Article pubs.acs.org/Langmuir
Water Droplet Spreading and Wicking on Nanostructured Surfaces Xue Chen, Jiannan Chen, Xiaolong Ouyang, Yu Song, Ruina Xu, and Peixue Jiang* Key Laboratory for Thermal Science and Power Engineering of Ministry of Education, Key Laboratory for CO2 Utilization and Reduction Technology of Beijing, Department of Thermal Engineering, Tsinghua University, Beijing 100084, China S Supporting Information *
ABSTRACT: Phase-change heat transfer on nanostructured surfaces is an efficient cooling method for high heat flux devices due to its superior wettability. Liquid droplet spreading and wicking effect then dominate the heat transfer. Therefore, this study investigates the flow behavior after a droplet touches a nanostructured surface focusing on the ZnO nanowire surface with three different nanowire sizes and two array types (regular and irregular). The spreading diameter and the wicking diameter are measured against time. The results show that the average spreading and wicking velocities on a regular nanostructured surface are both smaller than those on an irregular nanostructured surface and that the nanowire size affects the liquid spreading and capillary wicking.
■
INTRODUCTION Ongoing technological advancements have increased the need for miniaturization, integration, and extreme power dissipation rates across numerous industries. The main technology bottleneck for industries such as electronics, aeronautics, and nuclear power is cooling for high heat flux devices.1 A nanostructured surface has the advantages of a large heat-transfer area and good wettability.2 Therefore, utilizing the nanostructured surface is a viable solution for enhancing phase-change heat-transfer rates to overcome the cooling bottleneck.3,4 There has been extensive research on droplet spreading on solid and porous surfaces, with recent studies further investigating droplet behavior on nanostructured surfaces.5−10 As for a porous surface, Starov et al.11 experimentally investigated the spreading of silicon oil droplets on porous cellulose nitrate membrane filters and found that the spreading process can be divided into the fast spreading stage and the droplet base shrinking stage. Haidara et al.12 studied the wetting and imbibition of silicon oil droplets on nanoporous alumina membranes with 200 nm pore sizes using a CCD camera at 24 fps. They found that droplet spreading was strongly dependent on the viscosity regardless of whether the substrate was porous while the imbibition rate was significantly less dependent on the viscosity. Wang et al.13 observed water droplet spreading and imbibition on various superhydrophilic melt-blown poly(butylene terephthalate) fiber mats generated by controlling the hydrolysis times. The results revealed that the fiber mats with various fiber diameters and surface topographies showed different absorptions. Chao et al.14 experimentally investigated the spreading and imbibition of blood droplets (a non-Newton fluid) over filter paper and nitrocellulose membranes with pore sizes of 0.2, 3.0, and 8.0 μm. They measured the time evolution of droplet base radius, dynamic contact angle, and the wetted region radius for various conditions. Singh et al.15 investigated the spreading dynamics of a droplet on both plain and anodized nanoporous alumina surfaces with various pore distributions. They found that © 2017 American Chemical Society
the nanoporous structure significantly affected the substrate wettability and that randomly aligned nanoporous surfaces had the smallest contact angle and most complete spreading. Yang and Xu9 studied the droplet spreading on microporous and nanoporous surfaces with sintered copper particle diameters of 556 nm and 16 μm. Their results showed that the spreading process was composed of three distinct regimes according to the droplet neck height. Thus, the nanostructure improved the wettability, thereby accelerating the spreading. On the aspect of nanowire array surface, Fan et al.16,17 studied nanowire array surfaces by testing the wetting pattern of droplets on a surface with vertically aligned Si nanorod arrays using the glancing angle deposition technique. They investigated the spreading and the nanocarpet effect on the nanostructured surface with the diameter of nanowire near the substrate was 22 nm, that near the top was 103 nm, and the height was 890 nm. It was found that the wetting process can be divided into three stages: droplet impact, droplet spreading by gravity, and droplet spreading by precursor layer.17 Their subsequent research18 investigated the spreading of a droplet on similar Si nanorod array surfaces with an average height of 1106 nm. Their study measured the diameters of a spreading droplet and the precursor rim using a high-speed CCD camera with 210 fps. The results showed that the droplet behavior on vertically aligned Si nanorod arrays could be divided into four stages. Ahn et al.19 fabricated a nanostructure on a zircaloy surface with 20 nm average diameter and 1.62 μm high nanotubes. The phenomenon of droplet spreading and capillary wicking on nanotubes was described and divided into two stages: the inertial regime and the friction dominant regime. Although previous studies have investigated droplet behavior on nanostructured surfaces, there are still few studies of droplet Received: April 12, 2017 Revised: June 11, 2017 Published: June 13, 2017 6701
DOI: 10.1021/acs.langmuir.7b01223 Langmuir 2017, 33, 6701−6707
Article
Langmuir
Figure 1. Scanning electron microscope images of the nanostructured surfaces with the regular nanowires.
