Thomas Vidil1, Nicholas Hampu2, Marc A Hillmyer1. 1. Department of Chemistry, University of Minnesota, Minneapolis, Minnesota 55455, United States. 2. Department of Chemical Engineering and Materials Science, University of Minnesota, Minneapolis, Minnesota 55455, United States.
Abstract
A lamellar diblock polymer combining a cross-linkable segment with a chemically etchable segment was cross-linked above its order-disorder temperature (TODT) to kinetically trap the morphology associated with the fluctuating disordered state. After removal of the etchable block, evaluation of the resulting porous thermoset allows for an unprecedented experimental characterization of the trapped disordered phase. Through a combination of small-angle X-ray scattering, nitrogen sorption, scanning electron microscopy, and electron tomography experiments we demonstrate that the nanoporous structure exhibits a narrow pore size distribution and a high surface to volume ratio and is bicontinuous over a large sample area. Together with the processability of the polymeric starting material, the proposed system combines attractive attributes for many advanced applications. In particular, it was used to design new composite membranes for the ultrafiltration of water.
A lamellar diblock polymer combining a cross-linkable segment with a chemically etchable segment was cross-linked above its order-disorder temperature (TODT) to kinetically trap the morphology associated with the fluctuating disordered state. After removal of the etchable block, evaluation of the resulting porous thermoset allows for an unprecedented experimental characterization of the trapped disordered phase. Through a combination of small-angle X-ray scattering, nitrogen sorption, scanning electron microscopy, and electron tomography experiments we demonstrate that the nanoporous structure exhibits a narrow pore size distribution and a high surface to volume ratio and is bicontinuous over a large sample area. Together with the processability of the polymeric starting material, the proposed system combines attractive attributes for many advanced applications. In particular, it was used to design new composite membranes for the ultrafiltration of water.
Polymeric materials with nanoscopic
interpenetrating and percolating domains that exhibit high surface
to volume ratios are very useful for separation,[1] catalysis,[2] and energy technologies.[3] A rich variety of synthetic strategies to these
materials have been developed with the aim of combining two discrete
phases that form stable and interconnected domains.[4,5] Block
polymers have been extensively used for this purpose[6,7] as they provide self-assembled nanostructures with well-defined
geometries and domain sizes that can be readily controlled by tuning
the molar mass and the chemical composition.[5,8] However,
commonly observed equilibrium morphologies typically either require
additional postsynthetic processing to align the domains (cylinders
and lamellae)[9−11] or exist over a narrow range of composition (double
gyroid).[12] As a result, numerous reports
have outlined strategies to kinetically trap nonequilibrium, bicontinuous
morphologies.[13−15] In particular, microphase separation in cross-linkable
systems involving polymeric precursors (e.g., polymerization induced
microphase separation, PIMS,[13] or randomly
end-linked copolymer networks, RECNs[15])
has been successfully employed for the straightforward preparation
of large-area samples with cocontinuous nanodomains. A fine balance
between the kinetics of the phase separation process and the cross-linking
reaction ensures trapping of a nonequilibrium morphology comprised
of interconnected domains. The bicontinuity of these morphologies
has been correlated to their global disorder.[16]We now report a new strategy for the preparation of robust
bicontinuous
nanostructured materials by exploiting the well-known order–disorder
transition (ODT) of block polymers.[17] Previous
studies indicate that this transition is characterized by strong,
thermally induced local composition fluctuations that disrupt the
long-range order of the system.[18−21] As the order–disorder transition temperature
(TODT) is approached from the disordered
state, the fluctuations become large in amplitude and spatially correlated
over short distances, resulting in the persistence of local microphase
segregation despite the absence of global long-range order. Some view
the ODT as a “pattern transition” resulting in only
small changes in chain stretching and interfacial area, despite a
