Active lubricant-impregnated surfaces

(1)

J

Active Lubricant-Impregnated Surfaces

SACHUSETTS INSTITUTE

OF TECHNOLOGY

by

AUG 15 2014

Karim S. Khalil

B.S.E. Mechanical Engineering and Materials Science

LIBRARIES

Duke University (2012)

Submitted to the Department of Mechanical Engineering

in partial fulfillment of the requirements for the degree of

Master of Science in Mechanical Engineering

at the

MASSACHUSETTS INSTITUTE OF TECHNOLOGY

June 2104

@

Massachusetts Institute of Technology 2104. All rights reserved.

Signature redacted

A uthor

...

Department of Mechanical Engineering

May 9, 2104

Certified by...Signature

redacted

I

i~pipa

K

Varanasi

Doherty Chair in Ocean Utilization

Associate Professor of Mechanical Engineering

Thesis Supervisor

Signature redacted

Accepted by

...

Dave E. Hardt

Ralph E. and Eloise F. Cross Professor of Mechanical Engineering

Chairman, Department Graduate Committee

(2)

(3)

Active Lubricant-Impregnated Surfaces

by

Karim S. Khalil

Submitted to the Department of Mechanical Engineering on May 9, 2104, in partial fulfillment of the

requirements for the degree of

Master of Science in Mechanical Engineering

Abstract

This thesis presents the design and testing of actively controlled lubricant-impregnated surfaces for enhanced droplet mobility and manipulation. Droplet manipulation and mobility on non-wetting surfaces is of practical importance for diverse applications ranging from micro-fluidic devices, anti-icing, dropwise condensation and biomedical devices, however most of the time droplets are moved passively. The use of active ex-ternal fields has been explored via electric, acoustic and vibrational fields, yet moving

highly conductive and viscous fluids remains a challenge. Magnetic fields have been

used for droplet manipulation, however the fluid is usually functionalized to be mag-netic, and requires enormous fields of superconducting magnets when transitioning to diamagnetic materials such as water.

This thesis presents a new class of active surfaces by impregnating active fluids such as ferrofluids into a textured surface. Droplets on such ferrofluid-impregnated surfaces have extremely low hysteresis and high mobility such that they can be pro-pelled by applying relatively low magnetic fields. Our surface is able to manipulate a variety of materials including diamagnetic, conductive and highly viscous fluids, and additionally solid particles. The surface's droplet propulsion mechanism is described, and is demonstrated to operate independently of the fluid or solid's physical prop-erties that normally inhibit motion (such as conductivity, viscosity, magnetization). In addition, several previous methods for droplet manipulation require pre-fabricated channels that govern the path of the fluid, however we are able to achieve precise control of droplets on a free surface along complex paths, which allows for the use of a single surface for any number of lab-on-a-chip applications and designs.

Thesis Supervisor: Kripa K. Varanasi Title: Doherty Chair in Ocean Utilization Associate Professor of Mechanical Engineering

(4)

(5)

Acknowledgments

I would first like to thank my advisor Professor Kripa Varanasi. Without his expertise,

creativity and most importantly motivation, I could not have completed this work.

I'd like to thank all the members of the Varanasi Lab - current and past, that have heavily influenced my work with their relentless attention to detail. It is truly a pleasure to work alongside all of these amazing people. Specifically the great dis-cussions and support from Hyuk-Min, Dave, Gisele, Divya, Rajeev, Nada, Konrad, Navdeep and Lauren. A special thanks to Adam, Brian, Srinivas, Sushant and Seyed for being sources of inspiration for me and showing me how to succeed in the lab.

I would also like thank the MIT-KFUPM Center for Clean Water and Clean

Energy for financial support to complete this work.

I would not be where I am today without my parents. The selfless effort they have

put in to providing for my brothers and I has been a true source of motivation my entire life, and is what drives me to always follow through with all my goals. Thanks for everything Mom & Dad. Also, as the youngest of three, I truly have the two best older brothers one could ask for. Ahmad and Ayman will always be my role models. Lastly, I want to thank my friends. There is no possible way I can explain how much each of you mean to me in this short bit of writing. Thank you to all of my friends from home, Duke, and Boston for being such a big part of my life. To my Duke Maxwellians: there are way too many of you to list, but I'd like to say that you remain and will always be among my best friends and I'm ready for the next reunion trip to wherever. Special thanks to Sameer, Matt, Shaunak, Cohen, Brian, Flav and Dave, you guys are like brothers to me. Finally, I want to thank you Betsy, your unquestioning support and comfort along the way keeps me going and has had such a significant impact on me. I've had the best times with you, and you always know how and when to help, I can't thank you enough.

(6)

(7)

The field of interfacial science has evolved greatly over the past decade. Most popu-larly, the "lotus effect" which has been well studied, allows for droplets to roll easily along the lotus leaf surface, as seen in Figure 1-1. The leaf is able to stay pristinely clean despite growing in harsh, muddy conditions through its ability to "self-clean" in which water droplets roll off the surface dislodging and removing dirt from the surface.

Droplet mobility and manipulation on non-wetting surfaces has received widespread attention due to its significance in a variety of applications such as liquid repel-lency [3-9], anti-icing [7,10-18], dropwise condensation [19-23] and biomedical de-vices [24-26]. In these applications droplets are normally moved passively, most often under the force of gravity [4]. Active manipulation of droplets using electric fields [2, 27] and vibrational fields [28, 29] have been studied, yet moving highly-conductive and viscous fluids remains difficult.

