The high degree of control exercised in biological channels to regulate key processes including ionic flow and molecular transport across cell membranes is the ultimate inspiration and key motivator for the material scientist to mimic in synthetic materials [1
]. To achieve this, precise control over the structure, functional density, and ideally local placement is required [2
Ordered mesoporous silicas (OMSs) are an ideal support structure due to their high specific surface area, dimensional stability, and their uniform, ordered channel structure with tuneable pore size and geometry [3
]. Furthermore, the rich surface hydroxyl groups available on the silica surface are a convenient handle for various synthetic transformations [3
]. These properties have afforded OMSs significant research interest over the last several decades, attracting attention for application in a broad range of fields from catalysis to controlled drug delivery [2
The incorporation of polymeric architectures offers the greatest amount of versatility and a high degree of functionality; e.g., a high density of chargeable groups can be incorporated into the porous structure [9
]. The incorporation of a polymer coating further permits the intelligent design of hybrid materials. For example, the careful selection of monomers can endow the hybrid material with sensitivity towards a number of external stimuli including pH, redox potential, temperature, solvent, salt concentration, and even light to create “smart” materials, changing their transport behaviour by altering the ion–pore wall interactions [2
Such responsive materials possess great potential for various applications [5
]: catalysis and nanoreactors [10
], energy storage [11
], molecular separation and filtration [12
], controlled drug delivery and release [15
], and in microfluidic devices from chemical sensing and biosensing [2
] to lab-on-chip devices (pre-concentration, enrichment/extraction for analysis, e.g., pesticides) [16
Polymerisation in spatially confined mesoporous materials is a key technology that has been widely investigated over the last decade with advances in both soft chemistry and controlled/“living” radical polymerisation (CLRP) techniques [7
]. A number of CLRP techniques have been successfully applied to the functionalisation of mesoporous silica including atom transfer radical polymerisation (ATRP) [19
], reversible addition‒fragmentation chain transfer (RAFT) polymerisation [22
], nitroxide-mediated radical polymerisation (NMP) [23
], and photoiniferter-mediated polymerisation (PIMP) [23
] through a grafting from or surface-initiated approach. While the successful grafting of pre-formed polymer to the interior and exterior surface has been reported in large-pore (15 nm) membranes [25
], this approach is preferred for gated membranes with sterics limiting the molecular weight, grafting density, and pore filling achievable [8
The formation of polymer brushes from the interior surface of mesopores through surface-initiated polymerisation, while synthetically more viable than post-grafting pre-formed polymers to the surface, remains synthetically challenging when compared to planar or spherical surfaces [17
]. This is particularly true for pores less than 10 nm in diameter [17
]. Polymer-functionalised pores below 10 nm generally become inaccessible, and the grafted polymer is generally multimodal with a broad molecular weight distribution [17
Surface-initiated atom transfer radical polymerisation (SI-ATRP) remains the most widely investigated technique due to its wide range of polymerisable monomers with controllable molecular weight and narrow dispersities, chemical tolerance, and mild polymerisation conditions [3
]. While several techniques such as initiators for continuous activator regeneration (ICAR) [3
] and activators (re)generated by electron transfer (A(R)GET) [10
] have been implemented to reduce the concentration of the copper catalyst that contaminates the final material, ATRP systems are more onerous, requiring numerous components to mediate a controlled polymerisation. This could severely limit the pore size that can be functionalised in terms of the control achievable and the polymer content.
