Drop impact upon micro- and nanostructured superhydrophobic surfaces

0.57 MB·English

by  Peichun Tsai

Drop impact upon micro- and nanostructured superhydrophobic surfaces
Peichun Tsai‡, Sergio Pacheco§, Christophe Pirat‡,†, Leon Lefferts§, and Detlef Lohse‡.
‡Physics of Fluids Group; §Catalytic Processes and Materials Group,
Faculty of Science and Technology, University of Twente, 7500AE Enschede, The Netherlands.
†Present address: Laboratoire de Physique de la Mati`ere Condens´ee et Nanostructures,
Universit´e de Lyon; Univ. Lyon I, CNRS, UMR 5586, 69622 Villeurbanne, France
(Dated: January 27, 2009)
We experimentally investigate drop impact dynamics onto different superhydrophobic surfaces,
consistingofregularpolymericmicropatternsandroughcarbon nanofibers,with similar staticcon-
9 tact angles. The main control parameters are the Weber number We and the roughness of the
0 surface. AtsmallWe,i.e. smallimpactvelocity,theimpactevolutionsaresimilar forbothtypesof
0 substrates,exhibitingFakirstate,completebouncing,partialrebouncing,trappingofanairbubble,
2 jetting, and sticky vibrating water balls. At large We, splashing impacts emerge forming several
satellitedroplets,whicharemorepronouncedforthemultiscaleroughcarbonnanofiberjungles. The
n
results imply that the multiscale surface roughness at nanoscale plays a minor role in the impact
a
J eventsfor small We >120 but an important one for large We ?120. Finally, we find theeffect of
ambient air pressure to be negligible in theexplored parameter regime We >150.
7
2

] INTRODUCTION hydrophobic (hydrophilic) substrates. Recently, super-
n hydrophobicsurfacesthroughcontrolledmicrostructures
y
(see Fig. 1c) have been developed, advanced, and in-
d Drop impact on a solid surface is ubiquitous and cru-
vestigated for their water-repelling [4, 13, 14] and hy-
- cial in a variety of industrial processes. The interplay of
u drodynamic slip applications [15, 16]. A water droplet
several experimental conditions often makes the impact
l
f dynamics surprising and too complex to elucidate [1]. when deposited on a micropatterned hydrophobic sur-
s. face can exhibit a metastable state of heterogeneously
The control parameters include the impact velocity, liq-
c wetting with air trapped between the liquid and surface,
i uiddensity,viscosity,surfacetension,andthewettability
s forming a “Fakir” drop [17, 18]. A spontaneous transi-
y and roughness of the substrates. In the past decade, the
tion can occur from “Fakir” to homogeneously wetting
h investigationshaveactivelyfocusedonthe impactsupon
“Wenzel” state with a smaller contact angle, even in the
p superhydrophobicsubstrates [2, 3, 4, 5, 6, 7, 8, 9] due to
[ their emergenceandopportunities inawide rangeofap- caseofzeroimpactvelocity[19,20]. Atthebreakdownof

superhydrophobicity,intriguingfillingdynamicsofwater
1 plications: forinstance,self-cleaningandanti-coatingfor
infiltration was found to possess multiple timescales and
v lab-on-chipdevices,ink-jetprinting,sprayingtechniques,
to tune the shape of the wetted areas by the geometric
8 andcoatingprocesses. The mostcommonlystudied sub-
2 dimensions of the micropillars [21]. In drop impact ex-
strates are silane-coated microstructures and hydropho-
2 periment, the threshold of the impact velocity V mark-
bic microtextures. These samples, possessing a combi- c
4 ing the transition from Fakir to Wenzel state is found to
. nation of chemical hydrophobicity and physical ordering
1 depend on the geometric dimensions of the hydrophobic
roughness,displaysuperhydrophobicity–withlargestatic
0 microstructures [7, 9]. So far, geometric arrangements
contactangleabove150◦andsmallcontactanglehystere-
9
and the impact velocity have been the main control pa-
0 sis within 5◦.
rametersforthedropimpactontosuperhydrophobicsur-
:
v In this study, we investigate the dynamics of drop im- faces [3, 6]. The change in surrounding air pressure has
i pact not only on superhydrophobic microstructures of
X notyetbeenexploredinthis context. We willalsostudy
controlled roughness but also on multiscale rough sur-
this effect here.
