Adsorption Studies of a Polymerizable Surfactant by Optical

Sep 2, 2009 - Department of Chemical Engineering, Center for Complex Fluids Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213 ...
1 downloads 8 Views 3MB Size
pubs.acs.org/Langmuir © 2009 American Chemical Society

Adsorption Studies of a Polymerizable Surfactant by Optical Reflectivity and Quartz Crystal Microbalance Chris Hodges,*,† Simon Biggs,† and Lynn Walker‡ †

Institute of Particle Science and Engineering, University of Leeds, Leeds LS2 9JT, U.K., and ‡Department of Chemical Engineering, Center for Complex Fluids Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213 Received April 15, 2009. Revised Manuscript Received June 4, 2009

A quartz crystal microbalance (QCM) and an optical reflectometer (OR) have been used to investigate the adsorption behavior of two different variants of the surfactant-hydrotropic counterion system, alkane trimethylammonium vinylbenzoate (CnTVB), onto silica surfaces. The C18TVB variant, with a longer hydrocarbon tail, produced a threestage adsorption isotherm in the OR, whereas the C16TVB surfactant showed a two-stage adsorption isotherm. This was explained in terms of the greater degree of hydrophobicity of the C18 carbon chain requiring a significantly higher concentration of surfactant to be present on the surface before any further adsorption can occur. A concentration dependent adsorption rate was observed for both surfactants, with the faster adsorption rate being detected for C18TVB. The OR data showed that each surfactant could be completely rinsed off with the flow of water into the OR cell. This was not observed with the QCM data, where only a partial rinse off was seen. The difference between the two techniques was hypothesized to be due to the ability of the QCM to detect both interfacial and bulk behavior thus complicating the interpretation of the adsorption data.

Introduction The capacity of small-molecule surfactant films on surfaces to form nanoscale structures of different types, depending on the surface charge and surfactant concentration, has been widely reported.1,2 These surface structures are related to the structures reported in the bulk although the exact details of the interplay between the two is not yet clear. These self-ordering layers have considerable advantages at a commercial scale because of the time and cost saved over lithography and other techniques. In addition, such systems offer the possibility to coat complex surface shapes and particles as opposed to flat films and surfaces. The behavior of systems based on a series of cationic surfactant with hydrotropic counterion (CnTVB) is an interesting addition to this body of work as initial results indicate that various shapes are formed onto different surfaces,3 and bulk properties particularly at concentrations above the cmc have been characterized.4 However the nature of the way in which these structures adsorb has not been investigated to date. We have carried out characterization of the polymerized version of these surfactants, which results in polyelectrolyte-surfactant aggregates, both in the bulk5-8 and at solid-liquid interfaces.3,9,10 We have recently demonstrated that although the initial adsorption of the polymerized aggregates may be quite rapid, a significant period of time *To whom correspondence should be addressed. E-mail: c.s.hodges@ leeds.ac.uk. (1) Atkin, R.; Craig, V. S. J.; Wanless, E. J.; Biggs, S. Adv. Colloid Interface Sci. 2003, 103(3), 219–304. (2) Warr, G. G. Curr. Opin. Colloid Interface Sci. 2000, 5(1-2), 88–94. (3) Biggs, S.; Kline, S. R.; Walker, L. M. Langmuir 2004, 20(4), 1085–1094. (4) Kline, S. R. Langmuir 1999, 15(8), 2726–2732. (5) Gerber, M. J.; Kline, S. R.; Walker, L. M. Langmuir 2004, 20(20), 8510–8516. (6) Gerber, M. J.; Walker, L. M. Langmuir 2006, 22(3), 941–948. (7) Kuntz, D. M.; Walker, L. M. J. Phys. Chem. B 2007, 111(23), 6417–6424. (8) Kuntz, D. M.; Walker, L. M. Ind. Eng. Chem. Res. 2009, 48(5), 2430–2435. (9) Biggs, S.; Labarre, M.; Hodges, C.; Walker, L. M.; Webber, G. B. Langmuir 2007, 23(15), 8094–8102. (10) Biggs, S.; Walker, L. M.; Kline, S. R. Nano Lett. 2002, 2(12), 1409–1412. (11) Hodges, C. S.; Biggs, S.; Walker, L. M. Langmuir 2009, 25(8), 4484–4489.

