pubs.acs.org/Langmuir © 2010 American Chemical Society
Adsorption Kinetics of Laponite and Ludox Silica Nanoparticles onto a Deposited Poly(diallyldimethylammonium chloride) Layer Measured by a Quartz Crystal Microbalance and Optical Reflectometry Dan Xu,*,† Chris Hodges,† Yulong Ding,† Simon Biggs,† Anju Brooker,‡ and David York‡ †
School of Process, Environmental and Materials Engineering, Institute of Particle Science and Engineering, University of Leeds, Leeds, U.K. LS2 9JT, and ‡Procter & Gamble Technical Centre, Longbenton, Newcastle upon Tyne, U.K. NE12 9BZ Received August 2, 2010. Revised Manuscript Received October 8, 2010
A quartz crystal microbalance with dissipation (QCM-D) and an optical reflectometer (OR) have been used to investigate the adsorption behavior of Laponite and Ludox silica nanoparticles at the solid-liquid interface. The adsorption of both Laponite and Ludox silica onto poly(diallyldimethylammonium chloride) (PDADMAC)-coated surfaces over the first few seconds were studied by OR. Both types of nanoparticles adsorbed rapidly and obtained a stable adsorbed amount after only a few minutes. The rate of adsorption for both nanoparticle types was concentration dependent. The maximum adsorption rate of Ludox nanoparticles was found to be approximately five times faster than that for Laponite nanoparticles. The QCM data for the Laponite remained stable after the initial adsorption period at each concentration tested. The observed plateau values for the frequency shifts increased with increasing Laponite particle concentration. The QCM data for the Ludox nanoparticles had a more complex long-time behavior. In particular, the dissipation data at 3 ppm and 10 ppm Ludox increased slowly with time, never obtaining a stable value within the duration of the experiment. We postulate here that this is caused by slow structural rearrangements of the particles and the PDADMAC within the surface adsorbed layer. Furthermore, the QCM dissipation values were significantly smaller for Laponite when compared with those for Ludox for all nanoparticle concentrations, suggesting that the Laponite adsorbed layer is more compact and more rigidly bound than the Ludox adsorbed layer.
Introduction Hard coatings of ceramics are often used to provide good scratch or abrasive wear resistance, although the deposition of the ceramic onto a particular surface may be difficult and the coatings formed are often very brittle.1 One possible alternative route for abrasive resistant coatings is to use an adsorbed layer of colloidal particles, for example clays, to coat the surface instead. The formation of a colloidal particle coating is expected to be simpler than the deposition of a typical ceramic film, since the colloid need not be deposited at high temperatures, and the film thickness can potentially be better controlled than with a typical coating. Very fine colloids, or nanoparticles, are a very good choice for this because of their hardness, small size, and relatively large surface area. In addition, they may be used to reinforce other softer polymeric coatings. The ability to characterize these types of films for better control of the surfaces produced is limited to just a few techniques that can detect surface structural changes and local nanoparticle arrangements. These techniques include the quartz crystal microbalance (QCM), the optical reflectometer (OR), and atomic force microscopy (AFM). The first stage of examining these films is a general surface-averaged investigation by both QCM and OR to be able to compare both the optical properties of these films and the general mass adsorbed onto such surfaces. *To whom correspondence should be addressed. E-mail:
[email protected].
(1) Kingery, W. D., Introduction to Ceramics. John Wiley & Sons, Inc.: New York, 1976. (2) H€oo€k, F.; Voros, J.; Rodahl, M.; Kurrat, R.; Boni, P.; Ramsden, J. J.; Textor, M.; Spencer, N. D.; Tengvall, P.; Gold, J.; Kasemo, B. Colloids Surf., B 2002, 24, 155. (3) Zhou, A.; Muthuswamy, J. Sens. Actuators, B 2004, B101, 8. (4) Sakai, K.; Smith, E. G.; Webber, G.; Schatz, C.; Wanless, E. J.; B€ut€un, V.; Armes, S. P.; Biggs, S. J. Phys. Chem. B 2006, 110, 14744.
