Note: Descriptions are shown in the official language in which they were submitted.
CA 2963431 2017-04-06
METHOD OF PRODUCTION OF NANOPOROUS MEMBRANES FOR WATER
PURIFICATION FROM METAL IONS AT LOW DIFFERENTIAL PRESSURES
FIELD
The present disclosure relates to water permeable molecular sieves and
methods of production for use in purifying water contaminated with metal ions,
and
more particularly this disclosure provides method of producing water permeable
molecular sieves characterized by water permeability at low differential of
less than
100 kPa.
BACKGROUND
In recent years, the global demand for clean water has been continuing to
increase due to rapid growth of the world's population and industry. Reverse
osmosis (RO) has become a vital research area in water purification, with the
potential to decisively decrease the energy footprints associated with other
purification methods, including chemical precipitation, flocculation, ion
exchange,
and electro-membrane systems (see references 1, 2) The energetic cost of
purifying water by RO is determined by the differential pressure (typically
¨100
kPa) (see reference 3) required at the two faces of a porous membrane
traversed
by water. Membranes operating at lower differential pressures are in
tremendous
demand to limit the energetic cost of RO. Since water flux across a membrane
scales inversely with the membrane thickness, thinner membranes are vital to
decrease the operating pressure of RO processes at a given water flow, and
1
=
CA 2963431 2017-04-06
graphene-based materials have been identified as the ideal alternatives to
existing
RO membrane materials due to their thinness (see references 4, 5).
Among graphene-based materials, graphene oxide (GO) has been widely
explored for RO filtration. In principle, GO is capable of filtrating both
metal and
organic contaminants by size-based rejection (see references 6, 7). However,
graphene oxide-based filtration relies on water passing through the interlayer
spacing between GO layers which poses very stringent limitations on GO
filtration
devices: (i) typically, filtration occurs through the spacing between two
oxidized
graphene layers, which is tuneable only over a very limited range; (ii) there
are
little benefits in terms of increased flow rate due to graphene thinness,
because
flakes are not used along their thinnest dimension, but as sorbents along
their
lateral dimensions; and (iii) GO domains ideally need to be placed vertically
to
minimize pathway through the pores and ensure a good flow rate at acceptable
differential pressures, which may lead to water leaks between neighboring
domains and thus limit the effectiveness of filtration (see references 7, 8,
9). To
summarize, purification utilizing GO interlayers has inherent limitations, in
spite of
incremental research focused on improving the flow rate and decreasing the
filtration pressure.
Non-oxidized, single-layers of graphene have also been proposed as
potential alternatives to GO (see references 10,11). Water filters can be made
by
perforating single-layer graphene with sub-nanometre pores to produce
atomically
thin sieves. However, preparing continuous single-layers of non-oxidized and
porous graphene is high in cost and complexity, and offers limited promise in
terms
2
CA 2963431 2017-04-06
of scalability. It requires large-area single-layer graphene, normally
prepared by
energy-expensive chemical vapor deposition, combined with advanced procedures
to create pores, such as ion or electron beam irradiation in ultrahigh vacuum
(see
references 12, 13). In order to solve these issues, chemical etching of non-
oxidized graphene has been explored (see references 14, 15), but pores
fabricated
so far by these methods suffer from significant limitations, including
prohibitively
long etching times (1 week or more) or pores too large (>200 nm) to be used
for
water filtration of metal ion contaminants. A significant breakthrough in
terms of
design of water purification devices and efficient pore fabrication in
graphene-
based membranes is needed to exploit their unique thinness for water
purification.
It would be very advantageous to provide water permeable molecular sieve
characterized by water permeability at low differential pressures such that
water is
able to pass through the sieve with differential pressures applied across the
two
sides of significantly less than 100 kPa.
SUMMARY
The present disclosure provides a method producing a water permeable
molecular sieve, comprising:
a) providing a porous substrate having micron-size pores;
b) depositing non-porous 2D platelets onto a surface of the porous substrate
to seal, at the surface, pores in the porous substrate to form a layer of 2D
platelets;
c) depositing a curable sealing material onto the layer of 2D platelets and
any remaining exposed areas of the surface of the porous substrate and curing
the
3
CA 2963431 2017-04-06
curable sealing material in order to form a sealed layer on the surface of the
porous substrate to prevent water by-passing the non-porous 2D platelets and
passing through the porous substrate; and
d) producing an array of sub-nanopores through the sealed layer with the
array of sub-nanopores having a size to allow water to pass therethrough but
not
metal ions to give a water permeable molecular sieve characterized by water
permeability at low differential pressures.
In some embodiments the porous substrate having micron-size pores may
comprise any one of microporous TeflonTm, polytetrafluoroethylene,
polycarbonate,
nitrocellulose, anodized alumina, fritted glass, plastic grids and metallic
grids.
In some embodiments the non-porous 2D platelets may comprise any one
of graphene platelets, graphene oxide platelets, doped graphene platelets,
functionalized graphene platelets, boron nitride platelets, MoS2 platelets,
MoSe2
platelets, carbon platelets, carbon fibres, micro graphite platelets,
nanocrystalline
graphite platelets, nickel oxide platelets, nickel oxide tubules, silicon
whiskers, and
silicon platelets.
In some embodiments the curable sealing material may comprise any one
or combination of sol-gel processed materials, epoxy resins, vinyl glues
(vinyl
polymers), polyurethane, curable polymers and ceramics.
In some embodiments the sol-gel processed materials may comprise any
one or combination of alkoxides, silicates, acrylates, siloxanes, ormosils,
silica
gels, and sulfides.
4
CA 2963431 2017-04-06
,
In some embodiments the epoxy resins comprise any compounds that can
be produced by combining phenols, bisphenols or glycidylamines with
crosslinking
agents including, but not limited to epichlorohydrin, aminoplasts, phenoplasts
and
isocyanates.
In some embodiments the the vinyl glues (vinyl polymers) may comprise
any one or combination of polyvinyl alcohols, polyvinyl acetates, polyvinyl
chlorides, polyacrylonitriles and polyvinyl fluorides.
In some embodiments the curable polymers comprise any one or
combination of polyethylene, polyesters, polypropylene, polycarbonates, poly-
chitosan, polyurethanes, polyimides, and polyamides.
In some embodiments the ceramics may comprise any one or combination
of alumina, beryllia, barium titanate, bismuth strontium calcium copper oxide,
boron oxide, boron nitride, ceria, ferrite, lead zirconate titanate, magnesium
diboride, porcelain, silica, silicon aluminium oxynitride, silicon carbide,
silicon
nitride, strontium titanate, strontium aluminate, titania, titanium carbide,
yttria, zinc
oxide and zirconium dioxide.
In some embodiments the step d) of producing an array of sub-nanopores
through the sealed layer comprises any one or combination of chemical etching
of
selected sites on the sealed layer, laser irradiation of selected sites on the
sealed
layer, ion bombardment of selected sites on the sealed layer, neutron
bombardment of seleated sites on the sealed layer, electron bombardment of
selected sites on the sealed layer, plasma etching of selected sites on the
sealed
layer, and UV treatments of selected sites on the sealed layer.
5
CA 2963431 2017-04-06
The method may include a step of templating the curable sealing material
onto the surface of the 2D platelets and any exposed regions of the surface of
the
microporous substrate in order to protect regions on the sealed surface
through
which the array of nanopores are not to be produced.