Langmuir 2009, 25(19), 11503–11508

may be required for the system to come into equilibrium.11 This may have implications if these systems were to be applied to coat a range of surfaces. Here, we investigate the adsorption of the unpolymerized surfactant since this may lead to insights into the behavior of the polymerized aggregates. Although no published work is available for the adsorption of these CnTVB surfactants, there is significant literature characterizing the adsorption of surfactants based on the cetyltrimethylammonium (CTAþ) ion with different counterions. These papers include atomic force microscope (AFM) structural studies,12-14 quartz crystal microbalance (QCM) adsorption studies,15-17 optical reflectometry (OR) studies,18-21 neutron and X-ray scattering studies.22-24 Overall, these investigations have suggested that a range of different structures may be formed depending on the surface type and surfactant concentration, and that most of these structures are formed quite quickly over just a few minutes.25 The rinse off behavior of CTAC has been found from AFM studies25 to consist of at least two stages, (12) Manne, S.; Cleveland, J. P.; Gaub, H. E.; Stucky, G. D.; Hansma, P. K. Langmuir 1994, 10(12), 4409–4413. (13) Lamont, R. E.; Ducker, W. A. J. Am. Chem. Soc. 1998, 120(30), 7602–7607. (14) Ducker, W. A.; Wanless, E. J. Langmuir 1999, 15(1), 160–168. (15) Knag, M.; Sj€oblom, J.; Gulbrandsen, E. J. Dispersion Sci. Technol. 2005, 26 (2), 207–215. (16) Gutig, C.; Grady, B. P.; Striolo, A. Langmuir 2008, 24(9), 4806–4816. (17) Hodges, C.; Biggs, S. Adv. Powder Technol. 2007, 18(6), 615–629. (18) Atkin, R.; Craig, V. S. J.; Biggs, S. Langmuir 2000, 16(24), 9374–9380. (19) Fleming, B. D.; Biggs, S.; Wanless, E. J. J. Phys. Chem. B 2001, 105(39), 9537–9540. (20) Theodoly, O.; Casc~ao-Pereira, L.; Bergeron, V.; Radke, C. J. Langmuir 2005, 21(22), 10127–10139. (21) Velegol, S. B.; Fleming, B. D.; Biggs, S.; Wanless, E. J.; Tilton, R. D. Langmuir 2000, 16(6), 2548–2556. (22) Iampietro, D. J.; Brasher, L. L.; Kaler, E. W.; Stradner, A.; Glatter, O. J. Phys. Chem. B 1998, 102(17), 3105–3113. (23) Penfold, J.; Tucker, I.; Staples, E.; Thomas, R. K. Langmuir 2004, 20(17), 7177–7182. (24) McDermott, D. C.; McCarney, J.; Thomas, R. K.; Rennie, A. R. J. Colloid Interface Sci. 1994, 162(2), 304–310. (25) Liu, J.; Min, G.; Ducker, W. A. Langmuir 2001, 17(16), 4895–4903.

Published on Web 09/02/2009

DOI: 10.1021/la901321h

11503

Article

Hodges et al.

depending on the length of time given for the adsorption to take place. Published QCM data on CTAB26 indicate that the transition from stable monomers to spherical micelles occurs over a fairly wide concentration range (0.03 to 0.9 mM), suggesting that the formation of entropically stable spherical micelles at an interface is gradual rather than sudden: something that has been confirmed by AFM studies.27 The rinse off results from QCM experiments with high concentration CTAC 17 show an almost immediate rinse off of the surfactant from a silica surface demonstrating that the solubility of CTAC in water is very high, so that the energy required to desorb a CTAC molecule from the interface to the bulk is small. This richness in surface behavior is observed while only considering a relatively small range of counterion, primarily the hard anions of Br- or Cl-. Changing this counterion to a larger organic ion, a hydrotrope, impacts both the bulk behavior of the self-assembly as well as the adsorption behavior. In the bulk, the addition of hydrotopic counterions lead to the formation of elongated, thread-like or worm-like, micelles.28,29 Considerable work has been done on the counterions tosylate,30,31 salicylate and, relevant to this work, the polymerizeable vinyl benzoate.4,32 These surfactant-counterion pairs are essentially complex organic salts with strong binding between the surfactant and counterion; the tail length of the surfactant can be varied to alter the hydrophobicity while subtle changes in the counterion structure impact the binding. Here, we focus on a system prepared from a one-to-one molar ratio of alkane trimethylammonium (CnTAþ, where n denotes the number of carbons in the tail) and vinyl benzoate (VB-). These CnTVB surfactants do not require the addition of an electrolyte to form stable aggregate structures at an interface.3 We expect the dynamics near to the interface to be slower because of the larger counterion, allowing for more possible changes in conformation to occur.

