Visualization of Patterned Modiﬁed Surfaces in Condensation and Frosting States

: In this study, a novel, thorn-shaped, containing, hydrophilic, and hydrophobic surface is proposed to have a better condensate drainage characteristic and to delay the required time for frosting. By using a hydrophilic and hydrophobic mixed thorn-shaped surface created by screen printing, the design makes use of the di ﬀ erences in the wettability gradient to achieve rapid condensate drainage and to lengthen the time for frosting. The results of a frosting experiment indicated that the droplet adsorption and combination and discharge e ﬀ ect in the thorn sample were substantial. The drainage e ﬀ ect increased the surface renewal rate and inhibited ice layer growth on the thorn sample by 52.4% compared with that on pure copper surface. The heat transfer coe ﬃ cient of the thorn sample during frosting was approximately 16.2% higher than that of pure copper surface. In addition, the defrosting results indicated that the defrosting time of the thorn sample was almost equal to that of the pure copper sample. However, large droplets were easily stagnated at the structural junction due to contact angle hysteresis after defrosting.


Introduction
Heat exchangers have been widely used in modern industrial and mechanical equipment and play a key role in processing and manufacturing. Due to concerns regarding environmental pollution and energy utilization, improving the energy efficiency, particularly the fin frosting phenomenon of finned tube heat exchangers used in air conditioning and refrigeration systems, has become a crucial issue. In addition to reducing system efficiency, accumulation of a frost layer causes an increase in the air-side pressure drop and fan power consumption [1]. Moreover, melted frost water can reduce heat transfer and cause refrosting, which reduces energy efficiency appreciably. To solve the aforementioned problems, many scholars have studied defrosting technology for heat exchangers. The defrosting-related technologies can be classified into active and passive designs [2]. Active designs increase the defrosting capacity through the energy input method. Many researchers utilized electrohydrodynamic (EHD) technique to reduce the frost amount [3]. Although active designs can considerably improve the defrosting capacity, they also increase system complexity and consume additional energy.
Passive defrosting designs are mostly adopted through surface treatment such that no additional power is input into the system, and no additional components are required. Passive defrosting designs are a recent study trend. Liu et al. [4] developed a frost resistance coating, and their results revealed that applying the coating on metal surfaces can delay the initial frost crystal formation time by at least 15 min. The frost layer thickness was also reduced by 40%. When relative humidity (RH) of the environment was less than 60%, and the temperature of the flat surface was higher than −10 • C, the coated surface could remain unfrosted for 3 h. By contrast, the uncoated copper surface was completely gradient between both areas to drive the droplets from the β area to the α area. Finally, the α area acted as a drainage path to remove surface-condensed water for increasing the removal rate. The heat transfer coefficient was increased by 4.8 and 1.8 times compared with that for filmwise condensation and dropwise condensation, respectively.
Yang et al. [12] designed an interlaced pattern of parallel oblique angles to compare hydrophilic, superhydrophilic, and hydrophobic surfaces. They claimed that the maximum average Feret diameter of the droplets on a mixed surface was approximately 80% to 90% smaller than that on the hydrophobic surface. The heat transfer coefficient of the mixed surface was approximately 3% to 9% higher than that of the hydrophobic surface due to the movement and discharge mechanism. Moreover, the heat transfer coefficient of the mixed surface was approximately 6% to 16% higher than that of the hydrophilic surface. Yang et al. [13] proposed an improved star design that can effectively enhance the heat transfer performance by 15.3% to 30.4%. Alheshibri et al. [14] discovered that when a droplet was at the tip of the hydrophilic surface, it slowly moved toward the divergence. The larger the angle of divergence, the faster the droplet moved, and the surface tension produced by the structure became higher than the gravity at a certain angle. Mahapatra et al. [15] placed a superhydrophilic surface track in the center of a hydrophilic pattern and staggered parallel to the hydrophilic pattern. The heat transfer coefficient of the design at 29.9% filmwise area increased by 34.4% compared with that for completely dropwise condensation.
The aforementioned studies have explained the condensation heat transfer enhancement benefits of surface modification, and surface modification technology has been used in frosting studies. Liu et al. [16] investigated the effect of condensation on a superhydrophobic modified surface with a contact angle of 162 • compared with that on a pure copper surface. The results indicated that compared with the pure copper surface, the superhydrophobic modified surface delayed the freezing time by approximately 55 min. Moreover, the ice layer of the superhydrophobic modified surface was loose and light, exhibiting a daisy-like pattern. Kim and Lee [17] conducted defrosting experiments on superhydrophilic, pure copper, and superhydrophobic fin surfaces. They also observed a frosting delay phenomenon on the superhydrophobic fins. Furthermore, the finding revealed that the surface contact angle had no significant effect on the defrosting time.
According to the existing literatures, a patterned wettability area can effectively improve the benefits of condensate drainage; however, this method has rarely been applied to frosting studies. Therefore, the objective of this study is to provide a novel, thorn-shaped, containing, hydrophilic, and hydrophobic surface to have a better condensate drainage characteristic, and it will be shown subsequently that that surface can also delay the required frosting time. The design is made available through screen printing and uses the difference in wettability gradient for achieving surface characteristics such as rapid surface drainage of condensate and frost resistance. Image visualization technology was used to understand its mechanism in depth as the basis for optimal pattern design.

