206
M. H. Köhler et al.
Long before 2D membranes came to the spotlight of research on desalination,
membranes composed of CNTs were extensively studied as attractive materials [71].
The investigation of water transport through CNTs led to the discovery of new and
exciting properties, such as flow rate enhancement [1], which was seen as a very
useful feature for desalination purposes. In 2D materials, the high transport rate is
associated with their extremely low thickness of just a single atom.
Notwithstanding, the high water transport rate observed in CNTs is a consequence of the smooth hydrophobic inner core, which allows for uninterrupted water
molecules passage with negligible adsorption and almost no friction [75]. Indeed,
the enhancement flow factor—defined as the ratio between measured flow and the
ideal no-slip Poiseuille flow—puts classical hydrodynamic theory in check, once the
condition of zero interfacial fluid velocity does not necessarily hold at nanoscopic
length scales [58]. This means that while water permeability goes down as we
increase the thickness of 2D membranes, it is almost independent of the nanotube
length, making CNT membranes still commercially attractive. Secchi et al. [17] have
confirmed experimentally fast water transport behavior with nearly frictionless interfaces through carbon nanotubes. Besides these elevated speeds, high aspect ratio, and
easy functionalization—a prerequisite for desalination purposes, avoiding aggregation that harms ion selectivity and water flux—renders CNTs as a widely explored
nanomaterial in water purification research.
Among other categories, we could divide CNT-based membranes into two types
based on their configurations: freestanding and mixed with polymeric materials [69].
The former can be produced either with vertically aligned nanotubes—where water
is forced through inside them—or as buckypaper membranes—in short, a random
network of CNTs with the large specific surface area. This network is structurally
similar to the current commercial thin-film RO membrane composite in which CNTs
are mixed with the top layer polymer. Buckypaper CNT membrane poses as an
excellent alternative for desalination in distillation technology [76]. In contrast, Baek
et al. [77] successfully synthesized a vertically aligned CNT membrane with pore
diameters of ~ 4.8 nm and a pore density of 6.8 × 10
10 cm
−2 . They presented
improved performance (three times higher flux) compared to typical ultrafiltration
(UF) membranes. Additionally, vertically aligned CNT membranes can be obtained
from thermal and oxygen-plasma treatments of densified outer-wall CNTs [78]. The
advantage of the latter is that pore diameter can be readily varied (e.g., from 7 to
40 nm) through simple mechanical compression. CNT wall membranes obtained in
this fashion can deliver water permeability that approaches 30,000 L·m
−2 ·h
−1 ·bar
−1
and still avoid bacterial adhesion and biofilm formation. As in every other nanoscaled
membrane, the experimental challenges here lie in the production of well-defined
specific nanotube diameters needed for selectivity purposes, pore size homogeneity,
alignment, and agglomeration control [62]. MD simulations have suggested that in
order to achieve desalination capacity comparable to that of RO membranes the inner
diameter of nanotubes should be around 0.6 nm [79], and current state-of-the-art CNT
membrane synthesis is not able to meet these requirements. Also, there are concerns
about potential nanotoxicity in the aquatic environment [70, 80].
M. H. Köhler et al.
Long before 2D membranes came to the spotlight of research on desalination,
membranes composed of CNTs were extensively studied as attractive materials [71].
The investigation of water transport through CNTs led to the discovery of new and
exciting properties, such as flow rate enhancement [1], which was seen as a very
useful feature for desalination purposes. In 2D materials, the high transport rate is
associated with their extremely low thickness of just a single atom.
Notwithstanding, the high water transport rate observed in CNTs is a consequence of the smooth hydrophobic inner core, which allows for uninterrupted water
molecules passage with negligible adsorption and almost no friction [75]. Indeed,
the enhancement flow factor—defined as the ratio between measured flow and the
ideal no-slip Poiseuille flow—puts classical hydrodynamic theory in check, once the
condition of zero interfacial fluid velocity does not necessarily hold at nanoscopic
length scales [58]. This means that while water permeability goes down as we
increase the thickness of 2D membranes, it is almost independent of the nanotube
length, making CNT membranes still commercially attractive. Secchi et al. [17] have
confirmed experimentally fast water transport behavior with nearly frictionless interfaces through carbon nanotubes. Besides these elevated speeds, high aspect ratio, and
easy functionalization—a prerequisite for desalination purposes, avoiding aggregation that harms ion selectivity and water flux—renders CNTs as a widely explored
nanomaterial in water purification research.
Among other categories, we could divide CNT-based membranes into two types
based on their configurations: freestanding and mixed with polymeric materials [69].
The former can be produced either with vertically aligned nanotubes—where water
is forced through inside them—or as buckypaper membranes—in short, a random
network of CNTs with the large specific surface area. This network is structurally
similar to the current commercial thin-film RO membrane composite in which CNTs
are mixed with the top layer polymer. Buckypaper CNT membrane poses as an
excellent alternative for desalination in distillation technology [76]. In contrast, Baek
et al. [77] successfully synthesized a vertically aligned CNT membrane with pore
diameters of ~ 4.8 nm and a pore density of 6.8 × 10
10 cm
−2 . They presented
improved performance (three times higher flux) compared to typical ultrafiltration
(UF) membranes. Additionally, vertically aligned CNT membranes can be obtained
from thermal and oxygen-plasma treatments of densified outer-wall CNTs [78]. The
advantage of the latter is that pore diameter can be readily varied (e.g., from 7 to
40 nm) through simple mechanical compression. CNT wall membranes obtained in
this fashion can deliver water permeability that approaches 30,000 L·m
−2 ·h
−1 ·bar
−1
and still avoid bacterial adhesion and biofilm formation. As in every other nanoscaled
membrane, the experimental challenges here lie in the production of well-defined
specific nanotube diameters needed for selectivity purposes, pore size homogeneity,
alignment, and agglomeration control [62]. MD simulations have suggested that in
order to achieve desalination capacity comparable to that of RO membranes the inner
diameter of nanotubes should be around 0.6 nm [79], and current state-of-the-art CNT
membrane synthesis is not able to meet these requirements. Also, there are concerns
about potential nanotoxicity in the aquatic environment [70, 80].