Adsorption of the poly (oxyethylene) nonionic surfactant C12E5 to

Mark W. Rutland* and Tim J. Senden .... molecularly smooth over an area of the order of a square ... need not be molecularly smooth over large areas a...
0 downloads 0 Views 1MB Size
Langmuir 1993,9, 412-418

412

Articles Adsorption of the Poly(oxyethy1ene) Nonionic Surfactant C12E5 to Silica: A Study Using Atomic Force Microscopy Mark W. Rutland* and Tim J. Senden Department of Applied Mathematics, Research School of Physical Sciences, The Australian National University, GPO Box 4, Canberra, ACT 2601, Australia Received July 23,1992. In Final Form: October 29, 1992 The forces between a spherical silica particle of colloidal dimension and a smooth silica surface have been measured in aqueoussolution as a function of the concentrationof the nonionic surfactant pentakia(oxyethylene)dodecyl ether. In the absence of surfactant the interaction between silica surfaces was purely repulsive, being composed of electrostatic and hydration forces, whereas at low surfactant concentrations(=2 X 1 t 6M, about one-third the critical micelle concentration (cmc))the repulsion was replaced at short separations by an attractive force which pulled the surfaces into adhesive contact. At higher concentrations(=4 X 1 t 6M)the surfaces still experiencedan attractive force at small Separations, but the adhesion decreased markedly. At concentrationsabovethe cmc repulsivestericforcesare observed. A tentative scheme for the adsorption of poly(oxyethy1ene)surfactant to silica is presented.

Introduction The adsorptionof poly(oxyethy1ene)nonionic surfactant to silica has been the subject of much attention in the 1iterature.l-ll The main industrial application of this type of surfactant is in the area of detergency, although they are also used for stabilizing dispersions and have been investigated for use as flotation agents and tertiary oil recovery agents. A great deal of debate surrounds the mechanism of adsorption of the surfactant to the silica surface and the nature of the surface aggregate so formed. The very steepness of adsorption isotherms of nonionic8 on silica,'* and their proximity to the cmc (critical micelle concentration), implies that the adsorption takes place largely by an aggregation process rather than by a surface site binding mechanism. However, early views were that the adsorption process followed a very similar path to ionic surfactant adsorption, i.e., initial adsorption of monomers, then hydrophobic association, and finally the formation of a bilayer near the cmc (critical micelle concentration), with the hydrocarbon tails intercalating due to the large size of the headgro~p.l-~*~ More recently this approach has been questioned,and many techniqueshave since been employed to infer the structure of the so-called "solloid" (a surface aggregate of adsorbed surfactant or polymer).l3 Some of these are summarized briefly here.

* To whom correspondence should be addressed at the Institute for Surface Chemistry, P.O. Box 5607,114 86 Stockholm, Sweden. (1) Rupprecht, H. Kolloid 2.1971, 249, 1127. (2)Rupprecht, H.Prog. Colloid Polym. Sci. 1978, 65,29. (3)Seng, H.P.; Sell, P. J. 2'ert.de Deterg. 1977, 14, 4. (4)Furlong, D. N.;Aston, J. R. Colloids Surf. 1982, 4 , 121. (5) Partyka, S.;Zaini, S.; Lindheimer, M.; Brun, B. Colloids Surf. 1984, 12, 255. (6)Levitz, P.; Van Damme, H.; Keravis, D. J. Phys. Chem. 1984,88, 2228. (7)Lee, E.M.;et al. Chem. Phys. Lett. 1989, 162, 196. (8) Gu, T.; Zhu, B.-Y. Colloids Surf. 1990, 44, 81. (9)Cummins, P. G.; Staples, E.; Penfold, J. J.Phys. Chem. 1990,94, 3740. . .. (10)Gonzalez, G.; Travalloni-Louvisse, A. M.Langmuir 1989,5, 26. (11)Lindheimer, M.; Keh, E.; Zaini, S.;Partyka,S. J. Colloidlnterface Sci. 1990, 138, 83. (12)Ottewill, R. H.;Rochester, C. H.; Smith, A. L. Adsorption from Solution; Plenum Press: London, 1983. (13)Someeundaran, P.; Kunjappu, J. T. Colloids Surf. 1989,37, 245. ~~~~

