Coalescence and Particle Self-Assembly of Inkjet-Printed Colloidal Drops

Coalescence and Particle Self-assembly of Inkjet-printed Colloidal Drops

A Thesis

Submitted to the Faculty

Drexel University

Xin Yang

in partial fulfillment of the

requirements for the degree

Doctor of Philosophy

December 2014

iii

Acknowledgements

I would like to thank my advisor Prof. Ying Sun. During the last 3 years, she supports and guides me throughout the course of my Ph.D. research with her profound knowledge and experience. Her dedication and passion for researches greatly encourages me in pursuing my career goals. I greatly appreciate her contribution to my growth during my Ph.D. training. I would like to thank my parents, my girlfriend, my colleagues

(Brandon, Charles, Dani, Gang, Dong-Ook, Han, Min, Nate, and Viral) and friends

(Abraham, Mahamudur, Kewei, Patrick, Xiang, and Yontae).

I greatly appreciate Prof. Nicholas Cernansky, Prof. Bakhtier Farouk, Prof. Adam

Fontecchio, Prof. Frank Ji, Prof. Alan Lau, Prof. Christopher Li, Prof. Mathew McCarthy, and Prof. Hongseok (Moses) Noh for their critical assessments and positive suggestions during my candidacy exam, proposal and my defense.

I also appreciate the financial supports of National Science Foundation (Grant

CAREER-0968927 and CMMI-1200385).

TABLE OF CONTENTS

LIST OF TABLES ...... vii LIST OF FIGURES ...... viii CHAPTER 1 : INTRODUCTION ...... 1 1.1 Background ...... 1 1.2 Literature reviews ...... 2 1.2.1 Inkjet printing systems...... 2 1.2.2 Drop formation ...... 3 1.2.3 Drop impact and spreading ...... 6 1.2.4 Drop evaporation and particle deposition ...... 8 1.3 Objectives ...... 14 CHAPTER 2 : EXPERIMENTAL MTHODS ...... 16 2.1. Materials preparation...... 17 2.2. Process control ...... 19 2.3. In-situ observation ...... 20 2.3.1. Drop formation ...... 21 2.3.2. Drop impact, spreading, and evaporation ...... 21 2.3.3. Particle self-assembly ...... 22 CHAPTER 3 : OSCILLATION AND RECOIL OF SINGLE AND CONSECUTIVELY PRINTED DROPLETS ...... 23 3.1 Introduction ...... 23 3.2 Experimental method ...... 27 3.3 Results and discussion ...... 29 3.3.1 Spreading and oscillation of a single drop...... 29 3.3.2 Coalescence, recoil, and oscillation of two consecutively printed drops ...... 35 3.3.3 Comparison between drop impact on a dry surface and on a pre-deposited sessile drop ...... 39 3.4 Conclusion ...... 44 CHAPTER 4 : COALESCENCE, EVAPORATION AND PARTICLE DEPOSITION OF CONSECUTIVELY PRINTED COLLOIDAL DROPS ...... 46

4.1 Introduction ...... 46 4.2 Experimental methods ...... 49 4.3 Results and discussion ...... 50 4.3.1 Drop coalescence ...... 50 4.3.2 Drop evaporation and deposition morphology ...... 56 4.4 Conclusions ...... 65 CHAPTER 5 : COMPETITION BETWEEN RECEDING CONTACT LINE AND PARTICLE DEPOSITION IN DRYING COLLOIDAL DROP...... 67 5.1 Introduction ...... 67 5.2 Experimental methods ...... 69 5.3 Theoretical analysis ...... 70 5.3.1 Contact line receding velocity of an evaporating drop ...... 70 5.3.2 Particle deposition rate at contact line ...... 72 5.3.3 Ring spacing of the multi-ring structure ...... 73 5.4 Results and discussion ...... 75 5.4.1. Deposition morphology transition in droplets ...... 75 5.4.2 Interfacial instability near the drop contact line ...... 79 5.4.3 Effect of particle size ...... 83 5.4.4 Deposition morphology map ...... 84 5.5 Conclusion ...... 86 CHAPTER 6 : SELF-ASSEMBLY OF NANOPARTICLES IN EVAPORATING PARTICLE LADEN-DEMULSION DROPS ...... 88 6.1 Introduction ...... 88 6.2 Experimental method ...... 89 6.3 Results and discussion ...... 91 6.4 Conclusions ...... 97 CHAPTER 7 : SELF-ASSEMBLY OF NANOPARTICLES DIRECTED BY NANOPOROUS TEMPLATES ...... 99 7.1 Introduction ...... 99 7.2 Experimental methods ...... 101 7.3 Results and discussion ...... 101

7.3.1 Effect of annealing conditions on electrical properties and particle deposition morphologies ...... 101 7.3.2 Directed particle self-assembly into nanomesh ...... 103 7.3.3 Fabrication of Ag nanotubes with AAO membrane ...... 106 7.3.4 Fabrication of Ag nanotube forest on substrates ...... 108 7.4 Conclusion ...... 118 CHAPTER 8 : IMPACT OF DROPLETS ON LIQUID-INFUSED POROUS SURFACES ...... 119 8.1 Introduction ...... 119 8.2 Experimental methods ...... 120 8.3 Results and discussion ...... 122 8.4 Conclusion ...... 130 CHAPTER 9 : CONCLUSIONS AND RECOMMENDATIONS ...... 132 9.1 Conclusions ...... 132 9.1.1 Coalescence and particle deposition of two consecutively printed drops ...... 132 9.1.2 Beyond the drop size limitation: complex submicron structures in inkjet-printed drops ...... 134 9.1.3 Development and morphology control of Ag nanotube forest ...... 135 9.1.4 Droplet impacting on liquid-infused porous surfaces ...... 136 9.2 Recommendation for future work ...... 137 9.2.1 Oil film assisted particle self-assembly in inkjet-printed drops ...... 137 9.2.2 Printing of colloidal drops with ellipsoid particles for reducing the deposition area smaller than the drop size ...... 138 9.2.3 Large-scale and high throughput fabrication of submicron scale structures via particle self-assembly in inkjet printing processes ...... 142 9.2.4 Dielectrophoresis-directed particle self-assembly in evaporating drops ...... 145 9.2.5 Vapor-mediated droplet motion ...... 146 LIST OF REFERENCES ...... 150 VITA ...... 167

vii

LIST OF TABLES

Table 8.1 Re, We, and splashing parameter for the drops impacted surfaces...... 122

viii

LIST OF FIGURES

Figure 1.1 Physical stages of inkjet printing ...... 1

Figure 1.2 Schematics of a) continuous inkjet (CIJ) printer, b) and c) drop on demand (DOD) inkjet printers, b) thermal inkjet printer and c) piezoelectric inkjet printer1...... 3

Figure 1.3 A typical voltage waveform applied to the piezoelectric element in the nozzle of an inkjet printer...... 4

Figure 1.4 Sequence of images of DOD drop formation11...... 6

Figure 1.5 Schematic representation of the spread factors with time for a drop on a dry surface. The different lines correspond to an arbitrary choice of possible spreading histories, depending on the parameters of the impact17...... 8

Figure 1.6 Schematic diagrams of the three distinct stages in the evaporation process of a droplet on a substrate: (a) constant contact area mode, (b) constant contact angle mode, (c) mixed mode23...... 9

Figure 1.7 Comparison of the time-dependent weight based on the model predictions from Hu and Larson26 and the experimental results published by Birdi et al.25 for droplets of water with radii of 2.01, 2.53, 2.93 mm on glass at T = 22 °C...... 10

Figure 1.8 a) Liquid evaporative flux along the drop surface27. b) A dried colloidal drop with microspheres29. The scale bar is 1 cm...... 11

Figure 1.9 Categorized particle deposition shape as a function of the particle volume fraction and contact angle of the liquid with the substrate. The open circles are experimental results circles corresponding to convex shapes, and crosses show concave shapes30...... 12

Figure 1.10 Deposition of spheres and ellipsoids. Images of the final deposition of the ellipsoids a) and b) spheres after the evaporation. Insects are particle shapes, and α is the particle aspect ratio33...... 13

Figure 2.1 Schematic of the experimental setup for in-situ observation of droplet evaporation in inkjet printing...... 17

Figure 2.2 Schematic of the goniometer setup to measure contact angle of the liquid drop...... 19

Figure 2.3 Schematic of the humidity chamber...... 20

Figure 2.4 Shematic of flashphotography for visulization of the drop formation process...... 21

Figure 3.1 Schematic representation of droplet spreading radius as a function of time for drop impact on a hydrophilic substrate at low and high We numbers...... 24

Figure 3.2 Summary of the parameter space in a We-Oh plane for studies on inkjet- printed drops...... 27

Figure 3.3 Snapshots of single water droplets impact on substrates of different wettabilities. For all cases, the drop velocity is 0.77 m/s, the diameter is 60 µm, and We = 0.49...... 31

Figure 3.4 Single water droplets impact on substrates of different wettabilities. (a) Drop spreading radius versus time and (b) evolution of the drop height with time. The uncertainties of drop radius and height measurements are 0.83µm...... 32

Figure 3.5 Effects of viscosity and surface tension on drop spreading and oscillation on a hydrophobic surface. (a) Drop spreading radius as a function of time and (b) time evolution of the drop height. The uncertainties of drop radius and height measurements in this study are 0.83µm...... 34

Figure 3.6 Snapshots of a water drop impact on a sessile drop on substrates of different wettabilities. The drop impact velocity is 0.77 m/s, diameter is 60 µm, and We = 0.49. . 36

Figure 3.7 Snapshots of high viscous drops impact on a sessile drop of the same liquid sitting on hydrophobic substrates (θeq = 90°). The drop impact velocity is 0.77 m/s, diameter is 60 µm, WeW/G = 0.57, and WeW/G/I = 0.97...... 38

Figure 3.8 Effects of surface tension and viscosity on spreading of combined drops. (a) Drop spreading radius as a function of time and (b) time evolution of the drop height. The uncertainties of drop radius and height measurements are 0.83µm...... 39

Figure 3.9 Spreading and oscillation of single and combined droplets on hydrophobic substrates. (a) Drop spreading radius as function of time and (b) evolution of drop height with time. The uncertainties of drop radius and height measurements are 0.83µm...... 40

Figure 3.10 Scaled oscillation period of the water and W/G drops as a function of scaled drop mass...... 44

Figure 4.1 Side-view images for the impact and coalescence of two consecutively printed drops compared to the impact of a single drop on a dry surface during the same time frame.The jetting delay between two drops is 0.6 s and the drop spacing is 73.4 μm (i.e., 0.7 Dc, where Dc is the maximum spreading diameter of the first drop on the substrate). Other conditions for the two consecutively printed drops and the single drop are identical. For the case of a single drop, the drop continues to spread after 300 µs until it reaches and equilibrium contact angle of 45°...... 52

Figure 4.2 Schematic of drop impact, coalescence, carrier liquid evaporation, and particle deposition of two consecutively printed evaporating colloidal drops from side view (left) and bottom view (right) based on the medium drop spacing (0.5 Dc < d < 0.75 Dc) and long jetting delay (0.9 s) case...... 55

Figure 4.3 Snapshots of drop coalescence and particle deposition as a function of the scaled drop spacing for jetting delay of 0.9 s. (a) Short drop spacing (0.4 Dc), (b) medium drop spacing (0.7 Dc), and (c) long drop spacing (0.95 Dc). Drop spacing is defined as the distance between two consecutively printed drops and is scaled by the maximum wetting diameter of the first drop on the substrate (Dc)...... 57

Figure 4.4 Snapshots of drop coalescence and particle deposition as a function of jetting delay between consecutively printed drops. (a) 0.2s delay, (b) 0.6s delay, and (c) 0.9s jetting delay. Spacing between two drops are 0.60 Dc, 0.64 Dc, 0.70 Dc, respectively. ... 59

Figure 4.5 Scaled particle number density along the major axis of the deposit for jetting delays of 0.2, 0.6, and 0.9 s and medium separation distance between drops...... 61

Figure 4.6 Scaled radius of curvature on the second drop side of the merged drop as a function of the scaled drop spacing...... 63

Figure 4.7 Circularity of the coalesced drop as a function of the scaled drop spacing d Dc ...... 64

Figure 5.1 a) Schematic illustration of a drop receding on a hydrophilic substrate. b) Schematic step function of stick-slip motion of the contact line (solid lines) and the fitted linear curve (dash line) during drop receding. The dash line follows R ~ t by integrating Eq. (5-9)...... 72

Figure 5.2 Schematic illustration of the stick-slip motion of the contact line...... 74

xi dimensionless parameter, x =U U , where R corresponds to the evaporating drop p CL 1 radius when x = x * , the critical value to warrant pinning of the entire contact line. Due to the difficulty in measuring the particle deposition growth rate at a sub-micron scale, x versus R shown in Fig. 1d is a schematic illustration based on the analysis. The scale bars are 20 µm in (a) – (c) and 2 µm in (d)...... 76

Figure 5.4 Wavelength of the saw-toothed structures in multi-ring deposits for  = 0.5%

3 and the grid spacing of spider web for  = 0.25% as a function of R R0 resulted from Marangoni instability within single drops containing 20 nm particles and RH = 40%. The scale bars are 20 µm. Error bars are standard deviations from 15 measurements...... 80

Figure 5.5 Wavelength of the radial spokes as a function of 1 3 1- RH for  = 0.1% for drops containing 20 nm particles evaporating at relative humidities of 20%, 40% and 70%. Error bars are standard deviations from 5 measurements. The scale bars are 20 µm...... 81

Figure 5.6 Ring spacing of multi-ring structures as a function of ring radius for 20 nm particles at volume fractions of 0.25% and 0.5%. The relative humidity is 40%. The outmost ring is not included. Error bars are standard deviations from 8 measurements. The scale bars are 20 µm...... 83

Figure 5.7 Scanning electron microscopy images of particle deposition morphologies in colloidal drops containing 200 nm particles at particle loadings of (a) 0.5%, (b) 0.25%, and (c) 0.1%. Schematic illustrations of (d) multi-ring formation in drops containing particles 20 nm in diameter, and (e) radial spokes in drops with particles 200 nm in diameter. Scale bars are 20 µm in (a) to (c)...... 85

Figure 5.8 Deposition morphology map of inkjet-printed colloidal drop with 20 nm particles, dryed under 40% relative humidity showing multi-ring, spoke, spider web, foam and island patterns as a function of the radial position and the particle concentration. Ring deposition occurs for x > x * while the contact line is only partially pinned when x < x * . Error bars are standard deviations from 8 measurements...... 86

Figure 6.1 Experimental procedures for the fabrication of nanomesh in inkjet-printed particle-laden emulsion drops...... 91

Figure 6.2 Evaporation of an inkjet-printed particle-laden emulsion drop (0.5% particle volume fraction, oil (mesitylene) : water = 1 : 10, SDS concentration = 0.3%, RH = 40%)...... 92

xii

Figure 6.3 SEM images of particle deposition in particle-laden emulsion drops with different particle volume fractions. (a), (b) 0.04% in water, (c), (d) 0.5% in water, (e), (f) 1% in water. Oil (mesitylene) : water = 1: 10 v/v, SDS concentration 0.3% in water, RH = 40%. (b), (d) and (f) are zoom in images at the centers of (a), (c) and (e)...... 94

Figure 6.4 Assembly of oil drops in an evaporating particle-laden emulsion drop...... 95

Figure 6.5 Effect of oil/water ratio on the depositions of the particle-laden emulsion drops. (a) oil (mesitylene) : water = 1 : 10, (b) oil (mesitylene): water = 1 : 5, (c) oil (mesitylene): water = 1 : 1. The particle volume fraction in water = 1%, SDS concentration 0.3%, RH = 40%...... 96

Figure 6.6 Effect of relative humidity on particle deposition morphologies in the drop center. Insets show the deposition of the entire particle-laden emulsion drops. (a) RH = 40%, (b) RH = 70%. Oil (mesitylene): water = 1 : 10, surfactant concentration in water: 0.3%, and silica particle concentration: 1%...... 97

Figure 7.1 SEM images of an Anodic aluminum oxide nanoporous membrane with 55 nm pore diameter. Top view (a) and cross-sectional view (b)...... 100

Figure 7.2 Effect of annealing temperature on sheet resistance. (a) Dependence of annealing temperature on sheet resistance; (b) Ag particle deposition before sintering; (c) Ag particle deposition after sintering at 175 °C for 10 min...... 102

Figure 7.3 Effect of annealing on Ag particle deposition morphologies in AAO membrane. (a) before sintering. (b) after sinterinf for 175 °C for 10 min...... 103

Figure 7.4 Self-assembly of particle into nanomehses...... 104

Figure 7.5 Deposition morphologies of silver particles on a glass substrate. (a) A photograph of the Ag nanoparticle deposition; (b) nanomesh in the red rectangular; (c) irregular deposition in the green rectangular. Particle concentration is 1%, and the pore size of the AAO membrane is 100 nm...... 105

Figure 7.6 Deposition morphologies of nanomesh with different particle concentration and AAO membranes. (a) 3% Ag ink, 55 nm AAO pore diameter; (b) 1% Ag ink, 55 nm AAO pore diameter; (c) 1% Ag ink, 100 nm AAO pore diameter...... 106

Figure 7.7 Self-assembly of Ag nanoparticles into nanotubes in an AAO membrane. (a) Schematic of the formation of Ag nanotubes in AAO; (b) SEM images of the Ag nanotubes after removing the AAO template...... 107

xiii

Figure 7.8 Experimental procedures for fabricating Ag nanotubes on substrates...... 109

Figure 7.9 Effect of substrates and fabrication methods on the integrities of the Ag nanotube forests. Substreates: Glass = glass cover slip, K-T = Kapton tape, silicon = silicon wafer, K-FPC = Kapton FPC film, K-HN = Kapton HN film. A layer of 500 nm thick Ag was thermal evaporated on Kapton FPC film before casting the silver ink drop in sample #8. AAO membrane sizes: the pore sizes are labeled in the schematics; otherwise, membranes with 200 nm pore diameter are used...... 111

Figure 7.10 Three drying processes for Ag nanotubes. Blue: freeze-drying; green: dried in air; red: supercritical drying...... 113

Figure 7.11 Ag nanotube forest dried in air. The nanotubes are bent and self-assembled due to the capillary force on the air-water interface when drying in air...... 114

Figure 7.12 SEM images of Ag nanotube forest dried through freeze-drying process. (a) Overview of the Ag nanotubes, (b) zoom in picture of the nanotube assembly...... 116

Figure 7.13 Ag nanotube forest dried under supercritical drying with and without water removal in anhydrous ethanol before kept in supercritical drier. (a) top and (b) cross- sectional views of Ag nanotube forest fabricated with water removal in anhydrous ethanol; (c) overview of Ag nanotubes in (a) and (b); (d) Ag nanotube forest fabricated without solvent exchange in ethanol before kept in supercritical drier...... 117

Figure 8.1 Experimental setup for observation of drop impact on liquid-infused surfaces...... 121

Figure 8.2 Silicone oil drops impact on: (a) dry AAO, (b) silicone oil infused AAO. Impact velocity is 1 m/s...... 123

Figure 8.3 Silicone oil drop impact on AAO membranes pre-suffused with silicone oil at (a) 0.3 m/s, (b) 1 m/s, (c) 3 m/s...... 124

Figure 8.4 Time evolutions of silicone oil drops impacting on a silicone oil-infused AAO with different velocities. (a) Time evolutions for the positions of the drop front and rear edges; (b) time evolutions of the drop widths and lengths upon impact...... 126

Figure 8.5 Effect of substrates on the drop behaviors upon impact. (a) A silicone oil drop impacts on silicone oil-infused print paper. (b) SEM images of the top views for (b) the AAO membrane and (c) the print paper...... 127

xiv

Figure 8.6 Water drops impacting on the silicone oil-infused print paper with (a) 0.3 m/s, (b) 1 m/s, and (c) 3m/s...... 129

Figure 8.7 Time evolutions of silicone oil drops impacting on silicone oil-infused print paper with different velocities. (a) Time evolution for the positions of the drop front and rear edges; (b) time evolution of the drop widths and lengths upon impact...... 130

Figure 9.1 Evaporation of inkjet-printed colloidal drops on (a) a glass slide and (b) a silicone oil-coated a glass slide. The scale bars are 100 µm in (a) and 50 µm in (b). Particle diameter: 1 µm...... 138

Figure 9.2 Ellipsoid particles fabricated by shape modification from spherical particles...... 140

Figure 9.3 Schematics for the interaction between the contact line and particles in drops with spherical and ellipsoid particles. (a) Force analysis on spherical particles at the contact line; (b) evaporation and particle deposition process for drops with spherical particles; (c) forces on ellipsoid particle at contact line; (d) theoretical analysis of evaporation and particle deposition process for drops with ellipsoid particles...... 142

Figure 9.4 Various depositions in inkjet-printed drops. (a) Deposition of carboxylic- modified 20 nm polystyrene particles (Invitrogen) in a drop with 0.5% volume fraction on a plasma-treated glass slide. (b) Deposition of sulfate-modified 20 nm polystyrene particles (Invitrogen) in a drop with 0.5% volume fraction on a plasma-treated glass slide. (c) Deposition of two adjacent drop containing 0.5% volume fraction of 20 nm sulfate- modified particles on a plasma-treated glass slide. (d) Deposition of two coalescing drops containing 0.5% volume fraction of 20 nm sulfate-modified particles on a plasma-treated glass slide. (e) Deposition of a series of drops containing 0.25% volume fraction of 20 nm sulfate-modified particles with 200 µm drop spacing...... 144

Figure 9.5 Positive dielectrophoresis in evaporating droplets. (a) Schematic for experimental setup for a drop evaporating on a glass substrate with two electrodes. (b) Particle moving towards the gap between two electrodes at the beginning of the drop evaporation. (c) Particles assemble into closely-packed pearl chains in the gap during the evaporation. (d) The closely-packed pearl chain decouples at the last stage of evaporation...... 146

Figure 9.6 Motion of a water droplet upon approached with an acetone saturated swab on (a) an Argon plasma treated glass slide and (b) a glass slide as received...... 147

Figure 9.7 Hypothesis of droplet moving on plasma-treated glass, but keeps stationary on glass as received. (a) Analysis of acetone vapor diffusion and absorption by water drop

xv on hydrophilic surfaces. (b) Hypothesized liquid flow in a drop on a plasma-treated glass slide (contact angle < 5°). (c) Hypothesized liquid flow in a drop on a glass slide as received (contact angle = 19°)...... 149

xvi

ABSTRACT Coalescence and particle self-assembly of inkjet-printed colloidal drops Xin Yang Sun Y. Supervisor, Ph.D.

Inkjet printing has received extensive interest as a low-cost and scalable additive manufacturing process for varieties of applications, such as printable electronics, photovoltaics, and microbatteries. The objective of the current research is to advance the fundamental understandings of the transport phenomena in the inkjet printing processes to improve the performance of the inkjet-printed structures and devices.

To study the stability of the printed patterns, the recoil and oscillations of single and consecutively printed pure liquid drops with different surface tension and viscosity are studied on substrates with different wettabilities. For both single and combined drops, the oscillations decay faster on more wettable surfaces due to stronger viscous dissipation.

When a drop impacts on a sessile drop on hydrophobic substrate, the combined drop recoils twice resulted from the coalescence of the two drops. Whereas no recoil for the single drop impact. A single-degree-of-freedom model is developed for drop oscillation period and duration.

Moreover, the particle deposition dynamics in two coalescing drops is examined with different drop spacing and jetting frequencies. The circularity of the deposition decreases with the increasing drop spacing and decreasing jetting delay, and the radius of curvature at the second drop side firstly decreases and then increases with increasing drop spacing. The capillary flow induced by the local curvature variation in the coalescing drops redistributes particles inside a merged drop, causing suppression of the coffee-ring

xvii effect for the case of a high jetting frequency while resulting in a region of particle accumulation in the middle of the merged drop at a low jetting frequency.

Once colloidal drops are printed on very hydrophilic surfaces, the nanoparticles deposited into various structures, such as concentric multi-rings, radial spokes, spider web, foam, and island-like depositions, as a result of the competition between the receding contact line and particle deposition during drop drying. Based on Marangoni instability, the functional relationship between the characteristic lengths of the deposition and drop drying conditions are established.

Finally, using nanoporous templates, the Ag nanotube forests are fabricated. The effect of drying methods on the morphology of nanotube forests is examined.

CHAPTER 1 : INTRODUCTION 1.1 Background

Inkjet printing is a process in which ink droplets are generated and deposited on papers, plastics and other media to create images and characters. The concept originated in nineteenth century, and advanced in the period of 1960 – 1980 to become widely used in the graphic arts and newspaper industry1. In recent years, inkjet printing has attracted extensive research interests as a versatile manufacturing technique, because of its nature as a non-contact, reduced-waste, maskless, low-cost, and scalable approach for fabricating large area of devices, such as thin-film transistors2, 3, organic light emitting devices4, 5, solar cells6, 7, sensors and detectors8, 9. Hence, it is important to develop the fundamental understanding on the mechanisms of this technology.

The inkjet printing process could be distinguished into three stages as shown in Figure 1.1: droplet formation at the inkjet printhead, drop impact and spreading on the substrate, and functional materials self-assembly due to the evaporation of the carrier liquid.

Figure 1.1 Physical stages of inkjet printing

1.2 Literature reviews

1.2.1 Inkjet printing systems

In inkjet printing, there are two kinds of mechanisms to generate drops: continuous inkjet

(CIJ) printing and drop on demand (DOD) inkjet printing. CIJ printing, which generates drops with a diameter of approximately 100 µm, is used mostly for coding and marking applications10;

DOD dominates the graphics and text printing area with smaller drop diameters of 20 - 50 µm1.

As shown in Figure 1.2a, in continuous printing, the liquid passes through a micron-sized nozzle under pressure. The Rayleigh instability causes the liquid column to break up into drops.

These drops are charged at the nozzle, and the unwanted drops in the stream are deflected by a charged deflector into the gutter to be recycled. According the mechanisms of the deflection, a

CIJ printer could be equipped with a binary or multiple deflection system. Drops in a binary deflection system are either charged or uncharged, while they are charged and deflected at different levels in a multiple deflection CIJ printer. However, recycling the ink after exposure to the environment introduces the potential contamination of the ink, which may be harmful for the application in the material printing. Hence, CIJ printing potentially increases the waste of the ink1.

In contrast, DOD inkjet printer generates individual drops when needed as shown in

Figure 1.2b and 1.2c. In a DOD printer, the nozzle is firstly positioned by adjusting the relative location of the nozzle to the substrate, and then a pressure pulse is generated in the nozzle chamber to eject a droplet. Based on the mechanisms initiating the pulses in the nozzle, there are two major DOD systems: thermal inkjet in Figure 1.2b and piezoelectric inkjet in Figure 1.2c. In the thermal inkjet printer, when a drop is required, a pulse is sent to the thin film heater to heat the nearby liquid above the boiling temperature, which allows the formation of a bubble. The

3 bubble generates a pressure pulse to eject a drop. In the piezoelectric inkjet printer, the pressure pulse required for the ejection results from the mechanical actuation. The thermal DOD systems are miniaturized, and widely used in desktop printers. In industrial applications, a majority of the printers adopt the piezoelectric DOD systems, because the thermal DOD systems requires high vapor pressure fluids, and piezoelectric DOD is easy to control the actuation pulse governing the drop size and velocity of any liquid.

Figure 1.2 Schematics of a) continuous inkjet (CIJ) printer, b) and c) drop on demand (DOD) inkjet printers, b) thermal inkjet printer and c) piezoelectric inkjet printer1.

1.2.2 Drop formation

A typical waveform driving the piezoelectric DOD nozzle is shown in Figure 1.3. Upon applying the voltage, a piezoelectric actuator bends to increase the volume in the nozzle chamber allowing the ink to be propelled in, and the ink is settled during the constant voltage. The sudden decrease in voltage reduces the chamber volume and drives the ink jetted out of the nozzle. The

4 constant negative voltage allows the meniscus fluctuation in the chamber to be damped down. At the end of the process, the increase in voltage reduces the pressure in the chamber to keep the ink inside the nozzle without forming satellite drops.

Figure 1.3 A typical voltage waveform applied to the piezoelectric element in the nozzle of an inkjet printer.

To image the drop formation, a flash photography setup with 1 µs temporal resolution and 0.81 µm spatial resolution was developed by Dong et al11, 12. Figure 1.4 shows the complete process during the formation of a water/glycerol mixture (48/52 by weight) drop. Initially, the chamber volume reduction, induced by the retraction of the piezoelectric actuator, pushes the ink out of the nozzle to form a liquid column with a round leading edge (1-3 in Figure 1.4). After a short period, the rate of the liquid flowing out of the nozzle reduces and the liquid column is stretched (4-9 in Figure 1.4). After pinch-off, due to the capillary pressure imbalance in the

5 liquid drop, the tail of the drop at the pinch-off recoils towards the drop head (10-12 in Figure

1.4). As a result of the liquid thread shrinkage between the head and the tail, the ejected drop breaks up (13-16 in Figure 1.4), generating a primary drop (larger drop) and a satellite drop

(smaller drop). The satellite drop catches up and merges with the primary drop (17-20 in Figure

1.4). The excessive energy after merging induces oscillation of the air-liquid interface until it is damped down by viscous dissipation (21-28 in Figure 1.4). Finally, the equilibrium state is reached and the drop falls down as a sphere. Dong et al. also investigated the effect of waveform, liquid surface tension and viscosity on the formation of pure liquid drops11, 12. Increasing liquid viscosity leads to the liquid thread profile changing from flat to a parabolic shape at the early stage of ejection. The surface tension doesn’t play an important role until the stretching process.

