Environmental Engineering Reference
In-Depth Information
Another type of membrane utilizing CNTs is a composite CNT membrane, which has
a structure similar to traditional polymeric membranes but CNTs are mixed into a top
layer such as polyamide [94]. These composite membranes have demonstrated approxi-
mately a 1.5× lux increase over traditional RO membranes and their fabrication is rela-
tively uncomplicated [94]. In addition, functionalization of CNTs is predicted to improve
the salt rejection, and although this functionalization is expected to decrease the water
lux over unfunctionalized CNTs it can also permit use of larger-diameter CNTs [95].
Functionalization of CNTs has been demonstrated by chemical, biological, and “active”
functionalization. Chemical functionalities such as short-chained alkanes, long-chained
alkanes, long polypeptides, and highly charged dye have all been explored as gatekeep-
ers to allow selective control over ion transport as salt ions approach the CNT entrance.
For 7-nm CNTs, polypeptide functionalization gave a permeation ratio of small dye to
large dye of 3.6, while the ratio in the bulk solution was 1.6. Active gatekeepers have been
demonstrated with long quadra-charge dye molecules that can be controlled by an electric
ield [96].
A third type of membrane has also been fabricated with CNTs, which layers CNTs on
existing low-pressure membranes to mitigate fouling at the membrane surface [97]. Low-
pressure membranes layered with 50-80 nm CNTs were shown to have the best perfor-
mance, reducing fouling onset time by almost three times that of the untreated membrane,
allowing for less backwashing, chemical cleaning, or both during membrane operation
[97]. However, more investigation of CNT adhesion will be needed, as some of the CNTs
were removed during the backwashing cycle.
Although polymeric membranes are currently the most widely used large-scale mem-
branes, a myriad of new materials are being developed for rapid permeation, including
graphene, graphene oxide, zeolites, and boron nitride nanotubes [98-102]. These new
membranes separate by molecular sieving instead of by solution/diffusion that state-of-
the-art commercial polymeric membranes use [99]. Molecular dynamics simulations have
shown that nanoporous graphene membranes of 1 cm
2
area could theoretically deliver a
water lux of 6.7 l/day at atmospheric pressure with a 99% salt rejection in comparison to
current polymeric membranes of 1 cm
2
area that deliver between 0.001 and 0.055 l/day
at atmospheric pressure and similar salt rejections [99]. However, development of robust
manufacturing processes still proves challenging, although 30-in sheets have been fabri-
cated recently [7].
Researchers have developed graphene oxide membranes with thicknesses ranging from
0.1 to 10 μm, approximately 1 cm
2
in area, and as low as 4 Ǻ in pore size. These membranes
have demonstrated negligible difference between evaporation of water without a mem-
brane and permeation through the membrane for submicron thicknesses. For the 10 μm
thickness, permeation through the membrane was two times slower than unimpeded
evaporation [101]. The effects of oxidation chemistry and membrane processing such as
annealing on water permeation are still under investigation; however, slip low is sug-
gested as the transport mechanism through these membranes similar to low through
CNTs [101].
Molecular dynamics modeling has also been used to suggest that boron nitride nano-
tubes embedded within silicon nitride membrane would be an extremely effective mate-
rial for water desalination, rejecting both cations and anions while allowing water lux
at a rate of 1.6-10.7 molecules/ns [100]. In addition, this model predicted that rejection
would occur even under high concentration (up to 1 M) and high pressures (approximately
612 MPa) [100]. Another simulation with boron nitride nanotubes of 0.83 nm in diameter
gave a range of ~12-50 molecules/ns for a pressure range of 100-500 MPa with complete
