Journal of Membrane Science 588 (2019) 117185

Contents lists available at ScienceDirect

Journal of Membrane Science

journal homepage: www.elsevier.com/locate/memsci

Bio-inspired gas-entrapping membranes (GEMs) derived from common T

water-wet materials for green desalination
Ratul Das1, Sankara Arunachalam1, Zain Ahmad, Edelberto Manalastas, Himanshu Mishra∗
King Abdullah University of Science and Technology (KAUST), Water Desalination and Reuse Center (WDRC), Biological and Environmental Science and Engineering
(BESE) Division, Thuwal, 23955-6900, Saudi Arabia

A R T I C LE I N FO A B S T R A C T

Keywords: Widespread stress on global water supplies compels the need for low-cost and sustainable desalination processes.
Green desalination In this regard, desalination through membrane distillation (MD) can harness waste-grade heat or renewable
Membrane distillation energy. So far, the membranes for MD have been exclusively derived from intrinsically water-repellant materials
Perfluorocarbon-free - mostly perfluorocarbons. However, perfluorocarbons are limiting in terms of operational conditions, and they
Wetting and non-wetting materials
also introduce economic and environmental concerns. The development of perfluorocarbon-free MD membranes
Biomimetics
Reentrant textures
would likely address those challenges. Here, we report on the proof-of-concept for biomimetic gas-entrapping
Mass transfer membranes (GEMs) for MD derived from silica and poly(methyl methacrylate) (PMMA) that are water-wet
Heat transfer materials. We drew inspiration for our GEM design from the cuticles of springtails and hairs of Halobates ger-
manus, both of which exhibit mushroom-shaped (or reentrant) features. Accordingly, our GEMs comprise arrays
of microscale cylindrical pores with reentrant inlets and outlets that can robustly entrap air on submersion in
water. Our PMMA-GEMs yielded a vapor flux of J ≈ 1 L-m−2-h−1 while separating a solution of ∼0.6 M NaCl at
333 K from deionized water at 288 K under a cross-flow configuration. To our knowledge, this is the first-ever
demonstration of MD membranes derived from intrinsically water-wet materials, and these findings suggest that
the rational design of membranes towards greener and cheaper desalination processes is possible.

1. Introduction repellent materials for MD membranes, such as polypropylene [10],

Membrane distillation (MD) refers to a gamut of desalination pro-
cesses in which water vapor is extracted from hot seawater (feed side)
through water-repellent membranes and collected on the permeate side
[1–4]. If the permeate side comprises cold deionized water or a mild
vacuum, the process is, respectively, known as direct contact membrane
distillation (DCMD) or vacuum membrane distillation [4]. To reduce
the carbon footprint of desalination, MD processes could be coupled
with low-grade waste heat from industrial and natural sources, making
them an active area of research for interdisciplinary materials discovery
and process engineering [1,5]. MD requires membranes for the robust
separation of the (salty) feed side from the (pure) permeate side to
prevent liquid intrusion and to allow only pure water vapor to move to
the permeate side. So far, only membranes with water-repellent che-
micals/coatings, for example, made from or coated-with polytetra-
fluorethylene (PTFE), polyvinylidene difluoride (PVDF), and per-
fluorosilanes have achieved the necessary robust separation [4,6–9].
Researchers have also investigated perfluorocarbon-free yet water-
carbon nanotubes [11], bucky paper [12], nanoporous carbon [13],
graphene [14,15], and poly(ether imide) [16], but the commercially
available membranes are generally composed of PVDF or PTFE. How-
ever, perfluorinated chemicals are expensive, non-biodegradable, and
vulnerable to mechanical damage due to abrasion, elevated tempera-
tures, and harsh and frequent cleaning cycles to preempt fouling
[1,2,17–21]. Perfluorinated membranes also suffer from the time-de-
pendent degradation of their wetting properties in contact with water
[22] due to the adsorption of amphiphilic contaminants, driven by
hydrophobic interactions [23], leading eventually to the intrusion of
water in the pores [1,2]. Thus, membranes derived from intrinsically
water-wet robust materials that can realize the functions of per-
fluorinated membranes are desirable, although they have remained
unachievable thus far. One possible reason for why this problem has
remained unexplored among surface scientists is the conventional
wisdom that surface roughness monotonically increases/decreases the
wettability of hydrophilic/hydrophobic surfaces – sometimes, referred
to as the Wenzel's law [24], which is given by cosθr = r × cosθo , where r

∗
Corresponding author.
E-mail address: himanshu.mishra@kaust.edu.sa (H. Mishra).
1
Equal author contribution.

https://doi.org/10.1016/j.memsci.2019.117185
Received 19 April 2019; Received in revised form 7 June 2019; Accepted 11 June 2019
Available online 16 June 2019
0376-7388/ © 2019 Published by Elsevier B.V.