Langmuir 2010, 26(23), 18105–18112
Although QCM has been widely used to measure the adsorption kinetics of macromolecules such as proteins2,3 and polymer micelles,4,5 only a very limited amount of research has been reported for the deposition kinetics of nanoparticles with this technique.6-8 Chen and Elimelech6 studied the effect of both monovalent (NaCl) and divalent (CaCl2) electrolytes on the rate of fullerene (C60) deposition onto silica surfaces using the QCM. Their data showed that the rate of deposition increased with increasing salt concentrations. However, the deposition rate sharply reduced at/or above the critical coagulation concentration (CCC) of C60 because of the presence of large aggregates. Furthermore, Quevedo and Tufenkji7 used a QCM to investigate the effect of salt on carboxyl-terminated CdTe quantum dot deposition onto silica surfaces. Fatisson et al.8 showed that adsorption of TiO2 nanoparticles on silica surfaces took place at pH 3, and the desorption of TiO2 particles was observed at pH 9 due to the electrostatic repulsion dominating the forces between (5) Sakai, K.; Webber, G.; Vo, C. D.; Wanless, E. J.; Vamvakaki, M.; B€ut€un, V.; Armes, S. P.; Biggs, S. Langmuir 2008, 24, 116. (6) Chen, K. L.; Elimelech, M. Langmuir 2006, 22, 10994. (7) Quevedo, I. R.; Tufenkji, N. Environ. Sci. Technol. 2009, 43, 3176. (8) Fatisson, J.; Domingos, R. F.; Wilkinson, K. J.; Tufenkji, N. Langmuir 2009, 25, 6060. (9) Velegol, S. B.; Fleming, B. D.; Biggs, S.; Wanless, E. J.; Tilton, R. D. Langmuir 2000, 16, 2548. (10) Theodoly, O.; Casc~ao-Pereira, L.; Bergeron, V.; Radke, C. J.; Tilton, R. D. Langmuir 2001, 16, 2548. (11) Fleming, B. D.; Biggs, S.; Wanless, E. J. J. Phys. Chem. B 2001, 105, 9537. (12) Atkin, R.; Craig, V. S. J.; Biggs, S. Langmuir 2000, 16, 9374. (13) Hodges, C.; Biggs, S.; Walker, L. Langmuir 2009, 25, 4484. (14) Hodges, C.; Biggs, S.; Walker, L. Langmuir 2009, 25, 11503. (15) Dijt, J. C.; Cohen Stuart, M. A.; Hofman, J. E.; Fleer, G. J. Colloids Surf. 1990, 51, 141. (16) Dijt, J. C.; Cohen Stuart, M. A.; Fleer, G. J. Adv. Colloid Interface Sci. 1994, 50, 79.
Published on Web 11/12/2010
DOI: 10.1021/la103071c
18105
Article
Xu et al.
the particles and the silica surface. OR has also been applied to study the adsorption kinetics of surfactant,9-14 polymers,15,16 and nanoparticles.17-19 The study of nanoparticle adsorption at solidliquid interfaces by OR is very limited, and only a few researchers have investigated similar systems.17-19 Sennerfors and Tiberg20 used ellipsometry to study the interfacial properties of a polymer-particle system at a solid interface. They found that the adsorbed cationic polymer has a “flatter” conformation on the silica surface in a nanoparticle free system; however, the thickness of the adsorbed polymer layer was seen to increase in the presence of nanoparticles, and these changes were concentration dependent. They suggested that the change of polymer thickness is due to the competition between nanoparticles and the surface for polymer charges. Kelimann et al.19 (2006) studied the effect of substrates (bare silica and cellulose coated surfaces) and ionic strength on the adsorption of latex nanoparticles by OR. They found that the nature of the substrate is very important in determining the properties of the saturated layer of deposited positively charged nanosized latex particles, since the maximum coverage of latex particles was substantially higher for silica than for cellulose. Furthermore, the maximum coverage of particles was also shown to increase strongly with ionic strength of the solvent. B€ohmer et al.17,18 precoated a silica surface with poly(vinyl imidazole) (PVI) and used this surface to study the adsorption of silica particles onto a substrate. They found that as the particle size was increased, a small decrease in the initial adsorption rate occurred for both particle concentrations (10 and 100 ppm) examined. No change to the initial adsorption kinetics was found when salt was added to the solvent, but salt addition caused a significant increase in the measured adsorbed amount. Although both QCM and OR data can be converted to adsorbed mass, the absolute amount of adsorbed material is not necessarily consistent from these two techniques. Hodges, et al.14 discussed the differences between QCM and OR by study of the adsorption of alkane trimethylammonium vinylbenzoate (C16TVB) and octadecyl trimethylammonium vinylbenzoate (C18TVB). They found that the measured adsorbed amounts were higher from QCM than from OR, and they suggested that this is caused by the entrapped water within the adsorbed layers that can be detected by QCM but not by OR. To the best of our knowledge no research has been reported for the effect of particle shape on the kinetics of Laponite/silica adsorption on PDADMAC coated surfaces in liquid by using either QCM or OR. Many factors may influence the adsorption/desorption and adsorption kinetics of soft nanoparticles (e.g., polymer micelles) or solid nanoparticles, including the shape of the surfactant aggregates,21 the concentration of the particulates,14,18 the pH,4,22 and the ionic strength.8 In particular, Macakova et al.21 studied the effect of the different shapes of surfactant aggregates (i.e., bilayer, cylindrical, and spherical) on the amount of trapped water within the adsorbed surfactant layer using QCM-D and OR. They reported that the amount of trapped water, which is sensed by the QCM-D, was found to increase with the nanoroughness of the adsorbed surfactant layer caused by the shape of the surfactant aggregates. The estimated water content within the surfactant layer is up to