loss in long-range order.[22,23] Early works have also
postulated that the resulting structure is bicontinuous, with a morphology
that is topologically similar to other disordered, microphase separated
systems.[17,19,23] In two important
experimental studies, the persistence of microphase separation above TODT was demonstrated by the direct imaging of
the morphology of reactiveblock polymers cross-linked in the disordered
state.[24,25] However, there have been very few attempts
to characterize the three-dimensional connectivity of these domains
or exploit this interesting state in nanotechnology with the important
exception of Balsara et al., who demonstrated that the conductivity
of a lamellar diblock polymer loaded with a lithium salt increases
in the vicinity of TODT, most likely as
a consequence of the interconnectivity of the conductive domains induced
by composition fluctuations.[26] Consequently,
there is a strong motivation to further investigate the morphological
development in the vicinity of the TODT so that the ODT could be leveraged for the design, discovery, and
development of new materials.Our aim is to utilize the order–disorder
transition of reactiveblock polymers to prepare mesoporous materials with well-defined and
continuous three-dimensional network structures (Figure ). This approach involves heating
a reactivediblock polymer above its TODT in the presence of a thermally latent initiator that can trigger
the cross-linking reaction for curing temperatures Tcuring ∼ TODT, thus
kinetically trapping the disordered state. The concurrent use of a
chemically etchable block allows for the subsequent removal of one
of the nanodomains, affording a nanoporous structure that can be easily
characterized under ambient conditions. This method represents a new
and versatile strategy for the design of mesoporous materials that
retains all of the processability advantages of polymer-based systems.
To demonstrate a potential application of this work, we prepared composite
membranes for the ultrafiltration of water.
Figure 1
Strategy employed to
obtain a bicontinuous nanoporous structure
by chemical cross-linking of a diblock polymer in the disordered state.
The chemical structures of the components are represented at the top.
(a) Lamellar diblock copolymer with an etchable block (blue) and a
cross-linkable block (green) containing a thermally latent cross-linking
initiator. (b) This mixture is heated above its order–disorder
transition temperature (TODT) to adopt
a disordered bicontinuous morphology. (c) The thermolatent initiator
triggers the cross-linking reaction above TODT. (d) Once the cross-linkable phase is cured (darker green), the
etchable block (blue) is removed to obtain well-defined percolating
nanopores.
Strategy employed to
obtain a bicontinuous nanoporous structure
by chemical cross-linking of a diblock polymer in the disordered state.
The chemical structures of the components are represented at the top.
(a) Lamellar diblock copolymer with an etchable block (blue) and a
cross-linkable block (green) containing a thermally latent cross-linking
initiator. (b) This mixture is heated above its order–disorder
transition temperature (TODT) to adopt
a disordered bicontinuous morphology. (c) The thermolatent initiator
triggers the cross-linking reaction above TODT. (d) Once the cross-linkable phase is cured (darker green), the
etchable block (blue) is removed to obtain well-defined percolating
nanopores.We synthesized a block polymer
containing a chemically etchable
polylactide (PLA) and a cross-linkable block consisting of a statistical
copolymer of styrene (S) and glycidyl methacrylate (GMA) [PLA-b-P(S-s-GMA)] (Figure , top). The synthetic route starts with a
functionalized PLA precursor obtained by ring opening polymerization
of ±-lactide initiated by an hydroxyl-functionalized chain transfer
agent (CTA) for reversible addition–fragmentation chain transfer
(RAFT) polymerization (see the Supporting Information for synthetic details and supporting characterization data). A series
of near symmetric PLA-b-P(S-s-GMA)
samples with controlled molar masses and mole fractions of GMA in
the cross-linkable block as well as low values of dispersity (Đ ∼ 1.1) were obtained (Figures S4–S6 and Table S2). The diblocks are labeled
as PLA-α-P(S-s-GMA)-β-XGMA-γ where α and β are the molar masses (in kg mol–1) of the PLA and P(S-s-GMA) blocks,
respectively, and γ is the molar percentage of GMA in the P(S-s-GMA) block (XGMA). Two glass transition temperatures, Tg, were observed by differential scanning calorimetry
