Greener processes in the preparation of thin film nanocomposite membranes with diverse metal-organic frameworks for organic solvent nanofiltration

https://doi.org/10.1016/j.jiec.2019.04.057Get rights and content

Highlights

  • A greener process to prepare thin film nanocomposite membranes for nanofiltration.

  • Substitution of DMF by DMSO as an activating solvent.

  • Thin film nanocomposite membranes using UiO-66, ZIF-8 and ZIF-93.

  • Performance given by MOF porosity, microdefects and hydrophilic/hydrophobic character.

Abstract

The toxic solvent dimethylformamide has been replaced by the greener dimethylsulfoxide in the casting and activation processes for the preparation of thin film composite (TFC) membranes. This methodology has been validated with the use of MOFs ZIF-8, ZIF-93 and UiO-66 as fillers for thin-film nanocomposite (TFN) membranes. These membranes were successfully applied in the organic solvent nanofiltration of sunset yellow (SY) in methanol obtaining the highest permeance when UiO-66 and ZIF-93 were used with a value of 11 L m−2 h-1 bar-1. The permeance improvements are related to the MOF porosity, polyamide-MOF layer thickness and the hydrophilic/hydrophobic character of the membrane.

Introduction

Nanofiltration is a membrane separation process for liquids characterized by an operating pressure difference ranging from 5 to 40 bar and a molecular weight cut-off (MWCO) between 200 and 1000 Da [1]. While this technique has been widely used in water treatment processes [2], [3], [4], it has recently received much attention for its application with organic solvents, the so-called organic solvent nanofiltration (OSN) process, with important economic, environmental and safety benefits [1], [5].

The most competitive membranes in OSN are the so-called thin film composite (TFC) membranes, first developed by Peterson [6]. Although, thin film nanocomposite membranes (TFN, i.e. including fillers in the TFC membrane), first developed by Jeong et al. [7] for reverse osmosis, have also been widely used for OSN [5,8] obtaining an improvement in permeance without sacrificing rejection in comparison with TFC membranes. Nowadays several different nanoparticles are also used as fillers, namely TiO2 [[9], [10], [11]], MCM-41 silica [12], graphene oxide [13] and a limited range of metal-organic frameworks (MOFs) [14–16]. The foregoing is related to the tendency of modification and improvement of the thin film membrane as well as the synthesis and applications related to the MOF [[17], [18], [19]].

The main drawback in the fabrication of TFN membranes is that the polymer necessary to prepare the support is usually soluble in highly toxic organic solvents, such as N,N-dimethylformamide (DMF), N-methyl-2-pyrrolidone (NMP) or N,N-dimethylacetamide (DMAc) [20]. Moreover, in the case of TFC and MOF-TFN membranes, DMF is commonly used in the post-treatment as an activating solvent (either by bath, filtration or a combination of both treatments) [16,21,22]. In recent years, the principles of Green Chemistry are implanting [23] focusing on resource efficiency, nontoxicity and the environmentally friendly profile of solvents, and on the overall life cycle assessment of the product or process [24].

Solvent selection guides, in particular Sanofi’s [25], provide relevant information and rankings of commonly used solvents based on several features that must be considered when designing a “green” membrane. Health issues of solvents (acute, long-term and single target organ toxicity) are evaluated by REACH (Registration, Evaluation, Authorisation and Restriction of Chemical substances) and, in particular, DMF, DMAc and NMP are classified as substances of very high concern (SVHC) in a list that is updated twice a year [26]. Given other disadvantages of DMF such as corrosivity, a melting point at 18 °C and the formation of dimethylsulfide, REACH has proposed DMSO as an advisable substitute because of its low toxicity for human health. In turn, Sanofi classified DMSO as “substitution advisable” in contrast with DMF classified as “substitution requested”.

However, very few publications have presented strategies for reducing the impact of membrane production. Of these, da Silva Burgal et al. [27] used poly(ether ether ketone), a chemical resistant polymer that does not require cross-linking and dissolves in solvents that can be easily neutralized by water. Hua et al. [28] carried out the synthesis of the selective layer using water as the reaction medium instead of hexane. Figoli et al. [29] described many successful cases were DMSO was applied for the preparation of membranes. In particular, Soroko et al. [30]. developed a new route to synthesize TFC OSN membranes using DMSO as a polyimide (PI) solvent instead of DMF and Jimenez Solomon et al. [22] used DMSO as the activating solvent for TFC membrane post-treatment instead of DMF. DMSO and DMF can be considered as interchangeable because of their similar Hansen solubility parameters and ability to dissolve both polyimide and polyamide [29]. Furthermore, DMSO is the most environmentally friendly solvent among other PI diluents (DMF and NMP) in terms of its emissions and resource use. All these solvents were produced through the “methanol route”. Capello et al. [31] showed their Life Cycle Assessment and energy profiles, obtained by the Cumulative Primary Energy Demand (CED). DMF is produced in two steps requiring between 50 and 100 MJ-eq per kg of product. NMP requires four production steps and between 100 and 150 MJ-eq per kg of product. In contrast, DMSO is produced in only one step and causes the lowest CED with less than 50 MJ-eq per kg of product.