The goal of this work is to create surfaces that utilize active-external force fields to manipulate free droplets. The fundamentals governing such surfaces interaction with the external fields is explored, and the results of testing a variety of liquids on the surfaces are presented. This thesis will focus on the use of magnetic fields. Current methods for droplet manipulation using external fields are reviewed as well.

(12)

Figure 1-1: Water rolling off a lotus leaf demonstrating the lotus effect [4].

1.1 Motivations

1.1.1 Condensation Heat Transfer and Power Generation

Currently, steam power cycles account for approximately 50% of the energy that is produced globally. Condenser surfaces are relied on for providing low temperature

and pressure surfaces at the outlet of the turbines. Yet, as these surfaces are currently made from high surface energy metals such as titanium, they operate in a film-wise mode, in which a film of water forms and serves as a thermal insulator on the condenser surface. By designing the surface correctly, condensation can be operated in a dropwise mode in which the condensed water is shed rapidly from the surface in the form of individual drops, and one may achieve an increase in heat transfer coefficient by a factor of more than 10 [30]. This would correspond to an increase in overall plant efficiency of up to 5%. The surfaces investigated in this thesis could be later directly studied for enhanced condensation capabilities due to their ability of on-demand droplet shedding.

(13)

ground ground a.c. ground

Figure 1-2: Liquid-Liquid microfluidic system [2].

1.1.2 Lab-on-a-chip Droplet Manipulation

Delivering small amounts of liquid is important for diverse applications ranging from pharmaceutical systems to microfluidic devices for biochemistry and microsystems engineering [31, 32]. Droplet-based micro-fluidic devices have been found to be ben-eficial as they can reduce operating pressures and the need for bulky pumps with significant power consumption [33]. Typical lab-on-a-chip systems include prefabri-cated microchannels that may constrain the possible pathways to transport droplets containing materials of interest such as biological cells and proteins or suspended particles [34]. An area that is drawing much interest of late is the transportation of droplets on a free surface, but frequently encounter the drawback of material adhe-sion on the channel walls [35-39]. More recently, liquid-liquid microfluidic systems were successfully applied utilizing an alternating or constant electric field for liquid transport as shown in Figure 1-2 [2].

Yet the overall flexibility of a surface of method remains a challenge. The goal of this thesis is to present a surface with the ability to manipulate a large range of materials with no pre-determined design or function such that one could achieve any number of lab-on-a-chip applications or designs with the use of a single surface.

(14)

(15)

Chapter 2 Fundamentals of

Lubricant-Impregnated Surfaces

Non-wetting surfaces containing impregnated lubricating liquids have recently been shown to exhibit superior non-wetting behavior compared to typical superhydropho-bic surfaces. Lubricant-impregnated surfaces display low contact angle hysteresis [40-46], self-cleaning [47], dropwise condensation [48,49] icing [50-52] and anti-fouling properties [53-57].

On a traditional superhydrophobic surface, a stable air pocket is trapped within the surface (Figure 2-1) which creates a shear-free interface that diminishes the con-tact of the surface with the liquid that is being repelled. Maintaining stable air pockets, however, can be challenging as they can collapse by an externally applied pressure or even diffuse away into the liquid over time. It has been shown that by introducing a lubricant into a rough surface (Figure 2-2), stabilizing it by capillary forces arising from the microscopic texture, liquid droplets can move with ease, ex-hibiting a contact angle hysteresis of approximately 10.

These surfaces make up a unique four phase system (the repelled liquid, the sur-rounding gas, the lubricant oil, and the rough solid), and how they interact ultimately determines the contact line behavior of the liquid being repelled. The way with which the contact line of the liquid behaves on the surface will ultimately govern droplet pinning and hence its mobility on the surface.

(16)

Repelled Liquid

Figure 2-1: Schematic of a superhydrophobic surface.

Repelled Uquid

Figure 2-2: Schematic of a lubricant-impregnated surface.

2.1 Governing Interfacial Relations

The fundamental equations of interfacial science are key to understanding how a droplet will rest on a surface. For the case of superhydrophobic surfaces, or lubricant-impregnated surfaces, the interaction of all phases (fluid, solid, vapor) are governed

(17)

2.1.1 Young-Laplace Equation

The Young-Laplace equation relates the pressure difference across a curved surface of two static phases

AP + (2.1)

R1 R2

where R1 and R2 are the radii of curvature of the curved surface and -y is the surface tension (or more formally the interfacial tension between the two phases) [58]. This equation allows us to understand that along any curved surface, there will be a pres-sure difference between the concave and convex sides that is related to the interfacial tension. This pressure difference can be used to understand several phenomenon as-sociated with interfacial science and capillarity, including the classical capillary rise experiment where the capillary pressure of the curved fluid interface is balanced by a rise in fluid height (and gravitational energy).

2.1.2 Young Equation

Young's equation relates the contact angle of a resting droplet on a surface to the interfacial energies of the 3 phases.

YICOS6 = -Ysv - 7sl (2.2)

where -y is the interfacial tension between the 3 phases (1-liquid, v-vapor, s-solid) [58]. This equation can be explained as a horizontal force balance of the surface energies of all the phases for a droplet resting on a surface, as seen in Figure 2-3.

