1

An Experimental Investigation on the Effects of Surface Wettability on

Water Runback and Ice Accretion over an Airfoil Surface

Rye M. Waldman1, Haixing Li2, Hao Guo3, Linkai Li4 and Hui Hu5()

Department of Aerospace Engineering, Iowa State University, Ames, Iowa, 50010

Aircraft frequently encounter supercooled liquid water droplets when flying through the

cloud layer during take-off and landing, and the ice accretions on aircraft wings, control

surfaces, and instruments pose a safety and performance threat. Anti-icing and de-icing

strategies are important for managing the negative effects of ice accretion and extending the

operating envelope of aircraft in cold conditions. Active anti-icing and de-icing strategies

often utilize heating to melt ice or mechanical boot actuation to break off ice, and the power

used by these methods necessarily result in a drop in efficiency; therefore, passive strategies

for inhibiting ice accretions are highly desirable for improving the operation of aircraft in

icing conditions while reducing the operational cost. Several surface coatings have been

proposed as candidates for passive anti-/de-icing applications, such as superhydrophobic or

self-lubricating coatings. The differing chemistries of the coatings result in variations of the

surface wettability, thereby altering the water runback process and the ice accretion

geometry. In this paper, we present experiments performed in the ISU-UTAS Icing Research

Tunnel on water runback and ice accretion on NACA 0012 airfoils with varied surface

treatments. The airfoil surface is modified using different coatings to produce a range of

surface wettabilities. The water and ice accretion dynamics on the wing are quantified using

high-speed imaging. We discuss how the surface wettability alters the balance between the

water adhesion stresses and the aerodynamic stresses that drive the surface water transport

I. Introduction

Aircraft operating in cold conditions are at risk of ice accretion over aircraft wings that significantly degrade

flight performance and may pose a danger to safety [1]. To mitigate the detrimental effects of accreted ice, various

de-icing/anti-icing techniques are employed to eliminate, reduce, or prevent ice accretion on the aircraft. Traditional

approaches to de-icing have utilized heating or de-icing fluids to melt ice and mechanical devices to breakup and

shed the accreted ice. The energy requirements of the heating and mechanical anti-icing devices must be minimized

to realize an efficiency benefit from the applications of anti-icing techniques. Passive anti-icing techniques are

highly desirable because they may reduce or eliminate the cost penalty associated with active anti-icing techniques.

Recently, superhydrophobic surfaces have been proposed as candidates for anti-icing applications and experiments

have demonstrated that some superhydrophobic coats do have icephobic properties [2] , that droplets can bounce off

of cold superhydrophobic surfaces without phase change [3], and some authors assert that superhydrophobicity

directly implies anti-icing functionality [4]. The superhydrophobicity of the surface results from a combination of

chemical hydrophobicity with a micro- or nano-textured surface. The structure of the surface plays an important role

both in the wettability of the surface and in the ability of the surface to resist ice accretion [2]. An alternate surface

coating strategy attempts to reduce the surface adhesion strength by using a lubricating fluid impregnated in the

coating matrix [5,6]. In these coatings, a lubrication fluid (such as an oil) prevents a strong bond between the ice and

the surface. Diffusion replenishes any lubricant lost from the surface (e.g., lubricant is lost during an ice shedding

event), making these surfaces robust and self-healing.

1 Postdoctoral Research Associate, Department of Aerospace Engineering, AIAA member. 2 Graduate Student, Department of Aerospace Engineering. 3 Graduate Student, Department of Aerospace Engineering. 4 Graduate Student, Department of Aerospace Engineering. 5 Professor, Department of Aerospace Engineering, AIAA Associate Fellow, Email: huhui@iastate.edu

Dow

nloa

ded

by I

OW

A S

TA

TE

UN

IVE

RSI

TY

on

Nov

embe

r 27

, 201

6 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/6.2

016-

3139

8th AIAA Atmospheric and Space Environments Conference

13-17 June 2016, Washington, D.C.

AIAA 2016-3139

Copyright © 2016 by Rye M. Waldman, Haixing Li, Hao Guo, Linkai Li and Hui Hu. Published by the American Institute of Aeronautics and Astronautics, Inc., with permission.

AIAA AVIATION Forum

2

The choice of surface treatment on an airfoil will determine how impinging water interacts with the wing

surface. The adhesion forces between the water and the surface resist the aerodynamic stress that drive surface water

flows, therefore the surface chemistry of the wing plays a critical role in the dynamics of surface water runback and

the shape of ice accretions [7,8].