R. Das, et al.
is the surface roughness, defined as the ratio of the real surface area to
the projected surface area, and θr and θo are, respectively, the apparent
and as-placed contact angles on rough and smooth surfaces, assuming
that there is no entrapment of air at the solid-liquid interface. Water is
therefore expected to wet (or intrude) the pores of a rough hydrophilic
surface/membrane instantaneously, whereas rough hydrophobic sur-
faces can exhibit superhydrophobicity [25]. Here, contrary to conven-
tional expectations, we report on the conceptualization and develop-
ment of biomimetic gas-entrapping membranes (GEMs) derived from
water-wet materials that robustly entrap air based on their surface to-
pography and facilitate desalination.
We drew inspiration from seaskaters (Halobates germanus) [26,27]
and springtails (Collembola) [28] that have evolved strategies for sur-
viving accidental submersion in water by spontaneously entrapping
life-sustaining air (Fig. 1A–E). Despite the unrelated evolutionary tra-
jectories of these two insects, the hair of seaskaters and the cuticles of
springtails show striking similarities with mushroom-shaped topo-
graphies, which prevent liquids from intrusion. In fact, when mush-
room-shaped pillars were carved onto intrinsically wetting materials,
such as steel [29], silica [30], poly(ethylene glycol) dimethacrylate
[31], polyvinyl alcohol [32], and epoxy-based poly(disperse orange)
(PDO-3) [33], these materials exhibited liquid repellence, e.g., contact
angles of drops of water and non-polar liquids (e.g., hexadecane) in air
were θA > θr > 150° and θA - θR < 20°, where θA, θR, and θr were the
advancing, receding and as-placed or apparent contact angles
[30,33,34]. Building on these findings, we demonstrated that pillar-
based surfaces suffered a catastrophic loss of liquid-repellence upon
immersion in wetting liquids when the liquids could penetrate laterally
(parallel to the surface) through the spaces between the pillar-stems
[35]. To remedy this weakness, we introduced cavity-based micro-
textures with mushroom-shaped openings/inlets (from here onwards
we refer to them as "reentrant"), which robustly entrapped air upon
immersion in wetting liquids and exhibited omniphobicity, character-
ized by contact angles of water and hexadecane in air of θA > θr > 90°
2
Journal of Membrane Science 588 (2019) 117185
Fig. 1. (A–E) Representative scanning electron mi-
crographs of cuticles of springtails (Collembola)
showing pine-cone-shaped papillae decorated with
reentrant overhanging features. (F–I) Scanning
electron micrographs of our biomimetic silica-
GEMs. (F) Cross-sectional view, (G) Magnified cross-
section of a single pore, (H) and (I) Magnified views
of the reentrant edges at the inlets and outlets of
pores.
[35,36]. We also identified shapes (circular or non-circular), profiles
(simple, reentrant, or doubly reentrant microtextures), and corners
(sharp or smooth) of cavities on intrinsically wetting surfaces that led to
the robust entrapment of air and maximized resistance to liquid in-
trusion [37,38]. Given those design principles, here, we introduce gas-
entrapping membranes (GEMs) derived from water-wet materials
comprising vertically aligned cylindrical pores with reentrant inlets and
outlets for the robust entrapment of air upon immersion in water. In the
next section, we demonstrate the proof-of-concept for our silica-GEMs,
microfabricated using lithography and dry etching, and the quantifi-
cation of liquid entry pressures (LEPs). Subsequently, we translate our
design principles into a commercially available low-cost polymer using
scalable manufacturing techniques and demonstrate that GEMs can be
used for MD under a cross-flow configuration.
2. Experimental
2.1. Microfabrication of silica-GEMs
We microfabricated arrays of pores with reentrant features at the
inlets and outlets onto double-side-polished silicon wafers (p-
doped, < 100 > orientation, 4” diameter and 300 μm thick with 2-μm-
thick thermally grown oxide layers on both sides from Silicon Valley
Microelectronics) using photolithography and dry etching (Fig. 1F–I,
Figure S1, Section S1 in SI).
2.2. Set-up for leak-testing silica-GEMs
We designed our customized setup to test liquid intrusion into the
pores of silica-GEMs using SOLIDWORKS 2019 (Fig. 2). The setup was
3D printed using the BCN3D SIGMA R19 printer with polylactic acid as
the filament. Silica-GEMs separated a reservoir of saline water (∼0.6 M
NaCl that was labelled with green food coloring) from deionized water
(MilliQ Advantage A10), both maintained 293 K and 1 atm. We
R. Das, et al.
employed a Hanna portable conductivity meter (Model – HI98192) for
real-time monitoring of the electrical conductivity of the deionized
water reservoir.
2.3. Microfabrication of PMMA-GEMs
We started with 1.2-mm-thick PMMA (poly(methyl methacrylate))
sheets (McMASTER-CARR) and used a CNC drill (LPKF ProtoMat S103)
and a CO2 laser (Universal X-660, PLS6.75) for controlled drilling.
Details are provided in Section S6. The two layers were sandwiched
together by using UV adhesive (NOA 68, Thorlabs) under a high-in-
tensity UV lamp (Spectroline SB-100PC, 365 nm) (Figs. S2, S3, S4 and
Section S2, S3 in SI).
