Focal Plane Shift Imaging for the Analysis of ... - ACS Publications

Jul 22, 2016 - Page 1 ..... measure the radii and the center of mass of the initial droplets (Figure 2b) and after coalescence to check whether the in...
2 downloads 15 Views 6MB Size
This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.

Focal Plane Shift Imaging for the Analysis of Dynamic Wetting Processes Hyeongyun Cha,†,‡ Jae Min Chun,† Jesus Sotelo,† and Nenad Miljkovic*,†,‡ †

Department of Mechanical Science and Engineering, University of Illinois at Urbana−Champaign, Urbana, Illinois 61801, United States ‡ International Institute for Carbon Neutral Energy Research (WPI-I2CNER), Kyushu University, 744 Moto-oka, Nishi-ku, Fukuoka 819-0395, Japan S Supporting Information *

ABSTRACT: Droplet−surface interactions are common to a plethora of natural and industrial processes due to their ability to rapidly exchange energy, mass, and momentum. Droplets are particularly of interest due to their large surface-to-volume ratios and hence enhanced transport properties. For example, coalescence-induced droplet jumping on superhydrophobic surfaces has recently received much attention for its potential to enhance heat transfer, anti-icing, and self-cleaning performance by passively shedding microscale water droplets. To study droplet jumping, researchers typically use a two-camera setup to observe the out-ofplane droplet motion, with limited success due to the inability to resolve the depth dimension using two orthogonal cameras. Here we develop a single-camera technique capable of providing three-dimensional (3D) information through the use of focal plane manipulation, termed “focal plane shift imaging” (FPSI). We used FPSI to study the jumping process on superhydrophobic surfaces having a wide range of structure length scales (10 nm < l < 1 μm) and droplet radii (3 μm < R < 160 μm). We benchmarked the FPSI technique and studied the effects of droplet mismatch, multidroplet coalescence, and multihop coalescence on droplet jumping speed. Furthermore, we were able to resolve the full 3D trajectory of multiple jumping events, to show that, unlike previously theorized, the departure angle during droplet jumping is not a function of droplet mismatch or number of droplets coalescing prior to jumping. Rather, angular deviation arises due to in-plane motion postcoalescence governed by droplet pinning. The outcomes of this work both elucidate key fundamental aspects governing droplet jumping and provide a powerful imaging platform for the study of dynamic droplet processes that result in out-of-plane motion such as sliding, coalescence, or impact. KEYWORDS: focal plane shift imaging, jumping droplet, coalescence, condensation, droplet, heat transfer, superhydrophobic, hydrophobic, nanostructure

W

not capable of obtaining the initial droplet radii and jumping velocity, respectively. Furthermore, simultaneous two-camera imaging requires the jumping droplet trajectory to be in plane with the side-view camera, in addition to being costly and complex to implement. Side-view imaging using the depth-ofdefocus technique suffers similar limitations to the two-camera method; mainly droplet coalescence on the surface prior to departure must reside in the focal plane of the side-view camera in order to obtain initial conditions.40 The lack of successful imaging techniques has held back the fundamental understanding of droplet jumping, making it difficult to determine

hen two or more water droplets coalesce on a suitably designed superhydrophobic surface, the resulting droplet can jump away from the surface due to inertial−capillary energy conversion. The speed of the droplet departure scales with ∼ σ /ρR , where σ, ρ, and R are the surface tension, density, and the radii of the coalescing droplets prior to jumping, respectively.1−8 Droplet jumping on superhydrophobic surfaces has received significant attention due to its ability to fundamentally advance state-of-the-art technologies and enhance the performance of a variety of applications such as thermal diodes,9 anti-icing,10−13 selfcleaning,8,14−16 vapor chambers,17 energy harvesting,18−20 and condensation heat transfer.21−39 To characterize the jumping process, past experiments have utilized side view,40−44 top view,45−50 or simultaneous orthogonal two-camera imaging.2,51 Although qualitatively powerful, side and top view imaging are © 2016 American Chemical Society

Received: June 10, 2016 Accepted: July 22, 2016 Published: July 22, 2016 8223

DOI: 10.1021/acsnano.6b03859 ACS Nano 2016, 10, 8223−8232

Article

www.acsnano.org

Article

ACS Nano what governs the departure angle and how to rationally design surfaces to attain control of both jumping speed and direction. In this work, we develop a rather simple single-camera optical imaging technique to study the 3D trajectory of droplet jumping through focal plane manipulation (FPSI, focal plane shift imaging). We take advantage of the temporal lag prior to coalescence and departure of droplets, by initially observing the droplets prior to coalescence in plane with the surface, then shifting the imaging focal plane above the droplets and observing the departed droplet as it moves through the shifted plane. By calculating the initial center-of-mass of the droplets prior to coalescence and measuring the time taken for the departed droplet to pass through the shifted focal plane, we can determine the 3D droplet trajectory. To validate the accuracy and precision of the FPSI technique, we used it to study the jumping speed on superhydrophobic surfaces having a wide range of structure length scales (10 nm < l < 1 μm) and jumping droplet radii (3 μm < R < 160 μm), showing excellent agreement with previously reported studies utilizing twocamera techniques.2 After verifying the accuracy of FPSI, we studied the effects of initial droplet size mismatch and multiple droplet coalescence on the jumping droplet velocity, showing that multidroplet jumping has the potential to enhance the droplet departure speed. Furthermore, we show that, unlike previously predicted, the departure angle during jumping is weakly correlated to droplet mismatch or number of droplets coalescing. Rather, angular deviation from the surface normal arises due to in-plane motion governed by droplet pinning during coalescence. The outcomes of this work not only show the effects of droplet mismatch and multidroplet coalescence on the jumping velocity, but also establish a powerful imaging platform to probe whether dynamic wetting processes can be used in the future to provide avenues for improving the performance of technologies utilizing droplet motion.