(Tg(PLA) ∼ 55 °C and Tg(P(S-s-GMA) ∼ 70–80
°C, Figure S10, Table S3), implicating
microphase separation.Figure a shows
the temperature dependence of the low-frequency dynamic elastic shear
modulus (G′) for PLA-17-P(S-s-GMA)-11-XGMA-29. The TODT, 174 ± 5 °C, was identified as the onset of the discontinuous
drop in G′ resulting in a liquid-like response
(see the Supporting Information for details, Figure S11). The linear viscoelastic responses
of the melt at low frequency (0.01 ≤ ω ≤ 100 rad
s–1) for two temperatures above and below TODT are also shown in Figure a (insets). For T = 160
°C (T < TODT),
the moduli obey a G′ ∼ G″ ∼ ω1/2 power law typically observed
for lamellar ordering.[18] Above TODT at 190 °C the moduli scale as G′ ∼ ω1.5 and G″ ∼ ω1. The frequency dependence exponent
of G′ (1.5) is indicative of composition fluctuations
in disordered phases close to the ordering transition[20,27] where the longest relaxation time is shifted to lower frequencies
due to the collective motion of the segregated domains in the frequency
range of the experiment (G′ ∼ ω2.0 would be expected at even lower frequencies).
Figure 2
(a) Temperature
dependence of the low-frequency dynamic storage
modulus (G′) for PLA-17-P(S-s-GMA)-11-XGMA-29. Ramp rate = 1 °C min–1 (strain, ε = 1%, frequency, ω = 1 rad s–1). Insets: isothermal frequency sweeps (frequency, 0.01 ≤
ω ≤ 100 rad s–1) acquired at 160 and
190 °C. (b) SAXS patterns obtained for PLA-17-P(S-s-GMA)-11-XGMA-29. The sample was annealed 2 min at each
temperature. The TODT was determined as
174 ± 1 °C through more precise SAXS measurements (Figure S12). For T < TODT, higher ordering peaks are marked by inverse
triangles and are consistent with a lamellar morphology.
(a) Temperature
dependence of the low-frequency dynamic storage
modulus (G′) for PLA-17-P(S-s-GMA)-11-XGMA-29. Ramp rate = 1 °C min–1 (strain, ε = 1%, frequency, ω = 1 rad s–1). Insets: isothermal frequency sweeps (frequency, 0.01 ≤
ω ≤ 100 rad s–1) acquired at 160 and
190 °C. (b) SAXS patterns obtained for PLA-17-P(S-s-GMA)-11-XGMA-29. The sample was annealed 2 min at each
temperature. The TODT was determined as
174 ± 1 °C through more precise SAXS measurements (Figure S12). For T < TODT, higher ordering peaks are marked by inverse
triangles and are consistent with a lamellar morphology.Figure b illustrates
representative small-angle scattering (SAXS) data for PLA-17-P(S-s-GMA)-11-XGMA-29 upon heating from 25 to 220
°C. The SAXS pattern acquired at 25 °C exhibits a principal
scattering peak q* = 0.23 nm–1 and
higher order peaks 2q* and 3q* that
persist up to 160 °C and are consistent with a lamellar ordering
and a domain spacing d = 27 nm. Above 160 °C,
the principal scattering peak broadens and a single broad reflection
is evident at 180 °C. More precise SAXS measurements give TODT = 174 ± 1 °C (Figure S12), in agreement with TODT determined by rheological measurements. In a series of related block
polymers we demonstrated that TODT is
a decreasing function of the molar fraction of GMA in the cross-linkable
block, XGMA. This trend is consistent with the expectation
that inclusion of GMA in the PS block increases its solubility parameter,
making it more compatible with PLA (Table S4). We also showed that the tunable nature of TODT in this new block polymer system allows for independent
tuning of the cross-linking temperature and nanodomain size (Figures S13 and S14).Benzyl triphenylphosphonium
hexafluoroantimonate (BTPH, Figure top, the synthetic
procedure is available in the Supporting Information, Figures S15–S17) was used to
initiate the cationic polymerization of the pendant epoxide moieties.[28] BTPH was selected as the catalyst as its reaction
kinetics are appropriate for the temperature window around the TODT of PLA-17-P(S-s-GMA)-11-XGMA-29
and its high efficiency allows for low catalyst loadings. A film was
prepared by casting a THF solution of PLA-17-P(S-s-GMA)-11-XGMA-29 (24.7 wt %) and BTPH (0.3 wt %) (dried
for 1 d at room temperature and 1 d at 60 °C). The film was annealed
at 110 °C for 1 h, and complete removal of THF was confirmed
gravimetrically. SEC and SAXS analysis of the initiator-loaded system
indicated that the molar mass, dispersity, and morphology of the block
polymer were not impacted by these low-temperature drying procedures
(Figures S18 and S19), and independent 1H NMR analysis suggested that BTPH was located in both the
P(S-s-GMA) and PLA blocks prior to cross-linking.