Materials and Methods All water for these experiments came from a Milli-Q Elix RiOs Synthesis A10 system (resistivity 18.2 MΩ cm). The silica substrates used for the quartz crystal microbalance (QCM) work were cleaned as described previously.9 The QCM used was the D300 from Scientific and Medical Instruments Ltd. (Gatley, Cheshire, U.K.) that has a volume of 80 μL and is capable of measuring both the frequency of the silicon coated quartz crystal as well as the dissipation. The QCM frequency shifts observed in this paper were converted into adsorbed Sauerbrey masses using the Sauerbrey relation33 Δm = -(C/n)Δf, where Δm is the mass change from adsorption, Δf is the frequency change measured, n is the overtone number, and C is the crystal sensitivity given by (Fqvq)/(2f02) and is equal to 18 ng cm-2 Hz-1. In each QCM experiment, a few milliliters of sample was allowed to flow into the chamber before the flow was stopped. The adsorption of the sample to the substrate was then monitored for several hours. To observe desorption, 4 aliquots of approximately 3 mL of water were injected into the QCM cell. The optical reflectometer (OR) was purchased from the Laboratory of Physical Chemistry and Colloid Science at the Wageningen University, Germany. The sample cell of the reflectometer contained a volume of approximately 5 mL of liquid. The (26) Kawasaki, H.; Nishimura, K.; Arakawa, R. J. Phys. Chem. C 2007, 111(6), 2683–2690. (27) Liu, J.; Ducker, W. A. J. Phys. Chem. B 1999, 103(40), 8558–8567. (28) Candau, S. J.; Oda, R. Colloids Surf., A 2001, 183-185, 5–14. (29) Carver, M.; Smith, T. L.; Gee, J. C.; Delichere, A.; Caponetti, E.; Magid, L. J. Langmuir 1996, 12(3), 691–698. (30) Soltero, J. F. A.; Puig, J. E.; Manero, O. Langmuir 1996, 12(11), 2654–2662. (31) Truong, M. T.; Walker, L. M. Langmuir 2000, 16(21), 7991–7998. (32) Kline, S. J. Appl. Crystallogr. 2000, 33(3 pt1), 618–622. (33) Sauerbrey, G. Z. Phys 1959, 155(2), 206–222.

