA is shown by the shift
Controlled Association of Colloidal-Particle Mixtures
Langmuir, Vol. 25, No. 4, 2009 2117
Figure 5. Force between PEO-functionalized glass sphere and the COOH-silica plate as a function of pH in 0.1 NaCl. The pH 9 curve has been shifted in force for clarity and is also shown on an enlarged scale in the inset.
Figure 4. High-resolution XPS spectra of silicon plates with various coating films. N1s: The N1s is absent on both the native oxide and PEG-modified wafer and present on both wafers that were reacted with the amine silane. After deposition of PAA and EDC coupling, there is a shift to higher binding energy showing the formation of the amide bond. C1s: The adsorption and covalent attachment of PAA produce a new peak that is characteristic of carbonyl groups. The adsorption of the PEG silane is shown from the new ether peak.
of this peak to higher frequency. The adsorption of PAA is also demonstrated by the C1s shoulder at about 288 eV, and the adsorption of the PEO film is demonstrated by C1s peak at 286.5-287 eV. Force between Surfaces Can Be Controlled through pH. Forces between a PEO-coated silica sphere and a PAA-coated silicon plate in aqueous 0.1 M NaCl are shown in Figure 5. There is a repulsive force and no adhesion at pH 9, whereas there is an attractive force and a large adhesion at pH 3, which is consistent with the hypothesis that the PEO on the particle hydrogen bonds to the PAA on the plate at pH 3, but not at pH 9. An alternative hypothesis is that the attractive force is a van der Waals force and that control of the force is achieved by alteration of the double-layer force through changes in surface potential. It is clear that this is not the case because, if the attractive force were a van der Waals force, we would still expect the attractive force to be present at pH 9, which is not the case. On approach, the attractive force at pH 3 extends out to 14 nm, which is further than the repulsive force at pH 9. Thus, the repulsive force cannot mask the attractive force, and the attractive force must have changed. These measurements were performed at 0.1 M NaCl, where the double layer forces have very short range (Debye (23) NIST X-ray Photoelectron Spectroscopy Database. NIST Standard Reference Database 20, version 3.5. http://srdata.nist.gov/xps/ (accessed Jan 2009).
length 0.95 nm) compared to that of the measured forces; this allows the hydrogen-bonding force to be clearly resolved. Furthermore, the measured attractive force is much greater than the expected van der Waals forces. Lifshitz calculations reveal that the van der Waals force between amorphous silica solids is approximately described by a Hamaker constant in the range (0.46-0.8) × 10-20 J.24,25 Force measurements between silica surfaces under similar conditions confirm that this is an accurate estimate of the van der Waals force.26 The addition of a watery organic layer over the silica will reduce the van der Waals force but the silica-silica force can still be used as a guide. For example, using Bergstrom’s value of the Hamaker constant,24 if the attractive force at pH 3 were due to a van der Waals force alone, then for a spring constant of 0.1 N m-1 there should be an instability at about 2.4-nm separation. In fact, the instability on approach occurs at about 14 nm, showing a longer ranged and stronger force. Likewise, when the solids are separated, there is still an attractive force of almost 1 mN m-1 at 20-nm separation, whereas the van der Waals force between silica surfaces is about 0.002 mN m-1. In other words, the measured forces differ from the van der Waals force in range, magnitude, and pH dependendence. The observed forces are explained by a combination of doublelayer, polymer-entropic, and hydrogen-bonding forces. Potentiometric titrations on poly(acrylic acid) in solution show that ionization increases gradually from pH 3 to 10, with about half the acid groups dissociated at pH 5.5-6.5.27,28 When a high density of PAA is adsorbed to silicon nitride27 or rutile29 colloidal particles, the surface charging of the particles becomes similar to that of PAA. Likewise, we expect that the PAA-coated silica will have a very small charge at pH 3 and a large charge at pH 9. The absence of acidic or basic groups on the PEO means that the PEO-coated glass will have little or no charge at both pH 3 and 9. The large charge on the PAA at pH 9 explains the large repulsive force that is measured at pH 9: it is a combination of electrostatic double-layer repulsion caused by the charged (24) Bergstrom, L. AdV. Colloid Interface Sci. 1997, 70, 125–169. (25) Hough, D. B.; White, L. R. AdV. Colloid Interface Sci. 1980, 14, 3–41. (26) Valle-Delgado, J. J.; Molina-Bolivar, J. A.; Galisteo-Gonzalez, F.; GalvezRuiz, M. J.; Feiler, A.; Rutland, M. W. J. Chem. Phys. 2005, 123, 034708/1– 034708/12. (27) Hackley, V. A. J. Am. Ceram. Soc. 1997, 80, 2315–2325. (28) Nagasawa, M.; Murase, T.; Kondo, K. J. Phys. Chem. 1965, 69, 4005– 4012. (29) Gebhardt, J. E.; Fuerstenau, D. W. Colloids Surf. 1983, 7, 221–231.