In the present study, we investigate the role of surface wettability on the dynamics of ice accretion by

comparing measurements of water runback and ice formation on an airfoil coated with two different surface

treatments to produce a range of surface wettability. We first review how surface wettability affects adhesion forces.

Then, we present experiments performed in the ISU-UTAS Icing Research Tunnel on water runback and ice

accretion on a NACA 0012 airfoil model coated with an enamel paint and with a superhydrophobic treatment, as

well as measurements of the advancing and receding contact angles on each of these surfaces. The water and ice

accretion dynamics on the wing are quantified using high-speed imaging. We discuss how the surface wettability

alters the balance between the water adhesion stresses and the aerodynamic stresses that drive the surface water

transport.

II. Surface wettability

Wettability describes how a liquid interacts with a solid surface. A droplet sitting on a surface is a three-phase

system of the liquid in the drop, the solid surface, and the surround gas. The shape of the droplet is determined by

the liquid-solid, liquid-gas, and solid-gas interaction energies [9]. The fundamental equation of wetting is Young’s

equation, relating the interface energies of the phases to the thermodynamic equilibrium contact angle θY, which is

expressed as

cos .LG Y SG LS (1)

Here, γLS, γLG, and γSG are the liquid-solid, liquid-gas, and solid-gas surface tensions, respectively. The contact

angle determined by Young's equation applies when the droplet is under static conditions without external forces.

If an external force is applied to the droplet, such as aerodynamic drag resulting from wind blowing past the

droplet as shown in Figure 1a, or a body force with a component tangential to the surface as shown in Figure 1b,

then the droplet can remain pinned if the contact angle at the front and back sides of the droplet differ, thereby,

allowing capillary forces to hold the droplet in place. Beyond a critical value, the external force overcomes the

capillary forces pinning the droplet as the contact angles at the front and back reach their critical values—namely,

the advancing contact angle θadv, and receding contact angle θrec—and the droplet begins to slide.

The capillary force that resists droplet sliding can be estimated by considering the surface tension acting at

the contact line as shown schematically in Figure 2. Here, the droplet is approximated as a spherical cap with a

radius R, with a circular shaped solid-liquid interface with a radius of r = R sin(θM), where θM = (θadv+θrec)/2 is the

averaged angle of the advancing and receding contact angles. The component of the surface tension acting in the

plane of the surface fc is determined by the local contact angle, which must vary continuously traversing around the

contact line from θadv to θrec. To keep the analysis tractable, a simple model is used for the contact angle as a

function of the position around the contact line that has the requisite smooth contact angle variation and matches the

advancing and receding contact angles at the front and back of the contact line. This allows the component of the

capillary force that lies in the plane of the surface to be calculated as:

cos2 2

adv rec adv rec

, (2)

cosc LGf . (3)

Here, the angle φ is the angle relative to the direction of the motivating force (or direction of motion), and fc

is the component of the capillary force in the plane of the surface as depicted in Figure 2. By integrating about the

contact line the component of surface tension that lies in the surface plane, the net force acts against the external

force on the droplet Fcap is given by:

Dow

nloa

ded

by I

OW

A S

TA

TE

UN

IVE

RSI

TY

on

Nov

embe

r 27

, 201

6 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/6.2

016-

3139

3

2

0

1

cos cos cos d2 2

2 J sin2 2

sin sin , for 12 2 2

adv rec adv rec

cap LG

adv rec adv rec

LG

adv rec adv rec adv rec

LG

F R

R

R

, (4)

where J1 is the Bessel function of the first kind. For contact angle hysteresis less than 180º, J1(θ) can be

approximated as sin(θ)/2 with less than 10% error. Thus, the final form of Eqn. (4) adequately illustrates the

behavior of the capillary force which acts on a spherical cap droplet to resist translation along the surface. First, it is

important to note that the geometry of the problem places restrictions on the allowable range of contact angle;

namely, the contact angle must everywhere strictly lie between 0º and 180º. The mean contact angle (θadv+θrec)/2 is

further restricted by must lie between 0º and 180º, and the contact angle hysteresis (θadv–θrec) cannot be larger than

twice the distance between the mean contact angle and the nearest bound. Eqn. (4) shows that the capillary force

resisting droplet motion along the surface is maximized when the arguments of the sinus functions are 90º; namely,

the droplet sits with a mean contact angle of 90º and the largest possible contact angle hysteresis. Additionally, the

capillary force will vanish when either of the arguments of the sinus functions vanishes; namely, when the mean

contact angle approaches 0º or 180º, or when the contact angle hysteresis approaches 0º, the droplet will slide freely

over the surface. It is important to note that this equation for capillary force was derived with the assumptions of a

spherical-cap droplet and circular contact line, which is an improvement from the analysis of Quéré et al. [10] who

only considered the two-dimensional projection of the droplet shape and arrived at a simpler, conservative equation

for the capillary force resisting motion:

sin cos cos2

adv rec

cap LG adv recF R

(5)

It should be noted that, Eqn. (5) would overpredict the capillary force resisting droplet motion. Here, the

adhesion force resisting aerodynamic stress is determined by geometry of the water, R, the thermodynamics of the

surface chemistry and water, γLG, and the dynamic contact angles θadv, and θrec.

