3418
Langmuir 2009, 25, 3418-3424
Hydrogen Bonding and Interparticle Forces in Platelet r-Al2O3 Dispersions: Yield Stress and Zeta Potential Kay-Sen Khoo, E-Jen Teh, and Yee-Kwong Leong* Chemical and Process Engineering Program, School of Mechanical Engineering, Centre for Strategic Nano-Fabrication, The UniVersity of Western Australia, Crawley, Perth, Western Australia 6009, Australia
Ban Choon Ong School of Life Sciences and Chemical Technology, Ngee Ann Polytechnic, Singapore 599489 ReceiVed NoVember 2, 2008. ReVised Manuscript ReceiVed December 20, 2008 Adsorbed phosphate on smooth platelet R-Al2O3 particles at saturation surface coverage gives rise to strong interparticle attractive forces in dispersion. The maximum yield stress at the point of zero charge was increased by 2-fold. This was attributed to a high density of intermolecular hydrogen bonding between the adsorbed phosphate layers of the interacting particles. Adsorbed citrate at saturation surface coverage, however, reduced the maximum yield stress by 50%. It adsorbed to form a very effective steric barrier as intramolecular hydrogen bonding between -OH and the free terminal carboxylic group prevented strong interactions with other adsorbed citrate molecules residing on the second interacting particle. This steric barrier kept the interacting platelet particles further apart, thereby weakening the van der Waals attraction. The platelet R-Al2O3 dispersions were flocculated at all pH level. These dispersions displayed a maximum yield stress at the point of zero zeta potential at the pH ≈ 8.0. They also obeyed the yield stress-DLVO force model as characterized by a linear decrease in the yield stress with the square of the zeta potential.
Introduction Platelet R-Al2O3 particles display unusual reflective properties. As a result it is used as a cosmetic prepared in a paste-form. This cosmetic produces a soft focus optical effect that hides wrinkles. The manner in which this cosmetic paste is prepared and the way it spreads on the surface of skin is governed by its rheological behavior. Therefore, the ability to control its slurry rheological behavior is highly desirable and will improve the appeal of the cosmetic product to customers. The surface chemistry and rheology, in particular the yield stress behavior, of roughly spherical R-Al2O3 particle dispersions have been studied quite extensively.1-3 The surface chemistry was varied by pH and ionic strength and with a range of adsorbed additives of high and low molecular weight and of different molecular structure and architecture. In most of these cases the yield stress was employed to characterize the rheological behavior of the slurries. The yield stress measures the strength of the flocculated network structure. These dispersions obeyed the yield stress-DLVO force model1 where the yield stress decreases linearly with the square of the zeta potential. They also displayed a critical zeta potential value of 40 mV that was independent of particle size and concentration at an ionic strength of ∼0.05 M 1:1 electrolyte. The critical zeta potential parameter characterizes the transition from flocculated to disperse state. In the presence of adsorbed additives and depending upon their nature and structure, a range of non-DLVO forces such as steric and bridging2,3 were observed. However, the relationship between surface chemistry and rheology of platelet R-Al2O3 dispersions has not been studied systematically and in detail to provide a good understanding of * Corresponding author. E-mail:
[email protected]. (1) Zhou, Z. W.; Scales, P. J.; Boger, D. V. Chem. Eng. Sci. 2001, 56, 2901– 2920. (2) Leong, Y. K. Langmuir 2002, 18, 2448–2449. (3) Leong, Y. K. Phys. Chem. Chem. Phys. 2007, 9, 5608–5618.
