Langmuir 2006, 22, 8435-8442
8435
Characterizing the pH-Responsive Behavior of Thin Films of Diblock Copolymer Micelles at the Silica/Aqueous Solution Interface Kenichi Sakai,*,† Emelyn G. Smith,‡ Grant B. Webber,†,§ Murray Baker,‡ Erica J. Wanless,‡ Vural Bu¨tu¨n,| Steven P. Armes,⊥ and Simon Biggs† School of Process, EnVironmental and Materials Engineering, UniVersity of Leeds, Leeds LS2 9JT, United Kingdom, School of EnVironmental and Life Sciences, The UniVersity of Newcastle, Callaghan, New South Wales 2308, Australia, Department of Chemistry, Eskisehir Osmangazi UniVersity, Campus of Meselik, Eskisehir 26040, Turkey, and Department of Chemistry, Dainton Building, The UniVersity of Sheffield, Brook Hill, Sheffield S3 7HF, United Kingdom ReceiVed June 14, 2006. In Final Form: July 25, 2006 The pH-responsive behavior of cationic diblock poly(2-(dimethylamino)ethyl methacrylate)-block-poly(2(diethylamino)ethyl methacrylate) copolymer micelles adsorbed at the silica/aqueous solution interface has been characterized. The micellar morphology of this copolymer, initially adsorbed at pH 9, can be dramatically altered by lowering the solution pH. The original micelle-like morphology of the adsorbed copolymer chains at pH 9 completely disappears as the pH is decreased to 4, and a brush-like layer structure is produced. This change results from protonation of the copolymer chains: the subsequent electrostatic repulsions within the film drive the copolymer chains to expand into the aqueous phase. Returning the solution pH from 4 to 9 causes this brush-like layer to collapse, with atomic force microscopy images suggesting degradation of the film. Hence, the pH-responsive behavior of the copolymer film exhibits irreversible morphological changes. Measurements of the adsorbed/desorbed amounts of the copolymer film were conducted using both a quartz crystal microbalance with dissipation monitoring (QCM-D) and optical reflectometry (OR). After an initial rinse at both pH values, the OR adsorbed mass becomes almost constant during subsequent pH cycling, whereas the corresponding QCM-D adsorbed mass changes significantly but reversibly in response to the solution pH. Since the QCM-D measures a bound mass that moves in tandem with the surface, the discrepancy with the OR data is due to changes in the amount of bound water in the copolymer film as a result of the pH-induced changes in surface morphology. The larger effective mass observed at pH 4 suggests that the brush-like layer contains much more entrapped water than the micellar films at pH 9. The pH dependence of the contact angle of the adsorbed film is consistent with the changes observed using the other techniques, regardless of whether the solution pH is altered in situ or the aqueous solution is completely replaced. In fact, comparison of these two approaches provides direct evidence of the exposure of adsorbed micelle core blocks to the solution during pH cycling and the concomitant impact upon all the other measurements.
Introduction Recent advances in synthetic methodology allow the design of a wide range of novel diblock copolymers that undergo spontaneous self-assembly in aqueous solution in response to stimuli such as solution pH1,2 or temperature.3,4 Under appropriate conditions these copolymers form micelles whose physicochemical properties such as critical micelle concentration, size, and ζ potential may be controlled by adjusting the block composition, overall copolymer molecular weight, and functionality of the side chains. Such stimulus-responsive diblock copolymer micelles are being developed for drug delivery applications,5-8 as * To whom correspondence should be addressed. E-mail: k.sakai@ leeds.ac.uk. † University of Leeds. ‡ The University of Newcastle. § Current address: Particulate Fluids Processing Centre, Department of Chemical and Biomolecular Engineering, The University of Melbourne, Parkville, Victoria 3010, Australia. | Eskisehir Osmangazi University. ⊥ The University of Sheffield. (1) Liu, S.; Armes, S. P. Curr. Opin. Colloid Interface Sci. 2001, 6, 249. (2) Sumerlin, B. S.; Lowe, A. B.; Thomas, D. B.; McCormick, C. L. Macromolecules 2003, 36, 5982. (3) Aoshima, S.; Sugihara, S.; Shibayama, M.; Kanaoka, S. Macromol. Symp. 2004, 215, 151. (4) Wei, H.; Zhang, X.-Z.; Zhou, Y.; Cheng, S.-X.; Zhuo, R.-X. Biomaterials 2006, 27, 2028. (5) Ro¨sler, A.; Vandermeulen, G. W. M.; Klok H.-A. AdV. Drug DeliVery ReV. 2001, 53, 95.