III. Experimental methods

A. Surface treatments

Two different surface coatings were used in this study, representing an extreme difference in surface

wettability. Each coating was applied to a 38 × 38 mm2 square sample substrate for wettability testing, and also

applied to one half span of an airfoil model for wind tunnel testing. Each surface coating was applied to the test

substrate and the wind model using the same procedure to ensure consistency in the surface properties between the

different experiments.

A readily available all-weather protective spray-on enamel was used (Rustoleum, Flat Protective Enamel,

White) as the reference surface. The surface was prepared by first priming the surface with several coats of spray-on

sandable primer. The primer provides a strong adhesion of the enamel to the test surfaces, and with sanding, ensures

that the surfaces are uniform and free from imperfections. Several coats of the enamel were then applied to the

primed surfaces and the final surfaces were wet-sanded using 1000-grit sand paper. Wet sanding ensures that

microroughness of the surface finish are consistent from specimen to specimen, with a characteristic roughness scale

determined by the sandpaper grit size, which in this case is about 25 µm.

The second surface that was tested is a spray-on superhydrophobic coating (Hydrobead, Hydrobead Standard

and Hydrobead Enhancer). This coating consists of a single-step spray-on base coating which provides a

superhydrophobic base layer, and an optional spray-on top coat that further increases the hydrophobicity. In this

study, both the Standard and Enhancer coatings were used. Because the coating is translucent, the Hydrobead

coating was applied on top a Rustoleum enamel base coating to provide uniformity in imaging.

Dow

nloa

ded

by I

OW

A S

TA

TE

UN

IVE

RSI

TY

on

Nov

embe

r 27

, 201

6 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/6.2

016-

3139

4

B. Surface wettability measurements

To characterize the wettability of the surface coatings used in the wind tunnel experiments, the contact angle

of the advancing and receding contact lines of water droplets on each substrate was experimentally measured. The

experiment configuration is depicted in Figure 3. A high speed camera (PCO Tech, Dimax) using a 105 mm macro

lens (Nikon, 105mm Nikkor 2.8D) was positioned with a view from the side of the 38 mm substrate mounted on a

vertical translation stage. A syringe was used to pump deionized water through a needle mounted above the test

substrate, forming a drop that could be expanded and contracted, thereby creating advancing and receding contact

lines, respectively. A 20W led lamp (Dot Line RPS Studio, RS-5410) illuminating a piece of frosted glass provided

back-illumination to provide high-contrast images for digitizing the drop shape.

Each substrate was tested by forcing water through the needle using the syringe, thereby creating an

expanding droplet with an advancing contact line. Then, the syringe was used to pull the water back into the needle,

causing the droplet to shrink and the contact line to recede. The images recorded with the high speed camera were

analyzed using a custom MATLAB code. The digitization process is illustrated in Figure 4. The time sequence of

the droplet images were imported into MATLAB, where for each image, a line-detector identified the pixels on

substrate surface and the droplet edges. Then, linear and quadratic best-fit curves were computed for the substrate

surface and the droplet edge, respectively, using a least squares fit to the corresponding edge pixels. The intersection

of the two fit curves denotes the contact line location, while the relative angle between the lines at their intersection

determines the contact angle. By tracking the contact line position and the contact angle throughout the image

sequence, the contact angle could be computed as a function of the contact line speed.

C. Icing tunnel measurements

A comparative study by using the same airfoil model with different surface treatments (e.g., with vs. without

superhydrophobic coating) subjected to icing conditions was performed to examine the effects of the hydrophobicity

of the substrate surface on the droplet—surface interactions under icing conditions. The experimental study was

performed in the ISU-UTAS Icing Research Tunnel (ISU-UTAS-IRT) available in the Aerospace Engineering

Department of Iowa State University (ISU). ISU-UTAS-IRT has a test section of 16 × 16 inches2 (406 × 406 mm2)

in cross-section, and the walls of the test section are optically transparent. It has a capacity of generating a maximum

wind speed of 50 m/s and air temperature of –20º C. An array of eight pneumatic atomizing spray nozzles (Ikeuchi,

BIMV11002)_are installed at the entrance of the contraction to inject micro-sized water droplets (15–100µm in size)

into the airflow. By adjusting the water flow rate through the spray nozzles, the liquid water content (LWC) in the

icing wind tunnel could be changed up to 10 g/m3, which can be used to simulate various atmospheric icing

phenomena (e.g., from dry rime ice to extremely wet glaze ice conditions).

