Angewandte
Communications
Chemie
reconstruction, as anticipated. Furthermore, the rate of
surface reconstruction increased as the length of the
PDMAEMA block increased. This trend was reasonable, as
the collapsed PDMS layer initially at the top of the coating
had a certain thickness. As m decreased, the PDMAEMA
chains would be buried more deeply and more time was
required for them to break through the PDMS layer. Addi-
tionally, short PDMAEMA chains needed to stretch them-
selves to reach the coating surface, which was not favored
entropically.
In another experiment, we placed a series of P1-b-A -
m
2
coated fabric swatches (1.5 ꢁ 1.5 cm ) into vials containing
2
0 mL of a hexadecane (HD)-in-water emulsion that was
stabilized by sodium dodecyl sulfate at a HD volume fraction
fHD of 20%. Figure 1b shows a photograph of the emulsions
1
5 min after the addition of the P1-b-A -coated fabric
m
swatches. The extent of HD phase separation increased as
m increased. De-emulsification would have been possible
only if the PDMAEMA chains reached the coating surface.
Thus, these results suggested that the fabrics in contact with
the water-rich emulsion could also surface-reconstruct.
Another image was taken (Supporting Information, Fig-
ure S7) 19 h after fabric addition into the vials shown in
Figure 1b. No clear HD phase was observed in the vial
Figure 2. a) Photographs of a separation experiment immediately and
5 min after 20 mL of a dyed fHD =20% emulsion was added into the
right cell of the H-shaped device. b) Separation data for fabrics coated
with P1-b-Am at m=4, 18, 23, and 42. c) Data for three separations in
a series of seven consecutive trials using a P1-b-A -coated fabric.
containing a P1-b-A -coated swatch even then. This indicated
4
that this coating had limited de-emulsification ability.
1
8
In a real separation, the surface reconstruction on the
emulsion-contacting side should be followed by local de-
emulsification and the filling of the pores on the un-
reconstructed side by the coagulated oil, yielding an overall
Janus fabric. To determine if this Janus structure could be
sustained, we dispensed a water droplet on a HD-impreg-
nated P1-b-A - or P1-b-A105-coated fabric and then waited.
d) Data for HD separation from 300 mL of an fHD =1.00% emulsion.
e) Separation data for emulsions containing toluene and chloroben-
zene as the oily phase.
P1-b-A -coated fabric could not fully separate HD because it
4
had poor de-emulsification ability (Supporting Information,
1
8
Figure 1c compares photographs of droplets that were taken
Figure S7). The P1-b-A -, P1-b-A - and P1-b-A -coated
1
8
23
42
immediately and 80 min after their dispensing on a P1-b-A -
swatches allowed the full separation of HD (4.0 mL, theoret-
ically) because they sustained the desired Janus structure.
Since the separation was fastest with the P1-b-A -coated
1
8
coated and HD-impregnated fabric. The droplets were
essentially stable, exhibiting a minor contact angle decrease
from 88 ꢀ 28 to 82 ꢀ 28. This was in contrast with the behavior
1
8
fabric, only this fabric was further investigated. To probe its
robustness, we used the same swatch with an effective
on a P1-b-A1 -coated and HD-impregnated fabric, into which
05
2
a water droplet was absorbed within 2 s. Thus, the Janus
separation area of only 1.13 cm to perform seven consecutive
structure could be sustained only by the P1-b-A -coated
fabric but not by the P1-b-A105-coated fabric.
For oil separation, the two halves of an H-shaped cell were
separated by a coated fabric (Figure 2a). In one experiment,
separations. For each separation, the fabric was exposed to
the emulsion for about 15 min before the liquids were
removed from the feed and permeate sides and another
20 mL emulsion at fHD = 20% was added to the feed side.
Figure 2c compares data gathered from separations 1, 3, and
7. Even after seven trials, the filterꢀs performance did not
degrade. This demonstrated the robustness of the fabric.
To probe the separation limit for HD, we covered one
mouth of a three-neck round-bottom flask with a coated
swatch and then tilted the flask so that HD could be separated
from 300 mL of a fHD = 1.00% emulsion (Supporting Infor-
mation, Figure S1). The separated HD volume reached
2.7 mL by 1.5 h and 2.9 mL by 3 h (Figure 2d). The latter
value was close or identical to, within experimental error, the
3.0 mL of HD contained in the feed emulsion. Thus, the
residual HD in the final 297 mL emulsion was less than
0.1 mL, and the fabric had an impressive HD separation limit
of 0.034% (which is the same as the measurement error of the
instrument).
1
8
2
0 mL of an fHD = 20% emulsion was then added to the right
half-cell. Under vigorous stirring, the emulsion bombarded
the fabric and de-emulsification and separation started.
Figure 2b shows the increase over time of the volume of
collected HD in the left half-cell when fabric swatches coated
by P1-b-A with m = 4, 18, 23, and 42 were used. No data are
m
shown for the P1-b-A - and P1-b-A105-coated fabrics because
0
no permeation occurred with the former and both de-
emulsified HD and the emulsion permeated the latter.
Nothing permeated the P1-b-A -coated fabric because the
0
stable emulsion droplets, which had an average diameter of
3
.7 ꢀ 0.6 mm (Supporting Information, Figure S8) and were
surrounded by water, appeared as water to the fabric.
Emulsions also permeated the P1-b-A105-coated swatches
because such a fabric could not sustain a Janus structure. A
Angew. Chem. Int. Ed. 2016, 55, 1 – 5
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
These are not the final page numbers!