Figure 1. Top-view high-resolution field emission scanning electron micrographs of the (a) carbon nanotube surface coated with a 30 nm thick layer of P2i fluoropolymer, (b) aluminum oxide surface coated with a monolayer ( 7.5 μm resulted in immeasurable droplet radii changes of ∼100 nm in the time taken (∼2−5 s) between the measurement of initial conditions (Figure 2b) and jumping (Figure 2c,d). The required temporal resolution of the FPSI technique is contingent on the size of coalescing droplets and their jumping speed. In order to minimize the temporal error to 250 coalescence events) and determined to correlate with departing droplet radius and number of coalescing droplets (see Supporting Information, Section S3). The droplet time-of-flight was then calculated by subtracting the touch-to-liftoff time, τ, from the total time measured from touch to crossing of the jumping droplet through the shifted plane (Figure 2d). It is important to note the droplet growth rate on the superhydrophobic surfaces examined here was not fast enough to create a measurable 8225

DOI: 10.1021/acsnano.6b03859 ACS Nano 2016, 10, 8223−8232

Article

ACS Nano speeds exceeding 400 000 frames/s are required. For larger droplets (∼50 μm) having slower departure speeds (∼0.2 m/s), the required capture rates decrease to 40 000 frames/s. To verify the accuracy of our approach, we first used the FPSI technique to study the two-droplet jumping speed and benchmark our result to previous studies.2 Jumping Droplet Speed. Figure 3 shows elevated-focalplane time-lapse images of jumping caused by the coalescence of two equally sized droplets having radii R ≈ 17.4 ± 0.5 μm (see Supporting Information, Video S1). Although the image of the droplets prior to coalescence on the surface is blurry (out of focus), the time when the droplets touch can be identified by the disappearance of the blurred droplets (to within ±1 frame). The jumping speed was calculated by measuring the droplet travel distance and dividing by the time-of-flight. To calculate the travel distance, the center of mass of the coalescing droplet pair was obtained from the in-plane image by applying the conservation of mass to calculate the departing droplet radius. To rigorously verify that the calculated center of mass can be used as an initial location, we experimentally showed that the calculated center of mass just prior to coalescence matches well with the experimentally measured center of mass of the droplet immediately after coalescence. Interestingly, the calculated center of mass matched the experimental value even for coalescence between droplets having substantial size differences (mismatch), where Laplace pressure generated flows from the smaller to larger droplet are present (see Supporting Information, Section S4 and Video S2). The end-point in the trajectory analysis of the jumping droplet was obtained when the middle of the droplet (maximum radius) passed through the elevated focal plane, which is deemed appropriate considering that droplets quickly reach an equilibrium spherical shape after coalescence and departure due to the low droplet bond and capillary numbers (Bo = 1.4 × 10−5 and Ca = 8.7 × 10−3 for Rj = 10 μm). Furthermore, shifting of the focal plane at least 2 droplet diameters above the initial plane ensured that coalescence-induced capillary wave oscillations had enough time to dampen6 and were minimally present as the droplet passed through the elevated focal plane. To verify that only the observed droplets partook in the jumping process, and to eliminate the possibility of additional hidden droplets beneath the projection of larger droplets partaking in coalescence,39 we calculated the final droplet diameter from the initial droplet arrangement using the conservation of mass and ensured that it has excellent agreement ( 3 (red square), respectively. The dashed lines and shaded area in (a) represent the two-droplet analytical solution and numerical simulation (ref 2), respectively. Departing droplet radius was calculated by applying the conservation of mass from the experimentally measured radius of the initial droplets. The results show that the majority of experimentally measured jumping droplet speeds are bounded between the numerical prediction and that the speeds increase as the number of droplets increases, as expected. For n = 2, major speed reduction occurred when mismatch in radius of two initial droplets exceeded 40% due to the antisymmetric momentum transfer (red star). Error bars for droplet radius are not shown due to their being smaller than each data point. 8226

DOI: 10.1021/acsnano.6b03859 ACS Nano 2016, 10, 8223−8232

Article

ACS Nano

Figure 5. (a) Schematic defining the polar and azimuthal angles in three dimensions. Polar angle as a function of initial droplet size mismatch where the number of initial droplets are (b) 2 (inset: schematic of two-jumping-droplet coalescence with jumping droplet angle β), (c) 3 (inset: schematic of multiple jumping droplet coalescence with jumping droplet angle β), and (d) > 3. Azimuthal angle for n = 2 as a function of (e) initial droplet size mismatch and (f) departing droplet radius. Each data point corresponds to the jumping droplet speed data in Figure 4. Jumping droplet angle was measured by calculating the angle between initial and final location of the droplets with respect to the surface. The center of masses of initial droplets were measured with in-plane imaging (Figure 2b), and the center of mass of the final droplet was measured as the middle of the jumping droplet passes through the shifted focal plane (Figure 2d). Experimental results show that ≈86% of jumping droplets for n = 2 have β > 80° in spite of droplet size mismatch, indicating that droplet size mismatch prior to coalescence is not the factor governing angular deviation from the surface normal.

