Thin Walled Imprinted Polymer Beads Featuring Both Uniform and

Jul 13, 2012 - SIP combined with CRP allows polymer films with controllable .... Corporation, Boynton Beach, FL) automatic adsorption instrument. ... ...
1 downloads 0 Views 2MB Size
Article pubs.acs.org/cm

Thin Walled Imprinted Polymer Beads Featuring Both Uniform and Accessible Binding Sites Mahadeo R. Halhalli, Eric Schillinger, Carla S. A. Aureliano, and Börje Sellergren* INFU, Faculty of Chemistry, Technical University of Dortmund, Otto-Hahn-Str. 6, D-44221, Dortmund, Germany S Supporting Information *

ABSTRACT: A novel approach addressing the classical deficiencies of molecularly imprinted polymers (MIPs), that is, low binding capacity and nonuniform binding sites, is reported. The thin walled beads were produced in two steps by first grafting thin MIP films, under controlled (RAFT) or noncontrolled conditions, from porous silica beads following previously reported procedures. The resulting composites were compared in terms of film thickness, the grafted layer homogeneity, the effect of different support morphologies, and for their ability to recognize the template in chromatographic or static binding tests. Thus, using L-Phenylalanine anilide (L-PA) as template to imprint poly(MAA-co-EDMA) in such a way led to nanometer thick films where the resulting composite were able to selectively retain the template in relation to the thickness of the grafted film. In the second step, removing the silica supports from the above composites by etching, led to nanometer thin walled beads with structure, morphology and recognition properties strongly depending on grafting chemistry (RAFT or non-RAFT), monomer dilution and on the film thickness of the original composite. Thus whereas the thicker walled materials retained their mesoporous morphology and displayed enhanced enantioselectivity, load capacity, and higher surface areas compared to their composite precursors, the thin walled beads showed lower surface areas indicating network collapse. The thin walled beads prepared under dilute conditions in absence of RAFT displayed a perfectly uniform binding site distribution and a saturation capacity exceeding that of a conventional monolithic MIP. The beads prepared by RAFT control showed a further enhanced saturation capacity significantly exceeding that of the reference material. Finally, the reduced hydrophobic character of the thin walled materials indicated the existence of two separate pore systems with different pore wettabilities. KEYWORDS: molecular imprinting, grafting, surface initiatied polymerization, RAFT, template synthesis, isotherm, capacity



INTRODUCTION

controllable thickness, composition, and structure to be prepared. The concept of template synthesis allows on the other hand porous materials with different morphologies to be prepared.1 Here either an organic polymer may serve as a shape template for the synthesis of an inorganic porous network or alternatively an inorganic material serves as template for the synthesis of organic materials of defined morphology.6 In the latter, porous silica has been used as a sacrificial template for the synthesis of mesoporous organic polymer networks.7 Templating at a molecular level is referred to as molecular imprinting where the shape and functionality of individual molecules are molded into a network polymer.8−10 This approach has been used to generate porous materials exhibiting receptor-like affinity for a large variety of template structures. One recurring problem in the development of a new molecularly imprinted polymer (MIP) is the incompatibility between conditions that are

The ability to control the structure and composition of materials at a nanometer scale is key to a number of advanced functions within diverse areas such as drug delivery, diagnostics and sensing, molecular electronics, catalysis, separations, or as mimics of biological systems.1,2 Among the most promising approaches and techniques in nanomaterial design are on the one hand grafting and controlled radical polymerization (CRP)2,3 and on the other hand templated synthesis and molecular imprinting.1,4 For instance, starting from an inorganic support of known morphology, nanocomposites can be synthesized by grafting an organic polymer film onto the surface. Grafting can be performed following two different approaches, “grafting to” or “grafting from” where the latter, when performed on a surface, is referred to as surface inititated polymerization (SIP).3,5 SIP relies on surface immobilized initiator species or in situ generated radicals leading to reactions mainly between monomers and surface confined radicals. This is commonly preferred as it produces a high density of grafted chains. SIP combined with CRP allows polymer films with © 2012 American Chemical Society

Received: March 27, 2012 Revised: June 29, 2012 Published: July 13, 2012 2909

dx.doi.org/10.1021/cm300965t | Chem. Mater. 2012, 24, 2909−2919

Chemistry of Materials

Article

Figure 1. Grafting of L-phenylalanine anilide (L-PA) imprinted polymer films from a porous silica support modified with azoinitiator (A), or RAFT agent (R). (i) The grafting was performed using a common prepolymerization mixture to reach high conversions followed in (ii) by removal of the silica support by fluoride etching.

template can either be immobilized to the walls of the mold, or the template can be simply dissolved in the monomer mixture. Relatively few studies have reported the combination of controlled grafting and template synthesis. Walt et al. used atom transfer radical polymerization (ATRP) to graft thick polymer layers on porous silica.25 Etching away the silica template hollow spheres remained with a relatively thick shell with thickness larger than 175 nm. Arnold and Yang reported on thin walled coaxial nanotubes prepared by ATRP from initiator modified silica/silicone core shell nanowires,26 while Wang et al. applied a similar approach to produce molecular imprinted nanotube membranes for separation and sensing14,15 of TNT.16 Here we wish to report on the combination of the two approaches to generate a novel class of porous materials. Grafting a thin film onto a disposable support and subsequently removing the support we anticipated would leave behind a porous material with thin walls (Figure 1). The wall thickness will impact the stability of the framework leading to either gellike materials with a collapsed structure when the walls are thin or permanently porous morphologies when the walls are sufficiently thick. In the latter case the support removal will lead to an increase in surface area impacting the saturation capacity of the materials when used for adsorption under static conditions or in chromatography. If the grafting is performed by living CRP, multiple layers may be grafted exhibiting different composition, structure, and function. After removing the support the innermost layer (the first grafted layer) would be exposed within walls which thus would contain two nonequivalent surfaces. To test whether this approach could produce such effects we have started from an identical model system used in our

optimal to generate the templated binding sites at a molecular level with those leading to the desirable format or morphology at the nano- or microscopic level, that is, particle and pore sizes, surface areas and swelling properties. To circumvent these problems imprinting techniques relying on either grafting or so-called “hierarchical imprinting” have been developed. For instance, the “grafting from” technique has been used by several research groups to graft imprinted polymer layers on various substrates.11−20 We reported on “grafting from” techniques for the synthesis of MIP composite materials with improved kinetic properties.12,21 The first of these consisted in the use of immobilized azo-initiators which allowed the synthesis of thin film MIP composites with much improved mass transfer characteristics.12 Some control of the grafting was achieved by the use of immobilized dithiocarbamate initiators (commonly referred to as iniferters)22,21 or the use of chain transfer agents (RAFT agents).23,24 The living character of the iniferter polymerization was used for layer by layer grafting of different MIPs onto wide pore silica. Recently, we assessed the relative merits of the different grafting techniques for producing imprinted thin film composite materials.24 This led to the conclusion that azoinitiated polymerizations that were allowed to proceed to high conversion under dilute conditions or in presence of an excess of chain transfer agents (RAFT) resulted in films with an enhanced homogeneity and template binding affinity. Hierarchical imprinting on the other hand takes advantage of the morphology control offered by template synthesis to obtain imprinted polymer beads exhibiting molecular recognition properties combined with a predefined and unique morphology. Porous silica is used as a mold to control the particle size, shape, and porosity of the resulting imprinted polymer.7 The 2910

