Scalable micro/nanostructured superhydrophobic surface modifications for enhanced energy efficiency and heat transfer performance in stainless steel and titanium

Condensation refers to the change of a substance from a gaseous phase to a liquid phase, an example of which is the condensation of water vapor in nature. Condensation is used in many industries, such as energy generation and seawater desalination. On a general surface, filmwise condensation is the main phenomenon in which gaseous water vapor is condensed in the form of a film. However, film condensation acts as a factor that reduces energy efficiency as the liquid film formed on the surface interferes with heat transfer. A phenomenon opposite to film condensation is dropwise condensation, which is immediately separated after condensation in the form of droplets, and thus a film is not formed, greatly improving heat transfer efficiency. Because of these advantages, many studies have been conducted, and most studies have induced dropwise condensation by modifying the surface to be superhydrophobic. However, in the case of a superhydrophobic surface, it takes a lot of time and money in the process, so there is a great difficulty in increasing the area. Among them, stainless steel and titanium, which are most materials for industrial heat exchangers, have high robustness, so there are few studies on improving the condensation performance after surface modification due to the difficulty of processing. For this reason, there is a large gap between the currently conducted studies and the actual industry. Our research team succeeded in modifying the surface of a stainless steel and titanium tube the size of an actual heat exchanger into superhydrophobicity with a simple process. We confirmed that the condensation performance was improved on the superhydrophobic surface through experiments under various conditions. By comparing the improvement in the heat transfer performance of stainless steel and titanium under several conditions, the main cause of the performance improvement was proved. This study is expected to play a major role in the eco-friendly future industry where energy efficiency is important by improving the heat transfer performance of stainless steel and titanium, which are mainly used throughout the industry.


Introduction
As energy-related industries continue to expand, there is a growing focus on research that can improve energy efficiency [1]. One of the major topics in this area is phase change heat transfer, as most energy conversion processes involve this type of heat transfer, which can transfer a large amount of heat even with a small temperature difference. Phase change heat transfer includes both boiling and condensation phenomena, and numerous studies have been conducted on these topics [2][3][4][5]. The improvement of heat transfer through condensation, i.e., the increase of energy efficiency, is one of the important research goals in a wide variety of industries such as power generation [6][7][8], ocean desalination [9][10][11], refrigeration [12,13], small electronics [14,15], and other chemical fields [12]. Condensation occurs when vapor comes into contact with a surface or fluid that is at a temperature lower than the saturation temperature [16]. This phenomenon can be classified into two types: filmwise condensation and dropwise condensation, depending on the mechanism involved [17,18]. Dropwise condensation occurs in a discontinuous form, with condensation proceeding in the form of droplets. In contrast, filmwise condensation proceeds in the form of a continuous film. Filmwise condensation has lower heat transfer performance than dropwise condensation because the film formed during condensation creates resistance to heat transfer [19,20]. As filmwise condensation typically occurs on ordinary metal surfaces, efforts are being made in the industry to induce dropwise condensation by applying hydrophobic coating agents to surfaces [21,22].
Miljkovic et al [23] discuss the use of superhydrophobic nanostructured surfaces to enhance heat transfer performance during condensation. The authors demonstrate a simple and cost-effective method to achieve highly efficient jumping-droplet condensation heat transfer, which can be useful for applications such as atmospheric water harvesting and dehumidification. However, it is important to note that the findings of this paper are limited to low supersaturations (<1.12) and may not be applicable to higher supersaturations. Rykaczewski et al [24] investigate the condensation behavior of low-surface tension fluids on three types of omniphobic surfaces: smooth oleophobic, re-entrant superomniphobic, and lubricant-impregnated surfaces. The study demonstrates that properly engineered surfaces can promote dropwise condensation of low-surface tension fluids, resulting in a four to eight-fold improvement in the heat transfer coefficient. However, it is important to note that the study only investigates the condensation behavior of low-surface tension fluids on three types of omniphobic surfaces, and therefore, the findings may not be applicable to other types of surfaces or fluids. Additionally, the study is limited to laboratory experiments and may not fully capture real-world conditions.
High heat-resistant metals such as stainless steel and titanium are predominantly used in power generation applications involving intense heat. Although a significant amount of research has been carried out to improve the condensation performance of widely-used metals like copper and aluminum, studies focusing on the enhancement of stainless steel and titanium condensation performance are noticeably scarce. This is primarily because these metals have high mechanical strength and chemical stability, making surface modifications challenging and limiting the available methods for changing surface wetting properties. Wu et al [25] developed a superhydrophobic surface with microstructures on stainless steel utilizing a femtosecond laser. They achieved evenly spaced microstructures with a diameter of 180 μm, a gap of 30 μm, and a speed of 1 mm s −1 using a laser system capable of applying 800 nm central-wavelength pulses to a 15 mm diameter and 0.2 mm thick disk. Furthermore, they accomplished superhydrophobicity on the stainless steel surface through a silanization process. Lian et al [26] created microstructures on titanium surfaces employing a nanosecond laser beam. They produced a microstructure with a pulse duration of 100 ns, a power of 10 W, and a scanning speed of 500 mm s −1 , and achieved a superhydrophobic titanium surface by heating at 100°C for 24 h.
However, the aforementioned research has limitations as it cannot be implemented on a large scale due to the lengthy duration and very high costs associated with the procedure. Moreover, the above method can only be processed on a flat surface, meaning it cannot be applied to shapes having curvature, such as an industrial heat exchange tube. In other words, there are no study cases of improvement in condensation performance through stainless steel and titanium superhydrophobic surfaces applicable in real industrial environments.
In this study, our researchers aimed to achieve several objectives, all of which were accomplished: (1) Creation of a condensation performance test model for condensation performance testing.
(2) Optimization of the process to produce superhydrophobic stainless steel and titanium surface tubes, and success in surface modification.
(3) Conducting long-term repeated condensation experiments to confirm the robustness of the superhydrophobic surface and demonstrate its suitability for industrial applications.
(4) Derivation of the degree of heat transfer improvement the superhydrophobic surface offers compared to bare surface tubes.
(5) Attempt to understand the reasons for the difference in heat transfer enhancement between stainless steel and titanium by measuring the dynamic contact angle to obtain and compare the contact angle hysteresis values, analyzing the micro/nano structure through SEM imaging, and finally, revealing the correlation between the surface structure and the droplets generated during the condensation process.
Our research will be of great significance for increasing energy efficiency in many industries as well as in energy-related fields, which will become increasingly important in the future.