this relationship. There is a distinct lack of literature in this area. However, the rheological and processing behavior of platelet R-Al2O3 composites has been investigated to a greater extent.4,5 For spherical particles the interaction between adsorbed additives of the interacting particles is only significant within a small spherical cap area near the closest point of interaction between particles.3,6 Within this area the adsorbed molecules are close enough to interact with molecules in the adsorbed layer of the second particle. The size of this area depends upon the size of the adsorbed molecule and particle radius and is generally quite small. So the density of intermolecular interactions is small. However, with platelet particles this interaction area can be very large, especially when it is in a face-face interaction orientation. This is particularly important when it involves adsorbed additives with multiple functional groups. A high density of intermolecular interactions such as hydrogen bonding is likely, especially when the face-face orientations are perfectly aligned. However, in a flocculated 3-D network structure perfect face-face alignment is not expected. Intra- and intermolecular functional group interactions of adsorbed additives could affect the nature and strength of the surface forces in dispersions.7 With platelet-platelet interactions the greatly increased density of intermolecular functional group interactions is amplified in terms of its effect on the yield stress and surface forces of the dispersions. Clays and talc are materials with a similar platelike structure. There are several types of clay, including kaolinite, bentonite, and montmorillonite. However, these materials are normally amphoteric in surface property. The face charge is always negative, and the edge charge is positive. In water these materials form a “house of cards” microstructure. The edge of one particle forms an electrostatic attractive bond with the face of a second (4) Gautier, S.; Champion, E.; Bernache-Assollant, D.; Chartier, T. J. Euro. Ceram. Soc. 1999, 19, 469–479. (5) Cherian, I. K.; Kriven, W. M. Am. Ceram. Soc. Bull., 2001, 80/11, 60–67, 80/12, 57-63. (6) Leong, Y. K. J. Chem. Soc., Faraday Trans. 1997, 93, 105–109. (7) Leong, Y. K. Colloids Surf., A 2008, 325, 127–131.
10.1021/la8036204 CCC: $40.75 2009 American Chemical Society Published on Web 02/10/2009
Yield Stress and Zeta Potential
Langmuir, Vol. 25, No. 6, 2009 3419
Figure 1. SEM Image of platelet R-Al2O3 particles.
particle. The surface chemistry and rheology of various clay and talc slurries, including the effects of adsorbed additives such as hydrolyzable metal ions8-10 and polyelectrolytes,11 have been studied quite extensively. Laxton and Berg12 studied the relationship between the yield stress and zeta potential of bentonite and kaolinite clay slurries. They reported a linear relationship between yield stress and zeta potential squared but with a positive slope instead of the usual negative slope. They attributed the behavior to electrostatic attraction. However, they also noted that the zeta potential was calculated on the basis of spherical particles and assumed that it is a good approximation of the zeta potential of the platelets particle. In the presence of adsorbed additives non-DLVO forces come into play. With rough and approximately spherical particles, adsorbed phosphate13 and citrate2,13,14 were observed to give rise to strong steric interactions between particles. However, with low surface area or smooth platelet particles the effect of these additives on surface forces is still unclear.
Material and Methods The platelet R-Al2O3 powder was sourced from Antaria, formerly known as ANT (Advanced Nano Technology), Perth, Western Australia. This powder is used as an ingredient in cosmetics as it produces a soft-focus optical effect. The platelet characteristic is clearly seen in the SEM image shown in Figure 1. The average diameter of the platelet is 7.5 µm. This powder has a relatively low BET surface area, typically of 1.8 m2/g. This means that the particles have a relatively smooth surface. The additives used to change the (8) Johnson, S. B.; Dixon, D. R.; Scales, P. J. Colloids Surf., A 1999, 146, 281–291. (9) Bremmell, K. E.; Addai-Mensah, J. J. Colloid Interface Sci. 2005, 283, 385–391. (10) Mpofu, P.; Addai-Mensah, J.; Ralston, J. J. Colloid Interface Sci. 2003, 261, 349–359. (11) McFarlane, A. J.; Bremmell, K. E.; Addai-Mensah, J. Powder Technol. 2005, 160, 27–34. (12) Laxton, P. B.; Berg, J. C. J. Colloid Interface Sci. 2006, 296, 749–755. (13) Leong, Y. K.; Scales, P. J.; Healy, T. W.; Boger, D. V.; Buscall, R. J. Chem. Soc., Faraday Trans. 1993, 89, 2473–2478. (14) Biggs, S.; Scales, P. J.; Leong, Y. K.; Healy, T. W. J. Chem. Soc., Faraday Trans. 1995, 91, 2921–2928.