dx.doi.org/10.1021/cm300965t | Chem. Mater. 2012, 24, 2909−2919

Chemistry of Materials

Article

previous investigations.24 This refers to poly(MAA-co-EDMA) imprinted with an enantiomerically pure template (L-phenylalanine anilide (L-PA)). Imprinting effects are here straightforwardly assessed as the ability of the materials to discriminate between the template and its optical antipode as well as their affinity for the template. L-PA imprinted polymers were thus grafted from a common silica support using an immobilized azo initiator or by CRP from RAFT modified supports. Evaluation was performed by characterizing the pore structure, morphology, pore wettability, and template recognition of the polymers prior to and post silica removal.



Engineering, TU Dortmund. The samples were deposited on holders with a carbon foil without gold sputtering. Thermogravimetric Analysis (TGA). TGA was carried out using a TGAQ50 (TA Instruments, ESchborn, Germany). The sample (10− 15 mg) was placed in a platinum pan, which is suspended in a sensitive balance together with the reference pan. The sample was then heated in a furnace with a heating rate of 20 °C/min, under N2 atmosphere. Nitrogen Sorption. Nitrogen sorption measurements were performed on a Quantachrome Nova 4000e (Quantachrome Corporation, Boynton Beach, FL) automatic adsorption instrument. Prior to measurements, 100−150 mg of the samples were heated at 40−60 °C under high vacuum (10−5 Pa) for at least 12 h. The specific surface areas (S) were evaluated using the Brunauer−Emmett-Teller (BET) method, the specific pore volumes (Vp) following the Gurvitch method and the average pore diameter (Dp) using the Barrett− Joyner−Halenda (BJH) theory applied to the desorption branch of the isotherm. Energy Dispersive X-ray Analysis (EDX). This was performed using a SEM Hitachi S 4500 at the Fachbereich, Technische Chemie, Universität Dortmund. Swelling Tests. NMR tubes were filled during intermittent vibrations up to 1 cm (142 μL) with dry polymer particles and the solvent (1 mL) added. The particles were allowed to equilibrate in the solvent for 24 h, whereafter the volume of the swollen particles was measured. The volume swelling ratio was calculated as: Swelling ratio = bed volume swollen particles (Vs)/bed volume dry particles (Vd). Azo-Initiator Immobilization. The silica supports were modified with azoinitiator in two steps as previously reported.12 Silanization of rehydroxylated silica with APS was followed by condensation of the initiator ACPA with the surface amino groups to give the initiator modified supports listed in Supporting Information, Table S1. In 250 mL three-necked round-bottom flasks equipped with a condenser, an overhead stirrer and a dropping funnel, 6 g batches of rehydroxylated silica were suspended in 80 mL of dry toluene. The whole system was flushed with N2. According to the number of silanol groups on the silica surface (8 μmol/m2) the appropriate amounts of APS were added and the solution refluxed overnight at 110 °C. The products were filtered through glass funnels and washed with 2× 50 mL of toluene and 2× 50 mL of MeOH followed by drying in a vacuum oven at 40 °C for 24 h. For the coupling of ACPA, dry THF (250 mL) in a 500 mL threenecked round-bottom flask, equipped with a dropping funnel, on overhead stirrer, and a ethanol thermometer, was cooled to −78 °C using a liquid-nitrogen-ethanol bath. Under continuous N2 flow, azoinitiator (ACPA) (5.04 g; 18 mmol), ethylchloroformate (1.95 g; 18 mmol), and triethylamine (1.82 g; 18 mmol) were added to the flask. After stirring for 30 min at −78 °C, 25 g of amino-modified silica was added to the mixture, and the resulting suspension stirred for 3 h at −78 °C and then for 4 h at −10 °C. The product was filtered, washed with THF and MeOH, and dried under vacuum at room temperature. The product was then characterized by elemental microanalysis, FT-IR spectroscopy, and TGA, and the amount of coupled ligands was estimated. Immobilization of RAFT Agent. In a three-necked round-bottom flask (250 mL), equipped with a dropping funnel, an overhead stirrer, and an ethanol thermometer, 200 mL of dry THF was introduced, and the flask purged with nitrogen. For Si-APS2: 1.60 g (5.74 mmol) of 4cyanopentanoic acid dithiobenzoate, 0.62 g (5.75 mmol) of ethylchloroformate, and 0.58 g (5.75 mmol) of triethylamine and for Si500APS: 1.395 g of 4-cyanopentanoic acid dithiobenzoate, 543 mg of ethylchloroformate, and 506 mg of triethylamine were consecutively added. The mixture was then cooled at −78 °C using a liquidnitrogen-ethanol bath. After stirring for 30 min, aminomodified silica (Si-APS2: 15 g; Si500-APS: 25 g) was added to the mixture, and the suspension was stirred for 3 h at −78 °C and then for 4 h at −10 °C. The product was then filtered, washed with THF and MeOH, and dried under vacuum at room temperature. The surface density of RAFT agent calculated based on % mass loss by TGA was 0.72 μmol/ m2 (Si-RAFT) and 3.32 μmol/m2 (Si500-RAFT).