11504 DOI: 10.1021/la901321h

instrument may be operated in either a continuous flow condition where a fixed cone of liquid is aimed perpendicularly at the silicon wafer, or the liquid may be allowed to fill the cell, and then the flow switched off to enable the injected liquid to diffuse uniformly throughout the cell. For both cases the adsorption at the interface will be diffusion limited (the continuous flow case is called “stagnant point flow”). In our experiments the volume of polymerizable surfactant available for injection was limited, so we chose to inject continuously at 5 mL/min for 1 min after which time the reflectometer cell volume would be completely exchanged by the injected liquid. After this time, the flow was switched off and steady state diffusion was allowed to take place for a couple of hours or so, before water was then allowed to flow into the cell continuously for 5 min (also at 5 mL/min) to ensure thorough rinsing. A polarized red helium-neon laser at 632 nm strikes the sample surface near to the Brewster angle and the reflected light is monitored using a pair of photodetectors mounted at right angles to each other. If adsorption occurs at the sample-liquid interface, the reflected light will undergo a change in polarization that can be followed in real time during an experiment. Consequently, the OR is a useful instrument to measure the initial stages of adsorption as well as the slower conformational rearrangements at an interface. The silicon wafers for the optical reflectometry work (Silicon Valley Microelectronics Inc., U.S.A.) had been cut to show the Æ100æ crystal face and had a 115 nm thermally deposited oxide layer on the surface. A piece of silicon wafer approximately 1 cm in width and 3 cm in length was cut from this wafer for experiments. These pieces were then treated by a UV/ozone cleaner for 15 min and thoroughly rinsed with water before being inserted into the optical reflectometer. The chamber of the optical reflectometer (OR) was cleaned by repeatedly filling the cell with Decon 90 solution (Sigma-Aldrich) and then thoroughly rinsing the cell with water at least 4 times. The voltage changes recorded from the OR were converted into adsorbed masses using Γ = (ΔS/S)As where ΔS is the recorded voltage change in the initial set voltage, S, and S = Ip/Is, that is, the ratio of the parallel to the perpendicular polarized components of the reflected laser light. As is a sensitivity coefficient calculated from a four-layer optical model, and for our instrument this was found to be 38.8 mg/m2. The surfactants used in this study were prepared as previously described.1-4 The surfactant/counterion pair, denoted CnTVB, used for this work is synthesized from commercially available surfactants. Cetyltrimethylammonium bromide (C16TAB) was purchased from BDH Limited (Poole, U.K.). Octadecyltrimethyl ammonium chloride (C18TAC, Arquad 18-50) was donated by Akzo-Nobel (Chicago, IL). These materials were used as received. Before use each solution was lyophilized to remove solvent. Solutions of CnTAY (where Y = Br or Cl) were prepared for each surfactant. The first step is to replace the hard counterion (Br- or Cl-) of the surfactant with a hydroxide counterion (OH-) using DOWEX ion-exchange resin (99% monosphere 550 A˚ anion exchange resin) obtained from Sigma-Aldrich (St. Louis, MO). In the second step, CnTAOH is neutralized by 4-vinylbenzoic monomer (97%) from Sigma Alrich (St. Louis, MO), in which the OH- counterions are replaced by VB- counterions. CnTVB precipitates in a chilled water bath with high conversion, and the purity of the product has been verified by NMR.34 The result is a purified complex salt of the cationic surfactant (CnTAþ) and anionic counterion (VB-) which is stored as a dried powder until use.

Results and Discussion The OR data in Figure 1 demonstrate a possible three stage isotherm for C18TVB, but only a two stage isotherm for C16TVB. Every point in isotherms shown in Figure 1 was recorded only after a stable value in either the OR or the QCM was obtained. (34) Gerber, M. J., PhD Thesis, Carnegie Mellon University, 2006.

Langmuir 2009, 25(19), 11503–11508

Hodges et al.

Article

Figure 1. Isotherms for (a) C16TVB and (b) C18TVB obtained by QCM (red circles) and OR (black squares).

The stable region at very low C18TVB concentrations implies that once the longer C18TAþ ions have adsorbed there is a small energy barrier to be overcome for any further adsorption to occur. This makes sense in terms of our earlier published model of CnTVB adsorption,3 in that the CnTAþ ions must first charge reverse the silica surface for the vinyl benzoate ions (VB-) to adsorb. The charge density of C18TAþ over the silica surface is likely to be lower than for C16TAþ, and so more material is required before the surface has been fully charge reversed. Once the silica surface is sufficiently coated with C18TAþ the adsorption of the VB- ions appears to occur with smaller increases in bulk concentration than the C16TVB case. It would seem that the silica surface is sufficiently charge reversed for some VB- ions to adsorb even at very low concentrations of C16TVB, implying that a smaller entropy change is involved when C16TAþ adsorbs onto silica compared to C18TAþ. As a consequence, the OR isotherm for C16TVB is more gradual. Both systems reach a plateau of about 2 mg m-2 at higher bulk concentrations. The QCM data for C16TVB agree with the two stage isotherm observed by OR, although the adsorbed amounts are higher, probably because of absorbed water within the adsorbed layer. Absorbed water will result in QCM data that overestimates the adsorbed mass of the C16TVB present at the silica surface, since the absorbed water will also oscillate with the adsorbed layer of C16TVB, so the mass detected by the QCM will be from both the water and the C16TVB. The QCM data for C18TVB appear only to show a two stage isotherm, in contrast to the three stage OR isotherm just discussed. This may indicate that even at very low concentrations, the C18TAþ may couple more strongly with the surrounding water, thus showing a gradual increase in apparent adsorbed mass on the QCM as the added amount of C18TAþ is gradually increased. Both QCM data sets reach a plateau in adsorbed amount at a value of approximately 3 mg m-2. The difference between the OR data and the QCM data over this plateau region is most likely a result of included water within the adsorbed layer since the concentrations used in these experiments were not sufficient to affect the solution bulk properties.17 Since these surfactants are based on the CTAþ ion, it is expected that the adsorption behavior should be similar to that seen for more common surfactants such as cetyltrimethylammonium chloride (CTAC) or cetyltrimethylammonium bromide (CTAB), which have been reported at length in the literature as discussed above. Data by Atkin et al.18 showed that overall the adsorption process at concentrations below the cmc was described by a simple two-step isotherm, which they explain in terms of the “sticking ratio”, which is proportional to the concentration of surfactant present. However, at just one CTAB concentration (0.6 mM), Atkin et al. reported that after the initial rapid adsorption, they observed a slow increase in the amount adsorbed Langmuir 2009, 25(19), 11503–11508