For a particulate suspension, the particle loading does not significantly influence the drop formation when it is less than 10% volume fraction, but for higher volume fractions of 15 – 40%, the large number of particles results in a cone-like structure in the thinning region13.

Figure 1.4 Sequence of images of DOD drop formation11.

1.2.3 Drop impact and spreading

The behavior of a drop impacting on a substrate relates to multiple parameters, such as, impact velocity, drop size, liquid properties, substrate roughness, and wettability. The important

2 dimensionless numbers governing drop impact are Weber number We  DV0 / , Reynolds

2 number Re  DV0 /  , Ohnesorge number Oh   / D , and Bond number Bo  gD / ,

where ρ , μ , and σ are liquid density, viscosity and surface tension, respectively; D , and V0 are drop diameter and velocity in flight; and g is gravitational acceleration. Depending on the impact conditions, there are six major outcomes: deposition, prompt splash, corona splash, receding break-up, partial rebound and complete rebound14, 15. The spreading behavior of a

7 deposited millimeter-sized drop on the substrate could be divided into kinematic, spreading, relaxation, wetting, and equilibrium stages as shown in Figure 1.516, 17. Upon impact, the shape

of a drop resembles a truncated sphere, and the spreading factor β=/ D D0 increases with time

0.5 t until β =1, where D and D0 are the drop base diameter on the substrate and diameter in flight, respectively. During the spreading phase, a lamella is ejected from the drop base, and the spreading factor increases due to the drop inertia before impact. After the drop reaches its minimum height, the air-water interface starts to oscillate, and the liquid is pushed flowing both vertically and radially between the drop center and edge during the relaxation phase. Drop recoil may occur on more hydrophobic surfaces during this phase. The oscillation of the drop then gradually decays by viscous dissipation, and the shape of the drop reaches a spherical cap at the end of the relaxation phase. Depending on the surface wettability, the drop further spreads or recedes during the wetting phase under the effect of surface tension and liquid viscosity until the drop reaches its equilibrium contact angle. The drop base radius versus time for a non- evaporative drop spreading under capillary force on a complete wetting substrate follows

Tanner’s law Dt~ 0.1 18.

Instead of the millimeter-sized drops, the drop size used in inkjet printing is usually in the range of 10 to 150 µm1. To observe in-situ the impact and spreading behaviors of these pico-liter drops, Dong et al. developed a flash photography technique in an inkjet printing system11, 12, 19, which enables the studies for the effects of liquid surface tension, viscosity, impact velocity, drop size, and substrate wettabilities on the impact and spreading of these inkjet-printed drops17.

For the inkjet-printed micron-sized and millimeter sized drops with similar Weber and Reynolds number (Re ~ 16 and We ~ 32), the drop spreading behaves similarly in all the kinematic, spreading, and wetting phases. Similar conclusion are also found by Dam and Clerc20 and Dong,

8 et al19. If the drop contains particles, the increase in the particle loading increases the viscosity of the colloidal drop, which in turn slows down the drop spreading process21. Schiaffino and Sonin studied the drop impact with different Weber numbers22. At high We, the drop spreads radially by the impact-induced dynamic pressure gradient, but, at low We, the drop spreading is driven by capillary force.

1.2.4 Drop evaporation and particle deposition

For a pure liquid drop evaporating on a polymer substrate, three distinctive stages are observed23, as shown in Figure 1.6. Once the evaporation initiates, the drop contact line stays pinned. The contact area is fixed and the contact angle keeps decreasing until the static receding

9 contact angle is reached. This is the constant contact area mode. In the second stage, the drop recedes while the contact angle remains constant, which is known as the constant contact angle mode. In the last stage (mixed mode), both the contact angle and contact area decrease until the drop completely evaporates. Similar results are also observed on epoxy and glass substrates for water and decane drops by Bourges-Monnier and Shanahan24.

Figure 1.6 Schematic diagrams of the three distinct stages in the evaporation process of a droplet on a substrate: (a) constant contact area mode, (b) constant contact angle mode, (c) mixed mode23.

The weight evolution of the droplet during the evaporation is measured by Birdi et al25.

The weight of the drop is found to linearly decrease with time, as shown in Figure 1.7. Hu and

Larson theoretically modeled the diffusion driven evaporation of the sessile drops when the contact line is pinned, which shows good agreements with the experimental results26.

Besides the drop shape evolution, because the evaporation rate peaks near the contact line, the fluid flow from the drop center to the drop edge resulted from the pinned contact line, as shown in Figure 1.8a27. In a colloidal drop, this radial flow brings the liquid together with the particles to the drop edge, and deposited at the contact line. The accumulation of these particles enhances the contact line pinning28, and prolongs the constant contact area mode. After the drop has completely dried, a vast majority of the particles are deposited at the drop edge forming the

“coffee ring” effect on a hydrophilic substrate, as in Figure 1.8b27.

Figure 1.8 a) Liquid evaporative flux along the drop surface27. b) A dried colloidal drop with microspheres29. The scale bar is 1 cm.

The morphology of the coffee ring effect is also influenced by the properties of the inks and substrates. Kuncicky and Velev studied the effects of particle volume fraction and substrate wettability on the particle deposition morphologies,30 as shown in Figure 1.9. On a hydrophilic substrate, the constant contact area mode dominates, and the particles are carried to the drop edge to form a coffee ring, a concave shape deposition. However, on a hydrophobic substrate, the constant contact angle mode dictates the drop evaporation process, leaving a convex shape deposition on the substrate. Increasing the particle concentration for drops on both hydrophilic and hydrophobic substrates leads to more particles depositing at the drop centers, promoting the transition from a concave to convex shapes. This transition results from the lack of carrier liquid at the late stage of evaporation.

This coffee ring effect (non-uniform distribution of the functional materials) in the printed deposition seriously harms the performance of the inkjet-fabricated devices. Therefore, great efforts have been carried out to eliminate the coffee ring effect. Park and Moon suppressed coffee ring effect by generating a Marangoni flow inside the inkjet-printed drops31 with a mixture of two solvents. This ink is formulated with a low vapor pressure, low surface tension diethylene glycol and high surface tension, high vapor pressure water. Once a drop is printed on the substrate, the high vapor pressure, high surface tension liquid evaporates faster at the drop edge, which locally reduces the surface tension near the contact line. The surface tension

13 gradient between the drop edge and center enables the Marangoni flow in the opposite direction to the evaporative-driven flow, preventing particle accumulation at the drop contact line. Hence, uniform deposition is achieved. The Marangoni flow could also be generated by the non-uniform temperature distribution along the drop surface by controlling the thermal conductivities of the substrates and the carrier liquid32.

In contrast to the strong coffee ring deposition of spherical particles in evaporating drops, the coffee ring effect could also be suppressed in the drops containing ellipsoid particles33. When a sessile drop is placed on a substrate, the evaporative flow brings the particles to the drop edge.

Due to their shape anisotropy, the particles are strongly trapped at the air-water interface, and assemble into loosely-packed structures under the effect of long range capillary attractions33, 34.

This assembled particle network on the air-water interface avoids the particle accumulation at the contact line, which suppresses the coffee ring effect.

Yunker et al.33 also examined that for mixtures of ellipsoid particles and spherical particles. When the spherical particle diameter is smaller than the minor axis of the ellipsoid particle, they could move under or pass through open branched aggregations of the ellipsoid particles, and form the coffee ring at the drop edge. However, if the diameter of the spherical particles is larger than the minor axis of the ellipsoid particles, they cannot reach the drop edge.

Hence, the coffee-ring is suppressed.

1.3 Objectives

In this research, we aim to develop the fundamental understanding of the complex transport phenomena involved in the inkjet printing process. The proposed research focuses on the control of the impact, spreading, coalescence, evaporation and particle deposition in the inkjet-printed drops to develop novel methodologies for fabricating various structures applied in printable electronics and photovoltaics. The specific objectives of the present study are:

(a) Study the impact and spreading of inkjet-printed single droplets on a solid substrate with low Weber numbers (We ~ O(1)) in which both the drop inertia and capillary effect play important roles.

(b) Develop the droplet oscillation models to analyze and compare the oscillation of a drop impact on a hydrophobic substrate with and without a pre-deposited sessile drop. Also experimentally study the effect of liquid surface tension and viscosity on oscillation amplitude and frequency of these single and two consecutively printed drops.

(c) Investigate the coalescence process of the inkjet-printed pico-liter drops in which capillary forces and evaporation are the dominating mechanisms, and the gravitational effect is negligible.

(d) Examine the effect of jetting delay and the drop spacing on the colloidal drop coalescence, and particle deposition dynamics.

(e) Utilize the coffee ring phenomenon to explore inkjet printing technique as a methodology for fabricating nanoscale structures.

(f) Investigate the mechanism for the formation of various particle deposition morphologies inside inkjet-printed colloidal drops.

(g) Develop quantitative models for predicting the characteristic lengths of deposition patterns under different droplet drying conditions and colloidal properties.

(h) Study the evaporation, infiltration and particle deposition behaviors during inkjet printing process.

(i) Develop the fabrication process for the silver nanotube forest on solid surfaces.

(j) Investigate the behavior of droplets impacting on porous surfaces infused with the miscible/immiscible liquids.

CHAPTER 2 : EXPERIMENTAL MTHODS

For in-situ observation of drop evaporation and particle self-assembly processes of inkjet- printed colloidal drops from side and bottom views, a homemade experimental setup incorporating the printing and observation systems is used as shown in Fig. 2.1.

A waveform generator (JetDrive) is used to adjust the voltage waveform driving the piezoelectric printhead with a diameter of 50-60 μm (MicroFab MJ-Al-01) to generate pico-liter drops. The drops are printed on a substrate supported by a microscope stage (Ludl MAC 6000).

The printhead, substrate and the microscope stage are enclosed in a humidity chamber to control the drop evaporation conditions. A high resolution SensiCam QE CCD camera (Romulus,

Michigan) or high speed camera (Phantom V210) with a Navitar 12× Zoom lens (Rochester,

New York) and a halogen strobe light (Perkin Elmer) are used to capture side-view images for drop formation, impact, spreading, coalescence and evaporation processes. The videos of these processes are captured based on the flash photography technique introduced by Dong et al.11.

During evaporation of colloidal drops, the particle motion and deposition processes are observed by using a Zeiss inverted fluorescence microscope (Thornwood, New York) with a 40× oil objective. Bottom-view images are captured by a Sony XCL-5005CR CCD camera (Park Ridge,

New Jersey) at 10 frames per second. The waveform generator, two CCD cameras and the strobe light are synchronized with a delay generator (SRS DG645, Sunnyvale, California) to enable the flashphotography technique.

Figure 2.1 Schematic of the experimental setup for in-situ observation of droplet evaporation in inkjet printing.

2.1. Materials preparation

In this study, colloidal drops are printed to study the particle deposition dynamics in the inkjet printing processes. The stability of colloids depend on the particle surface property, concentration, and density etc. Hence, a stable ink without particle aggregation is the prerequisite for the inkjet printing studies and applications. When formulating the inks, particle diameters in the colloids should be less than 2% of the nozzle diameter to ensure stable jetting35. Increasing the particle size or concentration may dramatically increase the probability of nozzle clogging during printing. In this study, carboxylate-modified polystyrene beads of 1.1 µm, 100 nm and 20 nm, and sulfate-modified polystyrene beads of 20 nm and 200 nm are purchased from Invitrogen.

Silver ink with particle size of 8 – 15 nm (UTDAgIJ) is purchased from UTDots. These inks are diluted into different concentrations after 15 min of sonication in a water bath sonicator (Cole-

Parmer 8891) to perform different experiments. After dilution, the ink is sonicated for 5 min for homogenization.

Substrate silanization and plasma treatment are used to obtain substrates with various degrees of wettability. Glass microscope cover slips (Bellco, ~150μm thick) were consequently cleaned by sonicating in Sparkleen/DI water solution, ethanol and acetone for fifteen minutes each. Then, they were rinsed with DI-water and immediately dried in a stream of compressed air.

Argon gas plasma treatment for 2 minutes for both sides is accomplished at 18W and 250mtorr in Harrick Plasma PDC-32G. By soaking the plasma-treated substrates in a 5mM octadecyltrichlorosilane (OTS (CH3-(CH2)17-SiCl3)) solution in hexane (both from Sigma

Aldrich and used as received) for one hour followed by thermal annealing under vacuum at 60

°C for 1 hour, the silanization is completed, and the glass substrate becomes hydrophobic.

Substrates with various wettabilities are achieved by subjecting the already silanized surfaces to

Ar plasma treatment with different durations (2 seconds to 2 mins), and there is no measureable changes in the surface roughness for these substrates.

The static contact angle, static receding and advancing contact angles of droplets on various substrates are measured by a homemade goniometer setup. A micropipette is used to gently place a drop of water on top of a substrate to be measured. Side-view images of the drops are captured with a camera (Basler a601-f) equipped with a Navitar Zoom 7000 lens and a light source (Fiber-Lite MI-150). The images are measured with ImageJ to obtain the drop contact angles on substrates. To measure the static advancing/receding contact angle, a pipette is used to continuously add/remove liquid from the drop. The advancing/receding contact angle is obtained

19 by measuring the contact angle of the drop just before spreading/receding from series of captured images.

Figure 2.2 Schematic of the goniometer setup to measure contact angle of the liquid drop.

2.2. Process control

The piezoelectric printheads with 50 or 60 μm orifice diameter used in this research are from MicroFab Technologies Ltd. These printheads are driven by voltage pulses sent by Jetdrive in which the voltage waveform driving the jetting could be set. To ensure a stable jetting, negative pressures (whose values depend on the printed inks) is applied to the ink reservoir connected to the printhead. This negative pressure is controlled by tuning the pressure and vacuum in a pneumatic pressure regulator purchased from MicroFab. During the purging and cleaning, a positive pressure is applied to the ink reservoir through the pressure regulator, which constantly pushes liquid out of nozzle as a straight stream.

An enclosed humidity chamber housing the printhead, camera, strobe-light and microscope is used to control the drying conditions of printed droplets. The schematic of the system to control the environmental humidity is shown in Fig. 2.3. The dry compressed air is pumped into the chamber to reduce the relative humidity, or a beaker of water to carry moisture into the chamber for increasing the relative humidity. The humidity inside the chamber could be controlled between 10% and 90%, and the humidity in the chamber is measured with two OM-

DVTH humidity sensors.

Figure 2.3 Schematic of the humidity chamber.

2.3. In-situ observation

Three methods of drop observation are conducted for research. A) Side-view observation of the drop formation at the nozzle orifice; B) side-view observation of the drop impact, spreading and evaporation; C) bottom view observation of the drop evaporation and particle self-

21 assembly. The side-view and bottom-view cameras are synchronized with a drop ejection through a delay generator.

2.3.1. Drop formation

In experiments, the side-view observation of the drop formation is used to examine the printing instability, drop size and diameter. The side-view imaging system comprises a Sensicam

QE CCD camera coupled with a halogen strobe (Perkin Elmer) as shown in Figure 2.1. The mechanism for capturing the drop formation process is flashphotography shown in Figure 2.4.

Only one image is captured for each single drop, and the drop formation process could be obtained by continuously increasing the delay times of the side-view imaging system (camera and strobe) and the printhead by 10 µs. Drop diameter and speed in flight are measured with

IMAGEJ (http:// rsbweb.nih.gov/ij/).

Figure 2.4 Schematic of flashphotography for visualization of the drop formation process.

2.3.2. Drop impact, spreading, and evaporation

Side-view images for drop impact and spreading are captured with the same side-view imaging system. During the observation, the camera is focused on a location of a substrate where

22 a drop impacts. The camera runs at 10 Hz when using the high resolution Sensicam QE CCD camera, and 50 kHz for Phantom V210 high speed camera. After taking the images, the drop diameters on the substrate and contact angles are measured with IMAGEJ. However, during the last stage of evaporation, the drop contact angle becomes smaller than 5° which makes accurate measurement impossible.

2.3.3. Particle self-assembly

After drop impact on a substrate, the evaporation and particle self-assembly is observed by SONY XCL-5005CR camera installed on an inverted fluorescent microscope (Zeiss Axio

Observer). The resolution of captured images varies from 3.45 µm × 3.45 µm to 0.63 µ× 0.63

µm depending on the lenses. With a 40× oil objective lens, the system allows the observation of the deposition behaviors of individual fluorescent particles with 1.1 µm diameter. Fluorescent filter sets 38 endow GFP (excitement 450/490, and emission 500/550) and 45 HQ texas red

(excitement 540/580, and emission 595/665) are used. The camera is triggered by a delay generator at 10 fps to compare the results from side-view observations.

CHAPTER 3 : OSCILLATION AND RECOIL OF SINGLE AND CONSECUTIVELY PRINTED DROPLETS 3.1 Introduction

Drops impact on a dry or a wet surface can result in different morphologies, from spreading and splashing to receding and rebounding, depending on the impact velocity, drop size, and properties of the liquid and substrate36. Yarin14 carried out a detailed review of drop impact on a solid substrate and on a shallow liquid film. Majority of studies focused on high-Weber- number drop impact that leads to crown-shaped splash37-40. However, for applications such as inkjet printing and spray deposition, low-Weber-number impact leading to drop spreading or

"deposition" is a preferred process for material deposition. Drop spreading dynamics and the subsequent relaxation of low-Weber-number impact are important for a better control of the deposition morphology.

For the regime of "deposition" impact, the drop spreading dynamics can be distinguished into five consecutive phases: kinematic, spreading, relaxation, wetting, and equilibrium (see

Figure 3.1)16. In the kinematic phase upon drop impact, the drop spreading radius increases with

0.5 t until it reaches the drop in-flight radius. In the following drop spreading phase, the contact line advances under the influence of surface tension, viscosity, and drop impact velocity. When the drop impact kinetics is either consumed by viscous dissipation, transferred into surface energy, or the combination of both, the drop reaches its minimum height and the drop spreading radius at the end of the spreading phase is often referred to as the primary spreading radius.

With the drop transitioning into the relaxation phase, the air-water interface oscillates and pushes the liquid to flow both radially and vertically inside the drop. Drop recoil can occur during this phase and recoil becomes more significant with the decrease in substrate wettability41, 42 and

Ohnesorge number ( Oh   D0 ), or with the increase of the Weber number

( We  D U 2  )43. Here,  is the liquid density, D the drop in-flight diameter,  the liquid 0 0 viscosity, U the impact velocity, and  the surface tension. Drop oscillation then gradually decays by viscous dissipation and the drop reaches a spherical cap shape at the end of the relaxation phase. Additional spreading occurs in the wetting phase until the drop reaches its equilibrium contact angle. In the wetting phase, the drop spreading radius versus time for a non- evaporative drop on a very hydrophilic substrate follows the Tanner’s Law44, R ~ t 0.1 , as a result of the competition between surface tension and viscosity.

Kinematic Spreading Relaxation Wetting We ~ O(1) Interface Capillary We >> 1 R ~ f (Re, oscillation spreading

We, θeq, t) R ~ t0.5

pinned CL

Spreading radius Spreading substrate

Time

Figure 3.1 Schematic representation of droplet spreading radius as a function of time for drop impact on a hydrophilic substrate at low and high We numbers.

For a single droplet impact onto a sessile drop of We >> 1, drop spreading dynamics is expected to be similar to a single drop impact on a thin liquid film. In both cases, the condition of the substrate plays an important role13, 14. Fujimoto et al.45, 46compared drop impact onto a sessile drop with that on a dry surface. The results show that the threshold velocity for drop impact to produce crown-shaped splash decreases in the presence of a sessile drop, as compared to a dry surface.

Until now, most studies have been concerned with drop impact at large Re and We

47 numbers (e.g., Re  UD0  = 100 - 200 and We >> 1) . In such a regime, the inertia is more dominant than the surface energy. For inkjet-printed micron-sized droplets, strong surface tension effects upon impact12, 17, 20, 47, 48 lead to different droplet spreading and oscillation damping behaviors than millimeter-sized drops of large Re and We numbers. Figure 3.2 summarizes the parameter space of recent studies14-18 of inkjet-printed micron-sized drops in a

We-Oh plane. The regime of We < 1 and Oh < 1, where inertia, capillary, and viscosity simultaneously affect the drop impact and spreading, has only been explored by Son el al.47 and van Dam and Le Clerc20 using inkjet-printed water droplets on substrates with different wettabilities. The effects of surface tension and viscosity on drop oscillation and recoil, especially for consecutively printed drops, have not been investigated.

In this study, recoil and oscillation of inkjet-printed drops impact on a dry surface and on another sessile drop of the same liquid are examined using a high speed camera. Substrate wettability is systematically modified to reach 0 - 90 equilibrium contact angle while keeping surface roughness the same. Pico-liter drops of water, water/glycerin mixture (W/G), and water/glycerin/isopropanol mixture (W/G/I) are used to independently examine the effects of surface tension and viscosity. We focus on the low We number impact that results in drop

26 deposition. Schematic illustration of droplet spreading radius as a function of time for drop impact on a hydrophilic substrate at low and high We numbers is shown in Figure 3.1. Due to different drop inertia, the kinematic, spreading, relaxation, and wetting phases for We >> 1 and

We ~ O (1) are expected to be different. Since the kinematic phase usually takes only ~ 10 s, at most one data point can be obtained within this time frame by changing the delay time between drop ejection and camera exposure. In addition, the initial drop impact becomes less important in the wetting phase. Much attention of this study is paid to the spreading and relaxation phases.

The similarities and differences of drop impact on a dry surface and a sessile drop are presented.

A single-degree-of-freedom vibration model is used to predict the height oscillation of single and combined drops on a hydrophobic substrate. The oscillation frequency and damping amplitude obtained from model predictions are in good agreement with our experimental results of drop oscillation on a hydrophobic substrate with and without a pre-deposited sessile drop.

Figure 3.2 Summary of the parameter space in a We-Oh plane for studies on inkjet-printed drops.

3.2 Experimental method

In the experiments, pico-liter drops were generated using a piezoelectric-driven ink jetting nozzle with a diameter of 60 µm (MicroFab MJ-AL-01). A Phantom v12.1 high-speed camera (Wayne, New Jersey) with a Navitar 12 Zoom lens (Rochester, New York) and a microscope excitation light source (X-Cite 120Q) were used to capture drop impact, spreading, and relaxation on a substrate with and without a pre-deposited sessile drop. Drop ejection and camera system were synchronized with a delay generator (SRS DG645, Sunnyvale, California).

Two drops with controlled temporal delay, created using a waveform generator (MicroFab

Jetdrive) and a delay generator, were printed consecutively on top of each other. Images were taken at 50,000 fps with pixel resolution of 256 × 128 pixels at 1.667 µm/pixel. To obtain better temporal resolution, the high speed camera is combined with the flash photography technique12

28 using different delay times between the camera exposure and drop ejection (e.g., 5, 10, 15 µs), based on the high repeatability nature of the inkjet printing process.

Besides pure DI-water, a mixture of water and glycerin with a weight ratio of 52:48, and a mixture of water, glycerin, and isopropanol 53:34:13 in weight were used to independently examine the effects of viscosity and surface tension on drop relaxation. The water/glycerin mixture has a viscosity of 5 cP, five times that of water, but a surface tension of 68 mN/m similar to 72 mN/m for water and a density of 1.1 g/cm3 slightly higher than 1 g/cm3 for water.

Comparing the properties of the water/glycerin/isopropanol mixture (W/G/I) with the water/glycerin mixture (W/G), the surface tension decreases to 35 mN/m for W/G/I while the density becomes 1.04 g/cm3 and the viscosity remains about the same. The resulting We numbers for W, W/G and W/G/I drops are 0.49, 0.58, and 1.06, and the K values (K = WeOh-2/5) for W,

W/G, and W/G/I are 2.6, 1.6 and 2.59, respectively.

To obtain substrates with varying degrees of wettability, glass microscope cover slips

(Bellco, ~150μm thick) underwent a sequence of cleaning, silanization, and plasma treatment.

Substrates were cleaned by sonicating in Sparkleen-DI water solution, ethanol and acetone for fifteen minutes each, immediately dried in a stream of compressed air and treated with Argon gas plasma (2 minutes each side at 18W and 250 mtorr, Harrick Plasma PDC-32G). Silanization consisted of soaking the substrates for one hour in a 5mM solution of octadecyltrichlorosilane

(OTS (CH3-(CH2)17-SiCl3)) in hexane (both from Sigma Aldrich and used as received). This silanization was carried out inside a dry glove box with a controlled N2 atmosphere to minimize excessive polymerization of the water-sensitive OTS molecules. Varying degrees of wettability

(different equilibrium contact angle eq ) were obtained on the substrates by subjecting the already silanized surfaces to time-varying treatments of Ar plasma. In the current study, the time

29 of treatment from 2 to 60 seconds at a power level of 6W and 250 mtorr Argon pressure was used to produce systematically varied wettability of glass surfaces with no measurable change in surface roughness. Equilibrium contact angles of the substrates were measured by using a goniometer setup and the LB-ADSA drop analysis plug-in of Image J, where the drop surface was assumed to be a spherical cap and the gravitational effect was neglected due to a very small

2 Bond number (Bo = rgD0 s < 0.005, where g the gravitational acceleration).

The drop in-flight diameter and impact velocity were kept at 60µm and 0.77 m/s respectively. The ambient temperature maintained at 22 °C and the relative humidity was around

50 %for all experiments. For a drop impacting on a dry surface, the starting time (t = 0) is defined as the bottom of the drop touches the substrate. Similarly, t = 0 for the case of a drop impacting on a pre-deposited sessile drop is when the second drop contacts the top of the first drop.

3.3 Results and discussion

3.3.1 Spreading and oscillation of a single drop

A. Effect of substrate wettability

Figure 3.3 shows snapshots of a single DI-water droplet impact on substrates with

equilibrium contact angleeq of 0°, 45°, and 90°. Upon impact, the drop deforms under inertia until it becomes a disc like shape around t = 25 µs, where the drop reaches its minimum height.

At this time, the drop kinetic energy is partly converted into surface energy due to increased surface area and partly consumed by viscous dissipation. It has been reported that, at a high

Weber number (We >> 1), inertia dominates the drop impact and during the initial spreading

30 regime (the kinematic and spreading phases shown in Figure 3.1) the drop spreading kinetics, i.e., drop radius versus time, is independent of the substrate wettability17. However, for low-

Weber-number drop impact (e.g., We = 0.49 for DI-water droplets in this study), both inertia and capillary effects are important in the initial spreading regime. As shown in the drop spreading radius versus time plot of Figure 3.4a, in the initial fast spreading regime (from drop impact up to 18-33 µs, depending on the substrate wettability), the spreading radius R versus time plot shows three different slopes for wetting angles of 0°, 45°, and 90°. The uncertainties of drop radius and height measurements in this study are 0.83µm.

Both theoretical and empirical approaches have led to various correlations for the contact line mobility relationship. For example, for Ca < O(1), the Hoffman-Voinov-Tanner law44, 49, 50 gives

3 3  eq ~ Ca , where the capillary number Ca  u / and u is the contact line velocity. With

other conditions maintain the same, different values of eq result in different contact line

3 3 mobilities. The fastest capillary spreading occurs at eq  0, where the driving force, eq , is the largest. The dependence of drop spreading kinetics on the substrate wettability is a result of the capillary effect in the initial spreading regime.

Figure 3.3 Snapshots of single water droplets impact on substrates of different wettabilities. For all cases, the drop velocity is 0.77 m/s, the diameter is 60 µm, and We = 0.49.

100 70 60 80 50

60 m)

m) 40



30 40

W o W o

Height ( 20 Radius (  ~ 0 - single  ~ 0 - single eq eq 20 W o W o  = 45 - single  = 45 - single eq 10 eq W o  W= 90o - single  = 90 - single 0 eq 0 eq 10-1 100 101 102 103 104 105 106 0 50 100 150 200 250 300 Time (s) Time (s)

(a) (b)

At the end of the initial spreading regime, drop reaches its minimum height (e.g., t  25

s) where the kinetic energy reduces to zero. However, at this stage, the shape of the drop is far from equilibrium. Due to inertia, a large amount of liquid in the drop is squeezed to the contact line region leading to a large curvature at the air-water interface near the contact line, which locally increases the liquid pressure. A pressure gradient is hence induced between the center and the periphery of the drop pushing the liquid to flow inward to the drop center41 until the drop reaches a maximum height (t  45 s in Figure 3.3) with a large local curvature at the center of the drop surface. A reversed liquid pressure gradient is then set up and causes an outward liquid flow until a large local curvature builds up again at the contact line region to stop the liquid motion. Under the effect of the surface tension and inertia, the liquid inside the drop oscillates

33 back and forth between the contact line and the center of the drop in the relaxation phase until the oscillation is completely damped down by viscous dissipation and the drop shape reaches a spherical cap. It is important to note that, in contrast to the cases with We >> 1 where the contact line is often pinned during the drop relaxation phase of pico-liter drops17, for drops with We ~ O

(1), capillary spreading takes place during drop oscillation in this study. As shown in Fig. 3.4a

W W for the cases of eq ~ 0° and eq = 45°, during the relaxation phase (from the end of initial spreading to about t = 100 µs), the drop spreading radius increases while the height of the drop oscillates, as shown in Figure 3.4b.