One assumption that is inherent in the derivation of the Young equation is that it assumes a perfectly flat and rigid surface. This is not valid for even fairly smooth surfaces, and most certainly not for micro-textured surfaces like those studied in this work. This means that a drop will assume a wide spectrum of contact angles when placed on surfaces with varying morphologies. This is normally accounted for by considering a range of contact angles for a given surface (advancing and receding

(18)

.Y'v

ysv

Figure 2-3: Force balance in horizontal direction of a liquid droplet on a surface. Force balance is used to derive Young's equation.

contact angles). The difference between these surfaces is referred to the contact angle hysteresis of a given surface, which is generally a gauge of the droplet adhesion on a surface.

2.2 Surface Thermodynamics

It has been shown that the thermodynamically stable state of a lubricant impregnated in a rough solid depends on both the spreading coefficients of the two fluids, as well as the texture geometry [40]. When designing a lubricant-impregnated surface, it is important to understand what state will result from the choice of a lubricant and solid in order to properly repel the desired liquid.

2.2.1 Impregnation and Pinning

A microtextured surface lends itself to a critical contact angle below which the oil

will successfully impregnate the texture and remain held by capillary forces. The lubricant will impregnate a texture surface if,

(19)

where 0os(a) is the contact angle of lubricant oil (o) on the smooth solid (s) in the presence of air (a) and 0, is the critical contact angle for impregnation, given by:

C = cos-[(1 - #)/(r - #)] (2.4)

Here,

#

is the fraction of the projected area of the textured surface that is occupied

by a solid and r is the ratio of total surface area of the textured surface to its projected area. In the case of square micro-posts with width a, edge-to-edge spacing b and height

h,

#

is given by

#

= a2_/(a₊_b)2 _{andr= 1}

+ 4ah/(a + b)2_.

Secondly, the degree to which the lubricant fills the textured solid will govern the contact line pinning dynamics of a repelled liquid. One must not only consider the interfacial behavior of the oil in air, but also in the presence of the repelled liquid say water (w). As initially stated, the lubricant will impregnate the texture if its contact angle on the solid in the presence of air is below a critical contact angle. It is worth mentioning that the micro-post tops will remain uncovered or dry unless

Oos(a) = 0. In this case the solid posts are fully encapsulated by the lubricant. This is the only condition with which the oil will fully cover the solid. This is an important consideration because the oil serves as a cushion to reduce the contact area of the droplet and the solid. Yet if the post tops remain dry, they serve as pinning sites for the droplet on the solid (See Figure 2-4 and 2-5 for all possible impregnated states in the presence of air and water and the criteria for each case).

Yet these considerations do not fully account for the system at stake. As the repelled liquid is placed on the surface, a new phase introduces new interfacial tensions to consider. In order for the water to not displace the oil our new condition is:

0os (w) <; 0c (2.5)

where the new oil-solid contact angle is measured in the presence of water rather than air. If this condition is met, the oil will remain in the texture as water is placed on top. As discussed earlier, the texture posts remain dry and will act to pin the water droplet, unless they can be fully encapsulated by oil. This alters our encapsulation

(20)

dry

impregnated, dry posts

impregnated, encapsulated

o

_osta>

>0

_c

0<0

<0

osta> c

o

=

0

osta>

Figure 2-4: Impregnation criteria in the presence of a vapor (air).

impaled

impregnated, dry posts

impregnated, encapsulated

o

_*sCw)

>0

_c

0<0

_OS(w)

0 =

0

Figure 2-5: Impregnation criteria in the presence of a fluid (water). criteria for under the water droplet to os(W) = 0.

Using the knowledge given above, one may design and predict how well a lubricant-impregnated surface will shed a particular liquid.

(21)

S >0

Lubricant Cloak

Sow(a)

< 0

Figure 2-6: Diagrams representing positive and negative spreading coefficient Sow(a) (cloaking versus non-cloaking surfaces)

2.2.2 Cloaking

It has been shown that depending that lubricant oil may in fact spread over and "cloak" the droplet of repelled liquid. The criterion for whether cloaking occurs is given by the spreading coefficient of the oil on water in the presence of air:

Sow(a) =Ywa - _Ywo - 'Yoa (2.6)

where -y is the interfacial tension between the phases defined earlier and w re-ferring to water as the repelled liquid. If Sow(a) > 0 the oil will cloak the droplet

of repelled liquid, whereas Sow(a) < _{0 implies no cloaking will occur. For example,}

silicone oil for which Sow(a) ~ 6mNm-1 has been shown to be a cloaking lubricant.

Certain lubricants have been investigated that have negative spreading coefficients such as ionic liquid (1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl) imide)

(22)

for which Sow(a) ~ -5mNm- 1. These two cases are displayed in the schematics of Figure 2-6. Cloaking is important to consider not only for understanding the true physical state of the system, but also for the progressive loss of the cloaked lubricant through entrainment in the water droplets as they are shed from the surface.

(23)

Chapter 3 Ferrofluid-Impregnated Surfaces

The use of magnetic fields for droplet manipulation has also been recently stud-ied [59-61], but encounter similar issues regarding the magnetic properties of the fluid, and normally need the enormous magnetic fields of superconducting magnets to op-erate on diamagnetic fluids such as water. Here we present a lubricant-impregnated surface that is able to manipulate a variety of different liquids including diamag-netic, electrically conductive and even solid particles, using a rough surface that is impregnated with ferrofluid.