EXPERIMENTAL SECTION

Materials. N,N′-dicyclohexylcarbodiimide (DCC), trifluoroacetic acid, ethylene glycol dimethacrylate (EDMA), methacrylic acid (MAA), (3-aminopropyl)triethoxysilane (APS), and ethyl chloroformate were obtained from Aldrich (Deisenhofen, Germany). EDMA was purified by washing consecutively with 10% aqueous NaOH, water, brine, and finally water. After drying over MgSO4, pure, dry EDMA was obtained by distillation under reduced pressure. MAA was purified by distillation under reduced pressure. 1-Hydroxybenzotriazole hydrate (HOBt), 4,4′-azobis (4-cyanopentanoic acid) (ACPA), 3amino-quinoline, ethyl acetate, methanol, and acetic acid (AcOH) were obtained from Fluka (Deisenhofen, Germany). Phenylacetonitrile, N-bromosuccinimide, phenyl magnesium bromide (3 M solution in ethyl ether), deuterated chloroform, magnesium sulfate, benzoyl peroxide, carbon disulfide, carbon tetrachloride were acquired from Aldrich chemical company and were used as received. HPLC grade acetonitrile (MeCN) and methanol were purchased from Merck. The templates L -and D -phenylalanine anilide (L-PA, D-PA) were synthesized following a previously described procedure.12 Anhydrous solvents, dimethyl formamide (DMF), toluene, and tetrahydrofuran (THF) were purchased from Fluka and used as received. The following silicas were used as supports. For Si-APS-ACPA the support was Licrosphere Si100 with the following charactersitics: (S = 380m2/ g, Dp = 12 nm, Vp = 1.26 mL/g). The soluble RAFT agent (cyanobenzyl dithiobenzoate) was synthesized as previously reported.36 For the RAFT immobilization either mesoporous silica beads (Licrosphere Si100) (15 μm average particle size) with a surface area (S) of 320 m2/g (after treatment with 17% HCL); an average pore diameter (Dp) of 11.6 nm and an pore volume (Vp) of 1.28 mL/g or macroporous beads (Si500) (Fuji Silysia Japan) (30 μm average particle size) with a surface area (S) of 45 m2/g; an average pore diameter (Dp) of 47.5 nm, and an pore volume (Vp) of 0.81 mL/g) was used. The RAFT agent 4-cyanopentanoic acid dithiobenzoate was purchased from Stream chemicals. Apparatus and Methods. High Performance Liquid Chromatography (HPLC). The HPLC measurements were carried out on Hewlett-Packard HP 1050 instruments (Agilent Technologie, Waldbronn, Germany). Fluorescense Microscopy. The fluorescent labeled silica particles were investigated using a LEICA DM R fluorescence microscope HC (Benzheim, Germany). Elemental Analysis. Carbon, nitrogen, and sulfur contents were determined by elemental analysis at the Department of Organic Chemistry, Johannes Guttenberg Universität Mainz using a Heraeus CHN-rapid analyzer (Hanau, Germany). FT-IR Spectroscopy. This was performed using a NEXUS FT-IR spectrometer (Thermo Electron Corporation, Dreieich, Germany). The samples were prepared by adding the solid (∼ 2 mg, carefully dried) to KBr salt, this matrix was grinded and mixed with an agate mortar and pestle, and then pressed into a transparent disk or pellet at sufficiently high pressure Scanning Electron Microscopy (SEM). The particle morphology, size and size distribution were determined using a Hitachi H-S4500 FEG Microscope in secondary electron mode with an acceleration voltage of 1 kV at the Department of Biochemical and Chemical 2911

dx.doi.org/10.1021/cm300965t | Chem. Mater. 2012, 24, 2909−2919

Chemistry of Materials

Article

Nonlinear fitting of theoretical isotherms to experimental data was performed using Microcal Origin 5.0, and best fits were evaluated with the Fisher test where a higher F-value indicates a better fit.27 The adsorption isotherm models evaluated were Langmuir (eq 1), BiLangmuir (eq 2) and Freundlich (eq 3) where q* is the concentration in the stationary phase at equilibrium with concentration C, and C is the concentration in the mobile phase.

Grafting of Polymer from Azo-Modified Silica. The grafting was performed in specially designed tubes containing 1 g of azomodified silica particles (Si-ACPA) suspended in a mixture containing L-PA (8.0 mg (d = 1 nm), 14.0 mg (d = 2 nm) or 18.9 mg (d = 3 nm)), MAA (23.0 mg (d = 1 nm), 41.0 mg (d = 2 nm) or 54.1 mg (d = 3 nm)) and EDMA (264.5 mg, (d = 1 nm) 472.2 mg (d = 2 nm) or 623.1 mg (d = 3 nm)) in different volumes (0, 5, 10, 15, 20 mL) of dry toluene. These quantities correspond to a molar ratio L-PA/MAA/ EDMA of 1/8/40. After purging the mixture with nitrogen, polymerization was initiated by UV-irradiation at 15 °C for 24 h. After polymerization, the suspended particles were filtered through a sintered glass funnel and extensively washed with methanol/formic acid/water 80/15/5 (v/v/v) and pure methanol. The composites were finally dried in vacuum oven at 40 °C for 12 h. Grafting of Polymer from RAFT-Modified Silica. RAFT modified silica particles (Si-RAFT) (1 g) were suspended in a prepolymerization mixture containing L-PA (28.9 mg, 0.12 mmol), MAA (0.081 mL, 0.96 mmol), and EDMA (0.905 mL, 4.8 mmol) dissolved in 20 mL of dry toluene. The polymerization mixture was subjected to three freeze−thaw cycles under nitrogen whereafter the initiator ABDV (115 mg, (SiPR) and 29 mg, (SiPRE)) was added. This corresponds to a ratio of RAFT/initiator of 0.5 (SiPR) and 2 (SiPRE). Polymerization was initiated at 50 °C for 24 h. After polymerization the particles were filtered through a sintered glass funnel and washed with methanol/formic acid/water, 80:15:5(v/v/v) and pure methanol and then the polymer was dried under vacuum at 40 °C overnight. SiPR500, SiPRE500, and SiPREE500 were prepared in a similar manner, but the prepolymerization mixtured contained L-PA (5.2 mg, 0.022 mmol), MAA (0.015 mL, 0.173 mmol), and EDMA (0.163 mL, 0.865 mmol), and the RAFT/initiator ratio was adjusted to 0.3, 1.4, and 14, respectively. Silica Removal. Portions (1 g) of the composite materials were suspended in 10 mL of 3 M NH4HF2 (aq.) in Teflon flasks. The suspensions were shaken at room temperature for 2 days and then filtered through glass funnel. The resulting polymer was washed with water to remove unreacted NH4HF2 and dried in vacuum oven at 40 °C for 24 h. Evaluation of Binding Affinity and Selectivity in the Chromatographic Mode. The materials were slurry packed into stainless steel columns (33 × 4.6 mm SiPA, SiPR or 20 × 2 mm PA, PR) using MeOH/H2O 80:20 (v/v) as dispersing solvent. The composites prior to etching were packed at a maximum pressure of 200 bar using an air-driven fluid pump whereas the thin walled beads were packed under negative pressure using a vacuum pump. The flow rate was adjusted between 0.2 mL/min to 1 mL/min to achieve a similar linear velocity for all columns. 10 μL aliquots of 1 mM solutions of pure enantiomers or racemate were injected unless otherwise mentioned. The elution was monitored at 260 nm. The retention factors (kL and kD) and the separation factor (α) were calculated using the following formulas: kL = (tL − tO)/tO; kD = (tD − tO)/tO; a = kL/kD where tL is the retention time of the L-enantiomer, tD is the retention time of the D-enantiomer, and tO is the retention time of the void marker, acetone. Evaluation of Binding Affinity and Selectivity in the Static Mode. The L-PA imprinted composites (10 mg) and corresponding thin-walled materials (5 mg) were weighed into 2 mL rubbersealed vials. Solutions of D- or L-PA in acetonitrile (0.5 mL) made up to concentrations within the range (C = 0.05−5 mM) were added. After 24 h incubation at room temperature the supernatants were sampled (30 μL) and the aliquots diluted in 270 μL water and transferred to HPLC vials for measurement of unbound solute concentration by reversed phase HPLC. The HPLC system consisted of an Agilent HPLC 1100 series instrument (Agilent) equipped with a UV-DAD detector and an autosampler. The column was a reversed phase (C18) column (Phenomenex Luna C-18, 250 × 4.6 mm), the mobile phase: MeOH/H2O: 62/38 (0.2% TFA) and the detection performed by UV absorbance at 260 nm. The resulting peak areas were used to calculate the amount of bound analyte on the polymer (in μmol/g of polymer). Each data point is based on the average of at least two replicate measurements.