Figure 3.4b also shows that, upon drop impact, it takes a longer time (~ 270 µs) for the

viscous stress to dissipate the height oscillation of a drop impact on a hydrophobic substrate (eq

= 90°). In contrast, for a hydrophilic substrate ( ~ 0° or 45°), the stronger capillary spreading prior to the relaxation phase leads to a larger drop contact area on the substrate, which in turn increases the viscous dissipation51,52 and results in an relaxation phase of less than 100 µs. This can also be clearly observed in Figure 3.3, where the drop shape changes between t = 145 and

165 µs on the substrate of = 90° but not on the other two cases with ~0° and = 45°.

After drop oscillation, the capillary effect dominates further spreading of the drop in the wetting phase until it reaches equilibrium. As shown in Figure 3.4a, for a hydrophilic substrate of

~ 0°, the contact line advances faster in the wetting phase due to a stronger capillary driving

3 3 force,  eq . For a hydrophobic substrate of = 90°, however, the drop already reaches its equilibrium contact angle at the end of initial spreading and hence no further spreading is observed in the wetting phase.

B. Effect of surface tension and viscosity

The effects of surface tension and viscosity on drop spreading and oscillation are then

investigated based on a hydrophobic substrate ( eq = 90°). The drops of water (W), water/glycerin mixture (W/G), and water/glycerin/isopropanol mixture (W/G/I) are used. As shown in the drop spreading radius versus time plot of Figure 3.5a, at the end of the initial spreading regime where the drop reaches the minimum height, the water drop spreads the least

extent due to its high surface tension ( w = 72 mN/m). For the W/G/I drop of the same viscosity

5 as that of W/G (  W/G =  W/G/I = 510 Pa·s), its lower surface tension ( = 38mN/m) leads to farthest spreading among three drops.

40 70 0  = 900, single drop  = 90 , single drop eq 65 eq 35 60

30 55



25 45

Height ( Radius ( 40 20 W W W/G 35 W/G W/G/I 15 30 W/G/I 0 50 100 150 200 0 50 100 150 200 250 300 Time (s) Time (s)

(a) (b)

Similar to water, the W/G and W/G/I drops also oscillate after impacting on a hydrophobic substrate. As shown in Figure 3.5b, it takes about 250 µs for the height oscillation

5 of a water drop ( W 10 Pa·s) to damp down, while the oscillation of the W/G and W/G/I drops last less than 100 µs due to stronger viscous dissipation resulted from a higher liquid viscosity. The results also show that, by comparing the damping time of the W/G and W/G/I drops, surface tension does not play a significant role in oscillation damping for a single drop impacting on a hydrophobic substrate at a low Weber number.

3.3.2 Coalescence, recoil, and oscillation of two consecutively printed drops

A. Effect of substrate wettability

Now consider the scenario when a drop impacts on a sessile drop consisting of the same liquid. Figure 3.6 shows snapshots of a water drop of 30 µm in-flight radius impact on another

evaporating water drop sitting on substrates of eq ~0°, 45°, and 90°, respectively. Here, t = 0 is when two drops first touch each other and the impact velocity is at 0.77 m/s. As shown in Figure

3.6, upon impact of the second drop on top of the first drop (e.g., t = 4 µs), the local mean curvature of the air-water interface at the neck region between two drops is very large. The unbalanced surface tension and pressure in the neck region results in a fast expansion of the neck width53, pumping liquid from both drops into the neck region. The contact angle of the first drop

on substrates is hence decreased. For the substrate of eq = 90°, the contact angle of the first drop decreases below its static receding contact angle and drop recoil is observed at t = 20 - 30

µs, as shown in Figure 3.6 at t = 24 µs. However, recoil is not observed on hydrophilic substrates of ~ 0° and 45°. For such cases, the dynamic contact angle of the first drop is still larger than

36 the static receding contact angle. This can be explained by the increase of contact angle hysteresis with substrate hydrophilicity54.

Figure 3.6 Snapshots of a water drop impact on a sessile drop on substrates of different wettabilities. The drop impact velocity is 0.77 m/s, diameter is 60 µm, and We = 0.49.

It is also observed that, at t = 24 µs in Figure 3.6, a tip appears at the center of the drop

W surface for substrates of eq = 45° and 90°. Note that, for eq ~ 0°, similar phenomenon is also observed at around t = 30 µs. Because the drastic capillary driven expansion of the neck region is faster than the downward motion of the second drop, and the liquid in the center top region of the second drop consequently lags behind. This tip at the drop top surface results in a pressure gradient to accelerate the liquid to flow downward, and gives rise to an inertia pressure overshoot

37 which in turn pushes the liquid radially toward the edge of the drop22. This pressure overshoot

leads to a crater like depression of the combined drop on a hydrophilic substrate of eq ~ 0° at t =

64 - 84µs, as shown in Figure 3.6. On a hydrophobic substrate, however, the overshoot increases impact inertia causing the combined drop to spread more than the equilibrium spreading radius during initial spreading. As a result, the surface tension drives the liquid from periphery back to the center of the combined drop, causing the contact angle to be smaller than the static receding contact angle where a second recoil is observed at t = 64 – 84µs.

Similar to the effect of substrate wettability on drop oscillation time for a single drop, it takes a shorter time for the oscillation of a combined drop to be dissipated on a hydrophobic substrate compared to a hydrophilic one, i.e., the oscillation lasts 144 µs on a substrate of ~ 0° versus more than 204 µs on a substrate of = 90° as shown in Figure 3.6 for a combined water drop.

B. Effect of viscosity and surface tension

Figure 3.7 shows snapshots of W/G and W/G/I drops impact on a sessile drop of the same

liquid on a hydrophobic surface of eq = 90°. The corresponding drop spreading radius versus time and drop height versus time for W, W/G, W/G/I drops are compared in Figure 3.8. In contrast to water drop, no recoil and tip are observed for W/G and W/G/I drops at t = -3 –33 µs.

Higher viscosity of both W/G and W/G/I drop slows down droplet coalescence (i.e., neck region expansion)53, giving enough time for the second drop to provide the liquid needed for neck expansion. In addition, no second recoil is observed as a result of higher viscous dissipation and the lack of pressure overshoot leads to the absence of more than equilibrium spreading radius.

The two recoils for water drop and the lack of recoil in W/G and W/G/I drops can be clearly observed in the spreading radius versus time plot of Figure 3.8. Figure 3.7 also shows that, due to stronger viscous dissipation, the oscillations of the combined W/G and W/G/I drops are both damped down in ~ 217 µs much less than that of 500 µs for water drop as shown in the drop height versus time plot in Figure 3.8b. Similar oscillation decays for W/G and W/G/I also indicate that the effect of surface tension does not play a significant role in oscillation damping.

120 50 W W - Eq. (5) 45 105 W/G W/G - Eq. (5) 40 90 W/G/I W/G/I - Eq. (5)



 0 75

35 0  = 90 , combined drop  = 90 , combined drop eq eq 30 60

Height (

Radius ( 25 W 45 W/G 20 W/G/I 30 0 40 80 120 160 200 0 100 200 300 400 500 600 Time (s) Time (s)

(a) (b)

3.3.3 Comparison between drop impact on a dry surface and on a pre-deposited sessile drop

A. Recoil in drop impact on a sessile drop

When two water drops consecutively impact on a dry substrate, due to drop coalescence, the spreading and oscillation of the combined drop differ from that of a single drop. Since the jetting delay between two drops is 0.02s, much longer than the oscillation and spreading time for the single drop, the first drop has already equilibrated into a spherical cap before the impact of the second drop. Moreover, the jetting delay is only about 0.6% of the time for a single water drop to evaporate. The effect of evaporation of a water drop is hence neglected in this study.

Figure 3.9 shows the comparison between drop impact on a dry surface and on a sessile drop. In contrast to a single drop that continuously spreads on a hydrophobic substrate until it

40 reaches an equilibrium contact angle, two stages of drop recoil as discussed in Sec 3.3.2A are observed for the combined drop as a result of drop coalescence as shown in Figure 3.9. It has been reported that on a hydrophobic substrate, no recoil is observed for low-Weber-number drop impact on a dry surface, while increasing the Weber number can result in the emergence of drop recoil15. It can hence conclude that, drop impact on another sessile drop is capable of decreasing the threshold We number for drop recoil.

50 120 W - single 45 W - single - Eq. (5) W - combined 40 100 W - combined - Eq. (5)  = 900 eq

m) 35 m)



80 30

Height (

Radius ( 25  = 900 60 eq 20 W - single W - combined 15 40 0 50 100 150 200 0 100 200 300 400 500 600 Time (s) Time (s)

(a) (b)

B. A single-degree-of-freedom vibration model of drop height oscillation on a hydrophobic substrate

When a drop impacts on a surface or on a sessile drop, the single or combined drop deforms and spreads under the effects of inertia and surface tension, causing the drop to oscillate between minimum and maximum drop heights. During oscillation, energy is dissipated by viscous stress, which eventually brings the drop to rest. Treating drop oscillation on a hydrophobic substrate as a single-degree-of-freedom vibration system consisting of mass, spring, and damper, the inertia resulted from drop impact is given by

3  Finer ~ Rn H (3-1)

3 where Rn is the drop nominal radius and Rn  3V 4  , V is the volume of the single or combined drop, H is the drop height. When the drop shape deviates from a spherical cap, the surface tension force works to retain a spherical cap shape and it is scaled as

Fsur ~  H - Heq  (3-2)

During the drop oscillation phase, the viscous drag force for oscillation damping is scaled as

 Fvis ~ Rn H (3-3)

Therefore, the height oscillation of both single and combined drops governed by inertia, surface tension and viscous drag forces can be balanced by

    H  C1 2 H  C2 3 (H  H eq )  0 (3-4) Rn Rn where the coefficient C1 is related to oscillation amplitude, C2 is related to frequency, and the oscillation damping amplitude, H(t) , is given by

 -C (t-t ) 1 2 0 2Rn H (t)  H 0e (3-5)

42 where H is the initial amplitude of the oscillation when the drop first reaches its minimum 0 height at t0. During oscillation, liquid inside the drop flows both radically and vertically. On a hydrophilic substrate, the drop spreading radius is larger than that on a hydrophobic substrate, thus the stronger radical motion of the liquid plays a more important role in oscillation damping.

However, the single-degree-of-freedom model considered here is only appropriate for the drops oscillating on a hydrophobic substrate where the vertical flow is more dominant during oscillation.

C. Height oscillation

It is found that for C1= 6.28, the damping of the oscillation amplitude calculated by Eq.

(3-5) best fits to the experimental data of the height oscillation of a single drop on a hydrophobic substrate shown in Figure 3.9b. With the same value for C1, the evolution of the oscillation amplitude for the combined drops are calculated by using Eq. (3-5). Note that the effect of evaporation is neglected in the amplitude calculation. As shown in Figures 3.8b and 3.9b, the results of the single-degree-of-freedom vibration model is in good agreement with the experimentally observed drop oscillation dynamics on hydrophobic substrates for both single and combined water, W/G, and W/G/I drops. The single-degree-of-freedom vibration model, can be used to explain the effect of viscosity to mass ratio on oscillation damping, as shown in the exponent of Eq. (3-5). The model can also predict the slower oscillation damping of a combined water drop as compared with a single drop as shown in Figure 3.9b: the volume of the combined drop is twice as that of the single drop and a larger value of Rn leads to a larger oscillation amplitude in Eq. (3-5), thereby a longer damping time. Similarly, as shown in Figure 3.8b, faster oscillation damping of W/G and W/G/I drops compared with the water drop can also be

43 predicted with a larger viscosity value in Eq. (3-5). Moreover, the absence of surface tension term in Eq. (3-5) indicates that the surface tension does not play an important role in oscillation damping. This can be observed in Figure 3.5b and 3.8b for both single and combined W/G and

W/G/I drops.

In addition, the period of height oscillation can be rewritten as

1 R 3 T   C n (3-6) f 2 

3 and can be scaled with the oscillation period of a combined water drop TW  C2 W RW  W as

T R 3 /  R 3  n w w (3-7) TW  / w

Equation (3-7) indicates that, the mass and surface tension rather than viscosity are the major factors influencing the drop oscillation period. The scaled oscillation periods of the single and combined water and W/G drops are shown in Figure 3.10, based on measurements from

Figures 3.8b and 3.9b. It is found that the scaled periods of the single water drop, combined water drop of 0.02s jetting delay, combined water drop of 0.5s jetting delay, and combined W/G

drop of 0.02s jetting delay all fit fairly well with the calculated curve for a water drop (   w ) for the condition considered. As shown in Figure 3.10, although the viscosity of W/G is 5 times larger than that of water, its surface tension 94% of water leads to an oscillation period close to the water curve calculated from Eq. (3-7). The error bars in Figure 3.10 are standard deviations of several repeating experiments.

By neglecting the evaporation of the first W/G/I drop, the scaled period of the height

oscillation, T TW , based on Eq. (3-7) is 1.37. However, the experimentally measured scaled period is 1.07. The volatile isopropanol owning to a 5.7% volume decrease during the 0.02 s

44 jetting delay causes the drop density between W/G and W/G/I and the viscosity identical to W/G.

Thus, the damping rate, which is related to density, viscosity and drop volume, is not significantly affected by evaporation. However, the time-dependent concentration of the isopropanol significantly affects the surface tension of the combined drop, thereby affecting the oscillation period. Using Eq. (3-10) to estimate the oscillation period of a combined W/G/I drop without considering the effect of evaporation becomes inappropriate.

1.6  = 900 eq 1.2

W 0.8

T/T water drop - Eq. (7) 0.4 W - single W - combined (0.02s delay) W/G - combined (0.02s delay) 0.0 W - combined (0.5s delay) 0.0 0.5 1.0 1.5 3 3 R /( R ) W W

Figure 3.10 Scaled oscillation period of the water and W/G drops as a function of scaled drop mass.

3.4 Conclusion

In this paper, the recoil and oscillation of single and two consecutively printed drops on substrates of different wettabilities are examined by side-view images using a high speed camera.

The results show that, for a single droplet impact on a dry surface at We ~ O(1), both inertia and capillary effects play important roles in the initial spreading regime leading to substrate wettability dependent spreading kinetics before the droplet starts to oscillate. This observation is very different from the inertia-dominated drop impact at We >> 1, where variations in drop spreading kinetics with substrate wettability are negligible.

On a substrate of higher wettability, drop oscillation decays faster due to its stronger viscous dissipation resulted from a longer liquid oscillation path parallel to the substrate. Drops with a high viscosity (W/G and W/G/I) oscillate shorter in time than low viscous drops (water).

Similar results on the effect of wettability and viscosity on drop oscillation time are found for the combined drop after a drop impacts on a sessile drop of the same liquid.

It is found that, when a drop impacts on an evaporating sessile drop sitting on a hydrophobic substrate, as two drops coalesce and the combined drop subsequently spreads on the substrate, drop recoil twice due to the coalescence of the two drops. This stands in contrast to the absence of recoil for a single drop impact on a dry surface under the same condition, indicating that the threshold value for drop recoil is lowered in the presence of another sessile drop.

Furthermore, a single-degree-of-freedom vibration model for the height oscillation of the single and combined drops on a hydrophobic substrate is established. The model predictions are in good agreement with the experimental results of the drop relaxation phase. The results also show that, as the liquid viscosity increases or the drop mass decreases, oscillation damping is accelerated. The drop spreading and oscillation dynamics revealed here are important for inkjet printing, spray deposition, and other solution-based additive fabrication techniques.

CHAPTER 4 : COALESCENCE, EVAPORATION AND PARTICLE DEPOSITION OF CONSECUTIVELY PRINTED COLLOIDAL DROPS 4.1 Introduction

Deposition morphology of colloidal drops consisting of solution-processed functional materials is crucial in applications such as inkjet and gravure printing, and spray deposition of printable electronics, photovoltaics, and micro-batteries55-58. With the development of vastly available organic and inorganic ink materials consisting of nano-metallic particles, semiconductor quantum dots and nanowires, conducting polymers, among others, of different shapes and functionalities7, 59, 60, the challenge now is to effectively assemble these nano-scale building blocks into useful meso-scale structures61. Particle assembly by capillary interactions driven by the decrease in liquid-vapor (L-V) surface area as the particles approach each other has become an area of active research62, 63. Deposition of evaporating colloidal drops consisting of monodispersed spherical particles leads to the well-known coffee-ring effect as a result of the evaporatively-driven flow that carries particles to deposit in the vicinity of the drop contact line27,

29. The suppression of the coffee ring by shape-dependent capillary interactions has also been observed for deposition of colloidal drops consisting of ellipsoidal particles33. Understanding particle deposition from a single colloidal drop is important. However, drops are often used as building blocks for line and pattern printing where the interaction between drops plays an important role in determining the morphology of deposited functional materials.

A number of studies have been carried out to examine the dynamics of sessile drop coalescence by putting two drops of pure liquid slowly in contact with each other32, 64-67. The process of drop coalescence can be distinguished into three physical stages: (i) the initial stage where the edges of two droplets make contact and quickly form a thin liquid bridge which increases in width following a temporal power law; (ii) the intermediate stage where the

47 meniscus bridge relaxes, the contact line of the droplets begins to move (i.e., wetting on one side while dewetting on the other), and the curvature of the drop surface above the initial contact point changes from concave to convex; and (iii) the final stage where the combined drop relaxed toward a spherical cap, the minimum surface energy configuration. The dynamics of time- dependent width of the meniscus bridge between the two merging droplets32, 64, 68, and the perimeter67 and shape69 of the merged drop have been studied as a function of the liquid viscosity, surface tension, and density as well as the substrate wettability.

Drops impact on a dry or a wet surface can result in different morphologies, from spreading and splashing to receding and rebounding, depending on the impact velocity, drop size, properties of the liquid and substrate14. Spreading or "deposition" is observed for the case of a low-velocity impact. The impact and coalescence of a pure liquid drop onto a pre-deposited drop of the same liquid have been studied with a varying offset between the centers of two drops16, 70,

71. It was found that, for the case of no offset and in the low Weber number regime

2 ( We  D0U0  , where  is density, D0 the diameter, U 0 the velocity of the droplet before impact, and  the surface tension), axisymmetric impact and spreading lead to a merged drop of a spherical cap shape following the kinematic, spreading, and relaxation phases16.

The discussion until now is based on pure liquid drops where evaporation is negligible.

Recently, the deposition morphology of inkjet-printed solution or suspension into line patterns has been explored. For example, Duineveld72 studied the stability of inkjet-printed lines consisting of PEDOT:PSS, a conducting polymer, using a single-nozzle drop-on-demand inkjet printer. It was shown that the printed line becomes unstable and causes the bulging effect when its contact angle is larger than the static advancing contact angle. Stringer and Derby73 concluded that the stable line width of a printed line is bounded by the minimum line width determined by

48 the maximum drop spacing for stable coalescence and the upper bound determined by the minimum drop spacing below which a bulging instability occurs. Based on these analyses, appropriate printing parameters and ink/substrate properties can be selected to warrant optimal deposition uniformity and line definition. However, the studies of Duineveld72 and Stringer and

Derby73 are mostly concerned with the instability of the contact line. To the best of our knowledge, no information on the deposition dynamics of particles inside interacting colloidal drops is known.

In this paper, the interaction between two consecutively printed evaporating colloidal drops and the resulting particle deposition morphology are systematically studied as a function of the temporal delay and spatial spacing between drops. Pico-liter colloidal drops produced by an inkjet printer represent a common drop size widely used for high-resolution materials printing applications. In such length scales, capillary and evaporation are the dominating mechanisms, whereas gravitational effect is negligible. Three different regimes of drop spacing, d, are considered. The second drop is either printed (i) on the pre-wetted region of the substrate by the first drop (short drop spacing, i.e., 0 < d < 0.5 Dc, where Dc is the maximum spreading diameter of the first drop on the substrate), (ii) on a dry surface very close of the pinned contact line of the first drop (medium drop spacing, i.e., 0.5 Dc < d < 0.75 Dc), or (iii) on a dry surface away from the first drop so that the second drop is allowed to spread after the initial impact and then coalesces with the first drop (long drop spacing, i.e., 0.75 Dc < d < Dc). For the case of d > Dc, the interaction between drops is prohibited and the deposition of the second drop will be the same as the first drop. We focus on the regime where the drop jetting delay is comparable with or an order of magnitude smaller than the drop evaporation time scale. In addition, the relaxation time for two consecutively printed drops to completely coalesce in a saturated environment is on

49 the same order of magnitude as or slightly longer than the drop evaporation time. Due to the presence of colloidal particles and carrier liquid evaporation, drop coalescence is constrained. In such conditions, multiple mechanisms simultaneously determine the drop interaction and particle deposition. This makes a single model that correlates ink/substrate properties and jetting parameters with the final deposition morphology either impossible or very difficult. In the present study, by using a fluorescence microscope and a synchronized side-view camera, we directly observe in real time the interplay between the evaporation of the first drop, impact and spreading of the second drop, capillary relaxation, microflows, evaporation, and particle deposition of the merged drop for better control of the deposition morphology of inkjet-printed functional materials.

4.2 Experimental methods

To observe the particle movement and the evaporation of colloidal drops, a colloidal mixture of carboxylate-modiﬁed polystyrene ﬂuorescent beads (Invitrogen) of 1.1 µm diameter in DI-water, 0.2% by volume, is used in our inkjet printing experiments in Fig. 2.1. The mixture of water and particles is homogenized in sonicator (Cole-Parmer 8891) for 15 minutes before printing. The diameter of the drop is 68 μm and the velocity for the colloidal drop upon impact on cleaned glass substrates (Bellco, ~150 μm thick) is around 1.9 m/s. The corresponding Weber

number and Ohnesorge number ( Oh   D0 , where  is the liquid viscosity) are 2.1 and

0.014 respectively. The substrates are cleaned with a sequence of DI water, acetone, ethanol, and isopropanol for three times, and then, dried with compressed air. During the experiments, the ambient temperature and humidity are kept at 20˚C and 50% relative humidity. Image processing is performed in IMAGEJ (http://rsbweb.nih.gov/ij/).

Two colloidal drops with controlled jetting delay and spatial spacing are printed consecutively to observe interactions between them. The drop spacing is controlled by a high- resolution motorized microscope stage (Ludl) and the jetting delay between two drops is created by the JetDrive. For printing of colloidal drops, the jetting delays between two drops are 0.2 s,

0.6 s and 0.9 s, and the droop spacing varies from 38 μm to 105 μm. The Bond number

2 ( Bo  gD0  , where g is the gravitational acceleration) of our printed drops is less than 0.005 where the gravitational distortion of the drop shape is neglected.

4.3 Results and discussion

4.3.1 Drop coalescence

Side-view images for the coalescence of two consecutively printed evaporating colloidal drops on a cleaned glass substrate (Bellco, ~150 μm thick) are shown in Figure 4.1a, where the jetting delay between two drops is 0.6 s and the drop spacing is 73.4 μm (i.e., 0.7 Dc, where Dc is the maximum spreading diameter of the first drop on the substrate). Initial observation shows that it takes 1.3 s for a single drop to evaporate. The contact line of the first drop is pinned right after the drop impact and spreading. It is noted here that the static receding contact angle measured using a DI-water drop on a cleaned glass substrate is 23. However, the accumulation of particles along the drop edge prolongs the pinning of the contact line for the case of a colloidal drop. When a second drop impacts on the right side of the substrate (at t = 0 in Figure 4.1a) 0.6 s after the first drop is deposited, the inertia from its impact on the dry surface drives the second drop to get in contact with the first drop within 10 s. The merged drop continues to spread due to the interplay of inertia, viscous stress, and surface tension. The drop contact line becomes

51 pinned at t = 30 s and it is followed by both vertical and horizontal oscillations of the water-air interface. After the oscillations are damped by viscosity, the water-air interface exhibits an arc shape at t = 300 µs, where both left and right contact angles of the merged drop converge to a contact angle of 45. During this first 300 µs, the effect of evaporation, which has time scale on the order of seconds, can be neglected.

It is important to note that the relaxation time of the water-air interface of the merged drop is much shorter than that of a single drop impacting on a dry surface. As shown in Figure

4.1b, in the first 300 µs upon impact, the drop spreads on the dry surface and oscillates vertically and symmetrically where the contact angle fluctuates between 70 and 90. After 300 µs, the drop continues to spread and it takes over 1 ms for the single drop to relax to its equilibrium contact angle. It is shown in a comparison study of drop impacting on dry and wetted surfaces that both the impact morphology and spreading dynamics vary significantly with the type of surface74. Rioboo et al.74 also found that the drop spreading dynamics is less sensitive to the surface tension, impact velocity, and viscosity for a wetted surface as compared to a dry surface.

In the present study, however, the impact of the second drop is either on the edge of an evaporating drop or on a dry surface close to an evaporating drop, where the drop impact and spreading are expected to show combined characteristics of both dry and wetted surfaces.

(a) two consecutively printed drops

(b) a single drop

Andrieu et al.67 suggested that the rate of relaxation of slightly deformed drop to a spherical cap can be distinguished into three characteristic time scales: the inviscid inertial time

(te ~ µs), viscous inertial time (ti ~ ms) and capillary relaxation time (tb ~ µs). The inviscid

inertial time, te , during which the liquid is compressed and a shock wave is formed can be given by

ρR3 t = (4-1) e σ where R is the radius of the drop in flight. The viscosity then damps the oscillation on a time scale of

R 2 t = (4-2) i ν where ν is the kinematic viscosity. Finally, the viscous relaxation driven by the surface tension is

νρR t = (4-3) b σ

All three time scales are much smaller than the evaporative time scale (on the order of a second)

2 and are comparable to or smaller than the steady state diffusion time ( ~ ms), td  4R D . For a drop of 68 μm in diameter, the total relaxation time (te+ ti + tb) is about 1.2 ms. When the second drop impacts on the substrate 0.2, 0.6, or 0.9 s after the first drop is deposited, the first drop has already equilibrated and can act as a buffer to effectively damp the oscillation resulted from the impact of the second drop (e.g., the relaxation time of the merged drop reduces to 300 s from over 1 ms for a single drop impacting on a dry surface as discussed earlier).

Andrieu et al.67 also shows, for coalescence of two water drops of a similar size as the ones considered here (~50 μm in diameter), the relaxation time of the merged drop in a nitrogen environment saturated with water vapor is on the order of 5 s, 107 larger than the bulk capillary

relaxation time, tb . In such a case, the dynamics of drop coalescence is controlled by the receding motion of the contact line. The dissipation due to the gas-liquid phase transition near the contact line introduces a very small K ~ 10-8 Arrhenius factor in the relation between the driving force and the contact line velocity. In the present experiment, two drops coalesce at 20˚C and 50% relative humidity, where the merged drop evaporates in less than 3 seconds. The

54 subsequent particle accumulation at the contact line due to the evaporatively-driven flow enhances contact line pinning, and hence leads to incomplete drop coalescence before the carrier liquid completely evaporates.

Although the water-air interface appears to reach equilibrium at t = 300 µs from the side- view image shown in Figure 4.1a, the merged drop is still not completely coalesced, indicated by the shape of its footprint observed from the bottom-view image. Figure 4.2 shows schematic illustrations from both the side and bottom views of the drop impact, coalescence, carrier liquid evaporation, and particle deposition processes of two consecutively printed evaporating colloidal drops. As it can be seen, before the impact of the second drop, the contact line of the first drop is pinned and the evaporatively-driven flow drives particles to deposit near the contact line region and further enhances pinning. As the second drop lands on one side of the first drop, the inertia drives liquid to merge into the first drop. As the merged drop spreads and the water-air interface vibrates, it creates variations in local contact angle, relaxation velocity, and local radius of curvature. These local variations drive liquid to flow back and forth between two drops and bring particles along with it. Once the water-air interface reaches equilibrium, the curvature difference between the first and second drops drives the liquid to flow into the first drop side due to the local pressure difference p  p  p , where p is the pressure on the first drop side and p d1 d 2 d1 d 2 is the pressure on the second drop side and are both determined by averaging the local capillary pressure

1 1 p(x, y)   (  ) (4-4) R1 R2

Here R1 and R2 are the principal radii of curvature at any point (x, y) on the water-air interface.

Because the second drop only spreads partially before it merges with the first drop whose contact

55 line is pinned, the mean radius of curvature of the air-water interface on the second drop side is smaller than that of the first drop side. This curvature variation along the contact line leads to a flow inside the merged drop from the right to the left as shown in Figure 4.2e. When the merged drop continues to coalesce and evaporate, its footprint relaxes towards a circular shape while the evaporatively-driven flow brings particles to the contact line region and deposit there.

Side view Bottom view Figure 4.2 Schematic of drop impact, coalescence, carrier liquid evaporation, and particle deposition of two consecutively printed evaporating colloidal drops from side view (left) and bottom view (right) based on the medium drop spacing (0.5 Dc < d < 0.75 Dc) and long jetting delay (0.9 s) case.

