250
Langmuir 1999, 15, 250-258
Adsorption of Pentaethylene Glycol Monododecyl Ether at the Planar Polymer/Water Interface Studied by Specular Neutron Reflection V. A. Gilchrist and J. R. Lu* Department of Chemistry, University of Surrey, Guildford GU2 5XH, U.K.
E. Staples and P. Garrett Unilever Research Port Sunlight, Bebington, Quarry Road East, Wirral L63 3JW, U.K.
J. Penfold ISIS Neutron Facility, Rutherford Appleton Laboratory, CLRC, Chilton, Didcot OX11 0QX, U.K. Received August 21, 1998. In Final Form: October 27, 1998 The adsorption of pentaethylene glycol monododecyl ether (C12E5) onto the surface of poly(methyl methacrylate) (PMMA) has been studied by specular neutron reflection. The polymeric surface was formed by dip-coating a thin layer of PMMA onto a freshly polished silicon oxide block. The structure of the coated thin film was characterized by performing neutron reflectivity measurements under D2O and a mixture of H2O and D2O. The simultaneous model fitting to the measured neutron reflectivity profiles produced a thickness of 52 ( 3 Å for the polymer film. The PMMA density within the coated film was found to be identical to that of the corresponding bulk solid, indicating no appreciable water penetration into the polymeric film. That the structure of the polymer film can be modeled by uniform layer distribution suggests that the outer surface of the thin film was smooth within the experimental resolution. The surfactant adsorption isotherm was determined by making neutron reflectivity measurements in D2O using fully hydrogenated surfactant, hC12hE5. The amount of surfactant adsorbed was found to reach a plateau well below the critical micellar concentration of the surfactant. The limiting area per molecule at the cmc was 51( 3 Å2 and the layer thickness tended to 20 ( 3 Å, suggesting monolayer adsorption. Further detailed structural information of the adsorbed surfactant layer was obtained by using a combination of measurements with chain deuterated and fully hydrogenated surfactants and different water contrasts. The thickness of the dodecyl chain layer was found to be 4 ( 2 Å only, as compared with some 16 ( 3 Å for the ethoxylate headgroup layer, suggesting that the alkyl chains lay flat on the surface of the PMMA while the ethoxylate groups are extended into the aqueous solution. These results, together with our previous work on the adsorption of nonionic surfactant at the hydrophilic silicon oxide/water and hydrophobed silica/water interfaces, show that the structural conformation of the alkyl chains within the adsorbed surfactant layer is heavily affected by the nature of the solid substrate.
Introduction Polymeric colloidal dispersions are widely used in industrial applications such as in surface coatings, adhesives, paper additives, and pharmaceutical formulations. Because of the strong affinity of surfactants toward the surfaces of these colloidal particles, surfactants are widely used to tune the surface properties of the colloidal particles, e.g., surface charge, hydrophilicity, surface tension, and surface viscosity.1 Addition of surfactants can at the same time alter the aggregation state, sedimentation effect, and rheological behavior of the colloidal systems. It is hence desirable to understand the nature of the interaction between various surfactants and polymeric interfaces so that the interfacial properties of the polymeric colloidal suspensions can be effectively manipulated, leading to an improved technological processing. It has long been speculated that the extent of modification to the polymer surface after surfactant adsorption is dependent on the amount of the surfactant adsorbed * Please address all communications to Dr. J. R. Lu, Department of Chemistry, University of Surrey, Guildford GU2 5XH, U.K. (1) Blake, T. D. In Surfactants; Tadros, Th. F., Ed.; Academic Press: New York, 1983.
(surface excess) as well as the structure of the adsorbed layer.2-5 Although the level of adsorption can be estimated from a variety of techniques, the in situ structural conformation of the adsorbed surfactant layer has never been accessible. Surfactant adsorption at the polymeric solid/solution interface has traditionally been studied using polymeric particulate dispersions. Surface excess is estimated from the concentration difference before and after adsorption, the so-called depletion measurements.6,7 This usually requires the knowledge of the precise distribution of the particle sizes, and any uncertainty will affect the total surface area and hence the surface excess. The adsorption isotherms are often compared with different theoretical models, such as the Langmuir equation, the Kronberg equation,4 and the Zhu(2) Zisman, W. A. In Contact Angle, Wettability and Adhesion; Advances in Chemistry Series 43; American Chemistry Society: Washington, DC, 1964. (3) Kronberg, B.; Stenius, P. J. Colloid Interface Sci. 1984, 102, 410. (4) Steinby, K.; Silveston, R.; Kronberg, B. J. Colloid Interface Sci. 1993, 155, 70. (5) Haq, Z.; Thompson, L. Colloid Polym. Sci. 1982, 260, 212. (6) Romero-Cano, M. S.; Martin-Rodriguez, A.; Chauveteau, G.; de las Nieves, F. J. J. Colloid Interface Sci. 1998, 198, 2711. (7) Gau, C.-S.; Zografi, G. J. Colloid Interface Sci. 1990, 140, 1.