Experimental Framework
The diagram of the experimental system used in this study is displayed in Figure 1. The system consisted of an environmental chamber, a frost and defrost water circulation system, a data acquisition system, and a high-speed image system. The system was equipped with two thermostatic baths, namely the thermostatic baths for frosting (EYELA CCA-1111) and defrosting (YSC RC-10), which can offer controlled temperature ranges of −20 to 30 • C and −10 to 100 • C, respectively. Both thermostatic baths were supplied with ethylene glycol solution (60% wt), which can control the inlet water temperature and circulating water volume of the hydrocooling plate for allowing a stable condensate flow with a minimum control temperature of −20 • C. The water circulation system controlled the frost or defrost circuit through a manual three-way valve. A flow regulating valve and flow meter (YOKOGAWA-AXF002G) were installed at the tube inlet for adjustment and monitoring the flow. After passing through the condensation testing section of the test segment, the flow circulated back to the thermostatic bath and completed the condensate circuit. The flow meter had a measurement range of 0.1-8.0 m 3 /h and an accuracy of 0.35%. The cold plate of the experiment was placed in the environmental chamber (JOBHO CD-321). The four corners of the front side of the cold plate were fixed by screws so that samples could be easily replaced. The other exposed parts of the cold plate were insulated with polyethylene heat insulating material, and a camera was added in front of the sample to observe and record the surface phenomenon. To confirm the wind speed in the environmental chamber, a unidirectional hot-wire anemometer (Testo 480) was used to measure the airflow velocity from sampling of 20-points. The measuring points of the hot-wire anemometer are located at a virtual plane which is 50 mm apart from the sample. The vertical and horizontal spacing of measuring points are 50 mm and 60 mm, respectively. The hot-wire anemometer had a measurement range of 0 to 20 m/s and an accuracy of ±(0.03 m/s + 5%). The average airflow velocity was 0.85 m/s. . The four corners of the front side of the cold plate were fixed by screws so that samples could be easily replaced. The other exposed parts of the cold plate were insulated with polyethylene heat insulating material, and a camera was added in front of the sample to observe and record the surface phenomenon. To confirm the wind speed in the environmental chamber, a unidirectional hot-wire anemometer (Testo 480) was used to measure the airflow velocity from sampling of 20-points. The measuring points of the hot-wire anemometer are located at a virtual plane which is 50 mm apart from the sample. The vertical and horizontal spacing of measuring points are 50 mm and 60 mm, respectively. The hot-wire anemometer had a measurement range of 0 to 20 m/s and an accuracy of ±(0.03 m/s + 5%). The average airflow velocity was 0.85 m/s. To more accurately evaluate the condensing or frosting performance, nine T-type thermocouples were installed on the back of the test plate sample ( Figure 2) for measuring the temperatures at different locations on the bottom surface of the plate. In addition, a heat conduction calculation formula was used to evaluate the differences between the surface and bottom temperatures in the experiment (<0.01 °C). Thus, the heat transfer coefficient was directly calculated using the average temperature. The temperature was measured by the T-type thermocouple wires at the frost and defrost water cycle inlet and outlet, and the temperature signals were recorded separately. The temperature variations of the thermocouples were within 0.2 °C during the isothermal test. The thermocouples were calibrated using a quartz thermometer with an accuracy of 0.01 °C prior to testing for achieving a thermocouple accuracy of 0.1 °C. All the relevant information was converted by the data acquisition system (GRAPHTEC-GL840) and transmitted to the computer for processing and recording through a web interface. To capture the changes in the patterned surface condensation, frosting, and defrosting, an image system was set up to determine the relative relationship between the flow field change and the heat transfer performance. The camera used in this system was UPMOST-UPG650, which features a shooting speed of more than 30 fps. The camera was placed on a tripod to capture the pictures of condensation, frosting, and defrosting phase changes on the plate To more accurately evaluate the condensing or frosting performance, nine T-type thermocouples were installed on the back of the test plate sample ( Figure 2) for measuring the temperatures at different locations on the bottom surface of the plate. In addition, a heat conduction calculation formula was used to evaluate the differences between the surface and bottom temperatures in the experiment (<0.01 • C). Thus, the heat transfer coefficient was directly calculated using the average temperature. The temperature was measured by the T-type thermocouple wires at the frost and defrost water cycle inlet and outlet, and the temperature signals were recorded separately. The temperature variations of the thermocouples were within 0.2 • C during the isothermal test. The thermocouples were calibrated using a quartz thermometer with an accuracy of 0.01 • C prior to testing for achieving a thermocouple accuracy of 0.1 • C. All the relevant information was converted by the data acquisition system (GRAPHTEC-GL840) and transmitted to the computer for processing and recording through a web interface. To capture the changes in the patterned surface condensation, frosting, and defrosting, an image system was set up to determine the relative relationship between the flow field change and the heat transfer performance. The camera used in this system was UPMOST-UPG650, which features a shooting speed of more than 30 fps. The camera was placed on a tripod to capture the pictures of condensation, frosting, and defrosting phase changes on the plate under any operating conditions. Uncertainties in the reported experimental values were estimated by the method suggested by Moffat [18]. The highest uncertainties are 5.83% for the heat transfer coefficient.
under any operating conditions. Uncertainties in the reported experimental values were estimated by the method suggested by Moffat [18]. The highest uncertainties are 5.83% for the heat transfer coefficient.