r
a faces of carbon nanofiber jungles (CNFJs). The aim is
to compare the impact behaviors upon both types of su-
The analyzed superhydrophobic substrates consist of
perhydrophobicsurfacesandto examinethe effectofthe
carbon nanofiber jungles. Carbon filaments are formed
detailsoftheroughness. Moreover,complementarytore-
catalytically in metallic catalysts, particularly in Ni, Fe
cent studies of drop impact using superhydrophobic sur-
andCobasedcatalysts,usedfortheconversionofcarbon
faces,ourworkexaminesthe pressureofthesurrounding
containing gases, e.g. in steam reforming of hydrocar-
air, as an additional control parameter, which recently
bons and Fischer-Tropsch synthesis [22]. In these cases,
has been shown to affect drop impact upon dry wetting
the carbon filament formation was detrimental for oper-
surfaces [10, 11, 12].
ationastheypluggedreactorsanddeactivatedcatalysts.
Chemical and/or geometrical surface modification af- Fiber type carbon nanomaterials can be classified into
fects the solid-liquid interactions. This offers a conve- three types, namely, Carbon Nanofiber (CNF), Carbon
nient control in surface physics. For example, physi- Nanotube(CNT)andSingleWalledNanotube(SWNT).
cal roughness increases (decreases) the contact angle for In CNF the graphitic planes are oriented at an angle to
2
the central axis, thus exposing graphite edge planes. In
CNTthegraphiticplanesrunparalleltothecentralaxis,
in this state only basal planes are exposed. CNT is also
referredto asMulti-WalledCarbonnanotubes(MWNT)
orparallelcarbonnanofibers. Ifthefiberconsistsofonly
one graphene sheet that is oriented in the direction par-
allel to the fiber axis, it is called a single-walled carbon
nanotube. We referthereaderto Ref.[23,24,25]forthe
history, studies, and reviews of catalytically generated
CNFs.
The diameters of CNFs, in general, range from a few
to hundred nanometers and their lengths vary from mi-
crometers to millimeters [23]. The novel physical and
chemical properties of CNFs include high surface areas,
large elasticity, and low electrical resistivity and hence
make CNFs good candidates for catalysts, catalyst sup-
ports, and selective adsorption sites [25]. Particularly,
the highly porous layers of entangled nanofibers (CNF
jungles)aresuitableascatalystsupportsforliquidphase
reaction, allowing fast mass transfer to catalytic active
particles deposited on the CNFs [26, 27]. In this work
entangled layers of CNFs are growndirectly on iron and
stainless steel metal foils, and we experimentally study
the dynamicaleffect of waterdropimpinging uponthese
CNFs.
EXPERIMENTAL SECTION
Superhydrophobic samples
Oursuperhydrophobicsubstratesarecatalyticallysyn-
thesized CNFJs and micropatterned polymers. Fig. 1
shows representative scanning electron microscope
(SEM) images of used samples. As revealed by Fig. 1
(a)and(b),catalyticallygeneratedCNFJsarecomposed
ofentangledandbundledcarbonnano-filamentsforming
FIG.1: (a)and(b): SEMpicturesoftworepresentativeused
uncontrolled, microscopic roughness. In contrast, Fig. 1
samples of carbon nanofiber jungles, showing uncontrolled,
(c) shows the top-view of an ordered, microstructured multiple-scale roughness. The porous fiber-bundles span a
substrate comprising micro-pillars periodically placed in few tens microns and the fine fibersare submicron thick. (c)
a structured lattice. A SEM top-view image of a representative microstructured
Synthesis of Carbon Nanofiber surfaces. The substrate consisting of 5 µm wide pillars, yielding a well-
definedroughness bythegeometric regularity [28].
carbonnanofibers(CNFs)wereproducedbycatalyticva-
pordeposition(CVD)fromcarboncontaininggasesusing
a metallic catalyst[25]. Two types ofcatalytic materials
were employed: they are iron (99.99%, Alfa Aesar) and
stainlesssteelType304(Fe:Cr:Ni70:19:11%,AlfaAesar) The CVD reactor consists of a verticalquartz reactor,
foilsof0.1mmthick. Roundsamplesofthese metalfoils withaporousquartzplatecentrallyplacedtosupportthe
(10 mm in diameter) were prepared by an electric dis- metalfoils. The temperature was raisedfromroomtem-
charge wire-cutting machine (Agiecut Challenge 2, GF peratureto600◦Catarateof5K/min. Thesampleswere
AgieCharmilles). The foils were degreasedultrasonically firstpre-treatedinahydrogen/nitrogenmixture(20%H2
in acetone and dried at room temperature before loaded with 80% N2) with a total flow of 100 ml/min for 1 hr
into the CVD chamber. In addition, hydrogen and ni- at 600◦C. After the pre-treatment,ethylene was fed into
trogen (99.999% purity, Praxair), and ethylene (C2H4, the reactor (20% with N2/H2, 20%/60%) at 600◦C for
99.95% purity, Praxair) were used for the CNF forma- 2 hrs (with stainless steel foil) or 3 hrs (with iron foil).