FPSI technique easily resolved the initial conditions prior to jumping and showed that events having lower than expected departure speeds were due to droplet size mismatch. Particularly, droplets having a difference in radii of more than 40% (red stars) were susceptible to lower surface-to-kinetic energy conversion due to the lower surface-energy-to-volume ratio.7 Furthermore, droplet jumping on each surface was found to be structure length scale invariant, showing identical departure speed behavior for the CNT, Al2O3, and CuO structured samples (see Supporting Information, Section S5). Although the minimum jumping droplet departure size is governed by the structure length scale and contact angle hysteresis, the jumping droplet sizes characterized here (R > 3 μm) were well above the adhesion limit for jumping on microstructures (∼1 μm) having low-contact-angle hysteresis.53 After demonstrating the validity of FPSI using two-droplet coalescence, we studied the coalescence between more than two droplets, n > 2. Since the nucleation process is spatially random, jumping stemming from the coalescence of three or more droplets is frequent and remains a poorly understood phenomena. Figure 4b and c plot the jumping speed as a function of jumping droplet radius for three-droplet coalescence (n = 3, green triangles) and coalescence between more than three droplets (n > 3, red squares), respectively (see

Supporting Information, Video S3). Due to the enhanced surface-energy-to-volume ratio of three-droplet coalescence, the jumping speed was observed to increase for the entire range of departure radii studied here (3 μm < R < 150 μm). Interestingly, the results of three-droplet coalescence indicate that the capillary-to-inertial energy conversion mechanism is applicable. However, droplet coalescence between four or more droplets (Figure 4c) did not have the same enhancement as predicted by theory, showing higher scatter in the data. The lower than expected jumping speeds for n > 3 can be attributed to the serial nature of multidroplet coalescence. Depending on the initial position of droplets on the surface, the coalescence event may not proceed instantaneously among all droplets, as may be expected for two and three droplets. Coalescence may be initiated with two droplets and continue in a serial manner with other neighboring droplets prior to leaving the surface (see Supporting Information, Video S4). The internal flow field of this serial process can create out-of-phase momentum components and result in a decreased net momentum in the surface-normal direction depending on the droplet center-tocenter spacing, sizes, and initial positions. Note that the larger scatter for n > 3 cannot be attributed to a fundamental difference in the capillary-to-inertial energy conversion mechanism compared to the two-droplet case (Figure 4a). 8227

DOI: 10.1021/acsnano.6b03859 ACS Nano 2016, 10, 8223−8232

Article

ACS Nano

Figure 6. (a) Time-lapse in-plane imaging of multidroplet pinned motion. The images show six droplets with radii of 10.5, 10.3, 10.3, 7.3, 7.3, and 5.6 μm, coalescing and forming a final departing droplet of radius 16.3 μm. The initially coalesced droplets did not undergo jumping; rather they continued coalescing with neighboring droplets, eventually jumping from the surface. The measured horizontal component of the jumping droplet speed was 0.34 m/s, while the average horizontal jumping speed in Figure 4 was 0.04 m/s. The focal plane was focused on the center of the biggest droplet, and images were captured at a frame rate of 79 069 frames/s. Schematics of pinned droplet coalescence during (b) bridge formation with bridge velocities v−b and v+b and (c) just before departing with reflected momentum P−r . Droplet pinning at the liquid−solid interface (depicted by the × symbol) acts to generate a moment, M, in the horizontal plane and results in surface-to-kinetic energy transfer non-normal to the surface. (d) Schematic depicting the cancellation of non-normal momentum components (P−x and P+x ) during coalescence of two nonequal radii droplets (R2 > R1).

mismatch (100R1/R2, where R1 and R2 are the smallest and largest droplets, respectively) for 2, 3, and n coalescing droplets, respectively. Each data point corresponds to a jumping event characterized in the speed measurements shown in Figure 4. Interestingly, the experimental results for n = 2 show that more than 86% of jumping events have β > 80°, or approximately normal to the symmetry-breaking surface. Even cases of coalescence with droplets having radii mismatch of >40% resulted in jumping at polar angles β > 80°. The experimental data imply that, unlike previously theorized, droplet size mismatch prior to coalescence is not the factor governing angular deviation. For cases where large angular deviation did occur (β < 70°), we observed that droplet movement was initiated prior to droplet coalescence (see Supporting Information, Video S5). To study the mechanism further, we characterized β for n = 3 (Figure 5c) and n > 3 (Figure 5d), where the droplet size mismatch was calculated as the percent ratio of the smallest to largest droplets in the coalescing set. The experimental polar angle data showed larger scatter when compared to the n = 2 case (Figure 5a), with little correlation to radius mismatch. Similar to the two-droplet case, higher angular deviation events were observed to result when motion prior to jumping was noticed in the initial focal plane. The larger scatter in the multidrop data did not indicate any particular trend, but rather implicated serial coalescence as the primary mechanism for lateral motion generation prior to jumping.

The validity of the energy conversion mechanism is in quantitative agreement for the case of n = 3 and does not show any effect on initial droplet orientation prior to coalescence, indicating that the surface-mediated symmetry breaking and internal flow momentum generation in the surface-normal direction is scalable up to n droplets coalescing instantaneously (see Supporting Information, Section S6). The ability to observe the coalescence event from above the focal plane enables FPSI to accurately characterize jumping initiated by returning droplets to the surface due to external body or surface forces (Figure 4b,c, multihop, orange inverse triangles). Similar to the coalescence of nonequal radii droplets, multihop jumping events showed a decrease in jumping speed due to the adverse momentum of returning droplets.7 Jumping Droplet Angle. After verifying the fidelity of the FPSI technique to characterize jumping speeds for a variety of droplet configurations, we next investigated the governing mechanisms of jumping droplet angular deviation from the surface normal, characterized by the polar angle, β, and azimuthal angle, α (Figure 5a). Previous studies have hypothesized that angular deviation is governed by droplet mismatch during coalescence,2,40,54 with no experimental data to verify the theory due to a lack of imaging techniques. Here, the polar angle was measured by calculating the angle between initial and final location of the droplets with respect to the surface (insets of Figure 5). Figure 5b, c, and d show the jumping droplet polar angle as a function of droplet size 8228