nanoreactors for preparing nanosized materials,9 and as building blocks for the fabrication of smart surface coatings.10 To develop a smart surface coating, we have investigated the adsorption of poly(2-(dimethylamino)ethyl methacrylate)-blockpoly(2-(diethylamino)ethyl methacrylate) (PDMA-PDEA) diblock copolymer micelles at solid/aqueous solution interfaces.11-15 PDMA-PDEA shows stimulus-responsive behavior in aqueous solution: it is molecularly dissolved in acidic solution, while in alkaline solution it forms core-shell micelles with the PDEA chains located in the hydrophobic cores and the hydrophilic PDMA chains forming the cationic micelle coronas.16-18 These (6) Rodriguez-Herna´ndez, J.; Lecommandoux, S. J. Am. Chem. Soc. 2005, 127, 2026. (7) Che´cot, F.; Bruˆlet, A.; Oberdisse, J.; Gnanou, Y.; Mondain-Monval, O.; Lecommandoux, S. Langmuir 2005, 21, 4308. (8) Hruby, M.; Koo`a´k, EÅ .; Ulbrich, K. J. Controlled Release 2005, 103, 137. (9) Gohy, J.-F.; Lohmeijer, B. G. G.; Schubert, U. S. Chem.sEur. J. 2003, 9, 3472. (10) Russell, T. P. Science 2002, 297, 964. (11) Webber, G. B.; Wanless, E. J.; Bu¨tu¨n, V.; Armes, S. P.; Biggs, S. Nano Lett. 2002, 2, 1307. (12) Webber, G. B.; Wanless, E. J.; Armes, S. P.; Tang, Y.; Li, Y.; Biggs, S. AdV. Mater. 2004, 16, 1794. (13) Webber, G. B.; Wanless, E. J.; Armes, S. P.; Biggs, S. Faraday Discuss. 2005, 128, 193. (14) Sakai, K.; Smith, E. G.; Webber, G. B.; Schatz, C.; Wanless, E. J.; Bu¨tu¨n, V.; Armes, S. P.; Biggs, S. Langmuir 2006, 22, 5328. (15) Sakai, K.; Webber, G. B.; Smith, E. G.; Schatz, C.; Wanless, E. J.; Bu¨tu¨n, V.; Armes, S. P.; Biggs, S. J. Phys. Chem. B 2006, 110, 14744. (16) Bu¨tu¨n, V.; Billingham, N. C.; Armes, S. P. Chem. Commun. 1997, 671. (17) Lee, A. S.; Gast, A. P.; Bu¨tu¨n, V.; Armes, S. P. Macromolecules 1999, 32, 4302.
10.1021/la061708f CCC: $33.50 © 2006 American Chemical Society Published on Web 09/02/2006
8436 Langmuir, Vol. 22, No. 20, 2006
micelles spontaneously form an adsorbed layer at the mica/ aqueous solution interface due to favorable electrostatic interactions.12,13 Interestingly, atomic force microscopy (AFM) studies and surface fluorescence studies indicate that the adsorbed micelles observed at high pH “open up” at low pH due to protonation of the PDEA chains in the micelle cores to form a localized brush-like layer. Moreover, the original core-shell micelle structure appears to be re-formed on returning to high pH, at least for a mica substrate.12,13 Such pH-responsive behavior at the mica/aqueous solution interface may also be expected to occur on silica. However, in our previous investigation14 the adsorbed micellar morphologies on the two surfaces were found to be somewhat different: the adsorbed micelles form close-packed layers on mica, while more disordered layers are formed on silica. In addition, the estimated mean spacing of the adsorbed micelles on silica is significantly larger than that on mica. Therefore, one might anticipate that the pH-responsive behavior of the adsorbed copolymer micelles is modulated, at least to some extent, by the substrate. To clarify this, it is necessary to observe the morphological change of the diblock copolymer micelles adsorbed on silica in response to the solution pH and to compare the result with that of the mica case. Quantitative understanding of the pH-responsive behavior is also a goal of this investigation. Previous AFM studies suggested that cycling of the solution pH produces a reversible change in the layer morphology of the copolymer micelles adsorbed on mica, with little or no micelle desorption evident from the obtained images.12,13 This is consistent with the well-known observation that the desorption rate of polymers once adsorbed at a solid/ solution interface is usually slower than that of small surfactant molecules due to the multiple attachment points of the polymer chains. Such a slow desorption rate may allow for subsequent processing and advanced applications of these surface coatings where replacement of the solution phase is required. However, direct evidence for the restricted desorption and quantification of the subsequent pH-responsive behavior of the PDMA-PDEA class of copolymers have yet to be reported. Herein we characterize the pH-responsive behavior of the adsorbed PDMA-PDEA film at the silica/aqueous solution interface. The initial copolymer film was prepared at pH 9,15 and the solution pH was repeatedly changed to either 4 or 9. The morphological change of the adsorbed film during the pH cycling was observed using in situ soft-contact AFM. These changes in the adsorbed layer were characterized using contact angle and ζ potential (based on the streaming potential) measurements. Finally, the adsorbed/desorbed amounts of the copolymer in response to the solution pH were assessed using both a quartz crystal microbalance with dissipation monitoring (QCM-D) and optical reflectometry (OR). Experimental Section Materials. The PDMA-PDEA diblock copolymers were synthesized using group transfer polymerization (GTP) as described elsewhere.18 Two equivalent copolymers having slightly different chain lengths were used in this study: PDMA93-PDEA24 (Mn ) 19100, DMA content 79 mol %, PDI ) 1.13) and PDMA96-PDEA26 (Mn ) 21800, DMA content 79 mol %, PDI ) 1.09), where the subscripts refer to the mean degrees of polymerization of each block. These two copolymers are generically denoted by “PDMA9XPDEA2Y”. In our previous investigation,15 the micellar characteristics of these diblock copolymers in aqueous solution were presented. On the basis of dynamic light scattering (DLS) analysis, the micelle hydrodynamic diameter in aqueous 10 mmol‚dm-3 KNO3 solution at pH 9 was observed to be 22-27 nm. In addition, the aggregation (18) Bu¨tu¨n, V.; Armes, S. P.; Billingham, N. C. Polymer 2001, 42, 5993.