tion without further purification. During the heating process at the high temperature, the
3
decomposition of C2H4 led to the growth of carbon fila- controlthe pressure of the surrounding air. The dynam-
ments upon the metallic catalysts. The concentration of ics of drop impact was recorded by a high-speed camera
hydrogen and the total flow were kept the same during (Fastcam SA1, Photron) with a recording rate ranging
pre-treatment and deposition. Finally the ethylene and from1500to30000fps(framespersecond). Theimpact
hydrogengasstreamswereshutoffandthewholesystem velocity and the droplet size were determined from the
wascooleddownto roomtemperature under nitrogenat captured images.
a rate of 10 K/min. Several control parameters influence the impact dy-
Preparation of the polymeric microstructures. namics, for example, droplet size, liquid viscosity µ, im-
Micro-patternedsubstratesare composedofPDMSelas- pactvelocityV . Wedescribetheseeffectsintermsofdi-
i
tomer (Polydimethylsiloxance, RTV 615 rubber compo- mensionless numbers: the Weber number e, the ratio
W
nent A and curing agent B, GE Bayer Silicones). The ofkineticenergytosurfaceenergy,characterizingthede-
fabrication of precise and controllable microstructures formability of the droplet; the Reynolds number e, the
R
was achieved via a micro-molding method [28]. This ratio of inertia to viscosity effect:
technique is to cast a polymeric film from solution on a
ρRV2 ρRV
molding wafer of desirable micropatterns. The PDMS e= i , e= i. (1)
W σ R µ
films were obtained by mixing the rubber component
A with the curing agent B (10 : 1 wt/wt). The mix- Here R is the radius of the liquid drop, V is the impact
i
ture was degassed and then poured onto the mold ( 1 velocity, ρ is the liquid density, σ is the surface tension,
∼
mm) and cured in an oven for 3 hrs at 85◦C. The cast- and µ is the liquid viscosity. In this paper, we did not
ing molds consisted of diverse arrays of micro-patterns, simultaneouslychange eand ebyusingdifferentliq-
which can produce periodically arranged, round or rect- uids; instead, these paWrametersRare only different ways
angular polymeric micro-pillars of height h, diameter or to non-dimensionalize the velocity V of a milli-Q water
i
width w, and the interspacing a. With different lattice droplet. The Ohnesorge number, comparing viscous and
arrangements, h, a, and w, we controlled the roughness capillary forces, Oh = √ e/ e = µ/√ρRσ is small
of the polymeric microstructures. We characterize the O(10−3) in our experimWents.RIn addition, we control
surface roughness f by the ratio of the total surface a∼nd measure the pressure of the surrounding air, noted
R
area to that projected on the horizontal plane For the asP . Forthe data presentedhere wespecify the cases
air
experiments presented here, w = 5 µm, h = 6, 10, and under reduced air pressure, otherwise P = 101.3 kPa
air
20 µm,andaisvariedbetween1.27 µmand5 µm. The at the standard ambient pressure.
corresponding spans between 1.5 and 9.
f
R
Contact Angle Measurements. The contact an-
gle measurements were performed with a milli-Q water RESULTS AND DISCUSSION
dropletof4 µl,matchingthevolumeofimpactingdroplet
of 1 mmwideinradius,withtheLaplace-Youngfitting In this section, we compare impact events on CNFs
∼
method(OCA20,DataPhysics). Catalyticallygenerated and on microstructured surfaces with similar wettabil-
CNFJs exhibit random topography of multiscale rough- ity. Fig. 2 shows a small fraction of the phase space
ness, and thus we need to ensure the homogeneous wet- and the time evolution of the impact dynamics for wa-
tability of such surfaces when performing drop impact ter drops impacting upon CNFJs of θ = 155 3◦.