2.4. Membrane distillation with PMMA-GEMs under crossflow
Vapor fluxes from our PMMA-GEMs were measured in a crossflow
setup using a custom-built MD module with PMMA (Fig. S5 and Section
S4). For the feed-side, we used saline water (∼0.6 M NaCl) and labelled
with green food coloring, maintained at T = 333 K using a hot plate
(IKA RCT basic) and a flexible resistive foil heater (HT10K, Thorlabs),
controlled by a TC200 temperature controller (Thorlabs). Both streams
were recirculated at a rate of 10 ml/min using a Harvard Apparatus
syringe pump (PHD ULTRA™ 4400). The flux was collected in a col-
lection flask and measured using a mass balance (Sartorius Quintix 613-
1S) with a precision of 0.001 g.
3. Results and discussion
To demonstrate the feasibility of GEMs, we used photolithography
and dry etching techniques to create vertically-aligned pores with re-
entrant inlets and outlets on double-side-polished SiO2(2 μm)/Si
(300 μm) wafers that exhibited an intrinsic contact angle for the water/
vapor system of θo = 40° ± 2° (Fig. 1 F–I and S1, Section S1, Section
S5 and Table S1). We chose the design parameters based on our recent
investigation of the time-dependence of air inside silica cavities with
reentrant inlets that exhibited over eight orders of magnitude longer
entrapment compared with simple cylindrical cavities under hex-
adecane [37]. After iterating the microfabrication process, we realized
silica-GEMs with vertically-aligned pores with inlet and outlet dia-
meters of D = 100 μm, pitches of L = 400 μm, reentrant overhangs of
w = 18 μm, and lengths of h = 300 μm (Fig. 1 F–I).
To investigate the efficacy of silica-GEMs at entrapping air in the
context of MD, we built a customized module (Fig. 2A) for leak-testing
silica-GEMs, in which the membrane separated a salty solution (∼0.6 M
NaCl) from a reservoir containing deionized water at normal tem-
perature and pressure (NTP, T = 293 K, P = 1 atm). Our module fa-
cilitated in situ monitoring of the electrical conductivity (detection
limit: ± 0.01 μS/cm ) of the deionized water reservoir, allowing the
3
Journal of Membrane Science 588 (2019) 117185
Fig. 2. (A) Design of our custom-built cell used to
test the efficacy of silica-GEMs at separating salty
solutions (∼0.6 M NaCl) from pure water at
T = 293K, P = 1 atm. (B) Semi-logarithmic plot of
the electrical conductivity of the reservoir con-
taining deionized water monitored over time. Silica-
GEMs robustly entrapped air in every pore, ren-
dering the electrical conductivity to remain un-
changed for over six weeks, after which the ex-
periment was discontinued.
intrusion of water (even pore-level leakage) to be detected through the
membrane. Remarkably, under isothermal conditions, our silica-GEMs
robustly entrapped air for more than six weeks, after which the ex-
periment was discontinued (Fig. 2B); in contrast, silica membranes with
similar cylindrical holes, but without reentrant features at pore inlets
and outlets, failed instantaneously (Fig. S6, Supplementary Movie S1).
These results establish that it is possible to prevent the intrusion of
water into membranes made from water-wet materials by robustly
entrapping air inside the pores using our biomimetic approach.
Supplementary data related to this article can be found at https://
doi.org/10.1016/j.memsci.2019.117185.
To investigate the factors and mechanisms underlying the functions
of GEMs immersed in water, we employed confocal microscopy and
tracked the positions of water-vapor interfaces at the inlets/outlets of
silica-GEMs and compared them with silica surfaces with arrays of
cavities with reentrant inlets (Supplementary Section S6). It is well
known that the laser in confocal microscopy heats up water locally
[37,39]. For instance, in the case of our silica surfaces with arrays of
cavities with reentrant inlets immersed in water, even the mild laser
power of 0.6 mW led to the capillary condensation of water inside the
cavities, which displaced the trapped air; the primary air-water menisci
at the inlets of the cavities bulged toward the water, leading to the
intrusion of water in 3 h, characterized by the liquid meniscus depin-
ning from the reentrant inlets and touching pore walls or floor [37]. On
the other hand, if the laser power was increased five-fold to 3 mW, we
observed water intrusion within 1.7 h (Supplementary Figs. S7 and S8
and Section S6). In stark contrast, in the case of silica-GEMs, we found
that air-water interface at the reentrant edges of the pores remained flat
under the 3 mW laser power for the first 2 h, and started bulging slowly
afterwards, and eventually got intruded at 17 h (Fig. 3). We therefore
realized that during the initial 10 h, the water vapor arising from the
laser heating moved from the hot side to the cold side, where it con-
densed (Fig. 4). This led us to wonder why the pores of silica-GEMs did
not become clogged by the condensation of water on the pore-walls. We
note that for surfaces with intrinsic contact angles of water, θo < 10°,
the condensation of water leads to thick wetting films, whereas for
θo > 40°, water condenses as droplets [40,41]. Thus, water vapors
from the hot side condense as tiny drops on the pore-walls that grow
locally, but with limited advancement due to the pinning forces
[41–43], leaving the rest to condense onto the cold air-water interface
on the permeate-side. It was not possible for us to directly visualize of
the condensed water drops on the pore-walls and their collective be-
havior over time, because of the opaqueness of silicon, the overhanging
reentrant-edges, the high aspect ratio of the pores, the absence of dye in
the condensed water, and the low range of the z-stacking (z < 50 μm)
of our confocal laser scanning microscope (Supplementary Section S6).