Sakai et al. number in 10 mmol‚dm-3 Borax buffer solution at pH 8.5 was estimated to be 42 ( 7, as determined using static light scattering (SLS). Silicon wafers (purchased from Silicon Valley Microelectronics, California, with a predefined oxide layer of 115 nm) were used for AFM observations and contact angle and OR measurements. MenzelGla¨ser glass coverslips for an optical microscope (Germany) were employed for the streaming potential measurements. All other reagents were of analytical grade, and water was Millipore Milli-Q grade. Unless otherwise stated, all solutions were 10 mmol‚dm-3 in KNO3, and the solution pH was adjusted using KOH or HNO3. All copolymer solutions were used within 7 days of preparation, and the pH was adjusted immediately prior to use. Measurements. AFM. In situ imaging of the adsorbed PDMA9XPDEA2Y copolymer films on silica was performed with a Nanoscope III atomic force microscope (Veeco, CA). Cantilevers with an integral silicon nitride tip (NanoProbe, Veeco, CA) were used for all AFM experiments and were cleaned using UV irradiation (approximately 9 mW‚cm-2 at 254 nm) prior to use. The silicon wafers were subjected to a three-step treatment before use. This involved UV irradiation for 30 min, followed by ultrasonication in ethanol for 20 min. The silica was then stored in ethanol, rinsed just prior to use with Milli-Q water, and soaked in 10 wt % aqueous NaOH for 10 min, followed by copious rinsing with Milli-Q water to give a hydroxylated silica surface. A fresh piece of silica was used for each experiment and then discarded. The background electrolyte and copolymer solutions were passed through a syringe-mounted 0.2 µm filter (GHP Acrodisc, Pall Gelman Science, Michigan) as they were injected into the AFM fluid cell. Images were collected using the in situ soft-contact method.19 All images presented are deflection images and have been zero-orderflattened. Interaction forces between the cantilever and the copolymer film formed at the silica/aqueous solution interface were also monitored. At the beginning of each experiment, an electrolyte solution adjusted to pH 9 was injected into the cell and left for approximately 30 min to allow the substrate to reach charge equilibrium. After equilibration for at least 4 h in a copolymer solution (500 ppm, pH 9), the solution was repeatedly replaced by approximately 5 cm3 of electrolyte solution at either pH 9 or pH 4 every 2 h. All measurements were performed at 25 ( 2 °C. ζ Potential Based on Streaming Potential Measurements. An electrokinetic analyzer (Anton Paar, Austria) was used to estimate the ζ potential of the silica substrates coated with the PDMA9XPDEA2Y copolymer. The ζ potential values were calculated using the following equation:20 ζ)
dU η κ dp r0
(1)
where U is the streaming potential, p is the pressure, η is the viscosity of the sample solution, r is the relative dielectric permittivity of the solution, 0 is the vacuum permittivity, and κ is the specific conductivity of the solution. Glass coverslips were ultrasonicated for 15 min prior to use in an anionic/nonionic surfactant mixture (Decon 90, Decon Laboratories Ltd., U.K.), rinsed with Milli-Q water, irradiated with UV ozone for at least 10 min, and finally rinsed thoroughly with Milli-Q water. A commercial rectangular cell equipped with two Ag/AgCl electrodes was used for measuring the streaming potential. The two sheets of the silica plates were sandwiched into the cell using PTFE sealing and PTFE distance foils, and then a pH 9 electrolyte solution was passed into the cell. After equilibration, a copolymer solution (50 ppm, pH 9) was injected. After equilibration overnight, the solution pH was repeatedly changed to either 9 or 4 every hour, and the change in the ζ potential during this pH cycling was monitored. All data presented herein were collected at room temperature (approximately 25 °C). (19) Manne, S.; Cleveland, J. P.; Gaub, H. E.; Stucky, G. D.; Hansma, P. K. Langmuir 1994, 10, 4409. (20) Fairbrother, H.; Mastin, H. J. Chem. Soc. 1924, 75, 2318.