q* = qsbC /(1 + bC)

(1)

q* = qs1b1C /(1 + b1C) + qs2b2C /(1 + b2C)

(2)

q* = aC m

(3)

The Langmuir models (eqs 1−2) assume that one (eq 1) or two (eq 2) distinguishable classes of sites are present on the surface, each with saturation capacity qs and association constant b. The dissociation constant Kd was calculated as the inverse of b. The Freundlich isotherm (eq 3), on the other hand, assumes sites with a Gaussian distribution of binding strengths. Here the width of the Gaussian distribution describes the degree of heterogeneity, through the index m. Engelhardt Test of Hydrophobicity and Acidity. The Engelhardt test was applied to SiPA103, SiPR, SiPRE, SiPREE500 PA103, PR, PRE, PREE as well as a bulk polymer prepared in presence of RAFT (PRref). The materials were packed in stainless steel columns, and a commercially available C18 column (Phenomenex Luna C-18, 250 × 4.6 mm) was used as a reference. The properties of the columns are summarized in Supporting Information, Table S7. The Engelhard test was performed at room temperature, the mobile phase was MeOH/ H20 (55/45), and the flow rate was adjusted to achieve a similar linear velocity for all columns, that is, 1.0 mL/min for the C18, 0.5 mL/min for SiPR, SiPRE, SiPREE500 and 0.2 mL/min for SiPA, PA103, PR, PRE, PREE, and PRref. Thiourea was used as a void marker, the detector was set at 254 nm, and the test compounds were aniline (1 mg/mL), phenol (2 mg/mL), p-ethylaniline (2 mg/mL), N,N-dimethylaniline (0.4 mg/mL), ethylbenzoate (2 mg/mL), toluene (10 mg/mL), and ethylbenzene (10 mg/mL).



RESULTS AND DISCUSSIONS Materials Synthesis and Characterization. The concept of combining the “grafting from” technique with templated synthesis of materials was investigated using two different SIP protocols that we described in our previous report.24 The first (A) starts from silica containing an immobilized azoinitiator (Si-ACPA) which can be decomposed by either thermolysis or photolysis,28 whereas the second (R) relies on the use of a RAFT chain transfer agent immobilized to the silica surface via the R-group (Si-RAFT) (Figure 1, Supporting Information, Table S1).29 Grafting of polymer from the latter requires an external source of primary radicals which is provided by addition of a soluble initiator. The azoinitiator and RAFT modified silicas were synthesized according to previously reported protocols resulting in surface coverages of initiators in accordance with our previous investigations.24 Imprinted copolymers of methacrylic acid (MAA) and ethyleneglycol dimethacrylate (EDMA) were then grafted from the supports as shown in Figure 1, that is, in a 1:5 molar ratio of MAA to EDMA in presence of 5 mol % of Lphenylalanine anilide (L-PA) as chiral template and toluene as solvent. The quantity of monomer relative to the silica support Si-ACPA was adjusted as previously reported24 to result in films with approximately 1, 2, and 3 nm average thickness, and 5 different dilutions were investigated. As seen in Table 1 and Supporting Information, Figure S1−S2 results from characterization by FTIR, TGA, and elemental analysis all indicated successful formation of grafted films. In Figure 2 the apparent 2912

dx.doi.org/10.1021/cm300965t | Chem. Mater. 2012, 24, 2909−2919

Chemistry of Materials

Article

Table 1. Molecularly Imprinted Polymer Composites Prepared by Photoinitiated Grafting to High Monomer Conversion from Silica Modified with Azoinitiator

sample

a

SiPA01 SiPA02 SiPA03 SiPA51 SiPA52 SiPA53 SiPA101 SiPA102 SiPA103 SiPA151 SiPA152 SiPA153 SiPA201 SiPA202 SiPA203

volume toluene (mL)

nominal film thickness (nm)

%C

0 0 0 5 5 5 10 10 10 15 15 15 20 20 20

1 2 3 1 2 3 1 2 3 1 2 3 1 2 3

22.5 28.6 11.5 22.9 28.2 31.0 20.7 26.9 31.3 21.4 25.9 28.9 20.6 24.6 26.5

%N

mass loss (%)

conversionb (%)

2.56 2.15 3.23 2.40 2.08 1.72 2.52 2.16 1.72 2.49 2.24 1.82 2.52 2.33 1.98

36 45 19 36 44 51 33 43 53 34 42 49 33 40 46

79 76 3 82 72 75 68 70 83 70 66 68 64 57 60

a

The composites were prepared by photoinitiated grafting of a mixture of MAA and EDMA to high monomer conversion from silica modified with azoinitiator in presence of template L-PA. The molar ratio of LPA/MAA/EDMA was kept constant (1/8/40), and the total weight of monomer adjusted to obtain the nominal thicknesses as indicated. b Monomers converted into grafted polymer (%) calculated based on thermogravimetry assuming a contribution to the mass loss of 17% from the initiator modified silica Si-ACPA.

Figure 2. Thickness of grafted polymer films calculated based on the mass loss obtained by TGA. The composites were prepared by photoinitiated grafting to high monomer conversion from silica (Si100) modified with azoinitiator in presence of the indicated volumes of toluene (see Experimental Section). The amount of monomers was adjusted to obtain a nominal film thickness of 1 nm (blue bars), 2 nm (red bars), or 3 nm (green bars).

thickness, calculated from the TGA mass loss data, has been compared with the nominal thickness, estimated assuming the grafted film to consist of monomers forming a liquid film covering the inner pore walls of the material. Although the measured and nominal values correlated the former were systematically about 50% lower than the latter. This is likely the result of chains propagating and terminating in solution and thereby remaining non-attached to the support. Non-attached polymer could indeed be observed in the SEM images of the composites, most so for the nondilute systems (Figure 3 and Supporting Information, Figure S3). As seen in Table 1 (%C

Figure 3. Scanning electron micrographs of imprinted polymer composite prepared by RAFT mediated grafting corresponding to SiPR500 (A), azomodified grafting corresponding to SiPA203 (C) and SiPA201 (E), and the corresponding polymer beads PR500 (B); PA203 (D) and PA201 (F) after removal of the silica by etching.