The facility is instrumented with a data acquisition system to record all of the relevant experimental

conditions. Because ice accretion would quickly disrupt a pitot-static probe in the test section during icing

experiments, the test section freestream velocity is monitored indirectly via calibrated static pressure wall taps

mounted upstream of the contraction and at the test section entrance. The pressure drop across the contraction is

monitored using a differential pressure transducer (Setra, 239) and the static pressure is monitored at the test section

using an absolute pressure transducer (Setra, 270). The air temperature is monitored using an (Omega, DP-41) and a

K-type thermocouple mounted in the air, upstream of the spray system to avoid ice accretion. The mass flow of

water to the spray system was monitored using a water flow meter (Omega, FLR-1605A). All analog signals were

digitized using a 16-bit data acquisition card (National Instruments, USB-6218) and monitored using a MATLAB

script to compute and display the ISU-UTAS-IRT environmental conditions in real time.

Figure 5 gives the schematic of the experiment configuration used for the icing wind tunnel experiments.

Four led lamps (Dot Line RPS Studio, RS-5410 (x2), RS-5610, and RS-5620) were used to provide soft flood

illumination while a high-speed camera (PCO 1200HS, PCOtech) recorded images of the water and ice on the wing.

The camera was mounted perpendicular to the wing planform. The camera was calibrated by manually digitizing the

chord of the wing at the interface between the enamel and superhydrophobic coating. The timing of the images was

precisely controlled by triggering the camera from a digital delay generator (Berkeley Nucleonics Corporation,

BNC-575). The test model used in the present study is a rapid prototyped symmetric NACA0012 airfoil model with

a chord length of 152.4 mm. The test airfoil model spans the test section, therefore, pseudo 2D flow is assumed.

The airfoil model was tested at = 0° angle of attack, thereby, allowing the runback glaze icing process to

be observed from above the airfoil. In the experiments, the air temperature in the icing wind tunnel is set to T∞ = –

Dow

nloa

ded

by I

OW

A S

TA

TE

UN

IVE

RSI

TY

on

Nov

embe

r 27

, 201

6 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/6.2

016-

3139

5

8°C, –4°C, and 10°C (above freezing). The freestream velocity set to U∞ = 20, 30, 40 and 50 m/s. The mass flow

rate of water sprayed into the icing tunnel was adjusted according to the wind speed to achieve a liquid water

content of LWC = 4 g/m3, and with droplet sizes MVD ≈ 40μm. The experiment duration was limited by the 4 GB

memory capacity of the camera. The images were captured at 30 frames per second for a duration of 95 seconds at

full resolution (1280 × 1024 pixels2).

The experiments were conducted by choosing a freestream velocity and an air temperature, and allowing the

ISU-UTAS-IRT facility to equilibrate to the set conditions for 15 minutes. While the icing tunnel settled into the

specified experimental conditions, the spray system was adjusted to achieve the desired liquid water content. Here,

the air and water pressure regulators were adjusted while monitoring the liquid water content that was computed and

displayed in real time by a MATLAB code. The MATLAB code computed the liquid water content from the

freestream velocity and water volume flow rate, which were sampled via the data acquisition system to. After the

spray system was set for the trial, the water spray system was shut off and the airfoil model was installed in the

tunnel and allowed to reach thermal equilibrium with icing tunnel. After the icing tunnel and test model were

allowed to reach thermal equilibrium, the high speed-camera was armed and triggered simultaneously as the water

spray system was engaged.

The images were digitized using a MATLAB code implementing the algorithms described previously by

Waldman and Hu [7] for extracting ice and water runback features from time-resolved image sequences. For

convenience, we briefly describe the process here. Consider the sequence of images Ii(x,y) where the image intensity

at spatial location (x,y) is encoded, and i denotes the ith image in the sequence. Then, the time-derivative of the

image is denoted ΔIi(x,y) and is given by Eqn. (6).

1( , ) ( , ) ( , )i i iI x y I x y I x y (6)

( , ) ( , ) MedianFilter ( , )i i i

fI x y I x y I x y (7)

The image derivative indicates where ice or water is accumulating or changing configuration. A median filter is used

to help reject image noise due to lighting irregularities (Eqn. (7)). The activity in the image is due to the

accumulation and redistribution of the ice and water on the wing surface, therefore, by tracking the changes in the

image sequence, the ice accretion can be measured. Here, we compute the image activity as the square of the image

changes, where the changes are larger than a tolerance value chosen to suppress shot noise in the camera sensor.