Next, using a custom-built setup, we measured the liquid entry
pressure (LEP), defined here as the pressure at which the liquid stabi-
lized at the reentrant inlets/outlets of the pores of GEMs intrudes inside
R. Das, et al.
and touches the pore-walls (Supplementary Section S7 and Fig. S9). We
found that for GEMs with 100 μm inlets, LEP was ≈ 25 kPa. We modeled
the LEP by accounting for the compression of the air entrapped inside
the pores by assuming it to be an ideal gas and the ‘Laplace pressure’ of
the curved air-water interface as, PLEP =
P1 V1/ V2 + γLV × (1/ R1 + 1/ R2) × cos θo ≈ 27 kPa, where P1, V1, and P2 , V2
are the pressures and volumes of the air entrapped in the pores, re-
spectively, immediately after immersion in water and just before the
intrusion, γLV is the liquid-vapor interfacial tension, θo is the intrinsic
contact angle of the as-placed water droplet on a smooth and flat SiO2
surface, and R1, and R2 are the radii of curvatures of the air-water in-
terface before intrusion (Supplementary Figs. S10A–B and Table S2);
we ignored the hydrostatic pressure component of the water column of
height, l = 5 mm (P l = ρgl ≈ 0.05 kPa ) [44]. We note that significantly
higher LEPs might be achieved by reducing the diameter of the pore
4
Journal of Membrane Science 588 (2019) 117185
Fig. 3. Confocal micrographs of reentrant pores: (A)
Computer-enhanced 3D reconstructions of the air-
water interface at inlets of silica-GEMs with re-
entrant pores under a z ≈ 5 mm column of water at
laser power of 3 mW. (B–D) cross sectional views at
different time intervals. Intruding liquid menisci
were stabilized at the reentrant edges of the pores
for more than 8 h. Blue indicates the water, white
indicates the silica and grey indicates the silicon. (E)
Schematic of menisci stabilization at the pore-level;
due to the localized laser-heating during confocal
microscopy, water evaporates and condenses onto
other surfaces. The significantly reduced bulging
compared with cavities with reentrant inlets
(Supplementary Figs. S7 and S8 and Section S6)
indicates that most of the vapor condenses on the
air-water interface at the (colder) pore-outlet.
inlets, for instance, to the nanoscale. Next, we explored the potential of
GEMs for desalination.
MD exploits thermal energy to drive the extraction of water vapor
from the hot side to the cold side. Thus, in addition to the robust en-
trapment of air, MD membranes should minimize the loss of heat from
the hot to the cold side [8]; chemical stability and resistance to fouling
are other key attributes of MD [1,2,4,45]. To quantify the mass trans-
port of water vapor through silica-GEMs, we built a module that se-
parated 0.6 M NaCl maintained at T = 333 K as the surrogate feed from
deionized water maintained at T = 288 K as the surrogate permeate.
Even though silica-GEMs prevented liquid intrusion, they did not ex-
hibit measurable fluxes due to the high thermal conductivity of silicon
(k = 149 W-m−1-K−1) [46], which lowered the feed-side temperature
and heated the permeate side. In fact, this phenomenon, known as
temperature polarization, is well known in the MD literature and
Fig. 4. (A) Schematic representation of the manu-
facturing of PMMA-GEMs using CNC and a CO2
laser. (B) PMMA-GEMs under cross-flow in a MD
setup. (C) A pore-level view of vapor transport in
PMMA-GEMs: liquid water is stabilized at the re-
entrant inlet/outlet, allowing only vapor to be
transported. Water vapor evaporates from the hot-
side and condenses onto the pore-walls and the cold
air-water interface, with the latter being the larger
sink for the vapor.
R. Das, et al.
researchers have demonstrated it to be a significant factor in the ra-
tional design of MD membranes [8,19]. Our attempts to lower the
thermal conductivity of silica-GEMs by using a conformal coating of a
2-μm-thick layer of poly(p-xylylene) (thermal conductivity of
k = 0.001 W-m−1-K−1) [47] did not mitigate the temperature polar-
ization (Supplementary Section S8). We therefore decided to test our
design principles on another material system.
We translated the design principles to achieve GEMs from com-
mercially available sheets of poly(methyl methacrylate) (PMMA) that is
a water-wet material (θo = 77° ± 2°, (Supplementary Table S1).
PMMA has a significantly lower thermal conductivity, k = 0.19 W-m−1-
K−1 than silicon [46]. To create PMMA-GEMs, we developed a two-step
manufacturing process entailing drilling using a Computer Numerical
Control (CNC) system and a CO2 laser (Supplementary Section S2).