2118 Langmuir, Vol. 25, No. 4, 2009
carboxylate groups and repulsive steric interaction that arises from confinement of the charged PAA chains to a smaller volume. The hysteresis between the approach and retract curves shows that the PAA molecules relax more slowly than the time scale of the measurement (∼3 s). The force shows some hysteresis because the polymer film is slow to equilibrate. The repulsive force barrier explains the lack of heterocoagulation at pH 9 that is described later. At pH 3, the repulsive double-layer force is absent because almost all the carboxylate groups are discharged through bonding to protons. One would also expect that the uncharged PAA, on average, is closer to the solid than at pH 9 because of reduced repulsion between neighboring segments. This will reduce the range of any polymer entropic repulsion. We explain the apparently long-range attractive force at separations up to about 14 nm as follows. The PAA segments adopt a distribution of displacements from the solid. When a COOH group on one segment encounters an oxygen atom from the PEO attached to the glass particle, a hydrogen bond can form, thereby increasing the density of ethylene oxide monomers near the carboxylic acid-coated silica. This change in distribution of the segment results in an attractive force between the solids because the polymer has been stretched beyond its equilibrium conformation in the absence of the second solid. This effect is known as a polymer bridging force.30 The short-range hydrogen bond acts like a long-range attractive force because it is attached to a flexible tether. The aVerage length of the PAA in the alltrans conformation is 27 nm, so the attraction at 14 nm is well within the range of the polymer. When the solids are pulled or pushed together, many segments come in range of hydrogen bonds, so there is a strong adhesion. The measured force is characteristic of a polymer bridging force because the attractive force increases with separation (i.e., as the bridging polymers are stretched). There are also discontinuities in the retract curve that arise as individuals, or groups, of polymer chains are stretched beyond the adhesive strength of the hydrogen bond interaction and the hydrogen bond or bonds break. In Figure 5, these occur in the retract curve at 20 and 80 nm. These discontinuities suggest that some polymer chains are up to 80-nm long (i.e., up to three times the length of the average molecule). This distance is much greater than the average chain length because of the polydispersity in molecular mass of the PAA. The polymer bridging force that we observe is qualitatively similar to bridging force that has been measured previously when PAA is made to bridge between mica solids in the presence of divalent cations.31 Given these measured forces, we would expect mixtures of particles coated in PEO with particles coated in PAA to be stable at pH 9 but to undergo rapid heteroflocculation at pH 3. In fact, exactly this behavior was observed, as described in the next section. Note that earlier results from the Hammond group are in conflict with the results presented here.12 In a study of gold-coated surfaces, the Hammond group found that there was only very small adhesion (0.08 nN or less than 1% of the value measured here) between an ethylene oxide and a carboxylic acid-terminated surface. One hypothesis is that, in Hammond’s work, the adhesion is weaker because there are fewer hydrogen-bonding sites than in our work, only one per alkyl chain. However, this is not supported by their report of hydrogen bonding between a carboxylic acid-coated surface and amine- or carboxylic acid(30) Ji, H.; Hone, D.; Pincus, P. A.; Rossi, G. Macromolecules 1990, 23, 698–707. (31) Abraham, T.; Kumpulainen, A.; Xu, Z.; Rutland, M.; Claesson, P. M.; Masliyah, J. Langmuir 2001, 17, 8321–8327.
Starck and Ducker
Figure 6. Force between a PEO-functionalized glass sphere and an amine-functionalized plate at pH 3 and 9. Table 1. Average Values of 10 Measurements of Force between Silane-Coated Spheres and Platesa pH 3 9 3 9 3 9
sphere PEO PEO SiOH SIO-
plate
adhesion (mN/m)
jump-in distance (nm)
COOH COONH3+ NH2/NH3+ COOH COO-
1.1 ( 0.3 no adhesion 0.4 ( 0.1 0.27 ( 0.02 0.05 ( 0.02 0.04 ( 0.01
23 ( 6 no jump in 8(3 12.5 ( 1.7 no jump in no jump in
a The error is the standard deviation of about 10 measurements and does not include systematic errors arising from the spring constant and radius. The structure shown indicates the predominant structures at the interface; the films are always mixtures of different charged states.