SCA
±
experiments. The static contact angle θSCA of CNFJs In this small e-number regime (> 2.5), the impact
W
reported here was measuredat different locations on the events include non-bouncing Fakir droplet, complete re-
same sample, both before and after the drop impact ex- bound, sticky wetting ball, and partial rebound, gener-
periment. Thescatteringofthesedataisusedtoestimate allyinthissequenceas eisgraduallyincreased. When
W
the error in θSCA. a water droplet is gently deposited, due to small kinetic
energythe dropletmaintainsinaCassie-wetting“Fakir”
state with air trapped underneath the drop and thus a
Drop impact experiments high contact angle during the whole impacting process,
as shown in Fig. 2f. In this heterogeneous state, air is
The generalexperimentalprocedure consistsin releas- trapped underneath the droplet and between the porous
inganimpingingmilli-Qwaterdropletfromafineneedle fibers. The non-wetting nature of the sample limits the
(0.1mminnerdiameter)withasyringepump(PHD2000 spreadingofthedropandsometimesevenleadstoacom-
Infusion,HarvardApparatus)atdifferentheightstovary plete rebound (Fig. 2e). At high V ? 0.28ms−1 a wet-
i
theimpactspeeduponthesolidsurface. Thebalancebe- tingtransitioncanoccurasthekineticenergyovercomes
tween the surface tension anddroplet gravitationalforce thesurfaceenergy,associatedwiththeliquidsurfaceten-
setsthedropletsize,whichis 1mminradiuswithin5% sion, and the droplet turns to the completely wetting
≈
deviation. The whole experimental setup was enclosed Wenzelstate,inwhichitispinnedonthesurface(Figs.2c
in a chamber, connected with a vacuum pump so as to and d).
4
$'#
789
) $ partial rebound
6
)450 (c)
3!'#
2
1
)40+ ! 0ms 3.3ms 8ms 12ms 14ms 16ms
)3 sticky vibrating ball
6"'#
(d)
"
!"" !#" $"" $#" %"" %#" &"" &#"
()*+,-./0+123)450() 0ms 3.3ms 5.3ms 10.7ms 16ms 18ms
complete rebound
!"# 0
9>; (e)
;
!# !"!)
:
+ 0ms 2ms 8ms 12ms 14.7ms 17.3ms
;09!"!( non-bouncing
2
+
./0/5!"!' (f)
-
=/
4!"!% 0.33ms 3ms 5.7ms 8.3ms 15ms 21.7ms
.
9
#<
!
0
!"# !"#$ !"% !"%$ !"& !"&$ !"'
!#
*+,-./01234.5/670809+: ;
5
FIG. 2: (a) Fraction of the phase space of the impact dynamics upon a carbon nanofiber substrate (Fig. 1a) with a static
contact angle of (155±3)◦ at small We numbers, with the time evolutions of the impact events shown in (c)-(f). (b) shows
thecorresponding impact velocity Vi vs. theinversecontact time.
The average contact time before the droplet bounces
off the surface is 12.5 ms (based on the marked data
( ) in Fig. 2b). This average value is consistent with
•
the value 14 ms found by a recent study with a multi-
walledcarbonnanotubearraywithastaticcontactangle
of 163◦ [29]. The coexistence of different dynamical be-
haviors for 0.28 ms−1 > V > 0.35 ms−1 illustrates the
i
complexity of the impact process, showing the interplay
of several parameters and indicating that other factors
canaffectthephasespace. Forinstance,fromourresults
we note that the complete rebound regime happens at
higherV between0.5 ms−1 and1.1 ms−1 foramorehy-
i
drophobicCNFsubstratewithalargerθ =(163 4)◦.
SCA
±
Fig.3 showsthe phase diagramfor e between1 and
W
10, corresponding to high e, when a water droplet im-
R
pactsuponaCNFsurfaceofθ =(152 3)◦ shownin
SCA
±
Fig.1b. Inthis e-range,asdiscussedabove,theimpact
scenariocandisWplayacomplete bouncingorapinning of FIG. 3: (a) Impact dynamics upon a substrate of CNFJs,
(shown in Fig. 1b) with a static contact angle of (152±3)◦.
contact line resulting in wetting. Also, trapping an air
The insets show the snapshots of the impact events: sticky
bubble or jetting can happen due to the development of
ballattheWenzelstate((cid:4)),partialrebound(N),trappingof
an air cavity. Fig. 4 shows the detailed evolutions of an air bubble((cid:7)),and jetting (∗).