Parameter Measurement and Discussion
The distance between the front and back surfaces of the copper plate was 5 mm. The temperature difference between both surfaces of the copper plate can be calculated through (1) and (2). The results indicated that the temperature difference between both sides was only within 0.01 °C. Thus, the measured back surface temperature was assumed to be the front surface temperature.

= × × ∆ ∆
(∆T: temperature difference between the front and back surfaces of the copper plate)

Parameter Measurement and Discussion
The distance between the front and back surfaces of the copper plate was 5 mm. The temperature difference between both surfaces of the copper plate can be calculated through (1) and (2). The results indicated that the temperature difference between both sides was only within 0.01 • C. Thus, the measured back surface temperature was assumed to be the front surface temperature.
The temperature difference (T out − T in ) between the inlet and outlet of the ethylene glycol solution of the thermostatic bath was calculated to obtain the total heat absorbed by the cold plate. The heat convection coefficient on the air side (h) was estimated from the ambient temperature (T air ) and plate temperature (T s ) according to the following formula:

Patterned Modified Surface
The geometric size of the two samples tested in this study ( Figure 3) was 350 × 250 × 3 mm 3 (length × width × thickness). The two samples, namely a pure copper surface (as a control group) and patterned modified surface, were made of red copper with a heat conductivity of 387 W/m-K. Red copper has a hydrophilic surface. First, waterproof sandpaper was used to remove the oxide film on the surface, which was polished with a bench grinder. Finally, the surface of the copper plate was cleaned with industrial alcohol and a clean cloth with a contact angle of 65 • . The polished red copper sample was set at rest. Then, Teflon paint was poured into a predesigned thorn-shaped pattern stencil (the molding contact angle was 89 • ) for printing on the plate surface through manual pushing. The samples were then baked at 150 • C for 30 min, and the thorns were obtained after cooling. The temperature difference ( − ) between the inlet and outlet of the ethylene glycol solution of the thermostatic bath was calculated to obtain the total heat absorbed by the cold plate. The heat convection coefficient on the air side (h) was estimated from the ambient temperature ( ) and plate temperature ( ) according to the following formula:

Patterned Modified Surface
The geometric size of the two samples tested in this study ( Figure 3) was 350 × 250 × 3 mm 3 (length × width × thickness). The two samples, namely a pure copper surface (as a control group) and patterned modified surface, were made of red copper with a heat conductivity of 387 W/m-K. Red copper has a hydrophilic surface. First, waterproof sandpaper was used to remove the oxide film on the surface, which was polished with a bench grinder. Finally, the surface of the copper plate was cleaned with industrial alcohol and a clean cloth with a contact angle of 65°. The polished red copper sample was set at rest. Then, Teflon paint was poured into a predesigned thorn-shaped pattern stencil (the molding contact angle was 89°) for printing on the plate surface through manual pushing. The samples were then baked at 150 °C for 30 min, and the thorns were obtained after cooling.

