Article pubs.acs.org/Langmuir
Critical Salt Effects in the Swelling Behavior of a Weak Polybasic Brush Joshua D. Willott,† Timothy J. Murdoch,† Ben A. Humphreys,† Steve Edmondson,‡ Grant B. Webber,† and Erica J. Wanless*,† †
Priority Research Centre for Advanced Particle Processing and Transport, University of Newcastle, Callaghan, NSW 2308, Australia Department of Materials, Loughborough University, Loughborough LE11 3TU, United Kingdom
‡
S Supporting Information *
ABSTRACT: The swelling behavior of poly(2-(diethylamino)ethyl methacrylate) (PDEA) brushes in response to changes in solution pH and ionic strength has been investigated. The brushes were synthesized by ARGET ATRP methodology at the silica−aqueous solution interface via two different surface-bound initiator approaches: electrostatically adsorbed cationic macroinitiator and covalently anchored silane-based ATRP initiator moieties. The pH-response of these brushes is studied as a function of the solvated brush thickness in a constant flow regime that elucidates the intrinsic behavior of polymer brushes. In situ ellipsometry equilibrium measurements show the pH-induced brush swelling and collapse transitions are hysteretic in nature. Furthermore, high temporal resolution kinetic studies demonstrate that protonation and solvent ingress during swelling occur much faster than the brush charge neutralization and solvent expulsion during collapse. This hysteresis is attributed to the formation of a dense outer region or skin during collapse that retards solvent egress. Moreover, at a constant pH below its pKa, the PDEA brush exhibited a critical conformational change in the range 0.5−1 mM electrolyte, a range much narrower than predicted by the theory of the osmotic brush regime. This behavior is attributed to the hydrophobicity of the collapsed brush. The swelling and collapse kinetics for this salt-induced transition are nearly identical. This is in contrast to the asymmetry in the rate of the pH-induced response, suggesting an alternative mechanism for the two processes dependent on the nature of the environmental trigger.
■
INTRODUCTION Polymer brushes, formed by anchoring one end of a polymer chain to a surface, have been extensively used to modify the physical properties and chemistry of interfaces.1 Interfacial properties can be tailored by suitable polymer selection, allowing control over the resulting surface charge state, wettability, lubricity,2 and tuning of the response to changes in environmental conditions like temperature,3−6 pH,7−15 or ionic strength.16 Grafted brushes of polyelectrolytes, specifically weak polybases and polyacids, have attracted much attention due to their pH-responsive characteristics. When immersed in aqueous solution, such brushes become charged as a function of pH as dictated by the brush pKa.10,17 When considering a polybase, at pH values below the brush pKa the brush becomes charged due to protonation of the base groups on the brush, while at high pH (above the brush pKa) the brush becomes uncharged. When charged, a polyelectrolyte brush swells due to absorption of water and counterions into the brush. The reverse neutralization of the charged groups and expulsion of both the water and counterions from within the brush is accompanied by collapse of the brush layer. Together these transitions are known as the pH-response of the brush. The behavior of polyelectrolyte brushes is principally controlled by electrostatic repulsion between the charged monomers © 2014 American Chemical Society
and the osmotic pressure of the counterions. Consequently, screening of these interactions upon an increase in ionic strength represents a second mechanism, along with pH, of manipulating the structure and properties of surfaces modified with weak polyelectrolyte brushes. For such brushes raising the ionic strength decreases the range and strength of the repulsive electrostatic interactions, while simultaneously increasing the number of charged monomer groups in the polymer.18−20 Different brush behavioral regimes exist depending on the chain length, grafting density, the charge fraction, and the ionic strength of the solution.18,21 For polyacid brushes, the effect of ionic strength and pH on their behavior has been extensively investigated, both theoretically and empirically. Most experimental studies of weak polyacid brushes focus on poly(acrylic acid) and its derivatives and report primarily on the effects of pH and salt on equilibrium brush thickness.16,22−24 Polyacid brushes have been shown to exhibit the predicted nonmonotonic behavior as a function of added salt, where at high and low salt concentration the brush collapses.16,22−24 Rapid swelling and slow collapse kinetics of poly(methacrylic acid) Received: December 9, 2013 Revised: January 28, 2014 Published: January 29, 2014 1827