Poly(diallyldimethylammonium chloride) with a typical molecular weight of 100-200 kDa and KOH (reagent grade g 90%) was obtained from Sigma-Aldrich Company Ltd., Dorset, UK. The water used in all experiments was of Millipore Milli-Q grade, with a resistivity of 18.2 MΩ 3 cm (at 25 °C), and a total organic carbon concentration (TOC) of less than 5 ppb. Silica QCM sensors were purchased from Scientific & Medical Products Ltd., Gatley, UK. A stock suspension of 1000 ppm Laponite was prepared by suspending the required mass of Laponite RD powder (from Rockwood Additives Ltd., Widnes, Cheshire, UK) in a known volume of Milli-Q water. The preparation process is critical for Laponite dispersions,27 otherwise an aggregated system may result. The powder was added gradually, with continuous and vigorous stirring of the dispersion during the addition process. After all the Laponite had been added, the dispersion was left for several hours on a magnetic stirring plate and followed by 30 min in an ultrasonic bath to allow the particles to fully disperse. Both 10 ppm and 100 ppm Laponite suspensions were prepared by diluting the stock suspension in different amounts of Milli-Q water. The natural pH of Laponite suspensions is around pH 8-9; the pH of the Laponite suspension was then adjusted to pH 10.5. Poly(diallyldimethylammonium chloride) (PDADMAC) (20 mL, 1000 ppm) solution was prepared by suspending 0.1 g of PDADMAC (20 wt %) in 20 mL of degassed Milli-Q water which was adjusted to pH 10.5. AS-40 Ludox silica solutions were purchased from SigmaAldrich. The solutions were first diluted to the desired concentration before being cleaned by adding an ion-exchange resin (AG 501-X8 from BioRad Laboratories, CA). The ion-exchange resin was left in the Ludox solution overnight and then removed
(17) B€ohmer, M. R. J. Colloid Interface Sci. 1998, 197, 251. (18) B€ohmer, M. R.; van der Zeeuw, E. A.; Koper, G. J. M. J. Colloid Interface Sci. 1998, 197, 242. (19) Kleimann, J.; Lecoultre, G.; Papastavrou, G.; Jeanneret, S.; Galletto, P.; Koper, G. J. M.; Borkovec, M. J. Colloid Interface Sci. 2006, 303, 460. (20) Sennerfors, T.; Tiberg, F. J. Colloid Interface Sci. 2001, 238, 129. (21) Macakova, L.; Blomberg, E.; Claesson, P. M. Langmuir 2007, 23, 12436. (22) Sakai, K.; Smith, E. G.; Webber, G.; Baker, M.; Wanless, E. J.; B€ut€un, V.; Armes, S. P.; Biggs, S. J. Colloid Interface Sci. 2007, 314, 381.
(23) Xu, D.; Hodges, C.; Ding, Y.; Biggs, S.; Brooker, A.; David, Y. Langmuir 2010, 26, 8366. (24) Guzman, E.; Ritacco, H.; Rubio, J. E. F.; Ortega, R. Soft Matter 2009, 5, 2130. (25) Chen, J. S.; K€ohler, R.; Gutberlet, T.; M€ohwald, H.; Krastev, R. Soft Matter 2009, 5, 228. (26) Irigoyen, J.; Moya, S. E.; Iturri, J. J.; Llarena, I.; Azzaroni, O.; Donath, E. Langmuir 2009, 25, 3374. (27) Nicolai, T.; Cocard, S. Langmuir 2000, 16, 8189.
18106 DOI: 10.1021/la103071c
approximately 65%. These key findings can be used to consider the situation of our different shapes of solid nanoparticles. In our previous paper,23 Laponite adsorption onto a wide range of surfaces was investigated using QCM-D. Adsorption onto a positively charged surface, prepared by precoating a silica substrate with poly(diallyldimethylammonium chloride) (PDADMAC), was seen to provide reliable and rapid surface coating by the nanoparticles. Many layer-by-layer deposition studies24-26 have employed PDADMAC to create an initially positively charged surface, since PDADMAC is a strong polyelectrolyte and generally adsorbs easily to most negatively charged substrates. Previously, we only reported a relatively narrow range of nanoparticle concentrations,23 and these were studies only over comparatively short adsorption time. Several interesting open questions remain from our preliminary study, including (a) How can the nanoparticle concentration influence the adsorption kinetics at PDADMACliquid interfaces? (b) Can restructuring of interfacial PDADMAC-nanoparticle layers occur over time? (c) Does nanoparticle shape have an effect on the adsorption profiles? To address these questions, this paper reports the results from a study of the interfacial adsorption behavior of two types of nanoparticles onto PDADMAC-coated oxide surfaces. The adsorption kinetics, across a wide range of concentrations, for both Laponite and Ludox silica has been investigated using both QCM-D and OR.
Materials and Methods
Langmuir 2010, 26(23), 18105–18112
Xu et al.