4.3.2 Drop evaporation and deposition morphology

Figure 4.3 (a)-(c) shows evolution snapshots of two consecutively printed colloidal drops with a 0.9s delay time and short (0.4 Dc), medium (0.7 Dc), and long drop spacings (0.95 Dc), respectively, where Dc is the maximum spreading diameter of the first drop on the substrate. For the case with 0.4 Dc of spacing between two drops (Figure 4.3a), the second drop impacts on top of an evaporating drop at t = 0.9s and two drops merge immediately upon impact. It is shown from the side-view images that the water-air interface of the merged drop relaxes to an arc shape before t = 1.0 s (the third snapshot). Because the contact line of the first drop is pinned when the second drop impacts on it, the surface tension pulls the second drop to merge with the first drop and hence limits the spreading of the second drop. The merged drop shows an asymmetric footprint with a smaller radius of curvature on the second drop side and the curvature continuously varies along the perimeter of the merged drop at t = 1.0 s. In addition to the evaporatively-driven flow as observed in evaporation of single drops, the motion of the particles from the second drop to the first drop is observed due to the pressure imbalance that pushes the liquid to flow from the second drop (smaller radius of curvature) to the first drop side (larger radius of curvature) and drags particles along with it. The contact line of the merged drop is pinned and the particles move to the edge of the drop that results in a coffee-ring deposit, with a larger particle concentration on the first drop side.

As the spacing between two drops increases to 0.7 Dc (Figure 4.3b), the second drop impacts on a dry surface and spreads symmetrically before it gets in contact with the first drop.

The water-air interface of the merged drop relaxes to an arc shape and the footprint of the merged drop displays two deposition centers with a neck region connecting in the middle. The coffee-ring deposit that already formed along the pinned contact line of the first drop before it

57 coalesces with the second drop remains stable, which leaves a higher particle density region in the middle of the two drop centers.

(a) (b) (c) Figure 4.3 Snapshots of drop coalescence and particle deposition as a function of the scaled drop spacing for jetting delay of 0.9 s. (a) Short drop spacing (0.4 Dc), (b) medium drop spacing (0.7 Dc), and (c) long drop spacing (0.95 Dc). Drop spacing is defined as the distance between two consecutively printed drops and is scaled by the maximum wetting diameter of the first drop on the substrate (Dc).

When the spacing between two drops reaches 0.95 Dc, the second drop is allowed to spread more until its wetting front approaches the pinned contact line of the first drop. Once the merged drop coalesces, the imbalance in capillary pressure inside the drop drives the liquid to flow from the second to the first drop, resulting in local depinning of the contact line in the second drop side. The final deposition presents a higher particle density region in the original

58 first drop region compared to the second drop. As the separation distance between two drops increases, the area and length of the deposition increase.

Figure 4.4 shows evolution snapshots of two consecutively printed colloidal drops with similar separation distances, but with delay times of 0.2, 0.6, and 0.9s, respectively. The separation distance in all three cases are in the medium (0.6-0.7 Dc) separation regime where the second drop impacts on a dry surface and spreads briefly but symmetrically before it gets in contact with the first drop. For the case of a short jetting delay (Figure 4.4a), only a small amount of liquid in the first drop evaporates (the evaporation time for a single drop is 1.3 s) before it coalesces with the second drop. The two drops quickly merge and spread into a single drop of an elliptical footprint and the deposition is symmetric on both the first and second drop sides. No accumulation of particles along the contact line of the first drop is detected in the final deposition. The maximum dimension of the deposit perpendicular to the axis that connects the centers of the two drops (i.e., the major axis) is larger than the diameter of a single drop.

As the jetting delay between two drops increases, the effective total volume of the merged drop decreases due to the increase in evaporation time of the first drop. For the case of

0.6 s delay (Figure 4.4b), particle accumulation at the contact line of the first drop occurs before the second drop coalesces with it. The pinned contact line on the first drop side prevents spreading of the merged drop in the direction perpendicular to the major axis. Liquid flows from the second to the first drop due to curvature variation along the drop perimeter. The final deposition exhibits a smoothly connected pattern with a smaller radius of curvature on the

60 second drop side. When the jetting delay reaches 0.9 s, a coffee-ring pattern is already established in the first drop before the deposition of the second drop. The presence of the coffee ring from the first drop increases the non-uniformity of the deposit as compared with the cases of shorter jetting delay.

To quantify the particle distribution inside the merged drop, the scaled particle number density along the major axis of the deposit from two consecutively printed drops for the medium drop spacing cases (i.e., 0.5 Dc < d < 0.75 Dc where the second drop impacts on a dry surface but spreading is limited due to the presence of the first drop) and jetting delays of 0.2, 0.6, and 0.9 s is plotted in Figure 4.5. The particle number density is measured in a 36 μm wide stripe area along the major axis, x, of the deposit (where x = 0 denotes the edge of the first drop) and each

 grid is 18 μm long. The scaled particle number density is defined as n  Adropn p Agrid ntotal where

Agrid is the area of each 36 μm  18 μm grid along the major axis, Adrop is the area enclosed by

the drop contact line inside each grid, ntotal is the total number of particles inside the 36 μm stripe

area along the major axis, and n p is the number of particles inside each grid. The first data point of each case starts in the center of the first 36 μm  18 μm grid along the major axis, not at the exact edge of the deposition (x = 0). It can be observed for all three cases that the scaled particle number density peaks at both ends of the major axis, an indicator of the coffee-ring effect. In addition, the peak on the first drop side is higher than the second drop side as a result of the capillary flow due to curvature variation along the drop perimeter. For the case of a large jetting delay between two drops (i.e., 0.9 s delay), a third peak is observed in the middle of two drops where the pinned contact line of the first drop is located before the impact of the second drop

(i.e., x = Dc). However, for the cases of small jetting delays (0.2 and 0.6 s delays), the particle

61 number density in the center of the merged drop is fairly uniform. This implies that the deposition uniformity can be improved by increasing jetting frequency. It is also noted that the scaled number density is based on the average of three jetting experiments for each jetting delay and the standard divisions of all cases of the same jetting conditions are within 0.03 Dc. It might

 be argued that there is also a small increase in n for the case of 0.2 s jetting delay at x  0.6 Dc.

However, compared to a 31% increase in for the 3rd peak of the 0.9 s delay case at x  0.6 Dc based on the mean particle density along the major axis, this 11% increase in is not substantial enough to be viewed as a peak. It is important to note that, the long jetting delay of

0.9 s considered here is comparable to the 1.3 s evaporation time of a single drop.

third peak n*

Scaled major axis (x/Dc) Figure 4.5 Scaled particle number density along the major axis of the deposit for jetting delays of 0.2, 0.6, and 0.9 s and medium separation distance between drops.

 Scaled radii of curvature, r  r / Rc ( Rc  Dc 2 ), on the second drop side of the merged drop as a function of the scaled drop spacing for 0.2, 0.6, and 0.9 s jetting delays are shown in

Figure 4.6. The radii of curvature are measured based on the bottom-view images right after two consecutively printed colloidal drops coalesce. It is shown in Figure 4.6 that, for the delay time of 0.2 s, the scaled radius of curvature on the second drop side decreases monotonically with the scaled drop spacing. This is because, for a short jetting delay, two drops completely coalesce and a larger drop spacing leads to a more elongated shape for the merged drop. For the cases of longer jetting delays (0.6 and 0.9s), the radius of curvature on the second drop side decreases

 first as the merged drop becomes more elongated, and then r increases with the increase in drop spacing where the transition occurs at around 0.7 Dc, the drop distance where the curvature of the neck region between two drops changes from convex to concave. This indicates that two deposition centers are observed with a neck region connecting them for the case of large drop spacing and long jetting delay. In such cases, the second drop is allowed to spread more before it coalesces with the first drop, and hence increases with the increase in drop spacing. For the same drop spacing, a shorter jetting delay corresponds to a larger . Because a shorter jetting delay leads to a larger drop volume, the merged drop can further relax to a circular footprint before the carrier liquid completely evaporates.

Figure 4.6 Scaled radius of curvature on the second drop side of the merged drop as a function of the scaled drop spacing.

Figure 4.7 shows the circularity of the coalesced drop as a function of the scaled drop spacing for jetting delays of 0.2, 0.6, and 0.9 s. The circularity is measured based on the first bottom-view image after two drops coalesce, using circularity = 4  footprint area / (contact line perimeter)2. As it can be observed, higher circularities are obtained for shorter jetting delays compared with longer delays for all short, medium, and long drop spacing cases. This is because, for the same drop spacing, a short jetting delay allows for more liquid remained in the merged drop so that the neck region connecting two drops can spread more extensively to yield a more circular footprint that minimizes the surface energy. For the same jetting delay (which implies the same drop volume), the circularity of the merged drop decreases with the increase in drop spacing. This is because, especially for smaller jetting delays, the footprint of a merged drop exhibits an elliptical shape. If an equilibrium contact angle is assumed when the merged drop

64 fully relaxes, the height of the elliptical cap formed by the merged drop is the same for different drop spacing cases. This implies that the product of the major and minor axes of ellipse keeps constant for different drop spacings. When the drop spacing increases, the increase in the ratio between the major and minor axes results in a decrease in circularity of the merged drop as shown in Figure 4.7. For all three jetting delays considered in this study, the circularities are larger than 0.92 for the short drop spacing regime (0 < d < 0.5 Dc). However, with the increase in drop spacing for the cases of long jetting delays (0.6 and 0.9 s delays), as is shown in Figure 4.3c, a clear neck region is observed between the first and second drops and hence results in poor deposition circularity.

Figure 4.7 Circularity of the coalesced drop as a function of the scaled drop spacing d Dc .

4.4 Conclusions

The dynamics of drop interaction and particle deposition of two consecutively printed evaporating colloidal drops are observed in real time using a fluorescence microscope and a synchronized side-view camera. The second drop is printed either on pre-wetted region of the first drop (short drop spacing, i.e., 0 < d < 0.5 Dc), on a dry surface very close of the pinned contact line of the first drop (medium drop spacing, i.e., 0.5 Dc < d < 0.75 Dc), or on a dry surface away from the first drop so that the second drop is allowed to spread and then coalesces with the first drop (long drop spacing, i.e., 0.75 < d < Dc). The jetting delay considered here is either comparable with or an order of magnitude smaller than the drop evaporation time. The results show that the relaxation time of the water-air interface of the merged drop is shorter than that of a single drop impacting on a dry surface. As the drop spacing increases, the circularity of the coalesced drop decreases for all jetting delays. Moreover, for the drops interacting with a longer jetting delay, a coffee ring has already established before the impact of the second drop.

As the second drop impacts on the substrate and coalesces with the first drop, the capillary force and inertia drive the carrier liquid and suspended particles in the second drop toward the first drop resulting in more particles deposited on the first drop side. The particle number density along the major axis of the deposit shows two peaks at the two ends of the major axis for the cases with shorter jetting delays, but a third peak located at the contact line of the first drop is observed for a longer jetting delay. For a short jetting delay, the scaled radius of curvature on the second drop side of the deposit decreases monotonically with the scaled drop spacing, but it decreases first and then increases with the increase in drop spacing for larger delays.

In contrast to other drop coalescence studies, inertia, evaporation, and particle deposition all play important roles in the deposition morphology of the merged drop for the conditions

66 considered here. In addition, the present study reveals the deposition dynamics of individual particles in coalescing colloidal drops in real time when the drop contact line is pinned. By tuning the interplay of drop spreading, evaporation, capillary relaxation, and particle assembly processes, the deposition morphology of two consecutively printed colloidal drops can hence be controlled.

CHAPTER 5 : COMPETITION BETWEEN RECEDING CONTACT LINE AND PARTICLE DEPOSITION IN DRYING COLLOIDAL DROP 5.1 Introduction

Deposition morphology of colloidal drops containing solution-processed functional nanomaterials is crucial in applications such as inkjet and gravure printing, spray deposition of printable electronics, photovoltaics, and micro-batteries55-58. When a colloidal drop with a pinned contact line dries on a substrate, the solvent flow induced by the high evaporation rate at the contact line carries particles to the drop edge, forming the well-known coffee-ring deposit27. This evaporatively-driven, non-uniform distribution of functional materials in the deposit challenges the quality of printable electronics that often require uniform deposition patterns to achieve better performance. Several approaches have been introduced to suppress the coffee-ring effect by introducing bi-solvents of different surface tensions and vapor pressures to control the

Marangoni flow inside the drop31, 75 or by enhancing capillary interactions among asymmetrically-shaped particles (e.g., ellipsoids) on the drop surface33.

Contact line depinning can also occur during colloidal droplet evaporation, once the continuously decreasing contact angle reaches its threshold value for receding. The competition between the receding contact line and particle deposition near the contact line results in a variety of deposition patterns, from comb-like structures of stretched DNAs76 and multi-rings77-83, to spider web84, 85, and spokes29, 83, 86, 87. By moving a meniscus containing colloidal particles at different speeds, Prevo and Velev88 systematically examined the dependence of deposition morphologies (e.g. stripes, uniform layer, foam, and islands) on the contact line velocity in a substrate coating process.

Despite these studies on particle deposition morphologies, quantitative analysis to correlate the characteristic lengths of deposition patterns with droplet drying conditions and

68 colloidal suspension properties is still lacking. With the development of vastly available metallic, semiconductor, and dielectric nanoparticles, nanowires, and polymers as ink materials7, 59, 60, the challenge now is to effectively assemble these nanoscale building blocks into useful microstructures and devices61. Current materials printing techniques are limited to high-speed production of complex patterns with >10 µm resolution1, not compatible with applications that desire nanoscale features89. However, by carefully tuning the interplay of wetting, evaporation, and particle self-assembly, the formation of highly repeatable, ordered nanostructures inside inkjet-deposited colloidal drops can potentially be achieved.

In this paper, we report our systematic analysis of the complex morphology transitions inside inkjet-printed evaporating pico-liter droplets under different drying conditions, particle volume fractions, and particle sizes. As a result of the competition between contact line receding velocity and particle deposition growth rate, highly repeatable micron and sub-micron deposition structures such as multiple rings, spider web, radial spokes, saw-toothed structures, foams, and islands are observed. A dimensionless parameter that associates the radial deposition growth rate to contact line velocity has been identified to determine whether the entire contact line can be pinned to leave a continuous ring deposit. Both experimental measurement and model prediction of the multi-ring spacing as a function of local ring radius are presented and compared. The wavelengths of the spider web, radial spokes, and saw-toothed deposition structures are quantitatively analyzed using the Marangoni instability theory near the contact line. By varying the particle size while keeping other parameters the same, the effect of particle size on contact line pinning and subsequent deposition pattern is examined.

5.2 Experimental methods

Highly repeatable, pico-liter colloidal drops were generated by a piezoelectric inkjet printhead with an orifice diameter of 60 μm (MicroFab MJ-Al-01) driven by a waveform generator (JetDrive). A high-resolution (0.5 μm/pixel) camera system, consisting of a SensiCam

QE CCD camera (Romulus, Michigan), a Navitar 12 Zoom lens (Rochester, New York), and a halogen strobe light (Perkin Elmer), was used to capture side-view images of drop ejection.

Details of the experimental setup are described elsewhere28.

Colloidal suspensions of sulfate-modified polystyrene beads (Invitrogen) of 20 nm and

200nm in diameters with volume fractions of 0.1%, 0.25% and 0.5% in DI-water were used in the printing experiments. The mixtures of water and particles are homogenized in a sonicator

(Cole-Parmer 8891) for 5 minutes before printing.

The relative humidity (RH) levels were controlled at 20%, 40%, and 70% inside the custom-built humidity chamber and the ambient temperature was kept at 22 ºC for all experiments. Glass microscope cover slips (Bellco, ~150μm thick) were sequentially cleaned by sonicating in Sparkleen-DI water, ethanol and acetone, and then treated by Argon plasma (2 mins at 18W and 250 mtorr, Harrick Plasma PDC-32G). The root-mean-square (RMS) roughness of the plasma-treated glass slides measured by an atomic force microscope (Digital

Instruments/Veeco) is 0.5 nm, much smaller than the particle diameter of 20 and 200 nm.

Equilibrium contact angle of the cleaned glass substrates were measured to be close to 0º by using a custom-built goniometer and the LB-ADSA drop analysis plug-in of Image J

(http://rsbweb.nih.gov/ij/), where the drop surface is assumed to be a spherical cap and the

2 gravitational effect is neglected (the Bond number here is Bo  gD0 /  0.005 , where  is the

drop density, g the gravitational acceleration, D0 the drop diameter in flight, and  the drop surface tension).

5.3 Theoretical analysis

5.3.1 Contact line receding velocity of an evaporating drop

In colloidal drop drying processes, multi-ring depositions are formed due to the stick-slip motion of the contact line when it recedes as shown in Figure 5.1. This stick-slip motion can be described as a step function in the drop radius, R, vs evaporation time, t, plot (Figure 5.1b).

However, it is difficult to quantify the contact line velocity during each stick-slip motion, due to the repetitive acceleration and deceleration nature of the processes. Instead, we assume the contact line undergoes a continuous motion during receding, as shown in the dashed line of

Figure 5.1b, in order to correlate the contact line velocity with R.

For an inkjet-printed pico-liter drop ( Bo < 0.005), where the effect of gravity can be neglected, the drop shape is considered as a spherical cap with contact angle  and radius R shown in Figure 5.1a. When t  t0 , the volume of the drop can be written as

1 3 1 2 1 V (t0 )  R (t0 ) tan (t0 )[3  tan  (t0 )] (5-1) 6 2 2

On very hydrophilic substrates, the contact angle of the drop during receding approaches

2 zero and it follows tan [(t0 ) / 2]  3. Hence, the volume of the drop at can be simplified as

1 1 V (t )  R3 (t ) tan  (t ) for small (5-2) 0 2 0 2 0

Similarly, the volume of the drop at t  t0 t is given by

1 3 1 V (t0 )  R (t0  t) tan  (t0  t) for small  (5-3) 2 2

Utilizing Eqs. (3-13) and (3-14), it follows

d (t0 ) dV(t0 ) V (t0  t) V (t0 ) 1 dt 3 2 dR(t0 ) 1  lim  [ R (t0 )  3R (t0 ) tan  (t0 )] (5-4) dt t0 t 2 2 1 dt 2 2cos  (t0 ) 2

For an incompressible drop, dm  dV , where m and  are mass and density of a drop.

Hence, the drop contact line velocity at t  t0 is given by reorganizing Eq. (5-4)

d 0.5 dR 1 2 dm  (  dt R3) (5-5) 1 1 dt 3R2 tan   dt cos  2 2

For a very hydrophilic substrate, where the equilibrium contact angle of the drop is very

d small, the contact angle variation during drop receding is considered to be small, i.e., ~ 0 , dt which follows

dR 2 dm  ( ) (5-6) 1 dt 3R2 tan  dt 2

The volume variation of the drop due to solvent evaporation, dm/ dt , can be obtained by integrating the diffusion driven evaporation along the drop surface to yield26

dm  1.3RD(1 RH)c (5-7) dt v where D is the diffusivity of water vapor in air, RH is the ambient relative humidity, and cv is the saturated vapor concentration of the solvent.

Substituting Eq. (5-7) into Eq. (5-6), the instantaneous contact line velocity is given by

dR 1.3c D(1 RH) 1 U   v (5-8) CL 1 dt 3 tan  R l 2

Because the change in contact angle is trivial during drop receding on a very hydrophilic substrate, tan[(t) / 2] can be approximated as a constant and hence the contact line velocity inside a drying drop gives

UCL ~ 1/ R (5-9) which implies that the contact line velocity keeps increasing during drop receding.

5.3.2 Particle deposition rate at contact line

Based on Eq. (5-7), the volumetric solvent evaporation rate of the entire drop is

dV 1.3RD(1 RH)c   v (5-10) dt l

Because particles are accumulated and deposited along the drop edge, the volumetric particle deposition rate along the contact line gives 2RndU P or CP (dV / dt) /(1  ) , where d is the particle diameter, CP the particle volume fraction,  the deposition porosity, and n the

number of layers of particles in the deposition. The radial growth rate of particle deposition, Up , can hence be given by

1.3RD(1 RH)cv UP   CP (5-11) l

Experimental evidence shows that the particle concentration in the central region of the drop does not change during drying90. Here, we assume that the particle concentration, layer of the particles, and porosity in the deposition do not change during evaporation, and hence the radial growth rate of particle deposition, U P , is a constant during drop receding and is independent of the contact line radius, R , following

U P ~ const (5-12)

5.3.3 Ring spacing of the multi-ring structure

During drop receding, when the contact angle decreases below a critical value, the depinning forces surpass the pinning forces driving the contact line to depin from the original location ( R , ) and slip to a new location ( R R ,  ), where the contact line pins again and forms a new ring at R R , as shown in Figure 5.2. Here, R is the ring spacing at location R .

Assuming the drop is of spherical-cap shape both before and after slip and solvent evaporation is minimal during slip81, the volume of the drop can be written as

1 1 V  R3 tan  (5-13) 2 2 or

1 1 V   (R R)3 tan (  ) (5-14) 2 2

Hence, the ring spacing, R , can be expressed as a function of the instantaneous drop radius, R , as

R  CR (5-15)

1 1 1 1 tan   tan 2   tan   tan3  2 2 2 2 where C  (1 3 . Assuming  and   keep constant 1 1 tan   tan  2 2 at different ring positions79, 81, 91, the coefficient C becomes a constant during drop evaporation.

Hence, it follows

R ~ R (5-16)

Figure 5.2 Schematic illustration of the stick-slip motion of the contact line.

5.4 Results and discussion

5.4.1. Deposition morphology transition in droplets

Figure 5.3a-c shows the deposition morphologies of inkjet-printed drops on a cleaned glass substrate at 40% RH with 20 nm particles at volume fractions of 0.5%, 0.25% and 0.1%, respectively. It can be found that the deposition patterns transition from the edge to the center of the drop for all three cases, which are resulted from the competition between the receding contact line and particle deposition at the contact line. Both experimental evidence and models have shown that92, 93, for a pure liquid drop of spherical-cap shape evaporating on a hydrophilic substrate, the contact line velocity UCL continuously increases with the receding drop radius R.

During colloidal drop evaporation, particles inside the drop enhance the stick-slip motion of the contact line, i.e., the contact line radius undergoes a step function that decreases with time. In our study, assuming the particles do not alter the drying properties of the dilute drop of spherical-cap shape, the substrate is very hydrophilic, and the evaporation is diffusion-limited, the velocity of the receding contact line can be approximated as UCL ~1/ R(t) (see derivation of Eq. (5-9)), which implies that the contact line velocity gradually increases with the shrinking contact line.

Assuming the particle concentration in the remainder of the evaporating drop keeps constant

during drop evaporation, the radial deposition growth rate of particles, Up , is a constant (see Eqs.

(5-10) to (5-11)). Here, we introduce a dimensionless parameter,  Up UCL , to represent the competition between the particle deposition rate and contact line velocity. Figure 5.3d shows the linear increase of  with drop radius R from the drop center to the edge within a single deposit following   kR , using the case of particle volume fraction ϕ = 0.25% (Figure 5.3b) as an

76 example. Here, the slope k represents an interplay of the drop contact angle, the particle concentration and diameter, and the porosity and number of layers of the deposition.

Figure 5.3 Deposition morphologies inside colloidal drops dried under 40% relative humidity with 20 nm particles at different particle volume fractions: (a) 0.5%, (b) 0.25%, and (c) 0.1%. (d) deposition morphology transition in (b) as a function of the dimensionless parameter,  U /U , where R corresponds to the evaporating drop radius when    * , the critical P CL 1 value to warrant pinning of the entire contact line. Due to the difficulty in measuring the particle deposition growth rate at a sub-micron scale,  versus R shown in Fig. 1d is a schematic illustration based on the analysis. The scale bars are 20 µm in (a) – (c) and 2 µm in (d).

As shown in Figure 5.3b, for a drop with particle volume fraction of 0.25%, the initial pinning of the contact line induces the outmost ring-like deposit. As the drop continues to evaporate, the decreasing contact angle at the pinned contact line reaches the critical receding contact angle (a small but finite angle) so that the contact line depins from the outmost ring. At this time, the contact line speed UCL cannot keep up with the particle deposition growth rate Up , leading to the continuous accumulation of particles at the contact line region until they reach a critical number to pin the contact line again at a distance away from the outmost ring. A new ring is subsequently formed at this newly pinned contact line before another slip motion occurs as the drop evaporates29, 94-96. The repetitive stick-slip motion of the contact line drives nanoparticles to self-assemble into multi-ring structures for the regime of R  R1 in Figure 5.3d and near the drop

 edge in Figure 5.3b, where R1 corresponds to the radius when    , the critical value to warrant pinning of the entire contact line. For the case of  = 0.25%, R1 R0  0.77 , where R0 is the drop deposition radius. Note that the value of   is determined by the particle size, volume fraction, and wettability, solvent vapor pressure, relative humidity, and substrate wettability, among others. In this study, we focus on the effects of particle volume fraction, relative humidity and particle size while keeping other parameters unchanged.

 As the drop continues to recede, it yields   for the regime of R  R1 as shown in

Figure 5.3d. The contact line velocity keeps increasing but the particle deposition rate is not high enough to pin the entire contact line. This yields only parts of the contact line undergoing the stick-slip motion to form concentric rings, and other parts of the contact line are pinned to produce radial spokes, governed by the instability of the receding drop97, 98. The combination of stick-slip motion and instability near the contact line leads to the formation of the spider web

deposition at 0.77  R R  0.59 . Further receding of the contact line and the increasing of U 0 CL yield the deposition morphology transition into interconnected foam structures and finally isolated islands at the drop center where     , as shown in Figure 5.3d and Figure 5.3b.

Similar deposition morphology transitions are also observed for the particle volume fraction of 0.5%, as shown in Figure 5.3a. Due to the decrease of  during drop evaporation, the deposition morphology consecutively shows multi-ring, foam and islands from the drop edge to

the center where the transition radius R1 R0  0.29 . However, for the case of  = 0.1% (Figure

5.3c), the deposition morphologies are distinctly different from the ones with higher concentrations. Once the contact line depins from the outmost ring, due to the low particle volume fraction, the particles accumulated at the contact line region do not reach an enough number to form multi-ring structures. Instead, the particles accumulate and subsequently deposit periodically at the peaks of the sinusoidal contact line resulted from interfacial instability. The contact line depins locally while leaving periodic regions of the contact line pinned. As the drop evaporates, the magnitude of the finger-like instability grows and the continuously receding contact line pushes the particles to accumulate in the pinned regions and in turn enhances the local pinning of these regions. The radial spoke-like pattern of nanoparticles is hence formed near the edge of the deposit as shown in Figure 5.3c. As the contact line continues to recede, the deposition pattern transitions from spokes to foam and islands in the drop center. It is important to note that, the transition radius R R decreases with the particle volume fraction. 1 0

5.4.2 Interfacial instability near the drop contact line

When a droplet recedes on a very hydrophilic substrate, finger-like instability can be observed at the drop contact line97, where the thicker parts recede more slowly than the thinner parts leaving behind the radial deposition structures (spokes, spider web, saw-toothed structures) as shown in Figure 5.3. For diffusion-limit evaporation of a thin film on a solid substrate,

Marangoni instability arises as a result of the evaporation-induced thermal gradients98. For the case of an evaporating droplet, the wavelength of the instability follows

3 3   ~ 1/Ca  R/(j0) , where is the liquid viscosity,  the surface tension, R the drop

98 radius, and j0 the characteristic evaporation rate . Considering an aqueous drop evaporating on a very hydrophilic substrate with a contact angle θ ~ 0°, the characteristic evaporation rate can be

26, 97, 98 given by j0  0.65D(1- RH)cv / l , where D is the diffusivity of water vapor in air, RH the relative humidity, cv the saturated vapor concentration, and l water density. For the case

3 of constant RH, j0 is a constant, which follows the scaling law of  ~ R for Marangoni stress induced instability.

Figure 5.4 shows the experimentally measured dimensionless wavelength,  R0 , of instability induced radial structures (i.e., spacing of the peaks on saw-toothed structures in multi- ring deposits for  = 0.5% and grid spacing of the spider web shape for  = 0.25%) as a function

3 of R R0 inside single drops. For both particle volume fractions of 0.25% and 0.5%, the wavelength of radial structures increases from the drop center to the edge. Good linear fittings

3 of  R0 ~ R R0 for both volume fractions of 0.5% and 0.25% in Figure 3.20 indicate that instability, resulted from Marangoni stress in diffusion-limited evaporating droplets, is indeed

80 the driving force of particle deposition perpendicular to the contact line. It is important to note

3 that when 0.88  R R0  0.90 for volume fraction of  = 0.25%, the deposition pattern transitions from the spider-web shape to multi-ring structure, where the radial structures vanish.

Similar slopes of  R ~ 3 R R for particle volume fractions of  = 0.25% and 0.5% are due to 0 0 the fact that the same viscosity, surface tension, drop size, and substrate wettability are used in both experiments.

Figure 5.4 Wavelength of the saw-toothed structures in multi-ring deposits for  = 0.5% and the 3 grid spacing of spider web for  = 0.25% as a function of R R0 resulted from Marangoni instability within single drops containing 20 nm particles and RH = 40%. The scale bars are 20 µm. Error bars are standard deviations from 15 measurements.

Figure 5.5 shows the wavelength of radial spokes for colloidal drops dried at relative humidities of 20%, 40% and 70% for  = 0.1%. With the increase of humidity from 20% to 70%, the dimensionless wavelength of the spokes,  R0 , increases from 0.047 to 0.087. Recall that the

3 3 Marangoni instability near the contact line yields  ~ 1/Ca  R/(j0) and

26, 97, 98 j0 1.59D(1 RH)cv / l . A linear correlation relating the wavelength and relative humidity is hence given by  ~1 3 1 RH . As shown in Figure 5.5, good linearity between the

experimentally measured wavelengths of radial spokes  R0 (measured at R/R0 = 0.98) and

3 1 1 RH for relative humidities of 20%, 40% and 70% demonstrate is observed, consistent with the theoretical prediction.