particle surface chemistry were disodium hydrogen phosphate and citric acid. The dispersions for yield stress measurement were prepared by mixing a weighed amount of powder with a weighed amount of distilled water added with a few drops of 5 M NaOH and then sonicate with a sonic probe for about 1 min. The prepared dispersions were generally at an alkaline pH. The solid concentration of the prepared slurries ranged from 40 to 50 wt %. The 40 wt % slurries were chosen for the study with additives. The slurries with the citric and phosphoric acids were prepared in a similar manner to those without the additives. The additives were first dissolved in distilled water. More NaOH was added to neutralize the additives and to bring the pH of the prepared slurries to the alkaline starting pH of ∼12. The yield stress of the slurries was characterized using a vane technique.15 Two Brookfield DVII viscometers of different spring constants were used to conduct the vane measurement. The dispersions prepared for zeta potential measurement contained 5 wt % solids. Again they were prepared at an alkaline pH. The zeta potential was measured with a Colloidal Dynamic’s ZetaProbe. The equipment utilized an electroacoustic technique to determine the zeta potential. The equipment was operated in the potentiometric titration mode where the zeta potential was measured as a function of pH.
Results and Discussion Zeta Potential and Yield Stress of r-Al2O3 Dispersion. The effect of pH on the zeta potential behavior of 5 wt % platelet R-Al2O3 dispersion is shown in Figure 2. Repeat measurements with two fresh samples of the slurries showed almost identical zeta potential-pH behavior. Note the machine-calculated zeta potential is based on a theory developed for spherical particles and is therefore not a true potential of the platelet R-Al2O3 particles in water. However, it can be used as a representation of an “average zeta potential” of the particle.12 Figure 2 shows that the platelet R-Al2O3 particles have a point of zero charge (pI) at pH ≈ 8.0. The pI of the three samples are 7.72, 8.03, and 8.30, respectively. The two extreme values are within 0.5 pH unit of middle value. The negative zeta potential did not exceed 40.0 mV at pH (15) Nguyen, Q. D.; Boger, D. V. J. Rheol. 1983, 27, 321–349.
3420 Langmuir, Vol. 25, No. 6, 2009
Khoo et al.
The prepared dispersions at pH ≈ 12 displayed a relatively low yield stress that was quite insensitive to a small decrease in pH. In fact, the yield stress increased very gradually with decreasing pH up to 9.0, essentially forming a plateau. From pH 9.0, the yield stress increased sharply until it reached a maximum at around pH 7.0-8.0. After that pH the dispersions showed a zone of rapid yield stress decrease, followed by a zone of constant yield stress at pH below 4.0. The yield stress-DLVO force model for constant surface potential is given by:
τy ≈ -
Figure 2. Effect of pH on the zeta potential behavior of R-Al2O3 dispersions.
(
-κDo φ2 AH 2 κe 2πε εζ o a 12D2 (1 + e-κDo) o
below 12. Similarly, the positive zeta potential did not exceed 40 mV at pH less than 4.0. At pI the electrostatic repulsive force between particles is absent. Only the van der Waals attractive force is in play. The dispersion should be very strongly flocculated at this pH unless the van der Waals force is very weak. This occurs when the Hamaker constant of the particles in water is less than 1 kT. The strength of the flocculated network is determined by the nature and magnitude of the interparticles forces. The yield stress is a measure of the flocculated network strength and indirectly the strength of the interparticle force.2,13 The effect of pH on the yield stress behavior of 40, 45, and 50 wt % platelet R-Al2O3 suspensions is shown in Figure 3. For three dispersions there is no dispersed region where the yield stress is zero. The dispersions were flocculated in the entire pH range from 2 to 13. This behavior was very unlike that of other R-Al2O3 suspensions that are not platelet in shape.1,16 The maximum yield stress is located at the pH between 7 and 8.0 and increases with concentration. It increased from 19 Pa at 40 wt % solids to 53 Pa at 50 wt %. Within the limit of experimental error the maximum yield stress is located at the point of zero charge or pI. (16) Leong, Y. K.; Ong, B. C. J. Chem. Eng. Jpn. 2004, 37, 187–193.