Equation (8) defines the image activity, a, for the ith image in the sequence, using the square of the image changes

that are larger than the camera noise level, ε.

2 2

( , ) , ( , )( , )

0 otherwise

i ii f fI x y I x y

a x y

(8)

IV. Results and discussion

A. Surface wettability

The two surfaces tested in the study were chosen because of their significantly different surface wettabilities.

The difference between the two surfaces is obvious from the contact angle experiment images shown in Figure 6.

The enamel surface shows significant differences in the contact angle measured when the contact line is advancing

compared to receding. As expected, the superhydrophobic coating demonstrates a very large contact angle, and little

apparent change in the contact angle between the advancing and receding states.

The contact angle is plotted as a function of the contact line speed in Figure 7. The enamel surface shows a

static contact angle of about 65°, and the dynamic contact angle changes strongly with contact line speed over the

range of speeds tested. The largest advancing and receding contact line speeds correspond to contact angles of 104°

and 20°, respectively; however, the range of contact line speeds (±5 mm/s) accessible to the surface wettability

experiment is relatively limited. Nevertheless, the moderate static contact angle with an 84° contact angle hysteresis

indicates that the adhesion forces between the water and the enamel surface provide a relatively large resistance to

aerodynamic forces acting on a drop sitting on the surface.

Dow

nloa

ded

by I

OW

A S

TA

TE

UN

IVE

RSI

TY

on

Nov

embe

r 27

, 201

6 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/6.2

016-

3139

6

The superhydrophobic surface, however, shows very little contact angle hysteresis. The total contact angle

hysteresis is only 4° and the observed static contact angle is about 144°. The relative capillary forces experienced by

droplets with identical spherical cap radii sitting on the enamel and superhydrophobic coatings can be estimated

from the ratio of the capillary forces estimated Eqn. (4), namely,

, enamel enamel

, SHP

SHP

sin sin2 2

25

sin sin2 2

adv rec adv rec

cap

cap adv rec adv rec

F

F

. (9)

B. Icing tunnel measurements

Figure 8 and Figure 9 show a time series of the ice accretion on the different surface coatings. Figure 8

shows the process for a freestream velocity of U∞ = 20 m/s, air temperature of T∞ = –4°C, and = 0°. The right side

of the wing has the standard finely finished Rustoleum enamel coating and shows a typical slow process of water

runback and icing. The water collects at the leading edge, forms a water film that runs back and breaks up into

rivulets. The rivulets run back and freeze near quarter cord, leaving an ice bump at the end of each rivulet which

may continue to catch flow and accumulate water. The left side of the wing has the Hydrobead superhydrophobic

treatment with large contact angle and low contact angle hysteresis. Here, the only apparent accumulation of water is

immediately at the leading edge where the water is deposited. There are no rivulets forming, and there is no apparent

accumulation of water behind the leading edge. Here, the superhydrophobicity leads to very small adhesion forces,

so any liquid water drop sitting on the top of the airfoil is quickly carried away by the aerodynamic stresses.

Figure 9 shows the same wing as in Figure 8 with the air temperature of T∞ = –4°C, and = 0°; however, the

freestream velocity has been increased to U∞ = 40 m/s. On the right side of the wing with the Rustoleum coating, the

increased wind speed drives the surface water into many, thinner rivulets that runback further along the wing chord.

The left side of the wing with superhydrophobic treatment again suppresses the formation of rivulets. The water and

ice collection is constrained to the leading edge in the collection region. Further along the chord the water droplets

are quickly blown off of the wing surface by aerodynamic stresses.

The resulting ice formations shown in Figure 8 and Figure 9 indicate that reduced wettability of the

superhydrophobic coating clearly reduces the ice accretion on the surface of the airfoil model compared to the

enamel coating. However, once the ice forms at the leading edge, water and ice will continue accumulate in that

region. The characteristics of the accreted ice are altered by the surface properties underneath. The enamel coated

portion of the wings in Figure 8 and Figure 9 show markedly less roughness near the leading edge. Here, the

impinging water readily wets the surface as it runs back, leading to a smooth, thin layer of ice which covers a larger

extent of the airfoil chord. Conversely, water accumulating on the leading edge of the superhydrophobic half of the

airfoil model still exhibits large contact angle, increasing the height of the accumulated water and leading thicker ice

at the leading edge.