Figure 5.5 Wavelength of the radial spokes as a function of for  = 0.1% for drops containing 20 nm particles evaporating at relative humidities of 20%, 40% and 70%. Error bars are standard deviations from 5 measurements. The scale bars are 20 µm.

Figure 5.6 shows the ring spacing of multi-ring structures, R , as a function of ring radius, R, for 20 nm particles with particle volume fractions of  = 0.25% and 0.5% evaporated under 40% relative humidity. For both particle volume fractions, the ring spacing increases from the drop center to the edge and the R versus R curves show very good linearity. This linear relationship between ring spacing and drop radius is consistent with the model prediction,

R ~ R , (see Eq. (5-16)) that is based on a drop of spherical-cap shape right before and after a small slip motion on a very hydrophilic substrate where evaporation is neglected during slip. As discussed earlier, this stick-slip motion of the contact line is an incremental step of the globally increasing contact line velocity as the drop recedes. Figure 5.6 also shows that, a larger particle volume fraction results in larger ring spacing at the same drop radius, and this is resulted from the enhanced pinning force (accumulation of more particles at the contact line) and depinning energy barrier needed for each ring deposition at a higher particle concentration94.

5.4.3 Effect of particle size

Deposition patterns of colloid drops containing 200 nm particles with volume fractions of

0.5%, 0.25% and 0.1% are shown in Figure 5.7a-c, respectively. Similarly to those containing

20nm particles, ring-like deposits are formed for the cases of ϕ = 0.5% and 0.25% as shown in

Figure 5.7a and 5.7b. As evaporation continues, the depinning forces overwhelm the pinning forces leading to contact line depinning from the outmost ring. In contrast to the repetitive stick- slip motion of the contact line leading to the formation of concentric rings for the cases of 20 nm particles (Figure 5.7d), the contact line is only partially pinned for drops containing 200 nm particles resulting in radial spoke structures as shown in Figure 5.7a-c. We have shown, by examining the forces acting on particles close to the contact line, that it is the number of particles not the particle size which dictates particle pining28. For drops of the same particle volume

84 fraction but different particle sizes, there are 1000 times more 20 nm particles in a drop than 200 nm particles. As a result, the number of 200 nm particles remaining in the drop after the deposition of the outmost ring is not enough to initiate entire contact line pinning (shown in

Figure 5.7e). The contact line is only pinned locally to form a series of radial spokes for drops containing 200 nm particles shown in Figure 5.7a-c. For the case of ϕ = 0.1%, the number of 200 nm particles is so small that even a single ring deposit cannot be formed and the radial spokes break into segments as shown in Figure 5.7c.

5.4.4 Deposition morphology map

The deposition morphology map of multi-ring, spider web, spoke, foam and island structures from inkjet-printed evaporating colloidal drops as a function of particle concentration and radial position inside the drop is summarized in Figure 5.8. Ring deposition occurs in the regime     while the contact line is only partially pinned when     . With the increase in particle loading, the region for ring-like deposition increases, while the region resulting in island- like deposition decreases. For a drop with a given particle loading, as the drop evaporates, the contact line velocity keeps increasing, leading to deposition morphology transitions from the top to the bottom of the morphology map shown in Figure 5.8. At a higher particle volume fraction

(e.g.,  = 0.25% for 20 nm particles) corresponding to a high deposition growth rate, the deposition consequently shows multi-ring, spider web, foam, and islands from the edge to the drop center. However, for a drop with a lower particle volume fraction (e.g.,  = 0.1% for 20 nm particles) or a low deposition growth rate, the deposition consequently shows spoke, foam, and islands.

Figure 5.8 Deposition morphology map of inkjet-printed colloidal drop with 20 nm particles, dried under 40% relative humidity showing multi-ring, spoke, spider web, foam and island patterns as a function of the radial position and the particle concentration. Ring deposition occurs for    * while the contact line is only partially pinned when   * . Error bars are standard deviations from 8 measurements.

5.5 Conclusion

In this paper, we showed that nanoparticles self-assemble into highly repeatable deposition patterns, including concentric multi-ring, radial spokes, spider web, foam, and islands inside inkjet-printed pico-liter colloidal drops. The mechanism governing these deposition morphology transitions in freely evaporating droplets is found to be a result of the competition between the contact line velocity and particle deposition growth rate during drop receding. A

variety of particle volume fractions and drying conditions are used to effectively control UCL and

Up , thereby determining the deposition patterns and the critical morphology transition parameter

* x to warrant pinning of the entire contact line. Experimental measurements agree well with the model prediction on the linear correlation between multi-ring spacing R and ring radius R.

Marangoni instability near the contact line attributes to the radial spoke and saw-toothed structures whose wavelengths  agree well with the scaling laws of  ~ 3 R and  ~1 3 1 RH .

In addition, larger particles are found to suppress the multi-ring deposition as compared to smaller particles.

Besides revealing evaporatively-driven nanoparticle self-assembly mechanisms inside a colloidal drop, this study offers a simple, fast way to produce complex patterns from an inkjet- printed single drop, and hence extends inkjet printing applications to fabricating patterns at submicron scales. Future work will focus on the effects of solvent type, as well as substrate and particle wettability, on the deposition morphology map.

CHAPTER 6 : SELF-ASSEMBLY OF NANOPARTICLES IN EVAPORATING PARTICLE LADEN-DEMULSION DROPS 6.1 Introduction

Due to unique and fascinating properties, nanoscale structures, such as metamaterials99,

100 and microreactors100, have received growing interests in applications of optics89, electronics101, biological material manipulation102, and energy harvesting103, 104.

Current methodologies used for nanoscale fabrication processes include nanosphere lithography105, 106, electron-beam lithography107, ion-beam lithography108, 109, nanoimprinting110, and dip-pen nanolithography111. However, each of these technologies has certain limitations restricting their widespread applications. In the nanosphere lithography, the space among a monolayer or multilayer of closely-packed particles is used as template for fabricating nanoscale features. However, the formation of monolayer of particles is usually lack of long–range order, due to the tendency to nucleate uncorrelated domains on the substrate, yielding the grain boundaries112. Electron-beam lithography, ion-beam lithography, nanoimprinting, and dip-pen nanolithougraphy are not suitable for high throughput manufacturing because of their low speed, small patterning area, and high-cost of equipment112. Therefore, a low-cost and high throughput technique for fabricating nanostructures over a large area is needed.

Inkjet printing, which is compatible for roll-to-roll manufacturing, has been proved to be a scalable and low-cost additive manufacturing technology. However, the current materials printing techniques are limited to high-speed production of complex patterns with >10 µm line width1, not compatible with applications desiring nanoscale features89. Hence, nano-structured templates or masks need to be used in conjunction with inkjet printing processes for directing the self-assembly of functional materials into nanostructures in the micron-sized droplets. The objective of this research is to explore inkjet printing as a potential technique for fabricating

89 nanoscale structures, and to advance the fundamental understanding of involved complex transport phenomena. Emulsion has been used for synthesizing homogeneous polymer particles113, Janus particles114, and nanorings115 in bulk solutions. In these systems, the oil drops in bulk water serve as templates for polymerization to fabricate or modify particles. By employing particle-laden emulsion in inkjet printing processes, the evaporative nano-scale oil droplets could serve as templates for particle self-assembly during the evaporation of the microscale inkjet-printed drops, enabling the high throughput of nano-features with inkjet printing technology.

In this study, the fabrication of nanomesh structures in evaporating particle-laden emulsion drops is explored. The evaporation process and self-assembly of the oil drops are observed in-situ. To reveal the mechanism governing the formation of nanomesh structures in inkjet-printed drops, the effects of surfactant, particle and oil concentrations, and drop drying conditions on the particle deposition morphologies are discussed.

6.2 Experimental method

The experimental procedures and mechanisms for particle self-assembly in particle- laden emulsion drops are shown in Figure 6.1. To form the particle-laden emulsion ink for printing processes, surfactant Sodium dodecyl sulfate (SDS) from Fisher Scientific are added into aqueous suspensions containing different volume fractions of 12 nm silica particles

(SNOWTEX-C from Nissan Chemical), followed by the addition of evaporative oil, mesitylene

(Sigma-Aldrich), as shown in Figure 6.1a. The mixture is stirred for 1 hour before ultrasonication under a probe sonicator (Sonics VCX 130) with 100% amplitude for 60 min (20 seconds pulse with 60 second idle). After ultrasonication, the oil is dispersed in water as droplets,

90 and the coalescence of these oil drops is prohibited by the surfactants attached on the oil-water interface. The hydrophilic silica nanoparticles (diameter = 12 nm) are suspended in the water phase rather than trapped by the interface due to their small size as shown in Figure 6.1b. This particle-laden emulsion is printed on substrates allowing for the evaporation of the water and the oil, as in Figure 6.1c and d. The substrates used in this study are glass microscope cover slips

(Bellco, ~150μm thick) underwent a sequence of cleaning by sonicating in Sparkleen-DI water solution, ethanol and acetone for fifteen minutes each, and immediately dried in a stream of compressed air after rinsed with DI water. The evaporation process is taken with an inverted

Zeiss Axio Observer microscope conjunct with a Sony XCL 5005CR CCD camera. The final depositions are examined under Scanning Electron Microscope (Zeiss Supra 50VP).

Figure 6.1 Experimental procedures for the fabrication of nanomesh in inkjet-printed particle- laden emulsion drops.

6.3 Results and discussion

The evaporation process of a particle-laden emulsion drop is shown in Figure 6.2. The drop evaporation initiates at t = 0s. Under the effect of contact angle hysteresis and stronger evaporation near the contact line, the oil drops move towards the drop edge until stopped by the wedge-like air-water interface at t = 0.08s. As the evaporation continues, the reducing contact angle leads to the contact line depinning and recedes inward. At the same time, this receding contact line pushes the oil drops moving towards the drop center and self-assemble into closely- packed oil drop clusters at t = 0.56s. The surfactants on the oil-water interface prevent the

92 coalescence of these oil drops allowing the silica nanoparticles to deposit among them. After complete evaporation of the water and oil drops, mesh-like deposition is observed on the glass substrate.

Figure 6.2 Evaporation of an inkjet-printed particle-laden emulsion drop (0.5% particle volume fraction, oil (mesitylene) : water = 1 : 10, SDS concentration = 0.3%, RH = 40%).

The formation of nanomeshes in inkjet-printed particle-laden emulsion drops is influenced by the ink recopies. The effect of particle concentration on the deposition morphologies is shown in Figure 6.3. At low particle concentration of 0.04%, during the drop evaporation, the number of particles is not enough to self-assemble into closely-packed structures, leading to the formation of irregular deposition in Figure 6.3a and b. By increasing the particle volume fraction to 0.5%, particles starts to assemble around the oil drops at the last stage of evaporation (t = 0.56s in Figure 6.2), and the mesh-like deposition is observed in the center of the drop. However, this nanomesh structure is not complete, because there is not enough number of particles around the oil drops to prevent the coalescence of these oil drops at the end of the evaporation, when the water almost dries out. By further increasing the particle concentration to

1%, at the last stage of evaporation, large amount of particles self-assemble in the space among the oil droplets. These deposited particles prevent the oil drop coalescence, leading to complete nanomesh deposition in the drop center. Each pore in Figure 6.3f was occupied by an oil drop before complete evaporation. The pore diameter of the fabricated nanomesh ranges from ~10 nm to ~3 µm, and line width of the mesh is ~ 200 nm.

From Figure 6.3f, it can be observed that larger pores occur in the center of the mesh, while smaller pores are formed at the edge. This phenomenon results from the evaporatively- driven self-assembly process of the oil drops, as shown in Figure 6.4. After a particle-laden emulsion drop is printed on a substrate, the contact angle hysteresis and stronger evaporation at the contact line initiate the outward flow which carries the silica nanoparticles and oil drops towards the drop contact line. Due to wedge shape near the drop edge, smaller oil drops are able to move closer to the contact line, but larger ones are stopped by the air-water interface at locations far away from the contact line. As evaporation continues, the drop recedes, and the drop edge remains a wedge shape trapping the smaller oil drops closer to the contact line than the larger ones throughout the receding process. At the last stage of evaporation, all the oil drops are

95 collected to the center of the drops with larger oil drops occupying the central area, and smaller oil drops at the edge. Silica nanoparticles deposit in the space among these oil drops. Therefore, the larger pores are observed in the center of the deposition, and smaller pores are observed at the edge.

Figure 6.4 Assembly of oil drops in an evaporating particle-laden emulsion drop.

As show in Figure 6.3e, the nanomesh structures only occupies a portion of the entire deposition (r/R = 30.0% in Figure 6.5a) due to the limited amount of oil drops. To study the effect of oil concentration on the deposition morphology, the volume ratio of oil/water is increased from 1/10 to 1/1 while keeping the particle concentration in water constant at 1%, as shown in Figure 6.5. When the oil-to-water ratio is 1/10, the complete mesh-like structure is formed with r/R = 30.0%. By increasing the oil-to-water ratio to 1/5, the larger mesh with r/R =

40.2% is observed due to the increased number of oil drops in the particle-laden emulsion, however, many defects (or breakup of the nanomesh) is observed. This may result from the lack of enough particles to prevent the coalescence of the increased number of oil drops in the emulsion in the last stage of evaporation. Hence, several locations devoid of mesh-like structures

(or local breakup of the nanomesh) are observed in Figure 6.5b. Further increasing the oil-to- water ratio to 1/1, no mesh-like structures could be observed in the deposition as in Figure 6.5c.

The greatly increased oil-to-water ratio drastically increases the instability of the particle-laden emulsions. During the evaporation, instead of self-assembly in the drop center, the oil drops coalesce in the center of the water drop, leaving no mesh-like structures in the final deposition.

In addition to the recipe of the particle-laden emulsion, the drying condition also influences the particle deposition, especially the central region, as in Figure 6.6. For evaporating particle-laden emulsion drops, after the water is completely evaporated, the surfactant would no longer be stable on the surfaces of the small oil drops, leading to the coalescence of the oil drops.

However, when the relative humidity is as low as 40% in Figure 6.6a, particle deposition speed is fast enough to deposit around the oil drops before their coalescence, hence the nanomesh could be observed in the deposition. By increasing the relative humidity to 70%, the water drop

97 evaporation is slowed down, and particle deposition speed reduces accordingly. Lower particle deposition speed cannot catch up with the oil drop coalescence process, hence, no mesh-like deposition could be observed after the evaporation.

6.4 Conclusions

In this study, we showed that the dispersed oil drops in a particle-laden emulsion drop could serve as liquid templates for printing nanoscale structures, which advances the printing resolution from ~10 µm to ~100 nm scale. The in-situ observation of the particle self-assembly into nanomesh structures in an evaporating drop is conducted with an inverted microscope. The ink recipe and drop drying conditions are found to greatly influence the formation of mesh-like structures. Increasing the particle concentration could limit the oil drop coalescence leading to

98 more complete nanomesh. Increasing the oil concentration in the emulsion is found to increase the area of the mesh, however, the integrity of the mesh decreases due to the shortage of the particles to prevent the coalescence of the increased oil drops. Moreover, the water drop evaporation and receding contact line confines the dispersed oil drop in certain locations according to their sizes, enabling the larger pores in the center and smaller pores close to the edge of the nanomesh. Therefore, by tuning the ink recipes and drying conditions for the inkjet printing process, the printable nanomesh structure could be controlled.

CHAPTER 7 : SELF-ASSEMBLY OF NANOPARTICLES DIRECTED BY NANOPOROUS TEMPLATES 7.1 Introduction

Nanostructures fabricated on substrates, such as nanopillar arrays116-118, carbon nanotubes119, 120, and metallic/semiconductor nanowires121-123 attract great research interests due to their unique properties including large surface area, a surface topography well-separated from the underlying substrate, and large mechanical compliance116. These advantages promote the application of nanomaterials in various fields, such as biomimetic materials124, 125, supercapacitors126, 127, superhydrophobic surfaces128, solar cells129, 130, biomedical diagnostics131,

DNA separation132, heat exchangers,133 and anisotropic surfaces134.

In many applications, the one-dimensional filamentary materials are completed in a wet environment135, 136. In this environment, liquid surface tension, originated from the decreasing air-water interface during the solvent evaporation, often results in the collapse of these high- aspect ratio materials118. Beyond this disadvantage, surface tension could also serve as a versatile and scalable tool for capillary-driven self-assembly of the patterned nanostructures into functional features119. Via this capillary-driven self-assembly, complex 3D microstructures are obtained with various nanoscale filamentary materials, such as carbon nanotubes137-143,

Polydimethylsiloxane (PDMS)144, Poly(methyl methacrylate) (PMMA)136, 145, 146, Polyethylene terephthalate147, epoxy116, 118, copolymers136, polyurethane118, photoresists135, 148, hydrogel118, 149,

150, biological fibers150, Si151-154, Au155, ZnO156, and Cu153.

One-dimensional filamentary materials could often be achieved by nanotemplate-assisted synthesis in which materials are deposited within the pores of the template membranes by electrochemical, chemical and physical processes. This method is proved to be capable for achieving extraordinary low dimension (~3 nm) structures.157 Among various nanoporous

100 templates, the anodic aluminum oxide (AAO) membrane shows unique characteristics in well- controlled pore size, size distribution, high repeatability, and good biocompatibility with typical thicknesses on the order of 100 nm to 100 μm, as shown in Figure 7.1. Self-assembly of metallic, semiconductor, dielectric nanoparticles, nanowires, and polymers in nanoporous materials, such as AAO, in conjunction with low-cost and high throughput fabrication processes carries great promise in additive manufacturing of nanomaterials.

In this study, methodologies of fabricating Ag nanotubes, nanotubes forest and nanomesh on solid substrates are developed based on the evaporatively-driven self-assembly of Ag nanoparticles in AAO membranes. The effects of particles concentration, substrates, fabrication and drying methods on the morphologies of fabricated nanostructures are examined.

Figure 7.1 SEM images of an Anodic aluminum oxide nanoporous membrane with 55 nm pore diameter. Top view (a) and cross-sectional view (b).

101

7.2 Experimental methods

AAO membranes with pore diameter of 55 nm, 80 nm, 100 nm, and 150 nm are purchased from Synkera Technological Inc, while the membranes with 200 nm pore diameter are

Whatman Anodisc Inorganic membranes. Silver ink containing ~10 nm Ag particles are from UT

Dots, and the inks with different concentrations are achieved by diluting the stock ink with the

UT Dots solvent. Kapton-FPC (Dupont), Kapton-HN (Dupont), and glass cover slides (Bellco,

~150μm thick) are used as substrates for growing Ag nanomesh and nanotube forests. The substrates and AAO membranes used in this study are rinsed with DI water, ethanol, isopropanol and acetone 3 times, and immediately dried with compressed air. The plasma treatment was done in an Argon environment in a plasma cleaner for 2 min at 18W and 250 mtorr (Harrick Plasma

PDC-32G).

7.3 Results and discussion

7.3.1 Effect of annealing conditions on electrical properties and particle deposition morphologies

Silver ink is widely used for printing electronics due to its low electrical conductivities.

However, the electrical properties of the structures fabricated with the silver ink are greatly influenced by the annealing conditions. To measure the electrical properties of the silver ink, 12

µL silver drops with 5.75% volume fractions are placed on glass cover slips, and dried in air, followed by thermal annealing at different temperatures, as shown in Figure 7.2.

102

As shown in Figure 7.2a, by increasing the annealing temperature from 50 °C to 175 °C, and keeping the annealing duration as 10 min, the sheet resistance of the silver deposition reduces from 3.4 × 1010 Ω/sq to 2.3 Ω/sq. This dramatic decrease in sheet resistance resulted from stronger inter-particle binding induced by the increased annealing temperatures as shown in

Figure 7.2c. For the particles deposited in AAO membrane, thermal annealing conditions also greatly influences the morphologies of the deposition as in Figure 7.3. A drop of silver ink containing 5.75% Ag particles is placed on an Argon plasma treated AAO membrane with 80 nm pore diameter. After the drop is dried in air, Ag particles self-assemble into closely-packed cylinders in the AAO membrane as show in Figure 7.3a. Due to the spherical shape of these nanoparticles, the space among the particles is uniformly distributed in the pores, and the pores of the AAO are completely filled. However, after sintering the membranes at 175 °C for 10 minutes, the silver nanoparticles start to melt and merge together, and the space among the particles aggregates into larger voids. Due to the instabilities, the melted Ag particles segregated

103 into chains of particles with much larger diameters than the nanoparticles in the ink as in Figure

7.3b.

Figure 7.3 Effect of annealing on Ag particle deposition morphologies in AAO membrane. (a) before sintering. (b) after sinterinf for 175 °C for 10 min.

7.3.2 Directed particle self-assembly into nanomesh

Due to the unique structure, the top and bottom surfaces for the AAO membranes are surfaces with periodic holes as shown in Figure 7.1. These periodic holes could serve as nanoscale templates for the solvent evaporation and particle self-assembly. The experimental processes are shown in Figure 7.4. A drop of Ag ink is sandwiched between the plasma-treated glass and AAO membrane. Under the effect of surface tension, the AAO membrane is tightly attached on the glass substrate. Then, the AAO and glass is flipped over immediately and suspended for drying in air. Due to the capillary effect, the silver ink infiltrates into the pores of the AAO, and the solvent evaporation only happens through the pores. As the evaporation continues, the liquid level finally reduces to be below the AAO membrane. As the liquid contact line is pinned on the AAO membrane, solvent evaporation induces meniscus beneath the pores as

104 shown in the locally magnified schematic in Figure 7.4. Further evaporation leads to the meniscus contact with the glass, and the particles are trapped in the solvent under the solid surface of the AAO membrane. After complete evaporation, the AAO membrane is detached from the glass substrates, and the particles are found to self-assemble on the glass into a pattern replicating the solid portion on the surface of the AAO membrane (nanomesh structures) as in

Figure 7.5b.

Figure 7.4 Self-assembly of particle into nanomehses.

Figure 7.5 shows the deposition of the Ag nanoparticles on the glass surface fabricated with the process described in Figure 7.4. Due to the defects in the experiments (evaporation did not completely occur through the pores), nanomesh does not form on the entire area of the glass

105 that covered by the AAO membrane as in Figure 7.5a. The substrate covered by nanomesh (as

Figure 7.5b) shows a blue color, and the part covered by the irregular deposition (as in figure

7.5c) is almost transparent. This might result from the surface plasmonic phenomenon. The nanomesh of the Ag particles in Figure 7.5b almost replicates the surface of the AAO membrane.

However, because the solid surface of the AAO membrane is not completely smooth, the pore sizes of the nanomesh are not exactly the same as that of the AAO membranes.

By adjusting the particle concentrations and the pore size of AAO membranes, the deposition morphologies of the nanomesh structures could be adjusted accordingly, as in Figure

7.6. In Figure 7.6a, with 3% Ag concentration, the silver particles assemble into nanomesh structures whose line width ranges from ~80 nm to ~400 nm and the pore diameters ranges from

~190 nm to ~ 400 nm. The deviation of the pore diameter of the nanomesh from that of the AAO membrane (55 nm) is induced by the non-uniform pore sizes and rough surface of the AAO

106 membrane as shown in Figure 7.1. When the ink concentration is reduced to 1%, pore size of the deposition increases to a range from ~ 83 nm to ~550 nm, and the line width reduces to a range between 83 nm and 283 nm as in Figure 7.6b. The effect of AAO geometry on the deposition is shown in Figure 7.6b and c. By keeping the particle concentration the same, and increasing the

AAO membrane pore size from ~55 nm to ~100 nm the deposition pore sizes and line widths increased to ~350 to ~800 nm and ~ 113 nm to ~700 nm, respectively. To improve the uniformity of the Ag nanomesh deposition, AAO membranes with more regular pores and smooth surfaces are needed. This could be achieved by electrochemically polishing the surfaces of the aluminum foils before anodizing the nanoporous aluminum oxide structures on them.

7.3.3 Fabrication of Ag nanotubes with AAO membrane

Besides the nanomesh-like structures, AAO membranes are also used for the fabrication of nanotubes via a drop casting process. In experiments, the cleaned and Argon plasma treated

107

AAO membranes with 200 nm are suspended in air by two glass substrates, and a drop of Ag ink is casted on top of the AAO membrane. Under the capillary effect, the Ag ink infiltrates into the pores of the AAO. During the ink evaporation in the nanopores, as show in Figure 7.7a, stronger evaporation happens near the contact line bringing the Ag nanoparticles to the contact line, and deposited on the wall of the pores. After the solvent being completely dried, the silver particles are deposited into a tube in the AAO pore. Thermal sintering at 120 °C for 60 min enhances binding among these Ag particles. By dissolving the particle loaded AAO membrane in 1 M

NaOH for 2 hours, followed by removing the extra NaOH with DI water using a centrifuge, the

Ag nanotubes with outer diameter about 200 nm are observed under SEM as in Figure 7.7b.

These cleaned nanotubes aggregate into black clusters, and precipitate from the DI water, or the stock solvent purchased from UT Dots. This may be induced by the fact that the surface functional groups/surfactants are burnt during sintering, or washed away in the cleaning procedure. To improve the stability of the fabricated nanotubes in solvents, surfactants with thiol

108 groups or diblock copolymers that could strongly bind with Ag could be used to stabilize the suspension.

7.3.4 Fabrication of Ag nanotube forest on substrates

In addition to the nanotubes dispersed in solvents, nanotubes vertically grown on solid surfaces also attracts great interest in applications such as supercapacitors126, 127, biomimetic adhesives124, 125, solar cells130, superhydrophobic surfaces128, heat exchangers133, biomedical diagnostics131, and anisotropic surfaces134. In this study, we are developing a fabrication methodology for growing Ag nanotube forests on solid surfaces via an evaporatively-driven particle self-assembly process. The fabrication process developed in this study is shown in

Figure 7.8.

109

Figure 7.8 Experimental procedures for fabricating Ag nanotubes on substrates.

The substrate for holding the Ag nanotubes and AAO membrane are firstly cleaned by ultrasonication in Sparkleen/DI-water solution, ethanol, and acetone for 15 min each, followed by rinsing with DI water and dried with nitrogen. After treated in Argon plasma for 2 min, drops of Ag ink with 5.75% volume fraction are casted on an AAO membrane suspended in air and on a substrate. After being dried, the AAO membrane and substrate with Ag particle deposition are tightly clamped together. The paste-like silver particles on the substrates performs as an adhesive binding to the solid substrate with the silver loaded AAO. Then, sintering the clamped AAO and substrate at 120 °C for 60 min under vacuum enhances the binding of the Ag particles, and

110 allows the rigid nanotubes to be attached on the solid substrate. To remove the AAO template, the substrates with AAO attached on top are immersed into 1 M NaOH solution for 2 hours. The

AAO reacts with NaOH forming the Al(OH)3. Because it is insoluble in water, large amounts of

Al(OH)3 are trapped among the Ag nanotubes. In order to remove the undesired Al(OH)3, the substrates are sequentially immersed in DI water, ethanol/water (50/50) and isopropanol/water

(50/50) solutions for 2 hours each. To dry these Ag nanotube forests, three different methods, drying in air, freeze-drying and supercritical drying are used in this study.

To identify the optimal processes and materials to fabricate Ag nanotube forest, 5 kinds of substrates (glass, silicon wafer, Kapton Tape (KT), Kapton FPC and Kapton HN films) and 3 kinds of AAO membranes (pore size 200 nm, 150 nm and 80 nm) are compared. Photographs of these samples after each fabrication procedure are recorded in Figure 7.9 to evaluate the effect of each cleaning procedure on the integrity of the Ag nanotube forest. The red boxes in the figures shows the procedure during which the layers of the Ag nanotube forest starts to crack or breakup.

In #1 to #5, silver ink is dried on AAO with 200 nm pore size and different substrates, and then sintered together at 120 °C. Due to the bad adhesion of the silver to the glass, Kapton tape and silicon wafer, most of the fabricated silver nanotube forest fell off from the substrates during the cleaning processes in #1 - 3. The Ag nanotube forest shows the best adhesion to the Kapton FPC film. The Kapton FPC film is also used to fabricate the nanotube forest with AAO pore sizes of

150 and 200 nm shown as #6 and 7 in Figure 7.9, which shows complete nanotube forests on the surfaces. In sample #8, a layer of silver is thermally evaporated on the Kapton FPC film before casting the silver drop, and all other procedures are kept the same. All the Ag nanotube forests fabricated on Kapton FPC film are complete from the photographs in Figure 7.9.

111

Upon reviewing all the processes described in Figure 7.9, Kapton FPC film and AAO membranes with 200 nm pore diameter are chosen to fabricate the Ag nanotube forests. After the fabrication and cleaning procedures in Figure 7.8 and 7.9, the samples are dried with three different methods described in Figure 7.8, and the mechanism of these drying methods are shown

112 in Figure 7.10. The most common drying procedure is to dry the samples in air which is shown as green arrow in Figure 7.10, where the solvent trapped in the Ag nanotube forest are removed through solvent vaporization. For the freeze-drying process, shown as the blue arrow in Figure

7.10, the liquid is frozen and kept in a vacuum environment allowing the solid state solvent sublimating into the vapor, which avoids the collapse of the high-aspect ratio nanotube forest induced by the liquid-air interface. The third method, shown as the red arrow in Figure 7.10, is in contrast with the freeze drying process. Instead of sublimation from solid state, the liquid that Ag nanotubes are immersed in are kept in a high pressure and temperature chamber enabling the transformation of liquid solvents directly into gas by passing through the supercritical region.