Article
Figure 1. Schematic illustration of five-layer model used for light reflection. The values used for the optical constants of silicon, silica, polymer layer, nanoparticle layer, and water are indicated. by filtration. The Ludox solution pH was adjusted to pH 10.5 usingNaOH, and the solution ionic strength was 1 mM NaCl. The optical reflectometer (Laboratory of Physical Chemistry and Colloid Science, Wageningen University, Netherlands) has a polarized red He-Ne laser operating at 632.8 nm that is incident onto a silicon wafer near to the Brewster angle. The laser is reflected back toward a pair of photodetectors mounted at right angles to each other. The adsorbed material on the silicon wafer will lead to a change in polarization of the reflected laser beam, which is then monitored by the photodetectors. Perpendicularly to the silicon wafer, a flow of liquid enters the cell at approximately 5 mL/min. Once the flow has been established, a stable cone of liquid forms nears the silicon surface, within which there exists a stagnant point where the fluid motion is diffusion limited. The silicon wafers used for optical reflectometry (Silicon Valley Microelectronics, Inc.) have been cut by the manufacturer to show the [100] crystal face and had a 115 nm thermally deposited oxide layer on the surface. A piece of wafer approximately 1 cm in width and 3 cm in length was cut for experiments. These pieces were then treated with 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 was cleaned by repeatedly filling the cell with Decon 90 solution and then thoroughly rinsing the cell with water. This process was carried out four times. The voltage changes recorded from the optical reflectometer were converted into adsorbed amounts using Γ = (ΔS/S)/As, where ΔS is the recorded voltage change from the initial set voltage, S, and S = Ip/Is, where Ip is the parallel component and Is is the perpendicular component of the reflected laser light. As is a sensitivity coefficient calculated from a five-layer optical model.16 A schematic drawing and values for the refractive indices are shown in Figure 1. In the five-layer model, both PDADMAC and nanoparticle layers are treated as homogeneous. The refractive index increment with concentration, dn/dc = 0.058 g-1 cm3 was used for both types of nanoparticles, the calculation details can be found in Bohmer’s work.18 Shi et al.28 investigated layer-by-layer (LBL) films formed by the alternate deposition of negatively charged titanium(IV) bis(ammonium lactato) dihydroxide (TALH) and positively charged PDADMAC. They reported an average thickness of the dried PDADMAC layer to be 0.4 ( 0.1 nm using ellipsometry. On the basis of this value, we have used a value of 1 nm for the thickness of our fully hydrated PDADMAC layer in the optical model. The thickness of a monolayer of Ludox silica was taken to be that of a single Ludox silica sphere, 20 nm. However, a reasonable value for the thickness of the adsorbed Laponite layer is more difficult, due to the wide range of possible angles that the Laponite nanoparticles may deposit themselves with respect to the substrate. More details are discussed in the Results and Discussion section below. The effect of Laponite layer thickness on the value of the sensitivity coefficient, 1/As, was modeled, and is shown in Figure 2. The thickness range of the Laponite layer 0-25 nm was selected to mimic the two extreme circumstances: (1) no adsorption of Laponite nanoparticles and (2) vertical adsorption of each Laponite nanoparticle. Figure 2 shows that 1/As increases to a nearly constant value of 225 mg/m2 (28) Shi, X.; Cassagneau, T.; Caruso, F. Langmuir 2002, 18, 904.
Langmuir 2010, 26(23), 18105–18112
Figure 2. Effect of adsorbed layer thickness on the sensitivity coefficient, 1/As, for the adsorption from water of Laponite nanoparticles on silica. at a Laponite layer thickness as small as 0.2 nm. This suggests that 1/As is almost totally insensitive to the thickness of the adsorbed nanoparticle layer. Using the Q-Sense QCM instrument used in this work (Q-300, Q-sense AB), four harmonics are detected: F1 (fundamental harmonic), F3 (3rd overtone), F5 (5th overtone), and F7 (7th overtone) leading to eight parameters measured simultaneously (four frequencies and four dissipation values). Typically, frequency shifts (Δf ) and dissipation shifts (ΔD) at the third overtone are presented in this paper since this overtone usually has the best signal-to-noise ratio. For small adsorbed masses, changes in mass are related to changes in frequency of the oscillating crystal through the Sauerbrey relationship: Δm = -cΔf/n where Δf is frequency-shift, Δm is the mass added to the crystal, C is the mass sensitivity constant (C = 18 ng 3 cm-2 3 Hz-1 at 5 MHz) and n is the overtone number (n = 1, 3, 5, 7, ...). The Sauerbrey relation is valid for rigid, nonslip, evenly distributed, and sufficiently thin adsorbed layers. However, for soft or viscoelastic films that do not fully couple to the oscillating crystal, the Sauerbrey relationship underestimates the mass and thus another method of analysis developed by Kanazawa and Gordon29 is needed to fully characterize such materials. As well as measuring frequency changes, the QCM is also capable of detecting changes in the amount of energy lost per oscillation, or the dissipation. The dissipation (D) is an important factor in helping us to understand structural (viscoelastic) properties of an adsorbed material. The QCM can simultaneously measure both frequency and dissipation of the quartz crystal. D is defined as D = Elost/(2πEstored). The temperature controller of the QCM is capable of maintaining the temperature of the QCM cell to within (0.01 °C. The temperature of the solution in the chamber was maintained at 25 °C for all experiments. Prior to each experiment, all crystals used in this work were treated in a UV-ozone chamber for 20 min, followed by 10 min in diluted Decon (a commercial cleaner, mainly containing anionic surfactants) solution and 10 min in Milli-Q water using ultrasound. After the cleaning procedure, the frequency of QCM crystal after all the cleaning in air is observed to determine whether the adsorbed material is completely removed. The QCM chamber was then subsequently cleaned with two hot Decon and four hot Milli-Q water injections (∼60 °C). This treatment produced a very stable baseline for our measurements. All the Laponite and Ludox nanoparticle suspensions were made up by weight, and the concentrations are defined in ppm. However, since there is a considerable difference in density between the two samples, and to be able to directly compare the effect of nanoparticle shape on the nanoparticle adsorption, we convert the mass concentration of nanoparticle suspensions into the number (29) Kanazawa, K. K.; Gordon, J. G. Anal. Chim. Acta 1985, 175, 99.