In an embodiment this step of templating may include controlling a size and
location of the cured sealing material particles on the layer of 2D platelets
and any
remaining exposed areas of the surface of the porous substrate.
In another embodiment this step of templating may include selecting the
microporous substrate, the 2D platelet material and the curable sealant
material to
have a preselected combination of hydrophobicity and hyrdophillicity to
control the
selectivity of pore occlusion and/or the coverage of curable sealant material
on the
2D platelet surfaces and the microporous substrate.
A further understanding of the functional and advantageous aspects of the
present disclosure can be realized by reference to the following detailed
description and drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
Embodiments disclosed herein will be more fully understood from the
following detailed description thereof taken in connection with the
accompanying
drawings, which form a part of this application, and in which:
Figures 1(a) to 1(e) shows a schematic of the present fabrication process
of pore formation in which graphene/RNA solution is vacuum filtrated onto
micro-
porous Teflon TM supports, in which:
6
CA 2963431 2017-04-06
Figure 1(a) shows the deposition of graphene flakes onto a porous
substrate using vacuum filtration;
Figure 1(b) shows the structure obtained using the fabrication process of
Figure 1(a);
Figure 1(c) silica colloid is deposited to fill the Teflon TM pores and spaces
between graphene flakes;
Figure 1(d) shows copper deposition forms a sub monolayer of Cu on the
silica/FLG surface; and
Figure 1(e) shows Cu-assisted etching creates pores in the graphene
between silica islands that penetrate the full flake thickness.
Figure 2(a) shows an SEM micrograph of a bare microporous Teflon TM
substrate.
Figure 2(b) the SEM for the same Teflon TM substrate coated with graphene
flakes obtained from vacuum filtration of 24 L m-2 FLG suspension.
Figure 2(c) shows the SEM for the same substrate as in panels Figures
2(a) and 2(b) sealed with 3 L m-2 of colloidal silica and annealed at 250 C.
The
inset shows a higher resolution image of thermally nucleated np-Si02.
Figure 2(d) shows the SEM for the same substrate shown in panels a¨c are
coated with a nominally 1 nm thick Cu layer and ready for Cu-assisted etching.
Cu
deposits both on top np-Si02 and in the inter-particle spacing in direct
contact with
graphene flakes.
Figure 3(a) shows SEM and EDX images before Cu-assisted HNO3 etching
of a complete membrane, treated with a low (1.8 L m-2) amount of colloidal
silica.
7
CA 2963431 2017-04-06
Figure 3(b) shows the same membrane in Figure 3(a) after Cu-assisted
HNO3 etching.
Figure 3(c) shows SEM and EDX images of a complete membrane treated
with a high (4.8 L m-2) amount of colloidal silica before Cu-assisted etching.
Figure 3(d) shows the same membrane as in Figure 3(c) after pore etching
has been performed. Right panels show the elemental composition (in atomic %)
from all detected elements, except oxygen that may be in the form of both
bonded
and adsorbed atoms. At low colloidal silica, the increase in fluorine after
etching
occurs due to increased sensitivity of the EDX technique to the F-containing
microporous Teflon TM substrate, due to the formation of a network of bored
nanopores in the overlaying graphene flakes.
Figure 4(a) shows flow rate vs. silica volume measured at 83 kPa
differential pressure. The bottom panel shows the difference in flow rate
after
etching. The largest increase in flow rate within experimental error is
observed for
3 L m-2 of silica colloid deposited. This volume results in an ideal layer of
homogenously distributed silica nanoparticles that also effectively seals the
inter-
flake spacing. Solid lines are visual aids.
Figure 4(b) shows flow rate vs. differential pumping pressure for two
completed membranes with different colloid volume compared to a bare Teflon TM
support. The membranes are shown to have a measureable flow rate at operating
differential pressures as low as 34 kPa.
Figures 5(a) and 5(b) show the EPR spectra for Mn2+ and Fe3+ ions
respectively in aqueous solution compared to a bare EPR tube and glass
capillary
8
CA 2963431 2017-04-06
in Figure 5(b). The peak indicated with an asterisk (*) is from the EPR sample
holder.
Figures 5(c) and 5(d) show the ion concentrations of Mn2+ and Fe3+
respectively in aqueous solution after 0 to 5 passes through the optimized
graphene-based membranes, showing retention after a single pass that increases
with multiple passes for Fe3+ ions. Performance of the same porous Teflon TM
substrates prior to coating with graphene-based membranes are also shown as a
reference.
Figures 6(a) and 6(b) show relative ion concentration of Mn2+ and Fe3+
respectively after soaking of the both bare Teflon TM support and an optimized
graphene-based membrane. Both ion types are partially retained in the
membranes after 3 h of soaking, with no further observed uptake of ions within
experimental error over a 12 hour (h) period.
Figures 7(a) to 7(g) inclusively show a cross view of the different
components used in the fabrication of the present molecular sieves at each
step of
the fabrication process.
DETAILED DESCRIPTION
Various embodiments and aspects of the disclosure will be described with
reference to details discussed below. The following description and drawings
are
illustrative of the disclosure and are not to be construed as limiting the
disclosure.
The drawings are not to scale. Numerous specific details are described to
provide
a thorough understanding of various embodiments of the present disclosure.
9
CA 2963431 2017-04-06
However, in certain instances, well-known or conventional details are not
described in order to provide a concise discussion of embodiments of the
present
disclosure.
As used herein, the terms "comprises" and "comprising" are to be construed
as being inclusive and open ended, and not exclusive. Specifically, when used
in
the specification and claims, the terms "comprises" and "comprising" and
variations
thereof mean the specified features, steps or components are included. These
terms are not to be interpreted to exclude the presence of other features,
steps or
cornponents.
As used herein, the term "exemplary" means "serving as an example,
instance, or illustration," and should not be construed as preferred or
advantageous over other configurations disclosed herein.
As used herein, the terms "about" and "approximately" are meant to cover
variations that may exist in the upper and lower limits of the ranges of
values, such
as variations in properties, parameters, and dimensions.
As used herein, the phrase "sub-nanopores" means pores having a size
range between 0.01 nm and 100 nm.
As used herein, the phrase "porous substrate" refers to a porous material
which can be permeated by fluids, including: liquids, gases, sols, gels,
aerosols
and emulsions.
As used herein the phrase "curable sealing material" refers to a sealing
material that can be solidified by a treatment such as, but not limited to,
crosslinking by use of chemical cross linking agents, thermal treatment,
drying,
CA 2963431 2017-04-06
photo-induced cross linking by exposure to light of suitable wavelengths,
irradiation
by electrons, spraying, and chemical reaction between two (or more) species or
phases.
The curable sealing material may be any material forming a colloidal
suspension of aggregates (with or without the assistance of a surfactant) in a
liquid
solvent, including but not limited to water, that can be used to fill or
occlude the
substrate pores.
Non-limiting examples of curable sealing materials include, but is not limited
to silica particles, silica gels, polyimides, polyamides, silicates,
borosilicates, epoxy
resins, vinylic glues (vinyl polymers), silicones, urethanes, and
polyurethanes.
As used herein the phrase "platelet materials" refers to a layer or layers of
two-dimensional materials which is perforated by the methods disclosed herein
and let fluids move through the sub-nanopores perforated.
As used herein, the phrase "water permeable molecular sieves
characterized by water permeability at low differential pressures" means a
layer of
material that is impenetrable to water, except for a number of specific
locations,
through which water will be able to pass if differential pressures between 0
and
100 kPa are applied at the two sides of said layer.