Figure 2. Scanning electron microscope images of the nanostructured surfaces with the irregular nanowires.
spreading and wicking on nanowire array surfaces, especially with the aim to enhance the heat transfer on the nanostructured surface. Furthermore, previous studies have not significantly addressed the effects of the nanowire array type and nanowire size. In this paper, an observation experiment with optical direct measurement was performed to investigate droplet spreading and capillary wicking on nanostructured surfaces fabricated by the hydrothermal synthesis method with various nanowire array types and nanowire sizes.
■
Table 1. Nanostructured Surface
EXPERIMENTAL METHOD
surface model
average nanowire height
nanowire diameter at half height
array type
N2-1 N3-1 N4-1 N2-2 N3-2 N4-2
4.26 ± 0.21 μm 9.15 ± 0.38 μm 16.20 ± 0.42 μm 4.22 ± 0.34 μm 9.40 ± 0.56 μm 16.80 ± 1.20 μm
134 ± 20 nm 178 ± 30 nm 247 ± 42 nm 128 ± 28 nm 172 ± 30 nm 238 ± 51 nm
regular regular regular irregular irregular irregular
top with a resolution of 1024 × 640 pixels at 1000 fps. Another high-speed CCD camera (Photron Mini UX100, with a Nikkor AF-S 50 mm lens and a close-up ring) captured the spreading from the side view with a resolution of 1280 × 1024 pixels at 4000 fps. The two cameras were simultaneously triggered by the control software. The last image captured immediately before the droplet touched the surface was treated as the initial image, t = 0. (2) Auxiliary camera: A high-definition camera (Canon 7D with a Canon Marco EF 100 mm lens) captured images at several given time points. (3) Microsyringe pump: A Harvard pump (Pump 11 Pico Plus Elite with 0.35% precision) was used to generate uniform droplets with an injection volume of 2 μL per injection. (4) Diffuse light source: A LED diffuse light source (CCS PD3-3024-3-PI) provided backlight when taking images from the side. An optical calibration plate with the same thickness as the target substrate was used to calibrate and focus the camera on the optical bench. In the experiments, a droplet was generated by the microsyringe pump and was allowed to gently touch the target surface by slowly lowering the needle to minimize the inertia effect. Because of the short time for spreading and wicking, evaporation was neglected. Figure 4a shows a schematic of the droplets spreading and wicking on the nanostructured surface. The droplets spread over the nanowires as the liquid imbibed into the nanowire layer and wicked into the channels formed by interlaced nanowires. Figure 4b shows an image captured by the auxiliary camera. Rd is the droplet spreading radius (from center to contact line) and Rp refers to the capillary wicking radius (from center to wicking rim). A “precursor ring” formed between the three-phase contact line and the wicking rim. Rd and Rp were measured using an in-house MATLAB-based processing program that reduced the random disturbances and identified the droplet spreading and wicking rim boundaries.