(1)
where AH is the Hamaker constant of the particle in water, Do is the surface separation distance between the interacting particles, a is the particle radius, ε is the dielectric constant in water, εo is the permittivity of free space, ζ is the zeta potential (assumed equal to the surface potential) and κ is the Debye parameter. The number of particles per unit area scaled to (φ2)/(a2) where φ is the volume fraction of solids.17,18 At the point of transition from flocculated to dispersed state, the yield stress τy is zero so eq 1 gives the critical zeta potential,
ζcric )
Figure 3. Effect of pH on the yield stress behavior of platelet R-Al2O3 dispersion of different solids concentration.
)
AH(I + e-κDo) 24D2oπεεoκ e-κDo
The critical zeta potential appears to be only a function of the Hamaker constant, independent from the size, shape, and concentration of the dispersion.19 Many slurries obeyed one of the several yield stress-DLVO models.1,13,20,21 They displayed a linear relationship between yield stress and zeta potential squared with a negative slope. With amphoteric platelet bentonite and kaolinite slurries that formed “house-of-cards” structures, a positive slope was obtained instead.12 The maximum yield stress occurred at pI. So it is a function of the strength of the van der Waal attractive force and the number of particle-particle bonds per unit area, i.e. a function of volume fraction of particles. Do is expected to be constant and shortest in the flocculated state and should be independent of solids concentration. So the maximum yield stress is a function of the number of particle-particle bonds per unit area of the flocculated network or solids concentration. Figure 4 shows the relationship between the maximum yield stress and the solids volume fraction. Within the volume fraction range of 0.14 and 0.21, the maximum yield stress increased with the cube of the volume fraction. In contrast, the yield stress-DLVO force model given by eq 1 developed for spherical particles showed a dependence on the square of volume fraction. Suspensions that obeyed this yield stress-volume fraction square relationship were an exception rather than the norm. Normally the volume fraction exponent is much larger than 2. For approximately spherical ZrO2 dispersions the volume fraction exponent is 4.22 The yield stress-pH behavior of nonplate-like R-Al2O3 dispersions is quite different.1,16 These dispersions displayed a dispersed pH region at low and high pH where the electrostatic (17) Russel, W. B.; Saville, D. A.; Schowalter, W. R. Colloidal Dispersions; Cambridge University Press, Cambridge, 1989. (18) Larson, R. G. The Structure and Rheology of Complex Fluids; Oxford University Press, New York, 1999. (19) Leong, Y. K.; Ong, B. C. Powder Technol. 2003, 134, 249–254. (20) Hunter, R. J. AdV. Colloid Interface Sci. 1982, 17, 197–211. (21) Avramidis, K. S.; Turian, R. M. J. Colloid Interface Sci. 1991, 143, 54–68. (22) Leong, Y. K.; Scales, P. J.; Healy, T. W.; Boger, D. V. J. Am. Ceram. Soc. 1995, 78, 2209–2212.
Yield Stress and Zeta Potential
Figure 4. Relationship between maximum yield stress and solids volume fraction of platelet R-Al2O3 dispersion.
Figure 5. Plot of yield stress versus the square of zeta potential of platelet R-Al2O3 dispersions showing that the yield stress-DLVO force model is obeyed.
repulsive force dominates. The dispersed region occurred at pH less than 7 and more than 12 at ionic strength of 0.05 M of a 1:1 electrolyte solution. The critical zeta potential characterizing the flocculated-dispersed transition state was 40 mV. At the magnitude of 40 mV the electrostatic repulsive force is equal to the van der Waals attractive force at the maximum yield stress. The yield stress of the platelet R-Al2O3 dispersion in the basic region at pH > 10.0 was slightly larger than that in the acidic region at pH < 4. The magnitude of the zeta potential at pH 12 was generally less than 40 mV while the magnitude in the acidic region at pH ≈ 4 is slightly larger. The plot of yield stress versus zeta potential squared for the platelet R-Al2O3 dispersions is shown Figure 5. A linear relationship was observed for the 40 and 45 wt % R-Al2O3 dispersions. However, the 50% solids dispersion showed a considerable higher degree of scatter in the data, but these data can be approximated by a linear relationship. With nonplatelet R-Al2O3 dispersions, the square of the critical zeta potential is (1600 mV)2.1 With the platelet dispersion, the critical zeta potential was certainly dependent upon solids concentration and was much more than 40 mV. The nonconstant critical zeta potential may be due to the fact that the true zeta potential of this R-Al2O3 dispersion was not known. Also the theory of zeta potential for
Langmuir, Vol. 25, No. 6, 2009 3421
Figure 6. Effects of citric acid concentration on the zeta potential-pH behavior of 5 wt % platelet R-Al2O3 dispersions.