In its broadest ,aspect, the present disclosure provides a method producing
a water permeable molecular sieve for purifying water having metal ion
contaminants contained therein. The method involves depositing, onto a surface
of
a porous substrate having micron-size pores, non-porous 2D platelets in order
to
seal, at the surface, pores in the porous substrate to form a layer of 2D
platelets.
11
CA 2963431 2017-04-06
Since the platelets are typically small, thin and irregularly shaped, the may
coat the
surface of the substrate with one layer in places or a few layers in other
places
with the platelets overlapping in various locations on the surface. Since the
goal is
to completely seal the coated surface so that later nanopores can be produced
in
selected locations on the sealed layer, a curable sealing material is
deposited onto
the layer of 2D platelets and any remaining exposed areas of the surface of
the
porous substrate. The curable sealing material is cured in order to form a
sealed
layer on the surface of the porous substrate to prevent water by-passing the
non-
porous 2D platelets and passing through the porous substrate.
Once the porous substrate has been sealed with the sealing layer formed
by the platelets and the curable sealing material, so that water cannot pass
from
one side of the substrate to the other, an array of sub-nanopores is produced
through the sealed layer with the array of sub-nanopores having a size to
allow
water to pass therethrough but not metal ions to give a water permeable
molecular
sieve characterized by water permeability at low differential pressures.
The porous substrate having micron-size pores may include any one of
microporous TeflonTm, polytetrafluoroethylene, polycarbonate, nitrocellulose,
anodized alumina, fritted glass, plastic grids and metallic grids.
The non-porous 2D platelets may include any one of graphene platelets,
graphene oxide platelets, doped graphene platelets, functionalized graphene
platelets, boron nitride (BN) platelets, molybdenum sulphide (M0S2) platelets,
molybdenum selenide (MoSe2) platelets, carbon (C) platelets, carbon fibres,
micro
12
CA 2963431 2017-04-06
graphite platelets, nanocrystalline graphite platelets, nickel oxide (NiO)
platelets,
nickel oxide tubules, and silicon (Si) platelets and silicon whiskers.
The curable sealing material may be any one or combination of sol-gel
processed materials, epoxy resins, vinyl glues (vinyl polymers), polyurethane,
curable polymers and ceramics.
The above-mentioned sol-gel processed materials may be any one or
combination of alkoxides, silicates, acrylates, siloxanes, ormosils, silica
gels, and
sulfides.
The above-mentioned epoxy resins may be any compounds that can be
produced by combining phenols, bisphenols or glycidylamines with crosslinking
agents including, but not limited to epichlorohydrin, aminoplasts, phenoplasts
and
isocyanates.
The above-mentioned vinyl glues (vinyl polymers) may be any one or
combination of polyvinyl alcohols, polyvinyl acetates, polyvinyl chlorides,
polyacrylonitriles and polyvinyl fluorides.
The above-mentioned curable polymers may be any one or combination of
polyethylene, polyesters, polypropylene, polycarbonates, poly-chitosan,
polyurethanes, polyimides and polyamides.
The above-mentioned ceramics may comprise any one or combination of
alumina, beryllia, barium titanate, bismuth strontium calcium copper oxide,
boron
oxide, boron nitride, ceria, ferrite, lead zirconate titanate, magnesium
diboride,
porcelain, silica, silicon aluminium oxynitride, silicon carbide, silicon
nitride,
13
CA 2963431 2017-04-06
strontium titanate, strontium aluminate, titania, titanium carbide, yttria,
zinc oxide
and zirconium dioxide.
The step of producing an array of sub-nanopores through the sealed layer
comprises any one or combination of chemical etching (e.g. metal assisted) of
selected sites on the sealed layer, laser irradiation of selected sites on the
sealed
layer, ion bombardment of selected sites on the sealed layer, neutron
bombardment of selected sites on the sealed layer, electron bombardment of
selected sites on the sealed layer, plasma etching of selected sites on the
sealed
layer, and UV treatments of selected sites on the sealed layer.
The example below uses metal assisted chemical etching using
submonolayers of copper metal and it is discussed in detail. However, as noted
immediately above, other techniques may be used.
The general procedure for preparing the nanoporous membranes or sieves
is shown in Figures 7(a) to 7(h). Figure 7(a) illustrates a side view of a
microporous substrate 8 with micropores 9 extending therethrough, while Figure
7(b) shows a side view of the substrate 8 but with 2D platelets 11 covering or
occluding pores 12 while pores 9 remain unsealed. Figure 7(c) shows a side
view
of the porous substrate 8 after the curable sealing material has been
deposited on
top of the platelets 11 and as well the curable sealing material is deposited
onto
exposed areas on the surface of the porous material not covered by the
platelets
11 thus sealing the uncovered pores 9. The sealant, once cured, forms
particles 15
on the top surface of the platelets 11 to give covered areas 16 on the
platelets 11,
while any exposed pores 9 are filled by cured sealant particles 14. The cured
14
CA 2963431 2017-04-06
particles 15 are useful in that they prevent the covered areas 16 from being
perforated, thus particles 15 act as a template or pattering agent to direct
where
the nanopores are not located. Thus during the process of producing the
nanopores, uncovered areas 17 on the top of the platelets 11 are the areas in
which the nanopores may be produced. Figure 7(d) shows a side view of the
nanoporous sieve after production of the nanopores 19 in preselected areas
through the platelets 11.
Figure 7(e) shows a top view looking down towards the top surface of the
porous membrane showing the exposed areas 17 of the top surface of platelets
11
and the cured sealant particles 15 which act to protect the areas below them
from
being perforated, and also act as a template so that it is only exposed areas
17
that are able to be perforated.
Figure 7(f) shows a side view of one embodiment of a completed molecular
sieve in which porous substrate 25 is hydrophobic, platelets 23 are
hydrophilic and
sealant/templating particles 24 are hydrophilic. Figure 7(g) shows an
alternative
embodiment in which substrate 25 is hydrophobic, platelets 21 are hydrophobic
and sealant particles 24 are hydrophilic so that fewer sealant particles 24
cover the
platelets 21 compared to the embodiment in Figure 7(h). Having different
degrees
of hydrophobicity/hydrophilicity of the platelets, the substrate and the
sealant may
be used to produce sealed surfaces with different proportions of the sealant
located on the platelet and in the substrate pores which enables the pore-
filling
process to be selective.
CA 2963431 2017-04-06
In this aspect of the present method, selecting the microporous substrate,
the 2D platelet material and the curable sealant material to have a
preselected
combination of hydrophobicity and hydrophilicity provides control over the
coverage of the surface by the curable sealant material by making sure that
the
curable sealant material does not completely adhere on the surface of
perforable
platelets and does not prevent the subsequent perforation processes. Instead,
the
curable sealant material will completely adhere on the microporous substrate
and
occlude the apertures between the platelets. In this way, a fluid-tight
membrane, in
which the fluid to be filtrated will only pass through the pores that will be
perforated
on each platelet, is fabricated.
The 20 platelets can range in diameter from 10 nm to 1000 pm and in
thickness from about 0.1 nm to about 1 pm.
The cured sealer materials may have a size or diameter in a range from
about mm to about 100 pm.