Photoinitiated systems such as photoiniferter-mediated polymerisation (PIMP) are becoming increasingly popular in material functionalisation [18
]. PIMP is advantageous in its synthetic simplicity, requiring minimal reaction components and proceeding by the simple application of UV irradiation. PIMP is also versatile, being effectively applied in surface functionalisation with a number of functional monomers including styrene [29
], methyl methacrylate (MMA) [10
], 2-(dimethylamino)ethyl methacrylate (DMAEMA) [30
], [2-(methacryloyloxy)ethyl]trimethylammonium chloride (METAC) [24
], and zwitterionic monomers [19
]. Surface-initiated polymerisation (SIP) activated by UV irradiation also permits spatial and temporal control over the surface functionalisation to specify the location, degree of functionalisation, and even allowing surface patterning [18
]. For a more detailed overview of the various SIP techniques applied to functionalise surfaces and the impact of surface curvature, the reader is referred to a recent and comprehensive review by Klok and co-workers [18
While it is possible to functionalise small pores below 10 nm via SI-CLRP techniques, it is rather challenging to control the polymerisation kinetics, resulting in poorly accessible pores [17
]. Consequently, selective functionalisation of the external surface has taken precedence in many studies. For instance, no current reports have attempted the formation of block copolymers within mesoporous materials. Although several reports have incorporated block copolymers selectively to the exterior surface of mesoporous silica nanoparticles for gating and controlled drug release [15
]. To design and fabricate well-controlled uniform and responsive polymer coatings under the confines of mesopores below 10 nm, it is essential to understand how confinement impacts the polymerisation kinetics. Our research group previously reported an iniferter-initiated polymerisation of a zwitterionic monomer, carboxybetaine methacrylate (CBMA), from the interior and exterior surface of mesoporous silica with pores ranging from 2 to 115 nm in diameter, focussing on the effect of pore size, grafting density, and initiator on the SIP from iniferters that were post-grafted to the mesoporous silica surface through a terminal trimethoxysilane moiety [24
]. In this system, a certain degree of control over the polymer content was achieved; however, the pore filling ratio was limited under the conditions used, and the post-grafting strategy of the iniferter does not permit localised polymer functionalisation.
Herein, we report a simple one-step approach to functionalise mesoporous silica films with a responsive polymer through surface-initiated photoiniferter-mediated living radical polymerisation (SI-PIMP). We further explore the effect of confinement on the polymerisation kinetics of 2-(dimethylamino)ethyl methacrylate (DMAEMA), and its impact on the amount of polymer formed within the mesopores, including the use and effect of a sacrificial iniferter on film functionalisation, and how this impacts the wetting and ionic permselectivity behaviour of the hybrid materials. Our approach involves the incorporation of a photoiniferter, N
-(diethylamino)dithiocarbamoyl-benzyl(trimethoxy)silane) (SBDC), in a one-pot procedure through co-condensation with an inorganic precursor combining sol–gel chemistry and evaporation-induced self-assembly (EISA) to prepare a functional mesoporous thin film through dip-coating that is capable of undergoing polymerisation on exposure to UV light in the presence of a monomer as depicted in Scheme 1
. We will further demonstrate that the formation of block copolymers is possible under confinement.
2. Materials and Methods
All materials and solvents were purchased from Sigma-Aldrich (Darmstadt, Germany) and used as received unless stated otherwise. The impact of a sacrificial iniferter was examined by the addition of benzyl diethyldithiocarbamate (BDC), a structural mimic of N
-(diethylamino)dithiocarbamoyl-benzyl(trimethoxy)silane) (SBDC), which were synthesised according to a previous literature protocol [29
]. The monomer used to study the photoiniferter polymerisations from BDC and/or SBDC, 2-(dimethylamino)ethyl methacrylate (DMAEMA, 98%) was passed over neutral alumina immediately prior to use to remove the hydroquinone monomethyl ether (MEHQ) inhibitor, and the solvent N
-dimethylformamide (DMF, anhydrous, 99.8%) was used as received. For re-initiation studies, 2-(methacryloyloxy)ethyl phosphate (MEP) was used as received.