10.1021/la9810758 CCC: $18.00 © 1999 American Chemical Society Published on Web 12/11/1998
Adsorption of C12E5 onto PMMA
Gu equation,8 and information about the possible structural conformation of the adsorbed layer is indirectly inferred from the surface coverage together with the known volumes for surfactant molecules. This procedure can result in misleading interpretations. Neutron reflection is a technique which is able to probe the structural information at the solid/solution interface with resolution at the level of a few angstroms. Its combined use with the change of isotopic labeling of the chemical species across the interfacial region makes it very sensitive to the structural distributions of different components at the interface.9,10 We have shown in our previous work that, with the help of deuterium labeling, the technique is capable of revealing the structural composition for the mixed surfactant layers as well as the complex distributions for the surfactant/polymer mixtures.11,12 In this work we use the same technique to study the adsorption of surfactants onto the polymeric solid/ water interface. Of particular interest is the possible plasticization of the polymer surface by surfactants. This can be straightforwardly followed by the change of structural conformation of the polymeric films before and after the polymeric surfaces are in contact with surfactant solutions. The polymer surface used in this work was formed by coating a thin poly(methyl methacrylate) (PMMA) layer onto the surface of a freshly polished silicon oxide. A nonionic surfactant pentaethylene glycol monododecyl ether (C12E5) has been used in this study. Nonionic surfactants are widely used in many industrial processes, partly because their interfacial properties, e.g., surface activity, steric stabilization, can be easily modified by changing the size of the headgroup and partly because they are mild to biological organisms. There have been extensive studies on the interaction of nonionic surfactants with latex particles, with consideration of the effect of the molecular structure of surfactants and the consequence of the polymer surface polarity and charge density.1,5 The measurements of the in situ structure for the adsorbed surfactant layer will provide a useful basis for the interpretation of the role played by nonionic surfactants in stabilizing latex particulate dispersions. Experimental Section The large face of the silicon block (111) was polished using an Engis polishing machine to generate a fresh and smooth hydrophilic surface. The block was lapped on a copper plate with 3 µm diamond polishing fluid and on a pad with 1 µm diamond followed by 0.1 µm alumina fluids. The freshly polished surface was immersed in 5% neutral Decon solution diluted from the concentrated Decon purchased from Decon Laboratories, U.K., and ultrasonically cleaned for 30 min. This was followed by a further 30 min of ultrasonic cleaning in water. The block was then copiously rinsed and soaked in acid peroxide solution (100 mL of 98% H2SO4 in 600 mL of 25% H2O2) for 6 min at 120 °C.15 The block was then thoroughly rinsed with ultrapure water (UHQ) to remove acid and exposed to UV/ozone for 30 min to remove any traces of organic impurities.16 This procedure was (8) Zhu, B.-Y.; Gu, T. J. Chem. Soc., Faraday Trans. 1 1989, 85, 3813. (9) Lu, J. R.; Lee, E. M.; Thomas, R. K. Acta Crystallogr. 1996, A52, 42. (10) Lu, J. R.; Thomas, R. K. J. Chem. Soc., Faraday Trans. 1998, 94, 995. (11) Penfold, J.; Staples, E. J.; Tucker, I.; Thomas, R. K.; Simister, E. A.; Lu, J. R. J. Chem. Soc. Faraday Trans. 1996, 92, 1549. (12) Cooke, D. J.; Blondel, J. A. K.; Lu, J. R.; Thomas, R. K.; Wang, Y.; Han, B.; Yan, H.; Penfold, J. Langmuir 1998, 14, 1990. (13) Gilchrist, V. A.; Lu, J. R.; Staples, E. J.; Garrett, P. To be submitted for publication. (14) Keddie, J. L.; Jones, R. A. L.; Cory, R. A. Faraday Discuss. 1994, 98, 219. (15) Brzoska, J. B.; Shahidzadeh, N.; Rondelez, F. Nature 1992, 360, 719.
Langmuir, Vol. 15, No. 1, 1999 251 found to produce surfaces with reproducible thickness and roughness of the oxide layer, which were completely wetted by water. The thin PMMA film was coated onto the silicon oxide surface by dipping the freshly cleaned block, of the dimension of 12.5 × 5 × 2.5 cm3, into the polymer solution at 25 °C. To reduce the possible effect of mechanical vibration, the whole dipping device was set on an antivibration table. The silicon block was attached in a vertical position to the end of a rod on the pulley system, and the coating was started by dipping the block into the polymer solution followed by lifting it out. Prior to this work the effect of the different system conditions on the coating of the thin polymer films has been characterized.13 A quantitative relationship was established between the quality of the films, in terms of film thickness and outer surface roughness, and the coating parameters such as motor speed, solvent composition, and polymer concentration. The film used in this work was coated at a lifting speed of 0.14 cm min-1 in 1 wt % PMMA solution in tolune, resulting in a final film thickness of 50 ( 10 Å after annealing. Since the glass transition of the polymer was around 120 °C, annealing was done for about 12 h at 150 °C in a vacuum to remove residual solvent and was also found to improve the smoothness of the coated film.14 This procedure was found to form reproducible films on the solid surface, and for the conditions described above the difference in layer thickness was usually within 5 Å. The same coated surface was used throughout the experiment. The neutron reflection measurements were made on the white beam reflectometer CRISP at the Rutherford-Appleton Laboratory, Didcot, U.K.,17 using neutrons of wavelengths from 0.5 to 6.5 Å. The sample cell used was almost identical to that used by Fragneto et al. in ref 18 with the aqueous solution contained in a Teflon trough clamped against the polished surface of the silicon block. The collimated beam enters the end of the silicon block at a fixed angle, is reflected at a glancing angle from the solid/ water interface, and exits from the opposite end of the silicon block. Each reflectivity profile was measured at three different glancing angles, 0.35°, 0.8°, and 1.8°, and the combined results covered a range of momentum transfer, κ, between 0.012 and 0.3 Å -1 (κ ) (4π sin θ)/λ, where λ is the wavelength and θ is the glancing angle of incidence). The beam intensity was calibrated with respect to the intensity below the critical angle for total reflection at the silicon/D2O interface. A flat background determined by extrapolation to high values of momentum transfer was subtracted. For all the measurements the reflectivity profiles were essentially flat at κ > 0.2 Å -1, although the limiting signal at this point was dependent on the H2O/D2O ratio. The typical background for D2O runs was found to be 2 × 10-6 and that for H2O to be 3.5 × 10-6 (measured in terms of the reflectivity). The PMMA sample used in this work has a molecular weight of 101 000 (Aldrich). Its density in solid state at 25 °C is 1.19 g cm-3. Fully hydrogenated nonionic surfactant hC12hE5 (Fluka, 99%) was purified through a silicon column before use. Chain deuterated surfactant (dC12hE5) was synthesized by reacting deuterated bromododecane (Merck Sharp and Dohme, 98% D) with pentaethylene glycol (Fluka, 98+%), and the detailed description for the synthetic process is given in ref 19. H2O was processed through an Elgastat Ultrapure water system, and D2O was purchase from Flourochem (99.9% D). The surface tensions at 25 °C for H2O and D2O were constant at 71.5 ( 1 mN m-1. The physical constants for the surfactant and water used in the data analysis are given in Table 1. All the experiments were performed at 25 °C. The glassware and Teflon troughs for the reflection measurements were cleaned using alkaline detergent (Decon 90) followed by repeated washing in UHQ water. Neutron Reflection. The principle of specular neutron reflection resembles that of light reflection. Theories developed (16) Vig, J. R. J. Vac. Sci. Technol. 1985, A3, 1027. (17) Penfold, J.; Richardson, R. M.; Zarbakhsh, A.; Webster, J. W. P.; Bucknall, D. G.; Rennie, A. R.; Jones, R. A. L.; Cosgrove, T.; Thomas, R. K.; Higgins, J. S.; Fletcher, P. D. I.; Dickinson, E. J.; Roser, S. J.; McLure, I. A.; Hillman, A. R.; Richards, R. W.; Staples, E. J.; Burgess, A. N.; Simister, E. A. and White, J. W. J. Chem. Soc. Faraday Trans. 1997, 93, 3899. (18) Fragneto, G.; Lu, J. R.; McDermott, D. C.; Thomas, R. K.; Rennie, A. R.; Gallagher, P. D.; Satija, S. K. Langmuir 1996, 12, 477. (19) Lu, J. R.; Li, Z. X.; Thomas, R. K.; Staples, E. J.; Tucker, I.; Penfold, J. J. Phys. Chem. 1993, 97, 8012.