over more than 3 h. Contamination was excluded as being the reason for this slow increase. Instead it was linked with the transition from the first to the second step on the CTAB isotherm. The differences expected between the adsorption rate for CTAB and C16TVB was previously described3 as being largely due to the significant difference in the ion exchange due mainly to the difference in counterion size between the Br- and the vinyl benzoate ions (VB-). We would expect the VB- ions to only adsorb when the silica surface has first been charge reversed by the CTAþ, after which the rate of counterion adsorption will be diffusion limited. However, optical reflectivity was not used to investigate the kinetics of C16TVB in these earlier studies. Figure 2 illustrates the kinetics as observed by the OR over both short (Figure 2a) and longer (Figure 2b) time scales for C16TVB. Solutions are introduced into the OR cell at t = 10 min, and the adsorption was measured for several hours. The fast kinetics, over the first few minutes, is expected to be dominated by the adsorption of the CnTAþ ions according to our published data.3 Figure 2(a) shows that as higher concentrations of C16TVB are introduced into the cell the surface excess increases at a faster rate until a plateau is reached that is concentration dependent. The data at the lowest concentration (0.002 mg/mL) are barely detectable, indicating that at this concentration the density of molecules at the interface are just sufficient to locally alter the polarization of the reflected laser light from the OR. Because of the low level of adsorption, very little charge reversal will occur at the interface thus preventing any further adsorption of VB- ions. At 0.005 mg/mL the OR signal rises to a plateau within about 2 min, indicating that sufficient C16TAþ ions are adsorbed to significantly alter the polarization of the reflected laser light. By examining the raw data files it was observed that the vertical laser polarization (Is) was altered by the C16TAþ more than the parallel laser polarization (Ip). This is expected since the adsorption of relatively short and stiff C16TAþ ions at the silica surface should orient themselves near to the vertical away from the interface, and hence have little affect the lateral polarization of the reflected light. It should be pointed out that the flow rate in each of these experiments was the same within experimental error at 5 mL/min, so the differences in the rate of the increase in the OR signal must relate to the adsorption behavior of the CTAþ ions. Atkin et al.18 showed that at each concentration of CTAB that was used a common rate of adsorption was observed, although the time taken to achieve each plateau was different. The nearest case we observed to the long adsorption that Atkin et al. observed for CTAB was at 0.005 mg/mL of C16TVB (Figure 2), where there appears to be a gradual rise to a plateau over a almost 100 min. It seems that in this case (no micellization; cmc ∼ 0.3 mg/mL for C16TVB 4) the gradual rise in the OR signal must be from the limited diffusion of the VB- ions toward the charged reversed DOI: 10.1021/la901321h

11505

Article

Hodges et al.

Figure 2. Kinetics of C16TVB by OR (a) fast kinetics and (b) slow kinetics after injection of sample at ∼ t = 10 min. Data presented in (a) and (b) are at 0.002 mg/mL (black squares), 0.005 mg/mL (red circles), 0.01 mg/mL (green up-triangles), 0.025 mg/mL (blue down-triangles), and 0.083 mg/mL (dark yellow diamonds).