Experimental Principle
The goal of the experiment was to enhance the movement mechanism of the condensate by using different surface structures and pattern changes. According to the relevant literature, differences in the free energy of different wetted surfaces can cause small droplets to move from a less hydrophilic surface to a more hydrophilic surface. When the droplets were at the tip of the hydrophilic area, their force directions diverged. Moreover, the wedge-shaped angle opening of the droplets was set downward to allow the condensate in the Teflon area to move to the polished red copper area for a second time by using gravity to increase its moving force, as depicted in Figure 4a. Finally, the central vertical polished copper drain was designed for the droplets to move to the center. In addition to increasing the combined speed, it can quickly reach the maximum droplet size and slide down while sweeping other droplets on the drain to quickly reduce the surface heat resistance ( Figure 5).

Experimental Principle
The goal of the experiment was to enhance the movement mechanism of the condensate by using different surface structures and pattern changes. According to the relevant literature, differences in the free energy of different wetted surfaces can cause small droplets to move from a less hydrophilic surface to a more hydrophilic surface. When the droplets were at the tip of the hydrophilic area, their force directions diverged. Moreover, the wedge-shaped angle opening of the droplets was set downward to allow the condensate in the Teflon area to move to the polished red copper area for a second time by using gravity to increase its moving force, as depicted in Figure 4a. Finally, the central vertical polished copper drain was designed for the droplets to move to the center. In addition to increasing the combined speed, it can quickly reach the maximum droplet size and slide down while sweeping other droplets on the drain to quickly reduce the surface heat resistance ( Figure 5).

Results and Discussion
To evaluate the performance of the patterned modified surface under different designs, condensation, frosting, and defrosting experiments of different surface structure samples were conducted at an ambient temperature of 20 °C, 80% RH, and a flat surface temperature of −8.4 °C. The experimental image of the pure copper surface ( Figure 6) reveals a comprehensive condensation and frosting process. First, a condensation phenomenon occurred on the surface of the plate. For elapsed time of 30 minutes, the droplet accumulated to a threshold size, in which the gravity surpasses the adhesion force. Consequently, slippage began, and a condensation nucleus process occurred. The sliding path mostly comprised vertical strips. Although the aforementioned phenomenon was frequently repeated, a low contact angle of the pure copper surface resulted in high wettability, which caused some of the condensate to remain on the surface after renewal and the formation of irregular filmwise shapes. When time reaches 45 minutes, the liquid film was completely frozen, and surface renewal could not be performed. Finally, as the ice density strengthened during the ice layer complete growth period, the thickness of the ice layer gradually covered the frozen droplets.

