1 INTRODUCTION
Membrane separation technology with high energy efficiency and low carbon footprints has drawn increasing attention in a wide range of applications including petroleum refining, water purification and carbon capture, etc .[1-5] Two-dimensional (2D) lamellar membranes with controllable microstructure and well-aligned nanochannels make excellent candidates for devising fast molecule permeation and precise sieving.[6-8] Typically, membranes assembled by porous nanosheets offer continuous transfer highways in vertical direction, which further shortens mass transfer distance and permits fast transport.[9-11]Molecule transport through porous lamellar membranes should firstly dissolve into pore entrances on membrane surface and then diffuse through inner pores. It has been demonstrated that the resistance of molecular dissolution process accounts for 20% ~ 40% of the total transport resistance, and even 70% for membranes with sub-10 nm thickness.[12-14]Therefore, molecular dissolution process exerts huge and even decisive impact on molecular transport efficiency, which deserves the investigation of underlying molecular dissolution behaviors on membrane surface.
In general, experimental researches have confirmed that enhancing the affinity of membrane surface to molecules by grafting active sites or attaching compatible films could obviously improve molecular dissolution rate and hence permeation ability.[15-17] As a noteworthy example, water-capture capacity of over 300% was achieved for graphene oxide (GO) lamellar membrane by modifying tannic acid on membrane surface, which highly improve water permeance from 175 to 320 L m-2 h-1.[18]These observations are generally attributed to the strong molecule-membrane interactions that enrich molecules on membrane surface and promote them to enter into inner channels. In contrast, recent studies discovered that due to the strong interactions between water molecules and –COO groups on the pore rim of aquaporins, water permeance was 6 times lower than in carbon nanotube porins (CNTPs) which possessed similar channel size but hydrophobic pore rim.[19] This phenomenon parallels findings by simulation where the addition of four –COO groups to the rims of a 1.1-nm-diameter CNTPs reduced the water flux by 3 folds due to the enhanced water-rim interactions. These controversial viewpoints suggest that molecular dissolution behaviors need deep investigation, espeically for porous materials. On the other hand, expect for molecule-membrane interaction, other factors such as molecule-molecule interaction should also count much due to the confinement effect of molecules on pore entrances. However, the pores in current porous lamellar membrane are usually artificially created with random distribution and wide pore size distribution.[20-22]And the chemical groups on pore entrances are heterogeneous, which hampers the control of molecule-pore interactions on molecule level. More importantly, lamellar membranes generally possess homogeneous structure, and the high resistance of molecular diffusion in inner pores exerts huge impact on molecule permeance. This conceals the contribution of dissolution process on molecule permeance, hence impeding the exploration of underlying molecular dissolution behaviors.[20,23,24]Therefore, constructing hierarchical porous lamellar membranes with independently manipulated surface and support layers, and especially regular and controllable pore entrances on surface layer, should provide desired platform.
Metal-organic frameworks (MOFs) show great expectations as next-generation materials for porous membranes because of intrinsic characteristics including large porosity, well-defined pore structures, and tunable molecule-/ion- specific functional groups.[25-29] Significantly, the intrinsic pores (usually < 2 nm) in MOF nanosheets can be subtly tailored by controlling metallic node and organic linkers.[30-32] And the topology and chemical composition of pores are periodically distributed within the frameworks, which permits subtle manipulation of molecule-pore interaction on molecule level. Additionally, some MOFs can be synthesized in the form of ultrathin nanosheets with large lateral size (> 2 μm), holding promise for constructing porous lamellar membranes.[9,33,34] On the other hand, electrostatic atomization technology provides a novel strategy to evenly spray atomized nanodrops that carry desired nanosheets onto receiver to assemble ultrathin lamellar architectures.[35,36]In this way, hierarchical lamellar membranes with independently controlled surface and support layer can be facilely constructed.
Herein, hierarchical MOF lamellar membranes with ultrathin (~ 7 nm) surface layer and identical support layer (~ 553 nm) were preparedvia double-needled electrostatic atomization technology. Specifically, MOF nanosheets with hydrophobic (–CH3) pores were sprayed on Nylon substrate to construct support layers. And MOF nanosheets with hydrophilic (–NH2) or hydrophobic (–CH3) groups on pores were assembled to create surface layers, respectively. Based on these platforms, molecular dissolution behaviors on porous membrane surface was investigated in detail. Dissolution performances demonstrate that molecular dissolution behaviors are highly affected by molecule-molecule and molecule-pore interactions. Furthermore, the corresponding dissolution model equations were established in term of dissolution activation energy (E S) on membrane surface with hydrophilic and hydrophobic pore entrances. Importantly, membrane with hydrophilic pores on surface exerts strong interactions with polar solvents, contributing to lowE S and high permeance of over 270 L m-2 h-1 bar-1 for polar solvents. In addition, the regular pores within MOF lamellar membranes bring precise sieving with rejection of 99% for reactive red (1.6 × 1.4 nm). The elucidation of molecular dissolution behaviors on membrane surface is paramount to develop highly permeable and selective membranes for separation applications.
2 MATERIALS AND METHODS
2.1 Materials
NiCl2·6H2O (99.9%) and CoCl2·6H2O (99.9%) were purchased from Aladdin Reagent. Triethylamine (TEA), benzenedicarboxylic acid (BDC), 2-aminoterephthalic acid (NH2-BDC), 2-methylterephathalic acid (CH3-BDC) and diiodomethane were bought from Macklin Reagent. Organic solvents (methanol, ethanol, dimethyl sulfoxide (DMSO), N,N-dimethylformamide (DMF), acetonitrile, acetone, n-octane, cyclohexane, n-hexane, n-pentane and toluene) were obtained from Tianjin Kemiou Chemistry Reagent Co., Ltd. Nylon microfiltration substrates (0.2 μm pore size, 50 mm diameter) were bought from Tianjin Jinteng Experimental Equipment Co., Ltd. Polyethylene glycol (PEG, 200, 300, 400, 600, 800, 1000), Brilliant blue (BB), reactive red (RR), rhodamine B (RB), crystal violet (CV), and methylene blue (MB) were supplied by Aladdin Chemical Co., Ltd. These were used as received without further treatment, and deionized water was used throughout the experiment.