DOI: 10.1021/la103071c
18107
Article
Xu et al.
of nanoparticles per liter as shown in Table 1. A single Ludox silica nanoparticle is a sphere with a diameter of 20 nm; a single Laponite nanoparticle has a disk-like shape with a thickness of 1 nm and a diameter of 25 nm.
Results and Discussion 1. Initial Adsorption of Ludox and Laponite Nanoparticles. In Figure 3, the nanoparticle suspensions were injected at t = 5 min under the same flow conditions for 15 min. After this, Milli-Q water was injected at pH 10.5 to rinse off any excess nanoparticles from the bulk (data not shown). No difference in the adsorbed amounts was observed after rinsing at any particle concentration. Plateau values for the adsorbed amount as a function of nanoparticle concentration (Figure 3) show, for both nanoparticle types, an initial increase in the amount adsorbed with concentration. A maximum in the adsorbed amount is reached at a nanoparticle concentration of about 1015 nanoparticles/L for both types of nanoparticles. This strongly indicates that the maximum coverage of nanoparticles on the PDADMAC-coated surface is nearly insensitive to the shape of the nanoparticles. Figure 3 also demonstrates that the rate of initial adsorption for both types of particles is very rapid at the higher concentrations, with no more than 1 min being required to achieve a stable adsorbed mass. The lowest concentration of nanoparticle used here requires less than 5 min to reach the stable value. The initial adsorption rates of nanoparticles onto PDADMACcoated surfaces have been calculated for both types of nanoparticles from the data shown in Figure 3 and are presented in Figure 4. OR data are recorded from the stagnation point under constant flow rate conditions and so free diffusion conditions may be assumed to apply.
Figure 4 shows that the trends of the initial adsorption rate for both types of nanoparticles are approximately same. The initial adsorption rates of both Laponite and Ludox silica increase gradually with the number of nanoparticles in the bulk suspension until the upper limit is reached at about 1016 nanoparticles/L. However, the maximum adsorption rate of Laponite nanoparticles is about 1.3 times higher than Ludox nanoparticles. This difference may be caused by the nanoparticle shape. When Laponite nanoparticles approach the PDADMAC-coated layer, unlike Ludox, they have the opportunity to arrange themselves in a more favorable orientation as they approach the interface to achieve a faster diffusion rate, for example, by being more vertically oriented to the substrate. Figure 5a shows that the number of the adsorbed Laponite and Ludox nanoparticles increases with the number of nanoparticles in the bulk suspension until a plateau is reached at 1015 nanoparticles/L. The ratio between the maximum adsorbed amount of Ludox and that of Laponite was a factor of 5, correlating nicely with the ratio of the two nanoparticle masses for a monolayer of each particle. Assuming a monolayer of nanoparticles forms on the adsorbed PDADMAC layer, then the number of nanoparticles per unit area can be calculated as shown in Figure 5b. This indicates that there are typically more Laponite nanoparticles per unit area in a complete monolayer than in a complete Ludox monolayer. This suggests that a denser Laponite layer was formed on the PDADMAC than was the case for a Ludox layer. The formation of the denser Laponite layer may be aided by the polydispersity of the
Table 1. The Relationship between the Mass and Number Concentrations of Nanoparticle Suspensions Used in the QCM and OR Experiments within This Paper. The Approximate Number of Particles Relates to the Values Listed in the Figure Legends approximate approximate no. of no. of no. of Ludox no. of Ludox Laponite Laponite nanoparticle nanoparticles nanoparticles nanoparticles nanoparticles concn (ppm) per liter per liter per liter per liter 0.1 1 10 100 1000
1.1 1013 1.1 1014 1.1 1015 1.1 1016 1.1 1017
1013 1014 1015 1016 1017
9.3 1013 9.3 1014 9.3 1015 9.3 1016 9.3 1017
1014 1015 1016 1017 1018
Figure 4. The initial adsorption rate, dΓi/dt, of Laponite and Ludox silica nanoparticles on PDADMAC-coated surfaces versus the original number of nanoparticles per liter in the bulk suspension obtained from OR measurements.
Figure 3. OR data showing the adsorbed amounts of (a) Laponite and (b) Ludox nanoparticles onto a PDADMAC-coated substrate as a function of time. Data were recorded in water at pH 10.5 under stagnant point flow conditions. All the data presented here have been offset to zero after the PDADMAC had previously been adsorbed and rinsed with water. The nanoparticle suspensions were injected at t = 5 min while maintaining a constant pH. 18108 DOI: 10.1021/la103071c
Langmuir 2010, 26(23), 18105–18112
Xu et al.
Article
Figure 5. The number of adsorbed Laponite and Ludox silica nanoparticles on PDADMAC-coated surfaces versus the original number of nanoparticles per liter in the bulk suspension obtained from OR measurements.
Figure 6. QCM data from PDADMAC-coated surfaces in the presence of different concentrations of Laponite nanoparticles in water at pH 10.5: (a) third overtone frequency changes versus time, and (b) third overtone dissipation changes versus time. Each experiment starts with pH 10.5 water only. After 5 min, 5 mL of pH 10.5 Laponite suspension was injected and stabilized for approximately 160 min.
