Langmuir 2009, 25, 3981-3984
3981
Thin Water Layers on CaCO3 Particles Dispersed in Oil with Added Salts† Seung Yeon Lee,‡ Michael O’Sullivan,§ Alexander F. Routh,§ and Stuart M. Clarke*,‡ Department of Chemistry and Department of Chemical Engineering, BP Institute, UniVersity of Cambridge, Madingley Road, Cambridge CB3 0EZ, United Kingdom ReceiVed August 11, 2008. ReVised Manuscript ReceiVed NoVember 2, 2008 We report an investigation of the presence of thin water layers on calcium carbonate particles dispersed in cyclohexane using small-angle neutron scattering. We identify an adsorbed water layer and measure the thickness using contrast variation to optimize the sensitivity of the scattering to the water. We also report the variation in thickness of these water layers in the presence of salt solutions with the variation of salt concentration and valency. We conclude that thin water layers can be observed using SANS; however, the layers are thin and correspond to essentially the hydration of the particle surfaces.
Introduction Adsorption from liquids to solid surfaces is key to many academic and commercial issues because of their important role in many interfacial phenomena such as lubrication, wetting, detergency, and colloidal stabilization. For example, the colloidal stability of nanoparticles of calcium carbonates (CaCO3) is important in their use as additives in lubricant oils. These “overbased” engine oil additives are important in controlling acid formation in the fuel combustion process1-4 and often occur in the presence of metal ions of different types and water formed from the combustion process. Hence, this system provides an ideal opportunity to study the adsorption of water at the calcium carbonate/oil interface in the presence of a variety of species, in particular, aqueous solutions of ions with different valency and concentrations. Recently, small-angle neutron scattering (SANS) has been used to characterize core-shell particles of this type, using the contrast variation technique, in nonaqueous solvents.3-10 Quantitative information on the size, shape and structure of colloidal particles in suspensions can be provided with this approach. A crucial feature of SANS that makes it particularly useful is the different scattering of different isotopes of the same element, especially hydrogen (1H) and deuterium (2D), that permits contrast variation and contrast matching to optimize the sensitivity of the scattering to regions of interest. In particular, Markovic et al. † Part of the Neutron Reflectivity special issue. * Corresponding author. Tel: +00 44 1223 765700. Fax: +00 44 1223 765701. E-mail:
[email protected]. ‡ Department of Chemistry. § Department of Chemical Engineering.
(1) Tavacoli, J. W.; Dowding, P. J.; Steytler, D. C.; Barnes, D. J.; Routh, A. F. Langmuir 2008, 24, 3807–3813. (2) Hudson, L. K.; Eastoe, J.; Dowding, P. J. AdV. Colloid Interface Sci. 2006, 123, 425–431. (3) Markovic, I.; Ottewill, R. H.; Cebula, D. J.; Field, I.; Marsh, J. F. Colloid Polym. Sci. 1984, 262, 648–656. (4) Galsworthy, J.; Hammond, S.; Hone, D. Curr. Opin. Colloid Interface Sci. 2000, 5, 274–279. (5) Markovic, I.; Ottewill, R. H. Colloid Polym. Sci. 1986, 264, 65–76. (6) Strunz, P.; Mukherji, D.; Pigozzi, G.; Gilles, R.; Geue, T.; Pranzas, K. Appl. Phys. A 2007, 88, 277–284. (7) Seguin, C.; Eastoe, J.; Heenan, R. K.; Grillo, I. J. Colloid Interface Sci. 2007, 315, 714–720. (8) Penfold, J. Curr. Sci. 2000, 78, 1458–1466. (9) Wyslouzil, B. E.; Wilemski, G.; Strey, R.; Heath, C. H.; Dieregsweiler, U. Phys. Chem. Chem. Phys. 2006, 8, 54–57. (10) Ottewill, R. H. Small Angle Neutron Scattering in Colloidal Dispersions; Academic Press: New York, 1982; pp 143-163.
used SANS to characterize the CaCO3 particle system in nonaqueous solvents.3,5,11,12 They measured the sizes of core particles and the thicknesses of surfactant layers separately and showed the stable core/shell structure in both concentrated and dilute systems.
