Elsevier

Journal of Membrane Science

Volume 572, 15 February 2019, Pages 632-640
Journal of Membrane Science

Growing to shrink: Nano-tunable polystyrene brushes inside 5 nm mesopores

https://doi.org/10.1016/j.memsci.2018.11.058Get rights and content

Highlights

  • Nano-tuning of 5-nm γ-alumina pores by controlled brush growth is demonstrated.

  • Surface-initiated ATRP has been conducted in the most confined geometry to-date.

  • Pore brush growth and external brush growth have competing reaction kinetics.

  • Toluene permeability of 2.0 L m−2 h−1 bar−1 and retention of 330 Da are achieved.

Abstract

The development of controlled polymerization techniques in the last decade has enabled polymer brushes to be grown from inorganic substrates with precision. Less studied are brushes grown from concave geometries of high curvature, such as mesopores, despite their application potential in the separation sciences. The method used here, surface-initiated, activators-regenerated-by-electron-transfer, atom-transfer radical polymerization (SI-ARGET-ATRP), is used to grow a polystyrene brush grown from aluminum oxide pores of 5 nm diameter, to-date the most confined geometry in which ATRP has been conducted. The brush is characterized by TGA, AFM, and FTIR, the latter two methods applied specifically to the external brush. Additionally, permporometry as well as permeability and retention measurements are used to characterize the graft within the mesopores. We show that the brush length is tunable, that the brush length is solvent-dependent, and we also demonstrate the application potential of this hybrid material as an organic solvent nanofiltration membrane. This new class of membranes shows excellent performance: a toluene permeability of 2.0 L m−2 h−1 bar−1 accompanied by a 90% rejection of diphenylanthracene (MW 330 g mol−1).

Introduction

The union of inorganic and organic substances to realize hybrid materials of designed functionalities continues to attract interest from both industry and academia [1]. The application of polymer brushes on inorganic surfaces has proven a popular way to modify substrate surface properties. As scientists and engineers have developed finer control over the chemical composition, thickness, and architecture of polymer brushes, the number of possibilities and applications have grown. Highly specific hybrid materials have been created, examples include thermo-responsive, fluorescent nanocrystals and light-activated gates on the surface of mesoporous, hybrid-silica films [2], [3].

To create polymer brushes from inorganic substrates, one of two approaches can be applied. The first is to polymerize the brush from a surface-bound initiator molecule, this is called grafting-from. The other approach, termed grafting-to, is the binding of a pre-fabricated chain onto the target surface, often done with the help of an inorganic-organic linker already deposited on this surface. One of the main advantages of grafting-from over grafting-to is that there is no separate polymer synthesis step prior to grafting, decreasing the number of synthesis steps or dependence on a specialized commercial product. Another advantage are higher graft densities. This is due to the hindered diffusion of whole polymer chains to surface binding sites during grafting-to, while during grafting-from the monomer is relatively free to diffuse to active surface sites [1], [4], [5]. Before the advent of controlled polymerization methods, the separate synthesis step required by the grafting-to method meant greater control over the brush height and uniformity than could be achieved by a single grafting-from reaction. Nowadays, with controlled polymerization methods such as atom-transfer radical polymerization (ATRP), researchers are making dense, precisely-tuned brushes by grafting-from [6], [7].

The basis of ATRP lies in a high ratio of dormant chains to active chains, as seen in Scheme 1, where the equilibrium favors the left-hand side. This equilibrium requires the strict absence of oxygen, which if present will oxidize the metal complex. The benefits of this equilibrium are not only to slow down the reaction, resulting in lower, controllable molecular weights, but also to reduce the number of chain–chain terminations. However, inter-chain terminations still take place, removing the catalyst from the equilibrium cycle, as seen in Scheme 2. To remedy this, a reducing agent is introduced to renew the catalyst. The reducing agent also lowers the necessary catalyst concentration to ppm levels, as well as scavenging any oxygen left within the reaction mixture [8], [9]. This method, activators-regenerated-by-electron-transfer ATRP, or ARGET ATRP, allows for less specialized equipment than regular ATRP and enables grafting-from on larger surfaces, as performed in this work.