dx.doi.org/10.1021/la4047275 | Langmuir 2014, 30, 1827−1836
Langmuir
Article
regenerated by electron transfer (ARGET) ATRP is a recent variant of ATRP wherein a reducing agent, such as ascorbic acid, is used to generate the active, oxygen-sensitive Cu(I) catalyst species in situ by reduction of the oxidatively stable Cu(II) deactivator.40 The reducing agent is present in excess thus imparting a degree of oxygen tolerance to the polymerization, in contrast to more stringent conditions required in conventional ATRP. We have previously reported on the use of surface-initiated ARGET ATRP to synthesize PDEA brushes from colloidal silica.12,13 Therein, we observed the brushes to be highly charged and swollen due to protonation at pH values below the pKa of the free polymer of 7.3.41 The hydrodynamic diameter of the PDEA hybrid particles was observed to increase significantly as the pH was reduced from pH 7 to 4, even though the brush charge as recorded using electrophoretic zeta potential measurements only increased slightly. An increase in hydrodynamic diameter equates to a decrease in particle diffusion rate, which is caused by the brush absorbing more water and counterions leading to an increase in brush thickness. At pH values above the pKa of the free polymer the brush was observed to neutralize and the hybrid particles aggregated as the hydrophobicity now dominated any charge on the PDEA brushes. Significantly, this neutralization occurred over several pH units and a cationic zeta potential was still measured at pH 9. Thus, the association and dissociation of protons to the confined polymer chains occurs over a much broader pH-range than for untethered polybases. Our group has also published work on the kinetics and equilibrium pH-response of PDEA brushes grafted from planar surfaces, reporting that the collapse process was eight times slower than the corresponding swelling transition. These findings are consistent with theory on polyelectrolyte network swelling and collapse42 as well as the observed behavior of adsorbed PDEA microgel films43 and PMAA brushes.25 Herein we report on the equilibrium and pH-induced swelling and collapse kinetics of polybasic PDEA brushes at three different solution ionic strengths: 10, 50, and 100 mM. Such high temporal resolution kinetic data on the swelling behavior of polybasic brushes have not been widely reported. Furthermore, we have investigated the equilibrium thickness of a PDEA brush, maintained below its pKa, as the solution ionic strength is varied from 0.1 to 500 mM. Here we observe a significant change in brush conformation over a critical range of low ionic strength much narrower than predicted by theory. Differences in the kinetics for the pH-induced and electrolyteinduced changes in brush conformation suggest alternative swelling/collapse mechanisms dependent on trigger type.
(PMAA) brushes in response to pH changes have been recently reported by Cui et al.25 Here it is clear that the expulsion of water during brush collapse was hindered relative to the ingress of water during the swelling transition. While there are reports concerning the swelling, collapse, and equilibrium behavior of weak polybasic brushes in response to changes in pH,10,14,15,26−30 the current literature is lacking welldefined studies detailing the influence that ionic strength has on the empirical behavior of such brushes. Theory predicts that brush thickness varies non-monotonically with added inert electrolyte for a weak polyelectrolyte brush.18 When a weak polybasic brush is immersed in a hypothetical zero ionic strength solution, each charged base group is associated with a hydroxide counterion in order to preserve electroneutrality, and polymer hydrophobicity will affect the conformation of the brush. When the number of added salt ions is low the anions can replace the hydroxide counterions from within the brush, thus increasing the degree of dissociation inside the brush. Essentially, the local pH inside the brush becomes closer to that of the reservoir or bulk pH. As a consequence, the addition of small quantities of salt leads to a larger fraction of dissociated basic monomer units, resulting in an increase in the osmotic pressure inside the brush. This is known as the “osmotic-brush” regime and brush thickness increases with rising ionic strength with a predicted scaling exponent of +1/3.18 When the concentration of the bulk electrolyte approaches that of the counterions inside the brush, the fraction of charged monomers can no longer increase: the “salted-brush” regime. Now, any further rise in the solution ionic strength causes the brush to contract