pH-ResponsiVe BehaVior of Copolymer Micelles
Langmuir, Vol. 22, No. 20, 2006 8437
Contact Angle Measurements. A Dataphysics OCA 20 (Germany) instrument was employed to measure contact angles on inverted silica surfaces using the captive bubble technique. Freshly cleaned silica surfaces were placed in contact with filtered (0.2 µm, GHP Acrodisc) aqueous PDMA9X-PDEA2Y copolymer solutions at 500 ppm for 30 min to attain adsorption equilibrium. An air bubble (3 mm diameter) was then positioned on the underside of the surface using a hooked needle (0.5 mm external diameter). Average measurements from five advancing bubbles are reported. Contact angle data were collected during pH cycling experiments conducted with two different protocols: both in the presence and in the absence of copolymer. Initially, a clean surface was placed in contact with the polymer solution at pH 9, and after equilibration a single contact angle measurement was performed. Then the solution was replaced with copolymer-free electrolyte solution at pH 9 and the contact angle remeasured. The solution pH was then altered in situ, before equilibration for a further 30 min with gentle stirring prior to further measurement. At this point the two protocols deviate. In one experiment the solution pH was altered each time via complete replacement of the supernatant within the measurement cell. In the other experiment the pH was altered in situ by small additions of acid or base. QCM-D. A commercial (Q-Sense, Go¨teburg, Sweden) QCM-D was used to measure adsorbed amounts. The features of this instrument are described in greater detail elsewhere.21,22 The adsorbed mass can be calculated by applying the Sauerbrey equation:23 ∆f ) -
f02 f0 ∆m ) ∆m ) -C∆m Fqνq Fqtq
(2)
where ∆f is the change in resonant frequency, f0 is the resonant frequency of the “clean” substrate, Fq and νq are the specific density and shear wave velocity of the quartz, tq is the thickness of the quartz crystal, and ∆m is the mass of material adsorbed. It is noted that the mass measured using the QCM technique is the total mass of material coupled to the oscillation. Consequently, solvent molecules within the adsorbed layer will also contribute to the measured mass, and therefore, adsorbed amounts determined using the QCM-D usually exceed those measured using other analytical techniques such as ellipsometry24,25 or OR.26 Simultaneously, dissipation changes were also monitored to assess the relative rheological properties of the adsorbed layers. A single sensor crystal with a silica coating was used for all QCM-D experiments. The cleaning procedures of the sensor crystal have been previously described in detail.15 A copolymer solution (500 ppm, pH 9) was injected into the QCM-D inner cell after equilibration in electrolyte solution was achieved.15 After the system was left overnight, repeated pH cycling was conducted to probe the responsive behavior of the adsorbed film. In these experiments, 1 cm3 of either pH 4 or pH 9 electrolyte solution was injected every 15 min up to four times per pH change to ensure complete pH stability in the QCM-D inner cell. All measurements were performed at a constant temperature of 25.0 °C. OR. Copolymer adsorption to silica was measured by the OR technique as described by Dijt and co-workers;27 our instrumentation has been described in detail previously.28 The change in the output signal (∆S) was converted to an adsorbed amount (ΓOR) using eq 3, (21) Rodahl, M.; Ho¨o¨k, F.; Krozer, A.; Brzezinski, P.; Kasemo, B. ReV. Sci. Instrum. 1995, 66, 3924. (22) Rodahl, M.; Kasemo, B. ReV. Sci. Instrum. 1996, 67, 3238. (23) Sauerbrey, G. Z. Phys. 1959, 155, 206. (24) Ho¨o¨k, F.; Kasemo, B.; Nylander, T.; Fant, C.; Sott, K.; Elwing, H. Anal. Chem. 2001, 73, 5796. (25) Stålgren, J. J. R.; Eriksson, J.; Boschkova, K. J. Colloid Interface Sci. 2002, 253, 190. (26) Notley, S. M.; Biggs, S.; Craig, V. S. J.; Wågberg, L. Phys. Chem. Chem. Phys. 2004, 6, 2379. (27) Dijt, J. C.; Cohen Stuart, M. A.; Fleer, G. J. AdV. Colloid Interface Sci. 1994, 50, 79.
ΓOR )
∆S 1 S0 As
(3)
where S0 is the initial output value and As is the sensitivity factor, which is calculated from the optical properties of the s and p polarizations of the reflected laser beam. The instrument is housed in an incubator to maintain a constant temperature of 25 ( 2 °C. A typical experiment involves the PDMA9X-PDEA2Y copolymer (500 ppm, pH 9) being introduced from a gravity-fed line through a two-way valve after the surface has equilibrated with pH 9 electrolyte solution. After about 20 min of copolymer flow, the valve is turned and electrolyte solution at pH 9 is flowed through the cell. At this point the reservoir containing copolymer is switched to another containing electrolyte solution, and the line is thoroughly rinsed and then filled again with pH 4 electrolyte solution. At any given time hereafter, the valve can be turned and the pH of the solution changed. In this way a continuous experiment was carried out whereby the solution pH, and thus the adsorbed layer characteristics, could be cycled back and forth.
Results and Discussion Adsorbed Layer Morphology and ζ Potential. In situ softcontact AFM images of adsorbed PDMA9X-PDEA2Y films on silica are shown in Figure 1. Image a was observed after the adsorbed film was rinsed with an electrolyte solution at pH 9. One can clearly see surface micelle structures in this image. The physical characteristics of this initial adsorbed film have already been discussed in our previous papers.14,15 After these surface micelle images were recorded, pH 4 electrolyte solution was injected into the fluid cell. A typical AFM image after this pH change is given in Figure 1b. The adsorbed film morphology is significantly altered from that seen in Figure 1a: the micelle morphology observed at pH 9 completely disappears at pH 4 and is replaced by a much more featureless morphology. Interestingly, when the solution pH was returned to 9, the film morphology apparently became rougher than that originally observed at pH 4, as shown in Figure 1c. These AFM images reveal a dramatic and irreversible change in the adsorbed micelle film morphology on silica in response to pH cycling. An initially similar, but reversible, morphological change in the adsorbed copolymer film during pH cycling was also observed at the mica/aqueous solution interface using the same PDMA9XPDEA2Y copolymer under comparable experimental conditions.13 In this earlier paper we suggested that protonation of the adsorbed copolymer chains, due to the decrease in the solution pH, caused greater interchain electrostatic repulsion within the micellar film, leading to the extension of the copolymer chains into the solution phase. However, this extension may be partially restricted on mica due to the very close packing of the micelles on this highly anionic substrate. This enabled us to report the re-formation of the adsorbed micellar films on returning to the original solution pH: the size, shape, and lateral spacing of the adsorbed micelles observed after pH cycling were very similar to those observed in the original adsorbed micelles.13 While this final micellar structure was clearly observed on mica in our earlier work, on silica the original micellar structure is not apparent as shown in Figure 1c. As reported previously,14 the adsorbed micelles at pH 9 form close-packed layers on mica, while only loosely packed disordered layers are formed on silica, with relaxation of the coronal-forming PDMA chains to the surface being more pronounced in the latter case (see Figure 1a). Hence, there is a lower overall surface micelle density on silica at pH 9. Although more coronal-forming PDMA chains must be attached to the silica surface, chain (28) Atkin, R.; Craig, V. S. J.; Biggs, S. Langmuir 2000, 16, 9374.
