foulants as well, including bacteria ( 68 ), micro-
( 69 ) and macro-scale marine species ( 70 ), and
inorganic scale ( 71 ) (Fig. 4C). Different lubri-
cants, typically hydrophobic oils or water
( 61 , 62 , 72 ), may be incorporated as a pool of
liquid within a functionalized porous texture
( 58 , 62 , 73 ), or infused within a polymeric net-
work ( 58 , 62 , 74 ). However,^ticevalues for many
oil-containing lubricated surfaces can increase
after only a few icing–de-icing cycles ( 63 , 75 ).
This occurs primarily as a result of the par-
tial or complete loss of the lubricating liquid
( 75 – 78 ). Surfaces with an aqueous lubricating
layer can withstand multiple icing and de-
icing cycles, likely due to the continuous re-
plenishment of water from the environment
( 61 , 62 , 79 ). However,^ticevalues can exceed
100 kPa on these surfaces at temperatures be-
low−50°C because of the gradual freezing of
the aqueous lubricating layer ( 61 , 62 , 79 ). The
introduction of interfacial slip, facilitated by
the high surface mobility of grafted polymeric
chains, has also been shown to effectively re-
duce solid adhesion ( 38 ) and presents a prom-
ising avenue for future research.
One approach that moves away from mod-
ifying the chemistry or texture of the underly-
ing substrate involves understanding the role
of substrate modulus in the mechanics of solid
detachment. Previous work by Chaudhury and
Kim has shown that the shear stress required
to detach a rigid solid from a softer thin film is
given ast=A(WaG/t)1/2, whereAis an experi-
mental constant,Wais the work of adhesion,
Gis the shear modulus, andtis the thickness
of the film ( 80 ). Shear modulus minimization
can be an effective design strategy for reduc-
ing solid fouling, as the modulus of a coating
canbevariedoverasmanyasfiveordersof
magnitude ( 6 ). Thus, many literature reports
(Fig. 4C) have used low-modulus elastic coat-
ings to reduce surface adhesion by foulants
that are sufficiently large and rigid to deform
the coating. For example, low-modulus coat-
ings have been shown to appreciably reduce
the attachment of rigid-walled algal sporelings
( 81 ) and pseudo-barnacles ( 82 ) (Fig. 4C).
Recent work has shown that sufficiently soft
elastomers are intrinsically icephobic, irrespec-
tive of surface energy, with the ice adhesion
strength for soft elastomers^tice¼G^1 =^2 ( 63 ).
Thus, to enable lower values of^tice, one could
simply lower the shear modulus of a particular
elastomer by reducing its cross-link density.
However, as the cross-link density of a rubber
is lowered, its mechanical durability declines
as well.
To overcome this limitation, icephobic coat-
ings that displayed interfacial slippage ( 83 )
(i.e., a nonzero slip velocity) at the ice-coating
interface were developed ( 63 ). These coatings
were fabricated by the addition of oils or other
plasticizers below their miscibility limit with-
in different elastomers, and unlike lubricated
surfaces, do not possess a free oil layer ( 63 ). The
developed coatings were found to be extremely
durable, and some of them maintained their
low^ticevalues even after severe mechanical,
chemical, and thermal testing.
Another challenge with the design of ice-
shedding surfaces is the issue of scalability.
Even the best-performing icephobic systems
discussed above would require extremely high
forces to remove accreted ice from large struc-
tures such as bridges, ship hulls, and airplane
wings. Recently, Golovinet al. discussed mate-
rials that exhibit a low interfacial toughness
(LIT) with ice ( 59 ). Uniquely for LIT surfaces,
the force required to remove any adhered ice
is low and can be independent of interfacial
area because the delamination of ice depends
on the toughness of the interface (owing to
crack propagation) and not its actual shear
strength.
By understanding the role of different mate-
rial properties in determining the ice-material
interfacial toughness (G), they systematically
designed materials with toughness values close
to the theoretical limit (G~ 0.1 J/m^2 ). They
showed thatG≈^tice^2 t= 2 G, whereGis the shear
modulus andtis the thickness of the coating.