DOI: 10.1021/acsnano.6b03859 ACS Nano 2016, 10, 8223−8232

Article

ACS Nano Figure 5e and f show the jumping droplet azimuthal angle, α, as a function of droplet radius mismatch and two-droplet jumping radius, Rj, respectively. The azimuthal angle was defined as the angle between the plane defined by the surface normal and the vector connecting the center of mass of the bigger droplet, R2, to the departed droplet, Rj, with respect to the plane formed by the surface normal and vector connecting the center of masses of the initial coalescing droplet pair (Figure 5a). Values of α < 90° indicate that the droplet departed in the quarter-sphere containing the bigger droplet. Similar to the polar angle, the experimental results show that more than 86% of jumping events have either α > 150° or α < 30°, implying that droplets jump in the plane defined by the surface normal vector and vector connecting the droplet centers prior to coalescence. Furthermore, 42.5% of jumping events occurred with α > 150°, while 43.5% of jumping events occurred with α < 30°, indicating that the azimuthal direction of the droplet jumping is independent of both radii mismatch and jumping droplet radius. Multidroplet Pinned Motion. To better understand the mechanism of the random observation of droplet jumping with large polar angle deviations (β < 70°), we studied the multidrop coalescence process via classical high-speed optical microscopy in the initial focal plane only. We observed that for particular multidrop coalescence events the first serial coalescence event did not result in normal motion to the surface (jumping); rather the expansion of the radial liquid bridge during coalescence caused the droplet to touch and initiate coalescence with a tertiary nearby neighbor droplet (see Supporting Information, Video S6). The serial coalescence process continued until the droplet coalesced with all nearby neighbors, prior to departing the surface. Figure 6a shows timelapse images of one particular multidrop coalescence event with n = 6. The average horizontal speed of the droplet was determined to be ≈0.34 ± 0.1 m/s during the serial coalescence process. However, the average horizontal component speed of the jumping droplet data set in Figure 4 was only 0.04 m/s. The large discrepancy in horizontal speed components between the observed event (Figure 6a) and the jumping speed data (Figure 4) implicates pinning and in-plane energy release as the mechanism governing polar angle deviation rather than serial coalescence. Indeed, the pinning of droplets to the surface is common, and not all coalescence events result in droplet departure. From a statistical standpoint, the pinning mechanism is very consistent with the observed data. Assuming a constant probability of droplets residing on the surface that are pinned in specific locations at or near their contact line, multidrop coalescence events will have a higher probability of including one of these pinned droplets in their coalescence cascade. The larger probability manifests itself as higher scatter for the multidrop data when compared to two- or three-droplet data, consistent with the experimental observations of Figure 5. Furthermore, the random nature of large polar angle deviation events for the n = 2 cases (Figure 5b) suggests that serial coalescence or radius mismatch cannot be attributed to droplets jumping with low polar angles, but rather, pinning at the liquid−solid interface acts as a pivot, transferring the generated vertical force into a moment of force, M, and propelling the droplet in the horizontal plane (Figure 6b and c). Interestingly, angular deviation was not preferential on microscale surfaces when compared to nanoscale surfaces (see Supporting Information, Section S5). Although the measured contact angle hysteresis of the CNT surface (∼11°) was lower than that

of the CuO surface (∼16°), preference for contact line pinning remained similar. The pinning-induced lateral motion was further confirmed by the independence of the droplet polar departure angle on the multidrop arrangement prior to coalescence, i.e., linear, staggered, random (see Supporting Information, Section S6). Indeed, theoretical consideration of coalescence between nonequal radii droplets on zero adhesion surfaces reveals that any non-normal momentum components cancel each other out and droplet motion will be in the surface normal direction. Immediately after the interfaces of the two droplets make contact, the developing liquid bridge is accelerated radially from the point of contact due to the curvature difference (in the longitudinal direction) between the bridge radius (1/rb) and the droplets (1/R). Even for the case of droplets having different initial radii, this process will be present at early time scales. The components of momentum generated away from the coalescence plane will travel either away from the surface (P+) or toward the surface (P−). Assuming that the component of momentum traveling toward the surface, P−, reflects in a specular fashion, the in-plane momentum components, P−x and P+x , will cancel each other out (Figure 6d). From a purely theoretical standpoint, momentum conservation on a zero adhesion surface necessitates that the only way to generate a net droplet momentum postcoalescence is if the momentum is normal to the surface, as shear forces due to friction and pinning are nonexistent. Following this argument, we can see that in the limit of no pinning (zero adhesion), no lateral (inplane) momentum generation is possible even in the presence of Laplace pressure generated flows between nonequal radii droplets. As in the case of nonequal droplet coalescence in free space, Laplace pressure generated flows cannot generate a net momentum unless an external, nonsymmetric force is present. The inability to generate lateral momentum due to Laplace pressure generated flows corroborates our observation of twodroplet coalescence and jumping normal to the surface (β ≈ 90°) with large radii mismatch (>40%), as pinning of the contact line was not present for these observed events. In the future, it would be interesting to use the FPSI technique not only as a tool to study the droplet jumping process but as a development platform for structured surface geometries and droplet morphologies, which may enable directional pinning and passive or active control of droplet jumping direction. Furthermore, it would be interesting to utilize the FPSI technique to fundamentally answer questions pertaining to other problems involving dynamic motion of droplets such as sliding, coalescence, or impact. For example, the ability to quantitatively resolve the liquid−vapor interface during sliding as a droplet moves through the focal plane is a powerful tool to analyze the contact line shape and verify theoretical models. Furthermore, focal plane manipulation, in combination with careful spatial calibration, forms the basis of a microgoniometer capable of studying the evolved droplet morphology on smooth and structured surfaces as a function of droplet length scale.