Transferring the platelets onto the surface of the porous substrate may be
carried out using many techniques, including, but not limited to, vacuum
deposition, chemical deposition, transfer printing, roll-to-roll printing,
spin coating,
spray coating, dip coating and painting to mention a few. Similarly, the
curable
sealing material may be deposited using the above-mentioned techniques.
During deposition of the curable sealing material various techniques may be
used to control the locations of where the cured particles are located. Such
techniques include, but are not limited to, using a mask to cover areas it is
not
desired to coat while leaving exposed areas to be coated. The uncured liquid
16
CA 2963431 2017-04-06
sealing material may also be printed onto selected locations on the surface
using
various techniques.
EXAMPLE
The present method will be illustrated using the following non-limiting
example in which the porous substrate is platelets are graphene platelets, the
porous substrates were microporous Teflon TM substrates, the curable sealing
material was colloidal silica (Si02), and the process for producing the
nanopores in
the sealed layer was chemical etching achieved by depositing a sub monolayer
of
copper (Cu) on the top of the sealed layer and then employing Cu-assisted
chemical etching at the Cu sites thereby producing the nanopores. Thus, in
this
embodiment the present disclosure provides a new method and system for
producing graphene-based water purification membranes that combine the
advantages of porous, single-layer graphene filters, energy-efficient
filtration at
relatively low differential pressures, with the benefits of graphene oxide
membranes-high performance in terms of adsorption and filtration of
impurities.
The present purification method and system is based on weakly oxidized, few-
layer graphene flakes that can be produced in large amounts from graphite
using
surfactant-assisted exfoliation. The purification membranes disclosed herein
are
fabricated using a three-step method: (i) non-porous graphene flakes are
deposited on top of a highly porous substrate with micron-size holes; (ii)
spacing
between the flakes and the remaining open holes on the substrate are
completely
occluded by sintered colloidal silica to prevent water bypassing the graphene
17
CA 2963431 2017-04-06
flakes; and (iii) (sub)nanopores are drilled through weakly oxidized and
silica-free
inter-particle regions of graphene flakes, restoring high water permeability
at low
differential pressures, compared to nearly zero flow rate during step ii. The
use of
colloidal silica for both sealing material and masking for perforation enables
the
patterns of (sub)nanopores to be selected. Moreover, fully penetrating pores
generated by the present etching process are decisively advantageous over
interlayer spacing in GO membranes for increasing the water flow rate (see
references 9, 16). The present design avoids the requirements of high
differential
pressure typical of GO membranes due to their long water flow path length (see
reference 3).
With the use of the curable sealing material as a mask as well, the present
chemical etching process can be altered by other perforation methods such as
ion,
electron, neutron bombardment and plasma etching. In the alternative methods,
ions, electrons or neutrons are accelerated towards target and collide to
damage it.
A silica nanoparticle mask can sufficiently protect graphene flakes from any
damage while silica-free inter-particle regions of the graphene flakes are
physically
drilled through by the ion, electron or neutron beams in the methods mentioned
above.
Methods
Weakly oxidized FLG flakes or platelets were prepared from nano-
crystalline graphite (Sigma Aldrich, CAS 7782-42-5) using two treatment steps
that
were previously described by Sharifi et al. (see reference 18). The first
treatment
was conducted with nitric and sulfuric acid (HNO3:1-12SO4 1:3 volume ratio) to
18
CA 2963431 2017-04-06
promote dilation of graphite into multilayer flakes. An acid mixture with 200
mg L-1
of nano-crystalline graphite was treated for 24 hours using ultrasonication
(Branson, DHA-1000 ultrasonicator). After sonication, the slurry was diluted
to 10
times the original volume with distilled water, and filtrated onto a track-
etched
polycarbonate membrane (Millipore, HTTP0903, pore diameter 0.4 pm). The
membrane was then allowed to dry for 24 hours at room temperature and used in
the oxidizing treatment that follows. The second treatment utilized piranha
solution
(H202:H2SO41:4 volume ratio) to functionalize the graphite into hydrophilic
flakes
that can be dispersed in water.
Piranha solution with 30.4 mg L-1 of graphite from the previous exfoliation
step were mixed for 30 min. The resulting graphite Piranha solution was
diluted 10
times by volume with distilled water and vacuum filtrated onto the same type
of
polycarbonate membranes as in the previous step, and again dried for 24 hours.
In
order to prepare suspensions of weakly oxidized few layer graphene in water,
60
mg L-1 of weakly oxidized graphite flakes prepared as above, were dispersed
into
a 600 mg L1 watersolution of RNAVI (Sigma Aldrich, CAS 63231-63-0) from
torula utilis yeast. This RNA-based, surfactant-assisted exfoliation method of
graphite has been extensively discussed in Sharifi et al. (see reference 18).
The suspension is ultra-sonicated for 4 hours at room temperature with
RNA acting as a surfactant. The resulting suspension was stored at 2 C for 24
hours in a beaker to allow a sediment to precipitate, which consists of non-
exfoliated graphite. The top three quarters of the beaker was centrifuged at
6000
rpm for 1 hour (Fisher Scientific, Accuspin TM 400 centrifuge) and the
supernatant
19
CA 2963431 2017-04-06
was used for membrane preparation. To prepare the water purification devices,
40
mL of supernatant was deposited by vacuum filtration onto porous (5 pm average
pore size) 25 mm Teflon TM support discs (Chromatographic Specialties Inc.,
006FR2505).
Membranes were then annealed at 250 C for 10 min in air to dry and for 2
hours in nitrogen atmosphere, a treatment that in conjunction with the
following
strong acid process used to open the pores, is known to almost completely
eliminate RNA from our FLG graphene thin films.18 Variable amounts of
colloidal
silica (Syton HT-50, CAS 7631-86-9) were then filtrated through on Teflon TM
1.0 substrates. Amounts of silica colloid ranging from 1.8 L m-2 and 5.4 L
m-2 with 0.6
L m-2 steps were used to obtain a set of filtration membranes with different
properties. The colloidal silica treatment was followed by a second annealing
step
in a glove box (Nexus II, Vacuum Atmosphere Inc.) at 250 C for 2 hours under a
nitrogen atmosphere with less than 10 ppm of oxygen and moisture.
Thermal evaporation of copper was performed on our membranes in an
ultra-high vacuum chamber (base pressure 10-6 Pa) using Cu pellets (99.99%
purity, K.J. Lesker Inc.). The vacuum chamber is connected to the glove box to
prevent oxidation during sample loading. Pellets were positioned in alumina
crucibles (K.J. Lesker cat. no. EVA9A0) that were supported by tungsten basket
heaters (K.J. Lesker Inc., cat. no. EVB8B3030W). Heat was delivered to this
assembly at 7.5 V and 65 A by means of a Hewlett Packard 64660 DC power
supply and the thickness of the growing film was measured using a calibrated
quartz crystal oscillator connected to a Sycom STM-2 thickness monitor. A
nominal
CA 2963431 2017-04-06
thickness of 1.0 0.3 nm Cu was deposited at a chamber pressure of 2.5x10-6
mbar at a rate of 0.5 A s-1, with the sample stage temperature maintained
below
¨60 C. Immediately following copper deposition, metal-assisted etching was
performed in a 70% nitric acid bath (70%, Caledon, CAS 7697-37-2) for 10 min,
followed by cleaning in a distilled water bath 3 consecutive times for 10 min
each
time.