Surface Preparation. A ZnO nanowire array surface is highly photosensitive and the wettability can be changed by ultraviolet radiation.20 The electron−hole pairs are produced when ZnO is exposed to ultraviolet light, which then allows the oxygen vacancies to capture hydroxyl radicals more easily. Thus, the ZnO nanowire surface becomes more hydrophilic. The nanostructured surfaces were manufactured using hydrothermal synthesis method and the preparation process is shown as follows: First, a silicon substrate was coated with a thin Zn film (monocrystalline silicon N-type) using magnetron sputtering. Next, the substrate was suspended upside down in a zinc nitrate solution with stable temperature (90 °C), concentration (30 mM), and pH (pH 10.45, controlled by ammonia). Then the ZnO nanowire height (i.e., vertical distance from the substrate to the nanowire tip) was adjusted by controlling the growth time (one period of growth time is 6 h, for example, the N3-1 and N3-2 were obtained after two periods of growth time). Finally, the surfaces were exposed to ultraviolet light 2 h prior to experimentation. Two types of nanowire array were obtained using different magnetron sputtering modes. The nanowires on the substrate treated by direct current magnetron sputtering (DCMS) had good perpendicularity and alignment, as shown in Figure 1. The nanowires grown on the seed layer deposited by radio frequency magnetron sputtering (RFMS) had irregular structures, as shown in Figure 2. Table 1 lists the six nanostructures with the different nanowire array types and sizes tested in this paper. Experimental Setup and Process. Figure 3 shows the experimental system including the following parts: (1) Main cameras: A high-speed CCD camera (Baslar A504 K, with a Nikkor AF-S 50 mm lens and a close-up ring) captured the flow characteristics from the 6702
DOI: 10.1021/acs.langmuir.7b01223 Langmuir 2017, 33, 6701−6707
Article
Langmuir
Figure 3. Experimental system.
Figure 4. Droplet spreading and capillary wicking.
■
Figure 5. Side views of the droplet on the N3-2 surface.
RESULTS AND DISCUSSION Transportation Characteristics. The liquid transport after a droplet touches a nanostructured surface can be divided into
two parts: the spreading over the nanowires and the wicking in the nanowire layer. Figure 5 shows the droplet spreading process on the N3-2 surface from the side. Figure 6 shows the spreading 6703
DOI: 10.1021/acs.langmuir.7b01223 Langmuir 2017, 33, 6701−6707
Article
Langmuir
Figure 6. Top views of the spreading and wicking on the N3-2 surface (one-fourth of the original images).
line and then widens. After t = 2000 ms, Rd is generally constant with an ongoing increase in Rp. A video with 3/400 original play speed provided in the Supporting Information can illustrate the transportation clearly (see Video S1). In Fan et al.,16 both Rd and Rp on Si nanowires surface exhibited the dynamic scaling behavior, D = 2R ∝ tn/2. Starov et al.11 and Fan et al.16 both stated that the exponent n resulted from the capillary-driven spreading due to the contact line moving on the prewetted surface formed by the precursor rim. Power law fitting gave the exponent nd = 0.216 for the contact line and np = 0.272 for the precursor rim.16 Power law fits of Rd and Rp on the N3-2 surface in the present work gave nd = 0.378 and np = 0.324 during the capillary wicking stage. The differences are due to the different capillary wicking abilities of the N3-2 and Si nanowire surfaces in Fan et al.16 However, similar to the result in Fan et al.16 of nδ = 0.547, the width of the precursor ring, δ, also follows the power law, δ ∝ tnδ, with nδ = 0.544, which is consistent with the Washburn equation21 δ = (Reσμ cos θ/2)0.5t0.5 where σ is the liquid surface tension, μ is the liquid viscosity, and Re is the effective radius of the capillary as shown in the inset in Figure 7. Effect of Nanowire Array Type on Liquid Transport. The regular and irregular nanowire arrays were produced using different magnetron sputtering ways. Figure 8 shows the effect of the array type on the time evolution of Rd and Rp. Figure 8a−c indicates that all the irregular nanostructured surfaces have faster spreading and wicking rates for similar nanowire heights. Figure 8a shows that Rd for the N2-2 surface sharply increases to approach the