To systematically investigate the ice and water runback dynamics, the image activity, a, time series was

computed for each test condition according to Eqn. (8). These image activities are then integrated in time to produce

a cumulative image activity measurement, , of where ice and water accretion occur throughout the icing process:

1

( , ) ( , )i

i j

j

x y a x y

. (10)

The cumulative image activity is helpful for identifying some of the key difference in the ice accretion

processes on the different wing coatings. Figure 10 shows the cumulative image activity plots for LWC = 4 g/m3, T

= -8°C, and for freestream velocities U = 30 and 50 m/s. These image activity maps integrated over the first 75

seconds of icing conditions, and these plots clearly highlight the regions on the wing that accumulate the water and

ice. Most noticeably are the rivulets on the enamel-coated half wing, which are absent on the superhydrophobic-

coated half wing. The adhesion forces between the water and the superhydrophobic surface are so small, that the

aerodynamic forces eject water droplets from regions with pooling liquid water. The white arrows on the

superhydrophic plots in Figure 10 indicate regions with enhanced image activity because of localized water droplet

shedding. The initial ice that freezes at the leading edge in the impingement area shields subsequently collected

water from the hydrophobic underlying substrate. As this collected water runs back through the ice accreted at the

Dow

nloa

ded

by I

OW

A S

TA

TE

UN

IVE

RSI

TY

on

Nov

embe

r 27

, 201

6 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/6.2

016-

3139

7

front of the airfoil, the water runback is determined by the hydrophilic ice layer. When the liquid water reaches the

rear of the ice accretion it encounters the exposed superhydrophobic surface; hence, it reaches a gradient in the

surface wettability which retards the downstream flow. The collecting water pools up until it is ejected by

aerodynamic stresses. The superhydrophobicity of the exposed wing surface prevents the ejected water from wetting

the surface, resulting in water droplets that are rapidly advected off the wing. Under the current experimental

conditions, the 30 frames per second frame rate is too slow to capture the droplets in multiple frames, even at the

lowest wind speeds tested.

To study the runback dynamics, a spanwise-averaged measure of the water runback and ice accretion process

is computed by integrating the image activity, a, across the span at each point in time,

( ) ( , )i j

y

A x a x y . (11)

The resulting spanwise-averaged activity, A, shows the temporal evolution of the chordwise distribution of the

accretion activity. The spanwise-averaged activity is plotted for the enamel coating half-wing in Figure 11. Here, the

plots are shown for each of the freestream velocities U = 20, 30, 40 and 50 m/s, with LWC = 4 g/m3 and T = -8°C.

The activity on the wing in these conditions is predominately rivulet formation and runback. At the lowest speed,

rivulet formation and runback is relatively slow and during the 95 second trial, only a single rivulet is formed.

However, upon increasing the freestream velocity, the number of rivulets and the rivulet runback speed is enhanced.

Figure 11 shows that upon the initial impingement of water, rivulets are formed and run back along the wing. After

this initial rivulet formation, there is a pause as the ice accretion activity intensifies at the leading edge. At the

highest wind speed tested (50 m/s), there is a lull in the ice accretion activity in the rivulet runback region between

20 and 35 seconds into the trial. However, after 35 seconds, the rivulet formation and runback resumes along the

entire wing chord. As the freestream velocity decreases to 40 m/s, the time of decreased rivulet activity is delayed

and extended, whereas at even lower wind speeds (30 m/s) the rivulet formation and runback only appears to resume

near the end of the trial. Figure 10 highlights the difference between the 30 and 50 m/s trials. The cumulative image

activity in the 50 m/s trial after 75 seconds show that many rivulets are actively transporting water along the chord,

whereas the 30 m/s trial appears to be in the initial stages of forming a second round of rivulets.

The time history of the spanwise-averaged activity is used to digitize the initial rivulet runback, as indicated by the

white arrows in Figure 11. The digitized rivulet runback history for the velocities U = 20, 30, 40 and 50 m/s, with

LWC = 4 g/m3 and T =-4 and -8°C are presented in Figure 12. The digitized rivulet runback is plotted as chordwise

position vs time, therefore the slope of the curves indicate the rivulet runback speed. The rivulet runback speed

increases with increasing wind speed because of the increased aerodynamic stresses driving the flow. The rivulet

speed also increases at warmer temperatures. The convective cooling, which removes the latent heat of the freezing

water, depends directly on the air temperature. As the temperature is decreased, the removal of latent heat from the

water film is increased, therefore enhancing the rate of freezing. The enhanced rate of freezing leaves less liquid

water to flow along the surface, which thins the water film and reduces the size of the rivulet beads, thereby slowing

the rivulet runback rate.