SEM imaging was performed during all of the fabrication steps as discussed
below. SEM and EDX measurements were performed using a Zeiss LEO 1540XB
system. The surface conductivity of all samples was increased by depositing ¨1
nm of osmium prior to SEM and EDX measurements using a Filgen OPC8OT Os
plasma coater.
After Os coating, graphene-based membranes were mounted onto flat
sample holders using double sided carbon tape. Silver paint (Pelco Colloidal
Silver Liquid) was applied on the edge of the sample to the sample holder to
improve conductivity. Mounted samples were transferred to the vacuum chamber
attached to the instrument and pumped down to a pressure lower than 1.0 x 10-6
Torr. SEM morphology images were obtained by detecting secondary electrons at
1 kV with a magnification of 5 k and 10 k. EDX was collected at 10 kV with a
magnification of 1 k and 5 k. The flow rate through our graphene-based
membranes was characterized using a 25 mm vacuum filtration apparatus
(Wheaton, cat. no. 419327). The apparatus consisted of a 15 mL funnel pressed
to
a 25 mm glass support base and held together with an anodized aluminium clamp,
with the membrane mounted between the funnel and support base.
21
CA 2963431 2017-04-06
The glass support base was fed through a no. 5 silicone stopper that was
placed in the top opening of a 250 mL flask with a #2 side hose connection
that
was connected to a variable pressure vacuum source. The graphene-based
membranes were mounted between the funnel and glass support base in a unique
Teflon TM holder that was machined in house to fit both the glass support base
and
the 25 mm Teflon TM substrates tightly. Additionally a Teflon TM gasket
(Chromatographic Specialties Inc., N419338) was added between the Teflon TM
holder and Teflon TM substrate on the bottom of the graphene-based membrane to
prevent water from leaking around the membrane instead of passing through it.
With a membrane mounted in the filtration apparatus, flow rate tests were
conducted by drawing 25 mL of distilled water through the membranes at a
differential pressure in the 10 to 90 kPa range.
The pressure was measured using an inline vacuum gauge (Innova Equus
3620). EPR measurements were performed using a Jeol FA-200 EPR
spectrometer operating in the X-band at 9.1 MHz at room temperature. 10 mg L-1
solutions of MnCl2 (Sigma Aldrich, 99%, CAS 7773-01-5) and FeCl3 (Sigma
Aldrich, 97%, CAS 7705-08-0) were prepared by dissolving powder samples in
distilled water and mixing for 20 minutes by ultra-sonication. Metal ion
solutions
were measured in liquid form by inserting a fixed amount into a glass
capillary tube
with 1.0 mm inner diameter (4 pL for MnCl2 and 15 pL for FeCl3) using a micro
pipette. The filled capillary was then placed in a quartz glass EPR tube of 5
mm
inner diameter (Wilmad LabGlass, 710SQ-250M) and inserted in the microwave
cavity. EPR measurements were performed by first tuning the microwave cavity
22
CA 2963431 2017-04-06
using a sample of unfiltered metal ion solution as a reference, and keeping
the
tuning parameters fixed for all further measurements of filtered metal ion
solutions.
Electron paramagnetic resonance (EPR) is used herein as a highly selective
analytical technique to probe low metal ion concentrations in water, before
and
after purification through graphene-based membranes. EPR is capable of
detecting paramagnetic centres in very low amounts, on the order of parts-per-
billion. As a trace detection technique, EPR has significant advantages in
terms of
selectivity and sensitivity over common detection methods, including ion
conductivity measurements, (see reference 4). The EPR spectrum of each metal
ion has specific fingerprints, and each paramagnetic impurity can be
independently
monitored. Using EPR, we are able to show that our graphene-based membranes
retain metal ions through two cooperating mechanisms: size-based rejection and
sorption on the graphene surfaces or at the pore edges.
The parameters used for EPR measurements of Mn2+ were: 8 min sweep
time scanning from 275 to 375 mT, receiver gain of 200, 100 kHz modulation
frequency, 1.0 mT modulation width, 1.0 s time constant, 2 mW microwave power,
and microwave phase and coupling set to 339 and 441, respectively. For Fe3+
the
sweep time was 8 min scanning from 285 to 365 mT with a receiver gain of
10000,
a modulation frequency of 100 kHz with a modulation width of 1.4 mT, 3.0 s
time
constant, 5 mW microwave power, and microwave phase and coupling set to 709
and 441, respectively. The concentration of metal ions was determined by
integrating the EPR signal over the external magnetic field and calculating
the
integral area. The error in EPR measurements was quantitatively determined as
23
CA 2963431 2017-04-06
the difference between the integrated signal areas of subsequent, identical
measurements of the same sample.
Membrane fabrication
The entire fabrication process of our water purification devices is depicted
in
Figures 1(a) to 1(e). Purification membranes produced in accordance with
present
disclosure incorporate weakly-oxidized flakes of few-layer graphene (FLG), as
opposed of highly oxidized GO flakes (see reference 17), with the flakes made
nanoporous in post-deposition treatments. FLG flakes were prepared using
surfactant-assisted exfoliation of graphite in water, a method originally
developed
by Sharifi et al. (see reference 18). This method utilizes specific types of
ribonucleic acid (RNA) as efficient surfactants for graphite exfoliation, as
shown in
Figure 1(a). RNA of this type can be cultured in large amounts at low costs,
comparable to the costs of graphite starting material. Thus, the cost of RNA
has
very limited impact on the overall fabrication costs, and the method is
scalable for
producing large amounts of water-based, few-layer graphene suspensions (see
reference 18).
Substrates for the present graphene thin film deposition are microporous
Teflon TM disks through which aqueous suspensions of FLG are vacuum-filtrated
as
shown in Figure 1(a), to produce the structure shown in Figure 1(b).
Micropores in
the Teflon TM disks have diameters (-5 pm) that are smaller than the size of
FLG
flakes (-10 to 30 pm). Thus, filtration of suitable amounts of suspension
resulted in
an almost complete coverage of the substrate micropores, which was signalled
by
a strong (>90%) decrease of filtration rate. About 24 L m.-2 of FLG suspension
was
24
CA 2963431 2017-04-06
necessary for high substrate coverage by graphene flakes. However, even at
such
high coverage, water might leak through the membrane via the spacing between
partially overlapped graphene flakes juxtaposed to the substrate micropores.
Such
leaks need to be completely eliminated to make graphene thin films water-tight
such that water passes through nanopores to be drilled in graphene flakes
during
subsequent fabrication steps, and not through the larger inter-flake spacing
between partially overlapping graphene domains.
Scanning electron micrographs (SEM) of Teflon TM substrates before and
after coverage with not-yet nanoporous FLG flakes are shown in Figure 2(a) and
2(b), respectively. Comparison between the two Figure 2(a) and 2(b) shows that
substrate coverage by graphene is almost complete, with partially overlapped
FLG
flakes that are larger than the Teflon TM micropores, but with some spaces
still
remaining between the flakes as shown schematically in Figure 1(b). In order
to
make our graphene layer water-tight, colloidal silica was vacuum-filtrated
through
the membrane to fill any spaces between FLG flakes and occupy any micropores
that remained open in the Teflon TM support after graphene deposition. Samples
were subsequently annealed at 250 C to promote sintering of silica colloid,
resulting in the morphology seen in Figure 2(c).