3.1 Ferrofluid Properties

Ferrofluids, a colloidal suspension of ferromagnetic nanoparticles

(~

10 nm) in a

car-rier fluid stabilized by surfactant, have been studied extensively, but only rarely in its ability to manipulate free droplets [62-64]. The surface treatment of the ferromag-netic particles prevents agglomeration due to short range van der Waals forces, and Brownian motion prevents particle sedimentation in both gravitational and magnetic fields. In the absence of an applied field, the particles are randomly oriented giv-ing the fluid no net magnetization. Both paramagnetic and ferromagnetic materials align their magnetic dipole moment parallel to the direction of the applied field, yet ferromagnetic materials display a strong magnetic interaction between neighboring molecules, while paramagnetic materials display only weak interactions.

(24)

Magnetization theory for ferrofluids is derived utilizing the assumptions that fer-rofluids consist of non-interacting monodomain magnetic dipoles. The Langevin rela-tion for paramagnetic behavior describes the tendency for the dipole moment to align itself with the applied field, while also introducing the counteracting thermal energy, which acts to randomize the spatial orientation of the dipole moment:

= L(a) = coth(a) - 1 (3.1) M8a

,uomH

a = (3.2)

kbT

where M is the magnetization, M, is the saturation magnetization of the ferrofluid which corresponds to all dipoles being aligned with the field, A0 is the permeability of free space, m is the magnetic dipole moment of a particle, H is the applied magnetic field intensity and kbT is the thermal energy [62,64

3.2 Sample Preparation

3.2.1 Impregnation

A textured superhydrophobic surface was fabricated using typical photolithographic

techniques along with deep-reactive-ion etching (DRIE). Square micro posts were created with side length, edge-to-edge spacing and height 10 pm (See Figure 3-2). The micro post array was then treated with a coating of octadecyltrichlorosilane

(OTS) to render it hydrophobic. As described earlier, the critical contact angle with

which a fluid will impregnate the texture depends on a few geometric factors of the array. The critical contact angle:

(25)

CI\ CI C1" i'

Figure 3-1: Molecular structure of octadecyltrichlorosilane (OTS).

for the above described system requires a contact angle of 650. The ferrofluid used

(EMG901, Ferrotec,Inc.) exhibited an equilibrium contact angle of 20' on a flat

silicon substrate treated with OTS, which confirms the criteria for the ferrofluid im-pregnated the 10pm x 10pm square post array used to carry the liquid.

3.2.2 Cloaking

The spreading coefficient of the ferrofluid on water in the presence of air will determine whether the ferrofluid will rise out of the texture and cloak the water droplet. The spreading coefficient:

Sow(a) ='Ywa - ""Ywo -- Yoa (3.4)

for the ferrofluid is in fact positive which would predict cloaking behavior as a water droplet is placed on the surface.

Due to the dark color of the ferrofluid lubricant containing the magnetic nano particles, the cloaking film was observed and confirmed visually as shown in the photograph (Figure 3-3b).

(26)

Figure 3-2: SEM's of dry textured micro post array.

3.3 Droplet Manipulation

3.3.1 Initial Droplet Actuation Tests

Approaching a cylindrical permanent magnet (K&J Magnetics) to the surface and droplet attracts the magnetic particles to the region of highest magnetic field intensity,

(27)

ferrofluid cloak magnetic nanop cles

h

oil-based carrier fluid

Figure 3-3: a) Schematic of ferrofluid-impregnated surface b) Photograph confirming cloaking criteria.

along the centerline of the nearest magnetic pole. Macroscopically, this acts to distort the ferrofluid-air interface towards the magnet, thus translating the droplet along the surface towards the magnet. This is made possible by the extremely low contact-angle hysteresis

(~1 ).

As discussed in Chapter 2, the lubricant, solid, water interaction depends on the most stable thermodynamic state that the system can attain. When characterizing the pinning forces on the water droplet, the exposure of the micro post array tops needs to be considered. In order for the lubricant to fully coat the post tops in air, Oos(a) = 0. But, as measured earlier, the ferrofluid has a finite contact

(28)

C

Figure 3-4: c) Snapshots of droplet propulsion.

angle on OTS coated silicon (~20) which means the post tops will be exposed to the air. Yet, to explain the low contact angle hysteresis observed by a water droplet moving along this surface, we measured the contact angle of the ferrofluid on the OTS Si substrate in the presence of water to be ~Oo. This would explain that beneath the water droplet, the post tops are in fact being covered by the ferrofluid and act to lubricate the motion of the droplet along the surface, thus leading to virtually no pinning forces.

The permanent magnet was approached to the surface horizontally, and the droplet was propelled across the surface towards the magnet. The droplet position versus time

(29)

d

E

14E 12

-C

.o 10-U,) 0 8 ( 6 L2 ₄ 2 0 5 10 15 20

time/

s

Figure 3-5: d) Droplet position versus time for experiment.

was tracked and recorded, and shows that the droplet is in face accelerating towards the magnet, or the region of highest magnetic field intensity.