252 Langmuir, Vol. 15, No. 1, 1999
Gilchrist et al.
Table 1. Physical Constants for C12 E5 and Water Used in the Calculation species C12D25 (98% D) C12H25 (C2H4)5OH D2O H2O
ba × 105/Å Vb/Å3 F × 106/Å-2 241.2 -13.7 23.0 19.1 -1.7
350 350 330 30 30
6.9 -0.4 0.7 6.4 -0.6
fully extended lengthb
F)
16.7 16.7 19.5
for light reflection are more or less applicable to neutron reflection. Neutron reflection has two distinct advantages over reflection using light or laser sources. First, the wavelength used in neutron reflection is typically a few angstroms, and this is comparable to the dimension of many molecules and thus makes neutron measurements more sensitive to the structural distributions over this range of sizes. Second, neutron signal is related to the socalled scattering length or scattering amplitude, which is a physical constant and which varies from isotope to isotope. Isotopic substitution can be used to highlight the structural profile across the interface without altering the chemical composition of the system. The effectiveness of isotopic substitution can be further explained in the following. Specular neutron reflectivity, defined as the ratio of the intensities of the incoming to the exiting beams, is a function of the Fourier transform of the scattering length density distribution, F(z), normal to the surface or interface. F(z) is related to the chemical composition as follows9,10
∑n b
i i
(1)
where ni is the number density of element i and bi, its scattering length. The variation of values of bi from isotope to isotope allows us to label a given component located at the interface so that different reflectivity profiles can be obtained from a given chemical structure. This can be a great help in revealing the structural details of an interface. This is particularly the case for systems containing hydrogen atoms. For example, the scattering lengths of D and H are of opposite sign, and hence the scattering length density of water can be varied over a wide range, which can be used to highlight an adsorbed polymer or surfactant layer in different ways (contrast variation). Thus, when the solution is made from water consisting of approximately 2 mol of D2O and 3 mol of H2O, which has a scattering length density close to that of silicon, the specular reflectivity is predominantly from the coated polymer layer. Under these circumstances, neglecting the small contribution from the oxide layer on the silicon and assuming a model of a uniform layer distribution for the coated film, the volume fraction of the polymer within the coated layer can be obtained by fitting models to the experimentally measured profiles. The calculated reflectivity profile based on an assumed structural model and using the optical matrix formula20 is then compared with the measured one, and the structural parameters are subsequently modified in a least-squares iteration. The parameters used in the calculation are the thicknesses of the layers, τi, and the corresponding scattering length densities, Fi. Since the scattering length density of a given layer varies with isotopic composition, the fitting of a set of isotopic compositions to a single structural model greatly reduces the possibility of ambiguity in the interpretation, although it adds to the complexity of the fitting procedure. The volume fraction of a given component, e.g., PMMA, in the layer can be determined from the following equation
F ) Fpxp + Fw(1 - x)
bs + nwbw τA
(3)
where bs and bw are the scattering lengths for surfactant and water, and nw is the number of water molecules associated with each surfactant in the layer and is equal to
a See: Sears, V. F. Neutron News 1992, 3. 26. b denotes scattering length. b See Tanford, C. J. J. Phys. Chem. 1972, 76, 3020. V denotes fragment volume.
F)
molecule, A, can be obtained from the following equation, assuming that adsorption does not affect the structural composition of the polymer layer and the underlying oxide
(2)
where Fp and Fw are the scattering length densities for the polymer and water and xp is the volume fraction of the polymer in the layer. Subsequently, if there is a layer of surfactant adsorbed at the solid polymer/water interface, the area per surfactant (20) Born, M.; Wolf, E. Principles of Optics; Pergamon: Oxford, 1970.