Comparison of Figures 2(b) and 2(c) of demonstrates that not only has
silica cemented the interstitial spacing between partially overlapping
graphene
flakes, but also led to partial coverage of graphene by Si02 nanoparticles (np-
Si02)
that have thermally nucleated during annealing. As detailed in the inset of
Figure
2(c), a uniform distribution of np-Si02 of less than 50 nm in diameter is
formed on
CA 2963431 2017-04-06
top of our graphene flakes due to their hydrophobicity (see reference 19).
Inter-
particle spacing is often less than 10 nm, which indicates extensive graphene
coverage by np-Si02. As indicated in Figures 1(c) and 1(d), application of
colloidal
silica on top of our graphene thin films plays three critical roles in the
fabrication of
our devices. First, the sintered silica occludes the residual micropores that
remained open on the Teflon TM substrate after graphene deposition. Secondly,
the
silica seals the interflake spacing between partially overlapping graphene
domains.
When combined, these two functions prevent any leaks around the edge of
graphene flakes which are, in fact, water-tight. Thirdly, in addition, silica
nanoparticles nucleated on graphene form a mask that can be used for location
selective etching of the underlying graphene, with the formation of nanopores
only
on flake regions that are not coated by silica. This means the density of
resultant
(sub)nanopores can be controlled by using different size and different amounts
of
np-Si02, which changes np-Si02-free areas on graphene where (sub)nanopores
which are fabricated in the later process.
In order to make the graphene flakes nanoporous and to restore high water
permeability through the present membranes, metal-assisted etching by nitric
acid
was used, a technique that was originally proposed by Ramasse et al.
(reference
20) for graphene oxide and was here adapted for weakly oxidized graphene.
Metal
assisted etching is based on metal nanoparticles (e.g. copper) that oxidize in
the
proximity of graphene oxide due to the Bronsted-type properties of HNO3 see
reference 3).
26
CA 2963431 2017-04-06
It has been proposed that metal oxidation in the presence of GO promotes a
certain degree of additional oxidation of carbon, breaking some C-C bonds in
oxidized graphene with the formation of more C-0 and C¨OH groups (see
references 15, 20). Pores locally form in correspondence to such terminations,
where C-C bonds are broken. In our specific case, in which the preceding
oxidation of the carbon backbone is localized and weak, pores are anticipated
to
have very small (at the nanometre or even sub-nanometre scale) diameter.
Therefore, they may be suitable to filtrate metal ion impurities in water via
size-
based rejection.
Figure 2(d) shows the SEM image of one of the membranes after thermal
evaporation of nominally 1 nm thick copper layer on top. From this image, it
can be
observed that Cu deposits both on top of np-Si02 and silica-free inter-
particle
regions in which metal atoms are in direct contact with graphene. This
conformation of Cu deposition is demonstrated in Figure 1(d), in which Cu is
in
direct contact with FLG flakes, it assists HNO3 in etching locally oxidized
areas on
the graphene backbone, resulting in the formation of small pores over short
periods of time. As can be observed in Figure 3(b) to 3(d) there are a lot
more
pores in Figure 3(b) compared to Figure 3(d) so that these pores are too small
to
be resolved with SEM, indicating their size should be in nanometer range or
below.
Nanopore size in the present metal-assisted etching process is determined by
size
of the pristine oxidized centers in our FLG flakes, which is less than the
inter-
particle distances between np-Si02, not by the size of Cu patterns in contact
with
graphene.
27
CA 2963431 2017-04-06
The presence of (sub)nanometer pores in our FLG flakes, and the specific
conditions in which they form, could be inferred using energy-dispersive X-ray
spectroscopy (EDX). EDX images are shown in Figures 3(a) to 3(d) in
conjunction
with SEM scans of two distinct membranes. These membranes were treated with
small (Figures 3(a) and 3(b)) and large (Figures 3(c) and 3(d)) amounts of
silica
colloid, followed by annealing and etching in HNO3 under identical conditions.
When a relatively small amount of silica colloid was used as in Figure 3(a),
the
graphene flake surface was only partially covered with np-Si02. Under these
treatment conditions, a strong increase in the fluorine EDX signal originating
from
the underlying TeflonT,M substrate could be observed (Figure 3(b)) after Cu-
assisted membrane etching in HNO3. The subsequent increase in fluorine-to-
carbon ratio is clear evidence of decreased areal C atom concentration in
graphene flakes after etching, and shows their higher porosity. Conversely,
when a
larger amount of silica colloid was used to seal the membranes as in Figure
3(c),
the graphene flake surface was almost completely covered by np-Si02.
Although EDX mapping showed that the elemental composition of this
membrane prior to Cu-assisted HNO3 etching was not dissimilar from that shown
in
Figure 3(a), EDX in panel d shows no significant increase in the fluorine-to-
carbon
ratio after membrane etching. This suggests that no nanopores were formed in
membranes treated with excess silica colloid, and points to an optimal amount
of
colloid required to seal the membranes and simultaneously form a sparse
template
for pore etching. Unless otherwise stated, the characterization of filtration
28
CA 2963431 2017-04-06
performance that follows refers to devices treated with optimal amount of
silica
colloid (3 L per m2 of membrane area).
Testing flow rate and differential pressure
The flow rate of graphene-based membranes before and after nanopore
formation was tested as a function of the volume of silica used to seal the
spacing
between flakes in the graphene thin films and form a template for nanopore
fabrication. During tests in a dead-end filtration configuration, our
optimized
membrane demonstrated a flow rate of ¨20 L m-2 min-1, comparable to most
graphene oxide and commercial counterparts, (see references 1 to 7) but
operating at a differential pressure of Ap = 83 kPa which is significantly
lower than
differential pressures '(700-1500 kPa) (see reference 5) normally used in
state-of
the art RO purification systems. This represents a significant breakthrough
over
existing technologies in terms of their energy footprint because the energy
required
for water purification is proportional to Ap.
Figure 4(a) shows the water flow rate before nanopore fabrication by metal-
assisted HNO3 etching. Prior to making the graphene flakes nanoporous, the
flow
rate was approximately constant (-10 L m-2 min-1) when colloidal silica
treatment
of the membranes was performed with less than 3 L m-2. Between 3.6 L m-2 and
5.4 L m-2 of colloidal silica the flow rate monotonically decreases with
increasing
volume of silica colloid, trending to zero at the largest volumes at which the
membranes are completely water-tight. The monotonically decreasing trend of
water flow rate can be understood by considering that at low amounts of
colloidal
silica our devices were imperfectly sealed, which makes it impossible to
ensure
29
CA 2963431 2017-04-06
water passage through nanopores even though they have been drilled in the
graphene. Water continues to flow in large proportions through inter-flake
spacing
that are much larger and more frequent than nanopores in graphene flakes,
leading to less than optimal retention of impurities from water. Sealing of
the inter-
flake spacing becomes increasingly more efficient at increasing amounts of
colloidal silica used in the membrane treatment.
Figure 4(a) also shows the flow rates after metal-assisted HNO3 etching of
the devices. In this case corresponding to completed filtration membranes, the
water flow rate is no longer a monotonic function of the amount of colloid
used in
the sealing treatment. The water flow increases at low amounts of colloidal
silica
and undergoes a maximum at about 3 L m-2, at which the water flow rate doubles
with respect to the same membrane prior to pore nanofabrication. For amounts
of
silica colloid greater than 3 L m-2, metal-assisted HNO3 etching of the
membranes
does not lead to any significant gains in the filtration rate. These three
regimes can
be explained as follows: at low amounts of silica colloid, below 3L m-2,
nanopores
in graphene were fabricated by metal-assisted HNO3 etching, but the flow rate
was
not increased as dramatically because of the significantly larger openings
still
present in the form of inter-flake spacing that were imperfectly sealed by
colloidal
silica: water mostly passes through these spacing.