3.3.2 High-Magnification Droplet Studies

Now to better understand the interaction of the magnetic field and the cloaked droplets, we imaged single droplets resting on the surface at high magnification. In our experiments, we first approached the cylindrical magnet directly from above the droplet. The droplet is seen to deform toward the magnet, and the magnetic particles crowd to form local cone-like structures. The magnetic attractive force is locally bal-anced by restoring interfacial forces that act to hold these particles in the cloak. In this regime when the magnet is far enough, or the magnetic field is low, the cone-like structures move along the droplet when the magnet is moved. To further illustrate this point, as the region of highest magnetic field is moved, the cone-like structures move to that region (See Figure 3-6). This serves to act as a droplet propulsion mech-anism in which the cloaked film is able to distort and propel free droplets along the surface. As more particles are attracted towards the magnet, more particles move to

(30)

b

e-droplet orientation_ e-magnet

/

90- 080- 270-D 60- 50-(D 40- 230- S20-10 0 ( 10 20 30 40 50 60 70 80 e-magnet 90

Figure 3-6: a) Area of clumped particles b) Droplet orientation versus Magnet orien-tation.

regions of higher field intensity, ultimately causing droplet acceleration.

3.3.3 Jetting Transition

Cloaked droplets continue to deform along with the ferrofluid until the magnet reaches a critical distance at which the magnetic attractive force of the particles is greater than the interfacial forces that are stabilizing them in the cloak. The particles then physically detach from the film and agglomerate on the surface of the magnet, thus marking the onset of the "jetting" transition that has been previously reported.

[65]

Applying scaling analysis, one can roughly predict the required condition for which jetting transition occurs.As shown the natural scaling for the interfacial deflection of the ferrofluid:

h*2 XM*2_R* ₍₃₅₎

187ryL*6

where h 2 ~ O(10 3_{m) is a characteristic interfacial distortion during a jetting}

a

-0

0

-O

(31)

a

E N 0~0 ,,02.5 -2 2 ' 0 I GM 7.5 8 8.5 9 9.5 10 10.5 magnet distance d / mm

Figure 3-7: a) Jetting experiment. Droplet deformation versus magnet distance.

experiment, /u0 = 1.257 * 10-6m * kg * s~2* A2 is the permeability of free space, mag-netic susceptibility x = 6.79, permanent magnet magnetization M* ~ O(106_A/m), radius of the magnet R* = 3 * 10- 3m, characteristic volume of clumped particle area V* O(10- 0m3), interfacial tension = 0y .022N/m, and magnet distance from unperturbed interface L*. [65] Solving the above equation for L* utilizing the given values for the other quantities, we can predict a critical magnet distance of L* ~O(10- 2m) for the jetting transition to occur, which we confirm experimentally.

A high-speed camera was used to visualize this phenomenon. The droplet height was

recorded as the distance between the magnet and the surface was decreased. We ob-serve two different regimes in the droplet deformation as shown in the plot in Figure

3-7. When the distance between the magnet and droplet was larger than the critical

jetting length, the droplet deforms smoothly and reaches an equilibrium deformation (Figure 3-8c).

(32)

b

C

Figure 3-8: b) High-Speed images of fluid jet detaching from cloak/air interface in the jetting regime. c) Snapshots of droplet deformation as the magnet is slowly approached to the surface.

droplet height or deformation begins to oscillate due to the continuous forming and detaching of liquid jets (Figure 3-8b). This accounts for the jagged data points on the left portion of the graph. This observation illuminates a new design consideration for these surfaces. As the particles jet from the cloak, fluid and particles are permanently

removed from from the lubricant layer (and attaching themselves to the permanent magnet or magnetic field source). Continued use of the system without considering the jetting transition would ultimately reduce the lubricant layer low enough to re-move the surface's slippery qualities. Therefore, this thesis provides the guidelines for designing droplet manipulation systems that apply sufficient magnetic fields to translate droplets below the critical jetting transition distance to avoid lubricant loss.

3.3.4 Vertical Orientation

Experiments were performed to see how droplets could be arrested on vertically ori-ented surfaces. As the surfaces exhibit extremely low droplet adhesion and contact angle hysteresis, droplets are easily shed at high surface tilt angles. Yet when a mag-net is brought perpendicularly to the surface, as in Figure 3-9, the attraction between the droplet and the magnet is strong enough to suspend the droplet on the vertical

(33)

Figure 3-9: Droplet sliding down on a vertical surface. Magnet is able to suspend a droplet on the surface avoiding shedding by gravity.

surface. The active actuation of the surface allows for highly switchable non-wetting to wetting behavior and can be tuned by the applied magnetic field.

Upon removal of the droplet one can see that once the magnet is close enough to the surface, the ferrofluid begins to form spike-like structures directly on the surface (Figure 3-10). These structures, which one typically observes when a magnet is approached sufficiently close to a flat ferrofluid surface (normal-field instabilities), have been thoroughly studied in the past [62]. The shape of the ferrofluid-air interface distorts itself to minimize the total energy of the system. The spikes of the ferrofluid move out along the magnetic field lines and distort the fluid, which is also being resisted by gravity and surface tension. The spikes and valleys decrease magnetic

(34)

Figure 3-10: Magnet perpendicular to surface brought close enough to observe normal-field instabilities form on surface.

energy, yet are increasing gravitational energy by raising fluid, and surface energy by creating a larger surface area of the fluid. The equilibrium shape of the ferrofluid for a given magnetic field will be determined by the balance of these energies. This description is consistent with the concept that spikes will only form above a critical magnetic field strength, which we observe experimentally.

3.4 Potential Applications

To highlight the potential technological impact of this type of active surface we present experiments including free droplet coalescence, movement of viscous and conductive fluids as well as solid particles, and moving objects on complex curved paths. Simply introducing a magnet centered between two droplets may induce coalescence. The region of highest magnetic field intensity, directly beneath the magnet, attracts both droplets to center under the magnet, leading to coalescence (Figure 3-11).