nw )
τA - Vs Vw
(4)
where Vs and Vw are the molecular volumes for surfactant and water, respectively. In the calculation to be followed, Vs is taken to be 670 Å3 and Vw is taken to be 30 Å3. The surface excess Γ is related to A by
Γ)
1 NaA
(5)
Results (A) Structure of the Coated PMMA Layer. A native silicon oxide layer exists on the surface of a silicon block. Although the oxide layer is usually very thin on the freshly polished block surface, its composition contributes to the reflected signal. Thus, before the PMMA layer was coated, the thickness and composition of the bare oxide must be measured. The structure of the oxide layer is determined by making neutron reflectivity measurements at the solid/ water interface in the presence of D2O and D2O/H2O mixtures. This is particularly beneficial when the oxide layer contains defects or pinholes since the penetration of water alters the scattering length density of the layer. For example, if D2O is used, the scattering length density of the oxide layer is higher than that of the oxide (3.4 × 10-6 Å-2), and when H2 O is used, the scattering length density of the layer is lower than that of the oxide. The surface of the silicon block used in this work was characterized by measuring neutron reflectivity profiles in D2O, H2O and CM4 (D2 O:H2 O = 2:1). The simultaneous fits to the three reflectivity profiles give a thickness of 12 ( 3 Å and volume fraction of water of 0.25 ( 0.05. The structural composition of the coated polymer layer was determined in a similar manner. Figure 1 shows the reflectivity profiles measured in D2O and CM4. The reflectivity at the bare silicon oxide/D2O interface is also shown for comparison. It can be seen that the addition of the PMMA coating results in a broad interference fringe in the reflectivity profile. The extent of the deviation from the profile for the bare oxide/D2O interface is a measure of the thickness and density of the coated PMMA layer. We did not measure the reflectivity in H2O because model calculations based on the optical matrix formula showed that the reflectivity difference in the measurable κ range is rather small before and after coating if H2O was used as subphase. Comparison with the corresponding measurements at the bare oxide/water interface under different water contrasts suggests that the measurement in D2O is the most sensitive contrast to the structural dimension of the coated polymer film. Quantitative information about the structural distribution of the PMMA layer can be obtained by model fitting the two reflectivity profiles simultaneously. The continuous lines in Figure 1 were calculated assuming that the polymer layer was 52 ( 3 Å thick. The volume fraction of the polymer in the layer was found to be 0.95 from the D2O profile and 1.04 from the CM4 profile; giving an average of the volume fraction of 1 ( 0.05. The structural
Adsorption of C12E5 onto PMMA
Langmuir, Vol. 15, No. 1, 1999 253
Figure 1. Plots of neutron reflectivity profiles versus momentum transfer (κ) measured in D2 O (b) and CM4 (+). The continuous lines were calculated assuming the PMMA layer was 52 ( 3 Å and the fraction of water within the PMMA layer was less than 5%. The underlying oxide layer was taken to be 12 ( 3 Å thick and contained some 25% water. The reflectivity from the bare oxide/D2 O interface is shown as a dashed line for comparison.
composition of the oxide layer has been taken to be the same as the bare oxide; that is, coating did not affect the structure of the oxide layer and its water content. No roughness was required for the polymer/water interface, indicating that the outer polymer surface was very smooth. It should be noted that the calculated volume fraction of polymer and that of water are affected by the uncertainty in the scattering length density for PMMA, which in this case was just calculated from its known bulk density and molecular weight. Since molecular weight is an average figure and is affected by the degree of the polydispersity, which is unknown for this PMMA sample, this uncertainty will be directly related to the volume fraction of PMMA obtained for the layer. Nevertheless, the range of error obtained in the layer composition is entirely consistent within about 5%, comparable to the level of instrumental resolution. The results from neutron reflection thus suggest that within a few percent of error the coated film is composed of solid PMMA. It is possible that some PMMA fragments may penetrate into the defects in the oxide layer and the maximum volume fraction of the polymer in the oxide layer is 0.25 if the voids are filled completely. If this is the case, the average scattering length density for the oxide layer contains the contributions from the oxide, D2O, and the polymer, and the relationship can be expressed as
F ) xpFp + (0.25 - xp)Fw + 0.75FSiO2
(6)
where FSiO2 is the scattering length density of silicon oxide (3.4 × 10-6 Å-2). Thus the mixing of polymer into the oxide layer will effectively reduce the scattering length density for the layer. We have tried to refit the two reflectivity profiles shown in Figure 1 by lowering the scattering length density for the oxide layer while its thickness was retained at 12 Å. It was found that the scattering length density for the oxide layer could not be reduced to less than 3.8 × 10-6 Å -2 beyond which the fitting became obviously poor. According to eq 6 the value of F of 3.8 × 10-6 Å-2 is equivalent to about 7% polymer. It has to be
noted that in adjusting the quality of the fit the thickness of the PMMA layer had to be reduced by about 1-2 Å. This would suggest that although there is some uncertainty in distributing the polymer at the oxide/PMMA film interface both models appear to produce the same amount of PMMA. It can be seen from Figure 1 that the calculated reflectivity curves fit the measured ones well up to 0.12 Å-1, beyond which the fits are less good. Further tests have been made to examine if the use of roughness between different interfaces can improve the fitting over the high κ. This was done by introducing the roughness at each interface separately and at all three interfaces simultaneously. It was found that when roughness at each interface was less than 3 Å their effects to the reflectivities were weak. The use of greater roughness forced changes in the thicknesses for the two layers and in all cases did not lead to any improvement in the fitting at κ > 0.12 Å-1. It is nevertheless not surprising to reach such conclusion because the reflectivities above 0.12 Å-1 are around (1-5) × 10-6 and close to the level of uncertainty in the background subtraction. It can be summarized from the analysis given above that although the measurements do not offer a clear answer to the degree of penetration of the polymer into the oxide layer, the fittings clearly show that the outer surface of the coated film is within 2-3 Å smooth. The use of roughness at different interfaces did not show any obvious improvement in the fittings. Since a single-coated PMMA surface was used throughout the whole experiment, any possible change in the structure of the PMMA film would result in serious errors in the measured reflectivity profiles. The stability and reproducibility of the polymer film in contact with water were studied by measuring the reflectivity as a function of time at different stages of the experiment. Each reflectivity profile was measured over the whole κ range over a period of 40 min. The measurements were repeated at 2 h intervals, over a period of 12 h. No observable change in reflectivity profiles suggests that the coated film was very stable and no time dependent swelling or deterioration occurred. When the nonionic surfactant C12E5 was introduced into the sample cell the presence of surfactant molecules might deteriorate the polymer film by preferential adsorption onto the bare oxide surface or by penetration into the polymer thin layer. In the latter case the physical properties of the polymeric film could be altered and its structural composition might be different. The response of the thin polymer film to the presence of C12E5 was monitored during the whole course of the experiment. Figure 2 compares the reflectivity profiles between the PMMA coated solid/pure D2 O interface at the different stages of the experiment. Before these reflectivity profiles were recorded the sample cells were rinsed with pure water to remove the loosely adsorbed nonionic surfactant. No measurable difference between the profiles shown in Figure 2 suggests that the film remained intact during the whole process of the experiment. The result thus shows that the nonionic surfactant did not damage or penetrate the PMMA film. A further conclusion which can be drawn from the data shown in Figure 2 is that adsorption of the nonionic surfactant onto PMMA surface is completely reversible. (B) Surface Excess of C12E5. The amount of surfactant adsorbed at different bulk surfactant concentrations was determined from reflectivity measurements with the fully hydrogenated surfactant in D2O. The variation of the measured reflectivity profiles with [C12E5] is shown in Figure 3. It can be seen that adsorption at the critical micellar concentration (cmc) of C12E5 produces the greatest deviation from the reflectivity from the PMMA coated
254 Langmuir, Vol. 15, No. 1, 1999
Gilchrist et al.