113

Figure 7.10 Three drying processes for Ag nanotubes. Blue: freeze-drying; green: dried in air; red: supercritical drying.

Figure 7.11 shows the Ag nanotube forest dried in air. After the cleaning process, the Ag nanotube forest is immersed in water/isopropanol mixture. Exposing the sample in air allows the evaporation of the solvent, leading to the constant reduction of the liquid. As the evaporation continues, the liquid film thickness becomes smaller than the height of the nanotube forest, and the liquid will be trapped in the space among the nanotubes. Due to the liquid meniscus as shown in Figure 7.11, one component of the liquid-air interfacial tension force perpendicular to the nanotube surface cannot be balanced. Resulting from their small diameter, these Ag nanotubes are not stiff enough to resist this surface tension force, and Ag nanotubes start to bend and self- assembled together as shown in Figure 7.11. Therefore, the nanotubes self-assembled into

114 closely-packed bundles at the top, while the bottom of the nanotubes is still attached on the substrate. The sizes of the clusters of these Ag nanotube forests depend on the nucleation process during the drying of the solvent thin film. To suppress the bending and self-assembly of the Ag nanotube, freeze-drying and supercritical drying processes are used to avoid the presence of the liquid meniscus during the solvent removal.

Figure 7.11 Ag nanotube forest dried in air. The nanotubes are bent and self-assembled due to the capillary force on the air-water interface when drying in air.

Tert-butanol is chosen as the phase change material for the freeze-drying process to avoid capillary assembly of the nanotubes based on three reasons144. (1) Tert-butanol does not dissolve

Ag nanotubes and the Kapton FPC film; (2) the volume change between the liquid and solid state

115 around melting of the material is small, approximately 0.8%; (3) the melting point of tert-butanol is 25.4°C, slightly above room temperature, so the sublimation could be done in a vacuum environment at room temperature. In the experiment, after cleaning the Ag nanotube forest in water/IPA solution, the samples are immersed in tert-butanol for solvent exchange for 2 hours, followed by transferring the Ag nanotube forest into a beaker with 10 mL fresh tert-butanol. This beaker with tert-butanol is quickly frozen in a refrigerator, and kept in vacuum for sublimation.

In the entire process, the samples are transported in a small cap among different solutions to ensure that samples are always immersed in liquids. However, during the sublimation, liquid is often found on the sample, and the capillary self-assembly of the nanotube forest is still observed under SEM as shown in Figure 7.12. This resulted from water (from incomplete solvent exchange) contaminating the tert-butanol during the frozen process. Any small amount of water/IPA mixed brought into the tert-butanol will greatly change its melting temperature, leading to the difficulty in keeping the sample frozen throughout the entire sublimation process.

Hence, the presence of small amount of water results in the self-assembly of the Ag nanotubes forest as shown in Figure 7.12.

116

Figure 7.12 SEM images of Ag nanotube forest dried through freeze-drying process. (a) Overview of the Ag nanotubes, (b) zoom in picture of the nanotube assembly.

The third method for drying the samples is supercritical drying to prohibit the self- assembly of the Ag nanotube forest, as shown in Figure 7.8 and 7.10. After cleaning in water/IPA mixture, the Ag nanotube forest is immersed in anhydrous ethanol for 24 hours to remove the water trapped among the nanotubes. Then, the samples are immersed in a small beaker filled with fresh ethanol, and kept in an enclosed chamber. This chamber is filled with liquid carbon dioxide at 150 bar for solvent exchange for 2 days. By increasing the temperature to 40 °C, above the critical point, liquid carbon dioxide is vaporized. After releasing carbon dioxide vapor from the chamber, dried Ag nanotube forest samples are obtained. Figure 7.13a,

7.3b and 7.13c show the SEM images of the Ag nanotube forest obtained through the supercritical drying process. Because of the removal of the surface tension effect, a large area of

Ag nanotubes standing on the substrate is obtained as in Figure 7.13c. It is important to note that immersing the samples in ethanol to remove the water before the supercritical drying is very important to prevent the capillary self-assembly of the Ag nanotube forest. Any water brought to the supercritical drier will be trapped among the nanotubes in the supercritical drying process,

117 since it cannot be dissolved in carbon dioxide. This water remains on the Ag nanotube forest, and the water-carbon dioxide interfacial tension will induce the nanotube forest self-assembly after the carbon dioxide is vaporized. The morphology of the Ag nanotube forest without water removal in anhydrous ethanol before supercritical drying is shown in Figure 7.13d. It can be observed that Ag nanotubes self-assemble into bundles under the presence of the water.

Figure 7.13 Ag nanotube forest dried under supercritical drying with and without water removal in anhydrous ethanol before kept in supercritical drier. (a) top and (b) cross-sectional views of Ag nanotube forest fabricated with water removal in anhydrous ethanol; (c) overview of Ag nanotubes in (a) and (b); (d) Ag nanotube forest fabricated without solvent exchange in ethanol before kept in supercritical drier.

118

7.4 Conclusion

In this study, we demonstrate the fabrication of nanostructures with AAO nanoporous templates. The effect of annealing temperature on the electric properties and morphologies of the

Ag nanoparticle deposition in nanoporous templates AAO is examined. High annealing temperature results in the melting and re-solidifying of the Ag particles leading to a series of Ag beads formed in the nanopores. In contrast, the low sintering temperature only partially melts and bonds the Ag nanoparticles together ensuring the formation of the Ag nanotubes in the nanopores of the AAO membrane.

Based on the evaporatively-driven particle self-assembly in nanoporous AAO membranes, the methods for fabricating Ag nanotube forests on solid surfaces are developed. Among all the substrates tested, Kapton FPC polyimide film shows the best adhesion to Ag nanoparticles and best resistance to all the chemicals involved. The cleaning and drying processes are developed to remove the AAO templates from the silver nanotube forest. Three different drying processes, drying in air, freeze-drying, and supercritical drying are compared, and Ag nanotube forest dried supercritical drying is found to be most efficient in suppressing the nanotube self-assembly under surface tension effect.

119

CHAPTER 8 : IMPACT OF DROPLETS ON LIQUID-INFUSED POROUS SURFACES 8.1 Introduction

Superhydrophobic surfaces which incorporate micro-/nano-scale structures have attracted extensive attention in the last decade, for their unique characteristics, such as self-cleaning158-160, drag reduction161, 162, and condensation heat transfer163-165. Although promising, these surfaces shows limited oleophobicity with high contact angle hysteresis166, failure under pressure167-169, inability to self-heal upon damage and high production cost168, 170, 171. In contrast to traditional superhydrophobic surfaces, liquid-infused surfaces, inspired by the Nepenthes pitcher plants172, are introduced recently by Wong et al. in Harvard102. Such liquid-infused surfaces show superior performance in self-cleaning, anti-icing, anti-biofouling, and condensation heat transfer applications than the traditional superhydrophobic surfaces without infused liquid173-183. It is found that the surface tension and viscosity of the infused oil directly affects the sessile water drop sitting on the oil-infused surfaces184, 185. Also, the stability of the infused-oil is influenced the surface structures. On a plane surface, gravity, thermal capillary forces, elasticity, chemical reactions with the substrate, or substances moving on the surface usually enables the formation of unique instabilities on the liquid thin film186-191. The 3D micro-/nano-scale structures on the surfaces introduces complexity, compartmentalization and micro-/nano-scale effects maintaining a continuous film170, 171, 192-195. To our knowledge, all current studies on liquid-infused surfaces focus on the equilibrium processes, where drops are placed without inertia on surfaces infused with a different liquid. These surfaces are coated with hydrophobic coatings to prevent the droplet penetrating the liquid film to directly contact with the surfaces. However, there’s no study of droplets impacting on liquid-infused surfaces, especially for the liquid-infused superhydrophilic surfaces which contains the same liquid as the droplet.

120

In this study, we experimentally investigated the droplet impact dynamics on hydrophilic surfaces infused with the same/different liquids, to develop the fundamental understandings over the interaction between the droplets and liquid-infused porous surfaces. We demonstrated that when an oil drop impacts on a surface infused with the same oil, there exists a small window of impact velocity for the droplet sliding down the surface without coalescence into the liquid thin film. In contrast, water drop slides down the oil-infused porous surfaces with different morphologies upon impact with different impact velocities.

8.2 Experimental methods

To observe the drop impact behaviors on liquid-infused porous substrates, an experimental setup is built as shown in Figure 8.1. Liquid drops with diameter of 3.3 mm in flight are released by a micro-pipette. By releasing the drops from different heights, the velocities of drops impacting on surfaces are controlled as 0.3 m/s, 1 m/s and 3m/s. The

1/2 1/4 Reynolds number, Weber number and splashing parameter Kd = We Re are shown in Table

8.1. The liquid-infused porous surface is set to be 45° inclined, and a high speed camera

(Phantom V210) working at 4,000 frame per a second and a light source (Fiber-Lite MI 150) are used to capture the drop impact processes.

121

Figure 8.1 Experimental setup for observation of drop impact on liquid-infused surfaces.

Substrates used in this study are anodic aluminum oxide (AAO) membranes with 200 nm pore size (Purchased from GE) and print paper (from OfficeMax). The AAO membrane is rinsed with ethanol, acetone, isopropanol and DI water three times, and dried on a hotplate (Ceramic

Thermo Scientific) followed by Argon plasma treatment for 2 min at 18W and 250 mtorr

(Harrick Plasma PDC-32G) before use in experiment. The paper is prepared by Argon plasma treatment for 2 min without any liquid rinsing. The liquids of drops used in this study are DI water and silicone oil with 100 cP viscosity and 21.25 mN/m surface tension (Sigma-Aldrich).

To prepare the oil-infused substrates, silicone oil is sprayed over the surfaces of porous AAO and paper, and the oil-infused substrates are held on a 45° inclined platform in air for 2 hours for equilibration. After obtaining the videos, image processing and measurement are performed in

IMAGEJ (http://rsbweb.nih.gov/ij/).

122

Table 8.1 Re, We, and splashing parameter for the drops impacting surfaces.

silicone oil water

Impact velocity We Re Kd We Re Kd 0.3 m/s 13.5 9.6 6.5 4.1 990.0 11.4 1 m/s 150.2 31.9 29.1 45.8 3300.0 51.3 3 m/s 1351.5 95.7 115.0 412.5 9900.0 202.6

8.3 Results and discussion

Images comparing the single silicone oil drops impacting on dry AAO and silicone oil- infused AAO are shown in Figure 8.2. For a silicone oil drop impacting on a dry AAO substrate with impact velocity 1 m/s, the inertia drives the droplet to spread on the surface without receding or splashing. Because the substrate is inclined to 45°, inertia drives liquid to flow downward, and start to accumulate at the drop front edge (the bottom edge of the drop in Figure

8.2a). At t = 3.75 to 7.5 ms, this liquid accumulation forms a ring-like shape, and forms a pressure gradient inside the drop, which pushes the liquid flowing back to the rear edge (top edge of the drop in Figure 8.2a) at t = 10 ms. Due to the high viscosity of the silicone oil (100 cP), the drop interface oscillates between the front and rear edges, resulting from the liquid motion inside the drop which is quickly damped down by viscous dissipation, and a sessile drop is formed on the inclined AAO surface. Then, the drop starts to spread under the effect of surface tension.

Once the AAO membrane is infused with silicone oil, the behavior upon the impact of a drop with the same liquid becomes unique. Instead of splashing or coalescence into the infused liquid film, the droplet spreads, recedes, and, at the same time, slides down the substrate, as shown in

Figure 8.2b. This is resulted from a layer of air is trapped between the drop and oil film upon the impact. This air layer prevents the drop-film coalescence. Under the effect of inertia and gravity,

123 the oil drop slightly spreads, recedes and slides down on the inclined surface infused with silicone oil.

Figure 8.2 Silicone oil drops impact on: (a) dry AAO, (b) silicone oil infused AAO. Impact velocity is 1 m/s.

Figure 8.3 shows silicone oil drops impacting on the AAO membrane infused with silicone oil at different velocities. In figure 8.3b, as discussed above, the silicone oil drop impacts, slightly spreads, recedes and slides down the oil-infused AAO membrane at medium velocity of 1 m/s. However, at low (0.3 m/s) and high (3 m/s) velocities, both drops are found to spread on the surface and coalesce into the oil infused in the nanoporous AAO.

124

Figure 8.3 Silicone oil drop impact on AAO membranes pre-suffused with silicone oil at (a) 0.3 m/s, (b) 1 m/s, (c) 3 m/s.

Figure 8.4 shows the time evolution of drop front edge, rear edge, length and width upon impact on an oil-infused AAO membrane with different impact velocities. In Figure 8.4a, positions of the drop front and rear edges on the surface are the distance from the drop edges to the position where drop touches the surface (dashed line). It can be observed, that for drop impact on the surface at 0.3 m/s, the drop slowly approaches the oil film allowing enough time for the air to be expelled out of the gap between the two liquid layers, hence, the drop merges into the film. Due to small inertia, the front edge of this drop expends slowly under the capillary effect until t = 27.5 ms when the drop edge becomes difficult to measure due to the coalescence.

Different from the drops with high and low impact velocity, where the front and rear edges are located at two different side of the impact point as in Figure 8.4a, for drop impact with medium velocity of 1 m/s, the positions for both the front and rear edges of the drop increases in the same direction, which means that the drop slides down the oil-infused surface upon impact, due to the air film trapped between the drop and liquid film. This drop, as in Figure 8.4b, firstly spread

125 under the effect of inertia from t = 2.5 to 6.75 ms, and then gradually recedes into a spherical cap while sliding down. The drop recedes along the length direction is slower than that in the width direction probably because the drop is sliding down the surface under inertia which stretched the drop in the length direction into an ellipsoid shape as t = 12.5 – 20 ms in Figure 8.3b. During this process, the rear edge remains pinned on the inclined surface. For the impact with high velocity

(3 m/s, We = 1351.5), the high inertia compresses the air gap the between the drop and liquid film, and finally this film breaks up leading to the coalescence. Due to high inertia, this drop spreads much faster (as in Figure 8.4a) than the drop with low impact velocity (We = 13.5), as show in Figure 8.4a. Therefore, the drop could not recede into a spherical cap and slide down the surface. Therefore, for drops impacting on a surface infused with the same liquid, high and low impact velocity prohibit the drop receding and promote the drop merging into the thin film on the surface, but there exists a small window that the drop can spread, recede and slide on the surfaces.

126

Besides the drop impact velocity, structures of the oil-infused surfaces also play important roles for drop dynamics. The process of a drop impact on oil-infused print paper with

3 m/s velocity is shown in Figure 8.5a. Upon impact, the drop firstly spreads in the length direction, and the liquid starts to accumulate at the front edge at t = 2.5 ms due to inertia. This accumulated liquid keeps moving forward forming a neck region between fast advancing liquid front and slow spreading drops at t = 2.5 - 5 ms. This neck region keeps receding. Finally, the liquid front is separated from the rear part sliding down the surface, and the rear part of the drop spreads and coalesces into the liquid thin film. This phenomenon is in contrast to the drop impact, spreading and coalescence on oil-infused AAO surface, as shown in Figure 8.3c. This results from the different substrates used in the experiments. The scanning electron microscopy images

127 of the dry AAO and print paper are shown in Figure 8.5b and 8.5c. The AAO membrane shows relatively regular pores on its surface with nominal diameter of 200 nm. However, the print paper is composed by a series of fibers, and the pore sizes are irregular.

The above discussion is based on silicone oil drops impacting on porous substrates infused with the same silicone oil, where the drop behavior (deposit or slide) on this substrate is dependent on different impact velocities. When different immiscible liquids are used for the droplets and the infused-oil, the drop behaviors become quite different. The water drop keeps

128 sliding along the substrates infused with silicone oil upon impact with all three different velocities, but the drop morphologies are very different as shown in Figure 8.6. For drops with

0.3 m/s, the droplet spreads to its maximum diameter at 7.5 ms, and then recedes as shown in

Figure 8.7b. Similar length and width during the spreading and receding in Figure 8.7b means that the drop slides down the surface as a spherical cap, which is different from the spreading and coalescence of oil drop upon impact on the silicone oil-infused surface in Figure 8.3a. When the impact velocity increases to 1 m/s, the higher inertia leads to a greater drop more in the length direction as in Figure 8.7b, forming an elliptical shape during sliding down the surface as t = 7.5 ms – 50 ms in Figure 8.6. Under the effect of surface tension, the drop finally recedes into a spherical cap shape (similar width and length) at t ≈ 100 ms and slides down the surface. Further increasing the impact velocity to 3 m/s, the morphology evolution of the water drop becomes distinctively different. Upon impact, the drop firstly spreads similarly both in width and length directions up to t = 1.5 ms. Then, both the drop front and rear edges start to move downward along the substrate as shown in Figure 8.7a. The drop front moves much faster than the drop rear edge leading to the constantly increasing in drop length in Figure 8.7b, and a flat elliptical drop could be observed on the surface as t = 5 ms – 10 ms in Figure 8.6. Because of the surface tension of the water drop, the drop recedes in the width direction after 3.25 ms, while it keeps elongating in the length direction. The water drop forms a column shape at t = 15 ms. A small drop ejects from the front edge of this column drop, and the larger part further break up into two drops. By comparing Figure 8.7 and 8.4, it can be observed, that the oil drop slides down the surface only at medium impact velocity, and the water drop slides down the surface at all three velocities.

129

Figure 8.6 Water drops impacting on the silicone oil-infused print paper with (a) 0.3 m/s, (b) 1 m/s, and (c) 3m/s.

130

8.4 Conclusion

In this study, droplet dynamics upon impacting on inclined porous surfaces with and without infused-liquid is examined. On dry nanoporous AAO, the oil droplet deposit and spread on the surface without receding at medium impacting velocity, but on the oil-infused AAO, the same drop undergoes spreading, receding, and simultaneously sliding down on the nanoporous surface. Increasing and decreasing the impact velocity leads to the drop spreading under inertia and capillarity respectively without receding, and the spread out drop finally coalesces into the liquid infused in the AAO membrane. During the spreading phase, the drop widths and lengths increase similarly for drops impacting with all three velocities. However, during the receding phase of the drop with the medium impact velocity, the drop width firstly recedes faster than the length, but finally the drop slides along the surface with similar width and length. Moreover, structures of the substrates supporting the oil film also play important roles in drop dynamics. In

131 contrast to the drop merging in oil-infused AAO with high impact velocity, the oil drop breakup into two parts (one merges in and the other one slides down on) the oil-infused paper surface.

Furthermore, the dynamic morphologies of water drops impacting on oil-infused print paper are observed. The drops show similar width and length during sliding upon impacting with medium and low velocities, while the drop sliding down as a column and breakup into three pieces in a line for impacting with high velocity.

132

CHAPTER 9 : CONCLUSIONS AND RECOMMENDATIONS 9.1 Conclusions

Inkjet printing, as a low cost and high throughput additive manufacturing technology, is considered to be capable of revolutionizing the electronics manufacturing industry. However, the current challenge is to effectively control the self-assembly of the functional materials into useful structures for various devices. Therefore, it is important to develop the fundamental understanding for the inkjet printing process, including drop formation, impact, spreading, coalescence, evaporation and functional material self-assembly. This study is aimed to reveal the fundamental transport phenomena in inkjet-printed droplets.

9.1.1 Coalescence and particle deposition of two consecutively printed drops

For inkjet-printed drops impacting on a dry and impermeable surface at We ~ O(1), both inertia and capillary effects play important roles in the initial spreading regime leading to substrate wettability dependent spreading kinetics before the droplet starts to oscillate. This is in contrast to the inertia-dominated drop impact at We >> 1, where variations in drop spreading kinetics with substrate wettability are negligible.

The oscillation behaviors of the single and two printed drops depending on surface wettability, liquid surface tension, viscosity are investigated with a high speed camera. Drop oscillation decays faster on surfaces with higher wettability due to its stronger viscous dissipation resulting from a longer liquid oscillation path parallel to the substrate. Drops with a high viscosity oscillate shorter in time than low viscosity ones. The liquid surface tension does influence the drop oscillation frequency but not the oscillation period. Similar results on the effect of wettability and viscosity on drop oscillation time are found for the combined drop

133 resulting from a drop impacting on a sessile drop of the same liquid, but oscillation duration for the combined drop is longer than the single drop. To analyze the experimental observation, a single degree of freedom model analyzing the drop oscillation frequency (period) and amplitude decaying is developed, which are in good agreement with the experimental results during the drop relaxation phase.

When a drop impacts on a sessile drop sitting on a hydrophobic substrate, as two drops coalesce and the combined drop subsequently spreads on the substrate, the drop recoil twice due to the coalescence of the two drops. This is in contrast to the absence of recoil for a single drop impact on a dry surface under the same condition, indicating that the threshold value for drop recoil is lowered in the presence of another sessile drop.

By using a fluorescence microscope, the particle deposition behavior in two consecutively printed colloidal drops is observed with different drop spacing and jetting frequencies. As the drop spacing increases, the circularity of the coalesced drop decreases for all jetting delays, but the radius of curvature of the contact line at the second drop side firstly decreases, then increases. Jetting delay is also found to influence the particle deposition morphologies. Decreasing jetting delay enhances drop spreading by allowing more liquid in the combined drop, leading to the higher circularity for the deposition. The particle number density along the major axis of the deposition shows two peaks at the two ends of the major axis for the cases with shorter jetting delays, but a third peak located at the contact line of the first drop is observed for a longer jetting delay.

134

9.1.2 Beyond the drop size limitation: complex submicron structures in inkjet-printed drops

Based on the classic coffee ring effect formed in colloidal drops, we demonstrate that nanoparticles self-assemble into highly repeatable complex morphologies, including concentric multi-ring, radial spokes, spider web, foam, and islands inside inkjet-printed pico-liter colloidal drops. The formation of these complex depositions is induced by the competition between the drop contact line receding velocity and particle deposition growth rate during the evaporation which is modeled in this study. A critical morphology transition parameters relating to the contact line receding velocity and particle deposition growth rate is found to warrant pinning of the entire contact line and resulting the deposition morphology transition inside droplets.

The functional relationship between the deposition geometries and the drop drying dynamics are established. Model prediction on the linear correlation between multi-ring spacing and ring radius R, based on contact line stick-slip motion, agrees well with the experimental measurements. Marangoni instability near the contact line attributes to the radial spoke and saw- toothed structures whose wavelengths  agree well with the scaling laws of l ~ 3 R and l ~1 3 1- RH . Particle sizes are also critical for the deposition morphologies. Larger particles are found to suppress the multi-ring deposition, as compared to smaller particles, resulted from the smaller number of particles inside the drop.

The dispersed nano-scale oil droplets in the particle-laden emulsion are found to serve as liquid templates for inkjet printing nano-scale structures, which enables the fabrication of nanomesh with ~ 100 nm line width much smaller than the 60 µm drop diameter. Through the in- situ observation, the water drop evaporation and receding confines the dispersed oil drop in certain locations enabling the large pores in the center and smaller pores close to the edge of the deposition. The emulsion recipe and drop drying conditions are found to be critical for the

135 integrity of the nanomesh structures. Increasing the particle concentration limits of the oil drop coalescence leads to more complete nanomesh. Increasing the oil concentration in the emulsion is found to increase the area of the mesh, however, the integrity of the mesh decreases due to the lack of the particles to prevent the oil drop coalescence. Therefore, by tuning the ink recipe and drying conditions for the inkjet printing process, the printable nanomesh structure could be achieved.

Besides revealing evaporatively-driven nanoparticle self-assembly mechanisms inside colloidal drops, this study offers an easy and fast way to produce complex submicron structures from inkjet-printed drops, which may extend the inkjet printing for fabricating devices at submicron scales.

9.1.3 Development and morphology control of Ag nanotube forest

Nanoporous anodic aluminum oxide membranes are used as templates for self-assembly of Ag nanoparticles into various structures. For colloidal drops evaporating on an AAO template, the liquid infiltration brings solvent and particles into the nanopore. Due to the shape of the meniscus inside a pore, the particles are transported to the contact line, depositing into nanotubes in the cylindrical nanopores. The influences of thermal annealing temperature on morphologies of Ag nanoparticles trapped in AAO membranes are examined. High annealing temperature melts the Ag nanoparticle deposition, and re-solidified into a series of Ag beads throughout the nanopores, but the low annealing temperature only partially melt the particles and crosslink the

Ag nanoparticles into nanotubes.

Based on the evaporatively-driven self-assembly of Ag nanoparticles in AAO membranes, the fabrication process for Ag nanotube forest on solid surfaces is developed. Kapton FPC

136 polyimide film shows best adhesion to the Ag particles, and the NaOH, ethanol and isopropanol are used to remove the AAO templates from the Ag nanotube forests. Three different drying methods, drying in air, freeze-drying and supercritical drying, are compared. For samples dried in air, the continuously reduction of solvents due to the evaporation enables the self-assembly of nanotube forest into closely-packed bundles under air-liquid interfacial tension. This capillary- driven self-assembly could be suppressed by supercritical drying process.

9.1.4 Droplet impacting on liquid-infused porous surfaces

The behaviors of droplets impacting on inclined porous surfaces infused with liquids at three different velocities are examined. In contrast to the spreading and deposition on dry AAO membranes, the oil droplet spreads, recedes, and simultaneously slides down on the oil infused

AAO upon impact with medium velocity. Impacting with higher or lower velocities, the drop spreads under inertia or capillary force, and finally merges into the liquid infused porous surface.

Moreover, the dynamics of water drops impacting on oil-infused print paper are observed to examine the dynamics of drops impacting on surfaces with immiscible liquid. The drops possess similar widths and lengths during sliding upon impacting with medium and low velocities, while the drop slides down as a column and breaks up into three pieces in a line when impacting with high velocity.

137

9.2 Recommendation for future work

9.2.1 Oil film assisted particle self-assembly in inkjet-printed drops

Coffee ring effect happens due to the contact line pinning and non-uniform evaporation along the air-drop interface27, and the contact line pinning originates from the contact angle hysteresis induced by the heterogeneities in the surface energy or surface profiles (roughness)196,

197. Therefore, this contact angle hysteresis could be suppressed if the drop is evaporating on an ideally smooth surface, such as a droplet floating on a liquid film, where the drop could freely recedes during the evaporation. Hence, the coffee ring effect could be suppressed.

The preliminary results can be seen in Figure 9.1. The evaporation process of a colloidal drop printed on a dry glass slide is shown in Figure 9.1a, where the contact line is pinned during the evaporation. The particles are transported from the drop center to the edge forming the coffee ring effect. However on a silicone oil coated glass slide, the drop keeps receding and collecting particles to the drop center as in Figure 9.1b. No particle motion from drop center to the edge is found during the evaporation process. At the end, a monolayer of particles is formed on the substrate. Although this is unique, there are still several questions to be addressed to fully understand the phenomena described in Figure 9.1b. i) When an aqueous colloidal drop is printed on oil film, whether the drop contacts with oil or substrate is not clear. ii) Whether the particles are kept in water or trapped at the oil-water interface during the evaporation is not directly observed, since the particles tends to be trapped at oil-water interface to minimize the free energy of the entire system. iii) Current studies on the water drops sitting on oil films are based on millimeter-sized sessile drops184, 198. Whether these theoretical laws are applicable to the dynamic processes, like inkjet printing (drop evaporates in seconds), still needs more study.

138

Figure 9.1 Evaporation of inkjet-printed colloidal drops on (a) a glass slide and (b) a silicone oil- coated a glass slide. The scale bars are 100 µm in (a) and 50 µm in (b). Particle diameter: 1 µm.

9.2.2 Printing of colloidal drops with ellipsoid particles for reducing the deposition area smaller than the drop size

Numerous studies on the evaporation of colloidal drops focus on the spherical particles27,

29, 199. Once the drop is printed on hydrophilic surfaces, the deposition morphologies of these spherical particles exhibit the well-known coffee ring effect. This uneven deposition distribution usually harms the performance of printed lines and films. However, in evaporating drops with anisotropic particles, these particles could firmly attached on the air-water interfaces and form monolayer of a closely-packed particle network under the strong and long range capillary attraction33, 34. This strongly intertwined particle network trapped on the air-water interface prevents the particles jamming at the drop contact line, and suppresses the coffee ring effect33.

However, the previous studies failed to examine the drop contact line dynamics during the

139 evaporation of colloidal drops containing anisotropic particles, which in theory should be very different from the drops containing spherical particles.

The fabrication of the ellipsoid particles is revised from the existing methods200-202. The carboxylate-modified polystyrene particles of 1 µm in diameter are dispersed in a 10% w/w solution of polyvinyl alcohol (PVA) in DI water, followed by solidifying this particle-laden PVA gel in air. The dried gels are immersed in 140 °C vegetable oil (higher than 110 °C of the particle glass transition temperature) and stretched into desired aspect ratio, so the particles trapped in the

PVA gels are elongated into ellipsoid particles. By dissolving the PVA gel in 50/50 w/w mixture of isopropanol and DI water, and removing the PVA and isopropanol, as large amount of ellipsoid particles dispersed in water is obtained. The SEM images of fabricated ellipsoid particles are shown in Figure 9.2. It is important to note that the method we used for extracting of the ellipsoid particles from PVA solutions is different than the centrifuging process in literatures200-202. Once the PVA gel with ellipsoid particles is dissolved in water/isopropanol mixture, instead of using a centrifuge, the solvent is sucked through a syringe filter (200 nm pore size) into the syringe. Due to their sizes being larger than the filter pore size, the ellipsoid particles are separated by being blocked at one side of the filter, and the liquid goes through the filter into the syringe. Then, a new syringe filled with DI water is connected to the syringe filter, and the cleaned ellipsoid particles are re-dispersed by pushing the DI water flowing from the side without particles to the side with particles. This process is repeated three times to obtain a large quantity of the cleaned ellipsoid particle as shown in Figure 9.2.