Results and Discussion
To evaluate the performance of the patterned modified surface under different designs, condensation, frosting, and defrosting experiments of different surface structure samples were conducted at an ambient temperature of 20 • C, 80% RH, and a flat surface temperature of −8.4 • C. The experimental image of the pure copper surface ( Figure 6) reveals a comprehensive condensation and frosting process. First, a condensation phenomenon occurred on the surface of the plate. For elapsed time of 30 min, the droplet accumulated to a threshold size, in which the gravity surpasses the adhesion force. Consequently, slippage began, and a condensation nucleus process occurred. The sliding path mostly comprised vertical strips. Although the aforementioned phenomenon was frequently repeated, a low contact angle of the pure copper surface resulted in high wettability, which caused some of the condensate to remain on the surface after renewal and the formation of irregular filmwise shapes. When time reaches 45 min, the liquid film was completely frozen, and surface renewal could not be performed. Finally, as the ice density strengthened during the ice layer complete growth period, the thickness of the ice layer gradually covered the frozen droplets.
The thorns utilized the differences in free energy between different hydrophilic surfaces, the low wettability of the Teflon pattern, and the high wettability of pure copper surface. When the droplets in the Teflon pattern area touched the junction, parallel adsorption was observed onto the polished copper area, as displayed in Figure 7a. The droplet adsorption and combined force at the Teflon area exhibited an up-moving phenomenon hat overcame the influence of gravity (Figure 7b). Because the hydrophilic wedge angle was inclined downwards, the gravity force assisted the droplet in moving, which enhanced the gathering effect of the droplets on the central drainage channel, as illustrated in Figure 7c. When the droplets gradually merged at the central hydrophilic point to yield the maximum droplet size, the surface adhesion was then overpowered by gravity, which caused the droplet to slide down and simultaneously sweep away other droplets on the slip path, as displayed in Figure 8a. A special phenomenon of crossflow passage may occur when the sliding movement changes direction due to droplets on the non-conventional drainage channel, which increases the droplet merger rate as well as the sweeping area during sliding. Moreover, tiny droplets remained in the central drainage channel on the renewed surface mostly, whereas nearly no droplets adhered to the Teflon pattern area. Finally, the condensation process was repeated on the renewed surface, and the changes in size of the droplets are depicted in Figure 9. The thorns utilized the differences in free energy between different hydrophilic surfaces, the low wettability of the Teflon pattern, and the high wettability of pure copper surface. When the droplets in the Teflon pattern area touched the junction, parallel adsorption was observed onto the polished copper area, as displayed in Figure 7a. The droplet adsorption and combined force at the Teflon area exhibited an up-moving phenomenon hat overcame the influence of gravity (Figure 7b). Because the hydrophilic wedge angle was inclined downwards, the gravity force assisted the droplet in moving, which enhanced the gathering effect of the droplets on the central drainage channel, as illustrated in Figure 7c. When the droplets gradually merged at the central hydrophilic point to yield the maximum droplet size, the surface adhesion was then overpowered by gravity, which caused the droplet to slide down and simultaneously sweep away other droplets on the slip path, as displayed in Figure  8a. A special phenomenon of crossflow passage may occur when the sliding movement changes direction due to droplets on the non-conventional drainage channel, which increases the droplet merger rate as well as the sweeping area during sliding. Moreover, tiny droplets remained in the central drainage channel on the renewed surface mostly, whereas nearly no droplets adhered to the Teflon pattern area. Finally, the condensation process was repeated on the renewed surface, and the changes in size of the droplets are depicted in Figure 9.     Due to the surface tension of the thorns, the plate exhibited a frequent surface renewal phenomenon. When the condensed water slid down to the initial frost zone, the slipping phenomenon had the same renewal effect on the thin ice layer on the surface. The sliding condensate destroyed the initially grown ice crystals or the fragile ice layer, which forced the surface to refreeze and consequently resulted in an incomplete surface frosting phenomenon (Figure 10). When the ice layer was thick with a high distribution density, the slid droplets moved and spread across the porous media between the frozen droplets by capillary force, which enlarged the width of the sliding path and increased the renewed frost surface area. This phenomenon can delay the freezing time of the droplets, reduce the thickness of the ice layer and heat resistance, and improve the heat transfer coefficient, as illustrated in Figures 11-13. Due to the surface tension of the thorns, the plate exhibited a frequent surface renewal phenomenon. When the condensed water slid down to the initial frost zone, the slipping phenomenon had the same renewal effect on the thin ice layer on the surface. The sliding condensate destroyed the initially grown ice crystals or the fragile ice layer, which forced the surface to refreeze and consequently resulted in an incomplete surface frosting phenomenon (Figure 10). When the ice layer was thick with a high distribution density, the slid droplets moved and spread across the porous media between the frozen droplets by capillary force, which enlarged the width of the sliding path and increased the renewed frost surface area. This phenomenon can delay the freezing time of the droplets, reduce the thickness of the ice layer and heat resistance, and improve the heat transfer coefficient, as illustrated in Figures 11-13.      