Laponite suspension and the angular adsorption of Laponite nanoparticles at the interface. As mentioned in our previous work,23 the Laponite suspension was found to be composed of a range of small particles between 15 and 90 nm in diameter, but with a clear peak at about 30 nm using a Malvern NanoSeries ZetaSizer. The smaller Laponite nanoparticles are able to adsorb onto the PDADMAC layer in between the larger ones leading to a higher and more efficient coverage of the PDADMAC surface. In addition, when we examine our OR data more closely, we note that it is the perpendicular component of the reflected laser light that changes most when higher concentrations of Laponite nanoparticles are adsorbed. This suggests that at low particle concentrations most Laponite nanoparticles adsorb almost flat onto the PDADMAC surface, until the concentration of particles is increased near to the point where a complete layer is formed. At this point the Laponite nanoparticles are close enough to overlap each other causing a slight tilt to the mean orientation of the nanoparticles over the PDADMAC surface. This tilt is picked up by the OR as an increase in Is, the perpendicular component of the reflected laser intensity. 2. Adsorption of Laponite and Ludox over Longer Times. Since the OR is a continuous flow setup, a large volume of solution is required for measurements over long periods. In our QCM (a D300 from Q-Sense), once the solution has been injected, the cell is closed and the response of the crystal to the solution may be monitored for many hours if required. QCM was used as a complementary technique to OR and allows the experimenter to investigate the entire interfacial region, out into the bulk liquid to a distance dependent on the decay of the shear wave from the laterally Langmuir 2010, 26(23), 18105–18112
vibrating crystal surface (usually about 240 nm in water-like media). Since the QCM is a mass detector and sensitive to everything adsorbed onto the substrate including trapped solvent, whereas the adsorbed amount obtained by OR is often considered as the “dry” adsorbed mass (i.e., excluding bound solvent), the two techniques may be used in tandem. The adsorption profiles for both nanoparticle types obtained by QCM are presented in Figure 6 (Laponite) and Figure 7 (Ludox silica). In these experiments, the QCM was first cleaned as described above to achieve a clean silica surface and a stable background signal. A 1000-ppm PDADMAC solution (pH 10.5) was then used to coat the silica surfaces as described in our previous work.23 Preliminary experiments with PDADMAC showed only that a very stable layer was formed, and that this layer remained stable for at least several hours. For the nanoparticle experiments, the PDADMAC layer was thoroughly rinsed with water and a stable signal was achieved, Laponite suspensions at one of 1014, 1015, 3 1015, 1016, and 1017 nanoparticles/L were then injected into the QCM cell. Figure 6 shows the addition of Laponite at 5 min to the PDADMAC-coated surfaces and then the equilibration period for 160 min. The frequency shifts increase with particle concentration from 1014 to 3 1015 nanoparticles/L. Above 1015 nanoparticles/L, approximately the same frequency shift was observed for all concentrations. This suggests that approximately 1015 Laponite nanoparticles/L may be sufficient to form a complete monolayer over the PDADMAC substrate, in agreement with the OR data shown above. The rate at which the frequency shifts equilibrate after the injection of Laponite is clearly seen to vary DOI: 10.1021/la103071c
18109
Article
Xu et al.
Figure 7. QCM data from PDADMAC-coated surfaces in the presence of different concentrations of Ludox silica nanoparticles in water at pH 10.5: (a) third overtone frequency changes versus time, and (b) third overtone dissipation changes versus time. Each experiment starts with pH 10.5 water only. After 5 min, 5 mL of pH 10.5 Ludox silica suspension was injected and stabilized for approximately 160 min.
with particle concentration. At the lowest concentration (1014 Laponite nanoparticles/L) the adsorption appears to be rapid and settles quickly at a stable frequency. At 1015 Laponite nanoparticles/L, the frequency shift slowly increases with time and no stable frequency value was obtained during the experiment. At all higher particle concentrations the adsorption behavior is more complex, with several concentrations requiring 100 min or more to reach a stable frequency. The general features of the dissipation data as a function of time for Laponite, Figure 6b, support the observations just made from the frequency data. In general all the dissipation values are quite low, suggesting a dense, rigid, adsorbed layer, agreeing with our aforementioned conclusions from the OR data on Laponite layers. Figure 7a shows that the frequency shifts increase with Ludox silica concentration and appear to reach a maximum frequency shift at 3 1015 Ludox nanoparticles/L. As for Laponite, the frequency shift of Ludox nanoparticles is concentration dependent. At 1013 and 1014 Ludox nanoparticles/L, the frequency shifts are small but measurable, and a stable value of frequency was reached very quickly. The frequency shifts for 3 1014 and 1015 Ludox nanoparticles/L both rise continuously and do not achieve a stable value over the period of the experiment. The same size frequency shift is observed for both 3 1015 and 1016 Ludox nanoparticles/L, but the injection of 1016 Ludox nanoparticles/L reaches a stable frequency value almost immediately, whereas the 3 1015 Ludox nanoparticles/L requires a longer time (approximately 50 min) to reach a stable value. To understand why, we need to examine the corresponding dissipation data. The QCM dissipation data depicted in Figure 7b show more interesting behavior than the frequency data (Figure 7a). At low concentrations (1013 and 1014 nanoparticles/L), the dissipation increases with particle concentration to a stable value in about 20-30 min. At 3 1014 and 1015 nanoparticles/L, the dissipation is seen to increase slowly over the course of the entire experiment. However, if we zoom in on these two concentrations to show just the first 25 min of the experiment (not shown), we find that both concentrations appear to rapidly reach a stable dissipation value, in agreement with the OR data. These particular two concentrations were repeated several times due to the interesting nature of the dissipation, and each time a similar curve of approximately the same shape was obtained. These dissipation data suggest that either (or both) local rearrangements of the PDADMAC-Ludox layer are taking place or that adsorption of Ludox can continue to occur more and more slowly as fewer adsorption sites become available. This is possible if the PDADMAC surface roughens during Ludox adsorption, in which case a larger PDADMAC surface area would be exposed for more Ludox to adsorb onto, as well as the 18110 DOI: 10.1021/la103071c