maximum spreading radius before t = 12 ms, while Rd on the N2-1 surface increases more slowly. Furthermore, the maximum Rd on the N2-2 surface is slightly larger than that on the N2-1 surface. Rp on the N2-2 surface is always larger than Rp on the N2-1 surface. The precursor rims on the N2-1 and N2-2 surfaces both appear at t = 6 ms. There are many nanochannels within the ZnO nanowire array. Liquid penetration into the nanochannels depends on the balance between the capillary pressure and the viscous resistance. A smaller channel size has a larger capillary pressure and viscous resistance, while a larger channel size reduces the resistance but also reduces the capillary pressure. Therefore, a nanostructured surface with multiscale nanochannels has better penetration. Compared with the regular nanostructured array, the irregular nanowire array has a stronger capillary pressure due to the multiscale channels; hence, the average wicking rate with the irregular nanostructure is larger. During the spreading process, the wetted surface also changes the wettability at the three-phase contact line, which promotes the droplet spreading.
and wicking on the N3-2 surface from the top. The contact line and the wicking rim are both perfect circles, which confirm the good uniformity of the nanostructured surface. Figures 5 and 6 show that the droplet sharply deforms from a sphere to a spherical cap in 6 ms after touching the nanostructured surface. After t = 6 ms, both the spreading velocity and the spherical cap height both gradually decrease. When t = 6−10 ms, the liquid film fluctuates due to the strong attachment. After t = 10 ms, the spherical cap height and spreading diameter vary more slowly. There was no pinching neck or daughter droplet observed during the spreading process. The droplet spreading and wicking process can be divided into two stages: the free spreading and capillary wicking based on appearance of the precursor ring, as shown in Figure 7. Figure 7 also shows the variation of Rd and Rp with time for spreading on the N3-2 surface, the values of Rd and Rp averaged over two sets of measurements with each set along a pair of mutually perpendicular radii. Taking the spreading on N3-2 as an example, during the spreading stage (t < 10 ms in Figure 7), the droplet rapidly spreads after touching the surface due to the super hydrophilicity and the wicking has a similar radial velocity (neglecting the wicking in the vertical direction). Hence, only a three-phase contact line can be observed before t = 10 ms. In the capillary wicking stage (t ≥ 10 ms in Figure 7), the capillary wicking is still strong but the droplet spreading begins to decelerate; thus, the precursor ring appears ahead of the contact
Figure 7. Spreading process on N3-2. 6704
DOI: 10.1021/acs.langmuir.7b01223 Langmuir 2017, 33, 6701−6707
Article
Langmuir
Figure 8. Effect of array type on the droplet spreading and wicking.
Figure 9. Effect of nanowire size on the droplet spreading.
Effect of Nanowire Size on Fluid Transport. Figure 9 shows the effect of nanowire size on Rd. Here, the N3-1 and N3-2 surfaces have the largest Rd at any given point in time among the three different nanowire heights. Figure 10 shows the effect of nanowire size on Rp. Similarly, Rp on the N3-1 and N3-2 surfaces also have the largest average wicking velocities of all the nanostructures. Spreading and wicking on a nanostructured surface has a complicated flow behavior because the spreading is strongly dependent on the wicking inside the nanowire array. Faster wicking promotes
spreading while slower wicking reduces spreading. The nanowire growth time in the hydrothermal synthesis method used here to synthesize the nanowires has the strongest influence on the nanowire height, but it also has a minor effect on the nanowire diameter. The nanowire diameters of the N2-1 surface are smaller than those of the N3-1 and N4-1 surfaces. Therefore, the N2-1 surface has the largest nanochannels, which reduces the capillary wicking on the N2-1 surface, and then influences the spreading. The wicking in nanowires of the N4-1 surface is weak due to the large flow resistance that results from the narrow 6705
DOI: 10.1021/acs.langmuir.7b01223 Langmuir 2017, 33, 6701−6707
Article
Langmuir
Figure 10. Effect of nanowire size on the droplet wicking.