0743-746319312409-0412$04.0010

Levitz et al.6 used fluorescence decay of pyrene with TXlOO and -101 on spherosil and determined that at low coverages the adsorption was of a micellar nature but that at high coverages steric interactions caused the micelles to coalesce, forming a single, continuous bilayer-like structure. Lee et al., using neutron reflection' and hexakis(oxyethylene)dodecyl ether (C12&) on quartz, proposed a similarschemewhere "defectivebilayers" at low coverage fused into full bilayers at high coveragewith a hydrocarbon layer thickness of a single alkyl chain. In subsequent work Cummins et al.9 investigated the adsorption of Cl& to a ludox TM silical sol using SANS (small-angle neutron scattering). At very low coverages their resulta were inconclusive,but above a certain (low)coveragevalue they observed the adsorbate thickness to remain constant with increasing coverage. They proposed that the adsorption took place as "islands" of bilayers, which never attained better than 75% coverage (at very low coverages the adsorbate thickness was probably lower). Gu et al.8 through a kinetic treatment of adsorption isotherms on both narrow- and wide-pore silica using TXlOO propose that adsorption occurs by a single-step cooperative formation of aggregates on the surface, an average aggregation number of which may be determined from the adsorption isotherm, and that there is no initial monomeric "nucleation". Possibly some of the apparent conflictsin the literature may be explained by the work of Lindheimer et al., who used a modern microcalorimetric technique.ll From their study of the adsorption of a range of commercial poly(oxyethylene)alkyl ethers on spherosil silica they propose an adsorption scheme involvinga four-stage process. The first stage is a "nucleation stage" whereby individual monomers adsorb on silanol groups through hydrogen bonding. This is closely followed by a hydrophobic association,giving rise to a hydrophobic surface. However, as the concentration is raised further, but to a concentration corresponding to far less than 60% of maximum adsorption,the surfacetakes on a hydrophilicsurfaceagain due to the formation of preaggregates similar to those in the bulk. A t maximum coverage the surface aggegate is 0 1993 American Chemical Society