As displayed in Figure 14, the renewal rate of pure copper surface was lower than that of the thorns; however, the surface renewal phenomenon occurred quicker for polished red copper. Therefore, at the time 45 minutes, the red copper surface condensate had completely frozen, whereas the surface droplets of thorns began to freeze at 75 min and were completely frozen at 90 min. Finally, the ice layer produced after 2 hours of frosting was poured in a 30 °C ethylene glycol solution for defrosting and the melting phenomenon was recorded. As displayed in the image, because the polished copper surface had a low contact angle and high wettability, some condensate remained on the surface and exhibited an irregular filmwise state after defrosting. The surface of the thorns had a high surface contact angle and low wettability. Therefore, a strong repulsive force existed between the surface and the melted water When the thermal solution passed through the internal flow channel, the ice layer in contact with the plate surface as well as the ice surface exposed to the environment melted simultaneously. However, the heat conduction rate decreased because the ice layer was thicker at the middle. Moreover, the melted water in contact with the flat plate was removed, which caused the entire ice layer to be displaced downward. Finally, although the surface of the thorns was relatively clean after defrosting, a small amount of residual water was attached to the center of the red copper drainage channel due to the gravitational effect, which indicated that some hydrophobicity remained in the Teflon pattern area during and after the defrost process. As displayed in Figure 14, the renewal rate of pure copper surface was lower than that of the thorns; however, the surface renewal phenomenon occurred quicker for polished red copper. Therefore, at the time 45 min, the red copper surface condensate had completely frozen, whereas the surface droplets of thorns began to freeze at 75 min and were completely frozen at 90 min. Finally, the ice layer produced after 2 h of frosting was poured in a 30 • C ethylene glycol solution for defrosting and the melting phenomenon was recorded. As displayed in the image, because the polished copper surface had a low contact angle and high wettability, some condensate remained on the surface and exhibited an irregular filmwise state after defrosting. The surface of the thorns had a high surface contact angle and low wettability. Therefore, a strong repulsive force existed between the surface and the melted water When the thermal solution passed through the internal flow channel, the ice layer in contact with the plate surface as well as the ice surface exposed to the environment melted simultaneously. However, the heat conduction rate decreased because the ice layer was thicker at the middle. Moreover, the melted water in contact with the flat plate was removed, which caused the entire ice layer to be displaced downward. Finally, although the surface of the thorns was relatively clean after defrosting, a small amount of residual water was attached to the center of the red copper drainage channel due to the gravitational effect, which indicated that some hydrophobicity remained in the Teflon pattern area during and after the defrost process. However, some large droplets were stuck between the red copper and Teflon pattern areas because the difference in the surface free energy gradient between the two areas formed an active force opposite to that of the gravitational force, which led to excessive contact angle hysteresis, as illustrated in Figure 15. However, some large droplets were stuck between the red copper and Teflon pattern areas because the difference in the surface free energy gradient between the two areas formed an active force opposite to that of the gravitational force, which led to excessive contact angle hysteresis, as illustrated in Figure 15.   According to the abovementioned observations, three parameters are discussed, namely the heat transfer coefficient during the frosting period, time required for the ice layer to completely cover the surface, and time required for defrosting. Because the surface of the thorns exhibited an incompletely frozen state during the frosting period and steady and frequent condensate renewal occurred continuously, a high surface heat exchange efficiency was achieved ( Figure 16). Therefore, the heat transfer coefficient of the thorns was approximately 16.2% higher than that of the polished red copper sample. Furthermore, the time required for the complete growth of the ice layer was approximately 52.4% higher for the thorns than for polished red copper. Nonetheless, both samples had nearly identical defrosting time efficiencies.
Energies 2019, 12, x FOR PEER REVIEW 17 of 19 According to the abovementioned observations, three parameters are discussed, namely the heat transfer coefficient during the frosting period, time required for the ice layer to completely cover the surface, and time required for defrosting. Because the surface of the thorns exhibited an incompletely frozen state during the frosting period and steady and frequent condensate renewal occurred continuously, a high surface heat exchange efficiency was achieved ( Figure 16). Therefore, the heat transfer coefficient of the thorns was approximately 16.2% higher than that of the polished red copper sample. Furthermore, the time required for the complete growth of the ice layer was approximately 52.4% higher for the thorns than for polished red copper. Nonetheless, both samples had nearly identical defrosting time efficiencies.