Figure 3. Optical reflectivity data for the adsorption of C18TVB onto silica (a) fast kinetics and (b) slow kinetics. Concentrations in (a) and (b) are 0.002 mg/mL (black squares), 0.01 mg/mL (red circles), 0.0167 mg/mL (green up-triangles), 0.02 mg/mL (blue down-triangles), and 0.05 mg/mL (dark yellow diamonds).

silica when only a small concentration gradient exists. At all the higher concentrations examined, this slow adsorption process was not observed. We suggest that at higher concentrations there are sufficient VB- ions near to the interface to fill all the available adsorption sites almost immediately. It is likely that between 0.002 mg/mL and 0.005 mg/mL C16TVB some slow adsorption would also occur. The OR adsorption data for C18TVB are shown in Figure 3. The fast kinetics (Figure 3a) for C18TVB look similar to those observed for C16TVB, that is, there exists a concentrationdependent rise to a plateau over the first minute after injection of the surfactant. Figure 3(b) shows the slower kinetics for C18TVB, and immediately it is apparent that the adsorbed layer does not behave in a simple manner with time at each concentration. At the two lowest concentrations shown in the figure the adsorbed layer is essentially stable after a few minutes. However, when the concentration is increased to near 0.02 mg/mL the stabilization period is much slower, perhaps suggesting that there exists a significant concentration gradient near to the interface. Only when this concentration gradient equilibrates is the adsorbed layer stable. At the highest concentration, the concentration gradient has been reduced near to the interface because of the high availability of surfactant in the vicinity. The initial kinetics of adsorption are similar in the two systems, so we plot how quickly the adsorption is over the injection period for both variants in Figure 4. Over most of the concentration range for both variants of CnTVB, the adsorption rate increases steadily, with the suggestion of a maximum rate for C18TVB. The C18TVB data are consistently above the C16TVB data. This is consistent with the picture 11506 DOI: 10.1021/la901321h

Figure 4. Variation of the initial rate of adsorption of CnTVB into the optical reflectometer cell versus the bulk concentration of CnTVB in solution.

of the greater hydrophobicity of the C18TVB leading to a greater driving force to the surface. This increased repellency of the longer carbon chain for the solvent would encourage a more rapid adsorption with the silica surface. This also verifies that micellization is not occurring in the bulk and causing changes in adsorption behavior;the C18TVB system is considerably more viscous and would hinder diffusion to the interface if wormlike micelles had formed. In Figure 5, the kinetics measured in OR and QCM are compared for the C16TVB system. In all cases, the OR data represents one injection of sample, followed by equilibration and then rinse off with clean water. For the QCM experiments, multiple rinse off steps were employed which is observed as a Langmuir 2009, 25(19), 11503–11508

Hodges et al.

Article

Figure 5. OR kinetics data (solid black line) for C16TVB compared with QCM kinetics data (red dashed line) at (a) 0.002 mg/mL, (b) 0.05 mg/ mL, (c) 0.1 mg/mL, and (d) 0.5 mg/mL. The spikes visible on the QCM data are from the injection of water into the QCM cell. Each QCM experiment shows four water rinses toward the end.

sequence of spikes toward the end of the experiment. There is good agreement between the general trends observed for both the OR and the QCM data. At the lowest concentration shown (0.002 mg/mL) there is not enough C16TAþ present to charge reverse the silica surface, and neither the QCM nor the OR show any adsorption. At 0.05 mg/mL both the QCM and the OR demonstrate a rapid adsorption up to a stable value, but the degree of rinse off differs. The OR rinse off data show that all the adsorbed surfactant is removed, whereas the QCM rinse off data appear to show that some surfactant remains behind at the silica surface. Each concentration thereafter behaves in a similar way, and rinses back to the same value (ca. 0.5 mg m-2). The fact that the OR data rinse back to a zero adsorbed amount suggests that the QCM is probably measuring either included water or changes in the local coupling of the bulk liquid with the interface: something to which the OR is not sensitive. If the difference between the OR data and the QCM at equilibrium is considered to be as a result of included water within the adsorbed layer, then we may estimate that up to 40% of the layer is water once full surface coverage is achieved. The OR data shown in Figure 6 for C18TVB are mainly similar to the C16TVB OR data shown in Figure 5. A rapid adsorption is followed by a relatively stable plateau, except at intermediate concentrations where a slow equilibrium with the bulk concentration seems to occur. Again, the estimated water content within the C18TVB layer is about 40%. Note that the QCM appears not to see this slow equilibration, although the QCM data do sometimes show a small relaxation for a few minutes immediately after injection. This is most obvious in Figure 6(b) where both the QCM data and the OR data have an almost identical shape. Again, as for C16TVB the rinse off data observed from the OR Langmuir 2009, 25(19), 11503–11508