Moreover, variable-temperature SAXS experiments show that the TODT of PLA-17-P(S-s-GMA)-11-XGMA-29+BTPH(0.3 wt %) (TODT = 178
°C) is only 4 °C higher than the TODT of the sample without initiator (TODT = 174 °C) (Figures S20 and S21, Table S5). The TODT of the initiator-loaded
polymer (178 °C) is used as the reference TODT.Initiator-loaded samples were then cured at various
temperatures
in the proximity of TODT of PLA-17-P(S-s-GMA)-11-XGMA-29+BTPH(0.3 wt %) (see detailed
procedure in the Supporting Information). Samples were cured below TODT, Tcuring = 160 or 170 °C, and above TODT, Tcuring = 180,
190, 200, 210, or 220 °C. Curing times (tcuring) were chosen on the basis of estimations of the gel
times (tgel) through independent rheological
measurements (Figures S22 and S23, Table S6; typically we utilized tcuring >
3tgel). Thermogravimetric analysis was
used to
demonstrate that the polymer did not lose significant mass during
curing (Figure S24), and in all cases the
gel fraction of the cured samples was >90% (Table S7).The cured samples were first characterized by SAXS
experiments
at room temperature. For Tcuring = 160
°C i.e., 18 °C below TODT (ΔT = Tcuring – TODT = −18 °C), the room temperature
SAXS pattern exhibits a single sharp scattering peak (q*) and two higher order peaks (2q* and 3q*), indicating that a lamellar microstructure is kinetically
trapped during the curing step (Figure a, solid line). At Tcuring = 190 °C i.e., 12 °C above TODT (ΔT = +12 °C), a single broad reflection
peak is consistent with a microphase-separated, but disorganized,
structure (Figure b, solid line). These room temperature SAXS data are consistent with
the SAXS data for the initiator-free system acquired at temperatures
corresponding to the curing conditions (see Figure b).
Figure 3
Characterization of the structure of the materials
obtained by
curing PLA-17-P(S-s-GMA)-11-XGMA-29+BTPH(0.3
wt %) (TODT = 178 °C) at two different
temperatures: Tcuring = 160 °C (tcuring = 4 h) (a and b) and Tcuring = 190 °C (tcuring = 1 h) (c and d). (a) SAXS pattern acquired at room temperature
for the sample cured at 160 °C before (solid line) and after
(dashed line) etching in a basic solution. (b) Representative SEM
image of a cryofractured surface coated with Pt (∼2 nm) for
the sample cured at 160 °C after etching in a basic solution.
(c, d) These panels represent the corresponding SAXS patterns and
SEM image for the sample cross-linked at Tcuring = 190 °C.