Figure 2. Plots of neutron reflectivity profiles versus κ measured in D2O at the beginning (b), the middle ((+), and before the end (4) of the surfactant adsorption experiment. The continuous line was calculated assuming that the PMMA layer was 52 ( 3 Å and the fraction of water within the PMMA layer was less than 5%. The underlying oxide layer was taken to be 12 ( 3 Å thick and contained some 25% water. The polymer surface was rinsed with water many times to desorb surfactants before the reflectivity profiles were recorded. That the reflectivity profiles are identical shows that the PMMA surface remained intact and surfactant adsorption was completely reversible.
Figure 4. (a) Surface excess of C12E5 at the solid PMMA/water interface from neutron reflection (b) compared with that measured from particulate suspensions (+). (b) Variation of the thickness of the overall C12E5 layer at the solid PMMA/ water interface. Continuous lines are drawn through the points to guide the eye.
Figure 3. Plots of neutron reflectivity profiles measured in D2O at the surfactant concentration of 0 M (b), 1 × 10-5 M (]), 4 × 10-5 M (4), and 8 × 10-5 M (+). The continuous lines were calculated assuming uniform layer model distributions for the surfactant layers, and the corresponding surface excesses and layer thicknesses are summarized in Figure 4.
solid/D2O interface, indicating that adsorption attains the highest surface excess when the cmc is approached. As surfactant bulk concentration is lowered the amount of adsorption onto the PMMA/water interface is reduced and the reflectivity profile becomes close to that from the PMMA-coated solid/D2O interface. The solid lines through the reflectivity profiles in Figure 3 were calculated using a single uniform layer model to represent the distributions of the adsorbed surfactant layer. That the calculated curves fit the measured data so well indicates that the
simple model is a good representation of the actual distributions for the surfactant layers. In calculating the reflectivity profiles under different bulk concentrations of C12E5 the structural composition for the oxide layer was taken to be composed of 75% SiO2 and 25% water and the PMMA layer was taken to be unaffected. If 7% mixing of polymer into the oxide layer was considered together with a subsequently thinner layer of PMMA of 51 Å the thicknesses and the amount of surfactant in the surfactant layers were found to be consistent with those corresponding to the reflectivity profiles calculated in Figure 3. The variation of surface excess and layer thickness with bulk C12E5 concentration is shown in Figure 4. It can be seen that surface excess tends to a plateau at a concentration well below the cmc. A similar trend is seen for the variation of the surfactant layer thickness. The total surfactant layer is very thin at the low bulk concentration and gradually increases as the cmc is approached. There have been extensive studies of surfactant adsorption onto solid polymeric interfaces from aqueous and nonaqueous solutions. Most of these studies were made on polymeric
Adsorption of C12E5 onto PMMA
particulate suspensions where surface excess could be determined by measuring the change in surfactant concentration in bulk solution before and after adsorption. For comparison we show in Figure 4 the surface excess data from Gau et al.7 on the adsorption of C12E5 onto PMMA latex particles in aqueous solution. Two sets of data have very different shape and are overall not in particularly good agreement. The agreement is reasonably good at the low concentration region, but significant discrepancies appear at higher concentration. For the neutron reflection, the sensitivity increases with surfactant concentration, while for particulate systems the opposite trend holds. The amount of surfactant adsorbed on the surface of particles is usually calculated by determining the differences in bulk solutions before and after adsorption. At the low bulk concentration, adsorption causes a relatively large change in bulk concentration and since surface excess is directly proportional to the concentration difference the surface excess is therefore reasonably reliable. However, the reliability of the results starts to decrease with the increase of bulk concentration as a result of having to obtain the small difference between the two large numbers. It can be seen from Figure 4 that the deviation between the two sets of data starts from the region approximately at half cmc, where a sharp upturn in surface excess occurs in the plot by Gau et al. Their results imply that there is some kind of change either in adsorption mechanism or in the structural conformation of the surfactant layer. As the physical significance of such discontinuity has been much discussed in the literature it is necessary to discuss the reliability of the neutron data before the indication of the results by Gau et al. is assessed. Equations 3-5 define the relationship between scattering length density of the layer and the area per surfactant molecule. Uncertainty in F and τ can be estimated in terms of possible errors to A. As already indicated above, the sensitivity of the structural parameters used in the fits was strongly concentration dependent. At the concentration around the cmc, the fitting was easily sensitive to the variation of (0.1 × 10-6 Å-2 in F and (2 Å in τ. These errors could be converted into the uncertainty in A, which was found to be within (3 Å2. However, the sensitivity deteriorates as surface excess decreased. At [C12E5 ] ) 1.5 × 10-5 M the uncertain range in F was (0.6 × 10-6 Å-2 and in comparison for τ this was within (4 Å. The consequent effect when combining these errors leads to an enlarged error of (10 Å2 to A. This analysis shows that within the quoted error the results from neutron reflection are reasonably reliable and that the trend of neutron surface excess shown in Figure 4 is therefore convincing. The adsorption of a range of nonionic surfactants CnEm at the air/water and solid/water interfaces has been examined by neutron reflection. It is interesting to compare these surface excesses with the results of Gau et al. The limiting surface excess above the cmc from Gau et al. is about 4.5 × 10-6 mol m-2 at the solid PMMA/water interface, which corresponds to the limiting area per molecule of 37 Å2, as compared with a value of 51 Å2 from our neutron measurement. On the other hand, our limiting area per molecule of 51 Å2 is close to the limiting value of 50 ( 3 Å2 from the adsorption of C12E4 onto the hydrophobed silicon oxide/water interface studied by Fragneto et al.18 who modified the solid surface by chemically grafting a self-assembled monolayer of octadecyl trichlorosilane (OTS). Given that the hydrophobicity of the two solid surfaces is very different and that adsorption is hydrophobically driven, it is difficult to understand how the PMMA surface attracts the amount
Langmuir, Vol. 15, No. 1, 1999 255
Figure 5. Plots of neutron reflectivity profiles versus κ to show the adsorption of fully hydrogenated C12E5 in D2O (b), CM4 (+), and CMSi (4). The continuous lines were calculated assuming a 20 Å C12E5 layer on top of 52 Å PMMA layer. The underlying oxide layer was taken to be 12 ( 3 Å thick and contained some 25% water.