140

Figure 9.2 Ellipsoid particles fabricated by shape modification from spherical particles.

On hydrophilic surfaces, the evaporation of pure liquid drops exhibits three distinct stages: constant contact area mode, constant contact angle mode, and mixed mode23. The contact line keeps receding during the last two stages. However, if the drop contains particles, the drop evaporation behaviors will be altered due to the particle-contact interactions. The different interactions of the drop contact line with the spherical and ellipsoid particles are schematically shown in Figure 9.3, which may reduce the size of the printed deposition much smaller than the drop size. According to the force analysis on spherical particles at the drop contact line, the particle adhesion force to the substrate Fa (including van der Waals force and electrostatic force)

141

provides the friction force fFa that pin the particles at the contact line, while the surface tension

28 force Fs tends to bring the particles to the drop center as shown in Figure 9.3a. As the evaporation initiates, the contact angle hysteresis and non-uniform evaporation along the air- water interface activates the outward flow which brings particles to deposit at the drop edge.

These particles are strongly attached on the substrate and pin the drop contact line. The constant contact area mode is prolonged to almost the entire process. Finally, most of the particles are transported to the drop edge to form the coffee ring effect. The diameter of the deposition of the spherical particles would be the same as the diameter of the drop on the substrate as in Figure

9.3b. However, for drops containing ellipsoid particles, theoretically the evaporation process would be different, as in Figure 9.3d. Once a drop is printed, the particles are transported to the drop edge and strongly attached on the air-water interface due to their anisotropic shape33. The ellipsoid particles trapped on the air-water interface could not provide the force to be pinned at the contact line because of the absence of the particle-substrate interactions. Therefore, during the constant contact angle mode, the drop will start to recede without pinning from the particles as in Figure 9.3d. The receding contact line reduces the area of the air-water interface, and brings the particles to the drop center. Finally these ellipsoid particles are deposited on the substrate uniformly. Because the drop can freely recede on the substrate, the deposition of the ellipsoid particles could be smaller than the drop initial diameter on a partially wetting surface.

142

9.2.3 Large-scale and high throughput fabrication of submicron scale structures via particle self-assembly in inkjet printing processes

In Chapter 5, the nanoparticles in printed single colloidal drops self-assemble into various structures, e.g. concentric multi-ring, spider web, spokes, foam and islands, during the drop evaporation. The effects of particle concentration, particle size, and drop drying conditions have been studied. However, these phenomena are also related to the particle surface functional groups, and substrate wettability. As shown in Figure 9.7a and 9.7b, 20 nm polystyrene particles

143 with two different surface functional groups show distinct deposition morphologies. Carboxylic- modified particles assembled into broken arches in the deposition, while the sulfate-modified particles formed multi-ring structures on the hydrophilic surfaces. This difference may result from their different surface charges. The sulfate-modified and carboxylate-modified 20 nm polystyrene particles show surface charges of 0.085 and 0.715 milliequivalents/gram in a 2% stock solution, respectively. This difference in surface charge may greatly influence the particle- particle and particle-contact line interaction. However, the mechanism of these surface charges governing the deposition morphologies still needs to be explored.

Moreover, if two drops are printed close to each other, as in Figure 9.4c, the evaporation of a pico-liter drop may locally increase the relative humidity, and influences the adjacent drops resulting in the altered deposition morphologies. In contrast to the single drop in 9.4b that the irregular part of the deposition happens at the drop center, the irregular parts of the two adjacent drops drifted towards each other. Once the drop spacing is smaller than the diameter of a single drop, the two drops coalesce, and the deposition could be seen in Figure 9.4d. If the colloidal drops are printed into a line, the coalescence and evaporation of the entire line initiates the formation of even more complex structures. However, the effect of drop spacing and jetting delay on the evaporation and particle deposition in the printed lines needs more effort to explore.

144

Figure 9.4 Various depositions in inkjet-printed drops. (a) Deposition of carboxylic-modified 20 nm polystyrene particles (Invitrogen) in a drop with 0.5% volume fraction on a plasma-treated glass slide. (b) Deposition of sulfate-modified 20 nm polystyrene particles (Invitrogen) in a drop with 0.5% volume fraction on a plasma-treated glass slide. (c) Deposition of two adjacent drop containing 0.5% volume fraction of 20 nm sulfate-modified particles on a plasma-treated glass slide. (d) Deposition of two coalescing drops containing 0.5% volume fraction of 20 nm sulfate- modified particles on a plasma-treated glass slide. (e) Deposition of a series of drops containing 0.25% volume fraction of 20 nm sulfate-modified particles with 200 µm drop spacing.

145

9.2.4 Dielectrophoresis-directed particle self-assembly in evaporating drops

Dielectrophoresis (DEP) has been developed for a series of applications such as mineral separation, micropolishing, manipulation and dispersing of fluid droplets, and manipulation and self-assembly of micron/nanoparticles203. However, the manipulation of particles nowadays focuses on the assembly of particles inside the bulk solutions. As the demand for high throughput fabrication of submicron scale devices increases, utilizing the DEP force to control nanoparticle self-assembly in inkjet-printed drops (~ 10 µm) into submicron structures possesses great potential. In the evaporating drops, besides the outward flow bringing particles towards the drop edge, an extra DEP force is also able to control the particle motion inside the drop, hence the particle deposition morphologies. In the preliminary result in Figure 9.5, a colloidal drop containing 0.1% volume fraction of particles with 1 µm in diameter is placed between two electrodes fabricated with a 60 µm gap on a glass slide. Under the effect of DEP force, the particles start to accumulate in the gap region between the two electrodes as in Figure 9.5b. As the time increases, more and more particles come into the gap region, form closely-packed pearl chains, and hold the particles there as the dark column in Figure 9.5c. At the last stage of evaporation, the velocity of the outward flow greatly increases which drastically increases the drag forces on the particles. Once the force induced by the outward flow dominates the DEP force, the closely-packed pearl chains start to decouple as in Figure 9.5d. We observed the above phenomena, however, the mechanisms governing the particle self-assembly process, and effects of voltage, particle size, dielectric constant, and substrate wettabilty on the particle deposition behavior need more study to facilitate the inkjet printing of submicron scale structures assisted with the DEP effect.

146

Figure 9.5 Positive dielectrophoresis in evaporating droplets. (a) Schematic for experimental setup for a drop evaporating on a glass substrate with two electrodes. (b) Particle moving towards the gap between two electrodes at the beginning of the drop evaporation. (c) Particles assemble into closely-packed pearl chains in the gap during the evaporation. (d) The closely- packed pearl chain decouples at the last stage of evaporation.

9.2.5 Vapor-mediated droplet motion

Droplet manipulation attracts extensive research interests due to its promising applications in chemical analysis204, 205, particle synthesis206, cell encapsulation207, 208, and protein crystallization209-211. In these applications, complex microfluidic systems are fabricated to control the droplet formation and transportation212. Here, we introduce an easy method for freely controlling the droplet motion on a substrate, as shown in Figure 9.6. In Figure 9.6a, a DI water drop is placed on a glass slide treated by Argon plasma for 30 seconds (contact angle < 5°), and a cotton swab saturated with acetone is moving towards the water drop with a constant velocity.

When the swab is close enough to the water drop, the drop is found to spontaneously move away

147 from the swab. Similar phenomena could be observed when the swab approaching the droplet in any direction. However, if the DI water drop is placed on a glass slide as received (contact angle

= 19°), the droplet keeps stationary at any distance away from the swab, as shown in Figure 9.6b.

This phenomenon may provide a free and easy way to manipulate droplet motion on substrates in the microfluidic system.

Figure 9.6 Motion of a water droplet upon approached with an acetone saturated swab on (a) an Argon plasma treated glass slide and (b) a glass slide as received.

148

Our hypothesis on this phenomenon is shown in Figure 9.7. The vapor pressure of acetone is 229.5 mmHg at 25 °C compare to the water vapor pressure of 23.7 mmHg. The acetone quickly diffuses (or evaporates) into the air and forms an acetone vapor filed around the swab. The side of the drop close to the swab absorbs more acetone than the other side. Since the surface tension of acetone is 25.2 mN/m much smaller than the water surface tension of 72.0 mN/m, the side of drop close to the swab (with higher acetone concentration) shows lower surface tension than the other side. The surface tension gradient along the drop interface initiates a Marangoni flow from the side with lower surface tension to the side with high surface tension.

The accumulation of the liquid on the left side of the drop (in Figure 9.7a and 9.7b) locally increases the drop contact angle, and propels the drop to move towards the left. However, on a glass substrate without plasma treatment where the water contact angle is 19°, the drop is relatively thick. Although the Marangoni flow drives the liquid to flow from right to the left on the air-water interface, the thicker drop resulted from higher contact angle allowing the liquid flowing back to the right at the drop bottom as in Figure 9.7c. Therefore, the drop keeps stationary on a glass slide as received. Above is only the hypothesis for the mechanism, and the physics behind this phenomenon, e.g. the effects of liquid surface tension, viscosity and substrate structure on the droplet motion still need more effort to explore.

149

Figure 9.7 Hypothesis of droplet moving on plasma-treated glass, but keeps stationary on glass as received. (a) Analysis of acetone vapor diffusion and absorption by water drop on hydrophilic surfaces. (b) Hypothesized liquid flow in a drop on a plasma-treated glass slide (contact angle < 5°). (c) Hypothesized liquid flow in a drop on a glass slide as received (contact angle = 19°).

150

LIST OF REFERENCES

1. Derby, B., Inkjet Printing of Functional and Structural Materials: Fluid Property Requirements, Feature Stability, and Resolution. Annual Review of Materials Research, Vol 40 2010, 40, 395-414.

2. Liu, Y.; Cui, T. H., Polymer-based rectifying diodes on a glass substrate fabricated by ink-jet printing. Macromolecular Rapid Communications 2005, 26, (4), 289-292.

3. Molesa, S. E.; Volkman, S. K.; Redinger, D. R.; Vornbrock, A. F.; Subramanian, V. In A high-performance all-inkjetted organic transistor technology, Electron Devices Meeting, 2004. IEDM Technical Digest. IEEE International, 2004; IEEE: 2004; pp 1072-1074.

4. Singh, M.; Chae, H. S.; Froehlich, J. D.; Kondou, T.; Li, S.; Mochizuki, A.; Jabbour, G. E., Electroluminescence from printed stellate polyhedral oligomeric silsesquioxanes. Soft Matter 2009, 5, (16), 3002-3005.

5. Haverinen, H. M.; Myllyla, R. A.; Jabbour, G. E., Inkjet printing of light emitting quantum dots. Applied Physics Letters 2009, 94, (7).

6. Marin, V.; Holder, E.; Wienk, M. M.; Tekin, E.; Kozodaev, D.; Schubert, U. S., Ink-jet printing of electron donor/acceptor blends: Towards bulk heterojunction solar cells. Macromolecular Rapid Communications 2005, 26, (4), 319-324.

7. Hoth, C. N.; Choulis, S. A.; Schilinsky, P.; Brabec, C. J., High photovoltaic performance of inkjet-printed polymer: Fullerene blends. Advanced Materials 2007, 19, (22), 3973-3978.

8. Crowley, K.; O'Malley, E.; Morrin, A.; Smyth, M. R.; Killard, A. J., An aqueous ammonia sensor based on an inkjet-printed polyaniline nanoparticle-modified electrode. Analyst 2008, 133, (3), 391-399.

9. Bietsch, A.; Zhang, J. Y.; Hegner, M.; Lang, H. P.; Gerber, C., Rapid functionalization of cantilever array sensors by inkjet printing. Nanotechnology 2004, 15, (8), 873-880.

10. de Gans, B. J.; Duineveld, P. C.; Schubert, U. S., Inkjet printing of polymers: State of the art and future developments. Advanced Materials 2004, 16, (3), 203-213.

11. Dong, H. M.; Carr, W. W.; Morris, J. F., An experimental study of drop-on-demand drop formation. Physics of Fluids 2006, 18, (7).

12. Dong, H. M.; Carr, W. W.; Morris, J. F., Visualization of drop-on-demand inkjet: Drop formation and deposition. Review of Scientific Instruments 2006, 77, (8).

151

13. Furbank, R. J.; Morris, J. F., An experimental study of particle effects on drop formation. Physics of Fluids 2004, 16, (5), 1777-1790.

14. Yarin, A. L., Drop impact dynamics: Splashing, spreading, receding, bouncing... Annual Review of Fluid Mechanics 2006, 38, 159-192.

15. Rioboo, R.; Tropea, C.; Marengo, M., Outcomes from a drop impact on solid surfaces. Atomization and Sprays 2001, 11, (2), 155-165.

16. Rioboo, R.; Marengo, M.; Tropea, C., Time evolution of liquid drop impact onto solid, dry surfaces. Experiments in Fluids 2002, 33, (1), 112-124.

17. Jung, S.; Hutchings, I. M., The impact and spreading of a small liquid drop on a non- porous substrate over an extended time scale. Soft Matter 2012, 8, (9), 2686-2696.

18. Tanner, L., The spreading of silicone oil drops on horizontal surfaces. Journal of Physics D: Applied Physics 1979, 12, (9), 1473.

19. Dong, H. M.; Carr, W. W.; Bucknall, D. G.; Morris, J. F., Temporally-resolved inkjet drop impaction on surfaces. Aiche Journal 2007, 53, (10), 2606-2617.

20. van Dam, D. B.; Le Clerc, C., Experimental study of the impact of an ink-jet printed droplet on a solid substrate. Physics of Fluids 2004, 16, (9), 3403-3414.

21. Nicolas, M., Spreading of a drop of neutrally buoyant suspension. Journal of Fluid Mechanics 2005, 545, 271-280.

22. Schiaffino, S.; Sonin, A. A., Molten droplet deposition and solidification at low Weber numbers. Physics of Fluids 1997, 9, (11), 3172-3187.

23. Kim, J. H.; Ahn, S. I.; Kim, J. H.; Zin, W. C., Evaporation of water droplets on polymer surfaces. Langmuir 2007, 23, (11), 6163-6169.

24. Bourgesmonnier, C.; Shanahan, M. E. R., Influence of Evaporation on Contact-Angle. Langmuir 1995, 11, (7), 2820-2829.

25. Birdi, K. S.; Vu, D. T.; Winter, A., A Study of the Evaporation Rates of Small Water Drops Placed on a Solid-Surface. Journal of Physical Chemistry 1989, 93, (9), 3702-3703.

26. Hu, H.; Larson, R. G., Evaporation of a sessile droplet on a substrate. Journal of Physical Chemistry B 2002, 106, (6), 1334-1344.

27. Deegan, R. D.; Bakajin, O.; Dupont, T. F.; Huber, G.; Nagel, S. R.; Witten, T. A., Capillary flow as the cause of ring stains from dried liquid drops. Nature 1997, 389, (6653), 827- 829.

152

28. Chhasatia, V. H.; Sun, Y., Interaction of bi-dispersed particles with contact line in an evaporating colloidal drop. Soft Matter 2011, 7, (21), 10135-10143.

29. Deegan, R. D.; Bakajin, O.; Dupont, T. F.; Huber, G.; Nagel, S. R.; Witten, T. A., Contact line deposits in an evaporating drop. Physical Review E 2000, 62, (1), 756-765.

30. Kuncicky, D. M.; Velev, O. D., Surface-guided templating of particle assemblies inside drying sessile droplets. Langmuir 2008, 24, (4), 1371-1380.

31. Park, J.; Moon, J., Control of colloidal particle deposit patterns within picoliter droplets ejected by ink-jet printing. Langmuir 2006, 22, (8), 3506-3513.

32. Ristenpart, W. D.; Kim, P. G.; Domingues, C.; Wan, J.; Stone, H. A., Influence of substrate conductivity on circulation reversal in evaporating drops. Physical Review Letters 2007, 99, (23).

33. Yunker, P. J.; Still, T.; Lohr, M. A.; Yodh, A. G., Suppression of the coffee-ring effect by shape-dependent capillary interactions. Nature 2011, 476, (7360), 308-311.

34. Park, B. J.; Furst, E. M., Attractive interactions between colloids at the oil-water interface. Soft Matter 2011, 7, (17), 7676-7682.

35. Perelaer, J.; Smith, P. J.; Hendriks, C. E.; van den Berg, A. M. J.; Schubert, U. S., The preferential deposition of silica micro-particles at the boundary of inkjet-printed droplets. Soft Matter 2008, 4, (5), 1072-1078.

36. Rein, M., Phenomena of Liquid-Drop Impact on Solid and Liquid Surfaces. Fluid Dynamics Research 1993, 12, (2), 61-93.

37. Yarin, A. L.; Weiss, D. A., Impact of Drops on Solid-Surfaces - Self-Similar Capillary Waves, and Splashing as a New-Type of Kinematic Discontinuity. Journal of Fluid Mechanics 1995, 283, 141-173.

38. Levin, Z.; Hobbs, P. V., Splashing of Water Drops on Solid and Wetted Surfaces: Hydrodynamics and Charge Separation. Philosophical Transactions of the Royal Society of London. Series A, Mathematical and Physical Sciences 1971, 269, (1200), 555-585.

39. Weiss, D. A.; Yarin, A. L., Single drop impact onto liquid films: neck distortion, jetting, tiny bubble entrainment, and crown formation. Journal of Fluid Mechanics 1999, 385, 229-254.

40. Rioboo, R.; Bauthier, C.; Conti, J.; Voue, M.; De Coninck, J., Experimental investigation of splash and crown formation during single drop impact on wetted surfaces. Experiments in Fluids 2003, 35, (6), 648-652.

153

41. Fukai, J.; Shiiba, Y.; Yamamoto, T.; Miyatake, O.; Poulikakos, D.; Megaridis, C. M.; Zhao, Z., Wetting Effects on the Spreading of a Liquid Droplet Colliding with a Flat Surface - Experiment and Modeling. Physics of Fluids 1995, 7, (2), 236-247.

42. Sikalo, S.; Wilhelm, H.; Roisman, I.; Jakirlic, S.; Tropea, C., Dynamic contact angle of spreading droplets: experiments and simulations. Phys. Fluids 2005, 17, 062103.

43. Kim, H. Y.; Chun, J. H., The recoiling of liquid droplets upon collision with solid surfaces. Physics of Fluids 2001, 13, (3), 643-659.

44. Tanner, L., The spreading of silicone oil drops on horizontal surfaces. Journal of Physics D: Applied Physics 1979, 12, 1473.

45. Fujimoto, H.; Ogino, T.; Takuda, H.; Hatta, N., Collision of a droplet with a hemispherical static droplet on a solid. International Journal of Multiphase Flow 2001, 27, (7), 1227-1245.

46. Tong, A. Y.; Kasliwal, S.; Fujimoto, H., On the successive impingement of droplets onto a substrate. Numerical Heat Transfer Part a-Applications 2007, 52, (6), 531-548.

47. Son, Y.; Kim, C.; Yang, D. H.; Alm, D. J., Spreading of an inkjet droplet on a solid surface with a controlled contact angle at low weber and Reynolds numbers. Langmuir 2008, 24, (6), 2900-2907.

48. Hsiao, W. K.; Hoath, S. D.; Martin, G. D.; Hutchings, I. M., Ink Jet Printing for Direct Mask Deposition in Printed Circuit Board Fabrication. Journal of Imaging Science and Technology 2009, 53, (5).

49. Hoffman, R. L., A study of the advancing interface. I. Interface shape in liquid--gas systems. Journal of Colloid and Interface Science 1975, 50, (2), 228-241.

50. Voinov, O., Hydrodynamics of wetting. Fluid Dynamics 1976, 11, (5), 714-721.

51. Huh, C.; Scriven, L. E., Hydrodynamic model of steady movement of a solid/liquid/fluid contact line. Journal of Colloid and Interface Science 1971, 35, (1), 85-101.

52. Zhang, X. G.; Basaran, O. A., Dynamic surface tension effects in impact of a drop with a solid surface. Journal of Colloid and Interface Science 1997, 187, (1), 166-178.

53. Thoroddsen, S. T.; Takehara, K.; Etoh, T. G., The coalescence speed of a pendent and a sessile drop. Journal of Fluid Mechanics 2005, 527, 85-114.

54. Tadmor, R., Line energy and the relation between advancing, receding, and young contact angles. Langmuir 2004, 20, (18), 7659-7664.

154

55. Krebs, F. C.; Gevorgyan, S. A.; Alstrup, J., A roll-to-roll process to flexible polymer solar cells: model studies, manufacture and operational stability studies. Journal of Materials Chemistry 2009, 19, (30), 5442-5451.

56. Tekin, E.; Smith, P. J.; Schubert, U. S., Inkjet printing as a deposition and patterning tool for polymers and inorganic particles. Soft Matter 2008, 4, (4), 703-713.

57. Pudas, M.; Hagberg, J.; Leppavuori, S., Roller-type gravure offset printing of conductive inks for high-resolution printing on ceramic substrates. International Journal of Electronics 2005, 92, (5), 251-269.

58. Gaikwad, A. M.; Whiting, G. L.; Steingart, D. A.; Arias, A. C., Highly Flexible, Printed Alkaline Batteries Based on Mesh-Embedded Electrodes. Advanced Materials 2011, 23, (29), 3251-+.

59. Lee, H. H.; Chou, K. S.; Huang, K. C., Inkjet printing of nanosized silver colloids. Nanotechnology 2005, 16, (10), 2436-2441.

60. Rogers, J. A.; Bao, Z., Printed plastic electronics and paperlike displays. Journal of Polymer Science Part a-Polymer Chemistry 2002, 40, (20), 3327-3334.

61. Glotzer, S. C.; Solomon, M. J.; Kotov, N. A., Self-assembly: From nanoscale to microscale colloids. Aiche Journal 2004, 50, (12), 2978-2985.

62. Boncheva, M.; Whitesides, G. M., Templated self-assembly: Formation of folded structures by relaxation of pre-stressed, planar tapes. Advanced Materials 2005, 17, (5), 553-+.

63. Denkov, N. D.; Velev, O. D.; Kralchevsky, P. A.; Ivanov, I. B.; Yoshimura, H.; Nagayama, K., 2-Dimensional Crystallization. Nature 1993, 361, (6407), 26-26.

64. Kapur, N.; Gaskell, P. H., Morphology and dynamics of droplet coalescence on a surface. Physical Review E 2007, 75, (5).

65. Menchaca-Rocha, A.; Martinez-Davalos, A.; Nunez, R.; Popinet, S.; Zaleski, S., Coalescence of liquid drops by surface tension. Physical Review E 2001, 63, (4), art. no.-046309.

66. Nilsson, M. A.; Rothstein, J. P., The effect of contact angle hysteresis on droplet coalescence and mixing. Journal of Colloid and Interface Science 2011, 363, (2), 646-654.

67. Andrieu, C.; Beysens, D. A.; Nikolayev, V. S.; Pomeau, Y., Coalescence of sessile drops. Journal of Fluid Mechanics 2002, 453, 427-438.

68. Aarts, D. G. A. L.; Lekkerkerker, H. N. W.; Guo, H.; Wegdam, G. H.; Bonn, D., Hydrodynamics of droplet coalescence. Physical Review Letters 2005, 95, (16).

155

69. Borcia, R.; Bestehorn, M., On the coalescence of sessile drops with miscible liquids. European Physical Journal E 2011, 34, (8).

70. Castrejon-Pita, J. R.; Betton, E. S.; Kubiak, K. J.; Wilson, M. C. T.; Hutchings, I. M., The dynamics of the impact and coalescence of droplets on a solid surface. Biomicrofluidics 2011, 5, (1).

71. Fujimoto, H.; Tong, A. Y.; Takuda, H., Interaction phenomena of two water droplets successively impacting onto a solid surface. International Journal of Thermal Sciences 2008, 47, (3), 229-236.

72. Duineveld, P. C., The stability of ink-jet printed lines of liquid with zero receding contact angle on a homogeneous substrate. Journal of Fluid Mechanics 2003, 477, 175-200.

73. Stringer, J.; Derby, B., Formation and Stability of Lines Produced by Inkjet Printing. Langmuir 2010, 26, (12), 10365-10372.

74. Rioboo, R.; Marengo, M.; Cossali, G.; Tropea, C., Comparison of Droplet Impact: Dry and Wet Cases. ILASS-Europe 2000 2000.

75. Hu, H.; Larson, R. G., Marangoni effect reverses coffee-ring depositions. Journal of Physical Chemistry B 2006, 110, (14), 7090-7094.

76. Bensimon, A.; Simon, A.; Chiffaudel, A.; Croquette, V.; Heslot, F.; Bensimon, D., Alignment and Sensitive Detection of DNA by a Moving Interface. Science 1994, 265, (5181), 2096-2098.

77. Shmuylovich, L.; Shen, A. Q.; Stone, H. A., Surface morphology of drying latex films: Multiple ring formation. Langmuir 2002, 18, (9), 3441-3445.

78. Adachi, E.; Dimitrov, A. S.; Nagayama, K., Stripe Patterns Formed on a Glass-Surface during Droplet Evaporation. Langmuir 1995, 11, (4), 1057-1060.

79. Maheshwari, S.; Zhang, L.; Zhu, Y. X.; Chang, H. C., Coupling between precipitation and contact-line dynamics: Multiring stains and stick-slip motion. Physical Review Letters 2008, 100, (4).

80. Zhang, L.; Maheshwari, S.; Chang, H. C.; Zhu, Y. X., Evaporative self-assembly from complex DNA-colloid suspensions. Langmuir 2008, 24, (8), 3911-3917.

81. Xu, J.; Xia, J. F.; Hong, S. W.; Lin, Z. Q.; Qiu, F.; Yang, Y. L., Self-assembly of gradient concentric rings via solvent evaporation from a capillary bridge. Physical Review Letters 2006, 96, (6).

156

82. Moffat, J. R.; Sefiane, K.; Shanahan, M. E. R., Effect of TiO2 Nanoparticles on Contact Line Stick-Slip Behavior of Volatile Drops. Journal of Physical Chemistry B 2009, 113, (26), 8860-8866.

83. Han, W.; Byun, M.; Li, B.; Pang, X. C.; Lin, Z. Q., A Simple Route to Hierarchically Assembled Micelles and Inorganic Nanoparticles. Angewandte Chemie-International Edition 2012, 51, (50), 12588-12592.

84. Byun, M.; Laskowski, R. L.; He, M.; Qiu, F.; Jeffries-El, M.; Lin, Z. Q., Controlled evaporative self-assembly of hierarchically structured regioregular conjugated polymers. Soft Matter 2009, 5, (8), 1583-1586.

85. Hong, S. W.; Wang, J.; Lin, Z. Q., Evolution of Ordered Block Copolymer Serpentines into a Macroscopic, Hierarchically Ordered Web. Angewandte Chemie-International Edition 2009, 48, (44), 8356-8360.

86. Li, B.; Han, W.; Byun, M.; Zhu, L.; Zou, Q. Z.; Lin, Z. Q., Macroscopic Highly Aligned DNA Nanowires Created by Controlled Evaporative Self-Assembly. Acs Nano 2013, 7, (5), 4326-4333.

87. Xu, J.; Xia, J. F.; Lin, Z. Q., Evaporation-induced self-assembly of nanoparticles from a sphere-on-flat geometry. Angewandte Chemie-International Edition 2007, 46, (11), 1860-1863.

88. Prevo, B. G.; Velev, O. D., Controlled, rapid deposition of structured coatings from micro- and nanoparticle suspensions. Langmuir 2004, 20, (6), 2099-2107.

89. Shalaev, V. M., Optical negative-index metamaterials. Nature Photonics 2007, 1, (1), 41- 48.

90. Kajiya, T.; Kaneko, D.; Doi, M., Dynamical Visualization of "Coffee Stain Phenomenon" in Droplets of Polymer Solution via Fluorescent Microscopy. Langmuir 2008, 24, (21), 12369- 12374.

91. Tadmor, R., Line energy, line tension and drop size. Surface Science 2008, 602, (14), L108-L111.

92. Cachile, M.; Benichou, O.; Poulard, C.; Cazabat, A. M., Evaporating droplets. Langmuir 2002, 18, (21), 8070-8078.

93. Erbil, H. Y.; McHale, G.; Newton, M. I., Drop evaporation on solid surfaces: Constant contact angle mode. Langmuir 2002, 18, (7), 2636-2641.

157

94. Orejon, D.; Sefiane, K.; Shanahan, M. E. R., Stick-Slip of Evaporating Droplets: Substrate Hydrophobicity and Nanoparticle Concentration. Langmuir 2011, 27, (21), 12834- 12843.

95. Popov, Y. O., Evaporative deposition patterns: Spatial dimensions of the deposit. Physical Review E 2005, 71, (3).

96. Popov, Y. O.; Witten, T. A., Characteristic angles in the wetting of an angular region: Deposit growth. Physical Review E 2003, 68, (3).