Starting with PMMA sheets with 1.2 mm thickness, we drilled con-
centric circular pores with diameters D1 = 1 mm until the depth of
1 mm by CNC, followed by drilling smaller holes with diameters,
D2 = 16 μm, through the remaining 0.2 mm by the CO2-laser; the
center-to-center distance between the pores was L = 2 mm (Fig. 4A).
Subsequently, we realized PMMA-GEMs by bonding two sheets of
drilled PMMA by optical alignment, such that the smaller pores func-
tioned as the inlets and outlets. The porosity of those membranes, de-
fined as the volume of air trapped in the pores to the total membrane
volume (Section S3), was quite low, ϕ = 0.08, for instance, in com-
parison to that of standard MD membranes (0.2–0.9) [1–3]. Despite
their low porosity, PMMA-GEMs yielded fluxes of desalinated water in
the range of 1 L-m−2-h−1(LMH) for over 12 h while separating cross-
flowing streams of 0.6 M NaCl at T = 333 K and deionized water at
T = 288 K, both flowing at 10 ml-min−1 (Fig. 4, Supplementary Fig. S5
and Section S4), in comparison commercial membranes yield fluxes in
the range of 10–90 LMH [3,4,9,48]. By monitoring the electrical con-
ductivity of the permeate that remained unchanged at 20 μS/cm, we
estimated the salt rejection to be close to 100%. PMMA-GEMs fa-
cilitated the robust entrapment of air for over 90 h under those condi-
tions after which the experiment was discontinued. In contrast, PMMA
membranes with simple cylindrical pores with pore sizes in the range of
16–25 μm, i.e., without reentrant inlets and outlets, experienced instant
water intrusion upon immersion (Supplementary Fig. S6, Section S4,
and Movies S2 and S3). These findings demonstrate unambiguously the
potential of the reentrant topography in the context of liquid-vapor
extraction in MD.
Supplementary data related to this article can be found at https://
doi.org/10.1016/j.memsci.2019.117185.
To understand the mechanisms underlying the mass transport of
water vapor, we used Fuller-Schettler-Giddings’ empirical relation to
determine the diffusion coefficient of a binary mixture of ideal gases:
1
1.0133 × 10−7T1.75 1 1 2 than their surface chemistry, this green approach to desalination should
DAB = 2 ⎡ MA + ⎤ , where T is the mean temperature
⎡
1 1
⎤ ⎣ MB ⎦
P ⎢ (∑ v ) A3 + (∑ v )B3⎥
⎣ ⎦
in Kelvin, MA and MB are the molecular weights of the gaseous com-
ponents (water vapor and air), P is the total pressure in bar, and νA, νB
are the atomic diffusion volumes [49]; DAB = 2.75 ×10−5 m2-s−1 for
our experimental conditions. Using the molecular diffusion model
[50,51], considering the experimental vapor pressures [52] on the hot
and cold sides at 333 K and 288 K, respectively, and assuming no
temperature polarization, we estimated the flux under our experimental
conditions to be J≈ 0.6 L-m−2-h−1, which was in reasonable agreement
with our observations (Table S3, Supplementary Section S9).
We note here that the goal of this work was limited to introducing
the concept of and demonstrating the proof-of-concept for GEMs de-
rived from intrinsically water-wet materials for green desalination. We
anticipate that higher fluxes could be realized by enhancing the mem-
brane's porosity, choosing materials with even lower thermal con-
ductivity than PMMA, and reducing the length of the pores. However,
several questions pertaining to fundamental and applied aspects remain
to be answered by the interdisciplinary research community before
5
Journal of Membrane Science 588 (2019) 117185
GEMs could emerge as a viable mainstream technology. For instance, it
is important to realize that the thermodynamically stable state for any
hydrophilic porous material immersed in water is the fully-filled (or the
Wenzel) state [36,37]. Our biomimetic approach therefore relies on
robust kinetic barriers on the free energy landscape that prevent wet-
ting transitions from the air-entrapped (or Cassie-) states to the Wenzel
state [22,36,53,54]. In fact, due to the differences in the intrinsic
contact angles and shapes of SiO2-GEMs and PMMA-GEMs, the wetting
transitions taking place in them as they separate the feed-side from the
cold permeate might be dissimilar. In this context, systematic in-
vestigations of time-dependent pore-filling as a function of solid-liquid-
vapor interfacial tensions, pore characteristics, including shape, size,
length, and surface roughness, cross-flow conditions, and the tem-
peratures of the hot and cold air-water interfaces are needed. New
techniques such as neutron computed tomography and X-ray fluores-
cence might facilitate the direct observation of condensing water dro-
plets through opaque solids [55], if the spatial resolution could be en-
hanced. Concerted experimental and computational investigations are
also needed to investigate the temperature polarization, fouling, dur-
ability, and scalability characteristics of biomimetic GEMs.