of surfactant which is comparable to that adsorbed at the hydrophobic solid/water interface. Furthermore, the limiting area at cmc at the air/water interface was found to be 36 ( 2 Å2 for C12E3, 44 ( 2 Å2 for C12E4, and 48 ( 3 Å2 for C12E5. The limiting value from the work of Gau et al. would correspond to that for C12E3 instead of that for C12E5, a trend which is inconsistent with what is usually expected. (C) Structure of the Adsorbed C12E5 Layer. Many results reported elsewhere on the adsorption of surfactants onto polymeric solid surfaces have been interpreted in terms of possible differences in the structural conformation of surfactant layers. However, there have been no techniques which are sensitive enough to provide reliable information about the structure of the adsorbed surfactant layer at the solid polymer/water interface. The high depth resolution of neutron reflection combined with deuterium labeling offers sufficient sensitivity not only for distinguishing the solid surface from the adsorbed surfactant layer but also for revealing the detailed information of the structure of the surfactant layer. The contrast for the components across the interfacial region can be easily varied by labeling the solvent, the surfactant, and the polymer or any combination of the three. In this study we have made measurements by varying the contrast of water and by labeling the dodecyl chain of the surfactant (dC12hE5). We show in Figure 5 the reflectivity profiles of hC12hE5 measured at the cmc in three different water contrasts, D2O, CM4, and CMSi (H2O: D2O = 3:2). The variation of water contrast serves to highlight different aspects of the surfactant layer. The continuous lines through the measured profiles in Figure 5 were calculated by taking the surfactant layer to be 20 ( 2 Å thick. The volume fraction of the surfactant within the layer was calculated to be 0.61 ( 0.03. This was equivalent to a mean area per surfactant of 51 ( 3 Å2, as already discussed earlier, and the number of water molecules associated with each surfactant in the layer was about 12 ( 3. It is interesting to note that the minimum number of water molecules associated with each ethoxylate group is usually 2 when the adsorbed layer is saturated at the air/water interface. This result thus suggests that the nonionic surfactant layer at the PMMA/
256 Langmuir, Vol. 15, No. 1, 1999
Gilchrist et al. Table 2. Structural Parameters at the cmc Obtained from Fitting a Two-Layer Model isotope
A ( 3/Å2
τ1 ( 2/Å
τ2 ( 3/Å
(F1 ( 0.3) × 106/Å-2
(F2 ( 0.3) × 106/Å-2
dC12 hE5/D2 O dC12hE5/H2O dC12hE5/CM4 hC12hE5/D2O hC12hE5/CM4 hC12hE5/CMSi
51 52 50 51 53 52
4 4 4 4 3 4
15 15 17 17 17 17
6.8 6.5 -0.2 -0.2 -0.2 -0.1
3 0.4 1.7 3 1.6 1.1
Table 3. Structural Parameters for hC12hE5 in D2O Obtained from Fitting a Two-Layer Model
Figure 6. Plots of neutron reflectivity profiles versus κ to show the adsorption of chain deuterated C12E5 in D2 O (b), CM4 (4), and H2O (+). The continuous lines were calculated assuming a thickness of 4 ( 1 Å for the dodecyl chain layer and of 16 ( 3 Å for the ethoxylate headgroup layer. The PMMA layer was fixed at 52 Å, and the underlying oxide layer was fixed at 12 Å.
water interface is rather closely packed. Since the thickness and packing density of the surfactant are very much similar to the layers formed at the air/water interface and those at the hydrophobed solid/water interface, we can conclude that surfactants adsorbed onto the solid PMMA/ water interface form a monolayer. Further structural information inside the surfactant layer, e.g., the structural conformation of the alkyl chain and the ethoxylate headgroup, can be obtained by making measurements with partially labeled surfactant. We show in Figure 6 the reflectivity profiles for dC12hE5 measured at the cmc in D2O, CM4 and H2O. It can be seen that unlike the profiles shown in Figure 5 the labeling to the surfactant has changed the shape of the reflectivity profiles. The shape of the reflectivity profile for the measurement in H2O with dC12hE5 has changed most substantially. At this contrast the ethoxylate headgroup is more or less invisible because the scattering length density of the headgroup is very close to that of the H2O and what is then determined is the distribution of the alkyl chain. Because the dodecyl chain is deuterated, the scattering length density of the alkyl chain layer will be very high if there is a small amount of water mixed with it. We started the model fitting using a uniform layer model for the distribution of the alkyl chain. It was found that in order to produce the shape of the reflectivity profile in H2O the total dodecyl chain layer had to be as narrow as 4 ( 2 Å. The error quoted here reflects the actual tolerance of the fit to the thickness variation. The positions of the two broad interference fringes in the reflectivity curve are very sensitive to the change in the thickness of the alkyl chain layer. The solid line through the profile of dC12hE5 in H2O was calculated assuming that the alkyl chain region was 4 Å; the headgroup layer was some 16 Å. It was found that the model was very sensitive to the thickness of the dodecyl chain but less sensitive to that of the ethoxylate headgroup as F for the ethoxylate layer was close to that of water. The same model was used to calculate the reflectivity profiles for the measurements in D2O and CM4. In both cases the calculated curves fitted the measured ones well, as can be seen from Figure 6.
