surface can be facilitated by using super-
hydrophobic surfaces that display the so-
called“jumping droplet”effect, leading to
higher HTC and CHF values when com-
pared to hydrophobic surfaces ( 45 , 47 ). How-
ever, superhydrophobic textures can suffer
from condensate flooding at high super-
saturations (S> 1.54) ( 45 , 48 ). Hydrophobic,
lubricant–infused surfaces can similarly in-
crease droplet mobility, and thereby conden-
sation heat transport ( 49 ). Recent work has
also used hydrophobic, oil-infused, asymmetric,
bumpy surfaces for increasing condensation
efficiency ( 50 ). Hydrophilic lubricant–infused
surfaces can simultaneously enhance liquid
nucleation rates (thereby increasing HTC)
and droplet shedding (thereby increasing
CHF) ( 51 ). However, the lubricant on such
surfaces can be depleted over the long term
with continuous condensation. To avoid such
durability issues, tethered, hydrophilic, liquid-
like, brush surfaces with lowDqhave also
recently been developed for improving con-
densation heat transport ( 48 ). Irrespective of
the strategy adopted, the long-term stability
of the different coatings employed for increas-
ing phase-change heat transport needs to be
explored further under realistic operational
conditions.
Oil-water mixtures are generated from dif-
ferent sources such as petroleum extraction
and refining, textile and leather processing,
wastewater treatment, and fracking. Their
compositions range from free oil and water
to surfactant-stabilized oil-water emulsions.
Owing to the complexity, cost, and energy re-
quirements associated with current separa-
tion methods, membrane-based strategies for
oil-water separation have recently gained much
interest ( 52 ). Membranes with tailored wet-
tability, such as hydrophobic-oleophilic mem-
branes, allow for the lower–surface tension oil
to permeate through while preventing the pas-
sage of water (Fig. 3C) ( 23 , 53 ). However, such
membranes are not suitable for gravity-driven
oil-water separation as the water phase is like-
ly to contact the membrane first owing to its
higher density. Additionally, such hydrophobic
membranes are also prone to fouling by the oil
phase or any surfactants that may be present.
Membranes that are simultaneously hydro-
philic and oleophobic can overcome these lim-
itations ( 52 ). However, this combination of
surface wettability is counterintuitive, given
that the surface tension of oils is lower than
that of water. Recently, hygro-responsive (i.e.,
surfaces that can change their surface compo-
sition based on interactions with a contacting
liquid) membranes composed of metal meshes
or fabrics, coated with a mixture of poly-
ethylene glycol diacrylate (PEGDA) and fluo-
rodecyl polyhedral oligomeric silsesquioxane
(fluoroPOSS), were developed ( 5 ) (Fig. 3, D
and E). These membranes could reconfigure
their surface to become superhydrophilic or
superoleophobic depending on whether they
contacted the water or the oil phase, respective-
ly. The fabricated membranes could separate
a wide range of oil-water mixtures, including
both oil-in-water and water-in-oil emulsions
with >99.9% separation efficiency. The hygro-
responsive membranes could also be used
together with hydrophobic and oleophilic
membranes to achieve continuous oil-water
emulsion separation (Fig. 3F). A wide range
of selective wettability membranes have now
been developed for oil-water separation ( 52 ).
Given the numerous industrial and environ-
mental applications, this research topic is
expected to gain further prominence in the
coming years.Design principles for controlling
solid accretion
Solid foulants display substantial disparity
in terms of composition, chemical structure,
modulus, and the length scale of deposition
(Fig. 4A). Common hard foulants include ice,
inorganic scale, waxes and asphaltenes, and
natural gas hydrates, whereas soft foulants in-
clude bacteria, biofilms, and proteins. Numer-
ous surface modification strategies have been
used to reduce the attachment of different solid
foulants on a variety of underlying substrates.