(a) trapping an air bubble and (b) emitting a fast jet.
For our e-range between 2 and 8, on impact a sur-
W
face capillary wave is excited and the droplet deforms
5
FIG.4: Timeevolutionsoftheimpacteventsof(a)trappinganairbubbleatθSCA=(163±3)◦,We=5.3andRe=571,and
(b) jetting at θSCA=(152±3)◦, We =7.7 and Re =756.
like a pyramid around 1.5 ms. The oscillation of the drophobic surface consisting of regular micro-pillars of
surface deformation often makes a toroidal droplet pro- h = 37µm in height, and a = 3µm in the interspac-
ducing a cylinder-like cavity; see the snapshot at 3.2 ms ing distance between pillars [7]. This Ref. also reports
in Fig. 4a and that at 2.2 ms in Fig. 4b. Then the fol- thatthecritical e-numberabovewhichdropletsejectis
W
lowingdynamicsofthe cavityrestorationdeterminesthe about700foraflathydrophobicsolidandnofragmenta-
subsequent behaviors: in (a) at a lower V the surface tionofdropletsevenat e =1000foraflathydrophilic
i
W
wave restores the top of cavity sooner and closes it up solid. With these observations, the authors conjecture
producing a residual void; in (b) at a higher V the fast thattheroleofthefilmofairplaysacrucialrolebecause
i
collapseofthecavityproducesafastjet. Ourdatareveal the non-wetting Fakir state promotes an air film for su-
that the shooting jets can be as fine as 30µm in radius perhydrophobic solids. However, our investigations find
and as fast as 9 V . Similar observations underlying theeffectofairpressureonthesplashingtobenegligible
i
×
thesamemechanismhaverecentlybeeninvestigatedwith in this parameter regime; Fig. 5b reveals the same kind
ordered, superhydrophobic microstructures for the e- of splashing at a low air pressure,P =14.7 kPa, with
air
W
number ranging between 0.6 and 16 [6]. The analogy of similar e and e to those in Fig.5a. Our experiments
W R
the impact phenomena suggests that the formation and at different P with similar e numbers resemble the
air
W
the collapse of the air cavity may be insensitive to the same type of splashing. Hence, the data suggest that
details of surface roughness. the mechanism of aforementioned splashing behaviors is
duetotheeffectofsurfaceroughness,andnotduetothe
In the higher e-number range (between 90 and 140)
surrounding air.
W
splashing impacts occur as shown by Fig. 5 for CNFs.
Here, the “splashing” phenomenon refers to the forma- Fig. 6 shows a partial rebound for an impacting drop,
tion of many satellite droplets, which merge during the withsimilarcontrolparametersP =101.3 kPa, e=
air
W
spreading and/or contracting stages of the water film 120and e=2882,butnowforapolymericmicrostruc-
R
from our observations. In Fig. 5a under 1 atm, at 1 tured surface consisting of round pillars of a=1.27 µm,
ms tiny droplets radially emit while the water sheet still w = 5 µm and h = 10 µm. As observed from the snap-
spreads outwards with wavy perturbations, resembling shots, the rim is not destabilized in the same way as
“fingers”,attheedgeofwaterfilm. Severaldropletsorig- for the unstructured CNFJs (compare Fig. 5 and 6 at
inated from the wavy perimeter develop between 1.8 ms t 1 ms). Instead of perimetric fingers, wavy bumps
≈
and5.8mswhilethemainwatersheetcontracts,shaping develop and eventually merge together at the center to
into a partial rebound. A similar splashing phenomenon form an elongating water fountain. The pinning of the
has lately been observed at e 160 with a superhy- water wets the microstructured surface and breaks the
W ≈
6
FIG. 5: Snapshots of the splashing evolutions on CNFs of θSCA = (163±3)◦ at different air pressure (a) Pair = 101.3 kPa,
We =115.3 and Re =2784, and (b) Pair = 14.7 kPa, We = 141.7 and Re = 3060. The inset bars indicate a length scale
of 2 mm.