concentration × 10-5/M
A ( 3/Å2
12 (2 cmc) 6 (cmc) 4 1.5 1
50 51 54 70 ( 5 190 ( 30
(F1 ( 0.3) (F2 ( 0.3) τ1 ( 1/Å τ2 ( 3/Å × 106/Å-2 × 106/Å -2 3 4 5 3 2
19 17 18 14 7
-0.2 -0.2 0.0 0 0
3 3 3.3 3.1 4
Although structural parameters used in calculating the reflectivity profiles under these two contrasts were slightly altered, the thickness of the alkyl chain layer was always below 5 Å and that for the ethoxylate headgroup around 17 ( 3 Å. As the structure of the surfactant layers under other contrasts involving hC12hE5 have so far been fitted into a single uniform layer, it is useful to refit the data using the two layer model developed from the partially deuterated surfactant. This will help to check the consistency of the data and also to make the comparison more straightforward. Table 2 lists the structural parameters for the measurements at the cmc under different labeling and Table 3 lists those using hC12hE5 under different surfactant concentrations. As in the case of single layer model fitting the thicknesses for the oxide layer and the PMMA layer were taken to be constant at 12 and 52 Å, respectively. The compositions of the oxide and PMMA layers were also assumed to be constant with appropriate variation in scattering length densities to take into account the change in isotopic labeling. Since there are already two layers present (the oxide and PMMA film) at the interface the reliability of the fitted parameters for the surfactant layers obtained from the two layer fit will be reduced. The numbers shown in the two tables are the best fits and the variation in the thicknesses shown in Table 2 reflects the level of experimental error as well as the change in response to contrast variation. The main observation is that under all solution conditions the alkyl chain layer is always under 5 Å. The fitted scattering length densities tend to suggest that the volume fraction of the dodecyl chain in the alkyl chain layer is about 0.95 ( 0.05 and the volume fraction of the pentaethylene glycol groups in the headgroup layer is about 0.6 ( 0.1. The reliability of the division between the two layers is largely affected by the contrast in scattering length densities between them. We have already explained that in the case of dC12 hE5 good contrast was achieved due to the deuterium labeling to the dodecyl chain. In the case of hC12hE5 in D2O good contrast could also be obtained as the mixing of D2O with the headgroup helped to enhance the scattering length density for this part of the layer, as can be clearly seen in Tables 2 and 3. Discussion The dimension for the alkyl chain and the headgroup layers tends to suggest that the alkyl chain lies flat on the surface of PMMA and that the headgroups extend into bulk solution. This is a rather different picture from what
Adsorption of C12E5 onto PMMA
was observed from neutron reflection measurement for the adsorption of C12E4 at the alkylated hydrophobic solid/ water interface where the thicknesses for the alkyl chain and ethoxylate head were both found to be about 10 ( 3. The thicknesses for the two groups within the C12E4 layer thus indicated that both fragments adopted an average tilt angle of 50° from the surface normal. The difference in the structural conformation between the two systems must originate from the difference in the polarity of the surface. It is of interest to note that Thirtle et al.21 have recently studied the adsorption of nonionic surfactants CmEn at the hydroxylated solid/water interface using neutron reflection. The OH groups on the outer surface were formed by chemically grafting a monolayer of long chain alcohol onto the surface of silicon oxide. They have observed a total layer thickness of some 4 ( 1 Å for different CmEn, suggesting that the whole CmEn molecule lies completely flat on the hydroxylated surface, regardless of the size of the alkyl chain and headgroups. In view of these observations it can be said that the alkyl chain adopts a similar conformation at the PMMA/water interface as that at the hydroxylated solid/water interface, but the headgroups somehow take a different conformation. This is most likely due to the difference in the level of hydrogen bonding. At the PMMA surface, hydrogen bonding has to be established through water molecules, while at the hydroxylated solid surface direct hydrogen bonding can occur between the oxygens on the ethoxylate chains and OH groups on the solid surface. The latter must be energetically favored. The reliability of the information about the surfactant layer heavily hinges on the quality of the coated polymer layer. This includes the smoothness of the outer polymer surface and the extent of defects within the coated polymeric film in the form of gaps or voids. The characterization measurement on the thin film alone shows that within the experimental resolution of 5% the film is composed of 100% solid PMMA, suggesting that there are virtually no voids or gaps inside the film. Furthermore, no roughness required in the fitting between any interfaces suggests that the outer polymer surface is reasonably smooth. The wetting and spreading of PMMA thin films on the surface of silicon oxide have been widely investigated in the literature. Smooth films can be formed mainly as a result of the affinity of the solid substrate toward PMMA polymer. The spreading of the coated thin film is controlled by the physical behavior of the polymer at the air/film and film/substrate interfaces, which dictates the degree of smoothness on the outer surface and the packing density inside the film. It has been demonstrated that the features of the thin PMMA films are dominated by the interaction between the film and the solid substrate. One of the key parameters marking the change of physical nature of the thin film is its glassy phase transition temperature, Tg. Using ellipsometry, Keddie et al.14 have studied the effect of hydrophobicity of different solid substrates on Tg for various polymeric thin films. They observed that when the PMMA film was below 1000 Å, Tg showed a trend of increase with the decrease of the thickness for the coated thin PMMA films on the surface of silicon substrate. Over the thickness range 100-1000 Å Tg was found to increase by about 5°. This increase was attributed to the hydrogen bonding between the OH groups on the surface of the oxide and the oxygens on the ester groups. As film thickness decreases, the property of the thin film is increasingly affected by the level of (21) Thirtle, P. N.; Li, Z. X.; Rennie, A. R.; Satija, S. K.; Sung, L. P. Langmuir 1997, 13, 5451.