Our aim for a more sustainable nanofiltration process is to design greener processes for membrane preparation but also to achieve membranes able to resist and filter organic solvents. Here, we report the preparation of TFC and MOF-TFN membranes using DMSO, a greener solvent than traditional ones, both to dissolve the polymer and to activate the membrane. In addition, continuing with the development of MOFs as fillers in TFN membranes, we have incorporated ZIF-8 and two other MOFs that to date have not been used as fillers in OSN: ZIF-93 and UiO-66. These MOFs (see their structure and composition in Fig. 1a) differ in their pore size and hydrophilicity and provide the membrane with different performances in OSN applications.

Section snippets

MOFs preparation

The syntheses of ZIF-93 [35], ZIF-8 [36] and UiO-66 [37] were carried out as previously reported. The detailed procedures can be found in the Supporting Information.

Preparation of PI supports

The PI supports were prepared as follows: a dope solution of 24% (w/w) was prepared by dissolving P84® (HP polymer GmbH) in DMSO (99.5%, Scharlab) and stirring overnight. Once all the air bubbles disappeared, the solution was cast on a polypropylene non-woven backing material at a casting speed of 0.04 m s−1 using a casting knife

MOFs characterization

The XRD patterns of the MOFs prepared in this work (Fig. S1 in Supplementary material) reveal their purity and crystalline structure after comparing them with the simulated ZIF-8, ZIF-93 and UiO-66 XRD patterns. Fig. 1a–c shows the structure of these MOFs and their corresponding organic linker. The nanosized MOF crystals showed the expected morphology (Fig. 1d–f and Fig. S2 in Supplementary material) as described in the literature [35], [36], [37] and were obtained with a narrow particle size

Conclusions

The substitution of DMF as casting and activating solvent in the preparation of TFC and TFN membranes by the greener DMSO has been assessed. Characterization by SEM, AFM, contact angle and gel content of TFC membranes post-treated with both solvents revealed few differences between them. Nanofiltration of MeOH with Sunset Yellow produced similar results, these being slightly better when using DMSO owing to an increase in roughness.

The substitution of DMF by DMSO as an activating solvent was

Acknowledgment

Financial support from the Research Projects MAT2016-77290-R (MINECO/AEI/FEDER, UE) and T43-17R (the Aragón Government and the ESF) is gratefully acknowledged. Lorena Paseta would like to express her gratitude to the Spanish MINECO for the predoctoral grant (BES-2014-068287) awarded. All the microscopy work was done in the Laboratorio de Microscopías Avanzadas at the Instituto de Nanociencia de Aragón (LMA-INA). Finally, the authors would like to acknowledge the use of the Servicio General de

References (54)

  • J. Radjenović et al.

    Water Res.

    (2008)
  • K. Košutić et al.

    Sep. Purif. Technol.

    (2005)
  • C. Tang et al.

    Desalination

    (2002)
  • R.J. Petersen

    J. Membr. Sci.

    (1993)
  • B.-H. Jeong et al.

    J. Membr. Sci.

    (2007)
  • W.J. Lau et al.

    Water Res.

    (2015)
  • D. Emadzadeh et al.

    Chem. Eng. J.

    (2014)
  • I. Soroko et al.

    J. Membr. Sci.

    (2009)
  • M. Peyravi et al.

    Chem. Eng. J.

    (2014)
  • J. Yin et al.

    J. Membr. Sci.

    (2012)
  • J. Yin et al.

    Desalination

    (2016)
  • J. Duan et al.

    J. Membr. Sci.

    (2015)
  • H. Wang et al.

    Chem. Soc. Rev.

    (2019)
  • D. Jiang et al.

    Coord. Chem. Rev.

    (2018)
  • W. Xiong et al.

    Chemosphere

    (2018)
  • M.F. Jimenez Solomon et al.

    J. Membr. Sci.

    (2012)
  • L. Hou et al.

    Polym. Chem.

    (2016)
  • P. Gorgojo et al.

    Desalination

    (2014)
  • A.K. Ghosh et al.

    J. Membr. Sci.

    (2008)
  • L. Paseta et al.

    Org. Biomol. Chem.

    (2015)
  • C.M. Hansen

    Prog. Org. Coat.

    (2004)
  • C.H. Koo et al.

    Desalination

    (2012)
  • S. Roy et al.

    J. Membr. Sci.

    (2011)
  • L. Shao et al.

    J. Membr. Sci.

    (2014)
  • Y. Li et al.

    Sep. Purif. Technol.

    (2016)
  • S. Darvishmanesh et al.

    J. Membr. Sci.

    (2009)
  • P. Marchetti et al.

    Chem. Rev.

    (2014)
  • Cited by (48)

    • Sustainable and green membranes for chemical separations: A review

      2024, Separation and Purification Technology
    • Green and Sustainable Membranes: A review

      2023, Environmental Research
    View all citing articles on Scopus
    View full text