Molybdenum Disulfide and Tungsten Disulfide as Novel …
209
by individually stitching 2D grains of distinct crystallographic orientation. The size
of these nanopores is larger than that of the atomic vacancies within basal planes
as they are generally composed of multiple uncoordinated atoms. Grain-boundary
nanopores play a very important role in the membrane’s permeability, governing
both water permeation and ion rejection. The uniformly spaced vdW gaps in the
CVD-grown 2D MoS 2 also help the membrane to achieve high ionic sieving capability. In an experimental realization of few-layer 2D MoS 2 membranes, Li et al.
[12] found the physical sizes of the hydrated ions to be larger than the interlayer
vdW MoS 2 gap, which suggests that ion transportation is efficiently impeded. In
addition to this geometrical effect, the electrostatic interaction of atomic vacancies is
considered another major factor governing both water permeation and ionic sieving.
Theoretical works have suggested that salt ions encounter significantly high energetic and steric barriers when approaching sulfur vacancies, while water molecules
are relatively unaffected [60]. In this case, cations such as Na
+ will experience a
high Coulombic barrier due to the positively charged sulfur vacancies, which in
turn expose the hydrophilic Mo-rich sites to ignite a process of water attraction and
ion repulsion. Indeed, we have identified in our simulations that nanopores with
up to ~1 nm present strong ion rejection rates even at high pressures of ~100 MPa
[13]. Although experimentalists have shown good control of the naturally occurring
defects in CVD growth of TMDs, we still miss a systematic study on the vacancies
distribution where their effect, grain boundary, and vdW gaps would be properly
assessed.
(c) Imaging—In order to take advantage of all the prominent desalination and adsorption features of MoS 2 and WS 2 membranes, it is very important to establish a precise,
reproducible, rapid, and nondestructive method to effectively image nanopores and
grain boundaries. This method should be also independent of the membrane composition, doping, and defect reconstruction. Although 2D nanoporous membranes stand
as a new class of materials, the characterization of solid-state 2D nanostructures is
an old, always evolving area of research. Therefore, the long-term knowledge we
have built around this technology can be used to access atomic-scale information
about nanopore geometry and distribution along the membrane. For instance, we
have seen that transmission electron microscopy (TEM) directly resolves atomistic
details of defects but requires intensive and disruptive sample preparation [19, 88, 89].
Alternatively, we can use photoluminescence or Raman spectroscopy that provides
rapid and nondestructive probes of the electronic and vibrational properties of defective regions. Both techniques often manifest as red- or blue-shifted emission with
enhanced or suppressed intensities when compared to the response from a pristine
crystal, but it depends on multiple factors that affect local electronic properties such
as material composition, doping level, defect passivation, grain boundary geometry,
or edge terminations [90, 91]. Nonlinear optical spectroscopy can be highly sensitive to imaging imperfections, but their visualization can exhibit weak background
contrast [92] or require the use of chemical solvents [93]. We can alternatively boost
image contrast using dark-field (DF) microscopy. By comparing a bright-field linear
optical image of a TMD monolayer on quartz and a dark-field linear optical image
209
by individually stitching 2D grains of distinct crystallographic orientation. The size
of these nanopores is larger than that of the atomic vacancies within basal planes
as they are generally composed of multiple uncoordinated atoms. Grain-boundary
nanopores play a very important role in the membrane’s permeability, governing
both water permeation and ion rejection. The uniformly spaced vdW gaps in the
CVD-grown 2D MoS 2 also help the membrane to achieve high ionic sieving capability. In an experimental realization of few-layer 2D MoS 2 membranes, Li et al.
[12] found the physical sizes of the hydrated ions to be larger than the interlayer
vdW MoS 2 gap, which suggests that ion transportation is efficiently impeded. In
addition to this geometrical effect, the electrostatic interaction of atomic vacancies is
considered another major factor governing both water permeation and ionic sieving.
Theoretical works have suggested that salt ions encounter significantly high energetic and steric barriers when approaching sulfur vacancies, while water molecules
are relatively unaffected [60]. In this case, cations such as Na
+ will experience a
high Coulombic barrier due to the positively charged sulfur vacancies, which in
turn expose the hydrophilic Mo-rich sites to ignite a process of water attraction and
ion repulsion. Indeed, we have identified in our simulations that nanopores with
up to ~1 nm present strong ion rejection rates even at high pressures of ~100 MPa
[13]. Although experimentalists have shown good control of the naturally occurring
defects in CVD growth of TMDs, we still miss a systematic study on the vacancies
distribution where their effect, grain boundary, and vdW gaps would be properly
assessed.
(c) Imaging—In order to take advantage of all the prominent desalination and adsorption features of MoS 2 and WS 2 membranes, it is very important to establish a precise,
reproducible, rapid, and nondestructive method to effectively image nanopores and
grain boundaries. This method should be also independent of the membrane composition, doping, and defect reconstruction. Although 2D nanoporous membranes stand
as a new class of materials, the characterization of solid-state 2D nanostructures is
an old, always evolving area of research. Therefore, the long-term knowledge we
have built around this technology can be used to access atomic-scale information
about nanopore geometry and distribution along the membrane. For instance, we
have seen that transmission electron microscopy (TEM) directly resolves atomistic
details of defects but requires intensive and disruptive sample preparation [19, 88, 89].
Alternatively, we can use photoluminescence or Raman spectroscopy that provides
rapid and nondestructive probes of the electronic and vibrational properties of defective regions. Both techniques often manifest as red- or blue-shifted emission with
enhanced or suppressed intensities when compared to the response from a pristine
crystal, but it depends on multiple factors that affect local electronic properties such
as material composition, doping level, defect passivation, grain boundary geometry,
or edge terminations [90, 91]. Nonlinear optical spectroscopy can be highly sensitive to imaging imperfections, but their visualization can exhibit weak background
contrast [92] or require the use of chemical solvents [93]. We can alternatively boost
image contrast using dark-field (DF) microscopy. By comparing a bright-field linear
optical image of a TMD monolayer on quartz and a dark-field linear optical image