and mass loss) the conversion of monomer to grafted polymer decreased with increasing dilution, and this had a clear impact on the pore system parameters resulting from nitrogen sorption 2913

dx.doi.org/10.1021/cm300965t | Chem. Mater. 2012, 24, 2909−2919

Chemistry of Materials

Article

which in one sense is similar to the above-discussed dilution effect. This is also supported by the SEM images of such composites, showing a complete absence of nonspherical fragments (Figure 3). After characterization of the thin film composites we turned to the etching step, that is, the removal of the silica supports by fluoride catalyzed etching. This procedure has been developed previously for hierarchical imprinting to generate surface confined binding sites for various target molecules. Here the pore system is filled with the monomer mixture, polymerized, and thereafter the silica is etched away by treatment with fluoride. This leaves behind a rigid inverse polymer replica of the silica with a narrow pore size distribution and high surface area similar to the original silica mold. In contrast to hierarchical imprinting, removal of the silica from the thin film composites would leave behind thin walled beads whose integrity should depend on the thickness of the grafted films (Figure 1). Viewed simplistically, such uncollapsed beads should exhibit twice the surface area of the original silica mold. After the etching step the composition of the resulting particles was characterized by several complementary techniques (Table 1, Supporting Information, Table S2−S3, Figure S1, S4, S6). In view of the nominal elemental composition based on the monomer ratio (60%C) it is clear from the values in Supporting Information, Table S2 and Table S3 that the silica had been effectively removed from the polymers. The removal of the silica was further supported by the results from IR spectral analysis, TGA, and EDX analysis (Supporting Information, Figure S8). The IR spectra (Supporting Information, Figure S1) showed disappearance of the silica backbone bands at near 1100 cm−1 and a spectrum which was essentially superimposable on the spectrum of a corresponding solution polymerized reference polymer. The TGA showed mass losses of the etched samples of about 95% of their weights which should correspond to the removal of about 90% of the silica mold (Supporting Information, Table S2, Table S3). Figure 3 and Supporting Information, Figures S3−S6 show scanning electron micrographs of the precursor composite beads as well as the particles resulting from silica etching. A closer study leads to the following general observations. Whereas the composite particles in all cases are spherically shaped with a size independent of the graft density, the etched particles are smaller and of a size which is strongly dependent on the nominal thickness of the grafted films (Figure 3 and Supporting Information, Figure S4). For instance etching of the composite beads carrying the thinnest grafts (1 nm) (Figure 3E; Supporting Information, Figure S3A,D) led to a particle size reduction from about 20 to 5 μm (Figure 3F; Supporting Information, Figure S4A,D) and strong particle aggregation. The thicker walled beads resulting from the composites with the thicker films, however, seemed to retain the size of the silica mold (Figure3C; Supporting Information, Figure S3C,F versus Figure 3D; Supporting Information, Figure S4C,F). The beaded particles produced by RAFT mediated grafting (d = 3.9 nm) from a wider pore size support (Figure 3A,B) confirmed these results showing a particle shrinkage of roughly 50% upon etching. The shrinkage of the particles appeared to be reversible since they showed pronounced swelling in the presence of solvent resulting in the swelling factors reported in Figure 5. This further showed that the thinner walled beads swelled more than

measurements (Figure 4). The increase in surface area and pore volume in the case of the composites with 3 nm thick films and

Figure 4. Pore system parameters from BET for the SiPA composites with films adjusted to 1 (blue bars), 2 (red bars), or 3 (green bars) nm apparent thickness prepared in presence of the indicated volumes of toluene. The excessive surface area measured for the 1 nm composite prepared at zero dilution is likely caused by extensive formation of porous nongrafted polymer.

the absence of this trend for the thinner film composites highlights the importance of dilute systems for producing homogeneous grafts with enhanced functional group accessibility.30 To use more concentrated monomer systems, while avoiding diffusion limitations and pore blocking, we have performed the grafting using RAFT modified supports (Table 2). The primary Table 2. Molecularly Imprinted Polymer Composites Prepared by Thermally Initiated Grafting to High Monomer Conversion from Silica Modified with RAFT Agent samplesa

support

RAFT/ ABDV

nominal thickness (nm)

conversion (%)b

SiPR c SiPRE c SiPR500 d SiPRE500 d SiPREE500 d

Si100 Si100 Si500 Si500 Si500

0.5 2.0 0.3 1.4 14

3.1 3.1 3.9 3.9 3.9

70 70 60 73 74

a

The composites (SiPR) were prepared by thermally initiated grafting to high monomer conversion from (c) Si-RAFT or (d) Si500-RAFT using different ratios of RAFT agent to azoinitiator (ABDV). These were etched with fluoride to yield thin walled polymers PR. b Monomers converted into grafted polymer (%) calculated based on thermogravimetry assuming a contribution to the mass loss from the initiator modified silicas as specified in Supporting Information, Table S1.

effect of the RAFT agent is to reduce the effective concentration of active radicals and thereby reducing the rate of polymerization. Apart from imparting more control over the grafting step it introduces living chain ends in the grafted films which subsequently can be used for additional grafting or postreactions. Grafting was in this case performed by tuning the amount of monomer, to generate thicker films with 3−4 nm average thickness, and the initiator content to achieve the molar ratios of RAFT/ABDV given in Table 2 and Supporting Information, Table S3. In spite of the reduced rate of polymerization the monomer conversion with respect to the grafted films at high RAFT/ ABDV ratios did not appear to be affected (the elemental compositions and mass losses of the composites were similar). Instead the RAFT control seems to enhance the conversion of monomers to grafted polymer and the quality of the films, 2914

dx.doi.org/10.1021/cm300965t | Chem. Mater. 2012, 24, 2909−2919

Chemistry of Materials

Article

Figure 5. (A) Surface area in the dry state of composites (blue bars) and corresponding polymers (red bars) resulting after fluoride catalyzed etching (see Table 1 and Supporting Information, Table S2 for assignements) and (B) the polymer swelling ratios in MeCN.