Conclusions
In this study, a cold plate with a patterned hydrophobic and hydrophilic mixed surface was used. In addition, a patterned thorn sample was developed with reference to the results of the relevant literatures. Different types of coatings and chemical treatments were designed for the pattern so that it exhibited the condensate drainage effect on the surface without relying on external force and could control the temperature of the inlet solution to cause condensation and frosting on the surface of the sample. Moreover, a camera set in front of the sample was used to simultaneously record the surface images. Finally, the results for the patterned thorn sample were compared with those for the polished red copper sample, and the following conclusions were obtained: (1) Surface renewal was observed in the condensed part of the polished red copper sample.
Moreover, the central drainage effect of droplet adsorption and combination on the patterned thorn samples was significant and occurred frequently. (2) The thorns exhibited the mechanisms of droplet movement, combination, and central drainage, which increased the surface renewal rate and inhibited ice layer growth by 52.4% when compared with that for polished red copper. In addition, the heat transfer coefficient of the thorn sample increased by approximately 16.2% compared with that of polished red copper during the frosting period.

Conclusions
In this study, a cold plate with a patterned hydrophobic and hydrophilic mixed surface was used. In addition, a patterned thorn sample was developed with reference to the results of the relevant literatures. Different types of coatings and chemical treatments were designed for the pattern so that it exhibited the condensate drainage effect on the surface without relying on external force and could control the temperature of the inlet solution to cause condensation and frosting on the surface of the sample. Moreover, a camera set in front of the sample was used to simultaneously record the surface images. Finally, the results for the patterned thorn sample were compared with those for the polished red copper sample, and the following conclusions were obtained: (1) Surface renewal was observed in the condensed part of the polished red copper sample. Moreover, the central drainage effect of droplet adsorption and combination on the patterned thorn samples was significant and occurred frequently. (2) The thorns exhibited the mechanisms of droplet movement, combination, and central drainage, which increased the surface renewal rate and inhibited ice layer growth by 52.4% when compared with that for polished red copper. In addition, the heat transfer coefficient of the thorn sample increased by approximately 16.2% compared with that of polished red copper during the frosting period. (3) Although both samples had almost similar defrosting times, the residual water after defrosting did not form an irregular liquid film on the thorn sample, which had only a limited amount of adhered water, because it had a lower overall wettability and higher repulsive force on the melted water compared with the polished red copper sample. However, some of the large droplets were likely to stagnate at the structural junction due to contact angle hysteresis. surface X x-direction wet-grad gradient of wettability