Understanding the Effects of Surface Texturing on the Heat Transfer Characteristics of Spray Cooling

Abstract

Enhanced surfaces show improved heat transfer characteristics in the two-phase regime due to enhanced capillary action near the contact line region. The role of contact-line evaporation on spray impingement heat transfer is systematically studied by spraying de-ionized water on silicon substrates with micropillar arrays. The pillar height, the pillar diameter, and the pillar spacing of the micropillar array were varied from 5 to 50 μm while keeping the porosity constant at 0.75. An air-assisted nozzle was used to create a liquid spray with a Sauter Mean Diameter of ~ 35 μm depending on flow conditions. Most test runs were conducted at a water flow rate of 30 ml/min and an air-liquid mass flow rate ratio of ~ 0.57. Temperature and heat flux data were processed to estimate the macroscale heat transfer parameters, including critical heat flux (CHF), heat transfer coefficient (HTC), and peak cooling efficiency. The results show a continuous increase in the critical heat flux (CHF) as the pillar diameter is decreased until a pillar diameter of 5 μm. Below d = 5 μm (i.e., for d = 2 μm and 0.8 μm), a monotonous decrease in the heat transfer performance was observed. Likewise, the effects of pillar height were non-monotonic, with CHF and peak heat transfer coefficient attaining a maximum as the height-to-diameter ratio approaches unity (h/d ~ 1). The transition from Cassie–Baxter state to Wenzel state is critical for liquid penetration into the forest of micropillar array. Surfaces with 2-μm and 0.8-μm pillar arrays exhibit inferior cooling performance due to low liquid penetration.Silicon substrates containing 5-μm and 10-μm pillar arrays (const. h ~ 8 μm) with varying pillar spacing (p ~ 0.2d, 0.5d, 1d, 1.5d, 2d) were fabricated to investigate the role of pillar spacing on the spray cooling performance. Each pillar-array geometry tested has a CHF maximum for pitch-to-diameter ratio (p/d) of ~ 1. Based on the results obtained and insights gained, surfaces with pillar diameter (d) ≈ pillar spacing (p) ≈ pillar height (h) is preferred. Overall, 5-μm and 10-μm arrays with h/d ~ 1 yielded some of the maximum performance values for the given flow settings. The Cassie–Baxter state to Wenzel state transition theory stands valid for surfaces with small pillar spacing, i.e., p/d < 1. On the other extreme, in-plane Laplace pressures become essential to ensure complete wetting. Hence, pillar-arrayed surfaces with p/d > 1 show diminishing results. The effect of liquid flow rates and air flow rates were also investigated independently using textured surfaces. Values of CHF as high as 830 W/cm2 were achieved with a water flow rate of 82 ml/min, along with a cooling efficiency of 49%. High-speed images containing crucial information on the solid/liquid/vapor interface were obtained Total internal reflectance (TIR) technique. A sapphire substrate with an indium tin oxide (ITO) heater was used as a substrate. For the first time, high-temperature experiments (T > 120 ºC) were performed using a plain surface and 10-μm pillared surface. Images were processed to obtain contact-line parameters such as wetted area fraction, and contact-line perimeter per unit area. Time-averaged results reveal the superior performance of the 10-μm pillared surface is due to the enhanced wicking capabilities of the micropillar configuration at elevated heat flux values. The time-series data of wetted area fraction plotted for both plain and 10-μm surface show that the CHF occurs when the evaporation time scale is less than the re-wetting time scale. With wicking enhancement and faster in-plane capillary suction flows, the 10-μm surface was able to maintain the re-wetting time scale to be lower than the evaporation time scale, thereby delaying CHF to a value 35% greater than that of plain surface CHF. From the model presented, it is clear that the substrate approaches CHF when the heating time scale of the substrate is equivalent to the liquid imbibition time scale.