CONCLUSIONS In summary, we developed a single-camera optical imaging technique for the study of microscale dynamic droplet processes using focal plane shift imaging. We demonstrated the FPSI technique by studying the 3D trajectory of coalescence-induced droplet jumping on superhydrophobic surfaces having a wide range of structure length scales (10 nm < 8229

DOI: 10.1021/acsnano.6b03859 ACS Nano 2016, 10, 8223−8232

Article

ACS Nano l < 1 μm) and over a wide range of droplet jumping radii (3 μm < R < 160 μm). We used the FPSI technique to study the effects of initial droplet size mismatch and multiple droplet coalescence on the jumping droplet velocity, showing that multidroplet jumping has the potential to enhance the droplet departure speed. Furthermore, we showed that, unlike previously predicted, the departure angle during jumping is weakly correlated to droplet mismatch. Rather, angular deviation from the surface normal arises due to in-plane motion governed by droplet pinning during coalescence. The outcomes of this work not only elucidate fundamental aspects of droplet jumping dynamics but also establish a powerful imaging platform to probe dynamic wetting processes that can be used in the future to provide avenues for improving the performance of technologies utilizing droplet motion.

silane coating. Goniometric measurements (MCA-3, Kyowa Interface Science Ltd.) of ≈100 nL droplets on a smooth P2i-coated silicon wafer surface showed advancing contact angles of θa = 115 ± 2° and θr = 107 ± 3°, respectively. Surface Characterization. Contact angle measurements of ≈100 nL droplets on all samples were performed using a microgoniometer (MCA-3, Kyowa Interface Science). Field emission scanning electron micrographs (Hitachi model S-4800) were performed on all samples at an imaging voltage range of 3.0 to 5.0 kV. Experimental Setup. Jumping droplet condensation experiments were performed using a customized top-view optical light microscopy setup (Figure 2). A high-speed camera (Phantom, V711, Vision Research) was attached to the upright microscope (Eclipse LV100, Nikon) for top-view analysis, performing video recordings at variable frame rates up to 500 000 frames per second. Samples were horizontally mounted to a cold stage (TP104SC-mK2000A, Instec) using a thin layer of thermal grease (Omegatherm, Omega, thermal conductivity of 2.2 W/m·K) and cooled to the test temperature of Tw = 1 ± 0.5 °C, in a laboratory environment having air temperature Tair = 22 ± 0.5 °C and relative humidity ϕ = 28 ± 1% (Roscid Technologies, RO120). Illumination was supplied by an LED light source (SOLA SM II Light Engine, Lumencor). The LED light source was specifically chosen for its high-intensity, low-power consumption (2.5 W) and narrow spectral range (380−680 nm) in order to minimize heat generation at the surface due to light absorption. Furthermore, by manually reducing the condenser aperture diaphragm opening size and increasing the camera exposure time, we were able to minimize the amount of light energy needed for illumination and hence minimize local heating effects during condensation experiments. Imaging was performed with a 20× (TU Plan Fluor EPI, Nikon), 50× (TU Plan Fluor EPI, Nikon), or 100× (TU Plan Fluor EPI, Nikon) objective. All experiments were conducted at supersaturations, S = 1.02 ± 0.035, below the critical supersaturation, Sc ≈ 1.12, associated with surface flooding conditions for superhydrophobic surfaces (S = Pv/ Psat(Tw) < Sc). This was done in order to remain in the droplet jumping regime to study the coalescence and departure dynamics.

METHODS Surface Fabrication. CNTs were grown by chemical vapor deposition (CVD). Using electron beam deposition, a 20 nm thick Al2O3 diffusion barrier and a 5 nm thick film of Fe catalyst layer were deposited sequentially on silicon growth substrates. CNT growth was performed in a 2.54 cm quartz furnace tube. After purging in He/H2 atmosphere for 15 min, the silicon substrate was heated to 750 °C furnace temperature and annealed for 3 min under a flow of H2 and He at 400 and 100 sccm, respectively. Randomly aligned CNTs were then grown by flowing C2H4 at 200 sccm for 1 min. The thermally grown CNT had a typical outer diameter of d ≈ 7 nm. To create an Al2O3 nanostructured surface, the aluminum (Al) tabs (99.0% purity, 25 mm × 25 mm × 0.8 mm) were used as the test samples for the experiments. Each Al tab was ultrasonically cleaned in acetone (C3H6O, 99.0%) and ethanol (C2H5OH, 99.0%) for 5 min, followed by drying with clean nitrogen flow. Nanostructured Al2O3 surfaces with sharp, knife-like structures having length scales approaching ≈300 nm were formed by immersing tabs into hot water (90 °C) for an hour. To create the CuO nanostructured surface, commercially available copper (Cu) tabs (99.90% purity, 25 mm × 25 mm × 0.8 mm) were used as the test samples for the experiments. Each Cu tab was thoroughly rinsed with acetone, ethanol, isopropyl alcohol (IPA), and deionized (DI) water. The tabs were then dipped into a 5.0 M hydrochloric acid solution for 2 min to remove the native oxide film on the surface, rinsed with DI water, and dried with clean argon gas. Nanostructured CuO films were formed by immersing the cleaned tabs in a hot (96 ± 3 °C) alkaline solution composed of NaClO2, NaOH, Na3PO4·12H2O, and DI water (3.75:5:10:100 wt %). During the oxidation process, a thin (≈300 nm) Cu2O layer was formed and then reoxidized to form sharp, knife-like CuO oxide structures with heights of h ≈ 1 μm. Surface Functionalization. The P2i hydrophobic coating (Figure 1a and c) was achieved with plasma-enhanced vapor deposition. The process occurs under low pressure within a vacuum chamber at room temperature. The coating is introduced as a vapor and ionized. This process allows for the development of a highly conformal (≈30 nm thick) polymer layer, which forms a covalent bond with the CNT and CuO surfaces, making these extremely durable. Goniometric measurements (MCA-3, Kyowa Interface Science Ltd.) of ≈100 nL droplets on a smooth P2i-coated silicon wafer surface showed advancing and receding contact angles of θa = 124.3 ± 3.1° and θr = 112.6 ± 2.8°, respectively. Heptadecafluorodecyltrimethoxy-silane (CAS no. 83048-65-1) was deposited using vapor deposition (Figure 1b). Before deposition, each Al2O3 tab was dried in a clean N2 flow. Once dried, multiple Al2O3 tabs and a 50 mL beaker containing 1 mL of HTMS toluene (C7H8, 98.5%) solution (5 vol %) were sealed in a glass container and placed in an atmospheric pressure oven (Lindberg Blue M) at 80 °C for 3 h. The glass container was then allowed to naturally cool to room temperature for 1 h. The HTMS vapor phase depositions allowed for the development of a highly conformal (ca. monolayer thick) hydrophobic