ATRP–grown brushes on inorganic substrates are not new. For instance, PMMA was grown on the outside of imogolite cylinders [10] and patterning techniques have been developed for the selective growth of polymer on silicon wafers [11]. However, relative to grafting-from on planar to convex geometries, such as a wafer or nanoparticle, there exist only few examples of brush polymerization from concave surfaces, such as the inside of a pore, and even fewer of those are in the mesoporous (IUPAC: 2 < ϕ ≤ 50 nm, ϕ = pore diameter) range [12], [13], [14]. As the radius of curvature becomes more extreme, the growing brush becomes increasingly confined. Chu et al.[15] grew poly(3-(N-2-methacryloyloxyethyl-N,N-dimethyl) ammoniatopropanesulfonate) (PMAPS) brushes via ATRP from an initiator bound onto the surface of 200 nm diameter pores of anodic aluminum oxide. The study claims no impact on chain growth kinetics due to the concave geometry, though the retardation of monomer to chains growing in the center of the pores was noted. ARGET-ATRP was used by Cao et al. to grow poly(N-isopropylacrylamide) (PNIPAAm) from silica mesopores of 29 nm diameter [16]. Until now, the smallest concave surface successfully polymerized-from without eliminating the pore volume was the inside of mesoporous silica particles with a pore diameter of 15 nm [5]. The brush height could be varied within 1–2 nm, with a minimum attainable pore diameter of 6 nm. This was demonstrated with 3 different polymer brushes, polymethacrylate (PMMA), polyacrylonitrile (PAN) and polystyrene (PS), all on mesoporous (ϕ 15 nm) silica particles dispersed in solution [17], [18].

Here, we have grown brushes inside the smaller pores of 5 nm diameter of gamma aluminum oxide (γ-alumina, γ-Al2O3) particles and membranes. We demonstrate tunable pore shrinking by the controlled growth of polystyrene from surface-immobilized initiators. Styrene is polymerized via ARGET-ATRP, and we show that the reaction kinetics are affected not only by the reaction time and monomer concentration, but are also faster outside the confined, concave pore geometry. The pore diameter can be shrunk down to values below 2.0 nm, though this pore size is shown to be dependent on the solvent the polystyrene brush is in contact with. Emphasized is that the term “pore size” refers here to the openings created by the rapidly moving grafted polymer chains in the pores, which can be physically interpreted as the average diameter of the free volume elements represented as a cylinder [19]. The mesoporous systems developed here by applying polymer brushes are used to change the surface/pore property of the ceramic alumina. Possible industrial applications of mesoporous brushes include catalysis, as in [17], [18]. However here this new hybrid material is used as a membrane, a selectively barrier separating solutes of 330 g mol−1 from a given solvent.

Accordingly, the majority of γ-alumina samples used in this publication are not free-floating hollow cylinder particles as the examples cited above [5], [17], [18], but a collection of pores calcined together to form a uniform and selective layer, or mesoporous membrane. These pores are not perfectly cylindrical but rather a stacking of alumina particulates, creating a tortuous path through the selective membrane layer. By tuning the polystyrene brush growth in these 5-nm pores we show the application potential of bespoke organic solvent nanofiltration (OSN) membranes.

The membrane technology field termed organic solvent nanofiltration (OSN) strives to purify or concentrate compounds between 200 and 1000 g mol−1 from organic solvents. Since its inception in the early 2000s, marked by its first commercial success, the “max-dewax” process [20], OSN has received growing attention from the chemical industry and academia alike [21]. A focal point is membrane material development, fabricating membranes that can withstand organic solvents and elevated temperatures while maximizing performance. The advantage of ceramic membranes over polymeric membranes are their inertness to most organic solvents, notably apolar solvents such as toluene, which often chemically degrade or physically deform polymeric membranes or their supports [21]. However, the highly hydrophilic surfaces of ceramic oxides makes them ill-suited to OSN applications, and as such the surface modification of porous ceramic oxides with apolar groups has found success, such as the grafting-to of PDMS [19] and the Grignard-grafting of alkanes or phenyl group [22]. However, grafting-from of the ceramic mesopores and its potential advantages - facile tunability of both the brush size and its chemical character in one synthesis step - remain unexplored until now [23]. The two key metrics for membrane performance, permeability and retention, are measured in this work by the permeability of toluene and the retention of 9,10-diphenylanthracene (DPA, MW = 330 g mol−1). From these results we demonstrate the promising application potential of grafted-from mesopores in membrane technology, specifically in the field of organic solvent nanofiltration.