possible reptation of the PDADMAC onto the adsorbed Ludox nanoparticles. At 3 1015 nanoparticles/L, the dissipation behavior is similar to 3 1014 and 1015 nanoparticles/L, but the gradual increase in the dissipation is even slower. Finally at 1016 nanoparticles/L the dissipation appears to be stable over very long times. This would suggest that at high enough Ludox concentrations little rearrangement of the adsorbed layer is possible. Figure 8a shows that the dissipation values for adsorbed Laponite nanoparticles on PDADMAC increase gradually as the Laponite concentration increases, reaching a plateau region at 3 1015 nanoparticles/L. It is also clear from this plot that the values of dissipation for Laponite at each concentration do not change very much with time, since at each time plotted in Figure 8a the data almost completely overlap each other. This suggests that adsorbed Laponite particles are strongly bound onto the PDADMAC-coated surface and form an essentially rigid film (i.e., no further rearrangement takes place after adsorption). It is possible that the local stacking of the Laponite nanoparticles may change slightly as the particle concentration is increased, leading to the slightly higher dissipation values at high particle concentration, or that a general roughening of the PDADMAC-Laponite film occurs over this concentration range. However, we may conclude that once the Laponite nanoparticles have adsorbed at any concentration, they remain in a fixed conformation at the interface. By contrast the dissipation data for Ludox silica nanoparticles on PDADMAC presented in Figure 8b show significant changes both with particle concentration and with experiment time. The largest variation with time occurs at 1015 Ludox nanoparticles/L, and on either side of this concentration the adsorption appears to be more stable. This suggests that a significant degree of surface reorganization occurs at this particle concentration, possibly related to the degree of surface coverage by the Ludox nanoparticles since a complete monolayer of Ludox can be easily calculated to occur for particle concentrations between 1015 and 1016 nanoparticles/L depending on how the particles arrange locally on the PDADMAC. Since the Ludox silica nanoparticles are spheres, the degree of surface roughening from an adsorbed layer of Ludox particles may be expected to be larger than that from an adsorbed layer of Laponite disks, assuming that the disks preferentially adsorb flat face downward onto the PDADMAC. Thus, within these assumptions, we may expect the dissipation values for a complete surface layer of Ludox silica to be significantly larger than those obtained from adsorbed layers of Laponite, which is indeed what we find from Figure 8. Langmuir 2010, 26(23), 18105–18112
Xu et al.
Article
Figure 8. Dissipation changes as a function of nanoparticle concentration at three different times: 20, 100, and 160 min. (a) Laponite data. Note that the data at 100 min (red) are covered by the data at 160 min data (blue); (b) Ludox silica data.
Figure 9. Possible conformational arrangements of Ludox silica nanoparticles adsorbing onto a silica QCM surface that has been modified by the adsorption of PDADMAC.
As seen from the OR data, both types of nanoparticles can rapidly adsorb on the PDADMAC-coated surface and remain firmly adsorbed. However, the QCM data at some intermediate concentrations of Ludox silica show more complex behavior than the OR data, one reason for which may be the rearrangement of both the PDADMAC polyelectrolyte and the Ludox silica nanoparticles. On the basis of our analysis of both the QCM and OR data for Ludox, we present one possible set of conformations for three different particle concentrations as shown in Figure 9. Once the PDADMAC has been adsorbed onto the silica surface, the subsequent adsorption of Ludox at low particle concentrations, Figure 9a, shows a limited number of silica particles randomly adsorbed. This presents a slightly roughened interface to the bulk solution causing a small increase in the QCM dissipation data. At intermediate nanoparticle concentrations, Figure 9b, initially the Ludox nanoparticles adsorb on top of the PDADMAC. However, given enough time the PDADMAC and Ludox nanoparticles can both rearrange so that the surface becomes a mixed Ludox-PDADMAC layer. This mixing process will depend on several factors including the reptation rate of the PDADMAC, but it will most likely be a slow process, possibly accounting for the slow but significant dissipation shifts observed over the few hours of our experiments. Finally, at high Ludox nanoparticle concentrations, Figure 9c, all the available charge sites of the PDADMAC are occupied by the Ludox nanoparticles;minimal space is present between the Ludox particles;so any further rearrangement of the adsorbed material is hindered. Besides, as shown in Figure 8a, no significant change of Laponite dissipation was observed with a range of Laponite concentrations Langmuir 2010, 26(23), 18105–18112
over a few hours. We believe that this is due to the shape effect of these two types of nanoparticles. The adsorbed disk-like Laponite would have larger contact area with the PDADMAC layer than the adsorbed spherical Ludox nanoparticles. Therefore, this conformation of Laponite adsorption would limit the further rearrangement we just discussed for Ludox nanoparticles. Figure 10 directly compares the data from both OR and QCM showing adsorption isotherms of Laponite and Ludox silica on PDADMAC. A similar isotherm shape is obtained by both techniques for both types of nanoparticles. Within experimental uncertainty, the OR adsorption isotherm matches the QCM data at low particle concentrations, but lies below the QCM isotherm at higher particle concentrations for both types of nanoparticles. The differences between the QCM and the OR data are often attributed to included solvent within the adsorbed layer, which the QCM is sensitive to but the OR is not.14 This may be also attributed to other mechanisms which can be measured by QCM, such as the detection of the bulk, the trapped solvent within the adsorbed layer, surface roughening of the adsorbed layer, and the formation of multilayers. As mentioned earlier the QCM shear wave penetration depth into the bulk is approximately 240 nm. However, no significant frequency changes were observed after rinse-off with pH 10.5 water for either Laponite or Ludox (data not shown). This suggests that the frequency shifts were mainly dominated by the adsorbed nanoparticles at the solid-liquid interface and that little or no bulk effects were present over the entire particle concentration range tested. Macakova et al21 reported QCM data for nanoparticles that were either spherical or rod-shaped showing that while the DOI: 10.1021/la103071c
18111
Article
Xu et al.