Experimental Section Sample Preparation. CaCO3 nanoparticles were obtained from Infineum UK Ltd. They were prepared by diffusing CO2 gas through a mixture of a sulfonate surfactant, a mixture of hydrocarbon and polar solvents, and emulsion droplets of Ca(OH)2.13-17 The carbonate core was stabilized by a monolayer of alkyl aryl sulfonate calcium salt surfactants that was 210 mM kg-1 in the stock solution. A stock solution of amorphous CaCO3 was diluted to 7 wt % in H-cyclohexane and in mixed D/H-cyclohexane (64:36 to contrast match the oil to the particle core). Salt solutions (LiCl, NaCl, KCl, MgCl2 · 6H2O, CaCl2 · 2H2O, AlCl3 · 6H2O, Na2SO4, and MgSO4) were prepared in H2O and D2O to give solutions of 0.1, 0.01, and 0.001 M. The CaCO3 dispersions (3.0 g) were mixed with 1 mL of the salt solutions by inverting slowly to achieve good mixing but to avoid emulsification. Some water did remain at the bottom of the vials, indicating that the samples were saturated. The samples were put into 1.0 mm flight-path quartz cells for the SANS experiments. Small-Angle Neutron Scattering (SANS). D11 at Institut LaueLangevin (ILL), Grenoble, France, was used for this work. The beam size at the sample was 7 mm × 10 mm with an incident wavelength of 6 Å, corresponding to the maximum flux with a spread of 9%. The Q range was from 0.006 to 0.2 Å-1. Two detector distances were used: 2 and 8 m with collimation distances, respectively, of 5.5 m (protonated samples) or 8 m (deuterated samples) and 8 m for both H and D samples. The different collimation distances for H and D samples were chosen to obtain optimal scattering conditions, including the avoidance of detector saturation, and were treated separately in the normalization. Absolute intensities were calculated by using the measurement of a standard sample (1.0-mm-thick H2O) (11) Markovic, I.; Ottewill, R. H. Colloid Polym. Sci. 1986, 264, 454–462. (12) Ottewill, R. H.; Sinagra, E.; Macdonald, I. P.; Marsh, J. F.; Heenan, R. K. Colloid Polym. Sci. 1992, 270, 602–608. (13) Bandyopadhyaya, R.; Kumar, R.; Gandhi, K. S. Langmuir 2001, 17, 1015– 1029. (14) Dagaonkar, M. V.; Mehra, A.; Jain, R.; Heeres, H. J. Chem. Eng. Res. Des. 2004, 82, 1438–1443. (15) Kandori, K.; Konno, K.; Kitahara, A. J. Colloid Interface Sci. 1988, 122, 78–82. (16) Kang, S. H.; Hirasawa, I.; Kim, W. S.; Choi, C. K. J. Colloid Interface Sci. 2005, 288, 496–502. (17) Eastoe, J.; Paul, A.; Nave, S.; Steytler, D. C.; Robinsion, B. H.; Rumsey, E.; Thorpe, M.; Heenan, R. K. J. Am. Chem. Soc. 2001, 123, 988–989.
10.1021/la802616n CCC: $40.75 2009 American Chemical Society Published on Web 12/15/2008
3982 Langmuir, Vol. 25, No. 7, 2009
Lee et al.
Figure 1. Contrast matching and variation of samples in (a) protonated and (b) deuterated solvents used in this work.
Figure 2. Radius of the core (R1). The outer radius of the water layer is R2 ) R1 + water layer, and the external radius of the particle is R3 ) R2 + surfactant sheath layer.
at room temperature. All data was normalized to the scattering of the empty cell, the sample background, and water. The scattering modeling and data correction to the experimental data are given in the Supporting Information. Data was analyzed using the FISH program (Richard Heenan, ISIS, U.K.).18 Figure 1 shows the scattering length density (SLD) distributions of the samples. In the H-cyclohexane solvent, the SLD of the sulfonate surfactant matches that of the H-cyclohexane continuous phase (Figure 1a), hence this data set could be used to
determine the size and composition of the CaCO3 core of the particle using a simple core-only model. In D/H-cyclohexane, the SLD of the carbonate particle core matches that of the D/H-cyclohexane phase (Figure 1b). In the general case, the data was fitted with the more complex core/shell/shell model to encompass the most complex colloidal structures. Figure 2 defines the radii (R1, R2, and R3) and the SLDs (F1, F2, F3, and F4) used during the fits. In this work, dry samples refer to the CaCO3 particles in oil without any water present. Wet refers to the samples with pure water present. The SLDs of Hand D-cyclohexane were -0.28 × 1010 and 6.32 × 1010 cm-2, respectively.
Results and Discussion Dry and Wet. The data from the dry, core-only sample was fitted to a simple hard sphere model, referred to below as the core-only model. Scattering from the protonated surfactant sheath can be ignored at this stage in a hydrocarbon solvent because of their similar SLD values. The radius of the particle core, RHS, from these fits was found to be 26.3 Å, with a polydispersity of 0.2 and a core SLD of 4.24 × 1010 cm-2 as given in Table 1. This scattering-length density agrees reasonably well with that
Figure 3. Fits for the dry and wet carbonate particles in (a) H-cyclohexane and (b) D/H-cyclohexane. (a) Core-only model used for dry and core-shell model used for wet. (b, c) Core-shell model used for dry and core-shell-shell model used for wet. (c) Linear scaled graph displayed to show the difference between dry and wet at low Q in D/H-cyclohexane. The fitted particle polydispersity was found to be 0.15.