coated surfaces. Another hypothesis is that the reduced conformational freedom of thiol-monolayers limited the ability of the layers to organize into hydrogen-bonding conformations. However, we measure a strong adhesion between an amine monolayer and a PEO monolayer (as expected for a donor-acceptor combination), whereas Hammond’s group measures very weak or no adhesion for the same pair of groups, but with the amine present as linear poly(ethylene imine). These earlier results are hard to reconcile with the fact that ethers are known to be strong hydrogen bond acceptors32 and carboxylic acids and amines have hydrogen bond donors. In contrast, our observation of strong adhesion is consistent with both the observed instability of particle mixtures and adhesion of PEO particles to hydrogen donormodified silicon wafers. In contrast, if we do not attach PAA to the amine-functionalized plate, the force is attractive over the pH range 3-9, with a large adhesion (Figure 6). The amine group has hydrogen-bonding donors at both pH 3 and 9 and thus can always bind to the lone pair electrons on the oxygen on PEO providing that the groups are in hydrogen-bonding range. Close inspection of the approach curves in pH 3 shows a steep repulsive force at about 3 nm, which is probably due to electrostatic repulsion. There is some variation in the force measured between particles and plates, so it is useful to present data that is the average of many measurements, and this is shown in Table 1. These data show the same trends as above. The PEO-functionalized tip selectively binds to the PAA surface at low pH where the carboxylic acid groups have hydrogen-bonding donors. The PEOmodified probe always binds to the amine surface in the range of pH studied because it always has hydrogen-bonding donors. (32) Vinogradov, S. N.; Linnell, R. H. Hydrogen Bonding; Van Nostrand Reinhold: New York, 1971; p 121.
Controlled Association of Colloidal-Particle Mixtures
Langmuir, Vol. 25, No. 4, 2009 2119
Figure 7. AFM images of a PAA-modified plate that was exposed to PEO-coated particles at pH 9 and 3.
Figure 8. Photograph of particle suspension. At pH 4.5, a suspension of PAA particles is stable; likewise, the suspension of PEO particles is stable at pH 3.8. When the suspensions are mixed at pH 4.5, the particles quickly flocculate and sediment. The sediment is visible at the bottom of the vial When the pH of the mixture is increased to 7, and the suspension is shaken, there is no sediment and the suspension is stable.
The adhesion to the amine film is consistently stronger at pH 3 because the N-H bond is more polarized in -NH3+ than in -NH2. We also performed a control experiment where the glass sphere was not functionalized with PEO. There is only a weak adhesion at pH 3 because the oxygen in SiOH is a worse hydrogenbonding acceptor than the oxygen in PEO. Adhesion of Particles to Plates Can Be Controlled through pH. Additional evidence for the pH-dependent attraction between particle and plate was obtained from particle adhesion studies. The PAA-coated plates were dipped into a suspension of 0.3-µm PEO-coated particles, which was used for particle stability studies, and then rinsed thoroughly in 0.01 M NaCl solution. The particles were then imaged by AFM (Figure 7). Consistent with the force measurements, at high pH, where no adhesion was measured, the plate is completely free of particles, and at low pH, where AFM showed the strong adhesion, there are many particles adhering to the plate. Aggregation of Functionalized Particle Mixtures Can Be Controlled through pH. The silica particles are stable for days when coated with a thin film of PAA or PEO at about pH 4 and 0.01 M NaCl (Figure 8). The PEO particles are stabilized by steric repulsion of the highly hydrated PEO chains. At pH 4.5, the PAA molecules have a mixture of COOH and COO- groups.
There are sufficient COO- groups to stabilize the particle suspension through double-layer repulsion and also a contribution from steric repulsion. When the particles are mixed in equal quantity the particle mixture immediately flocculates and then sediments (Figure 8). This shows that there is no longer a significant activation barrier between particles and that a strong adhesion occurs (consistent with Figure 5). The stability of the suspensions of pure particles and the instability of the mixture prove that the different particles adhere at pH 4.5 and that we must be observing heteroaggregation. The simple explanation is that there are sufficient COOH groups to hydrogen bond to the PEO as shown by the surface force measurements. The doublelayer repulsion caused by the interaction between the COOgroups and the neutral particles is, by experiment, insufficient to stabilize the particles. This is consistent with the surface force measurement that showed that, when the particle approached the plate, an attractive bridging force extended, on average, 23 nm from the particle surface. In contrast, the characteristic length of the double-layer force, the Debye length, is only 3 nm in the suspension. Therefore, the double-layer force does not have sufficient range to screen the polymer hydrogen bond bridging force. In contrast, when the two types of particles are mixed at pH 7, the suspension is stable, because there are few undissociated COOH groups remaining on the PAA to form hydrogen bonds and a larger electrostatic double-layer repulsion. Note also that, in all photographs, there is little adhesion of particles to the glass vials at pH 7. At this pH, the glass is negatively charged. Neither type of particle is positive, so adhesion should rely on hydrogen bonding. At pH 7, there are few hydrogen-bonding donors on either particle type so there is no adhesion. Figure 9 shows that the properties of the particle mixture are reversible with respect to pH. If we mix the particles at pH 2.3, they immediately coagulate. Resuspension of the particles can be easily achieved by raising the pH of the suspension to 7. Of course, the particles flocculate again when the pH is decreased a second time. Coagulation is also reversible when the pH starts at pH 4.5. This reversibility supports the idea that the particle attraction is caused by hydrogen bonding and is a very useful property for controlling suspension properties. Figure 10 shows the influence of pH with greater resolution. As acid is added to decrease the pH, flocculation begins between pH 5 and 4.5 and is mainly complete by pH 3.8. This is consistent with the reported pK of poly(acrylic acid): about half of the acid groups are protonated at pH 6 and the remainder become
2120 Langmuir, Vol. 25, No. 4, 2009
Starck and Ducker
flocculation, whereas the mixture is stable at pH 7. Thus, the heterocoagulation method can be used over a range of temperatures.