■
spacing of the nanowires. The reduction in the nanochannel size raises both the flow resistance and the capillary pressure, but the reduction in nanochannel size has the greater impact on rising flow resistance than capillary pressure. The nanochannels on the N3-1 surface have moderate sizes that have a relatively strong capillary force and small resistance. Hence, the fluid permeates faster into the capillaries than on the other two surfaces. Furthermore, the wicking along the perpendicular direction in thicker nanowire layer depresses the wicking in the radial direction to a certain extent. However, for the irregular nanostructured surface, the interlaced nanowires reduce the impact of nanowire growth on the pore size. Hence, Rd for the N4-2 surface is only slightly less than that for the N3-2 surface, while the differences between the N3-1 and N4-1 surfaces are much greater as shown in Figure 9b. For the maximum Rd, Figure 9a shows that Rd on the N3-1 surface continues to increase for t larger than 1500 ms while Rd on the N4-1 surface stabilizes at t = 816 ms and Rd on the N2-1 surface reaches a maximum at t = 192 ms. In Figure 9b, Rd for the N4-2 and N3-2 surfaces do not reach their maximum spreading diameters within 1500 ms, while Rd for the N2-1 surface reaches dmax at t = 32 ms.
*E-mail:
[email protected]. ORCID
Ruina Xu: 0000-0001-8561-560X Peixue Jiang: 0000-0002-9777-8075 Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors acknowledge the support from the National Natural Science Foundation of China Science Fund for Creative Research Groups (No. 51621062).
■
REFERENCES
(1) Gurrum, S. P.; Suman, S. K.; Joshi, Y. K.; Fedorov, A. G. Thermal Issues in Next Generation Integrated Circuits. IEEE Trans. Device Mater. Reliab. 2004, 4, 709−715. (2) Webb, R. L. The Evolution of Enhanced Surface Geometries for Nucleate Boiling. Heat Transfer Eng. 1981, 2, 46−49. (3) Zhang, Z.; Jiang, P. X.; Christopher, D. M.; Liang, X. G. Experimental Investigation of Spray Cooling on Micro-, Nano- and Hybrid-structured Surfaces. Int. J. Heat Mass Transfer 2015, 80, 26−37. (4) Chen, J. N.; Zhang, Z.; Xu, R. N.; Ouyang, X. L.; Jiang, P. X. Numerical Investigation of the Flow Dynamics and Evaporative Cooling of Water Droplets Impinging onto Heated Surfaces: An Effective Approach to Identify Spray Cooling Mechanisms. Langmuir 2016, 32, 9135−9155. (5) Bird, J. C.; Mandre, S.; Stone, H. A. Short-Time Dynamics of Partial Wetting. Phys. Rev. Lett. 2008, 100, 1151−1156. (6) Winkels, K. G.; Weijs, J. H.; Eddi, A.; Snoeijer, J. H. Initial Spreading of Low-viscosity Drops on Partially Wetting Surfaces. Phys. Rev. E 2012, 85, 055301. (7) Ding, H.; Li, E. Q.; Zhang, F. H.; Sui, Y.; Spelt, P. D. M.; Thoroddsen, S. T. Propagation of Capillary Waves and Ejection of Small Droplets in Rapid Droplet Spreading. J. Fluid Mech. 2012, 697, 92−114. (8) Chen, L. Q.; Bonaccurso, E.; Shanahan, M. E. R. Inertial to Viscoelastic Transition in Early Drop Spreading on Soft Surfaces. Langmuir 2013, 29, 1893−1898. (9) Yang, W. L.; Xu, J. L. Drop Spreading and Penetrating on Micro/ nano Particle Sintering Porous with Multiscale Structure. Colloids Surf., A 2017, 516, 9−22. (10) Lee, J. B.; Derome, D.; Carmeliet, J. Drop Impact on Natural Porous Stones. J. Colloid Interface Sci. 2016, 469, 147−156. (11) Starov, V. M.; Kostvintsev, S. R.; Sobolev, V. D.; Velarde, M. G.; Zhdanov, S. A. Spreading of Liquid Drops over Dry Porous Layers: Complete Wetting Case. J. Colloid Interface Sci. 2002, 252, 397−408.