4. Conclusions
To summarize, we adopted a biomimetic approach to overcome the
vulnerabilities of MD membranes made from or coated with water-re-
pellent chemicals, such as perfluorocarbons. Inspired by the cuticles of
springtails and hairs of seaskaters, we first microfabricated gas-en-
trapping membranes (GEMs) from water-wet SiO2/Si wafers. Our silica-
GEMs comprised cylindrical pores with reentrant inlets and outlets,
which robustly entrapped air for over six weeks upon immersion in
water at NTP. To demonstrate the proof-of-concept for GEMs using a
low-cost and water-wet material with low thermal conductivity, we
chose PMMA. We developed a two-step drilling process using CNC and
CO2-laser platforms to achieve PMMA-GEMs. The resulting PMMA-
GEMs robustly separated solutions of ∼0.6 M NaCl solution at 333 K
from deionized water at 288 K, yielding a flux of desalinated water of J
≈ 1 L-m−2-h−1 for over 12 h and facilitated robust entrapment of air
until 90 h under those conditions. Thus, our biomimetic GEMs, derived
from common water-wet materials, can perform the crucial function of
perfluorinated MD membranes, i.e. the robust entapment of air while
separating feed- and permeate-sides. We note that the PMMA-GEMs
with the porosity of ϕ = 0.08 were used to demonstrate the proof-of-
concept and that new fabrication methods are needed to increase the
membranes' porosity and to decrease pores' inlet diameters to, respec-
tively, achieve higher fluxes and LEPs [1,3,6,9]. Since the functions of
our GEMs as MD membranes are derived from their geometry rather
also overcome the typical drawbacks of perfluorocarbon-based mem-
branes, including cost and vulnerability to abrasion, elevated tem-
peratures, and harsh cleaning protocols. The benefits of using low-cost
perfluorocarbon-free materials in MD membranes are obvious, e.g., the
costs of PMMA and polyethyleneterephthalate (PET) are 80–90% lower
than the cost of perfluoroethylene [20]. We conclude with the hope that
our findings might be useful in unlocking the potential for using in-
expensive and common materials for green desalination and for liquid-
vapor extraction in general.
Declarations of interest
RD, SA, and HM have filed an International patent, Application No.
PCT/IB2019/054548.
Funding
The authors thank KAUST Baseline Research Funding (BAS/1/1070-
01-01).
R. Das, et al.
Acknowledgements
We thank Mr. Ulrich Buttner and Mr. Ahad A. Sayed from the
KAUST Core Labs for their assistance with microfabrication. We thank
Mr. Adair Gallo Junior for assistance with the MD experiments; Dr. Dina
Garatly for assistance in preparing Fig. 3 and Figure S2; Dr. Eddy
Domingues (Universidade de Aveiro, Portugal) for providing recipes for
the dry etching of silica and silicon layers. The authors also thank Dr.
Virginia Unkefer (KAUST) for assistance in editing the manuscript. We
also thank Prof. Wei Song Hwang (National University of Singapore)
and Prof. Lanna Cheng (Scripps Institution of Oceanography) for kindly
providing us with specimens of springtails whose images we have
presented in Fig. 1A–E. The co-authors thank Mr. Ivan Gromicho, Sci-
entific Illustrator at KAUST, for preparing Fig. 4.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.memsci.2019.117185.
References
[1] A. Deshmukh, C. Boo, V. Karanikola, S. Lin, A.P. Straub, T. Tong, D.M. Warsinger,
M. Elimelech, Membrane distillation at the water-energy nexus: limits, opportu-
nities, and challenges, Energy Environ. Sci. 11 (2018) 1177–1196.
[2] M. Rezaei, D.M. Warsinger, J.H. Lienhard V, M.C. Duke, T. Matsuura,
W.M. Samhaber, Wetting phenomena in membrane distillation: mechanisms, re-
versal, and prevention, Water Res. 139 (2018) 329–352.
[3] E. Drioli, A. Ali, F. Macedonio, Membrane distillation: recent developments and
perspectives, Desalination 356 (2015) 56–84.
[4] M. Khayet, T. Matsuura, Membrane Distillation: Principles and Applications, 1 ed.,
Elsevier, 2011.
[5] N. Ghaffour, S. Lattemann, T. Missimer, K.C. Ng, S. Sinha, G. Amy, Renewable
energy-driven innovative energy-efficient desalination technologies, Appl. Energy
136 (2014) 1155–1165.
[6] S. Feng, Z. Zhong, Y. Wang, W. Xing, E. Drioli, Progress and perspectives in PTFE
membrane: preparation, modification, and applications, J. Membr. Sci. 549 (2018)
332–349.
[7] L. Eykens, K. De Sitter, C. Dotremont, L. Pinoy, B. Van der Bruggen, Membrane
synthesis for membrane distillation: a review, Separ. Purif. Technol. 182 (2017)
36–51.
[8] N. Subramanian, A. Qamar, A. Alsaadi, A. Gallo, M.G. Ridwan, J.-G. Lee, S. Pillai,
S. Arunachalam, D. Anjum, F. Sharipov, N. Ghaffour, H. Mishra, Evaluating the
potential of superhydrophobic nanoporous alumina membranes for direct contact
membrane distillation, J. Colloid Interface Sci. 533 (2019) 723–732.
[9] J.R. Werber, C.O. Osuji, M. Elimelech, Materials for next-generation desalination
and water purification membranes, Nature Reviews Materials 1 (2016) 16018.