8438 Langmuir, Vol. 22, No. 20, 2006
Figure 1. Soft-contact 1 × 1 µm2 AFM images of PDMA9XPDEA2Y layers on silica adsorbed from a 500 ppm, 10 mmol‚dm-3 KNO3 aqueous solution at pH 9. Images were recorded (a) after rinsing with 10 mmol‚dm-3 KNO3 electrolyte solution at pH 9, (b) after rinsing with electrolyte solution at pH 4, and (c) after the solution pH was returned to pH 9.
extension upon protonation at pH 4 may occur more effectively (Figure 1b) than on mica because of the larger micelle surface area exposed to the solution. This collapsed and more highly spread micelle conformation on silica may prevent complete re-formation of the micellar film morphology on returning to pH 9 from pH 4 (Figure 1c), where some desorption is expected. Adsorption and desorption are quantified below using the QCM-D and OR techniques. Throughout each AFM experiment, the interaction force curves between the cantilever tip and the copolymer film formed on silica were also recorded. Typical results are given in Figure 2. After replacement of the copolymer solution by electrolyte solution at pH 9 (see curve a), a purely repulsive interaction was observed commencing at an apparent separation of approximately 18 nm. This repulsion increases almost exponentially with decreasing separation, and the decay length was estimated to be 12 nm. This length is significantly longer than the Debye length expected in a 10 mmol‚dm-3 simple 1:1 electrolyte solution (3
Sakai et al.
nm).29 This suggests that the observed repulsion is electrosteric in nature, rather than purely electrostatic.30 A similar electrosteric repulsion was observed even when the solution pH was changed to 4 (see curve b). The force curve now shows an initial repulsion from an apparent separation of approximately 19 nm, and the decay length was estimated to be 10 nm. These values are in good agreement with the corresponding values obtained from curve a within experimental error. Hence, the two sets of force data agree quite well at larger surface separations. This result compares with that reported in our previous paper, where a significant increase in the range of the electrosteric repulsion was observed on mica after the pH was lowered from 9 to 4.12 Again, this suggests a different pH-responsive behavior for the adsorbed copolymer micelles on silica compared to that observed on mica.12,13 On the basis of the film morphology observed at pH 4 (see Figure 1b), an extension of the adsorbed copolymer chains as a result of this pH change seems very likely, and as a result, the range of the electrosteric repulsion is expected to increase. However, a direct comparison between the force curve data at pH 9 and 4 on silica is complicated by significant copolymer desorption (vide ultra). Furthermore, it is known to be difficult to define absolute separations for AFM force data, and this complicates comparison between different data sets. There is one key difference between force curves a and b at shorter apparent separations (below about 4 nm): the magnitude of the force curve at pH 9 is significantly larger than that at pH 4. At such short separations the repulsion represents compression of the adsorbed layer. At pH 9 the adsorbed micelle cores are highly constrained and are not able to move easily in response to the approaching tip. Therefore, the tip experiences increased resistance as it interacts more strongly with the adsorbed film. On the other hand, the adsorbed micelles “open out” at pH 4 due to protonation of the copolymer chains causing interchain electrostatic repulsions. Indeed, desorption of the adsorbed copolymers is expected to occur during this structural transformation. The obtained force curve data may indicate, therefore, that the bulk chain density of the normally extended brush-like layer at pH 4 is less than that observed for the adsorbed micelles at pH 9. Returning the pH to 9 results in an attractive interaction (see curve c), again indicating the irreversibility of the pH-responsive behavior for the adsorbed copolymer micelles on silica. The existence of this attractive force also limits the use of the softcontact imaging technique and contributes to the poorly resolved image observed in Figure 1c. Also shown in Figure 2 are force curve data obtained when the solution pH was returned once again to 4 (see curve d). Interestingly, a repulsive interaction was again observed (the decay length was calculated to be 9 nm), although the magnitude of the repulsion is much smaller compared with that of curves a and b. This interaction must result from chain expansion as a result of reprotonation of the adsorbed copolymer chains. The reduced magnitude again suggests that desorption of some copolymer chains most likely occurred during pH cycling. To further clarify the pH-responsive behavior at the silica/ aqueous solution interface, ζ potential values of the copolymer films on silica were estimated from streaming potential measurements. The copolymer film was established from a 50 ppm aqueous PDMA9X-PDEA2Y solution at pH 9. This copolymer (29) Israelachvili, J. N. Intermolecular and Surface Forces, 2nd ed.; Academic Press: San Diego, 1992; Chapter 12. (30) Bremmell, K. E.; Jameson, G. J.; Biggs, S. Colloids Surf., A 1998, 139, 199.
pH-ResponsiVe BehaVior of Copolymer Micelles
Langmuir, Vol. 22, No. 20, 2006 8439
Figure 2. Interaction force curves between an AFM cantilever and PDMA9X-PDEA2Y adsorbed layers on silica. These data were recorded after replacement of the 500 ppm PDMA9X-PDEA2Y copolymer solution by (a) 10 mmol‚dm-3 KNO3 at pH 9, (b) electrolyte solution at pH 4, (c) electrolyte solution at pH 9, and (d) electrolyte solution at pH 4. The inset shows the log-normal plots for curves a (circles), b (triangles), and d (squares).