The precursor solutions for dip-coating mesoporous films were prepared using the inorganic precursor tetraethoxysilane (TEOS, Acros Organics, 98%, Geel, Belgium) alone or in combination with an organosilane such as SBDC, in dry tetrahydrofuran (THF, Merck, SeccoSolv, Darmstadt, Germany) or absolute ethanol (EtOH, Merck, ≥99.5%), with a hydrochloric acid catalyst (HCl, 37%) and Milli-Q water. Pluronic® F-127 (F127, BioReagent, Sigma-Aldrich, 13,800 g mol−1) was used as a structuring agent in the sol, which was used as received at a molar ratio of 0.0050 or 0.0075 relative to the silane precursor to control the porosity of the mesoporous films prepared. The microscope slides (VWR, glass, cut edges) were cleaned in a mixture of detergent and water prior to rinsing with ethanol and drying under ambient conditions. The silicon wafers (Si-Mat, Kaufering, Germany, 100 mm diameter, 525 ± 25 µm thickness, type P/Bor, <100> orientation, CZ growth method, 2–5 Ω resistivity, polished on 1 side) and indium tin oxide (ITO, Delta Technologies, Ltd., Loveland, CO, USA, polished float glass, 150 × 150 × 1.1 mm, SiO2 passivated/Indium Tin Oxide coated one surface, RS = 4–8 Ω, cut edges) were cut to an appropriate size using a diamond cutter, cleaned using technical grade ethanol, and dried under ambient conditions prior to dip-coating.
Ellipsometry measurements of mesoporous films prepared on silicon wafer substrates (Si-Mat, Kaufering, Germany) were obtained from at least three spots per film to determine the average film thickness and refractive index using a Nanofilm EP3 imaging ellipsometer equipped with a 658 nm laser. Measurements were recorded at a fixed humidity of 15% at room temperature, and the angle of incidence (AOI) was varied from 38° to 66° in one zone in 2° increments. The EP4 analysis software supplied with the instrument was used to calculate the film thickness and refractive index from the measured amplitude and phase difference, Ψ and Δ, respectively. The data was fit using a one-layer box model, assuming an oxide layer on the silicon wafer was present between the substrate and mesoporous film, and this was fixed at 2.2 nm from previous measurements and fitting. The film thickness was allowed to vary between 10 and 300 nm, with a refractive index between 1 and 1.7. The Brüggemann effective medium approximation was used to calculate the volume porosity from the fitted refractive index [32
]. Detailed calculations are provided in the Supplementary Materials (Section 5, Scheme S1)
2.2.2. Cyclic Voltammetry (CV)
The electrochemical and transport properties of the mesoporous materials was investigated via cyclic voltammetry using and as negative and positive probe molecules, respectively. Measurements were recorded using a Metrohm Autolab PGSTAT302N potentiostat (Metrohm, Utrecht, The Netherlands). Mesoporous films prepared on ITO substrates were characterised using a 2 mM solution of either the positive or negative probe molecule in a 100 mM KCl electrolyte solution. The pH-dependent permselectivity was investigated by adjusting the solution pH between 2 and 11 by the addition of either aqueous NaOH or HCl to the prepared solutions. The pH was determined using pH-Fix 0–14 indicator sticks (Laborbedarf, Article No. 0549, Carl Roth, Karlsruhe, Germany). A Ag/AgCl reference electrode (BASi RE-6) and graphite counter electrode was used in the sample cell. The electrode area was 0.21 cm2; however, all measurements reported here correspond to the peak currents in μA. Each pH was measured using a scan rate sequence of 200, 100, 25, 300, 1000, and 200 mV s‒1, with each scan rate cycled three times. All voltammograms shown here were measured at 100 mV s‒1 from the last cycle recorded, and bare ITO was measured as a reference on each measurement day for comparison.
2.2.3. Static Contact Angle (CA)
CA measurements of the prepared films were measured using a DataPhysics OCA35 instrument (DataPhysics Instruments GmbH, Filderstadt, Germany) and evaluated using SCA20 software (Version 4.5.15 build 1064, DataPhysics Instruments GmbH, Filderstadt, Germany, 1998–2013) through the sessile drop method. All measurements were conducted within a climate-controlled room with a temperature of 25 °C and a relative humidity of 50%. Contact angles were determined within 30 s of applying a 2 µL water droplet onto the surface at a rate of 2 µL s‒1. All values reported arise from averaging at least 3–5 measurements across each film surface, and are reported with a standard deviation. Before each film was measured, it was stored under the measurement conditions within the climate-controlled room for several hours or overnight.