Figure 4B compiles and highlights different
surface design strategies that are likely to re-
duce the accretion of a given solid foulant,
based on the foulant length scale and modu-
lus. The success of each of these strategies has
been quantified in Fig. 4C, which compiles
the reduction in solid foulant attachment
achieved using different surface modification
techniques for a variety of solid foulants, on
different underlying substrates. It is clear from
the data that the utility of each surface design
strategy is strongly dependent on the prop-
erties of the foulant, and that a single design
strategy that works across a broad range of
different foulants and fouling length scales is
thus far missing.
One of the most commonly used strategies
to lower the adhesion or accretion of a range
of hard and soft foulants is surface energy (gSV)
optimization (Fig. 4, B and C) ( 6 , 54 ). However,
the effectiveness of this approach in prevent-
ing solid fouling is limited by the relatively
narrow range over which surface energy can
be varied ( 6 ). One specific application where
variation ingSVis particularly useful is the
repulsion of soft biological foulants. Baier
showed that the attachment of different bio-
logical foulants, such as bacteria, on surfaces
is minimized over a narrow range of surface
energies, typically between 20 and 30 mN/m
( 54 , 55 ). Another report ( 56 )alsofoundarange
of solid surface energies that yielded a mini-
mum in the amount of protein adsorbed from
milk, even though in this case, the minimumwas obtained over a different surface energy
range (gSV= 30 to 35 mN/m).
Ice accretion has adverse effects on the
operation of a range of commercial and resi-
dential activities ( 57 , 58 ). In this section, we
study accreted ice as a model solid foulant be-
cause, depending on the environmental con-
ditions, ice displays a wide disparity in terms
of its structure, modulus (1.7 to 9.1 GPa), den-
sity (0.08 to 0.9 g/cm^3 ), and length scales of
fouling (approximately square nanometers to
square meters). Different types of ice include
glaze, rime, frost, snow, or a combination of
these diverse forms. To combat the accretion
of these different forms of ice, a wide range
of surface modification technologies have
been studied. These developed technologies
have found some success in reducing the ac-
cretionofarangeofotherhardandsoftfoul-
ants as well (Fig. 4C). We highlight the overlap
between the design of surfaces to reduce ice
accretion and other solid foulants on the basis
of the similarities in their modulus and ac-
creting length scales below.
Figure 5A shows the ice adhesion strength
values reported in literature categorized by
coating material, testing methodology, and
testing temperature. The ice adhesion values
in these reports were measured over different
length scales of accreted ice (Fig. 5B). Recent
work has shown that the measured ice adhe-
sion strength can be a function of the accreted
area ( 59 ). Hence, in Fig. 5A we report the
“apparent”ice adhesion strengths (tice), de-
fined as the force required for ice detachment
per unit area for both small and large inter-
facial areas. For small areas (typically a few
square centimeters)tice¼^tice, where^ticeis
the shear strength for the ice-substrate in-
terface, or the ice adhesion strength. For larger
areas,ticecan be <<^tice. Typically,^ticevalues
for structural materials such as metals and
ceramics are ~1000 kPa. Icephobic surfaces
are defined as surfaces for which^tice< 100 kPa
( 60 ). However, the values of^ticerequired for the
passive shedding of ice in different applications
can be an order of magnitude lower ( 58 , 61 , 62 ).
The effects of varying the substrate surface
energy (gSV) on^ticehave been extensively
studied (Fig. 4C). It was shown previously
that on different, high-modulus solids,^tice=
BgLV(1 + cosqrec), whereBis an experimen-
tal constant,gLVis the surface tension of
water, andqrecis the receding water contact
angle ( 13 , 63 ). Lowering the solid surface
energy (gSV) increasesqrec, reduces the prac-
tical work of adhesion, and consequently
lowers^tice( 13 , 57 ). For nontextured surfaces,
the maximum water receding contact angle
qrec~125°. This leads to a minimum value of
^tice~150 kPa.
^ticevalues can be lowered below 100 kPa by
using textured superhydrophobic surfaces
(SHSs), which can readily achieveqrec>150°.Dhyaniet al.,Science 373 , eaba5010 (2021) 16 July 2021 5 of 13
RESEARCH | REVIEW