Conversely, at high amounts of silica colloid, above 3 L m-2, few or no
nanopores can be etched in graphene because the flakes are almost completely
coated by np-Si02. Consequently, copper does not adhere to the graphene
surface
in this case, which leads to insufficient pore fabrication. The case of
treatment with
CA 2963431 2017-04-06
3 L m-2 of silica colloid is found to be ideal to both seal most of the inter-
flake
spacing, as well as form a nanostructured silica mask on the surface to guide
metal-assisted HNO3 etching of pores.
The performance curve of the present optimized membranes, expressing
the water flow rate as a function of applied differential pressure, is shown
in Figure
4(b). Significant water flow occurs at differential pressures as low as 30 to
40 kPa
in completed membranes, which corroborates their very low energy footprint.
Bare
Teflon TM substrates show a much higher flow rate at any differential pressure
and
therefore do not limit the flow rate through the graphene membrane in which
most
of the water purification occurs. Figure 4(a) also shows that water flow
through
ultrathin graphene layers is negligibly small prior to pore opening (upper
panel),
but allows for a flow rate of 23 L m-2 min-1 at 83 kPa after pore fabrication
which is
comparable to ultrathin graphene oxide membranes, but at a lower operating
differential pressure (see reference 3). It is known that metal-assisted HNO3
etching can produce bored nano- or subnano-pores up to at least twenty (20)
graphene layers (see reference 20). This fact combined with the large increase
in
fluorine EDX signal after etching observed in Figure 3 suggests that large
amounts of very small but fully penetrating pores have been drilled. We thus
speculate that operation at very low differential pressure is made possible by
path
lengths shorter than those in GO membranes, through nanopores that penetrate
the entire thickness of FLG, such that water flow is not restricted to
interlayer
spacing as in the case of GO laminates (see references 6, 16).
Water purification from Mn2+ and Fe3+ ions
31
CA 2963431 2017-04-06
In studies of water purification from Mn2+ and Fe3+ ions disclosed herein,
EPR has been used to directly probe the concentration of paramagnetic metal
ions
in water prior to and after purification with membranes. The intensity of the
EPR
signal is proportional to the first derivative of microwave absorption by the
paramagnetic metal ion system under examination, due to optical transitions
between unpaired spin states parallel and antiparallel with respect to an
external
dc magnetic field (see reference 21). Therefore the EPR spectrum integrated
over
the external dc magnetic field is proportional to the number of paramagnetic
metal
ions in our water samples both before and after filtration. Fe3+ and Mn2+ ions
are
paramagnetic, stable in water, and produce detectable EPR signals at room
temperature (see reference 22). Both ion types exhibit EPR spectra with a six-
fold
peak structure, resulting from overall electron configurations with electron
spin S =
5/2. Although Fe2+ may also be present in water in limited amounts, it forms a
quadruplet spin complex (S = 2) that is only weakly paramagnetic, and
therefore is
not detectable by EPR at room temperature. Fe3+ and Mn2+ in part-per-million
concentrations were used to test the effectiveness of the present purification
process, and the difference in EPR intensity before and after purification is
a good
measure of our membrane effectiveness to independently retain these two ion
types.
Fe3+ and Mn2+ ions in solution were obtained from MnCl2 and FeCl3
dissolved in water. Both feed solutions were passed through our devices up to
five
times to investigate the effectiveness of multiple purification steps on the
same
solution. Multiple-pass filtration was also useful to gain insight into the
nature of the
32
CA 2963431 2017-04-06
dominant ion retention mechanisms of our graphene-based membranes. The
mechanisms can be either (i) pore diameter-based rejection, due to sub-
nanopores
comparable in size to the size of metal ions and their hydration shells, or
(ii) ion
adsorption, due to capture at nanopore edges. Although fluorine EDX maps in
Figures 3(a) to 3(d) show clear evidence of nanoporosity in the graphene
flakes,
with pore size below the detection limit of SEM (i.e. <5 nm), no direct
estimate of
the pore diameter could be obtained. Thus, an indirect estimate by analysis of
the
filtration process is imperative in our case. A decrease in ion concentration
after a
single pass is more likely due to filtration by pore diameter-based rejection,
while
increasing ion removal in subsequent passes is more likely attributable to
adsorption at nanopore edges.
Figures 5(a) and 5(b) show the EPR spectra of 10 mg L-1 aqueous feed
solutions of Mn2+ and Fe3+, respectively. The panels compare the spectra
measured before purification with our optimized graphene-based membranes and
after five consecutive passes. For both types of ions, Figures 5(c) and 5(d)
show
that the performance of our porous graphene membranes are superior to bare
Teflon TM supports. Furthermore, for both Mn2+ and Fe3+ ions, some degree of
retention is observed even after a single pass through the membrane, with 25
4% of Mn2+ and 20 6% of Fe3+ of ions trapped. Multiple filtration passes
show
that additional fractions of Fe3+ ions could be trapped, and their
concentration in
water can be reduced to 45 6% of the pristine value.
Compared to the retention of increasing amounts of Fe3+ ions with
increasing number of passes, water purification from Mn2+ ions using our
33
CA 2963431 2017-04-06
membranes (Figure 5(c)) shows a trend of slightly increased concentration
after
the first pass. This may be explained by considering two distinct mechanisms
of
absorption and pore-based rejection that remove ions from solution. Metal ions
may be either absorbed through ionic bonding or trapped sterically within
membrane pores, as disclosed in reference 5. Considering these mechanisms and
the trend in Figure 5(c), the membrane is saturated to Mn2+ absorption after
one
filtration pass, including a fraction of ions trapped in the membrane and
Teflon TM
support. The fraction of ions weakly trapped may then be circulated into the
permeate solution during additional filtration passes. From this result it can
be
inferred that water purification from Mn2+ ions occurs via a combination of
pore
diameter-based rejection and absorption. In contrast, the concentration Fe3+
ions
decreases non-reversibly with multiple filtration passes through a graphene-
based
membrane, indicating Fe3+ ions are retained more readily by ion adsorption at
nanopore edges.
In order to further discriminate from the two proposed mechanisms of pore-
based rejection and absorption of metal ions, our graphene-based membranes
have been immersed for up to 12 h in 10 mg L-1 in aqueous solutions of Mn2+
and
Fe3+. Clearly, no pore-size based ion retention can occur during the immersion
process, while ion adsorption at the membrane surface, or at the pore edge,
may
still take place. The changes of metal ion concentrations in the water bath
over
time are shown in Figures 6(a) and 6(b) for Mn2+ and Fe3+ ions, respectively.
The
concentration of Mn2+ ions reaches an asymptote at 90 4% of the pristine
fraction
after the first three (3) h of immersion. This indicates relatively low
adsorption of
34
CA 2963431 2017-04-06
Mn2+ ions. Conversely, the concentration of Fe3+ ions continues to decrease
until
70 4% at 9 hour (h). This indicates that absorption of Fe3+ ions at the
membrane
surface is more prevalent than absorption of Mn2+ ions. Consequently, data
shown
in Figures 6(a) and 6(b) shows that a greater fraction of Fe3+ can be absorbed
compared to Mn2+. This result supports the finding from Figure 5(c) that
recirculation of Mn2+ between the graphene-based membrane and ion solution
occurs due to Mn2+ ions retention through pore diameter-based rejection, with
weak adhesion of these ions at the surface. In contrast, Fe3+ ions do not show
evidence of recirculation (Figure 6(b)), and remain more steadily absorbed in
the
membrane over time.