2.2.4. Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy (ATR-FTIR)
Infrared spectra of the prepared mesoporous films on glass or ITO substrates were performed using a Perkin Elmer Instrument One Spectrum FT-IR Spectrometer equipped with a Universal ATR Polarization Accessory (Waltham, MA, USA). All spectra shown were averaged from at least 3 spots and normalised to the Si–O–Si band at ~1085 cm‒1. The spectra were recorded using the Spectrum Software (Version 10.5.4.738, PerkinElmer, Inc., Waltham, MA, USA, 2016) between 4000 and 650 cm−1 with a resolution of 4 cm−1, and a background correction was automatically applied. All further data corrections were performed in OriginPro9 (ADDITIVE Soft- und Hardware für Technik und Wissenschaft GmbH, Friedrichsdorf, Germany, 2012).
2.2.5. Transmission Electron Microscopy (TEM)
Electron micrographs were recorded using a Philips FEI CM-20 transmission electron microscope (Philips, Amsterdam, The Netherlands) equipped with a LAB-6 cathode and Olympus CCD camera, and with a maximum resolution of 2.3 Å operating at an accelerating voltage of 200 kV. Samples were prepared by scratching the mesoporous films off the substrate and dispersing in filtered ethanol with 10 min of sonication before being drop cast onto 3.05 mm Cu grids (mesh size 200) with a Lacey carbon film (Plano GmbH, article number S166-2). The samples were then dried under ambient conditions.
2.2.6. Scanning Electron Microscopy (SEM)
SEM spectra were obtained using a Philips XL30 FEG scanning electron microscope equipped with a tungsten cathode and a back scattered electron yttrium aluminium garnet (BSE YAG) detector with an accelerating voltage of 15–25 kV, a 30 µm aperture, and a spot size of 4–5. The samples were sputter-coated with a 5 nm coating of Pt/Pd. The digital micrographs were recorded over a range of magnifications at a working distance of approximately 10 nm using an SE2 detector.
2.2.7. 1H Nuclear Magnetic Resonance (1H NMR) Spectroscopy
1H NMR spectra were recorded on a 300 MHz Bruker AVANCE NMR spectrometer (Billerica, MA, USA) in deuterated chloroform (CDCl3, 99.8% D), methanol (MeOD-d4, 99.8% D) or dimethyl sulfoxide (DMSO-d6, 99.8% D) unless specified otherwise. The solvent residual peak was used as an internal reference.
2.2.8. Size Exclusion Chromatography (SEC)
SEC (Agilent Scientific Instruments, Santa Clara, CA, USA) was performed using a PSS-Agilent 1200 equipped with a 1200 Agilent RID and an autosampler (100 µL). The filtered samples were passed through a GRAM 10 µm pre-column (8 × 50 mm) and a PSS GRAM LINEAR 10 µm column (8 × 300 mm) using N,N-dimethylformamide (DMF) with 3 g L‒1 LiCl at a flow rate of 0.5 mL min‒1 at 25 °C. The molecular weight and molecular weight distribution was determined using WinGPC UniChrom software (Version 8.20 build 4815, PSS Polymer Standards Service GmbH, Mainz, Germay, 1992–2014) by conventional calibration with linear poly(methyl methacrylate) standards ranging between 2.2 × 102 and 1.6 × 106 g·mol‒1.
2.2.9. Krypton BET Adsorption
Krypton sorption and desorption measurements were performed at 77 K using an Autosorb iQ2 from Quantachrome. The samples were degassed for 12 h prior to measurement, and the BET surface area was calculated in the relative pressure range of 0.1 to 0.3.