Discussion The objective of particle mixing should be considered again in light of these results. The aim is to satisfy the following criteria in particle mixtures: (1) mixing of two different particle types such that each particle A is near particle B and (2) the formation of a dense particle mixture. As discussed in the Introduction, it is sometimes difficult to achieve these criteria simultaneously because if the force is attractive, then rapid coagulation occurs and a low density structure is formed. If one set of particles is coated in PEO and the other is coated with carboxylic acid groups, then a high degree of control over the particles is achieved. The particles can be mixed at pH 7, where all the interactions are repulsive, so that the particles can be purified by filtration. The pH can be slowly decreased to coagulate and sediment or densify the suspension. This allows time for the particles to obtain maximum density. The densification occurs through the combination of two types of particles, achieving the primary objective. The organization of the final mixture can be controlled though the pH. At higher pH (4.5-5), the suspension is slower to coagulate and the particles have more time to achieve maximum density. Alternately (not shown here), the particles can be pressure filtered at high pH to cause densification. Only a small amount of acid needs to be added (pH change from 5 to 4) and only a small ion (H+) needs to diffuse to the particle surface to produce an attractive network. The approach of pressure filtering then decreasing the pH would result in a more random mixture, without such a strong bias for each particle to be in close proximity to particles of the other type.
Conclusions Figure 9. Photographs of suspensions showing that the particle mixture can be resuspended by increasing the pH and then coagulated by decreasing the pH (i.e., the suspension properties are reversible after coagulation at pH 2.3 or 3.8).
Figure 10. Photograph of suspensions of PAA-PEO mixtures at pH 2.3, 3.8, 4.5, 5, and 7.
protonated over the range pH 6 to 4.28 The protonation decreases the double-layer force and allows hydrogen bonds to form. Suspension Properties Are Similar at 25 and 80 °C. Hydrogen bonding is normally weaker at higher temperatures because the formation of most hydrogen bonds is associated with negative entropy.32 If flocs are held together by hydrogen bonding, then one might expect weaker flocculation. In fact, the flocculation at 80 °C is very similar to the behavior at 25 °C: when the ethylene oxide-coated particles are mixed with the carboxylic acid-coated particles at pH 4, there is immediate
The results in this article describe a method for adhering together two types of particles. In the method, one particle should be coated in an ethylene oxide layer and the other should be coated in a carboxylic acid layer. Each of these sets of particles is stable in solution over a wide range of pH and salt concentrations. This allows for purification of the particles by filtration. The mixture of particles is also stable at neutral pH, which allows good mixing of the particles. When the pH is decreased, the particles undergo heterocoagulation due to hydrogen bonding between the ethylene oxide and the carboxylic acid groups. This hydrogen-bonding mechanism was demonstrated by direct force measurement. The number of hydrogen bonds per surface area can be titrated though control of the pH, and therefore the strength of binding of one particle to another can be controlled. This allows one to avoid rapid coagulation that produces low density, disorganized sediment or to achieve rapid coagulation if required. Many recipes exist for coupling ethylene oxide and carboxylic acids to interfaces, so this technique should have broad applicability. By using particles that are smaller than the length scale of property control (e.g., smaller than the wavelength of light for optical properties), this method should allow production of composites that can have mixture or product properties. Acknowledgment. We thank Thomas Gegenback for measuring the XPS spectra and assigning the peaks. This research was supported under the Australian Research Council’s Discovery Project funding scheme (DP0664051). W.D. is the recipient of anAustralianResearchCouncilFederationFellowship(FF0348620). LA8027258