Figure 3. Change in ζ potential of a PDMA9X-PDEA2Y thin film adsorbed on silica during pH cycling. The ζ potential of the bare silica plate at pH 9 was measured to be -53 ( 3 mV.14 The ζ potential was measured (I) in the presence of unadsorbed copolymer at pH 9, (II) during pH cycling in the presence of unadsorbed copolymer, and (III) during pH cycling after the removal of nonadsorbed copolymer. The error in the ζ potential was estimated to be (4 mV.
concentration has been shown to give a saturation level of adsorption corresponding to micelle adsorption.15 The estimated ζ potential values are shown in Figure 3. As reported in our previous paper,14 the ζ potential of the pristine silica surface at pH 9 was determined to be -53 ( 3 mV. As expected, the formation of the micellar thin film results in charge reversal, and only positive ζ potential values were recorded at both pH 4 and pH 9. The ζ potential data clearly show the pH-responsive behavior: the lower ζ potentials at pH 9 (20 ( 4 mV) are shifted to higher values at pH 4 (46 ( 1 mV), and this change in the ζ potential is evidently reversible. These results reflect a difference in the degree of protonation of the copolymer chains on the silica surface: in acidic solutions the copolymer chains are highly protonated (so the chains form an extended brush-like layer), whereas in alkaline solution their degree of protonation is much lower. These differences are equivalent in both the presence and absence of additional copolymer in solution. These changes in the overall charge of the copolymer thin film during pH cycling support the observed differences in the force curve data described above. Adsorbed Amount and Film Hydrophobicity. Changes in the adsorbed amount of the diblock copolymer during pH cycling are shown in Figure 4a as monitored using both the QCM-D and
Figure 4. pH cycling of the PDMA9X-PDEA2Y copolymer films adsorbed on silica from 500 ppm, 10 mmol‚dm-3 KNO3 aqueous solution at pH 9 monitored by both OR (open circles) and the QCM-D (closed circles). The mass values in region I were measured in the copolymer solution prior to pH cycling. The values were then monitored during pH cycling, in which the supernatant was replaced by fresh electrolyte solutions at either pH 4 or pH 9 (region III). (a) indicates the adsorbed mass data, while (b) shows the corresponding QCM-D dissipation data. Also shown in (a) are contact angle data recorded using an analogous experimental protocol (open triangles). The error for the QCM-D and OR data is approximately 10%, as estimated from the initial adsorption measurements. The error for the contact angle measurements is estimated within (4°.
OR. Also shown in this figure are the corresponding contact angle data. In Figure 4b the corresponding QCM-D dissipation data are given. The initial concentration was 500 ppm copolymer, and pH 9 was used for each experiment. The details for the initial adsorption were described in our previous paper.15 After equilibration, the copolymer film was repeatedly flushed with electrolyte solutions adjusted to either pH 4 or pH 9. It is emphasized, therefore, that all data presented in this figure were
8440 Langmuir, Vol. 22, No. 20, 2006
Sakai et al.
Table 1. Water Content of Diblock PDMA9X-PDEA2Y Copolymer Films on Silica at 25 °C water content ΓQCMa ΓORb (mg‚m-2) (mg‚m-2) (mg‚m-2) (wt %) initial copolymer film at pH 9 after the first rinse at pH 9 pH cycling at pH 9c pH cycling at pH 4c
15.1 13.2 4.4 7.9
2.6 1.9 0.6 0.5
12.5 11.3 3.8 7.4
83 86 87 94
a Adsorbed amount of PDMA9X-PDEA2Y on silica monitored with the QCM-D. b Adsorbed amount of PDMA9X-PDEA2Y on silica monitored with OR. c Averaged from multiple cycles at each pH. ΓQCM values were calculated using the final three values during the pH cycling.