0.2 ms 0.4 ms 0.9 ms 2.2 ms 3.7 ms 4.5 ms 5.7 ms
6.2 ms 6.9 ms 8.9 ms 10.1 ms 17.7 ms 19.4 ms 20.7 ms
1 mm
FIG. 6: Time evolutions of impact dynamics upon an ordered, microstructured polymeric substrate of θSCA = (147±3)◦ at
Pair = 101.3 kPa, We = 120 and Re = 2882. This surface is comprised of round pillars of a = 1.27 µm, w = 5 µm and
h=10 µm arranged in a rectangular lattice and yields the roughness R =5.
f
fountain while surface tension is in action, resulting in face roughness of CNFs display pronounced formations
a few breaking-up droplets. Consistent with the results of satellite droplets, comparable with the impacts upon
in Ref. [7], our experiment at e = 147, e = 3524 ordered micropatterned substrates under alike parame-
W R
and P = 101.3 kPa shows that a few ( 5) satel- ters. This finding suggests that the smaller length scale
air
≈
lite droplets develop during the contraction of the water ofcarbonfibersmayenhancetheinstabilityofperimetric
sheet upon a micropatterned substrate of a = 5 µm, wavy fingers and thus produce stronger fragmentations.
w = 5 µm and h = 20 µm with a roughness = 2.8.
f
ThecomparisonsbetweenthesedifferentexperiRmentsre- Fig.7displaysthephasespaceofwaterimpactdynam-
veal that roughness has a profound effect on splashing ics onto polymeric microstructures of θSCA =(147 3)◦
±
impacts. In addition, the multiscale, uncontrolled sur- explored at different air pressure Pair. Different col-
ors depict different surface roughness controlled by
f
R
7
CONCLUSIONS
Rf6
s,
s In comparison with the drop impacts performed upon
e 4
n microstructured hydrophobic surfaces, for e > 10 and
h W
ug 2 under the atmospheric pressure impacts upon CNFs of
o multiscale roughness reveal similar dynamical behav-
R
0 iors including Cassie-wetting Fakir droplet, complete re-
1 2
10 We 10 bound, partial rebound, jetting, and trapping bubbles.
)120 This comparison implies that impact evolutions are in-
a
P100 sensitive to the details of the roughness of superhy-
k
e ( 80 dropohobic surfaces at small We-number. Rather than
ur 60 a bouncing-pinning transition, driven by a threshold of
ss 40 impact velocity, as described in Ref [7] for regular mi-
e
r crostructures,we noticedthe coexistence ofcomplete re-
P 20
101 We 102 bound and sticky wetting droplet for CNFs at 1.1 >
e > 1.9. This observation indicates that the details
6000 Wof roughness can influence the transitional boundary in
the phase space of impact events.
4000
e Partial rebounds occur in the e-range between 10
R W
2000 and130forwaterdropshittinguponmicropatternedsub-
strates with h = 20 µm and a roughness ranging from
0 2 to 10. Air pressure and the surface roughness have
1 2
10 We 10 hardly any effect in this partially rebouncing regime. At
higher e ? 120, profound splashing impacts forming
W
severalsatellitedropletstakeplaceforsuperhydrophobic
FIG. 7: Phase space of water drop impact onto polymeric
microstructures. Differentcolorsindicatedifferentroughness, CNFs. This splashing mechanism is not suppressed by
R =1.9(green), 2.6(magenta), 3.0(black), and5.0(cyan), thedecreaseinairpressure,whichwasfoundtodebilitate
f
determinedbythegeometricarrangementsofthemicropillars thecoronasplashingonsmoothwettingsurfaces[10,12].
of h = 10µm. Different symbols present various dynami- Beyond e ? 140, in contrast to the violent splashing
cal impact behaviors: jetting (∗),complete rebound(◦), par- W
events on the roughCNFJs, on imcrostruture substrates
tial rebound(△), stickyball (2), formation of afew satellite
thefragmentationismuchlesspronounced,revealingthe
dropletswhilethemainimpactrevealsapartialrebound((cid:3))
or a sticky wetting ball ((cid:1)). importance of sub-micron roughness for splashing.
ACKNOWLEDGMENTS
The authors thank Alisia M. Peters and Hrudya Nair
for their helps on the preparations of the polymeric mi-
crostructures. We also gratefully acknowledge the Mem-
braneScienceandTechnologyGroupattheUniversityof
the geometric arrangement of micropillars of h=10µm.