Characterization of the structure of the materials
obtained by
curing PLA-17-P(S-s-GMA)-11-XGMA-29+BTPH(0.3
wt %) (TODT = 178 °C) at two different
temperatures: Tcuring = 160 °C (tcuring = 4 h) (a and b) and Tcuring = 190 °C (tcuring = 1 h) (c and d). (a) SAXS pattern acquired at room temperature
for the sample cured at 160 °C before (solid line) and after
(dashed line) etching in a basic solution. (b) Representative SEM
image of a cryofractured surface coated with Pt (∼2 nm) for
the sample cured at 160 °C after etching in a basic solution.
(c, d) These panels represent the corresponding SAXS patterns and
SEM image for the sample cross-linked at Tcuring = 190 °C.For the sample cross-linked
in the ordered state at 160 °C,
variable temperature SAXS experiments show that the microstructure
remains ordered at temperatures as high as 250 °C, i.e., higher
than the TODT of the original non-cross-linked
system (Figure S25). The sample cross-linked
in the disordered state is unable to recover lamellar ordering post-cross-linking
at temperatures below the TODT of the
unreacted system. Even after thermal annealing at 120 °C (i.e.,
well below the original TODT and above
the highest Tg of the pristine diblock)
for 15 h (Figure S26) the sample did not
recover the lamellar microstructure. Together, these observations
confirm that the efficiency of the cross-linking reaction is sufficient
to kinetically and irreversibly trap the microstructure adopted at
the curing temperature. In both cases the domain spacing is only slightly
affected by the curing reaction (within 8% of spacing prior to curing, Figure b and Table S8). For most other curing temperatures
the SAXS results correlate well with the observations made for Tcuring = 160 °C when ΔT < 0 and for Tcuring = 190 °C
when ΔT > 0 (see Figure S27 and Table S8). A notable exception is the sample cured
at 180 °C, where the SAXS pattern exhibits a sharp principal
scattering peak despite a curing temperature above the TODT of the non-cross-linked system. This unexpected behavior
is attributed to its close proximity to the TODT (ΔT = +2 °C), and further discussion
of this case is in the Supporting Information (Figure S27B).The PLA was subsequently
removed in the cured samples by immersing
the monoliths in a 0.5 M methanol (40% by volume)/water solution of
NaOH. Complete removal of PLA was confirmed by gravimetric analysis
with a mass loss in close agreement with the weight fraction of PLA
in the block polymer (Table S9), the absence
of the PLA C=O vibration by IR spectroscopy (Figure S28), and the absence of a glass transition corresponding
to PLA in the DSC thermogram of the etched materials (Figure S29). The removal of PLA was accompanied
by a dramatic change of the SAXS pattern of the samples cross-linked
below TODT (ΔT <
0) with a total loss of the scattering characteristic of the lamellar
ordering, likely due to the collapsing of the nanoporous structure
(Tcuring = 160 °C, ΔT = −18 °C, Figure a, dashed line, and Tcuring = 170 °C, ΔT = −8
°C, Figure S27a), consistent with
previous reports of partial collapsing in nanoporous lamellar microstructure.[29,30] Conversely, for all the samples cross-linked above 180 °C,
the SAXS patterns retain a broad reflection post etching consistent
with a disordered structure (Tcuring =
190 °C, ΔT = +12 °C, Figure c, dashed line, and Tcuring = 200 °C, 210 °C, and 220 °C,
ΔT = +22 °C, +32 °C, and +42 °C, Figure S27c–e) and the domain spacing
is within 9–14% of the spacing prior to etching (Table S10). The dramatic difference in stability
for the porous microstructures obtained from samples cross-linked
above and below TODT is likely due to
a difference in terms of the continuity of their porous network. Indeed,
while the etched lamellar structures result in large 2-dimensional
sheets that are only separated by voided domains and eventually collapse,[29] bicontinuous network structures (e.g., gyroid[31] or samples from PIMS,[13] RCENs,[16] and bicontinuous microemulsions[32]) are known for providing three-dimensional structures
that percolate the entire sample and are stable provided that their