2.3.1. Synthesis of N,N-(Diethylamino)dithiocarbamoyl-benzyl(trimethoxy)silane) (SBDC), a Silane Terminated Iniferter
-diethyldithiocarbamate trihydrate (STC) was dissolved in an excess of hot methanol to remove insoluble impurities. The solution was then dried over molecular sieves before being concentrated on a rotary evaporator and dried under high vacuum to remove any water or residual solvent. The purified sodium N
-diethyldithiocarbamate (1.51 g, 8.81 mmol) was dissolved in dry THF (10 mL) before being added dropwise via syringe to a solution of p
-(chloromethyl)-phenyltrimethoxysilane (1.93 mL, 8.76 mmol) in dry THF (10 mL) under a nitrogen atmosphere. The solution was then stirred at room temperature for 4 h before being filtered to remove the sodium chloride and dried under reduced pressure to remove the THF to yield crude SDBC (2) as a pale yellow liquid (76% crude yield). The crude product was then vacuum distilled at 160 °C using a Kugelrohr to yield pure SBDC as a yellow liquid (64% yield). 1H
NMR (300 MHz, CDCl3
, ppm): δ 7.53 (d, 2H, C6
), 7.35 (d, 2H, C6
), 4.49 (s, 2H, –CH2
S–), 3.98 (q, 2H, –NCH2
–), 3.66 (q, 2H, –NCH2
–), 3.55 (s, 9H, –Si(OCH3
), 1.22 (t, 6H, –NCH2
). MS (ES)+
(%) calcd. for C15
359, found 359 (100) [M∙
, 360 (23) [M + 1]+
, 361 (16) [M + 2]+
, 362 (4) [M + 3]+
, 363 (1) [M + 4]+
. Assigned NMR and MS spectra are provided in the Supplementary Materials (Figures S1 and S4)
2.3.2. Synthesis of Benzyl Diethyldithiocarbamate (BDC), a Sacrificial Iniferter and SBDC Mimic
-diethyldithiocarbamate (1.50 g, 8.77 mmol) was dissolved in dry THF (10 mL). This solution was then added dropwise via syringe to a solution of benzyl chloride (1105 μL, 9.62 mmol) in dry THF (10 mL) under a nitrogen atmosphere. The solution was then stirred for 3 h before being filtered to remove the sodium chloride. The filtrate was then dried under reduced pressure to remove the THF to yield crude BDC as a dark yellow-brown liquid (66% yield). The crude product was then vacuum distilled at 60–100 16C at 5 × 10−1
mbar using a Kugelrohr. 1H
NMR (300 MHz, CDCl3
, ppm): δ 7.33–7.16 (m, 5H, C6H5
), 4.47 (s, 2H, –SCH2
Ph), 3.97 (q, 2H, –NCH2
), 3.65 (q, 2H, –NCH2
), 1.21 (t, 6H, –NCH2
). MS (ES)+
(%) calcd. for C15
239, found 239 (100) [M∙
, 240 (16) [M + 1]+
, 241 (10) [M + 2]+
, 242 (1) [M + 3]+
. Assigned NMR and MS spectra are provided in the Supplementary Materials (Figures S2, S3 and S5)
2.3.3. Preparation of Mesoporous Silica Thin Films
The mesoporous silica films were prepared via sol–gel chemistry using tetraethoxysilane (TEOS) as an inorganic precursor either alone or in combination with a functional organosilane (SBDC) in order to perform a photo-initiated polymerisation. The sol contained an amphiphilic triblock copolymer, F127, which undergoes micellisation upon solvent evaporation resulting in the formation of a porous inorganic network. The precursor solutions were stirred for 24 h under ambient conditions before being used to prepare films through evaporation-induced self-assembly (EISA). Dip-coating was performed in a climate-controlled chamber at a temperature of 25 °C and a relative humidity of 50%, at a withdrawal speed of 2 mm s−1
unless stated otherwise. The precursor solution used to prepare mesoporous silica with 5 mol % SBDC (5SBDCSi) contained a final TEOS:SDBC:F127:THF:H2
O:HCl molar ratio of 0.95:0.05:0.0075:20:5.2:0.28, and these films were prepared at a reduced temperature of 15 °C. After aging the films for 24 h under climate-controlled conditions, the films were subjected to a stabilising thermal treatment: 60 °C for 24 h, 130 °C for 24 h, and then finally at 200 °C for 1–2 h. The F127 template was then removed by immersion into a 0.01 M HCl ethanoic solution for 3 days with stirring. Detailed optimisation of the film preparation conditions including the withdrawal speed, temperature, and relative humidity are provided in the Supplementary Materials (Figures S8 and S9)
, as well as the full structural characterisation of the final 5SBDCSi films used in this study (Figure S10
2.3.4. Surface-Initiated Photoiniferter-Mediated Polymerisation (SI-PIMP) from SBDC-Functionalised Mesoporous Silica