[10] M. Gryta, Influence of polypropylene membrane surface porosity on the perfor-
mance of membrane distillation process, J. Membr. Sci. 287 (2007) 67–78.
[11] A.K. An, E.J. Lee, J.X. Guo, S. Jeong, J.G. Lee, N. Ghaffour, Enhanced vapor
transport in membrane distillation via functionalized carbon nanotubes anchored
into electrospun nanofibres, Sci. Rep. 7 (2017).
[12] L.F. Dumée, K. Sears, J. Schütz, N. Finn, C. Huynh, S. Hawkins, M. Duke, S. Gray,
Characterization and evaluation of carbon nanotube Bucky-Paper membranes for
direct contact membrane distillation, J. Membr. Sci. 351 (2010) 36–43.
[13] W. Chen, S. Chen, T. Liang, Q. Zhang, Z. Fan, H. Yin, K.-W. Huang, X. Zhang, Z. Lai,
P. Sheng, High-flux water desalination with interfacial salt sieving effect in nano-
porous carbon composite membranes, Nat. Nanotechnol. 13 (2018) 345–350.
[14] Y.C. Woo, L.D. Tijing, W.-G. Shim, J.-S. Choi, S.-H. Kim, T. He, E. Drioli, H.K. Shon,
Water desalination using graphene-enhanced electrospun nanofiber membrane via
air gap membrane distillation, J. Membr. Sci. 520 (2016) 99–110.
[15] B. Bera, N. Shahidzadeh, H. Mishra, L.A. Belyaeva, G.F. Schneider, D. Bonn, Wetting
of water on graphene nanopowders of different thicknesses, Appl. Phys. Lett. 112
(2018) 151606.
[16] M.A. Hammami, J.G. Croissant, L. Francis, S.K. Alsaiari, D.H. Anjum, N. Ghaffour,
N.M. Khashab, Engineering hydrophobic organosilica nanoparticle-doped nanofi-
bers for enhanced and fouling resistant membrane distillation, ACS Appl. Mater.
Interfaces 9 (2017) 1737–1745.
[17] T. Verho, C. Bower, P. Andrew, S. Franssila, O. Ikkala, R.H.A. Ras, Mechanically
durable superhydrophobic surfaces, Adv. Mater. 23 (2011) 673–678.
[18] L. Boinovich, A.M. Emelyanenko, A.S. Pashinin, Analysis of long-term durability of
superhydrophobic properties under continuous contact with water, ACS Appl.
Mater. Interfaces 2 (2010) 1754–1758.
[19] Z.D. Hendren, J. Brant, M.R. Wiesner, Surface modification of nanostructured
ceramic membranes for direct contact membrane distillation, J. Membr. Sci. 331
(2009) 1–10.
[20] M.F. Ashby, K. Johnson, Materials and Design: the Art and Science of Material
Selection in Product Design, (2014).
Journal of Membrane Science 588 (2019) 117185
[21] A.B. Lindstrom, M.J. Strynar, E.L. Libelo, Polyfluorinated Compounds: Past,
Present, and Future, ACS Publications, 2011.
[22] H. Mishra, A.M. Schrader, D.W. Lee, A. Gallo, S.Y. Chen, Y. Kaufman, S. Das,
J.N. Israelachvili, Time-dependent wetting behavior of PDMS surfaces with bioin-
spired, hierarchical structures, ACS Appl. Mater. Interfaces 8 (2016) 8168–8174.
[23] D. Chandler, Interfaces and the driving force of hydrophobic assembly, Nature 437
(2005) 640–647.
[24] R.N. Wenzel, Surface roughness and contact angle, J. Phys. Colloid Chem. 53 (1949)
1466–1467.
[25] P.-G. de Gennes, F. Brodchard-Wyart, D. Quere, Capillarity and Wetting
Phenomena: Drops, Bubbles, Pearls, Waves, edition, Springer, 2004 2003.
[26] L. Cheng, Marine and freshwater skaters: differences in surface fine structures,
Nature 242 (1973) 132.
[27] N.M. Andersen, L. Cheng, The marine insect Halobates (Heteroptera: gerridae):
biology, adaptations, distribution, and phylogeny, Oceanogr. Mar. Biol. Annu. Rev.
42 (2004) 119–180.
[28] R. Hensel, C. Neinhuis, C. Werner, The springtail cuticle as a blueprint for omni-
phobic surfaces, Chem. Soc. Rev. 45 (2016) 323–341.
[29] P.B. Weisensee, E.J. Torrealba, M. Raleigh, A.M. Jacobi, W.P. King, Hydrophobic
and oleophobic re-entrant steel microstructures fabricated using micro electrical
discharge machining, J. Micromech. Microeng. 24 (2014).
[30] T.Y. Liu, C.J. Kim, Turning a surface superrepellent even to completely wetting
liquids, Science 346 (2014) 1096–1100.
[31] R. Hensel, A. Finn, R. Helbig, H.G. Braun, C. Neinhuis, W.J. Fischer, C. Werner,
Biologically inspired omniphobic surfaces by reverse imprint lithography, Adv.