since screening of the electrostatic repulsion between monomer groups becomes significant.31 Here weak polyelectrolyte brushes behave similarly to strong or constantly charged polyelectrolyte brushes, i.e., a decrease in brush thickness which is predicted to scale with increasing salt to the −1/3 power.18 An example of a weakly polybasic polymer is poly(2(dimethylamino)ethyl methacrylate) (PDMA), a tertiary amine methacrylate, which displays classic pH-responsive behavior with a pKa in the vicinity of physiological pH.29,32−34 Brushes of other closely related tertiary amine methacrylate polybasic polymers such as poly(2-diethylamino)ethyl methacrylate) (PDEA) and poly(2-(diisopropylamino)ethyl methacrylate) (PDPA) have been shown to display similar pH-responsive behavior.10,15,27,35−38 However, when uncharged, these polymers are more hydrophobic, adding a hydrophilic−hydrophobic solubility balance that is not present when studying PDMA. A study on PDMA brushes anchored to hydrophobic poly(methyl methacrylate) latex particles showed that the hydrodynamic diameter of the hybrid particles passed through a maximum as the concentration of added sodium chloride was varied within the range of 0−1000 mM.26 At high ionic strength, the reduction in the particle hydrodynamic diameter corresponded to the PDMA brush thickness decreasing as the electrostatic repulsion was screened in accordance with the salted brush regime, and water molecules and counterions were expelled from within the brush layer. Recent advances in controlled radical polymerization have allowed polymer brushes to be easily grafted from surfaces with excellent control over characteristics like brush length and dispersity. Atom transfer radical polymerization (ATRP) is widely used for synthesizing polymer brushes because of its controlled nature, the range of compatible monomers, and the ability to use relatively mild reaction conditions.39 However, a major drawback of ATRP is its high oxygen sensitivity. Activators
■
EXPERIMENTAL SECTION
Poly(2-(diethylamino)ethyl methacrylate) (PDEA) brushes were synthesized using an optimized ARGET ATRP protocol as described previously.12 Two brushes were used for these experiments, both grown on oxidized silicon wafers. However, the nature of the surfacebound initiator used to synthesize the two brushes was different. One brush was grown from an electrostatically adsorbed cationic ATRP macroinitiator while the other was synthesized from a covalently bound silane-based ATRP initiator monolayer. Materials. Silicon wafers were purchased from Silicon Valley Microelectronics, Santa Clara, CA. A cationic poly(2-(dimethylamino)ethyl methacrylate-stat-glycerol monomethacrylate) based water-soluble macroinitiator was used, the details of which have been reported previously;13 see also Supporting Information Figure S1. Silane-initiator functionalization reagents including (3-aminopropyl)triethoxysilane 1828
dx.doi.org/10.1021/la4047275 | Langmuir 2014, 30, 1827−1836
Langmuir
Article
index of the solution from that of pure water is minimal.49 As such, the variance in the final fitted brush thickness is small when using the different solution refractive indices, thereby justifying the use of this value. The dry brush thickness measurements are an average of three different positions on the wafer surface and were made at three angles of incidence (60°, 52°, and 40°) to improve measurement accuracy. In Situ Ellipsometry on PDEA Brush-Modified Wafers. In situ ellipsometric measurements were performing using a Nanofilm SL fluid cell with optical glass windows at a laser beam angle of incidence of 60°. The cell had a trapezoidal geometry with an internal volume of 0.7 mL and exposed sample area of 15 mm × 22 mm. Only a single angle of incidence was available as the beam must be perpendicular to the cell windows. Following the dry ellipsometric measurement described above, the fluid cell was fitted and, after ensuring that concordant dry thickness measurements could be attained (adjusting position and alignment if necessary), aqueous solution was flowed into the cell as desired. All reported in situ ellipsometry measurements were made under a constant flow of aqueous electrolyte solution between 3.5 and 5.2 mL min−1, and the pH was maintained over time scales of at least 25 min until a steady state was achieved before the solution pH was changed. For the kinetic experiments the flow rate was fixed at 4.3 mL min−1. These flow rates were selected to ensure rapid change of the pH of the solution inside the fluid cell (with a fluid residence time of 12 s or less and well below 1 min to switch the cell pH between 9 and 4), and in order to work in a constant supply regime where the pH inside the