nonspherical particles and its relationship with particle shape have not been established. Additives Effects. Apart from changing the surface chemistry of the particles adsorbed additives also gave rise to a range of non-DLVO surface forces such as steric, bridging, hydrophobic, or entropic. There is a body of literature on surface forces due to adsorbed additives in aqueous dispersions of near spherical particles detected by a yield stress rheological technique.2,3,6,7,13,23 However there is no study involving platelet particle dispersions. In this study the effect of anionic phosphate and citrate were evaluated. The effect of citric acid or citrate on the zeta potential-pH behavior of the platelet alumina dispersion is shown in Figure 6. Citric acid shifted the zeta potential-pH curve to a lower pH. The pH of zero zeta potential of 8.0 is shifted to 4.5 at 0.1 dwb % citric acid and to 4.2 at 1.0 dwb % citric acid. The small change in the pH of zero zeta potential between the different citric acid concentrations suggests that the alumina platelet particles are already saturated with adsorbed citrate molecules at 0.1 dwb % concentration. Adsorption of citrate commenced at pH just below 10 as indicated by the zeta potential being significantly more negative than that without citrate added. Adsorption of negative charged additives at pH 10 delayed the arrival of the point of zero charge to a lower pH. The effect of phosphate on the zeta potential-pH behavior is very similar to that of citric acid. See Figure 7. For concentrations of 0.1-1.0 dwb % phosphate, the pH of zero zeta potential varies slightly from 5 to 6.0. Adsorption of phosphate begins significantly at pH below 10 as reflected by the zeta potential being much more negative than that with no phosphate added. At 0.1 dwb % phosphate the surface of platelet alumina particles must be nearly saturated with phosphate. Adsorbed citrate had quite a profound effect on the yield stress of platelet R-Al2O3 dispersions. It reduced the yield stress very significantly. See Figure 8. The maximum yield stress at pH of zero zeta potential was decreased by as much as 50%. It decreased from 20 Pa to 12 Pa at 0.25 dwb % citrate and to 10 Pa at 1.0 dwb % citrate. The relative insensitivity of the yield stress to pH is illustrated with the data for 1.0 dwb % citrate. The yield stress at pH 12 was 8 Pa increased to only 10 Pa at pH 4.5. Also citrate did not bring about complete dispersion of the slurry at pH far away from the point of zero zeta potential. Adsorbed citrate is (23) Johnson, S. B.; Brown, G. E.; Healy, T. W.; Scales, P. J. Langmuir 2005, 21, 6356–6365.
3422 Langmuir, Vol. 25, No. 6, 2009
Figure 7. Effects of phosphate concentration on the zeta potential-pH behavior of 5 wt % platelet R-Al2O3 dispersions.
Khoo et al.
Figure 9. Effect of pH on the speciation of citric acid.
Figure 8. Effect of citric acid concentration on the yield stress-pH behavior of 40 wt % platelet R-Al2O3 dispersions.
known to form a steric barrier on the interacting particles of roughly spherical shape.2,13,14 This work also showed that adsorbed citrate forms a very effective steric barrier, keeping the interacting platelet particles further apart. As a result, the van der Waals attractive force is weaker, and hence, a low yield stress. Assuming 100% adsorption, the surface coverage at 1 dwb % citric acid is 6.3 Å2/molecule. Citric acid consists of three backbone carbons with a carboxylic acid group attached to each of these carbons. The middle carbon also has one hydroxyl group attached to it. This surface coverage is much greater than monolayer coverage for the carboxylate headgroup of 22 Å2/ molecule. So there must be a significant amount of free citric acid in solution, possibly as much as 80% remaining and not adsorbed. At 0.25 dwb % citric acid the surface coverage is 25 Å2/molecule which is much closer to monolayer coverage if only one carboxylate group of the acid is adsorbed. At saturation surface coverage, the intermolecular and intramolecular interactions between functional groups of the adsorbed molecules have a profound effect on the nature and magnitude of the interparticle or surface forces. These intermolecular and intramolecular interactions are dependent upon the dominant citric acid species present. Figure 9 shows the species diagram for citric acid. At the pH of zero zeta potential (pH 4-5) of the dispersions the main
Figure 10. Cartoon of an adsorbed monoionic citric acid molecule on a platelet R-Al2O3 particle. The monoionic citric acid conformation and its intramolecular hydrogen bonding were predicted from the modeling with ChemOffice Ultra 10 software subjected to MM2 energy minimization conditions. The effect of water solvent was not considered.