Langmuir, Vol. 15, No. 1, 1999 257
hydrogen bonding. The affinity of the PMMA layer toward the surface of silicon oxide has also been demonstrated by the preferential segregation of PMMA fragments onto the oxide surface in the symmetric diblock copolymers of polystyrene-poly(methyl methacrylate) (PS-PMMA) when the film was heated above its Tg. In the current study the structural conformation of the adsorbed surfactant layer actually works as a probe for detecting the smoothness of the outer polymer surface. Since the thickness of the dodecyl chain layer features the minimal dimension required for the sideways-on alkyl chain conformation, the observation thus indicates that the polymer surface underneath the surfactant alkyl chain is not so uneven. Otherwise, the staggered outer polymer surface would broaden the width of the alkyl chain distribution. Since the size of a water molecule is about 3 Å in diameter, the result indicates that the outer surface of the coated thin PMMA film is indeed very smooth. Such a surface is ideal for exploring the features of interaction between surfactants and polymeric interfaces without the complication arising from the uncertainty about the structure of the polymer surface itself. The wetting of various solid substrates by surfactants has been studied extensively. The interest has been focused on quantifying the thermodynamic relationship between contact angle and the extent of surfactant adsorption. For very nonpolar solids such as paraffin and polyethylene, it has been established that the relationship between the advancing contact angle and the surface excess can be represented by the combined application of the Gibbs and Young equations7,22
∂(γLV cos θ) ΓSV - ΓSL ) ∂γLV ΓLV
(7)
where γLV is the surface tension of surfactant solution at the vapor/solution interface, θ is the contact angle, and ΓSV, ΓSL, and ΓLV are the surface excesses of the surfactant at the vapor/solid, solid/liquid, and liquid/vapor interfaces, respectively. The term γLV is the measure of adhesion tension.22 For the adsorption of hydrocarbon surfactants onto these low energy solid surfaces a plot of γLV cos θ versus γLV has been shown to give a straight line with a slope around -1. It has been suggested that for surfactant adsorption ΓSV ) 0, because surfactant is hardly volatile at room temperature. Thus a constant slope of -1 in eq 7 would indicate that ΓSL and ΓLV are approximately equal. This has indeed been found to be the case for the adsorption of hydrocarbon surfactants on paraffin,23-25 suggesting that adsorption at hydrophobic surfaces follows the combined relation of Gibbs and Young equations established in eq 7. For the adsorption of nonionic surfactants CnEm onto PMMA Gau et al.7 found that the slopes for the plots of γLV cos θ versus γLV were very close to zero and the slopes were between 0 and -1 for the adsorption on PS surface. Since the contact angle is about 70° for PMMA, 90° for PS and 110° for paraffin, this observation would suggest that the absolute value of the slope increases with the hydrophobicity of the solid substrates. If ΓSV is again taken to be zero, the trend of the contact angle would suggest that ΓSL/ΓLV increases with the hydrophobicity of the solid surfaces and is in all cases less than or equal to 1. (22) Pyter, R. A.; Zografi, G.; Mukerjee, P. J. Colloid Interface Sci. 1982, 89, 144. (23) Lucassen-Reynders, E. H. J. Phys. Chem. 1963, 67, 969. (24) Bargeman, D.; van Voorst Vader, F. J. Colloid Interface Sci. 1973, 42, 467. (25) Wolfram, E. Kolloid-Z 1966, 211, 84.
258 Langmuir, Vol. 15, No. 1, 1999
The adsorption of C12E5 at the air/water interface has been studied by neutron reflection,26 and the surface excesses are thus available for comparison with the corresponding values at the solid PMMA/water interface obtained from this work. Although ΓLV and ΓSL from neutron measurements can easily suffer from a few percent of error, these results are intrinsically more reliable than both ΓLV from the Gibbs equation combined with surface tension measurements and ΓSL extracted from the depletion of surfactant concentration in particulate suspensions.7 At the concentrations of 7 × 10-5 M (just above the cmc), 4 × 10-5 M, 1.5 × 10-5 M, and 1 × 10-5 M for C12E5, ΓSL/ΓLV was found to be 0.94, 0.95, 0.92, and 0.6. Thus, apart from the last value at the lowest surfactant concentration where the measurement at the solid PMMA/ water interface suffers from a much greater degree of error, the data show that within an error of 10% the ratio of ΓSL/ΓLV is close to 1. The results thus demonstrate that the feature of adsorption at the semipolar PMMA/water interface is more or less similar to what was observed for the hydrophobic solid/water interface. Further, if the assumption about γSV still holds the predicted slope for the plot of γLV cos θ versus γLV would have to be -1. That the wetting measurements of Gau et al.7 are inconsistent with the trend predicted by neutron surface excesses may underline that either θ or γLV or a combination of both in their work suffer from large experimental errors. Conclusions We have demonstrated that information about the adsorption of surfactants at the solid polymer/water interface can be obtained from neutron reflection. The key step in performing neutron experiments is to coat a thin layer of polymeric film onto a neutron transparent substrate. That the density of the coated PMMA film is similar to that of the bulk PMMA at the same temperature suggests that the structural information derived from the thin polymeric film/water interface is directly comparable (26) Lu, J. R.; Li, Z. X.; Thomas, R. K.; Binks, B. P.; Crichton, D.; Fletcher, P. D. I.; McNab, J. R.; Penfold, J. J. Phys. Chem. B 1998, 102, 5785.
Gilchrist et al.
with the interfacial systems which are of direct practical interest. Although the coating of a thin polymeric film onto the solid substrate can reduce the sensitivity of the neutron measurements, the use of contrast variation to water and surfactants provides sufficient sensitivity. A single-coated PMMA surface was used during the whole experiment. This has enabled us to compare the level of adsorption under different concentrations rather straightforwardly. That this has been possible simply because the PMMA layer was robust and the interaction between the nonionic surfactant and PMMA film was strictly through physical adsorption. There was no indication of penetration of surfactant into the polymer film. The structural measurements from this work show that adsorption of C12E5 onto the PMMA/water interface forms a surfactant monolayer of some 20 Å thick. The dodecyl chains adopt a sideways-on conformation with thickness of the alkyl chain layer of some 3-5 Å. In contrast, the dimension of the ethoxylate headgroup layer is much greater, with the thickness around 17 ( 3 Å. The structural conformation adopted by the alkyl chain is clearly different from that of the nonionic surfactant at the hydrophobic solid/water interface where both alkyl chain and the ethoxylate headgroups adopt a headways-on conformation with an average tilt angle of 50° from the surface normal. That the conformation of the alkyl chain layer at the PMMA/water interface is similar to that at the hydroxylated solid/water interface suggests that the conformation of the alkyl chains is controlled by the polarity of the solid surface. Since the polarity of the solid surface is usually represented by contact angle, the result from this work tends to suggest that there is an apparent relationship between the conformation of the alkyl chains and the variation of contact angles on the solid substrates. Acknowledgment. We thank the Engineering and Physical Sciences Research Council (EPSRC) for support. V.A.G. thanks the Unilever Port Sunlight Research Laboratory for partial studentship. We also thank all four reviewers for their helpful comments and suggestions. LA9810758