V. Conclusions

Two surface coatings with very different surface wettabilities were tested on a symmetric airfoil in an icing

wind tunnel. Measurements of the advancing and receding contact angles on each surface suggest that the

superhydrophobic coating has about 1/25th the capillary forces resisting the motion of a water droplet across its

surface when compared to the enamel coating. Because of the drastic difference in surface wettability, a comparison

of the water runback and ice accretion on the two different surfaces, while subjected to the same icing conditions,

results in dramatically differing behaviors. The runback and icing behavior on an airfoil strongly depends on the

wetting properties of the airfoil surface. The surface wettability modifies the forces resisting the aerodynamic forces

driving surface water transport. High speed imaging of the surfaces in icing conditions show that the

superhydrophobicity suppresses the formation of rivulets, which is one of the prominent icing features present on the

enamel wing. Instead of forming rivulets, which then run back along the chord of the wing, water droplets are

ejected from liquid pools, which rapidly roll of the wing at speeds faster than can be resolved in the present

measurements. Dynamic measurements of the rivulets formed on the enamel wing show that wind speed and

temperature play an important role in the rivulet runback dynamics. Future work will investigate the surface water

Dow

nloa

ded

by I

OW

A S

TA

TE

UN

IVE

RSI

TY

on

Nov

embe

r 27

, 201

6 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/6.2

016-

3139

8

runback and ice accretion on additional surface treatments with various intermediate wettability characteristics to

determine how the airfoil surface chemistry affects the water runback and ice accretion at different icing conditions.

VI. Acknowledgments

The research work is partially supported by National Aeronautical and Space Administration (NASA) Grant

number NNX12AC21A with Mr. Mark Potapczuk as the technical officer and Iowa Space Grant Consortium (ISGC)

Base Program for Aircraft Icing Studies with Dr. Charisse Buising as the Director. The support of National Science

Foundation (NSF) under award numbers of CBET-1064196 and CBET- 1435590 is also gratefully acknowledged.

VII. References

[1] Bragg, M. B., Broeren, A. P., and Blumenthal, L. A., “Iced-airfoil aerodynamics,” Progress in Aerospace Sciences, vol.

41, Jul. 2005, pp. 323–362.

[2] Cao, L., Jones, A. K., Sikka, V. K., Wu, J., and Gao, D., “Anti-icing superhydrophobic coatings.,” Langmuir : the ACS

journal of surfaces and colloids, vol. 25, Nov. 2009, pp. 12444–8.

[3] Maitra, T., Tiwari, M. K., Antonini, C., Schoch, P., Jung, S., Eberle, P., and Poulikakos, D., “On the nanoengineering of

superhydrophobic and impalement resistant surface textures below the freezing temperature,” Nano Letters, vol. 14,

2014, pp. 172–182.

[4] Vorobyev, A. Y., and Guo, C., “Multifunctional surfaces produced by femtosecond laser pulses,” Journal of Applied

Physics, vol. 117, Jan. 2015, p. 033103.

[5] Epstein, A. K., Wong, T.-S., Belisle, R. a, Boggs, E. M., and Aizenberg, J., “Liquid-infused structured surfaces with

exceptional anti-biofouling performance.,” Proceedings of the National Academy of Sciences of the United States of

America, vol. 109, Aug. 2012, pp. 13182–7.

[6] Zhu, L., Xue, J., Wang, Y., Chen, Q., Ding, J., and Wang, Q., “Ice-phobic coatings based on silicon-oil-infused

polydimethylsiloxane.,” ACS applied materials & interfaces, vol. 5, 2013, pp. 4053–62.

[7] Waldman, R., and Hu, H., “High-Speed Imaging to Quantify Transient Ice Accretion Process over an Airfoil,” Journal

of Aircraft, vol. 53, 2016, pp. 369–377.

[8] Zhang, K., Wei, T., and Hu, H., “An experimental investigation on the surface water transport process over an airfoil by

using a digital image projection technique,” Experiments in Fluids, vol. 56, 2015, p. 173.

[9] Dorrer, C., and Rühe, J., “Some thoughts on superhydrophobic wetting,” Soft Matter, vol. 5, 2009, p. 51.

[10] Quéré, D., Azzopardi, M., and Delattre, L., “Drops at rest on a tilted plane,” Langmuir, vol. 14, 1998, pp. 2213–2216.