measured when the solution pH was altered by complete replacement of the supernatant. This means that any copolymer chains desorbed from the adsorbed layer are not able to readsorb onto the thin film during pH cycling. OR studies indicated that the response of the copolymer film during pH cycling was rapid, since the adsorbed mass reaches a new equilibrium value within a few minutes. The responses in the QCM-D adsorbed mass and the corresponding dissipation were similarly fast, but for these measurements each injection of fresh KNO3 solution was limited to 1 cm3 per 15 min to maintain temperature stability in the QCM-D inner cell (whereas the OR setup permits a continuous flow of electrolyte solution). Thus, to ensure the complete pH change, the injection had to be repeated 2-4 times. The values plotted in Figure 4 are the equilibrium amounts at each step. The adsorbed mass monitored with both the QCM-D and OR decreases significantly after the first (at pH 9) and second (at pH 4) rinses. This is reasonable, because loosely bound copolymers on the silica surface would be desorbed from the interface during these rinses. The adsorbed masses indicated by the QCM-D and OR before and after the first rinse at pH 9 are listed in Table 1. It is emphasized that the OR mass corresponds to the actual mass of adsorbed copolymer, whereas the recorded QCM-D mass also includes the entrained water molecules. Combining the QCM-D and OR data allows the degree of hydration within the adsorbed layer to be estimated (Table 1). The degree of hydration after the first rinse at pH 9 is slightly higher than that before rinsing. However, this difference is deemed to be within experimental error for our setup. Similarly, the hydrophobicity of the adsorbed PDMA9XPDEA2Y layer was significantly changed after the first pH 9 rinse (Figure 4a). The contact angle of the original micellar film at pH 9 was determined to be 17 ( 4° in the presence of the copolymer supernatant. When the copolymer supernatant was replaced by the copolymer-free electrolyte solution at pH 9, the contact angle value was increased to 32 ( 4°. This increase is due to the fact that, at pH 9, the diblock PDMA9X-PDEA2Y copolymer adsorbs not only at the silica/aqueous solution interface but also at the air/solution interface. This latter adsorption results in a reduction in the surface tension, as reported in our previous papers.16,18 Consequently, when the copolymer solution is replaced by the higher surface tension electrolyte solution, Young’s equation, see eq 4,
cos θ )
γSV - γSL γLV
(4)
predicts the observed increase in the three-phase contact angle. Here, θ is the contact angle and, for example, γSV is the interfacial tension at the solid/vapor interface. Moreover, desorption of the adsorbed copolymers during this replacement of the copolymer
supernatant by the copolymer-free solution may influence the hydrophobicity of the adsorbed copolymer layer. During the subsequent pH cycling the OR mass remains effectively constant, indicating that further loss of copolymer from the thin film is negligible. On the other hand, the QCM-D adsorbed mass clearly shows reversible pH-responsive behavior in response to the solution pH: the mass significantly increases when the solution pH decreases from 9 to 4, and vice versa. Therefore, combining the QCM-D and OR mass data suggests that, while the adsorbed amount of the copolymer is almost constant during the pH cycling, the mass of the water molecules associated within the copolymer film is dramatically affected by the solution pH. The average values of the adsorbed masses from multiple cycles at each pH are summarized in Table 1. It is obvious that the degree of hydration of the copolymer films at pH 4 is greater than that at pH 9. Indeed, the mass of water in the adsorbed film at pH 9 (3.8 mg‚m-2) is almost doubled when the solution pH is changed to 4 (7.4 mg‚m-2). When one considers the measured changes in the ζ potential of the copolymer thin film (Figure 3) and the captive bubble contact angle (Figure 4a) during pH cycling, these observed differences in water content are readily rationalized. When immersed in an acidic solution, the copolymer thin film has a much higher cationic charge density, and hence lower contact angle, than when immersed in an alkaline solution. Thus, the copolymer layer is substantially more hydrophilic when in an acidic environment and will evidently incorporate more water molecules within the film. Although a gradual decrease in the QCM-D dissipation data (see Figure 4b) is observed during multiple pH cycles, the reversible pH-responsive behavior is nevertheless obvious: dissipation increases dramatically when the solution pH decreases to 4 and returns to a lower level as the pH increases to 9. As discussed above, on the basis of the AFM images and the force results, the PDMA9X-PDEA2Y diblock copolymer films adopt a normally extended brush-like morphology at pH 4 due to the protonation of the copolymer chains and the subsequent electrostatic repulsion within the film. This extension (in other words, the swelling) of the film will require greater solvation of the copolymer chains and hence result in higher dissipation because the higher water content of the copolymer film enhances the damping effect on the oscillating QCM-D sensor. On the other hand, the dissipation decreases at pH 9, supporting the hypothesis that the localized brush-like layer collapses at pH 9. This constriction of the brush-like layer would necessarily result in less entrained water. The gradual decrease in the dissipation observed during multiple pH cycles may suggest gradual degradation of the adsorbed film. Finally, it is worth noting that, although there is an apparent loss of structural integrity indicated in the AFM images, all the other data reported up to this point suggest that the layer characteristics are reversible after the initial equilibration rinses are complete. Change in Film Hydrophobicity during Continuous pH Variation. Contact angle measurements of the adsorbed copolymer films on silica have also been performed in an alternative manner, Figure 5. As previously, after equilibration of the original copolymer thin film (500 ppm, pH 9), the supernatant was replaced by electrolyte solution at pH 9, thereby removing any unadsorbed unimers and/or micelles. In the current case, however, the pH was adjusted by additions of a small amount of acid (or base) to the measurement chamber without entire replacement of the supernatant. In this way copolymer desorbed from the interface during the pH changes was not removed from the system. As discussed previously, the contact angle was increased when the
pH-ResponsiVe BehaVior of Copolymer Micelles
Figure 5. Variation in the contact angle of an adsorbed layer of PDMA9X-PDEA2Y as a function of pH on silica. The film was adsorbed at pH 9 from 500 ppm solution in 10 mmol‚dm-3 KNO3. Contact angle measurements were then made in a bulk copolymer solution (open symbol) and in 10 mmol‚dm-3 KNO3 supernatant (closed symbols). The pH was adjusted stepwise down to pH 4 (s) and then directly back to pH 9 (---). These data therefore correspond to region II in Figure 2, where desorbed copolymer chains remain in the supernatant. The solid and dashed lines are not mathematical fits to the data, but merely lines to guide the eye.