thermal and mechanical properties are appropriate. Thus, the SAXS
results for the etched material shown in Figure c are consistent with a cured sample containing
bicontinuous domains: one consisting of mesopores and one of the cross-linked
P(S-s-GMA) polymer.Scanning electron microscopy
(SEM) observations of cryofractured
and coated (Pt, ∼2 nm) surfaces of the samples are consistent
with the results obtained from SAXS experiments. In agreement with
the lamellar ordering indicated by the SAXS patterns in Figure a, sheet-like and evidently
collapsed layered objects are observed in Figure b for the sample cured at Tcuring = 160 °C (ΔT = −18
°C). The SEM image is also consistent with the lack of strong
scattering in Figure a for the etched sample. Similar images were obtained at Tcuring = 170 °C (see Figure S30a). For the sample cured at Tcuring= 190 °C (ΔT = +15 °C)
the SEM micrograph clearly reveals a nanoporous structure with interconnected
pores homogeneously distributed throughout the material. Similar images
were obtained for other curing temperatures above the ODT (see Figure S30c–e for Tcuring = 200, 210, and 220 °C). This structure closely
resembles the morphologies of other polymeric materials with percolating
nanoporous network obtained from PIMS[13] and RECNs[16] or using polymeric bicontinuous
microemulsions.[32] This supports well the
idea that the morphology of the disordered state of a symmetric diblock
polymer is truly bicontinuous.Transmission electron microscopy
(TEM) of a microsection of porous
sample cured at 190 °C is also consistent with a homogeneous
disordered network of pores (Figure S32). The bicontinuity of the nanostructure was further assessed using
TEM tomography.[33] A tilt series of TEM
micrographs of the same sample was collected for a layer-by-layer
reconstruction of the 3D structure (Figure S33 and Movie S1). The tomogram reveals a
network of highly branched channels that traverse the thickness of
the sample and supports the idea of a bicontinuous network of pores.
A computational reconstruction of the volume also supports highly
interconnected porous domains (Figure S34).Nitrogen sorption was further used to quantitatively analyze
the
porosity. As expected from their collapsed lamellar structures, the
samples cross-linked below TODT (ΔT < 0 °C) exhibit essentially featureless nitrogen
adsorption isotherms (Figure S35a,b), indicating
that the materials are not porous. For all the curing temperatures
above TODT (ΔT >
0 °C), the materials produced type IV isotherms with H2 hysteresis
(Figure S35c–g). This again supports
the idea that the porous structures associated with composition fluctuations
in the disordered state are bicontinuous. For Tcuring ranging from 190 to 220 °C (12 °C ≤
ΔT ≤ 42 °C), the pore size distributions,
modeled using the adsorption branch of the isotherms and a quenched
solid density functional theory kernel (QSDFT),[34] are monomodal with a sharp peak centered on 10 nm (Figure S36). This correlates well with the values
of the domain spacing determined by SAXS d ∼
22 nm and the overall volume fraction of the sacrificial PLA domains
of 54% (Table S4). Nanoporous samples prepared
using a block polymer with a smaller PLA block, PLA-10- P(S-s-GMA)-10-XGMA-14+BTPH(0.3 wt %) (Figures S37–S42), gave pore diameters
of about 7–8 nm, indicating that the pore diameter can be controlled
by the molar mass of the PLA block.Figure represents
the pore volume determined at P/P0 = 0.95 as well as the surface area estimated by a Brunauer–Emmett–Teller
(BET) analysis as a function of ΔT = Tcuring – TODT, i.e., the relative distance between the curing conditions and the
ODT in °C.[35] Maxima in both sets of
data are reached for Tcuring = 190 °C
with a total pore volume of 0.47 cm3 g–1 and an estimated surface area of 190 m2 g–1. These values are essentially the same at higher curing temperature.
The material cured at 180 °C exhibits a low total pore volume
of 0.22 cm3 g–1 and a surface area of
90 m2 g–1. Again, this is attributed
to the close proximity to the TODT (ΔT = +2 °C, see Figures S35c and S36c).