ASSOCIATED CONTENT S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsnano.6b03859. Further information about the FPSI protocols and design guidelines, touch-to-liftoff time, initial center of mass, and independence of the droplet departure behavior on surface structure length scale and initial droplet arrangement (PDF) Six videos showing the water vapor condensation processes and droplet jumping, Video S1 (AVI) Video S2 (AVI) Video S3 (AVI) Video S4 (AVI) Video S5 (AVI) Video S6 (AVI)

AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Author Contributions

H.C. and N.M. conceived the initial idea for this research. N.M. guided the work. H.C., J.M.C., and J.S. fabricated, functionalized, and characterized the experimental samples and carried out the experiments. H.C. analyzed the data. H.C. and N.M. carried out the theoretical analysis. All authors were responsible for writing the paper and have given approval to the final version of the manuscript. 8230

DOI: 10.1021/acsnano.6b03859 ACS Nano 2016, 10, 8223−8232

Article

ACS Nano Notes

(16) Wang, Q.; Yao, X.; Liu, H.; Quéré, D.; Jiang, L. Self-Removal of Condensed Water on the Legs of Water Striders. Proc. Natl. Acad. Sci. U. S. A. 2015, 112, 9247−9252. (17) Boreyko, J. B.; Chen, C. H. Vapor Chambers with JumpingDrop Liquid Return from Superhydrophobic Condensers. Int. J. Heat Mass Transfer 2013, 61, 409−418. (18) Preston, D. J.; Miljkovic, N.; Wang, E. N.; Enright, R. Jumping Droplet Electrostatic Charging and Effect on Vapor Drag. J. Heat Transfer 2014, 136, 080909−080909. (19) Miljkovic, N.; Preston, D. J.; Enright, R.; Wang, E. N. JumpingDroplet Electrostatic Energy Harvesting. Appl. Phys. Lett. 2014, 105.01311110.1063/1.4886798 (20) Miljkovic, N.; Preston, D. J.; Enright, R.; Wang, E. N. Electrostatic Charging of Jumping Droplets. Nat. Commun. 2013, 4.10.1038/ncomms3517 (21) Enright, R.; Miljkovic, N.; Alvarado, J. L.; Kim, K.; Rose, J. W. Dropwise Condensation on Micro- and Nanostructured Surfaces. Nanoscale Microscale Thermophys. Eng. 2014, 18, 223−250. (22) Miljkovic, N.; Wang, E. N. Condensation Heat Transfer on Superhydrophobic Surfaces. MRS Bull. 2013, 38, 397−406. (23) Enright, R.; Miljkovic, N.; Dou, N.; Nam, Y.; Wang, E. N. Condensation on Superhydrophobic Copper Oxide Nanostructures. J. Heat Transfer 2013, 135, 091304. (24) 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, 1776−85. (25) Miljkovic, N.; Enright, R.; Wang, E. N. Growth Dynamics During Dropwise Condensation on Nanostructured Superhydrophobic Surfaces. 3rd Micro/Nanoscale Heat & Mass Transfer International Conference, 2012. (26) Miljkovic, N.; Enright, R.; Wang, E. N. Modeling and Optimization of Superhydrophobic Condensation. J. Heat Transfer 2013, 135.42710.1115/1.4024597 (27) Cheng, J.; Vandadi, A.; Chen, C. L. Condensation Heat Transfer on Two-Tier Superhydrophobic Surfaces. Appl. Phys. Lett. 2012, 101, 131909-1−131909-4. (28) Olceroglu, E.; Hsieh, C. Y.; Rahman, M. M.; Lau, K. K. S.; McCarthy, M. Full-Field Dynamic Characterization of Superhydrophobic Condensation on Biotemplated Nanostructured Surfaces. Langmuir 2014, 30, 7556−7566. (29) Olceroglu, E.; King, S. M.; Rahman, M. M.; McCarthy, M. Biotemplated Superhydrophobic Surfaces for Enhanced Dropwise Condensation. Proceedings of the Asme International Mechanical Engineering Congress and Exposition - 2012; Vol 7, Pts A-D 2013, pp 2809−2815. (30) Attinger, D.; Frankiewicz, C.; Betz, A. R.; Schutzius, T. M.; Ganguly, R.; Das, A.; Kim, C.-J.; Megaridis, C. M. Surface Engineering for Phase Change Heat Transfer: A Review. MRS Energy Sustainability 2014, 1.10.1557/mre.2014.9 (31) Chen, X. M.; Weibel, J. A.; Garimella, S. V. Exploiting Microscale Roughness on Hierarchical Superhydrophobic Copper Surfaces for Enhanced Dropwise Condensation. Adv. Mater. Interfaces 2015, 2.n/a10.1002/admi.201570015 (32) Hou, Y. M.; Yu, M.; Chen, X. M.; Wang, Z. K.; Yao, S. H. Recurrent Filmwise and Dropwise Condensation on a Beetle Mimetic Surface. ACS Nano 2015, 9, 71−81. (33) Miljkovic, N.; Enright, R.; Nam, Y.; Lopez, K.; Dou, N.; Sack, J.; Wang, E. N. Jumping-Droplet-Enhanced Condensation on Scalable Superhydrophobic Nanostructured Surfaces. Nano Lett. 2013, 13, 179−87. (34) Birbarah, P.; Li, Z.; Pauls, A.; Miljkovic, N. A Comprehensive Model of Electric-Field-Enhanced Jumping-Droplet Condensation on Superhydrophobic Surfaces. Langmuir 2015, 31, 7885−7896. (35) Lo, C.-W.; Wang, C.-C.; Lu, M.-C. Spatial Control of Heterogeneous Nucleation on the Superhydrophobic Nanowire Array. Adv. Funct. Mater. 2014, 24, 1211−1217.