species (77%) is the monoionic citric acid, H2A-. This species became a neutral molecule when adsorbed on a positive surface site. Thus, this species was not capable of forming an electrostatic bridging bond with another particle. Figure 10 shows a monoionic citric acid molecule adsorbed on a positively charged surface. This monoionic citric acid was produced by molecular modeling with ChemOffice Ultra 10 subjected to MM2 minimum energy condition. The modeling shows the presence of intramolecular hydrogen bonding between the -OH group and the terminal undissociated carboxylic acid group in carbon 3. This hydrogen bond rendered the terminal carboxylic acid group unable to give rise to strong particle bridging interactions. Intramolecular hydrogen bonding also hindered intermolecular hydrogen bonding between citric acid molecules on the same adsorbed layer or with another molecule in the adsorbed layer of an interacting
Yield Stress and Zeta Potential
Figure 11. Effect of phosphate concentration on the yield stress-pH behavior of 40 wt % platelet R-Al2O3 dispersions.
particle. As a result adsorbed citric acid functioned very effectively as a steric layer at the pH of maximum yield stress. This layer weakened the van der Waal forces interacting between particles by increasing the interparticle separation. The interparticle separation is increased by the thickness of two adsorbed citrate molecules. At saturation coverage of greater than 0.25 dwb % citric acid, this separation distance should not be dependent upon citric acid concentration. This may explain the small difference in the yield stress at any given pH. The effect of phosphate on the yield stress-pH behavior of platelet R-Al2O3 dispersion is completely unlike the effect of citric acid. See Figure 11. Phosphate caused the yield stress to be much higher. For example, the maximum yield stress increased from 20 Pa at 0.0 dwb % phosphate to 40 Pa at 0.25 dwb % phosphate and to 45 Pa at 1.0 dwb % phosphate. The maximum yield stress is located at the pH of zero zeta potential of ∼4.5. Like zeta potential, the yield stress-pH behavior is not very sensitive to phosphate concentration ranging from 0.25 to 1.0 dwb %. Assuming 100% adsorption the surface coverage is ∼12 Å2/ molecule for 0.25 dwb % phosphate. This phosphate concentration is more than that required to saturate the surface of platelet R-Al2O3 particles. Complete surface coverage should occur at a lower concentration of 0.12 dwb % phosphate based on 100% adsorption assumption. However, from the slight variation of the pH of zero zeta potential with phosphate concentration shown in Figure 7, complete coverage should occur at the concentration beyond which the point of zero zeta potential remained constant. This concentration is above 0.12 dwb % phosphate. Adsorbed phosphate must have given rise to an additional attractive force. The surface of the particles was nearly saturated with phosphate at concentration of 0.2 dwb % phosphate. The additional force was most likely due to hydrogen bonding. Each of the interacting particles contained a layer of adsorbed phosphate. Hydrogen bonding is a relatively weak force. Kjellin and Claesson observed a weak attractive force involving adsorbed tetra(ethylene oxide)dodecyl amide and stated specifically that they were not prepared to attribute this attraction to hydrogen bonding involving the amide group.24 More recently, 0.8 nN was attributed to hydrogen bonding contribution to the total AFM adhesion force of Staphylococcal bacteria to a hydrophilic glass (24) Kjellin, U. R. M.; Claesson, P. M. Langmuir 2002, 18, 6754–6763.
