Elsevier

Journal of Power Sources

Volume 406, 1 December 2018, Pages 35-41
Journal of Power Sources

A highly selective proton exchange membrane with highly ordered, vertically aligned, and subnanosized 1D channels for redox flow batteries

https://doi.org/10.1016/j.jpowsour.2018.10.043Get rights and content

Highlights

  • A selective proton exchange membrane concept was proposed for flow batteries.

  • The concept was proofed by a silica membrane with subnanosized channels.

  • The present membrane offers high ion selectivity in flow batteries.

  • The battery assembled with the membrane shows improved performance.

Abstract

Among the most critical issues associated with redox flow batteries (RFBs), one is the undesirable crossover of redox-active species through the membrane, which is typically polymer-based. To address this issue, we report an inorganic mesoporous silica film based nanochannel membrane with highly ordered, vertically aligned, and subnanosized channels. Unlike conventional polymer-based membranes that have ion transport channels with fixed sizes between 2 and 4 nm in diameter, the ion transport channels of the membrane are controllable in size and uniform in size distribution, enabling the membrane to separate the ions that are larger than the critical value. Experimental results prove the expected effect that the pristine nanochannel membrane (with the channel size of ∼0.5 nm) grown on the ITO glass can completely separate protons and vanadium ions. After ITO etching, the free-standing membrane maintains extremely low vanadium-ion permeability that is four orders of magnitude lower than that of Nafion membranes, suggesting that it is a promising candidate for redox flow batteries. Moreover, the membrane fabrication is convenient, inexpensive, and does not rely on sophisticated facilities or conditions, providing potential in scalable productions.

Introduction

Storage of electricity harvested by solar panels and wind turbines in batteries provides an attractive solution to the issue associated with the mismatch between the intermittent supply of these renewable resources and variable demand [[1], [2], [3], [4]]. Redox flow batteries (RFBs), in which energy-bearing materials are stored in external tanks separated from power packs, are now enjoying a renaissance due to their advantage of being able to store a large capacity of electrical energy efficiently and at relatively low cost [[5], [6], [7], [8], [9], [10], [11], [12], [13], [14]]. With the development of zinc/chlorine redox flow batteries, RFBs were first developed in the 1960s [15]. Since then, various types of flow batteries based on different chemistries (e.g. Fe/Cr, all vanadium, Zn/Br) have been studied and demonstrated in Multi-MWh scale [[16], [17], [18]]. Recently, novel flow battery systems (e.g. AQDS/Br, Li/TEMPO, DHAQ/Fe), which are mainly based on organic redox couples or solvents, were demonstrated to be promising due to the high energy density and low active material cost [[19], [20], [21], [22]].

In both traditional aqueous RFBs and non-aqueous flow battery systems, the membrane (or the separator) is a key component. It is not only responsible for the separation of positive electrolytes and negative electrolytes, but also provides an ion conducting pathway between these two electrolytes during the charge-discharge process [[23], [24], [25]]. The performance of a flow battery, characterized by its efficiency and cyclability, is highly dependent on membrane properties. Ideally, a membrane is required to have high ion conductivity to minimize the ohmic loss as well as low permeability of redox-active ions/molecules to reduce the cross-contamination [[23], [24], [25]]. Hence, selectivity between redox-active species (such as metal ions and organic molecules) and charge carriers (such as protons) is critically important for a membrane, as well as high conductivity and stability. The most widely studied and used membranes are perfluorinated cation membranes, i.e. Nafion® membrane, as they offer high ionic conductivity and excellent stability [[26], [27], [28]]. However, this type of membrane is not without its faults. A critical drawback is the crossover of redox-active ions/molecules through the membrane, which is primarily caused by relatively large ion transport channel sizes (typically 2–4 nm). The membrane's high cost is also a major barrier for its further application. Other alternatives include the hydrocarbon ion exchange membrane, which has clear advantages in selectivity and cost, but lose out on poor stability, leading to short battery lifespan and low cyclability. Thus, there is an imperative need to develop membranes with high selectivity, low cost and high durability [[29], [30], [31]].

Porous membranes, which were designed based on the concept of differentiating the redox-active species and charge carriers (e.g. protons) by their sizes and aimed at allowing the free motion of charge carries while avoiding crossover of the redox-active species, have recently been under the spotlight [[32], [33], [34], [35]]. This idea has been demonstrated in vanadium redox flow batteries, in which the vanadium ions and protons could be separated by tuning the pore sizes of the membrane. However, the ion conduction capability and selectivity still need further improvement since the morphology of the membrane developed has not been optimized to achieve the desired performance. For instance, porous membranes that have been previously developed typically exhibited highly curved pores, which were more effective in selectivity. However, the transport pathway of charge carriers (protons) became much lengthier, thus leading to larger ionic transport resistance. In addition, uniform pore sizes for curved pores are extremely difficult to create. Optimizing membrane morphology, which involves tuning the pore shape, size, tortuosity and interconnectivity is vital for achieving high membrane performance.