The authors declare no competing financial interest.

ACKNOWLEDGMENTS We gratefully acknowledge Professor Constantine M. Megaridis of UIC for fruitful discussions regarding the mechanism governing droplet jumping angular deviation. The authors gratefully acknowledge the funding support from the Office of Naval Research (ONR) with Dr. Mark Spector as the program manager. The authors gratefully acknowledge the support of the International Institute for Carbon Neutral Energy Research (WPI-I2CNER), sponsored by the Japanese Ministry of Education, Culture, Sports, Science and Technology. We are grateful to P2i for the hydrophobic layer depositions. We gratefully acknowledge funding support from the Air Conditioning and Refrigeration Center (ACRC), an NSFfounded I/UCRC at UIUC. Electron microscopy was carried out in part in the Frederick Seitz Materials Research Laboratory Central Facilities, University of Illinois. REFERENCES (1) Boreyko, J. B.; Chen, C. H. Self-Propelled Dropwise Condensate on Superhydrophobic Surfaces. Phys. Rev. Lett. 2009, 103, 184501. (2) Enright, R.; Miljkovic, N.; Sprittles, J.; Nolan, K.; Mitchell, R.; Wang, E. N. How Coalescing Droplets Jump. ACS Nano 2014, 8, 10352−10362. (3) Liu, F. J.; Ghigliotti, G.; Feng, J. J.; Chen, C. H. Numerical Simulations of Self-Propelled Jumping upon Drop Coalescence on Non-Wetting Surfaces. J. Fluid Mech. 2014, 752, 39−65. (4) Liu, F. J.; Ghigliotti, G.; Feng, J. J.; Chen, C. H. Self-Propelled Jumping upon Drop Coalescence on Leidenfrost Surfaces. J. Fluid Mech. 2014, 752, 22−38. (5) Nam, Y.; Kim, H.; Shin, S. Energy and Hydrodynamic Analyses of Coalescence-Induced Jumping Droplets. Appl. Phys. Lett. 2013, 103.16160110.1063/1.4825273 (6) Nam, Y.; Seo, D.; Lee, C.; Shin, S. Droplet Coalescence on Water Repellant Surfaces. Soft Matter 2015, 11, 154−160. (7) Kim, M.-K.; Cha, H.; Birbarah, P.; Chavan, S.; Zhong, C.; Xu, Y.; Miljkovic, N. Enhanced Jumping-Droplet Departure. Langmuir 2015, 31, 13452−13466. (8) Watson, G. S.; Schwarzkopf, L.; Cribb, B. W.; Myhra, S.; Gellender, M.; Watson, J. A. Removal Mechanisms of Dew via SelfPropulsion off the Gecko Skin. J. R. Soc., Interface 2015, 122014139610.1098/rsif.2014.1396. (9) Boreyko, J. B.; Zhao, Y. J.; Chen, C. H. Planar Jumping-Drop Thermal Diodes. Appl. Phys. Lett. 2011, 99, 234105. (10) Boreyko, J. B.; Collier, P. C. Delayed Frost Growth on JumpingDrop Superhydrophobic Surfaces. ACS Nano 2013, 7, 1618−1627. (11) Chen, X. M.; Ma, R. Y.; Zhou, H. B.; Zhou, X. F.; Che, L. F.; Yao, S. H.; Wang, Z. K. Activating the Microscale Edge Effect in a Hierarchical Surface for Frosting Suppression and Defrosting Promotion. Sci. Rep. 2013, 310.1038/srep02515. (12) Lv, J. Y.; Song, Y. L.; Jiang, L.; Wang, J. J. Bio-Inspired Strategies for Anti-Icing. ACS Nano 2014, 8, 3152−3169. (13) Zhang, Q. L.; He, M.; Chen, J.; Wang, J. J.; Song, Y. L.; Jiang, L. Anti-Icing Surfaces Based on Enhanced Self-Propelled Jumping of Condensed Water Microdroplets. Chem. Commun. 2013, 49, 4516− 4518. (14) Wisdom, K. M.; Watson, J. A.; Qua, X.; Liua, F.; Watson, G. S.; Chen, C. H. Self-Cleaning of Superhydrophobic Surfaces by SelfPropelled Jumping Condensate. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, 7992−7997. (15) Watson, G. S.; Gellender, M.; Watson, J. A. Self-Propulsion of Dew Drops on Lotus Leaves: A Potential Mechanism for Self Cleaning. Biofouling 2014, 30, 427−434. 8231