Dow

nloa

ded

by I

OW

A S

TA

TE

UN

IVE

RSI

TY

on

Nov

embe

r 27

, 201

6 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/6.2

016-

3139

9

(a) (b)

Figure 1. Typical scenarios where droplets are subjected to forces with a component tangential to the surface. Under

tangential forces, droplets exhibit contact angle hysteresis. Figure 1a shows droplet motivated by aerodynamic

forces along the surface, while Figure 1b shows a droplet subjected to a body force with a component along the

surface.

Figure 2. The pinning forces acting at the contact line on a droplet that resist droplet sliding.

Figure 3. Experiment setup used to measure surface coating wetting properties. A syringe is used to create an

expanding/contracting droplet on the test substrate. A highspeed camera captures back-lit images for image

processing.

Dow

nloa

ded

by I

OW

A S

TA

TE

UN

IVE

RSI

TY

on

Nov

embe

r 27

, 201

6 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/6.2

016-

3139

10

Figure 4. Droplet contact angle and contact line speed measurement process. A line-detector identifies the pixels that

lie on substrate surface and the droplet edge. Linear and quadratic best-fit curves are computed for the substrate

surface and the droplet edge, respectively, using a least squares fit to the corresponding edge pixels. The intersection

of the two fit curves denotes the contact line location, while the relative angle between the lines at their intersection

determines the contact angle.

Figure 5. Experimental configuration of the ice accretion experiments.

Dow

nloa

ded

by I

OW

A S

TA

TE

UN

IVE

RSI

TY

on

Nov

embe

r 27

, 201

6 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/6.2

016-

3139

11

Advancing

Receding

(a) Rustoleum enamel coating

Advancing

Receding

(b) Hydrobead superhydrophobic coating

Figure 6. Advancing and receding contact angle images for (a) the Rustoleum enamel surface coating and (b) the

superhydrophobic surface coating.

Dow

nloa

ded

by I

OW

A S

TA

TE

UN

IVE

RSI

TY

on

Nov

embe

r 27

, 201

6 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/6.2

016-

3139

12

Figure 7. Droplet contact angle is plotted as a function of contact line speed. The circles denote the Rustoleum

enamel surface and the squares denote the Hydrobead superhydrophobic (SHP) surface.

Dow

nloa

ded

by I

OW

A S

TA

TE

UN

IVE

RSI

TY

on

Nov

embe

r 27

, 201

6 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/6.2

016-

3139

13

Figure 8. Time sequence of ice accretion on a wing with half of the span treated with a Hydrobead

superhydrophobic coating. The freestream velocity is U∞ = 20 m/s, and the air temperature is T∞ = -4°C, and the

airfoil model set at = 0°. The black arrow denotes the freestream direction.

Dow

nloa

ded

by I

OW

A S

TA

TE

UN

IVE

RSI

TY

on

Nov

embe

r 27

, 201

6 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/6.2

016-

3139

14

Figure 9. Time sequence of ice accretion on a wing with half of the span treated with a Hydrobead

superhydrophobic coating. The freestream velocity is U∞ = 40 m/s, and the air temperature is T∞ = -4°C, and the

airfoil model set at = 0°. The black arrow denotes the freestream direction.

Dow

nloa

ded

by I

OW

A S

TA

TE

UN

IVE

RSI

TY

on

Nov

embe

r 27

, 201

6 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/6.2

016-

3139

15

Superhydrophobic Enamel

U=30 m/s

U=50 m/s

Figure 10. Cumulative image activity plots for LWC = 4 g/m3, T = -8°C, after 75 seconds in icing conditions. These

time-integrated plots highlight the regions that accumulate water and ice, as well as the paths of rivulets. The

superhydrophobic coated model does not form rivulets; however, water droplets ejected from pooled regions are

visible (denoted by white arrow). The black arrow denotes the freestream direction.

Dow

nloa

ded

by I

OW

A S

TA

TE

UN

IVE

RSI

TY

on

Nov

embe

r 27

, 201

6 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/6.2

016-

3139

16

Enamel

U=20 m/s

U=30 m/s

U=40 m/s

U=50 m/s

Figure 11. Spanwise averaged image activity plots for LWC = 4 g/m3, T = -8°C. These spatially-averaged plots

show the temporal dynamics of the ice accretion process. The initial rivulet runback is indicated by the white

arrows. The black arrow denotes the freestream direction.

Figure 12. Initial rivulet runback vs time for the enamel coated half-wing. The rivulet runback speed increases with

increasing wind speed. The rivulet runback speed decreases with decreasing temperature

Dow

nloa

ded

by I

OW

A S

TA

TE

UN

IVE

RSI

TY

on

Nov

embe

r 27

, 201

6 | h

ttp://

arc.

aiaa

.org

| D

OI:

10.

2514

/6.2

016-

3139