Figure 6. Retention factors (kL) for L-PA (A) and enantiomer separation factors (α) (B) obtained in the chromatographic mode after separate injection of stock solutions of D- and L-PA (10 μL of 0.5 mM) on composites with films adjusted to 1 (blue bars), 2 (red bars), or 3 (green bars) nm nominal thickness, prepared in presence of the indicated volumes of toluene. Mobile phase: MeCN/sodium acetate buffer 0.01M, pH 4.8:70/30 (v/v).

the thicker walled beads, resulting in swollen state diameters which approached the size of the composite precursor. The above results collectively suggest that the thicker walled beads should feature permanent porosity. This is clearly revealed viewing the dry state surface area of the materials. The polymers containing the thickest grafted films (Figure 2. Dilution with 10 mL; 2 and 3 nm films) prepared at intermediate dilution indeed featured much enhanced surface areas (Figure 5A). Particular attention should be paid to the 3 nm beads which display the highest surface area of all materials, even exceeding that of the corresponding composite. This implies that adjustment of the nominal film thickness of the composites allows tuning of the volume changes of the corresponding etched polymer beads. To the best of our knowledge this represents a new approach to engineer hard/ soft properties of network polymers which does not involve the cross-linking level or distribution of cross-links.31 Chromatographic Tests. To obtain an overview of the binding affinities, capacities, and mass transfer properties of the various materials they were packed in small columns and investigated by liquid chromatography for their ability to retain L-PA and its optical antipode D-PA using MeCN as mobile phase. An initial comparison of the SiPA series of composites revealed a strong influence of the monomer dilution and film thickness on the chromatographic retentivity and selectivity. Hence, increasing dilution and nominal film thickness led to a general improvement of the chromatographic properties manifested in an increased retention factor and separation factor. Figure 6 shows the corresponding results for the composites reported in Table 1 and Supporting Information, Table S2. Composites prepared from concentrated monomer solutions feature much lower retention factors and enantioselectivities compared to the composites prepared in the dilute system, this in spite of the higher carbon content of the former. We have attributed this to the faster polymerization occurring in the more concentrated system.24 This rapidly leads to locally high viscosities and blocking of pores, hence the low chromatographic retention factors. This contrasts with grafting under more dilute conditions where monomers can diffuse more freely and access the inner pore system of the material. The trend of increasing pore volume and surface area with increasing dilution offers support for this explanation (Figure 4). In view of the above results we decided to exclusively include materials prepared under dilutions to 10, 15, or 20 mL of toluene, for a more detailed chromatographic investigation and in the subsequent binding tests. Figure 7 shows a comparison of

the sample load dependence of retentivity and enantioselectivity for the composites and thin walled materials prepared using 15 mL of toluene as diluent. First of all we noticed that the column back pressure was similar for all materials indicating that the polymers, in spite of their different morphologies, remained robust and were not significantly compressed under the chromatographic conditions. All columns display a trend of decreasing retention and enantioselectivity with sample load in agreement with previous reports. This is the result of nonlinear chromatography with constant overloading of low-abundant high energy binding sites. Whereas the shape of the curves for the 1 nm film composites and thin walled polymers are nearly superimposable, the thicker film materials behave differently. Removal of the silica support here give rise to beads displaying much enhanced retentivity and, for the 3 nm films, enhanced enantioselectivity. In principle this behavior could have a simple explanation. The loss of the silica ballast will lead to a MIP presenting a higher functional capacity and density of imprinted sites. If this would be the sole reason for the enhanced retentivity however, all materials regardless of wall thickness should display this behavior. The result is more likely related to the enhanced binding site accessibility resulting from silica removal, an effect which should be more marked for the thicker films where accessibility is poor from the start. Indeed, the beads (50 nm pore size) featuring thick films (d = 3.9 nm) prepared by RAFT mediated grafting display even larger changes in retentivity post silica removal (Supporting Information, Figure S9). Removal of silica opens up a new pore system providing new diffusional paths for the solutes to reach the MIP sites. The mass transfer in the beads is reflected in the number of theoretical plates and peak asymmetry (Figure 8 and Supporting Information, Figure S10). The etched materials feature more than 2× higher plate numbers and lower peak asymmetry factors compared to the materials prior to etching, all in all supporting the above assumptions. Saturation Binding Experiments. Insights into the relative binding energies and abundance of imprinted sites were obtained by equilibrium32 partitioning experiments using acetonitrile as solvent. Thus, quantifying the equilibrium free concentration of solute (Cfree) by HPLC, the bound amount q could be determined and plots of q versus Cfree giving the binding curves of the template L-PA and of its optical antipode D-PA for the different imprinted polymer complements. In 2915

dx.doi.org/10.1021/cm300965t | Chem. Mater. 2012, 24, 2909−2919

Chemistry of Materials

Article

Figure 7. Plot of retention factors (k) (A−C) and enantiomer separation factors (D−F) versus sample load per column cross section area (n/A) for L-PA (filled symbols) and D-PA (open symbols) on columns packed with composite beads (blue triangles) and beads after etching (red squares) for SiPA151 and PA151 (A and D); SiPA152 and PA152 (B and E) and SiPA153 and PA153 (C and F). Mobile phase: MeCN. The nominal film thickness (d) has been indicated.

discontinuous and slightly sigmoidal in the low concentration regime with a higher final saturation capacity (Figure 10). The isotherms were subsequently fitted to mono-Langmuir, biLangmuir, and Freundlich isotherm models33 resulting in the isotherm parameters given in Supporting Information, Tables S4−S6. The Fisher values in Table 3 reflect which of the models provides the best fit to a particular isotherm, a higher number indicating a better fit. Binding Site Homogeneity. Giving the sigmoidal isotherm shapes, the materials prepared by RAFT polymerization (PR, PRE, and PRref) are either poorly fitted by all models, or are best fitted by the FI model (PRref). The non-RAFT beads behave differently. Here, the bulk reference polymer (PAref) shows a typical MIP binding curve indicating a heterogeneous binding site distribution. This isotherm is best fitted with the Freundlich isotherm model. In contrast to this, the binding sites of the corresponding composite and thin walled beads PA 20 2 (Supporting Information, Figure S11A) and PA203 (Figure 9A) appear homogeneous with a smooth saturation behavior and a high F-value for the mono-Langmuir fitting. In search for an explanation for this behavior two effects should be considered. As invoked in the previous section, silica removal will open up new pores leading to enhanced access to the imprinted sites from both sides of the wall. Hence, this will remove the influence of the silica−polymer interface as a source of hetereogeneity (different microenvironments, chain stiffness, silica surface interactions, etc.). Second, the increased dilution

Figure 8. Number of theoretical plates (A) and peak asymmetry factors (B) for D-PA (blue bars) and L-PA (red bars) at comparable sample loads (30 nmol/cm2) on columns packed with the indicated materials. Mobile phase: MeCN/sodium acetate buffer (0.01M), pH 4.8: 70/30 (v/v).

addition to the beaded polymers, conventional crushed MIP monoliths (PAref and PRref), prepared by solution polymerization of an identical monomer composition, were included as controls. As seen in Figure 9 the shape of the isotherms depended strongly on whether the polymers had been prepared by RAFT mediated grafting or not. The non-RAFT composites prepared at higher dilution and the corresponding thin walled polymers featured a clear saturation behavior whereas the shape of the isotherms obtained using the RAFT materials appeared 2916

dx.doi.org/10.1021/cm300965t | Chem. Mater. 2012, 24, 2909−2919

Chemistry of Materials

Article

Figure 9. Equilibrium binding isotherms of D- (open symbols) and L-PA (solid symbols) (A) for composite beads SiPA203 (red triangles) and the resulting beads after etching PA203 (blue squares); (B) a bulk reference polymer PAref; (C) composite beads SiPR (red triangles), the resulting beads after etching PR (blue squares) and PRE (green circles); (D) a bulk RAFT reference polymer PRref. Solvent: MeCN.