Adsorption of Poly(oxyethy1ene)Surfactant to Silica equivalent to that of the bulk. This work appeared to be in agreement with the contact angle work of Gonzalez et d.l0on quartz using TXlOO in terms of the observation of a hydrophobic-hydrophilic transition at low-medium coverage. The surface force apparatus (SFA)has been previously used to study surface forces between surfaces with adsorbed surfactant, and the effect of the adsorbate on the surface force has been monitored. It is possible to draw conclusionsabout the structure of the adsorbate from changesin the surface force and thickness of the absorbed l a ~ e r . l ~The - ~ ~technique has traditionally been limited to the study of muscovite mice surfaces and mica surfaces coated with Langmuir-Blodgett (LB) films due to the requirement that the surfaces under examination be molecularly smooth over an area of the order of a square centimeter (although recently new developments have enabled the study of a broader range of substrate^^^-^^). Surface forces may now also be measured using the recently developed atomic force microscope27which has very recently been applied to colloidal systems.28 The instrument possesses the dual advantages that surfaces need not be molecularly smooth over large areas and that the substrates need not be limited by refractive index and thickness for interferometric measurements. The major disadvantage of this technique is that no absolute zero of separation can be obtained and must be inferred from the force measurement. Thus, the two techniques of surface force measurement are largely complementary. In this paper we present results of measurements of surface forces between a spherical silica particle and a flat silica surface in aqueous solutions of the nonionic.surfactant pentakis(oxyethylene) dodecyl ether (C12E5). Materials and Methods was obtained from Nikko Chemicals and used without Mmand purification. Literature values for the cmc are 4 X 6.5 X 10-5M.30 Analytical grade NaCl was roasted at 500 OC to remove any organic contaminants. Water was treated by reverse osmosis (Krystal Klear, Memtec) and then treated with ion exchange and activated charcoal (Milli-Q Plus, Millipore). It typically had a pH of 6. The atomic force microscope (AFM) was a commercial instrument (the Nanoscope 11)purchased from Nanoscope, Santa Barbara, CA. Sample Preparation. Measurements were performed between silica-glass spheres (Polysciences Inc., Warrington, PA) and a silica substrate made by oxidizing a polished silicon wafer (14) Pashley, R. M.; Israelachvili, J. N. Colloids Surf. 1981,2, 169. (15) Israelachvili, J. N.; Pashley, R. M. J . Colloid Interface Sci. 1984, 98,500. (16) Clawon, P. M.; Kjellander, R.; Stenius, P.; Christenson, H. K. J. Chem. SOC.,Faraday Trans. 1 1986,82,2735. (17) Herder, P. C.; Claesson, P. M.; Herder, C. E. J. Colloid Interface Sci. 1987, 119, 155. (18) Kbkicheff, P.; Christenson, H. K.; Ninham, B. W. Colloids Surf. 1989,40, 31. (19) Herder, P. C. J. Colloid Interface Sci. 1990, 134, 336. (20) Herder,C. E.;Claesson, P. M.; Herder, P. C. J . Chem.Soc.,Faraday Trans. 1 1989,85,1933. (21) Rutland, M. W.; Christenson, H. K. Langmuir 1990,6,1083. (22) Rutland, M. W.; Waltermo, A.; Claesson, P. M. Langmuir 1992, 8, 176. (23) Horn, R. G.; Smith, D. T.;Haller, W. Chem. Phys. Lett. 1989,162, 404. (24) Parker, J. L.; Christenson, H. K. J. Chem. Phys. 1988'88,8013. (25) Parker, J. L.; Cho, D. L.; Claesson, P. M. J . Phys. Chem. 1989,93, 6121. (26) Parker, J. L.; et al. J. Colloid Interface Sci. 1990,134, 449. (27) Binnig, G.; Quate, C.; Gerber, G. Phys. Reu. Lett. 1986,56,930. (28) Ducker, W. A.; Senden, T. J.; Pashley, R. M. Nature (London) 1991,353, 239. (29) Mukerjee, P.; Mysels, K. J. Critical Micelle Concentrations of

Aqueous Surfactant Systems; U.S. National Bureau of Standards: Washington, DC, 1970. (30) Deguchi, K.; Meguro, K. J. Colloid Interface Sci. 1972,38,596.