Figure 4
Plot of the pore volume (filled squares) and the surface
area (empty
squares) as a function of ΔT = Tcuring – TODT, for
PLA-17-P(S-s-GMA)-11-XGMA-29+BTPH(0.3
wt %) cross-linked at seven different temperatures (Tcuring = 160, 170, 180, 190, 200, 210, and 220 °C).
The cured samples are subsequently etched in a basic solution and
characterized by nitrogen sorption experiments. The pore volumes are
determined at P/P0 =
0.95, and the surface areas are estimated on the basis of a Brunauer–Emmett–Teller
(BET) analysis. Error bars represent the typical range of the data
obtained from three independent measurements for the sample cross-linked
at 190 °C.
Plot of the pore volume (filled squares) and the surface
area (empty
squares) as a function of ΔT = Tcuring – TODT, for
PLA-17-P(S-s-GMA)-11-XGMA-29+BTPH(0.3
wt %) cross-linked at seven different temperatures (Tcuring = 160, 170, 180, 190, 200, 210, and 220 °C).
The cured samples are subsequently etched in a basic solution and
characterized by nitrogen sorption experiments. The pore volumes are
determined at P/P0 =
0.95, and the surface areas are estimated on the basis of a Brunauer–Emmett–Teller
(BET) analysis. Error bars represent the typical range of the data
obtained from three independent measurements for the sample cross-linked
at 190 °C.The system described
in this report provides nanoporous material
with narrow pore size distribution while retaining all of the processability
advantages associated with polymeric materials. In particular, it
can be solution processed to fabricate nanoporous thin film membranes
that have the potential to combine both high permeability and high
size selectivity, two appealing attributes for ultrafiltration applications.[36] To demonstrate this, an 8 wt % solution of PLA-17-P(S-s-GMA)-11-XGMA-29+BTPH(0.3 wt %) in chlorobenzene
was coated onto a water-filled poly(ether sulfone) (PES) membrane
using a wire-wound rod.[37] The resulting
film was dried and cured above TODT (Tcuring = 190 °C, Figure top, see the Supporting Information for experimental details). After selective etching
of PLA, SEM imaging of the composite membrane demonstrates that the
cross-linked polymer forms a ∼500 nm homogeneous layer (Figure S43) that effectively covers the pores
of the PES support (Figures a and 5b). In Figure b the nanopores from the block polymer coating
can clearly be seen within the larger support pores.
Figure 5
Characterization of the
composite membrane obtained by coating
an 8 wt % solution of PLA-17-P(S-s-GMA)-11-XGMA-29+BTPH(0.3
wt %) (TODT = 178 °C) in chlorobenzene
onto a PES support membrane followed by curing at Tcuring = 190 °C and etching with a basic solution.
The strategy employed for the design of the membrane and the rejection
experiments is represented at the top. (a) SEM micrograph of the bare
PES support membrane and (b) SEM micrograph of the nanoporous selective
layer obtained after coating the PES support with PLA-17-P(S-s-GMA)-11-XGMA-29+BTPH(0.3 wt %) followed by curing at Tcuring = 190 °C and PLA etching. (c) UV–vis
absorbance results for the feed solution (solid line) of fluorescent
TRITC-dextran (0.5 mg mL–1, M =
150 kg mol–1, Rh ∼
7 nm) and the filtrate (dashed line) obtained after passing the feed
solution through the composite membrane. Rejection was calculated
as 98% based on the ratio of the absorbance of the feed solution to
the absorption of the filtrate at 521 nm. Photographs of the feed
solution (1) and the filtrate (2) are included in the inset.
Characterization of the
composite membrane obtained by coating
an 8 wt % solution of PLA-17-P(S-s-GMA)-11-XGMA-29+BTPH(0.3
wt %) (TODT = 178 °C) in chlorobenzene
onto a PES support membrane followed by curing at Tcuring = 190 °C and etching with a basic solution.