cell is maintained at the desired level by controlling the pH of the external electrolyte reservoir.14 For overnight storage between measurements, the fluid cell was left filled with electrolyte solution at unadjusted pH (∼5.5) with the pump turned off. The brush was maintained in an aqueous environment for the duration of all experiments. Nulling one-zone measurements, to determine the ellipsometric quantities ψ and Δ, were performed every 15 s. Each ψ and Δ pair recorded at a given time can be modeled to determine brush thickness. The in situ ellipsometric data were fitted using a multilayer model consisting of sequentially; a 1 mm silicon layer, a 2.5 nm silica layer, a linear effective medium approximation (EMA, a method of interpolating the dielectric properties, such as refractive index, for a layer of mixed composition) layer of water and polymer of unknown thickness and composition, and a final ambient water layer.10 Each layer in the model is assumed to have a constant density throughout the thickness, i.e., as a slab. This study was conducted with a single wavelength and fixed angle instrument, making more detailed modeling of the solvated brush layer, such as multislab or gradient slab models, unreliable due to the greater number of unknowns to be fitted. In previous work on swollen brushes by Edmondson et al., it has been shown that although using a slab model can lead to inaccuracies in the absolute fitted thickness relative to gradient layer models (which more accurately describe the polymer density profile), the trends in the data are expected to be equivalent.50 In this previous work, thicknesses calculated using a graded layer with an exponential density decay and those calculated with a slab model were closely correlated. Fitted thicknesses from the graded model (using a root-mean-square measure of layer thickness) varied between 72% and 89% of the single slab thickness, with the thicknesses from both models having a very good linear correlation with an R2 = 0.995. Since this prior work dealt with brushes of a very similar thickness range and swelling ratios to the current work, we have good confidence that the single slab model closely reproduces the true changes in brush thickness.
(APTES), tetrahydrofuran (THF), triethylamine (Et3N), and 2-bromoisobutyryl bromide (BIBB) were purchased from SigmaAldrich. The THF and triethylamine were dried over 4 Å molecular sieves before use. All aqueous in situ ellipsometry measurements were conducted in the presence of potassium nitrate electrolyte (Asia Pacific Specialty Chemicals Ltd., >99%, used as received). Solution pH values were accurate to ±0.1 pH unit, and adjustments were made by adding the minimum amount of 0.01 or 0.1 M nitric acid (RCI Labscan Ltd., 70%) or potassium hydroxide (Chem-Supply Pty. Ltd., AR grade, used as received) to a feed reservoir of electrolyte solution, which was subsequently flowed into the sample chamber through air-impermeable Tygon tubing and a peristaltic pump. Milli-Q water was used to prepare all solutions. All other reagents were purchased from SigmaAldrich and used as received. Wafer Preparation and Initiator Functionalization. Pieces of silicon wafer with a 2.5 ± 1 nm natural silica layer (measured by ellipsometry) were rendered hydrophilic by irradiation for 20 min with UV/ozone before being sonicated in Milli-Q water for 20 min. The water was replaced every 5 min. Subsequently, the wafers were rinsed with Milli-Q water before being immersed in 10 wt % sodium hydroxide solution for 30 s and further rinsed with copious amounts of Milli-Q water. Cleaned wafers were treated in one of two ways, listed below, in order to introduce initiator functionality (bromine initiating sites) to the surface. ATRP Macroinitiator Adsorption. Cleaned wafers were immersed in 1 mg mL−1 macroinitiator solution for 17 h to allow adsorption to reach equilibrium. We have previously measured using a quartz crystal microbalance that at this concentration the adsorbed macroinitiator amount was 0.8 mg m−2.14 Furthermore, we have shown using solvent relaxation nuclear magnetic resonance spectroscopy that this value is in the plateau region of the isotherm.13 Macroinitiator-modified wafers were removed from the macroinitiator solution and rinsed with copious amounts of Milli-Q water and blown dry under a stream of nitrogen gas before being transferred to the polymerization solution for PDEA brush synthesis. The dry thickness of the macroinitiator layer was measured by ellipsometry to be less than 1 nm.44 Silane ATRP Initiator. Cleaned wafers were placed in a clean vacuum desiccator along with a small vial containing 10 drops of (3-aminopropyl)triethoxysilane (APTES). The desiccator was then evacuated to