Langmuir, Vol. 25, No. 6, 2009 3423
Figure 12. Effect of pH on the speciation of phosphoric acid.
surface.25 The strength of 0.8 nN was obtained by applying Poisson analysis to multiple AFM adhesion peaks allowing for separation of hydrogen bonding from other nonspecific interaction forces. This adhesion force should depend upon the number of hydrogen bonds formed between the bacteria and the hydrophilic glass surface. Figure 12 shows the species diagram for phosphoric acid. At the pH of zero zeta potential of 4.5 the main species (98%) was H2PO4-. Adsorption of this monoionic species will produce neutral species with two -OH groups that can participate in hydrogen bonding with a similar species adsorbed on the second interacting particle. The cartoon showing hydrogen bonding between adsorbed phosphate species of two interacting alumina particles is shown in Figure 13. A pair of this adsorbed phosphate species interacting can generate as many as 4 hydrogen (O---H) bonds. However, the diagram shows only two such bonds. The more of these bonds, the stronger will be the net interparticle attractive forces. Despite the weak nature of the hydrogen bond the 100% increase in the maximum yield stress suggests that the density of hydrogen bonding must be very high. With face-face interaction a high density of hydrogen bonding is expected, especially in conditions of saturation coverage of phosphate. The H2PO4- structure in the diagram was produced by molecular modeling with ChemOffice Ultra 10.0, subject to MM2 minimum energy conditions. The attachment of phosphate ions to the particle surface and the hydrogen bonding were drawn to show the interaction. Adsorbed phosphate was earlier reported to function as a highly effective steric barrier in dispersions of rough and nonideal spherical oxide particles.13,23 It caused a very significant reduction in the maximum yield stress of the dispersions of more than 50%. With these particles the phosphate needs to be adsorbed within a spherical cap area near the closest point of interaction for it to be close enough to reach another phosphate molecule adsorbed on the second interacting particle.3,6 This spherical cap area, a function of particle radius and the adsorbed phosphate molecule length, is very small. It accounts for much less than a fraction of 1% of the total particle surface area. Thus, the number of hydrogen bonds formed within this area is small. This number will be further reduced if the particle surface is very rough. The surface roughness is indicated by the BET area. One of the oxides used has a BET area of 16 m2/g.13 Adsorbed (25) Boks, N. P.; Busscher, H. J.; van der Mei, H. C.; Norde, W. Langmuir 2008, 24, 12990–12994.
3424 Langmuir, Vol. 25, No. 6, 2009
Khoo et al.
Figure 13. Cartoon showing adsorbed monoionic phosphates participating in intermolecular hydrogen bonding. The phosphate monoions conformation was modeled using the ChemOffice Ultra 10 software subjected to MM2 energy minimization conditions.
phosphate located in the trough and crevices could be too far away from a second phosphate molecule adsorbed on the second particle. Under these circumstances, the effect of hydrogen bonding on increasing the strength of the interparticle attractive force is small or negligible. Yield stress of concentrated suspension is an ideal method for probing small intermolecular forces such as hydrogen bonding. Concentrated suspensions have millions of particle interactions, and so any small intermolecular force is significantly magnified. This is reflected by a very large increase in the maximum yield stress. It may even be possible to distinguish the strength of the different types of hydrogen bonds involving different functional groups.
Conclusion The rheological behavior of platelet R-Al2O3 dispersions showed that it is flocculated at all pH levels. The maximum yield stress is located at the pH of zero zeta potential. Adsorbed citrate and phosphate additives change the zeta potential and yield stress
behavior of the dispersions but did not bring about complete dispersion at any pH level. Adsorbed citrate reduced the maximum yield stress by as much as 50%. Hence, it functioned as a steric layer keeping the flocculated particles further apart. Adsorbed phosphate increased the maximum yield stress by as much as a 100%. It generated an additional attractive force between particles. This additional attractive force is attributed to hydrogen bonding between phosphate species adsorbed on the interacting particles. Acknowledgment. We acknowledge the support of the University of Western Australia via a USF grant for the purchase of the ZetaProbe. We also acknowledge Antaria for the small financial contribution towards the purchase of this ZetaProbe and the provision of the raw materials for this investigation. B.C.O. was a visiting fellow at the University of Western Australia. For making this a better paper we thank the reviewer for the useful comments and suggestions and Sabbia Tilli for manuscript editing. LA8036204