BDC (0.057 g, 0.24 mmol) and DMAEMA (1.0 mL, 5.9 mmol) were dissolved in DMF (8 mL), and the reaction mixture deoxygenated by nitrogen bubbling under light protection. The solution was then transferred via a deoxygenated syringe to a sealed flask containing the SBDC-functionalised mesoporous films on either glass, silicon wafer, or ITO substrates, which had been previously dried in vacuo, and deoxygenated using 3 vacuum–nitrogen cycles before being placed under a slight underpressure. The flask was then irradiated with UV light (320–400 nm) for various periods of time before exposure to air to quench the polymerisation. The free polymer formed was collected, dried under reduced pressure, and characterised by 1H
NMR spectroscopy and SEC to determine the conversion and molecular weight, respectively. The films were extracted in THF at least 3 times to remove any physisorbed material before drying under ambient conditions for further characterisation. Unless otherwise stated, the DMAEMA:BDC molar ratio was maintained at 25:1, with a monomer to solvent ratio of 1:8 (v
). Investigations in the absence of sacrificial iniferter were carried out in a similar fashion with an identical monomer concentration. A Lumatec UV-Technik Superlite 410 lamp was used as the light source in all photoiniferter polymerisations unless stated otherwise. The wavelength ranged from 320 to 400 nm (UVA) with an intensity of ~6 mW cm‒2
(see Supplementary Materials, Figure S6
). Each polymerisation was carried out in a specially designed Schlenk flask with rectangular sides, and positioned ~10 cm away from the light source in an aluminium foil-lined box.
2.3.5. CO2 Plasma Treatment
The organic functional groups present on the exterior surface of the mesoporous silica films were removed by applying a CO2
plasma treatment using a Diener Femto plasma cleaner (Diener electronic, Ebhausen, Germany) at a pressure of 0.4 mbar and power of 20% for 12 s according to a previously published protocol by Babu et al. [34
The synthetic ease of preparing organic‒inorganic hybrid materials via SI-PIMP from a mesoporous silica support to pore blocking in pores less than 10 nm in diameter was demonstrated. The successful incorporation of an iniferter moiety, SBDC, through co-condensation provides a simple one-pot method to prepare functional mesoporous materials capable of further functionalisation by SIP simply by exposure to appropriate monomers and UV light. Moreover, co-condensation offers the potential to locally functionalise mesoporous materials, which is of high-importance in achieving fine control over ionic permselectivity with potential applications in self-sustained sensing and separation devices, including lab-on-chip devices. We will elaborate on this aspect in future reports.
We further demonstrated that the SIP from the inner surface is diffusion-limited, with significantly less polymer incorporated when polymer is able to grow from both the exterior and interior surface. This makes PIMP more suited for application under confinement when compared to ATRP since fewer reaction components are required to access the initiating sites, allowing the polymer content to be controlled in pores below 10 nm; PIMP also proceeds in the absence of any transition metal catalyst that ultimately pollutes the final material.
By the incorporation of a dual responsive polymer, PDMAEMA, we illustrated that the permselectivity towards both positively and negatively-charged species can be selectively controlled and fine-tuned by varying the polymer content and solution pH. We further showed that the formation of block copolymers under confinement is possible, a feat which has not yet been reported. The use of a sacrificial iniferter under high dilution was also investigated; sacrificial iniferters offer the possibility to control the molecular weight of the polymer grafted from the exterior surface, and potentially the amount of polymer incorporated within the pores in a controlled fashion by providing a sufficient concentration of stable radicals to deactivate surface-fixed polymer after prolonged polymerisation times. The amount of polymer incorporated within the pores can also be tuned by varying the initial monomer concentration, and polymerisation time.