shows complete removal of the surfactant from the silica, whereas the QCM data for C16TVB only rinse back to 0.5 mg/m2. It is interesting that in Figures 6(a) and 6(b) the QCM data for C18TVB do not seem to rinse back at all, compared with the significant change observed in the OR data. One interpretation of this result is that as water is passed through the QCM cell, the degree of coupling between the silica surface and the bulk changes, as was mentioned for the C16TVB data in Figure 5. At higher concentrations the QCM show more significant rinse off behavior, although the data never rinse back to zero. The OR data imply that C18TVB is easily removed from the silica surface across all concentrations, which is what was seen for C16TVB in Figure 5. However this contrasts with the behavior of the polymerized versions of these polymers15 where only partial removal of the aggregates from the silica surface was observed. Atkin et al.18 observed a fairly quick rinse off over just a few minutes for CTAB, which agrees with the data shown here, although published AFM data on the rinse off behavior of CTAB demonstrated a two stage removal process.25 It was shown from this AFM work that the duration of the adsorption onto a particular surface affects how easily the surfactant could be rinsed off. We suggest that the degree of surfactant removal in our CnTVB system is dependent on both the flow conditions of the experiment as well as the length of time the surfactant is allowed to adsorb onto the interface. Previously our group17 has shown QCM rinse off data demonstrating complete removal of CTAC from silica even when the concentration of the surfactant is high. This implies that something different happens in the QCM with CnTVB that is not present in the CTAB or CTAC systems. Perhaps the smaller counterion allows for better coupling between the adsorbed layer DOI: 10.1021/la901321h

11507

Article

Hodges et al.

Figure 6. OR kinetics data (black solid lines) and QCM kinetics data (red dashed lines) for C18TVB at (a) 0.005 mg/mL, (b) 0.01 mg/mL, (c) 0.02 mg/mL, and (d) 0.1 mg/mL. The spikes visible toward the end of the QCM experiments are from the water rinsing.

and the bulk compared to the poorer coupling with the larger VBcounterion. Thus we conclude this section by stating that it is always useful to run OR experiments alongside the QCM experiments to examine both the amount of absorbed water within the adsorbed layer and the ability of the layer to be removed by rinsing. This is in principle because the QCM is intrinsically a more complex signal than the OR reflection, since the QCM is sensitive to both the interfacial behavior of the adsorbent, as well as the fluid bulk properties.

Conclusions This paper has examined the adsorption behavior onto silica of the unpolymerized surfactants C16TVB and C18TVB with both the optical reflectivity and the quartz crystal microbalance. In general the two techniques agree with each other in terms of the overall adsorption behavior of each surfactant. However, the OR appears to be more sensitive to differences in the way that each surfactant adsorbs, demonstrating a three stage adsorption isotherm for C18TVB and a two stage isotherm for C16TVB. The difference between the two isotherms was described as being due to the increased degree of hydrophobicity of the C18TAþ compared with the slightly shorter C16TAþ. The short-term kinetics data obtained from the OR show a concentration dependent rate

11508 DOI: 10.1021/la901321h

of adsorption onto silica, with a faster rate observed for C18TVB than for C16TVB, in agreement with the greater degree of hydrophobicity for C18TAþ. The longer time kinetics obtained by OR was very simple for C16TVB, just demonstrating a stable amount adsorbed followed by a rapid and total removal of the surfactant after the addition of water to the system. There was some relaxation behavior observed with C18TVB in the OR data at intermediate concentrations, although the rinse off behavior was similar to that observed for C16TVB. However, although the slow adsorption kinetics QCM data for each surfactant in general agreed with the OR data, the rinse off behavior was quite different from that seen in the OR. Very little change was observed in the QCM signal on the addition of water at the low and intermediate concentrations. Even at the higher concentrations, the QCM data did not rinse back to zero. It was suggested that the reason for this difference between the OR rinse off and the rinse off data observed with the QCM was from the effects of interfacial coupling with the bulk that the QCM is sensitive to but the OR does not detect. Acknowledgment. L.M.W. acknowledges partial support from NSF CBET 0608864 during the duration of this project. Surfactant samples were prepared and bulk properties characterized by Dr. Michael J. Gerber and Dr. Danny M. Kuntz.

Langmuir 2009, 25(19), 11503–11508