supernatant containing nonadsorbed copolymer chains was replaced with copolymer-free electrolyte solution at a constant pH of 9. The contact angle then significantly increased as the solution pH was decreased incrementally, reaching a maximum in the range of 50-55° at a solution pH of ∼6.5. Further reduction in the solution pH to 4 yields a contact angle lower than that originally observed in alkaline conditions, in accordance with the data obtained when the pH is cycled by complete supernatant replacement (see Figure 4a). On the basis of these results, we propose that the unexpected increase in the contact angle in the neutral pH regime is due either to desorption of some of the copolymer chains from the top portion of the adsorbed micelles or to the onset of dissociation of the surface-adsorbed micelles. Consequently, the core-forming PDEA blocks, which are significantly more hydrophobic than the coronal-forming PDMA blocks, are partially exposed to the solution phase, and hence, the contact angle increases in the intermediate pH range. Since the dissociation constant (pKa) of the DEA unit is reported to be 7.3,31 the hydrophilicity of the core-forming PDEA blocks is expected to be gradually increased when the solution pH is decreased. This increase in the hydrophilicity of the PDEA blocks may enable such exposure against the thermodynamically unfavorable contact between the relatively hydrophobic micelle core and the aqueous solution. Indeed, in a further OR experiment, the desorbed amount upon stepwise reduction of the pH from 9 to 6.5 recorded the loss of about 32% of the adsorbed material (visible in Figure 1). This supports the exposure of the PDEA core blocks in the intermediate pH range. The subsequent decreasing contact angle to a minimum at pH 4 (17 ( 2°) is reasonable given the formation of the highly charged brush-like copolymer layer, which has the highest affinity for the aqueous phase. Interestingly, when the pH is then restored directly to 9, the contact angle returns to approximately the same value recorded prior to the pH cycle. Again, it is emphasized that in this experiment the supernatant was not replaced once pH cycling had commenced; hence, copolymer chains desorbed from the interface are not removed from the system. It is likely that chains desorbed from the micellar layer during pH cycling are therefore able to readsorb onto the thin film. This readsorption is driven (31) Merle, Y. J. Phys. Chem. 1987, 91, 3092.
Langmuir, Vol. 22, No. 20, 2006 8441
by the hydrophobic nature of the PDEA residues under such alkaline conditions: they will preferentially associate with hydrophobic domains within the surface-bound layer. This reversibility in contact angle is in conflict with the observed behavior when nonadsorbed copolymer is removed (see Figure 4a), where significant hysteresis is evident when the solution pH is returned to 9. The increased contact angle of 48° indicates that the PDEA core blocks remain partially exposed on returning to pH 9 in those studies where the experimental protocol necessitates complete replacement of the supernatant during pH cycling, i.e., AFM, QCM-D measurements, and OR. Therefore, the attractive interaction observed in Figure 2c may indicate an interaction between the AFM cantilever tip and the hydrophobic PDEA core blocks exposed to the solution phase. Also, the observed decrease in adsorbed amount after the initial rinses at pH 9 and 4 must be a result of the removal of desorbed copolymer. Further experiments in which the desorbed copolymer is not removed should quantify readsorption when the supernatant is returned to alkaline conditions.
Conclusions The pH-responsive behavior of the PDMA-PDEA diblock copolymer micelles adsorbed at the silica/aqueous solution interface has been characterized. Protonation of the adsorbed copolymers at pH 4 causes electrostatic repulsion of the copolymer chains in the adsorbed micelles and the subsequent extension of the chains into the solution phase. As a result, the adsorbed micelles at pH 9 are observed to open up at pH 4, forming an extended brush-like layer. This localized brush-like layer observed at low pH collapses as the solution pH returns to 9, and the micellar structure is partially re-formed. However, the resolution and clarity of the obtained image is much less than that of the original micelle images at pH 9, suggesting that this pH-responsive behavior is structurally irreversible on silica. The ζ potential values of the PDMA-PDEA film indicate reversible charging and discharging during pH cycling, within experimental error. The adsorbed/desorbed amounts of the copolymer have also been estimated using the QCM-D and OR. After the initial rinse at both pH 9 and pH 4, the OR adsorbed mass is almost constant during subsequent pH cycling, whereas the corresponding QCM-D adsorbed mass exhibits pronounced pH-responsive behavior: the mass increases significantly as the solution pH decreases to 4 and returns to the lower level when the pH is restored to 9. These results indicate that the normally extended brush-like layer at pH 4 entraps many more water molecules than the adsorbed micelles at pH 9. This higher degree of hydration at pH 4 is supported by the higher QCM dissipation of the adsorbed layer at this pH. Contact angle measurements indicate that the transformation of the layer to an extended brush-like conformation is preceded by a transitional morphology in which the relatively lowly charged hydrophobic PDEA blocks are exposed to the solution phase prior to full protonation of the tertiary amine residues. This desorption of PDMA-PDEA chains from the top portion of the adsorbed micelles results in the following phenomena: (i) the structural irreversibility of the adsorbed layer during pH cycling, (ii) the attractive interaction at pH 9 after pH cycling, (iii) the increase in the contact angle at the intermediate pH range, and (iv) the difference in the contact angle at pH 9 before and after pH cycling when the desorbed copolymer is removed from the system. When the solution pH is changed without complete replacement of the supernatant, the contact angle is observed to be reversible, within error, over a pH 9-4-9 cycle, indicating
8442 Langmuir, Vol. 22, No. 20, 2006
that desorbed copolymer chains do in fact readsorb onto the surface film. Acknowledgment. The EPSRC is thanked for the linked Research Grants GR/S60419 and GR/S60402 awarded to S.P.A. and S.B., respectively. The ARC is thanked for Research Grant
Sakai et al.
DP0343783. The Newcastle University School of Engineering is thanked for access to the contact angle instrument. S.P.A. is the recipient of a five-year Royal Society-Wolfson Research Merit Award. LA061708F