Langmuir, Vol. 9, No. 2,1993 413

Figure 1. Colloid probe. A silica sphere is attached to the pyramidic silicon nitride tip with a small amount of glue. The debris results from the evaporation of solution after an experiment. Unused probes do not show such debris. toa depth of 35nm in purified oxygen at 920 "C. Before mounting in the AFM, both surfaces were exposed to a water plasma (10 W a t a water vapor pressure of 50 mTorr for 30 s) to clean them and to ensure a high density of surface silanol group^.^^*^^ (It has been shown for mica surfaces that plasma treatment does not increase the surface ro~ghness.~~) The silica spheres were attached to microfabricated AFM cantilevers (V-shaped single cantilever springs with integrated silicon nitride tip, Park Scientific, Mountain View, CA) using the epoxy resin Epikote 1004 (Shell). The glue was applied using a thin copper wire on a heating stage under an optical microscope, and the sphere was positioned using another clean wire. On removal from the stage the glue set. Care was taken not to coat the sphere or the reflective gold coating of the cantilever with glue. The glue has been extensivelyused for a related use in the surfaceforce apparatus of Israelachvili,and there is no dissolution in aqueous systems. The sphere attached to a cantilever is henceforth referred to as a "colloid probe". Figure 1 shows a scanning electron microscopy (SEM) image of a colloid probe used in one of these experiments. A more detailed account of colloid probe preparation is given by Ducker et aL3* The maximum peak to peak roughness of the spheres is 3 nm.32 AFM imaging of the silica substrate indicates that the highest asperities were 0.7 nm above mean height with a standard deviation of 0.2 nm.32 Force Measurements. Force measurements were performed by determining the deflection of the cantilever, which, as the spring constant is known, gives the force. The deflection is monitored by movement of a laser beam reflected from the back of the cantilever across a split photodiode. The separation of the surfaces is altered by moving the flat silica substrate relative to the cantilever using a piezoelectric crystal. It was not possible to determine the spring constant experimentally so the manufacturer's specifications were used (0.58 N/m). However, errors in the spring constant only scale the force linearly and do not change the shape of the force profile. The absolute values of the potentials would be affected by an error in this parameter, but since it is changes in potential which are important in the interpretation, rather than the absolute values, this is not a grave concern. Zero force is calculated by determining the regime where there is no change in deflection as the separation is altered. Zero separation is determined from the high force regime where the deflection becomes linear with expansion of the piezoelectric crystal. This regime is referred to as "constant compliance" and is not necessarily silica-silicacontactif there is astrongly adsorbed layer. The gradient of the diode response to the "piezo" expansion at constant compliance also calibrates the deflection of the cantilever in terms of force (using Hooke's law). (31) Senden, T. J.; Ducker, W. A. Langmuir 1992,8,733. (32) Ducker, W. A.; Senden, T. J.; Pashley, R. M. Langmuir 1992,8.

414 Langmuir, Vol. 9, No. 2, 1993 The Nanoscope software produces a "pixel by pixel" screen file on the diode deflection as a function of piezo expansion, and the analysis of this file gives force versus distance curves. As a result of this technique there is a certain digitization error, which is highlighted in Figure 6. In the other force curves the data have been "packed" by averaging every three points determined from the screen file. A more detailed explanation of the force measuring technique is given by Ducker et al.32 Force measurements were performed at mom temperature, i.e., 22 i 3 OC, and were compared with DLVO theory. The double-layer force was calculated assuming identical surface charge using the nonlinear Poisson-Boltzmann (PB) approximation, neglecting ion correlation effects according to the algorithm of Chan et al.33 Although the sign of the charge cannot be directly obtained from the force curve, all potentials were assumed to be negative and the magnitude of the potential was referred to. When no attraction was observed, the van der Waals contribution was neglected. When the gradient of the force, aF/ ax, exceeds the spring constant, k,the surfaces are in an unstable regime and jump to a position of mechanical stability. This leads to what is referred to as a "jump-in", when steeply attractive forces at short separations overcome repulsive forces. When a jump-in was observed,the van der Waals force was approximated with the nonretarded van der Waals interaction for identical surfaces across water with a somewhat arbitrary Hamaker J. This value was chosen since literature constant of 1 X values for the Hamaker constant of silica across water are 0.85 X to 1.7 X J.M (See Appendix). If the surfaces are adhesive, an outward jump can also be observed. As a separating voltage is applied to the piezo the surfaces stay in contact due to their mutual adhesion and at some point (the force minimum) the gradient of the force becomes greater than the spring constant and the surfaces jump apart to the separation they would have reached in the absence of any adhesion. Thus, the longer the jump-out, thegreater the strength of the adhesion. This gives the scaled pull-off force F(O)IR,and the surface energy y can be estimated by the simple formula

F(O)/R = ay (1) where a varies between 3 ' (for ~ soft surfaces) and 47r (for hard surfaces).&38 The substrates used in this experiment fall into the category usually described as elastically "nondeforming", so the value 4r is used. A certain error is incurred in the measurement of outward jumps due to the "roll" of the surface prior to jumping. This arises from the use of a single cantilever spring-and is a phenomeon which could be almost completely avoided by the use of a double cantilever as in the SFA.39," In addition, it has recently been shown that even "hard surfaces" deform locally under the influence of adhesive loads, further complicating the issue.41