results in a slower rate of propagation giving the monomers time to diffuse into the pores prior to reaction. This per se should result in more homogenously grafted films. Which of these two effects is dominating appears by considering the results of two additional materials: the precursor composite to PA203 (SiPA203) and the thin walled materials prepared from more concentrated monomer solutions (PA102 and PA103). In contrast to the materials prepared from more dilute monomer solutions (PA202 and PA203) the binding curves corresponding to PA102 and PA103 are more shallow and are better fitted with the biLangmuir or Freundlich isotherm models. Studying the influence of wall thickness, the binding sites in the thicker walled beads appeared more uniform regardless of the dilution (Figure 9A, Supporting Information, Figure S11C) and exhibited furthermore a pronounced enantioselectivity. On the other hand, the thinner walled beads showed no enantioselectivity in the static binding test. This agrees with our previous finding of a critical thickness for enantiomeric discrimination. Collectively, these results offer sufficient evidence to explain the origin of the enhanced homogeneity displayed by PA202 and especially PA203. As seen in Figure 9 and in Table 3, also the isotherm corresponding to SiPA203 is best fitted by a monoLangmuir isotherm model. The fact that the silica precursor displays a uniform site distribution implies that monomer dilution is the decisive parameter controlling the film homogeneity. Hence a careful tuning of dilution can remove the main source of binding site heterogeneity in MIPs, leading to a material displaying a perfectly uniform distribution of binding sites. This is in agreement with the chromatographic test results shown in Figure 7 and stresses the importance of this key parameter for the formation of discriminative sites. Binding Capacity. The saturation capacities of all thin walled materials clearly exceeded those of the corresponding composites. This is obviously a consequence of removing the silica ballast which contributes weight only and should not contribute to binding especially when its surface is covered with

Figure 10. Adsorption saturation capacities (qs) for D-PA (red bars), L-PA (blue bars) obtained for the materials in Supporting Information, Tables S4−S6 estimated from mono-Langmuir curve fitting (PA202, PA203), bi-Langmuir curve fitting (PAref), or estimated as the q-value at the highest concentration of L-PA (PR, PRE, and PRref). The dashed line represents the theoretical saturation capacity assuming a quantitative yield of imprinted sites.

Table 3. Fischer Values Obtained by Fitting the L-PA Binding Curves in Figure 9 and Supporting Information, Figure S11 to mono-Langmuir (mono-LI), bi-Langmuir (biLI), or Freundlich isotherm models (FI)a

a

polymer code

mono-Langmuir

bi-Langmuir

Freundlich

PR PRE PA202 PA203 PA102 PA103 SiPA203 PAref PRref

1382 356 1135 3415 732 458 897 341 328

1306 88 535 1138 1215 344 320 683 746

1567 225 362 215 1115 386 183 942 1208

See Supporting Information, Tables S4−S6.

2917

dx.doi.org/10.1021/cm300965t | Chem. Mater. 2012, 24, 2909−2919

Chemistry of Materials

Article

a homogeneous polymer film. However the capacity increase, especially for the RAFT grafted composites, is higher than this theoretical value (for the RAFT materials: expected about 2.7×, found about 4×) and hence they also exceeded those of the bulk reference materials PAref and PRref (Figure 10). Particularly noteworthy is the capacity shown by the MIP prepared by RAFT polymerization using a high RAFT/initiator ratio. This attains a Q-value of nearly 120 μmol/g, which is close to the nominal capacity of the material, that is, the capacity for a MIP where every template molecule has given rise to an accessible binding site. The enantioselective contribution to binding here amounts to nearly 30 μmol/g. Surface Acidity and Hydrophobicity. In view of the higher capacity and selectivity of the thin walled MIPs compared to the bulk monoliths and the composites, it is tempting to interpret this result as being due to an increase in surface area after silica etching. This would allow access to the imprinted film from a pore newly created by the silica removal in addition to the original pore defined by the silica mesopore system. One question that arises concerns the nature of the interfaces post silica removal. After grafting of the polymer films, capped end groups should reside near the pore walls and could serve as reactive groups for continued grafting. What about the new pore wall? The silica polymer interface is covalently anchored and therefore the fluoride etching should create new end groups on the emerging wall. To demonstrate with other means that the resulting pore walls are different we turned to the so-called Engelhardt test. This is commonly used in chromatography for characterizing reversed phase columns with respect to their surface hydrophobicity.34 The assumption is here that the RAFT grafted composites will leave behind a more hydrophobic surface because of the capping of chains by dithioester groups as schematically drawn in Figure 1.35 We therefore characterized the packed columns with respect to retention and separation of a test mixture of solutes with different basicity and hydrophobicity. This would be informative of the abundance of accessible COOH groups as well as the general hydropobicity of the materials. First it became obvious that all the MIPs retained the basic test solutes more than the reversed phase reference column: this agrees with the presence of COOH groups in the backbone (Supporting Information, Figure S12). Figure 11 otherwise supports the view discussed above. Whereas the RAFT composites exhibit a rather similar hydrophobicity as the C18 column, this character dissapears upon the HF treatment. Obviously this may be the result of a chemical hydrophilization induced by the strong hydrolytic treatment, but the absence of this effect when the treatment was performed on a bulk MIP prepared by RAFT (PRref versus

PRrefH) shows that it has another origin.35 More likely here is that the enhanced hydrophilicity is due to the newly created pore system after silica removal which, in view of the lower abundance of dithioester moieties, is more hydrophilic (Scheme 1). Scheme 1



CONCLUSIONS Thin molecularly imprinted polymer (MIP) films were grafted from porous silica using immobilized azoinitiators or RAFT chain transfer agents. Removing the silica supports from the above composites by etching, led to nanometer thin walled beads with structure, morphology, and recognition properties strongly depending on the film thickness of the original composite. The resulting polymer beads prepared using immobilized azoinitiators showed a perfectly uniform distribution of binding sites whereas those prepared by controlled grafting showed a record high saturation capacity. Hence, the method of combining templated synthesis, surface initiated polymerization, and controlled polymerization techniques successfully addresses the main deficiencies of previous imprinting techniques in terms of imprinting efficiency, binding capacity, and a binding site heterogeneity. In forthcoming publications we will explore the living properties of the grafting allowing an improved control of the surface properties of the two pore systems. This suggests a new concept for the engineering of nanostructured materials for dedicated separations, catalysis, or transport.



ASSOCIATED CONTENT

S Supporting Information *

Further details are given in Tables S1−S7 and Figures S1−S16. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors wish to acknowledge the financial support from the European Commission under FP7 − Marie Curie Actions, contract PITN-GA-2008-214226 [NEMOPUR] and from the Deutsche Forschungsgemeinschaft DFG (Se777/5-2).