Additionally, the motion of conductive liquids (1M NaCl aqueous solution) and

highly viscous dielectric liquids (100% glycerol) was achieved. The manipulation of

low surface tension fluids could be achieved with this method utilizing the appropriate immiscible lubricant, which can be a challenge with typical superhydrophobic surfaces

(35)

a

Figure 3-11: a) Two water droplets (dye used to color droplets) are placed on ferrofluid-infused surface. Magnet lowered vertically directly between droplets and causes them to move toward one another. Coalesence of droplets occurs and is cap-tured just before droplets fully mix.

unless they have complex textures.

[6]

The proposed material along with magnets can also be used for handling solid objects. It is possible to translate a 5 mm glass bead using this material and a magnet. All of these experiments can be seen in Figure

3-12.

Finally, for lab-on-a-chip style applications, pre-fabricated microfluidic channels or embedded electrodes [66] would be used to move and mix droplets along com-plex paths, yet this surface's droplet actuation mechanism naturally does not have constraints on path geometries (Figure 3-13). Therefore, this ferrofluid-impregnated

(36)

b

Figure 3-12: b) Droplets of various fluids and one solid are actuated and displaced when a permanent magnet is approached to the surface.

C

Figure 3-13: c) Water droplet being dragged along an s-curve as the magnet is held beneath the surface.

surface provides a framework for free surface manipulation of a broad range of liquids with various physicochemical, electrical, and magnetic properties.

(37)

Chapter 4 Conclusions and Recommendations

To conclude, the focus of this thesis was the design and testing of active lubricant-impregnated surfaces for the manipulation of droplets. A silicon micro post array was treated with octadecyltrichlorosilane (OTS) and then impregnated with a ferrofluid. The resulting surface interacts dynamically with applied non-uniform fields by per-manents magnets. Droplets placed on the surface would deform and begin to displace towards the magnet, taking advantage of the low droplet adhesion and contact angle hysteresis of the impregnated surface.

The mechanism behind droplet propulsion was further studied as single droplets were inspected under high-magnification and high-speed video. The ferrofluid is able to cloak up onto the droplet, and then encounter particle-magnet attraction, which would then induce droplet motion along the surface. This mechanism does not rely on the typical properties of materials that can inhibit active methods such as con-ductivity, viscosity or magnetization.

This research is largely motivated by a need for droplet manipulation methods that do not have limitations on the type of fluid being transported. Low-surface ten-sion fluids are difficult to shed in a dropwise manner, such as in condensers, which is one area that this type of surface could provide a direct solution. In addition, this method's droplet actuation mechanism naturally does not have constraints on path geometries which would prove to be useful in lab-on-a-chip applications where pre-fabricated microfluidic channels or embedded electrodes are normally required.

(38)

This would allow for the use of a single surface for any number of lab-on-chip appli-cations and designs. Therefore, ferrofluid-impregnated surfaces provide a framework for free surface manipulation of a broad range of liquids with various physiochemical, electrical, and magnetic properties.

(39)

Appendix A

Experimental Methods and

Materials

A.1 Fabrication of Surfaces

Square microposts were etched in silicon using a standard photolithography process followed by deep reactive ion etching. Each post has a width, height, and edge-to-edge spacing of 10pm. The samples were then cleaned in a Piranha solution and treated with a low-energy silane. The samples were infused with ferrofluid by dip coating followed by nitrogen gas purging to remove any excess fluid not held by capillary forces.

A.1.1 Silanization

A self-assembled monolayer of octadecyltrichlorosilane (OTS) was deposited by

solu-tion deposisolu-tion onto the silicon substrates. This process involved dissolving the OTS into a solvent (toluene) and then reacting this mixture with a water/toluene emul-sion. Once the water comes into contact with the OTS molecules in the surrounding toluene, the reaction begins. A monolayer of OTS is then deposited onto the surface while the mixture is kept under vigorous sonication. Inevitably due to gravity, re-acted molecules will sediment during the reaction process and form extra layers on

(40)

top of the monolayer. After 2 minutes of reaction, the sample is taken out of the reaction bath and rinsed in acetone. The acetone rinse is meant to remove the excess

OTS layers laying on top of the monolayer. Once the surface was dry, it exhibited

hydrophobic properties (advancing/receding angles are 1100

/

96' respectively for a

flat silicon substrate).

A.2 Droplet Orientation-Angle Experiments

A 10pA drop was deposited on one of our surfaces. A cylindrical magnet (K&J

Magnet-ics) of radius 3 mm was slowly approached to the droplet at several angles relative to the droplet vertical centerline. Droplet deformation was then analyzed using ImageJ software analysis.

A.3 Jetting Experiments

A 10ld drop was deposited on one of our surfaces. A cylindrical magnet (K&J

Mag-netics) of radius 3 mm was slowly lowered directly over the droplet using an x-y con-trolled stage. The droplet height and magnet distance from the unperturbed interface were tracked using ImageJ as the attraction began to deform the droplet towards the magnet. A high-speed camera was used to capture the jetting phenomena (Photron

SA1).

A.4 Vertical Orientation Experiments

A 10pl drop was deposited on one of our surfaces that was attached vertically to

an x-y micro-controller stage. A cylindrical magnet (K&J Magnetics) of radius 3 mm was attached to a separate x-y micro controller stage such that it could be controllably approached horizontally and perpendicularly to the surface. Droplets were then allowed to slide down the vertical surface, and the magnet distance from the surface was varied such that eventually the droplet was completely arrested upon

(41)

falling.