The different mechanisms of water purification from these two types of
metal ions can be understood by considering that Mn2+ has an approximately 10%
larger hydration shell as a consequence of greater screening by water
molecules
and a larger ionic radius. This explains why Mn2+ concentration is more easily
decreased than Fe3+ by pore diameter-based rejection after only one
filtration. By
contrast Fe3+ may be more easily captured by electrostatic adsorption due to
reduced screening between the metal ion at centre of the hydration shell and
oxygen functional groups in pore edges of weakly oxidized FLG. The EPR results
agree with this hypothesis in which multiple filtrations can remove more Fe3+
compared to Mn2+, retaining up to 55 6% of the Fe3+ ion concentration in the
membranes. Considering these two mechanisms of purification, EPR results
suggest that the typical sub-nanopore diameter in our graphene-based
CA 2963431 2017-04-06
membranes is in between the hydration shell diameters of Fe3+ (0.40 nm) and
Mn2+
(0.44 nm) (see reference 23).
Conclusions
In conclusion, an exemplary non limiting example of a molecular sieve has
been produced and characterized herein. More particularly, graphene-based
water
purification membranes have been assembled on Teflon TM substrates with large
porosity using a fabrication process that is highly scalable and low in both
complexity and cost. These membranes have been assembled using nanoporous
few-layer graphene platelets that combine the advantages of porous single-
layer
graphene, offering energy efficient water filtration at relatively low
differential
pressures, and highly oxidized graphene oxide, which offers high performance
in
terms of adsorption and impurity filtration. A step in the fabrication process
utilized
colloidal silica to seal the spacing between partially overlapped graphene
flakes to
produce the water-tight membranes prior to drilling pores in few-layer
graphene. In
this way, purification occurs by water passing through the nanopores which
were
fabricated by Cu-assisted etching in nitric acid. This unique design was used
to
demonstrate water purification from metal ions through a combination of sub-
nanopore filtration in the case of Mn2+ ions, or adsorption at pore edges in
the case
of Fe3+ ions that possess a hydration shell diameter smaller than Mn2+ ions.
A significant competitive advantage of our graphene-based membranes that
makes them uniquely positioned to solve the problem of high energy costs
associated with water filtration by RO is the low differential operating
pressure, in
the range of 80-90 kPa, at which high flow rates (-20 L m-2 m1n-1) can be
36
CA 2963431 2017-04-06
obtained. This is significantly lower than the pressures (700-1500 kPa) used
in
state-of-the-art commercial filtration systems based on reverse osmosis, and
corresponds to a large reduction in the energy footprint for filter operation.
By
careful control over the pore drilling conditions allows pores of a fixed size
and
even distribution to be fabricated reproducibly. In this regime, filters
exhibiting pore
filtration can be used to remove any ionic contaminant from aqueous solution,
with
an efficacy that depends on the size of the hydration shell for both size
based
rejection and adsorption. With these properties our porous graphene-based
membranes could have applications in removing contaminants from water at a
much lower cost compared to conventional methods such as reverse osmosis
membrane filtration.
The foregoing description of the preferred embodiments of the invention has
been presented to illustrate the principles of the invention and not to limit
the
invention to the particular embodiment illustrated. It is intended that the
scope of
the invention be defined by all of the embodiments encompassed within the
following claims and their equivalents.
37
CA 2963431 2017-04-06
References
1. S. L. Loo, A. G. Fane, W. B. Krantz and T. T. Lim, Water Res., 2012, 46,
3125.
2. H. Li, L. Zoua, L. Pan and Z. Sun, Sep. Purif. Technol., 2010, 75, 8.
3. H. M. Hegab and L. Zou, J. Membr. Sc., 2015, 484, 95.
4. R. K. Joshi, P. Carbone, F. C. Wang, V. G. Kravets, Y. Su, I. V.
Grigorieva,
H. A. Wu, A. K Geim and R. R. Nair, Science, 2014, 343, 752.
5. A. Aghigh, V. Alizadeh, H. Y. Wong, M. S. Islam, N. Amin and M. Zaman,
Desalination, 2015, 365, 389.
6. S. C. O'Hern, M. S. H. Boutilier, J.C. Idrobo, Y. Song, J. Kong, T.
Laoui, M.
Atieh and R. Karnik, ACS Nano, 2013, 7, 428.
7. G. Z. Kyzas, E. A. Deliyanni and K. A. Matis, J. Chem. Technol.
Biotechnol., 2014, 89, 196.
8. H. Wang, X. Yuan, Y. Wu, H. Huang, X. Peng, G. Zeng, H. Zhong, J. Liang
and M. Ren, Adv. Colloid Interface Sci., 2013, 195-196, 19-40.
9. N. Wei, X. Peng and Z. Xu, ACS App!. Mater. Interfaces, 2014, 6, 5877.
10. S. C. O'Hern, M. S. H. Boutilier, J. C. ldrobo, Y. Song, J. Kong, T.
Laoui, M.
Atieh and R. Karnik, Nano Lett., 2014, 14, 1234.
11. S. P. Koenig, L. Wang, J. Pellegrino and J. S. Bunch, Nat.
Nanotechnol.,
2012, 7, 728.
12. S. P. Surwade, S. N. Smirnov, I. V. Vlassiouk, R. R. Unocic, G. M.
Veith, S.
Dai and S. M. Mahurin, Nat. Nanotechnol., 2015, 10, 459.
13. K. He, A. W. Robertson, C. Gong, C. S. Allen, Q. Xu, H. Zandbergen, J.
C.
Grossman, A. I. Kirklanda and J. H. Warner, Nanoscale, 2015, 7, 11602.
14. D. Zhou, Y. Cui, P.-W. Xiao, M.-Y. Jiang and B.-H. Han, Nat. Commun.,
2014, 2, 1.
15. H. Cao, X. Zhou, C. Zheng and Z. Liu, Carbon, 2015, 89,41.
16. R. K. Joshi, S. Alwarappan, M. Yoshimura, V. Sahajwalla and Y. Nishina,
App!. Mater. Today, 2015, 1, 1.
38
CA 2963431 2017-04-06
17. G. Eda, G. Fanchini and M. Chhowalla, Nat. Nanotechnol., 2008, 3, 270.
18. F. Sharifi, R. Bauld, M. S. Ahmed and G. Fanchini, Small, 2012, 8, 699.
19. X. Zhang, S. Wan, J. Pu, L. Wang and X. Liu, J. Mater. Chem., 2011, 21,
12251.
20. Q. M. Ramasse, R. Zan, U. Bangert, D. W. Boukhvalov, Y.-W. Son and K.
S. Novoselov, ACS Nano, 2012, 6, 4063.
21. C. P. Poole, Electron Spin Resonance: A Comprehensive Treatise on
Experimental Techniques, Dover Publications, Inc., Mineola, New York,
USA, 1996. '
22. S. K. Misra, S. Diehl, D. Tipikin and J. H. Freed, J. Magn. Reson.,
2010,
205, 14.
23. I. Persson, Pure App!. Chem., 2010, 82, 1901.
,
39