In this work, we designed and fabricated a nanochannel membrane containing highly ordered, vertically aligned nanochannels that are under 1 nm in diameter size (denoted as nanochannel membrane hereafter), as shown in Fig. 1. Preparation of the nanochannel membrane involves a mesoporous silica film to form the inorganic framework, which results in the ordered and vertically aligned nanochannels. A proton-conducting monomolecular layer is subsequently assembled onto the surfaces of the nanochannels, which shrinks the pore size to ∼0.5 nm and allows the Grotthuss transport mechanism of protons. A vertically aligned channel structure enables high mobility of charge carriers (protons), while the uniform ∼ 0.5 nm channel size will ensure extraordinary selectivity. In addition, this membrane is ultrathin, at only 200 nm, minimizing ohmic resistance and vastly increasing the applications for high power flow batteries. Post fabrication, we conducted a series of characterizations on the membrane morphology, proton conductivity, as well as vanadium ion permeability. Results show that the pristine membrane grown on ITO glass is completely impermeable to vanadium ions, and the vanadium ion permeability of the free-standing membrane after ITO etching maintains extremely low which is four orders of magnitude lower than that of conventional Nafion membranes. This result suggests that the nanochannel membrane is highly suitable for flow battery applications and shows marked improvement in performance compared with that of commercially available membranes. In the further testing by integrating the present membrane onto commercially available membranes as a substrate and installing into a vanadium redox flow battery, we demonstrated a coulombic efficiency of 98.4% and energy efficiency of 82.9% at a current density of 160 mA cm−2. These remarkable results bolster our nanochannel membrane design concept and indicate its applicability in flow battery applications.

Section snippets

Preparation of the membrane

A silica nanochannel membrane (i.e. mesoporous silica film) was prepared on ITO glass by using the Stöber-solution growth approach [45]. The ITO was first treated with 10 M NaOH solution at room temperature for 10 h to cleanse organic residues and then rinsed with deionized water. Subsequently, the ITO glass was immersed into a solution that contained 80 mg of cetyltrimethylammonium bromide (CTAB), 15 mL of ethanol, 35 mL of water, and 4.7 μL of concentrated ammonia aqueous solution (28 wt.%),

Preparation of the nanochannel membrane

The silica nanochannel membrane is fabricated by the Stöber-solution growth approach, which uses self-assembled cylindrical cetyltrimethylammonium bromide (CTAB) micelles as a template. The growth time was set to 72 h. At the conclusion of this process, the silica membrane possesses order-arranged and vertically aligned nanochannels which mimic the structure of cylindrical micelles [36]. The prepared silica nanochannel membrane as-deposited on the ITO glass is transparent with the area of

Conclusion

In summary, we report an inorganic-framework-based proton exchange membrane which exhibits vertically aligned and subnanosized channels. Due to its ultrathin (200 nm), uniform and ultrasmall channels (∼0.5 nm), the membrane is promising for redox flow battery applications. High selectivity between vanadium ions and protons ensures virtually no vanadium crossover. The experimental results show that the pristine membrane grown on ITO glass is completely impermeable to vanadium ions, and the

Acknowledgements

The work described in this paper was fully supported by a grant from the Research Grants Council of the Hong Kong Special Administrative Region, China (Project No. 16213414).

References (49)

  • E. Sum et al.

    J. Power Sources

    (1985)
  • P. Leung et al.

    J. Power Sources

    (2011)
  • P. Zhao et al.

    J. Power Sources

    (2006)
  • Q. Luo et al.

    J. Membr. Sci.

    (2008)
  • Q. Zheng et al.

    J. Power Sources

    (2016)
  • G. Codina et al.

    J. Power Sources

    (1994)
  • X. Zhou et al.

    Electrochim. Acta

    (2015)
  • J. Xi et al.

    J. Power Sources

    (2007)
  • T. Mohammadi et al.

    J. Membr. Sci.

    (1995)
  • T. Mohammadi et al.

    J. Membr. Sci.

    (1995)
  • D. Reed et al.

    J. Power Sources

    (2015)
  • T. Luo et al.

    J. Power Sources

    (2016)
  • X. Yan et al.

    J. Power Sources

    (2016)
  • Q. Luo et al.

    J. Membr. Sci.

    (2008)
  • X. Yan et al.

    J. Power Sources

    (2016)
  • Y. Zeng et al.

    J. Power Sources

    (2016)
  • C.P. De Leon et al.

    J. Power Sources

    (2006)
  • M. Skyllas-Kazacos et al.

    J. Electrochem. Soc.

    (1986)
  • C. Ding et al.

    J. Phys. Chem. Lett.

    (2013)
  • B. Dunn et al.

    Science

    (2011)
  • Z. Yang et al.

    Chem. Rev.

    (2011)
  • M. Winter et al.

    Chem. Rev.

    (2004)
  • P. Leung et al.

    RSC Adv.

    (2012)
  • F.C. Walsh et al.

    ChemPlusChem

    (2015)
  • Cited by (0)

    View full text