Twente for the usage of the contact angle measurement
Similar impact phenomena to those upon CNFs are ob-
device.
served, including jetting, complete rebound, and partial
rebound for small e > 10. For 70 > e> 110 par-
W W
tial rebound occurs with a high deformability stretch-
ing the water droplet into a elongating column. As dis-
cussedinRef.[7],athresholdofimpactvelocitymarking
[1] A. L. Yarin,Annu.Rev.Fluid Mech. 38, 159 (2006).
a transition from complete bouncing to wetting or par- [2] D. Qu´er´e, Ann.Rev.Materials Research 38, 71 (2008).
tially pinned drop is identified as a function of the size [3] D. Richard and D.Qu´er´e, EPL 50, 769 (2000).
ofthesuperhydrophobicmicrotextures. Interestingly,for [4] D. Richard, C. Clanet, and D. Qu´er´e, Nature 417, 811
the small roughness = 2, we observe coexistence of (2002).
f
R [5] A.L.Biance,C.Clanet,andD.Qu´er´e,Phys.Rev.E69,
different dynamical processes in some range of impact
016301 (2004).
velocities. At large e (between 200 and 300) a few
W [6] D.Bartolo,C.Josserand,andD.Bonn,Phys.Rev.Lett.
satellite droplets form while the main impact can par-
96, 124501 (2006).
tially rebound ((cid:3)) or behave like a sticky wetting ball [7] M.Reyssat,A.P´epin,F.Marty,Y.Chen,andD.Qu´er´e,
(shown by (cid:1)). EPL 74, 306 (2006).
8
[8] Y.-C. Jungand B. Bhushan, Langmuir 24, 6262 (2008). [21] C. Pirat, M. Sbragaglia, A. M. Peters, B. M. Borkent,
[9] J. Hyv¨aluomaand J. Timonen, EPL 83, 64002 (2008). R. G. H. Lammertink, M. Wessling, and D. Lohse, EPL
[10] L. Xu, W. W. Zhang, and S. R. Nagel, Phys. Rev. Lett. 81, 66002 (2008).
94, 184505 (2005). [22] K.P.D.JongandJ.W.Geus,Catal. Rev.Sci.Eng.42,
[11] L. Xu,Phys.Rev.E 75, 056316 (2007). 481 (2000).
[12] L. Xu, L. Barcos, and S. R. Nagel, Phys. Rev. E 76, [23] A. V.Melechko, V. I.Merkulov, T. E. McKnight, M. A.
066311 (2007). Guillorn,K.L.Klein,D.H.Lowndes,andM.L.Simpson,
[13] T.Onda,S.Shibuichi,N.Satoh,andK.Tsujii,Langmuir J. Appl.Phys. 97, 041301 (2005).
12, 2125 (1996). [24] K.P.deJongandJ.W.Geus,Catal. Rev.Sci.Eng. 42,
[14] A.Lafuma and D. Qu´er´e,Nat. Mater. 2, 457 (2003). 481 (2000).
[15] A. Steinberger, C. Cottin-Bizonne, P. Kleimann, and [25] N. M. Rodriguez, J. Mater. Res. 8, 3233 (1993).
E. Charlaix, NatureMaterials 6, 665 (2007). [26] M.J.LedouxandC.Pham-Huu,Catal.Today102–103,
[16] E.Lauga,M.P.Brenner,andH.A.Stone,Microfluidics: 2 (2005).
The No-Slip Boundary Condition (Springer, New-York, [27] M.L.Toebes, T.A.Nijhuis,J.H´ajek,J.H.Bitter,A.J.
2007), chap. 19. vanDillen,D.Y.Murzin,andK.P.deJong,Chem.Eng.
[17] D.Qu´er´e, Nat. Mater. 1, 14 (2002). Sci. 60, 5682 (2005).
[18] D.Qu´er´e, Rep.Prog. Phys. 68, 2495 (2005). [28] L. Vogelaar, R. G. H. Lammertink, and M. Wessling,
[19] M. Sbragaglia, A. M. Peters, C. Pirat, B. M. Borkent, Langmuir 22, 3125 (2006).
R.G.H.Lammertink,M.Wessling,andD.Lohse,Phys. [29] Z. Wang, C. Lopez, A. Hirsa, and N. Koratkar, Appl.
Rev.Lett. 99, 156001 (2007). Phys. Lett. 91, 023105 (2007).
[20] L. Courbin, E. Denieul, E. Dressaire, M. Roper, A. Aj-
dari, and H. A.Stone, Nat. Mater. 6, 661 (2007).