Section snippets

Materials

The solvents toluene (anhydrous, 99.8%), anisole (anhydrous, 99.7%), isopropyl alcohol (99.5%), ethanol (absolute) and hexane (anhydrous, 95%); the chemicals copper(II) chloride (99.999%), 4,4′-dinonyl-2,2′-dipyridyl (dNbpy, 97%) chlorotrimethylsilane (TMCS, ≥99%), tin(II) 2-ethylhexanoate (Sn(EH)2, 92.5–100.0%) and 9,10-diphenylanthracene (DPA, 97%) were all purchased from Sigma-Aldrich and used as received. Styrene (≥ 99%, containing 50 ppm of 4-tert-butylcatechol inhibitor) was purified

Results and discussion

The same set of reactions was performed on two porous substrate geometries, γ-alumina discs and γ-alumina particles. The discs consist of 2 layers, on top is a thin selective layer of 3 μm-thick γ-alumina with pores of ϕ = 5.6 nm, as measured by permporometry, supported underneath by 2 mm of α-aluminum oxide with ϕ = 80 nm pores. The particles are porous particles entirely composed of γ-alumina. The average pore diameter of the ungrafted particles was calculated by the Brunauer–Emmett–Teller

Conclusion

For the first time, tunable brushes were grown from the mesopores (pore size 5.2 nm) of γ-alumina membranes. By means of ARGET-ATRP reactions, polystyrene brushes were grown on both particles and discs of γ-alumina. The different geometries caused different results: opposite to the discs, there was almost no brush growth inside the pores of the γ-alumina particles at the same reaction conditions. Inside the pores of the discs and for a given monomer to initiator ratio (400:1), the amount of

Acknowledgements

This work is part of the research program titled ‘Modular Functionalized Ceramic Nanofiltration Membranes’ (BL-20-10), which is taking place within the framework of the Institute for Sustainable Process Technology (ISPT, the Netherlands) and is jointly financed by the Netherlands Organization for Scientific Research (NWO, the Netherlands) and ISPT. The authors declare no competing financial interest. The authors would like to thank Dr. Lucie Grebíková for AFM imaging and Ing. Clemens Padberg

References (43)

  • C.R. Tanardi et al.

    Coupled-PDMS grafted mesoporous γ-alumina membranes for solvent nanofiltration

    Sep. Purif. Technol.

    (2016)
  • S.R. Hosseinabadi et al.

    Solvent-membrane-solute interactions in organic solvent nanofiltration (OSN) for Grignard functionalised ceramic membranes: explanation via Spiegler-Kedem theory

    J. Membr. Sci.

    (2016)
  • L.S. White

    Transport properties of a polyimide solvent resistant nanofiltration membrane

    J. Membr. Sci.

    (2002)
  • P. Silva et al.

    Solvent transport in organic solvent nanofiltration membranes

    J. Membr. Sci.

    (2005)
  • O. Azzaroni

    Polymer brushes here, there, and everywhere: recent advances in their practical applications and emerging opportunities in multiple research fields

    J. Polym. Sci. Part A Polym. Chem.

    (2012)
  • W. Wu et al.

    Thermo-responsive and fluorescent cellulose nanocrystals grafted with polymer brushes

    J. Mater. Chem. A

    (2015)
  • A. Brunsen et al.

    Light-activated gating and permselectivity in interfacial architectures combining “caged” polymer brushes and mesoporous thin films

    Chem. Commun.

    (2012)
  • K. Saha et al.

    Gold nanoparticles in chemical and biological sensing

    Chem. Rev.

    (2012)
  • J.J. Keating et al.

    Polymer brushes for membrane separations: a review

    ACS Appl. Mater. Interfaces

    (2016)
  • J. Pyun et al.

    Synthesis of polymer brushes using atom transfer radical polymerization

    Macromol. Rapid Commun.

    (2003)
  • K. Matyjaszewski et al.

    Grafting from surfaces for “everyone”: ARGET ATRP in the presence of air

    Langmuir

    (2007)
  • Cited by (16)

    • Development of a novel multi-scale structured superhydrophobic nanofiber membrane with enhanced thermal efficiency and high flux for membrane distillation

      2021, Desalination
      Citation Excerpt :

      The absorption peaks at 3000–3100 cm−1 are the CH aromatic symmetric stretching vibrations [43], whereas the peaks at 2870–3100 cm−1 also correspond to the CH symmetric stretching vibration and asymmetric stretching vibration of CH2 group [44]. Additionally, a new band could be observed at 695 cm−1 after the addition of PS which was assigned to the CH bend peak of mono-substituted aromatics [44,45]. The results demonstrated that PS was successfully introduced into the fibers.

    View all citing articles on Scopus
    View full text