■
CONCLUSIONS The behaviors of droplet spreading and capillary wicking on nanostructured surface with different array types and nanowire sizes were described and analyzed quantitatively using a visualization measurement. The nanostructured surface with irregular nanowires gave faster spreading and wicking rates than the surfaces with regular nanowires due to the irregular nanochannel size distribution that promoted the spreading and wicking. The nanostructured surface with the medium height nanowires had the largest average spreading and wicking velocities of the three nanowire lengths studied in this work.
■
AUTHOR INFORMATION
Corresponding Author
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.7b01223. Video S1 shows the transportation process when a droplet touches the N2-2 surface with top view (AVI) 6706
DOI: 10.1021/acs.langmuir.7b01223 Langmuir 2017, 33, 6701−6707
Article
Langmuir (12) Haidara, H.; Lebeau, B.; Grzelakowski, C.; Vonna, L.; Biguenet, F.; Vidal, L. Competitive Spreading versus Imbibition of Polymer Liquid Drops in Nanoporous Membranes: Scaling Behavior with Viscosity. Langmuir 2008, 24, 4209−4214. (13) Wang, Z.; Espín, L.; Bates, F. S.; Kumar, S.; Macosko, C. W. Water droplet Spreading and Imbibition on Superhydrophilic Poly (Butylene Terephthalate) Melt-blown Fiber Mats. Chem. Eng. Sci. 2016, 146, 104− 114. (14) Chao, T. C.; Arjmandi-Tash, O.; Das, D. B.; Starov, M. V. Simultaneous Spreading and Imbibition of Blood Droplets over Porous Substrates in The Case of Partial Wetting. Colloids Surf., A 2016, 505, 9− 17. (15) Singh, S. K.; Khandekar, S.; Pratap, D.; Ramakrishna, S. A. Wetting Dynamics and Evaporation of Sessile Droplets on Nano-porous Alumina Surfaces. Colloids Surf., A 2013, 432, 71−81. (16) Fan, J. G.; Tang, X. J.; Zhao, Y. P. Water Contact Angles of Vertically Aligned Si Nanorod Arrays. Nanotechnology 2004, 15, 501− 504. (17) Fan, J. G.; Dyer, D.; Zhang, G.; Zhao, Y. P. Nanocarpet Effect: Pattern Formation during the Wetting of Vertically Aligned Nanorod Arrays. Nano Lett. 2004, 4, 2133−2138. (18) Fan, J. G.; Zhao, Y. P. Spreading of a Water Droplet on a Vertically Aligned Si Nanorod Array Surface. Appl. Phys. Lett. 2007, 90, 013102. (19) Ahn, H. S.; Park, G.; Kim, J.; Kim, M. H. Wicking and Spreading of Water Droplets on Nanotubes. Langmuir 2012, 28, 2614−2619. (20) Sun, R. D.; Nakajima, A.; Fujishima, A.; Watanabe, T.; Hashimoto, K. Photoinduced Surface Wettability Conversion of ZnO and TiO2 Thin Films. J. Phys. Chem. B 2001, 105, 1984−1990. (21) Washburn, E. W. The Dynamics of Capillary Flow. Phys. Rev. 1921, 17, 273−283.
6707
DOI: 10.1021/acs.langmuir.7b01223 Langmuir 2017, 33, 6701−6707