Thin Water Layers on CaCO3 Particles
Langmuir, Vol. 25, No. 7, 2009 3983
Table 1. Experimentally Determined Structural Parameters for Wet and Dry Carbonate Particlesa fit name
F1/1010cm-2
F2/1010cm-2
F3/1010cm-2
core only D2O sheath H2O
4.25 4.25 4.25 4.25
-0.28 5.71 -0.06 -0.53
-0.28 4.2 -0.06
a
F4/1010cm-2
water layer/Å
surfactant sheath/Å
2.7 4.2
2.7
8.8 7.8
Representive uncertainties from SANS fits: R1 (core only), (0.1 Å; R2 and R3, (0.4 Å, and the shell thickness (water, sheath layers), (0.5 Å.
above. Given that the protonated surfactant has a similar SLD to that of the medium, the contribution from the external sheath layer was not included in these fits. For this contrast, both the surfactant and the added water are expected to be in the shell layer and will contribute to the total SLD of the shell. The best fit indicated an SLD of this water layer that corresponds to a volume fraction ratio of 0.9:0.1 for D2O and surfactant in the shell layer. In the fittings, an averaged SLD of the surfactants was considered to be -0.28 × 1010 cm-2, including both the polar headgroup and hydrocarbon tails. The effective thickness of the thin water layer on the surface of the core particle was determined to be 2.7 Å.
Figure 4. Schematic of the proposed formation of the thin water layer and surfactant sheath on carbonate particles. (a) Core only in Hcyclohexane, (b) surfactant sheath and core in D/H-cyclohexane, (c) thin water layer and surfactant sheath on carbonate particles in H-cyclohexane, and (d) D/H-cyclohexane.
expected from bulk CaCO3. The best fit is shown in Figure 3. The total scattering is found to correspond to a volume fraction of particles of 0.007, which compares with the as-supplied volume fraction of 0.009. A core-shell model was used to fit the D2O-saturated samples in H-cyclohexane, referred to below as the D2O model. The parameters used to fit the core alone were kept as those determined
Figure 5. Thickness of water (H2O) and salt layers in D/H-cyclohexane.
For the dry particles dispersed in D/H-cyclohexane, the core-shell model was used and is referred to below as the sheath model. The value of the hard sphere radius, RHS was again established as 26.3 Å, as determined above. Given that the solvent could penetrate the surfactant sheath shell, the SLD of the surfactant layer was fitted by including some D/H solvent SLD to surfactant SLD and was determined by the best fit to correspond to a volume fraction ratio of 0.95:0.05 surfactant to D/H solvent. The thickness of the surfactant sheath layer on the surface of the core particle was calculated to be 8.8 Å. Figure 3 shows the fit, and the results are listed in Table 1. A core-shell-shell model for hard spheres was used to fit the SANS data of the H2O-saturated samples in D/H-cyclohexane, where all parts of the system have different SLD values. Again, the fitting was required to be consistent with the results of the other contrasts above. This is referred to below as the H2O model. The thickness of the thin water layer on the surface of the core particle was again determined to be 2.7 Å. Figure 3 illustrates the fit, and the results are listed in Table 1.
3984 Langmuir, Vol. 25, No. 7, 2009
Dependence on Valency and Concentrations of Added Salts. A significant number of samples with various salt types and concentrations were measured and fitted in the same way as discussed above: monovalent, divalent, and trivalent cations, monovalent and divalent anions, and ions with the same valence but different sizes were used. To restrict the number of fitted variables, all those except those for the water layers were kept as those used for the relevant fits of the wet samples above. Although we note that D2O samples in H-cyclohexane are more sensitive to the thickness of the water layer than H2O added to the samples in D/H cyclohexane, we have fitted both contrasts. The results are summarized in Figure 5. In general, we note that the water layer thickness increases with decreasing salt concentration for essentially all samples, independent of cation or anion valency. The thickness of the pure water layer is also included in this graph, and it is smaller than that of 0.001 M salinity systems. In considering the variation of behavior with cation/anion type for the 0.001 M systems, we note that there are no obvious trends in behavior, although there is some evidence that divalent cations have a slightly thicker water layer than the monovalent ions but trivalent aluminum does not follow this trend. However, high-valency ions often do not exist in solution as single hydrated species but tend to form more extensive complexes; this makes comparison more difficult. At the high salt concentration of 0.1 M, there is some evidence for smaller cations to have slightly thicker water layers, possibly reflecting their higher charge density and water-binding ability (Li+ < Na+ < K+). In addition, we note that Mg2+ with higher
Lee et al.
valence than K+ also had a thicker water layer, but it is difficult to draw significant conclusions at present.
Summary SANS has been used to investigate the presence of water layers at the surface of calcium carbonate particles. The thickness of the layer as a function of salt valency and concentration has also been studied. Using a number of different contrasts, we find that there is evidence for the formation of a water layer in this system; however, the thicknesses of the water layers were all rather small, and hence we conclude that they essentially represent the hydration of surface charges. Eight different salts were used to vary the ionic strength in H2O or D2O. In general, we find a repeatable trend with concentration in the water layer thickness but rather less obvious variations with cation/anion valency. Acknowledgment. We thank BP for financial support for this work, the ILL staff and scientists for the allocation of beam time and technical assistance with the SANS measurements, Infineum UK Ltd for donating the samples, and Richard Heenan for helpful discussions. Supporting Information Available: Details of the scattering models used and the data correction procedures. This material is available free of charge via the Internet at http://pubs.acs.org. LA802616N (18) Heenan R. K. FISH Data Analysis Program; Rutherford Appleton Laboratory: Didcot, U.K., 1989.