97. Poulard, C.; Benichou, O.; Cazabat, A. M., Freely receding evaporating droplets. Langmuir 2003, 19, (21), 8828-8834.

98. Sultan, E.; Boudaoud, A.; Ben Amar, M., Evaporation of a thin film: diffusion of the vapour and Marangoni instabilities. Journal of Fluid Mechanics 2005, 543, 183-202.

99. Yen, T. J.; Padilla, W. J.; Fang, N.; Vier, D. C.; Smith, D. R.; Pendry, J. B.; Basov, D. N.; Zhang, X., Terahertz magnetic response from artificial materials. Science 2004, 303, (5663), 1494-1496.

100. Peterson, D. S.; Rohr, T.; Svec, F.; Frechet, J. M. J., Enzymatic microreactor-on-a-chip: Protein mapping using trypsin immobilized on porous polymer monoliths molded in channels of microfluidic devices. Analytical Chemistry 2002, 74, (16), 4081-4088.

101. Adams, J. J.; Duoss, E. B.; Malkowski, T. F.; Motala, M. J.; Ahn, B. Y.; Nuzzo, R. G.; Bernhard, J. T.; Lewis, J. A., Conformal Printing of Electrically Small Antennas on Three- Dimensional Surfaces. Advanced Materials 2011, 23, (11), 1335-1340.

102. Wood, N. R.; Wolsiefer, A. I.; Cohn, R. W.; Williams, S. J., Dielectrophoretic trapping of nanoparticles with an electrokinetic nanoprobe. Electrophoresis 2013, 34, (13), 1922-1930.

103. Chen, X. B.; Li, C.; Gratzel, M.; Kostecki, R.; Mao, S. S., Nanomaterials for renewable energy production and storage. Chemical Society Reviews 2012, 41, (23), 7909-7937.

104. Bao, N.; Shen, L.; Takata, T.; Domen, K., Self-templated synthesis of nanoporous CdS nanostructures for highly efficient photocatalytic hydrogen production under visible light. Chemistry of Materials 2007, 20, (1), 110-117.

105. Haynes, C. L.; Van Duyne, R. P., Nanosphere lithography: A versatile nanofabrication tool for studies of size-dependent nanoparticle optics. Journal of Physical Chemistry B 2001, 105, (24), 5599-5611.

106. Vogel, N.; Weiss, C. K.; Landfester, K., From soft to hard: the generation of functional and complex colloidal monolayers for nanolithography. Soft Matter 2012, 8, (15), 4044-4061.

158

107. Groves, T. R.; Pickard, D.; Rafferty, B.; Crosland, N.; Adam, D.; Schubert, G., Maskless electron beam lithography: prospects, progress, and challenges. Microelectronic Engineering 2002, 61–62, (0), 285-293.

108. Parikh, D.; Craver, B.; Nounu, H. N.; Fong, F.-O.; Wolfe, J. C., Nanoscale Pattern Definition on Nonplanar Surfaces Using Ion Beam Proximity Lithography and Conformal Plasma-Deposited Resist. Microelectromechanical Systems, Journal of 2008, 17, (3), 735-740.

109. Matsui, S.; Ochiai, Y., Focused ion beam applications to solid state devices. Nanotechnology 1996, 7, (3), 247-258.

110. Schift, H., Nanoimprint lithography: an old story in modern times? A review. Journal of Vacuum Science & Technology B 2008, 26, (2), 458-480.

111. Lee, K. B.; Park, S. J.; Mirkin, C. A.; Smith, J. C.; Mrksich, M., Protein nanoarrays generated by dip-pen nanolithography. Science 2002, 295, (5560), 1702-1705.

112. Seo, J. H.; Park, J. H.; Kim, S. I.; Park, B. J.; Ma, Z. Q.; Choi, J.; Ju, B. K., Nanopatterning by Laser Interference Lithography: Applications to Optical Devices. Journal of Nanoscience and Nanotechnology 2014, 14, (2), 1521-1532.

113. Ishibashi, Y.; Asada, S.; Igaki, H.; Yamada, S.; Tsuruta, S.; Shimura, M.; Kakita, K.; Ono, A.; Sakata, M., Results of proficiency testing for determination of dioxins in fly-ash (2nd round) based on ISO/IEC guide 43-1. Bunseki Kagaku 2005, 54, (2), 167-173.

114. Landis, E. C.; Hamers, R. J., Covalent Grafting of Redox-Active Molecules to Vertically Aligned Carbon Nanofiber Arrays via "Click" Chemistry. Chemistry of Materials 2009, 21, (4), 724-730.

115. Liu, J. W.; Li, J.; Sedhain, A.; Lin, J. Y.; Jiang, H. X., Structure and Photoluminescence Study of TiO(2) Nanoneedle Texture along Vertically Aligned Carbon Nanofiber Arrays. Journal of Physical Chemistry C 2008, 112, (44), 17127-17132.

116. Chandra, D.; Yang, S., Stability of High-Aspect-Ratio Micropillar Arrays against Adhesive and Capillary Forces. Accounts of Chemical Research 2010, 43, (8), 1080-1091.

117. Simovski, C. R.; Belov, P. A.; Atrashchenko, A. V.; Kivshar, Y. S., Wire Metamaterials: Physics and Applications. Advanced Materials 2012, 24, (31), 4229-4248.

118. Grinthal, A.; Kang, S. H.; Epstein, A. K.; Aizenberg, M.; Khan, M.; Aizenberg, J., Steering nanofibers: An integrative approach to bio-inspired fiber fabrication and assembly. Nano Today 2012, 7, (1), 35-52.

159

119. De Volder, M.; Hart, A. J., Engineering Hierarchical Nanostructures by Elastocapillary Self-Assembly. Angewandte Chemie-International Edition 2013, 52, (9), 2412-2425.

120. Coleman, J. N.; Khan, U.; Blau, W. J.; Gun'ko, Y. K., Small but strong: A review of the mechanical properties of carbon nanotube-polymer composites. Carbon 2006, 44, (9), 1624-1652.

121. Kawata, S.; Ono, A.; Verma, P., Subwavelength colour imaging with a metallic nanolens. Nature Photonics 2008, 2, (7), 438-442.

122. Ono, A.; Kato, J.; Kawata, S., Subwavelength optical imaging through a metallic nanorod array. Physical Review Letters 2005, 95, (26).

123. Nordlander, P., Plasmonics - Subwavelength imaging in colour. Nature Photonics 2008, 2, (7), 387-388.

124. Gao, H. J.; Wang, X.; Yao, H. M.; Gorb, S.; Arzt, E., Mechanics of hierarchical adhesion structures of geckos. Mechanics of Materials 2005, 37, (2-3), 275-285.

125. Qu, L. T.; Dai, L. M.; Stone, M.; Xia, Z. H.; Wang, Z. L., Carbon nanotube arrays with strong shear binding-on and easy normal lifting-off. Science 2008, 322, (5899), 238-242.

126. Izadi-Najafabadi, A.; Yasuda, S.; Kobashi, K.; Yamada, T.; Futaba, D. N.; Hatori, H.; Yumura, M.; Iijima, S.; Hata, K., Extracting the Full Potential of Single-Walled Carbon Nanotubes as Durable Supercapacitor Electrodes Operable at 4 V with High Power and Energy Density. Advanced Materials 2010, 22, (35), E235-E241.

127. Futaba, D. N.; Hata, K.; Yamada, T.; Hiraoka, T.; Hayamizu, Y.; Kakudate, Y.; Tanaike, O.; Hatori, H.; Yumura, M.; Iijima, S., Shape-engineerable and highly densely packed single- walled carbon nanotubes and their application as super-capacitor electrodes. Nature Materials 2006, 5, (12), 987-994.

128. Lau, K. K. S.; Bico, J.; Teo, K. B. K.; Chhowalla, M.; Amaratunga, G. A. J.; Milne, W. I.; McKinley, G. H.; Gleason, K. K., Superhydrophobic carbon nanotube forests. Nano Letters 2003, 3, (12), 1701-1705.

129. Pint, C. L.; Takei, K.; Kapadia, R.; Zheng, M.; Ford, A. C.; Zhang, J. J.; Jamshidi, A.; Bardhan, R.; Urban, J. J.; Wu, M.; Ager, J. W.; Oye, M. M.; Javey, A., Rationally Designed, Three-Dimensional Carbon Nanotube Back-Contacts for Efficient Solar Devices. Advanced Energy Materials 2011, 1, (6), 1040-1045.

130. Mor, G. K.; Shankar, K.; Paulose, M.; Varghese, O. K.; Grimes, C. A., Use of highly- ordered TiO2 nanotube arrays in dye-sensitized solar cells. Nano Letters 2006, 6, (2), 215-218.

160

131. Fachin, F.; Chen, G. D.; Toner, M.; Wardle, B. L., Integration of Bulk Nanoporous Elements in Microfluidic Devices With Application to Biomedical Diagnostics. Journal of Microelectromechanical Systems 2011, 20, (6), 1428-1438.

132. Kaji, N.; Tezuka, Y.; Takamura, Y.; Ueda, M.; Nishimoto, T.; Nakanishi, H.; Horiike, Y.; Baba, Y., Separation of long DNA molecules by quartz nanopillar chips under a direct current electric field. Analytical Chemistry 2004, 76, (1), 15-22.

133. Mingo, N.; Broido, D. A., Carbon nanotube ballistic thermal conductance and its limits. Physical Review Letters 2005, 95, (9).

134. Tawfick, S.; De Volder, M.; Copic, D.; Park, S. J.; Oliver, C. R.; Polsen, E. S.; Roberts, M. J.; Hart, A. J., Engineering of Micro- and Nanostructured Surfaces with Anisotropic Geometries and Properties. Advanced Materials 2012, 24, (13), 1628-1674.

135. Tanaka, T.; Morigami, M.; Atoda, N., Mechanism of Resist Pattern Collapse during Development Process. Japanese Journal of Applied Physics Part 1-Regular Papers Short Notes & Review Papers 1993, 32, (12B), 6059-6064.

136. Nguyen-Vu, T. D. B.; Chen, H.; Cassell, A. M.; Andrews, R.; Meyyappan, M.; Li, J., Vertically aligned carbon nanofiber arrays: An advance toward electrical-neural interfaces. small 2006, 2, (1), 89-94.

137. De Volder, M. F. L.; Park, S. J.; Tawfick, S. H.; Vidaud, D. O.; Hart, A. J., Fabrication and electrical integration of robust carbon nanotube micropillars by self-directed elastocapillary densification. Journal of Micromechanics and Microengineering 2011, 21, (4).

138. Journet, C.; Moulinet, S.; Ybert, C.; Purcell, S. T.; Bocquet, L., Contact angle measurements on superhydrophobic carbon nanotube forests: Effect of fluid pressure. Europhysics Letters 2005, 71, (1), 104-109.

139. Chakrapani, N.; Wei, B. Q.; Carrillo, A.; Ajayan, P. M.; Kane, R. S., Capillarity-driven assembly of two-dimensional cellular carbon nanotube foams. Proceedings of the National Academy of Sciences of the United States of America 2004, 101, (12), 4009-4012.

140. Liu, H.; Li, S. H.; Zhai, J.; Li, H. J.; Zheng, Q. S.; Jiang, L.; Zhu, D. B., Self-assembly of large-scale micropatterns on aligned carbon nanotube films. Angewandte Chemie-International Edition 2004, 43, (9), 1146-1149.

141. Tawfick, S.; Hart, A. J.; De Volder, M., Capillary bending of Janus carbon nanotube micropillars. Nanoscale 2012, 4, (13), 3852-3856.

142. De Volder, M. F. L.; Tawfick, S.; Park, S. J.; Hart, A. J., Corrugated Carbon Nanotube Microstructures with Geometrically Tunable Compliance. Acs Nano 2011, 5, (9), 7310-7317.

161

143. De Volder, M.; Tawfick, S. H.; Park, S. J.; Copic, D.; Zhao, Z. Z.; Lu, W.; Hart, A. J., Diverse 3D Microarchitectures Made by Capillary Forming of Carbon Nanotubes. Advanced Materials 2010, 22, (39), 4384-4389.

144. Chen, B.; Seidel, S.; Hori, H.; Gupta, M., Self-Assembly of Pillars Modified with Vapor Deposited Polymer Coatings. Acs Applied Materials & Interfaces 2011, 3, (11), 4201-4205.

145. Duan, H. G.; Berggren, K. K., Directed Self-Assembly at the 10 nm Scale by Using Capillary Force-Induced Nanocohesion. Nano Letters 2010, 10, (9), 3710-3716.

146. Duan, H. G.; Yang, J. K. W.; Berggren, K. K., Controlled Collapse of High-Aspect-Ratio Nanostructures. small 2011, 7, (18), 2661-2668.

147. Ishikawa, A.; Sakata, M.; Kawai, A., Meniscus analysis in micro gap during liquid drying process. Journal of Photopolymer Science and Technology 2004, 17, (3), 457-460.

148. Choi, S.; Yan, M.; Adesida, I., Fabrication of triangular nanochannels using the collapse of hydrogen silsesquioxane resists. Applied Physics Letters 2008, 93, (16).

149. Chandra, D.; Yang, S.; Soshinsky, A. A.; Gambogi, R. J., Biomimetic Ultrathin Whitening by Capillary-Force-Induced Random Clustering of Hydrogel Micropillar Arrays. Acs Applied Materials & Interfaces 2009, 1, (8), 1698-1704.

150. Bico, J.; Roman, B.; Moulin, L.; Boudaoud, A., Elastocapillary coalescence in wet hair. Nature 2004, 432, (7018), 690-690.

151. Cai, H.; Xu, Y.; He, P. G.; Fang, Y. Z., Indicator free DNA hybridization detection by impedance measurement based on the DNA-doped conducting polymer film formed on the carbon nanotube modified electrode. Electroanalysis 2003, 15, (23-24), 1864-1870.

152. Zhao, Y. P.; Fan, J. G., Clusters of bundled nanorods in nanocarpet effect. Applied Physics Letters 2006, 88, (10).

153. Chen, R.; Lu, M. C.; Srinivasan, V.; Wang, Z.; Cho, H. H.; Majumdar, A., Nanowires for Enhanced Boiling Heat Transfer. Nano Letters 2009, 9, (2), 548-553.

154. Namatsu, H.; Kurihara, K.; Nagase, M.; Iwadate, K.; Murase, K., Dimensional Limitations of Silicon Nanolines Resulting from Pattern Distortion Due to Surface-Tension of Rinse Water. Applied Physics Letters 1995, 66, (20), 2655-2657.

155. Hill, J. J.; Haller, K.; Gelfand, B.; Ziegler, K. J., Eliminating Capillary Coalescence of Nanowire Arrays with Applied Electric Fields. Acs Applied Materials & Interfaces 2010, 2, (7), 1992-1998.

162

156. Dev, A.; Chaudhuri, S., Uniform large-scale growth of micropatterned arrays of ZnO nanowires synthesized by a surfactant assisted approach. Nanotechnology 2007, 18, (17).

157. Yu, Z.; McKnight, T. E.; Ericson, M. N.; Melechko, A. V.; Simpson, M. L.; Morrison, B., Vertically aligned carbon nanofiber arrays record electrophysiological signals from hippocampal slices. Nano Letters 2007, 7, (8), 2188-2195.

158. Blossey, R., Self-cleaning surfaces - virtual realities. Nature Materials 2003, 2, (5), 301- 306.

159. Nakajima, A.; Hashimoto, K.; Watanabe, T.; Takai, K.; Yamauchi, G.; Fujishima, A., Transparent superhydrophobic thin films with self-cleaning properties. Langmuir 2000, 16, (17), 7044-7047.

160. Miwa, M.; Nakajima, A.; Fujishima, A.; Hashimoto, K.; Watanabe, T., Effects of the surface roughness on sliding angles of water droplets on superhydrophobic surfaces. Langmuir 2000, 16, (13), 5754-5760.

161. Rothstein, J. P., Slip on Superhydrophobic Surfaces. Annual Review of Fluid Mechanics 2010, 42, 89-109.

162. Lee, C.; Choi, C. H.; Kim, C. J., Structured surfaces for a giant liquid slip. Physical Review Letters 2008, 101, (6).

163. Dietz, C.; Rykaczewski, K.; Fedorov, A. G.; Joshi, Y., Visualization of droplet departure on a superhydrophobic surface and implications to heat transfer enhancement during dropwise condensation. Applied Physics Letters 2010, 97, (3).

164. Miljkovic, N.; Enright, R.; Wang, E. N., Effect of Droplet Morphology on Growth Dynamics and Heat Transfer during Condensation on Superhydrophobic Nanostructured Surfaces. Acs Nano 2012, 6, (2), 1776-1785.

165. Boreyko, J. B.; Chen, C. H., Self-Propelled Dropwise Condensate on Superhydrophobic Surfaces. Physical Review Letters 2009, 103, (18).

166. Tuteja, A.; Choi, W.; Mabry, J. M.; McKinley, G. H.; Cohen, R. E., Robust omniphobic surfaces. Proceedings of the National Academy of Sciences of the United States of America 2008, 105, (47), 18200-18205.

167. Nguyen, T. P. N.; Brunet, P.; Coffinier, Y.; Boukherroub, R., Quantitative Testing of Robustness on Superomniphobic Surfaces by Drop Impact. Langmuir 2010, 26, (23), 18369- 18373.

168. Bocquet, L.; Lauga, E., A smooth future? Nature Materials 2011, 10, (5), 334-337.

163

169. Poetes, R.; Holtzmann, K.; Franze, K.; Steiner, U., Metastable Underwater Superhydrophobicity. Physical Review Letters 2010, 105, (16).

170. Quere, D., Wetting and roughness. Annual Review of Materials Research 2008, 38, 71-99.

171. Quere, D., Non-sticking drops. Reports on Progress in Physics 2005, 68, (11), 2495-2532.

172. Bohn, H. F.; Federle, W., Insect aquaplaning: Nepenthes pitcher plants capture prey with the peristome, a fully wettable water-lubricated anisotropic surface. Proceedings of the National Academy of Sciences of the United States of America 2004, 101, (39), 14138-14143.

173. Kim, P.; Wong, T. S.; Alvarenga, J.; Kreder, M. J.; Adorno-Martinez, W. E.; Aizenberg, J., Liquid-Infused Nanostructured Surfaces with Extreme Anti-Ice and Anti-Frost Performance. Acs Nano 2012, 6, (8), 6569-6577.

174. Epstein, A. K.; Wong, T. S.; Belisle, R. A.; Boggs, E. M.; Aizenberg, J., Liquid-infused structured surfaces with exceptional anti-biofouling performance. Proceedings of the National Academy of Sciences of the United States of America 2012, 109, (33), 13182-13187.

175. Yang, L. H.; Quan, X. J.; Cheng, P.; Cheng, Z. M., A free energy model and availability analysis for onset of condensation on rigid and liquid surfaces in moist air. International journal of heat and mass transfer 2014, 78, 460-467.

176. Leslie, D. C.; Waterhouse, A.; Berthet, J. B.; Valentin, T. M.; Watters, A. L.; Jain, A.; Kim, P.; Hatton, B. D.; Nedder, A.; Donovan, K.; Super, E. H.; Howell, C.; Johnson, C. P.; Vu, T. L.; Bolgen, D. E.; Rifai, S.; Hansen, A. R.; Aizenberg, M.; Super, M.; Aizenberg, J.; Ingber, D. E., A bioinspired omniphobic surface coating on medical devices prevents thrombosis and biofouling. Nat. Biotech. 2014, 32, 1134-1140.

177. Li, J. S.; Kleintschek, T.; Rieder, A.; Cheng, Y.; Baumbach, T.; Obst, U.; Schwartz, T.; Levkin, P. A., Hydrophobic Liquid-Infused Porous Polymer Surfaces for Antibacterial Applications. Acs Applied Materials & Interfaces 2013, 5, (14), 6704-6711.

178. Chen, J.; Dou, R. M.; Cui, D. P.; Zhang, Q. L.; Zhang, Y. F.; Xu, F. J.; Zhou, X.; Wang, J. J.; Song, Y. L.; Jiang, L., Robust Prototypical Anti-icing Coatings with a Self-lubricating Liquid Water Layer between Ice and Substrate. Acs Applied Materials & Interfaces 2013, 5, (10), 4026- 4030.

179. Ueda, E.; Levkin, P. A., Micropatterning Hydrophobic Liquid on a Porous Polymer Surface for Long-Term Selective Cell-Repellency. Advanced Healthcare Materials 2013, 2, (11), 1425-1429.

180. Stone, H. A., Ice-Phobic Surfaces That Are Wet. Acs Nano 2012, 6, (8), 6536-6540.

164

181. Anand, S.; Paxson, A. T.; Dhiman, R.; Smith, J. D.; Varanasi, K. K., Enhanced Condensation on Lubricant-Impregnated Nanotextured Surfaces. Acs Nano 2012, 6, (11), 10122- 10129.

182. Rykaczewski, K.; Paxson, A. T.; Staymates, M.; Walker, M. L.; Sun, X. D.; Anand, S.; Srinivasan, S.; McKinley, G. H.; Chinn, J.; Scott, J. H. J.; Varanasi, K. K., Dropwise Condensation of Low Surface Tension Fluids on Omniphobic Surfaces. Scientific Reports 2014, 4.

183. Xiao, R.; Miljkovic, N.; Enright, R.; Wang, E. N., Immersion Condensation on Oil- Infused Heterogeneous Surfaces for Enhanced Heat Transfer. Scientific Reports 2013, 3.

184. Smith, J. D.; Dhiman, R.; Anand, S.; Reza-Garduno, E.; Cohen, R. E.; McKinley, G. H.; Varanasi, K. K., Droplet mobility on lubricant-impregnated surfaces. Soft Matter 2013, 9, (6), 1772-1780.

185. Daniel, D.; Mankin, M. N.; Belisle, R. A.; Wong, T.-S.; Aizenberg, J., Lubricant-infused micro/nano-structured surfaces with tunable dynamic omniphobicity at high temperatures. Applied Physics Letters 2013, 102, (231603).

186. Oron, A.; Rosenau, P., On a Nonlinear Thermocapillary Effect in Thin Liquid Layers. Journal of Fluid Mechanics 1994, 273, 361-374.

187. Hamraoui, A.; Cachile, M.; Poulard, C.; Cazabat, A. M., Fingering phenomena during spreading of surfactant solutions. Colloids and Surfaces a-Physicochemical and Engineering Aspects 2004, 250, (1-3), 215-221.

188. Craster, R. V.; Matar, O. K., Dynamics and stability of thin liquid films. Reviews of Modern Physics 2009, 81, (3), 1131-1198.

189. Becker, J.; Grun, G.; Seemann, R.; Mantz, H.; Jacobs, K.; Mecke, K. R.; Blossey, R., Complex dewetting scenarios captured by thin-film models. Nature Materials 2003, 2, (1), 59-63.

190. Park, C.; Nosoko, T., Three‐dimensional wave dynamics on a falling film and associated mass transfer. Aiche Journal 2003, 49, (11), 2715-2727.

191. Schneemilch, M.; Cazabat, A. M., Shock separation in wetting films driven by thermal gradients. Langmuir 2000, 16, (25), 9850-9856.

192. Rauscher, M.; Dietrich, S., Wetting phenomena in nanofluidics. Annual Review of Materials Research 2008, 38, 143-172.

193. Bocquet, L.; Charlaix, E., Nanofluidics, from bulk to interfaces. Chemical Society Reviews 2010, 39, (3), 1073-1095.

165

194. Solomon, B. R.; Khalil, K. S.; Varanasi, K. K., Drag Reduction using Lubricant- Impregnated Surfaces in Viscous Laminar Flow. Langmuir 2014, 30, (36), 10970-10976.

195. Grinthal, A.; Aizenberg, J., Mobile Interfaces: Liquids as a Perfect Structural Material for Multifunctional, Antifouling Surfaces. Chemistry of Materials 2014, 26, (1), 698-708.

196. Joanny, J. F.; Degennes, P. G., A Model for Contact-Angle Hysteresis. Journal of Chemical Physics 1984, 81, (1), 552-562.

197. Nadkarni, G. D.; Garoff, S., An Investigation of Microscopic Aspects of Contact-Angle Hysteresis - Pinning of the Contact Line on a Single Defect. Europhysics Letters 1992, 20, (6), 523-528.

198. Lafuma, A.; Quere, D., Slippery pre-suffused surfaces. Epl 2011, 96, (5).

199. Yang, X.; Li, C. Y.; Sun, Y., From multi-ring to spider web and radial spoke: competition between the receding contact line and particle deposition in a drying colloidal drop. Soft Matter 2014, 10, (25), 4458-4463.

200. Keville, K. M.; Franses, E. I.; Caruthers, J. M., Preparation and Characterization of Monodisperse Polymer Microspheroids. Journal of Colloid and Interface Science 1991, 144, (1), 103-126.

201. Ho, C. C.; Keller, A.; Odell, J. A.; Ottewill, R. H., Monodisperse Ellipsoidal Polystyrene Latex-Particles - Preparation and Characterization. Polymer International 1993, 30, (2), 207-211.

202. Ho, C. C.; Keller, A.; Odell, J. A.; Ottewill, R. H., Preparation of Monodisperse Ellipsoidal Polystyrene Particles. Colloid and Polymer Science 1993, 271, (5), 469-479.

203. Pethig, R., Review Article-Dielectrophoresis: Status of the theory, technology, and applications. Biomicrofluidics 2010, 4, (2).

204. Zheng, B.; Tice, J. D.; Ismagilov, R. F., Formation of droplets of in microfluidic channels alternating composition and applications to indexing of concentrations in droplet-based assays. Analytical Chemistry 2004, 76, (17), 4977-4982.

205. Chen, H.; Fang, Q.; Yin, X.-F.; Fang, Z.-L., Microfluidic chip-based liquid-liquid extraction and preconcentration using a subnanoliter-droplet trapping technique. Lab on a Chip 2005, 5, (7), 719-725.

206. Nie, Z. H.; Seo, M. S.; Xu, S. Q.; Lewis, P. C.; Mok, M.; Kumacheva, E.; Whitesides, G. M.; Garstecki, P.; Stone, H. A., Emulsification in a microfluidic flow-focusing device: effect of the viscosities of the liquids. Microfluidics and Nanofluidics 2008, 5, (5), 585-594.

166

207. Hufnagel, H.; Huebner, A.; Gulch, C.; Guse, K.; Abell, C.; Hollfelder, F., An integrated cell culture lab on a chip: modular microdevices for cultivation of mammalian cells and delivery into microfluidic microdroplets. Lab on a Chip 2009, 9, (11), 1576-1582.

208. Huebner, A.; Bratton, D.; Whyte, G.; Yang, M.; deMello, A. J.; Abell, C.; Hollfelder, F., Static microdroplet arrays: a microfluidic device for droplet trapping, incubation and release for enzymatic and cell-based assays. Lab on a Chip 2009, 9, (5), 692-698.

209. Zheng, B.; Roach, L. S.; Ismagilov, R. F., Screening of protein crystallization conditions on a microfluidic chip using nanoliter-size droplets. Journal of the American Chemical Society 2003, 125, (37), 11170-11171.

210. Zheng, B.; Tice, J. D.; Ismagilov, R. F., Formation of Arrayed Droplets by Soft Lithography and Two-Phase Fluid Flow, and Application in Protein Crystallization. Advanced Materials 2004, 16, (15), 1365-1368.

211. Chen, D. L.; Gerdts, C. J.; Ismagilov, R. F., Using microfluidics to observe the effect of mixing on nucleation of protein crystals. Journal of the American Chemical Society 2005, 127, (27), 9672-9673.

212. Yang, C. G.; Xu, Z. R.; Wang, J. H., Manipulation of droplets in microfluidic systems. Trac-Trends in Analytical Chemistry 2010, 29, (2), 141-157.

167

VITA

XinYang

EDUCATION

 Ph.D. Mechanical Engineering, Drexel University, Philadelphia, PA, December 2014.

 M.S. Refrigeration and Cryogenics Engieering, Xi’an Jiaotong University, China, June 2011.

 B.S. Thermal and Power Engineering, Xi’an Jiaotong University, China, June 2008

SELECTED PUBLICATIONS 1. Yang, X., Li C., Sun Y., From multi-ring to spider web and radial spoke: competition between the receding contact line and particle deposition in a drying colloidal drop, Soft matter 10.25 (2014): 4458-4463. 2. Yang, X., Chhasatia V., Sun Y., Oscillation and recoil of single and consecutively printed droplets, Langmuir 29.7 (2013): 2185-2192. 3. Yang, Xin, Chhasatia V., Shah J., Sun Y., Coalescence, evaporation and particle deposition of consecutively printed colloidal drops, Soft matter 8.35 (2012): 9205-9213. 4. Yang, X., Feng J., Chang Y., Peng X., Experimental Study of Oil-Gas Cyclone Separator Performance in Oil-Injection Compressor System, Advanced Materials Research 383 (2012): 6436-6442. 5. Feng, J., Yang X., Gao X., Chang, Y., Numerical Simulation and Experiment on Oil-Gas Cyclone Separator, Journal of Xi'an Jiaotong University 7 (2011): 012. 6. Gao, X., Feng, J., Yang, X, Chang, Y., Peng, X., The research on the performance of oil-gas cyclone separators in oil injected compressor systems with considering the collision and breakup of oil droplets, Proceedings of International Compressor Engineering Conference at Purdue: 2012, NO. 1407.

7. Chang Y., He B., Yang X., Ma Y., Simulation and development of trans-critical CO2 rolling piston compressor, Proceedings of International Compressor Engineering Conference at Purdue: 2010, NO. 1253.