Results and Discussion Figures 2 4 show the surface force as a function of distance between a spherical silica colloid probe and a silica substrate, in a background electrolyte of approximately3 X 1W M NaCl in various concentrations of C12E5. Representative force curves have been shown in each case. The data were captured at 2 Hz; i.e., a complete cycle in and out with a piezo took 0.5 8. The measured force F is (33) Chan, D. Y.; Paehley, R. M.; White, L. R. J. ColloidZnterface Sci. 1980, 77,283. (34) Hunter, R. J. Foundatiom of Colloid Science; Oxford University Press: Oxford, 1987; pp 1-222. (35) Johnson, K. L.; Kendall, K.; Roberta,A. D. h o c . R.SOC. London 1971, A324,301. (36) Muller, V. M.; Yushchenko, V. S.; Derjaguin, B. V. J . Colloid Interface Sci. 1989. 92. 92. (37) Chrietenson;H.X.; Clawson, P. M. J. Colloid Interface Sci. 1990, 139, 589. (38) Neumann, A. W.; Moy, E. J. Colloid Interface Sci. 1990,139,591. (39) Jones, R. V. J. Sci. Imtrum. 1951, 38. (40) Christeneon, H. K. J. Colloid Interface Sei. 1988, 121, 170. (41) Parker, J. L.; Attard, P. J. Phys. Chem., in press.

Rutland and Senden 1oo00,

1

0

10

20

90

40

60

Distance (nm)

Figure 2. Forces between a colloidalsilica probe and an oxidized silicon wafer immersed in a 3 X lo-' M NaCl solution plotted on a logarithmic scale. The squares are the average of every three points obtained from a screen file of the photodiode response w piezo expansion and converted to force distance curves by the method outlined in the text. The solid lines are a fit using the nonlinear form of the Poisson-Boltzmann equation according to the algorithm of Chan et ala3*The upper line assumes interaction at constant charge and the lower at constant potential. The parameters of the fit are $ = -70 mV and ~ - 1= 19.0 nm which are in good agreement with previous work. The force may not be fitted with the full DLVO theory as there is an additional repulsive force at short range (50 wt %)F9 it has recently been shown that the sponge phase L3 and La phase can swell to as much as 99 7% water at high temperature (>50 "C).W de Gennes61 has predicted that in "presmectic fluids" the proximity of two walls could induce a phase transition at some critical separation, resulting in an oscillatory but overall repulsive force, in much the same way that thermotropic liquid crystals experiencea local reduction in the phase transition temperature near a surface.B2 This possibility can be discounted,however, as the oscillationsshould correspond (55) Clunie, J. S.;Corkhill,J. M.; Goodman,J. R.; Symone, P. C.;Tate, J. R. J. Chem. SOC., Faraday Trans. 1 1967,63,2839. (56) Clunie.J. S.;Goodman,J. R.; Symons,P. C.J . Chem. Soc.,Faraday Trans. 1 1969,65,287. (57) Ieraelachvili, J. N.; Wennerstrbm, H. Langmuir 1990, 6, 873. (58) Helm.. A. C.:. Ieraelachvili.. J. N.:. McGuirrrran.. P. M. Science 1989. 246, gig. (59) Mitchell,D. J.;Tiddy,F. J.T.;Waring,L.;Boetock,T.;MacDonald, M. P. J. Chem. SOC.,Faraday Trans. 1 1983. (60)Strey,R.;Schom&cker,R.;Roux,D.;Nallet,F.;Olaeon, U. J. Chem. SOC., Faraday Trans. 1990,86, 2253. (61) de Gennes, P. G. Langmuir 1990,6, 1448. (62) Ocko, B.M.; Braelau, A.; Pershan, P. S.;Ala-Nielsen, J.; Deutsch, M. Phys. Rev. Lett. 1986, 57,94. 1 -