Figure 11. Hydrophobicity index, estimated as the ratio of retention factors of ethylbenzene to toluene, of columns packed with the indicated materials. Mobile phase: MeOH/H20: 55/45 (v/v). 2918

dx.doi.org/10.1021/cm300965t | Chem. Mater. 2012, 24, 2909−2919

Chemistry of Materials



Article

similar in all networks independent of wall thickness. The relative intensity is also comparable to conventional solution polymerized MIPs which typically display 20−30% unreacted double bonds. (32) The uptake of L-PA versus time (Supporting Information, Figure S16) shows that the system has reached equilibrium after 24 h of incubation. (33) Shimizu, K. D. Binding isotherms. In Molecularly Imprinted Materials: Science and Technology, Yan, M., Ramström, O., Eds.; Marcel Dekker: New York, 2005; pp 419−434. (34) Engelhardt, H.; Jungheim, M. Chromatographia 1990, 29, 59− 68. (35) The dithioester RAFT group was stable under the fluoride etching conditions as shown by the retention of the characterstic pink color of the composite beads and of a soluble RAFT agent (Supporting Information, Figure S13−S15). The stability of the soluble RAFT agent was further verified by 13C NMR (Supporting Information, Figure S15). (36) Gregory, A. M.; Thurecht, K. J.; Howdle, S. M. Macromolecules 2008, 41, 1215−1222.

REFERENCES

(1) Jaroniec, M.; Schüth, F., Eds.; Special Issue: Templated Materials. Chem. Mater. 2008, 20, (3), 599−600. (2) Pyun, J.; Matyjaszewski, K. Chem. Mater. 2001, 13, 3446−3448. (3) Barbey, R.; Lavanant, L.; Paripovic, D.; Schuewer, N.; Sugnaux, C.; Tugulu, S.; Klok, H.-A. Chem. Rev. 2009, 109 (11), 5437−5527. (4) Johnson, S. A.; Ollivier, P. J.; Mallouk, T. E. Science 1999, 283, 963−965. (5) Prucker, O.; Rühe, J. Macromolecules 1998, 31, 592−601. (6) Hentze, H.-P.; Antonietti, M. Curr. Opin. Solid State Mater. Sci. 2001, 5, 343−353. (7) Titirici, M. M.; Hall, A. H.; Sellergren, B. Chem. Mater. 2002, 14, 21−23. (8) Alexander, C.; Andersson, H. S.; Andersson, L. I.; Ansell, R. J.; Kirsch, N.; Nicholls, I. A.; O’Mahony, J.; Whitcombe, M. J. J. Mol. Recogn. 2006, 19 (2), 106−180. (9) Ye, L.; Mosbach, K. Chem. Mater. 2008, 20 (3), 859−868. (10) Sellergren, B.; Hall, A. J. Molecularly Imprinted Polymers. In Supramolecular Chemistry: from Molecules to Nanomaterials; Steed, J. W., Gale, P. A., Eds.; John Wiley & Sons Ltd: Chichester, U.K., 2012; pp 3255−3282. (11) Piletsky, S. A.; Matuschewski, H.; Schedler, U.; Wilpert, A.; Piletska, E. V.; Thiele, T. A.; Ulbricht, M. Macromolecules 2000, 33 (8), 3092−3098. (12) Sulitzky, C.; Rückert, B.; Hall, A. J.; Lanza, F.; Unger, K.; Sellergren, B., G Macromolecules 2002, 35, 79−91. (13) Wei, X.; Li, X.; Husson, S. M. Biomacromolecules 2005, 6, 1113− 1121. (14) Li, Y.; Yang, H.-H.; You, Q.-H.; Zhuang, Z.-X.; Wang, X.-R. Anal. Chem. 2006, 78 (1), 317−320. (15) Wang, H.-J.; Zhou, W.-H.; Yin, X.-F.; Zhuang, Z.-X.; Yang, H.H.; Wang, X.-R. J. Am. Chem. Soc. 2006, 128 (50), 15954−15955. (16) Gao, D.; Zhang, Z.; Wu, M.; Xie, C.; Guan, G.; Wang, D. J. Am. Chem. Soc. 2007, 129 (25), 7859−66. (17) Wei, X.; Husson, S. M. Ind. Eng. Chem. Res. 2007, 46 (7), 2117− 2124. (18) Perez-Moral, N.; Mayes, A. G. Macromol. Rapid Commun. 2007, 28 (22), 2170−2175. (19) Lu, C.-H.; Zhou, W.-H.; Han, B.; Yang, H.-H.; Chen, X.; Wang, X.-R. Anal. Chem. 2007, 79 (14), 5457−5461. (20) Gonzato, C.; Courty, M.; Pasetto, P.; Haupt, K. Adv. Funct. Mater. 2011, 21 (20), 3947−3953. (21) Sellergren, B.; Ruckert, B.; Hall, A. J. Adv. Mater. 2002, 14 (17), 1204−1208. (22) Rückert, B.; Hall, A. J.; Sellergren, B. J. Mater. Chem. 2002, 12, 2275−2280. (23) Titirici, M.-M.; Sellergren, B. Chem. Mater. 2006, 18 (7), 1773− 1779. (24) Halhalli, M. R.; Aureliano, C. S. A.; Schillinger, E.; Sulitzky, C.; Titirici, M. M.; Sellergren, B. Polym. Chem. 2012, 3, 1033−1042. (25) Mandal, T. K.; Fleming, M. S.; Walt, D. S. Chem. Mater. 2000, 12, 3481−3487. (26) Mulvihill, M. J.; Rupert, B. L.; He, R.; Hochbaum, A.; Arnold, J.; Yang, P. J. Am. Chem. Soc. 2005, 127 (46), 16040−16041. (27) Chen, Y.; Kele, M.; Quinones, I.; Sellergren, B.; Guiochon, G. J. Chromatogr. 2001, 927, 1−17. (28) Prucker, O.; Rühe, J. Macromolecules 1998, 31, 602−613. (29) Li, C.; Han, J.; Ryu, C. Y.; Benicewicz, B. C. Macromolecules 2006, 39 (9), 3175−3183. (30) The enhanced binding site accessibility of the composite prepared under more dilute conditions was confirmed by coupling of a fluorescent probe (3-aminoquinoline) to the accessible carboxylic acid groups of the polymers. The resulting materials were characterised by fluorescence microscopy and elemental analysis (Supporting Information, Figure S7). (31) The FTIR spectra in Supporting Information, Figure S1 are informative of remaining unsaturations in the networks after etching. These feature a CC stretching vibration at 1638 cm−1 originating from pendent double bonds. The relative intensity of this signal is 2919

dx.doi.org/10.1021/cm300965t | Chem. Mater. 2012, 24, 2909−2919