Mater. 26 (2014) 2029–2033.
[32] P. Zhu, T. Kong, X. Tang, L. Wang, Well-defined porous membranes for robust
omniphobic surfaces via microfluidic emulsion templating, Nat. Commun. 8 (2017)
15823.
[33] J. Choi, W. Jo, S.Y. Lee, Y.S. Jung, S.-H. Kim, H.-T. Kim, Flexible and robust su-
peromniphobic surfaces created by localized photofluidization of azopolymer pil-
lars, ACS Nano 11 (2017) 7821–7828.
[34] R. Tadmor, P.S. Yadav, As-placed contact angles for sessile drops, J. Colloid
Interface Sci. 317 (2008) 241–246.
[35] E.M. Domingues, S. Arunachalam, H. Mishra, Doubly reentrant cavities prevent
catastrophic wetting transitions on intrinsically wetting surfaces, ACS Appl. Mater.
Interfaces 9 (2017) 21532–21538.
[36] Y. Kaufman, S.-Y. Chen, H. Mishra, A.M. Schrader, D.W. Lee, S. Das,
S.H. Donaldson, J.N. Israelachvili, Simple-to-Apply wetting model to predict ther-
modynamically stable and metastable contact angles on textured/rough/patterned
surfaces, J. Phys. Chem. C 121 (2017) 5642–5656.
[37] E.M. Domingues, S. Arunachalam, J. Nauruzbayeva, H. Mishra, Biomimetic coating-
free surfaces for long-term entrapment of air under wetting liquids, Nat. Commun. 9
(2018) 3606.
[38] S. Arunachalam, R. Das, J. Nauruzbayeva, E.M. Domingues, H. Mishra, Assessing
omniphobicity by immersion, J. Colloid Interface Sci. 534 (2019) 156–162.
[39] M.C. Xu, G.U. Sun, C.J. Kim, Infinite lifetime of underwater superhydrophobic
states, Phys. Rev. Lett. 113 (2014).
[40] D. Beysens, Dew nucleation and growth, Compt. Rendus Phys. 7 (2006) 1082–1100.
[41] H. Zhao, D. Beysens, From droplet growth to film growth on a heterogeneous sur-
face: condensation associated with a wettability gradient, Langmuir 11 (1995)
627–634.
[42] P.G. de Gennes, Wetting: statics and dynamics, Rev. Mod. Phys. 57 (1985) 827–863.
[43] D.F. Cheng, T.J. McCarthy, Using the fact that wetting is contact line dependent,
Langmuir 27 (2011) 3693–3697.
[44] B. Emami, H.V. Tafreshi, M. Gad-el-Hak, G.C. Tepper, Predicting shape and stability
of air–water interface on superhydrophobic surfaces comprised of pores with ar-
bitrary shapes and depths, Appl. Phys. Lett. 100 (2012) 013104.
[45] L. Eykens, K. De Sitter, C. Dotremont, L. Pinoy, B. Van der Bruggen, How to opti-
mize the membrane properties for membrane distillation: a review, Ind. Eng. Chem.
Res. 55 (2016) 9333–9343.
[46] J.A. Dean, Lange's Handbook of Chemistry, McGraw-Hill, Inc., New york; London,
1999.
[47] J.B. Fortin, T.-M. Lu, Chemical Vapor Deposition Polymerization: the Growth and
Properties of Parylene Thin Films, Springer Science & Business Media, 2003.
[48] F.Y. Huang, A. Arning, Performance comparison between polyvinylidene fluoride
and polytetrafluoroethylene hollow fiber membranes for direct contact membrane
distillation, Membranes 9 (2019) 52.
[49] E.N. Fuller, P.D. Schettler, J.C. Giddings, New method for prediction of binary gas-
phase diffusion coefficients, Ind. Eng. Chem. 58 (1966) 18–27.
[50] R. Schofield, A. Fane, C. Fell, Heat and mass transfer in membrane distillation, J.
Membr. Sci. 33 (1987) 299–313.
[51] T. Sherwood, R. Pigford, C. Wilke, Mass Transfer, McGraw-Hill, New York, 1975.
[52] W. Haynes, CRC Handbook of Chemistry and Physics, CRC Press, 1999.
[53] D. Seo, A.M. Schrader, S.-Y. Chen, Y. Kaufman, T.R. Cristiani, S.H. Page,
P.H. Koenig, Y. Gizaw, D.W. Lee, J.N. Israelachvili, Rates of cavity filling by liquids,
Proceedings of the National Academy of Sciences, 2018.
[54] N.A. Patankar, Hydrophobicity of surfaces with cavities: making hydrophobic
substrates from hydrophilic materials? J. Adhes. Sci. Technol. 23 (2009) 413–433.
[55] F. Salvemini, F. Grazzi, A. Agostino, R. Iannaccone, F. Civita, S. Hartmann,
E. Lehmann, M. Zoppi, Non-invasive characterization through X-ray fluorescence
and neutron radiography of an ancient Japanese lacquer, Archaeological and
Anthropological Sciences 5 (2013) 197–204.