Figure 10. QCM and OR adsorption isotherms of nanoparticles of (a) Laponite and (b) Ludox adsorbed onto a PDADMAC-coated surface.
QCM-adsorbed mass (i.e., frequency shifts) remained nearly constant, significant differences in the QCM dissipation data were found for each nanoparticle type. Although their nanoparticles were soft-surfactant based systems, the effect of shape on the measured QCM adsorbed mass is significant in the present context. They suggested that the main reason for the observed differences between these adsorbed soft nanoparticles is due to the mechanically trapped water within the adsorbed layer. The spherical surfactant aggregates resulted in the largest amount of trapped water, with the rod-shaped nanoparticles being next, and a simple bilayer contained the least amount of trapped water. In our case, both the OR and the QCM data change significantly for each nanoparticle type (Laponite and Ludox), perhaps due to a larger effect of our more complex substrate (PDADMAC) compared to that of the solid silica substrate used by Macakova et al.21
Conclusions The adsorption kinetics of Laponite/Ludox silica nanoparticles onto PDADMAC over a wide range of particle concentrations has been studied by QCM and OR. These data have confirmed our previous results23 suggesting that the adsorption of Laponite/ Ludox silica is nanoparticle-concentration dependent and the amount increases with the nanoparticle concentration as a typical S-shaped isotherm. The OR data show that the rapid initial adsorption rates for both Laponite and Ludox silica on a preadsorbed PDADMAC surface reach stable values within a few minutes at the lower particle concentrations, and in less than a minute at higher nanoparticle concentrations. The initial adsorption rate of both types of nanoparticles increased with nanoparticle concentration, with the maximum rate of adsorption for Laponite nanoparticles being approximately 1.3 times faster than that for Ludox nanoparticles. This faster adsorption rate may be attributed to the shape of the nanoparticles, where the disk-like Laponite nanoparticles arrange themselves during the adsorption process into a more favorable conformation (e.g., edge on) as the nanoparticle concentration is increased. The Ludox nanoparticles,
18112 DOI: 10.1021/la103071c
being spherical, do not have this option and so the diffusion of the Ludox nanoparticles toward the surface may be hindered at high Ludox particle concentrations. In general over shorter periods of time (the first few minutes of adsorption) the QCM data agree with the OR data, both in terms of the general adsorption trends with particle concentration and in terms of the relative differences observed between the two nanoparticle types. However over longer time periods (hours), the QCM data demonstrated some interesting slowly changing dissipation data for the adsorption of Ludox nanoparticles. These slow changes in dissipation were not observed during the adsorption of Laponite nanoparticles. The reason for these slow dissipation changes was attributed to surface rearrangement of the Ludox-PDADMAC system, possibly involving some reptation of the PDADMAC polyelectrolyte onto and around the surface of the Ludox nanoparticles. Overall, the QCM data also showed that the adsorption of Laponite resulted in smaller mean dissipation values than was found for Ludox silica across all nanoparticle concentrations measured. This suggests that the Laponite nanoparticles are adsorbed into a more compact formation on the PDADMAC surface than the Ludox silica nanoparticles. This would also explain why surface rearrangement would be more difficult with Laponite than with Ludox. Future work studying the local structural rearrangement of these nanoparticle-polyelectrolyte systems at the solid-liquid interface may be investigated using, for example, atomic force microscopy under identical conditions to those used here. This would increase our understanding of the adsorption mechanisms for each particle type. Other experiments examining the effect of polymer chain lengths or salt concentration on Laponite/Ludox adsorption in the cationic polyelectrolyte system would also be worthy of investigation. Acknowledgment. The EPSRC and Proctor and Gamble are thanked for the Research Grant EP/F000464/1 awarded to Yulong Ding and Simon Biggs.
Langmuir 2010, 26(23), 18105–18112