The design and ice-shedding behavior of LIT
materials is notably different from that of ice-
phobic materials with thinner, higher-modulus
coatings yielding lower values ofG. Other re-
lated work has shown that low-modulus fillers,
air voids and substructures within a polymeric
coating, and even air pockets in SHSs, can
act as stress concentrators at the coating-ice
interface, promoting crack formation and
propagation, and thereby allowing for easier
shedding of accumulated ice ( 84 – 86 ). The pro-
motion of crack formation can similarly be a
successful strategy against other solid foulants
as well ( 87 , 88 ).
Iced surfaces can experience wide variations
in temperatures, humidity, and, notably, forces
experienced for ice detachment. A wide range
of tests have been developed to estimate the
adhesion of ice under these differing scenarios,
which has often led to contrary or confusing
results. For example, when ice detachment
was facilitated by centrifugal shear forces, the
reported ice-metal adhesion strengths were
one-quarter the values reported via other
mode II–shear (push off, zero-degree cone)
and mode I–tensile (blister) tests (Fig. 5A)
( 9 , 15 , 58 , 89 – 91 ). Although coating compo-
sition and ice structure differences between
different tests can account for some variability
inticevalues, this variability can still be seen
when these variables are kept somewhat con-
sistent ( 92 ). Such issues have thus far pre-
vented the development of a standardized
testing methodology for evaluating ice-shedding
coatings ( 57 , 92 ). Additionally, differences in
ice volumes and iced areas (spanning two to
four orders of magnitude) used during testingcan lead to sizable variations in apparent ice
adhesion strength values for the same surface,
depending on whether the failure is dominated
by interfacial strength or interfacial toughness
(Fig. 5B).
For the case of crystalline foulants such as
ice ( 44 ), scale ( 93 ), and clathrate hydrates ( 94 ),
surface fouling initiates via the nucleation and
growth of the foulant crystals. Recently, much
attention has been focused on developing anti-
icing surfaces, i.e., surfaces that can retard the
nucleation and/or growth of ice crystals on a
surface at a given temperature. Figure 5C is a
compilation of water droplet freezing delay
times over surfaces with a range of wettability
reported in the literature ( 95 – 105 ). It can be
seen that typically, the droplet freezing time
increases with increasing water contact angle.
This is consistent with classical nucleation
theory and numerous experiments that show
that the nucleation rate within a liquid droplet
is dependent on both the surface wettability
and roughness ( 44 , 96 ). The minimum nuclei
radius needed for the stable growth of a crystal
within a droplet is given byrc¼ D^2 Ggvsl;where
rcis the critical nuclei radius,gslis the solid-
liquid interfacial tension, andDGv¼DHvðÞTTmm T
is the volumetric Gibbs free energy difference
between bulk crystallized solid and bulk liquid
( 44 ). For an ice-water interface, this relation
yieldsrc;ice< 10 nm at temperatures <−5°C
( 96 , 105 , 106 ). It has been shown that for sur-
face roughnessr≤rc;iceat any given tem-
perature, freezing of water droplets can be
substantially delayed, even for hydrophilic
surfaces ( 96 , 105 ). Additionally, the small liquid-
solid contact area for textured superhydro-
phobic surfaces reduces the thermal transport
between the cold surface and the water droplet,
which can, in turn, retard the ice nucleation
rates ( 98 , 99 ). However, no appreciable delay
in nucleation rates was found on SHSs when
the atmosphere surrounding the droplet was
also cooled to the surface temperature ( 96 ).
Some surfaces reported in the literature
( 96 , 105 ) can engender specific enthalpic in-
teractions with water molecules (circled data
in Fig. 5C) to alter ice nucleation rates. On such
surfaces, cooling liquid water droplets below
freezing temperatures can lead to quasi-liquid
states with ice-like properties ( 96 , 105 ). This in
turn leads to an increase in the droplet freez-
ing time and a suppression in the ice nucleation
temperature ( 96 , 105 ). Similar enthalpic inter-
actions to retard ice nucleation and growth
have been manifested through the use of am-
phiphilic or charged materials such as poly-
electrolyte brushes ( 95 , 102 ), hydrogels ( 79 ),
charged crystals ( 107 , 108 ), or even natural anti-
freeze proteins ( 103 , 104 ), which are known to
bind to ice crystal faces ( 109 ).
Multiple studies have reported the ability of
similar hygro-responsive surfaces to resist foul-
ing by different soft foulants such as proteins,Dhyaniet al.,Science 373 , eaba5010 (2021) 16 July 2021 8 of 13
RESEARCH | REVIEW