The strategy employed for the design of the membrane and the rejection
experiments is represented at the top. (a) SEM micrograph of the bare
PES support membrane and (b) SEM micrograph of the nanoporous selective
layer obtained after coating the PES support with PLA-17-P(S-s-GMA)-11-XGMA-29+BTPH(0.3 wt %) followed by curing at Tcuring = 190 °C and PLA etching. (c) UV–vis
absorbance results for the feed solution (solid line) of fluorescent
TRITC-dextran (0.5 mg mL–1, M =
150 kg mol–1, Rh ∼
7 nm) and the filtrate (dashed line) obtained after passing the feed
solution through the composite membrane. Rejection was calculated
as 98% based on the ratio of the absorbance of the feed solution to
the absorption of the filtrate at 521 nm. Photographs of the feed
solution (1) and the filtrate (2) are included in the inset.Pure water permeability of the
composite membrane was determined
to be 7 L m–2 h–1 bar–1 as compared to 3700 L m–2 h–1 bar–1 for the bare PES support (∼100 nm
nominal pore size) based on a linear fit of flux vs ΔP (Figures S44 and S45b). The possibility to use
the membrane for water ultrafiltration was first investigated by measuring
the rejection of a solute with a hydrodynamic radius, Rh, greater than the dimensions expected for the pores
of the selective layer. On the basis of the average pore radius measured
by N2 sorption for porous monoliths obtained from the corresponding
polymer (∼5 nm), we used a 0.5 mg mL–1 solution
of TRITC-dextran (Mw = 155 kDa, Rh ∼ 7 nm), a polysaccharide labeled with
a fluorescent dye.[38] UV–vis analysis
of the filtrate passed through the membrane demonstrates that 98%
of the solute was rejected (Figure c). The size selectivity of the membrane was further
investigated by using a mixed feed of dextran standards with molar
masses ranging from 5 kg mol–1 to 410 kg mol–1.[39] Aqueous SEC of the
filtrate indicates that the membrane displayed a sharp molecular weight
cutoff around 50 kg mol–1, i.e., Rh ∼5 nm (Figures S45 and S46). Again, this is consistent with the average pore diameter expected
for the selective layer. The same polymer was also spin coated onto
the PES membrane to provide a thinner selective layer (150 nm, Figure S47), resulting in a significantly higher
permeability (196 L m–2 h–1 bar–1, Figure S48) and only
a slight decrease of the TRITC-dextran rejection (96%, Figure S49).In this work we demonstrate
that the ODT of a diblock polymer comprised
of a cross-linkable block and an etchable block can be used to design
new mesoporous materials with narrow pore size distributions and high
surface to volume ratios. The characterization of the porous structure
provides an unprecedented set of data confirming that the morphology
of the disordered state is microphase separated and bicontinuous over
a large sample area. The solution processability of the proposed system
was demonstrated by directly coating the cross-linkable diblock on
top of a commercial PES membrane. After curing and etching, the resulting
composite membrane exhibits high permeability and sharp molecular
weight cutoff that are suited for the ultrafiltration of water. Together
with the chemical versatility and tunability of block polymers (e.g.,
using triblocks and other architectures) and the variety of cross-linking
reactions available in thermoset technology, this new strategy may
have utility for a large spectrum of advanced applications.
Authors: William A Phillip; Mark Amendt; Brandon O'Neill; Liang Chen; Marc A Hillmyer; Edward L Cussler Journal: ACS Appl Mater Interfaces Date: 2009-02 Impact factor: 9.229
Authors: Li Li; Lars Schulte; Lydia D Clausen; Kristian M Hansen; Gunnar E Jonsson; Sokol Ndoni Journal: ACS Nano Date: 2011-09-14 Impact factor: 15.881
Authors: Carlos M Portela; A Vidyasagar; Sebastian Krödel; Tamara Weissenbach; Daryl W Yee; Julia R Greer; Dennis M Kochmann Journal: Proc Natl Acad Sci U S A Date: 2020-03-04 Impact factor: 11.205