Molybdenum Disulfide and Tungsten Disulfide as Novel …
197
approach is the mechanism in which we can perform computational simulations that
take into account thermodynamic and dynamic behaviors of nanofluidic systems.
It means that this approach can be viewed as a bridge between the quantum realm
(hidden in atomic interactions and sizes) and classical hydrodynamics.
While there are clear benefits to using MD simulations, there are also major
challenges. Computational materials scientists have worked hard to design general,
accurate, and reliable molecular and atomistic models. For example: in the case of
water, the model chosen to represent atomic geometries and interaction parameters
is the seed in which the whole dynamic relies on so that the physics can emerge
following the classical equations of motion. The mainstream model in MD simulations is based on the Lennard–Jones potential plus a Coulombic term. The task of
finding the right model—which is generally suitable only for a handful of specific
systems—has haunted theoretical physicists and chemists since the first computer
simulation of liquid water with the Bernal–Fowler model in 1969. Even with the
help of new experimental data and theories being developed, water is still notoriously hard to model and remains relatively poorly understood, with several anomalous properties—behaviors which contradict general theories on the liquid state of
matter.
The interest in nanometric desalination systems has led us to question how the
key properties of saline solutions can be captured in a computer simulation (e.g.,
mixing different interaction parameters). One of the most widely used methodologies
to classically simulate saltwater desalination consists of creating a box with the
membrane located between two reservoirs, each one pressed by a piston—usually
graphene-made. This imposes a controlled pressure gradient in water as illustrated
in Fig. 2a. Here we can highlight that (i) as we apply different pressures in each
reservoir and (ii) the solutions are at different concentrations, the system is not under
thermodynamic equilibrium. In fact, these simulations aim to find water transport
and salt rejection rates acquired at a steady-state flow, which is only possible in nonequilibrium states. Of course, that as the simulation runs the system will eventually
reach equilibrium. But to mimic an actual nanoscaled RO scenario the simulation
should be far from this point.
In order to filter water using 2D membranes, we can use either the interlayer
spacing, forcing water to flow through the structure gaps, or drill holes (nanopores)
in the membrane. In any case, it is important to ensure that the nanopore size allows
only for the flow of water molecules. Interestingly, nanopores in 2D materials such as
TMDs will naturally appear during the growth process. Point defects, grain boundaries, and van der Waals (vdW) gaps, among other structural deformations, have
been observed in CVD grown of MoS 2 and WS 2 monolayers [19–21]. At first glance,
this might seem like a disadvantage, but the truth is that these “flaws” lead to the
emergence of nanopores and nanogaps that can be used to desalinate water and
remove heavy metals, biomolecules, and other pollutants. In the case of MoS 2 , for
instance, these intrinsic defects appear in high concentrations (~10
13 cm
−2 for sulfur
vacancies [22]). However, these “natural” pores are randomly distributed with varied
sizes, which pose a challenge for scalability—a very important aspect in large-scale
production.
197
approach is the mechanism in which we can perform computational simulations that
take into account thermodynamic and dynamic behaviors of nanofluidic systems.
It means that this approach can be viewed as a bridge between the quantum realm
(hidden in atomic interactions and sizes) and classical hydrodynamics.
While there are clear benefits to using MD simulations, there are also major
challenges. Computational materials scientists have worked hard to design general,
accurate, and reliable molecular and atomistic models. For example: in the case of
water, the model chosen to represent atomic geometries and interaction parameters
is the seed in which the whole dynamic relies on so that the physics can emerge
following the classical equations of motion. The mainstream model in MD simulations is based on the Lennard–Jones potential plus a Coulombic term. The task of
finding the right model—which is generally suitable only for a handful of specific
systems—has haunted theoretical physicists and chemists since the first computer
simulation of liquid water with the Bernal–Fowler model in 1969. Even with the
help of new experimental data and theories being developed, water is still notoriously hard to model and remains relatively poorly understood, with several anomalous properties—behaviors which contradict general theories on the liquid state of
matter.
The interest in nanometric desalination systems has led us to question how the
key properties of saline solutions can be captured in a computer simulation (e.g.,
mixing different interaction parameters). One of the most widely used methodologies
to classically simulate saltwater desalination consists of creating a box with the
membrane located between two reservoirs, each one pressed by a piston—usually
graphene-made. This imposes a controlled pressure gradient in water as illustrated
in Fig. 2a. Here we can highlight that (i) as we apply different pressures in each
reservoir and (ii) the solutions are at different concentrations, the system is not under
thermodynamic equilibrium. In fact, these simulations aim to find water transport
and salt rejection rates acquired at a steady-state flow, which is only possible in nonequilibrium states. Of course, that as the simulation runs the system will eventually
reach equilibrium. But to mimic an actual nanoscaled RO scenario the simulation
should be far from this point.
In order to filter water using 2D membranes, we can use either the interlayer
spacing, forcing water to flow through the structure gaps, or drill holes (nanopores)
in the membrane. In any case, it is important to ensure that the nanopore size allows
only for the flow of water molecules. Interestingly, nanopores in 2D materials such as
TMDs will naturally appear during the growth process. Point defects, grain boundaries, and van der Waals (vdW) gaps, among other structural deformations, have
been observed in CVD grown of MoS 2 and WS 2 monolayers [19–21]. At first glance,
this might seem like a disadvantage, but the truth is that these “flaws” lead to the
emergence of nanopores and nanogaps that can be used to desalinate water and
remove heavy metals, biomolecules, and other pollutants. In the case of MoS 2 , for
instance, these intrinsic defects appear in high concentrations (~10
13 cm
−2 for sulfur
vacancies [22]). However, these “natural” pores are randomly distributed with varied
sizes, which pose a challenge for scalability—a very important aspect in large-scale
production.