Langmuir, Vol. 9, No. 2, 1993 417 to the width of the bilayer, and would be expected to have a longer range. (2) Micelles are adsorbing to the solloid surface. This explanation is also unlikely as, once again, the range of the jumps in Figure 6 would appear to be too small. Surface Potentials. The observed trend in the surface potential tends to support the type of adsorption propounded above. It is not obvious why the surfacepotential should increase when nonionic surfactant is first added (on mica the potential was observed to decrease, but this was explained in terms of the different adsorption sites of metal and hydronium ionsz1). One of the few things one can say is that it is not the result of ionic contaminants. An anionic polyelectrolyte (for example) would not adsorb due to electrostatic repulsion, and a cationic contaminant would reduce the potential. However,the gradualdecrease in potential as the surface aggregatesbecome more densely packed can easily be explained in terms of the lowering of silanol dissociation due to a locally reduced dielectric constant and the fact that the plane of charge, which is presumably still at the silica surface, is not at the contact position. The dramatic drop in potential between 4 X and lo-* M (-90 to -44 mV) is then due to the aggregates coalescing from discrete patches (which expose the silica surfaces to the solution in places) into bilayers (which shield the surface from the solution). The fact that the surfaces retained some residual chargewould seem to indicate that the bilayer was at least slightly defective. Conclusions An atomic force microscope has been used to measure surface forces between a colloidal-sized particle and a smooth silica substrate, and the effect of nonionic surfactant upon these forces has been studied. It would appear that at very low concentrations (-one-third of the cmc) adsorption was sufficientto produce a sparse, weakly hydrophobic layer. A further increase in concentration leads to a reduction in adhesion and the screening of this layer by the adsorption of small aggregates which present a hydrophilic surface to the solution. This is in contrast to what one would expect from considering "classical" (ionic)surfactant adsorption to a (charged)surface where one would first expect an increase in hydrophobicity and growth of the surface monolayer. At concentrations slightly above the cmc the aggregates initially appear to have the character of closely packed regions of intercalated bilayers, which then undergo a transition to classicalbilayer structures. This study also serves to demonstrate the wide application of this technique in surface science, particularly in the study of forces on colloidal-sized particles. Acknowledgment. The authors wishto thankTommy Nylander and Stefan Welin for providing the oxidized silicon wafers and John Parker for the use of the plasma device. Per Claesson, Ric Pashley, Hugo Christenson,and Barry Ninham are thanked for valuable discussions. Appendix

It is very difficult to rigorously calculate avan der Waals force for the case of an adsorbed surfactant layer on silica as in Figures 3,4, and 6 due to uncertainty about the layer thicknesses, position of the "true" silica surface, nature of the silica substrates, and how to treat the adsorbed layer. Figure 8 shows how the effective Hamaker constant will behave as a function of separation of two amorphoussilica

Rutland a d Senden

418 Langmuir, Vol. 9, No.2, 1993 0.6

.......

I

calculated for a 14341 system including retardation effe~ta?~* in different amounta of electrolyte. It is assumed that the oxyethylene chains, particularly for the 1.0-nm layer would be in the water film and can be approximated as water. D = 0 corresponds to contact of the two layers. Such effects do not influence the interpretations of this work, however. The calculations were performed by Dr. P.Y. Kbkicheff.

warn

I

Separation (nm)

Figure 8.

surfaces each bearing (a) a 1.0-nm layer of dodecane and (b) a 5.0-nm layer of dodecane if the interaction is

(63) Lifshitz, E. M. Eksp. Teor. Fiz. 1966,29,94. (64)Lifshitz, E. M. Sou. Phys.-JETP (Engl. Tronel.) 1966,2,73. (65) Dzyaloshinskii, I. E.; Lifshitz, E. M.; Pitaevskii, L. P. Adu. Phys. 1961, 10, 165. (66)Ninham, B. W.; Pareegian, V. A. J. Chem. Phys. 1970,62,4578. (67) Pareegian, V. A.; Ninham, B. W.Biophys. J. 1970,10,646. (68)Mahanty, J. & Ninham, B. W. Dispersion Forces; Academic Press: New York, 1976.