A.5 Advanced Path Experiments

A 10pl drop was deposited on one of our surfaces. A cylindrical magnet (K&J

Mag-netics) of radius 3 mm was held beneath the surface so that the system could be imaged from above. Photoshop was used to overlay the first and last frame to show drops final position, and to illustrate the s-path the droplet moved along.

(42)

(43)

Appendix B

Additional Studies

B.1 Sub-Surface Magnet Configurations

As displayed in the advanced path experiments, the magnet configuration used on the surface-droplet can significantly alter the behavior of both the surface and the droplet. This section offers some insight and some brief observations when the magnet was approached from beneath the surface.

As a magnet is approached from beneath the ferrofluid-impregnated surface, normal-field instabilities are observed. These shark-fin needle like instabilities are well studied and documented in ferrofluids [62]. As previously discussed in this thesis, a mecha-nism for droplet propulsion when the magnet is approached from above the surface is the lubricant-cloak on the droplet. When the magnet is approached from below, as seen in the Figure B-1, the water droplet is pulled from beneath and distorted along its contact line seemingly by the ferrofluid spikes. The majority of this thesis discusses the ability for a magnet from above to distort the cloak and droplet, yet it is noted that droplets can be manipulated from distortions along the contact line which is observed when a magnet is approached from beneath the surface.

(44)

Figure B-1: Water Droplet is moved by magnet that is placed beneath the surface.

B.2 Ice Mobility Tests

Some preliminary tests with ice were taken on ferrofluid-impregnated surfaces. Due to complications with the Peltier stage used to cool the water droplet, the ice was frozen

(45)

on a separate surface and then placed onto a room temperature surface ferrofluid-impregnated surface. Although the ice would begin to melt immediately upon contact with the surface, it still offers an insightful experiment and observation on how this surface interacts with semi-solid materials. The complications of the Peltier stage can be attributed to an embedded magnetic stirrer. The embedded magnet would interact with the ferrofluid and cause it to cease up when the surface was placed on the stage, rendering the surface unusable.

(46)

(47)

References

[1] A. Otten, S. Herminghaus, Langmuir 20, 2405 (2004).

[2] 0. D. Velev, B. G. Prevo, K. H. Bhatt, Nature 426, 515 (2003).

[3]

P.-G. d. Gennes, Capillarity and wetting phenomena: drops, bubbles, pearls,

waves (Springer, New York, 2004).

[4] D. Quere, Annual Review of Materials Research 38, 71 (2008).

[5] X. Deng, L. Mammen, H.-J. Butt, D. Vollmer, Science 335, 67 (2012). [6] A. Tuteja, et al., Science 318, 1618 (2007).

[7] J. C. Bird, R. Dhiman, H.-M. Kwon, K. K. Varanasi, Nature 503, 385 (2013). [8] C. Duez, C. Ybert, C. Clanet, L. Bocquet, Nature Physics 3, 180 (2007). [9] T. Deng, et al., Applied Physics Letters 94, 133109 (2009).

[10] J. Lv, Y. Song, L. Jiang, J. Wang, A CS Nano 8, 3152 (2014). [11] L. Mishchenko, et al., ACS Nano 4, 7699 (2010).

[12] A. J. Meuler, G. H. McKinley, R. E. Cohen, ACS Nano 4, 7048 (2010).

[13] S. Jung, M. K. Tiwari, N. V. Doan, D. Poulikakos, Nature Communications 3, 615 (2012).

[14] T. Maitra, et al., Nano Letters 14, 172 (2014).

[15] P. Guo, et al., Advanced Materials 24, 2642 (2012). [16] J. B. Boreyko, C. P. Collier, ACS Nano 7, 1618 (2013).

[17] K. K. Varanasi, M. Hsu, N. Bhate, W. Yang, T. Deng, Applied Physics Letters 95, 094101 (2009).

[18] K. K. Varanasi, T. Deng, J. D. Smith, M. Hsu, N. Bhate, Applied Physics Letters 97, 234102 (2010).

(48)

[20] N. A. Patankar, Soft Matter 6, 1613 (2010).

[21] C. Dorrer, J. Rhe, Advanced Materials 20, 159 (2008). [22] D. M. Anderson, et al., A CS Nano 6, 3262 (2012).

[23] N. Miljkovic, et al., Nano Letters 13, 179 (2013).

[24] Y. Koc, et al., Lab on a Chip 8, 582 (2008).

[25] W. Choi, et al., Lab on a Chip 11, 3970 (2011).

[26] P. Tabeling, Current Opinion in Biotechnology 25, 129 (2014).

[27] D. R. Link, et al., Angewandte Chemie International Edition 45, 2556 (2006). [28] S. Daniel, M. K. Chaudhury, P.-G. de Gennes, Langmuir 21, 4240 (2005).

[29] P. Brunet, J. Eggers, R. Deegan, Physical Review Letters 99 (2007).

[30] E. Schmidt, W. Schurig, W. Sellschopp, Technische Mechanik und

Thermody-namik 1, 53 (1930).

[31] G. M. Whitesides, Nature 442, 368 (2006).

[32] P. S. Dittrich, A. Manz, Nature Reviews Drug Discovery 5, 210 (2006).

[33] R. Seemann, M. Brinkmann, T. Pfohl, S. Herminghaus, Reports on Progress in

Physics 75, 016601 (2012).