DOI: 10.1021/acsnano.6b03859 ACS Nano 2016, 10, 8223−8232

Article

ACS Nano (36) Rykaczewski, K. Microdroplet Growth Mechanism during Water Condensation on Superhydrophobic Surfaces. Langmuir 2012, 28, 7720−7729. (37) Rykaczewski, K.; Osborn, W. A.; Chinn, J.; Walker, M. L.; Scott, J. H. J.; Jones, W.; Hao, C.; Yaod, S.; Wang, Z. How Nanorough Is Rough Enough To Make a Surface Superhydrophobic during Water Condensation? Soft Matter 2012, 8, 8786−8794. (38) Rykaczewski, K.; Scott, J. H. J. Methodology for Imaging Nanoto-Microscale Water Condensation Dynamics on Complex Nanostructures. ACS Nano 2011, 5, 5962−5968. (39) Rykaczewski, K.; Scott, J. H. J.; Rajauria, S.; Chinn, J.; Chinn, A. M.; Jones, W. Three Dimensional Aspects of Droplet Coalescence during Dropwise Condensation on Superhydrophobic Surfaces. Soft Matter 2011, 7, 8749−8752. (40) Chen, X.; Patel, R. S.; Weibel, J. A.; Garimella, S. V. Coalescence-Induced Jumping of Multiple Condensate Droplets on Hierarchical Superhydrophobic Surfaces. Sci. Rep. 2016, 6, 18649. (41) Liu, T. Q.; Sun, W.; Sun, X. Y.; Ai, H. R. Mechanism Study of Condensed Drops Jumping on Super-Hydrophobic Surfaces. Colloids Surf., A 2012, 414, 366−374. (42) Peng, B.; Wang, S.; Lan, Z.; Xu, W.; Wen, R.; Ma, X. Analysis of Droplet Jumping Phenomenon with Lattice Boltzmann Simulation of Droplet Coalescence. Appl. Phys. Lett. 2013, 102, 151601. (43) Lv, C. J.; Hao, P. F.; Yao, Z. H.; Song, Y.; Zhang, X. W.; He, F. Condensation and Jumping Relay of Droplets on Lotus Leaf. Appl. Phys. Lett. 2013, 103, 021601. (44) Tian, J.; Zhu, J.; Guo, H.-Y.; Li, J.; Feng, X.-Q.; Gao, X. Efficient Self-Propelling of Small-Scale Condensed Microdrops by Closely Packed ZnO Nanoneedles. J. Phys. Chem. Lett. 2014, 5, 2084−2088. (45) Qu, X.; Boreyko, J. B.; Liu, F.; Agapov, R. L.; Lavrik, N. V.; Retterer, S. T.; Feng, J. J.; Collier, C. P.; Chen, C.-H. Self-Propelled Sweeping Removal of Dropwise Condensate. Appl. Phys. Lett. 2015, 106, 221601. (46) Lv, C.; Hao, P.; Yao, Z.; Niu, F. Departure of Condensation Droplets on Superhydrophobic Surfaces. Langmuir 2015, 31, 2414− 2420. (47) Ö lçeroğlu, E.; McCarthy, M. Self-Organization of Microscale Condensate for Delayed Flooding of Nanostructured Superhydrophobic Surfaces. ACS Appl. Mater. Interfaces 2016, 8, 5729−5736. (48) Rykaczewski, K.; Paxson, A. T.; Anand, S.; Chen, X.; Wang, Z.; Varanasi, K. K. Multimode Multidrop Serial Coalescence Effects during Condensation on Hierarchical Superhydrophobic Surfaces. Langmuir 2013, 29, 881−891. (49) 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. Appl. Phys. Lett. 2010, 97, 033104. (50) Chen, X.; Wu, J.; Ma, R.; Hua, M.; Koratkar, N.; Yao, S.; Wang, Z. Nanograssed Micropyramidal Architectures for Continuous Dropwise Condensation. Adv. Funct. Mater. 2011, 21, 4617−4623. (51) Yanagisawa, K.; Sakai, M.; Isobe, T.; Matsushita, S.; Nakajima, A. Investigation of Droplet Jumping on Superhydrophobic Coatings during Dew Condensation by the Observation from Two Directions. Appl. Surf. Sci. 2014, 315, 212−221. (52) Chavan, S.; Cha, H.; Orejon, D.; Nawaz, K.; Singla, N.; Yeung, Y. F.; Park, D.; Kang, D. H.; Chang, Y.; Takata, Y.; Miljkovic, N. Heat Transfer through a Condensate Droplet on Hydrophobic and Nanostructured Superhydrophobic Surfaces. Langmuir 2016, DOI: 10.1021/acs.langmuir.6b01903. (53) Cha, H.; Xu, C.; Sotelo, J.; Chun, J. M.; Yokoyama, Y.; Enright, R.; Miljkovic, N. Coalescence-Induced Nanodroplet Jumping 2016, in review. (54) Wang, F.-C.; Yang, F.; Zhao, Y.-P. Size Effect on the Coalescence-Induced Self-Propelled Droplet. Appl. Phys. Lett. 2011, 98, 053112.

8232

DOI: 10.1021/acsnano.6b03859 ACS Nano 2016, 10, 8223−8232