[34] T. Nisisako, T. Torii, T. Higuchi, Lab on a Chip 2, 24 (2002).

[35] M. K. Chaudhury, G. M. Whitesides, Science 256, 1539 (1992).

[36] C. Rosslee, N. L. Abbott, Current Opinion in Colloid & Interface Science 5, 81

(2000).

[37] W.-B. Song, Z. Ding, C. Son, B. Ziaie, Applied Physics Letters 90, 092501 (2007). [38] S.-W. Lee, P. E. Laibinis, Journal of the American Chemical Society 122, 5395

(2000).

[39] S. Daniel, M. K. Chaudhury, Langmuir 18, 3404 (2002).

[40] J. D. Smith, et al., Soft Matter 9, 1772 (2013).

[41] A. Eifert, D. Paulssen, S. N. Varanakkottu, T. Baier, S. Hardt, Advanced Mate-rials Interfaces (2014).

[42] A. Carlson, P. Kim, G. Amberg, H. A. Stone, EPL (Europhysics Letters) 104, 34008 (2013).

(49)

[43] X. Huang, J. D. Chrisman, N. S. Zacharia, ACS Macro Letters 2, 826 (2013). [44] T.-S. Wong, et al., Nature 477, 443 (2011).

[45] D. Querb, Reports on Progress in Physics 68, 2495 (2005).

[46] A. Lafuma, D. Quere, EPL (Europhysics Letters) 96, 56001 (2011).

[47] H. Liu, P. Zhang, M. Liu, S. Wang, L. Jiang, Advanced Materials 25, 4477

(2013).

[48] S. Anand, A. T. Paxson, R. Dhiman, J. D. Smith, K. K. Varanasi, ACS Nano 6, 10122 (2012).

[49] R. Xiao, N. Miljkovic, R. Enright, E. N. Wang, Scientific Reports 3 (2013). [50] S. B. Subramanyam, K. Rykaczewski, K. K. Varanasi, Langmuir 29, 13414

(2013).

[51] P. Kim, et al., A CS Nano 6, 6569 (2012).

[52] L. Chen, A. Geissler, E. Bonaccurso, K. Zhang, A CS Applied Materials &

Inter-faces (2014).

[53] A. K. Epstein, T.-S. Wong, R. A. Belisle, E. M. Boggs, J. Aizenberg, Proceedings

of the National Academy of Sciences 109, 13182 (2012).

[54] S. B. Subramanyam, G. Azimi, K. K. Varanasi, Advanced Materials Interfaces 1 (2014).

[55] J. Li, et al., ACS Applied Materials & Interfaces 5, 6704 (2013). [56] E. Ueda, P. A. Levkin, Advanced Healthcare Materials 2, 1425 (2013). [57] L. Xiao, et al., ACS Applied Materials & Interfaces 5, 10074 (2013). [58] R. J. Good, Journal of Adhesion Science and Technology 6, 1269 (1992).

[59] I. F. Lyuksyutov, D. G. Naugle, K. D. D. Rathnayaka, Applied Physics Letters 85, 1817 (2004).

[60] E. Beaugnon, R. Tournier, Nature 349, 470 (1991).

[61] Y. Ikezoe, N. Hirota, J. Nakagawa, K. Kitazawa, Nature 393, 749 (1998).

[62] R. E. Rosensweig, Annual Review of Fluid Mechanics 19, 437 (1987).

[63] R. E. Rosensweig, Scientific American 247, 136 (1982).

[64] T. A. Franklin, Ferrofluid flow phenomena, Master's thesis, Massachusetts Insti-tute of Technology, Cambridge, MA (2003).

(50)

[65] S. S. H. Tsai, I. M. Griffiths, Z. Li, P. Kim, H. A. Stone, Soft Matter 9, 8600 (2013).

Active lubricant-impregnated surfaces

Active Lubricant-Impregnated Surfaces

by

AUG 15 2014

Karim S. Khalil

B.S.E. Mechanical Engineering and Materials Science

LIBRARIES

Duke University (2012)

Submitted to the Department of Mechanical Engineering

in partial fulfillment of the requirements for the degree of

Master of Science in Mechanical Engineering

at the

MASSACHUSETTS INSTITUTE OF TECHNOLOGY

June 2104

@

Massachusetts Institute of Technology 2104. All rights reserved.

Signature redacted

A uthor

...

...

Department of Mechanical Engineering

May 9, 2104

Certified by...Signature

redacted

I

i~pipa

K

Varanasi

Doherty Chair in Ocean Utilization

Associate Professor of Mechanical Engineering

Thesis Supervisor

Signature redacted

Accepted by

...

Dave E. Hardt

Ralph E. and Eloise F. Cross Professor of Mechanical Engineering

Chairman, Department Graduate Committee

Active Lubricant-Impregnated Surfaces

by

Karim S. Khalil

Abstract

Acknowledgments

Contents

List of Figures

Chapter 1

Introduction

1.1

Motivations

1.1.1

Condensation Heat Transfer and Power Generation

1.1.2

Lab-on-a-chip Droplet Manipulation

Chapter 2

Fundamentals of

Lubricant-Impregnated Surfaces

2.1

Governing Interfacial Relations

2.1.1

Young-Laplace Equation

2.1.2

Young Equation

2.2

Surface Thermodynamics

2.2.1

Impregnation and Pinning

#

#

#

dry

impregnated, dry posts

impregnated, encapsulated

o

>0

0<0

<0

o

=

0

impaled

impregnated, dry posts