Data reduction
The Energy Balance equation is typically employed to determine the amount of heat transferred, and it requires the temperature difference between the inlet and outlet of the heat transfer tube. However, in this experiment, the measured temperature difference is too negligible. Instead, we've opted to measure the volume of condensate that falls from the heat transfer tube, which will allow us to compute the overall heat transfer coefficient. The equations necessary for this calculation are laid out below.
We have modified the latent heat equation, which we refer to as equation (3), to calculate equation (2). The computation of equation (3) requires the surface temperature of the heat transfer tube. However, due to the superhydrophobic nature of the tube surface used in this experiment, it is impossible to physically measure the surface temperature with a thermocouple. To overcome this challenge, we have employed the Log Mean Temperature Difference (LMTD) method and an iterative calculation approach to estimate the surface temperature. Assuming about the surface temperature allows us to calculate equation (3) and derive the overall heat transfer coefficient from equation (4). We then follow the iteration calculation to ultimately determine the condensation heat transfer coefficient.

Experimental equipment
The experimental setup for studying the external condensation heat transfer is shown in figure 1(a). The primary objective of this experiment is to observe the type of condensation that occurs on the exterior of the tube and to understand how the heat transfer performance alters when the tube is placed within a chamber, with steam flowing around the outside of the tube and coolant circulating within it. A visualization window situated in the middle of the chamber facilitates real-time observation of the condensation process. A model of this process is depicted in figure 1(b). Figures 1(c), (d) show images before and after the condensation experiment, respectively. In the case of the tube treated with a superhydrophobic surface, which will be elaborated on later, droplets form on the surface, indicating dropwise condensation.
The system features an open loop designed for counter-flow operation with a cooling loop and a steam generator loop. (figure 2(a)) The tubes used in most of the experimental loop are 1-inch stainless steel tubes, while the primary loop of the steam generator uses 0.5-inch tubes. The test section is a rectangular stainless steel chamber with dimensions W 1200 × L 500 × H 500 mm. A 1200 mm width is used to ensure fully developed flow within the heat exchanger tubes. A thickness of 10 mm is chosen to withstand external pressure at 0.02 bar vacuum conditions. Level gauges are attached to the bottom of the test section to measure the flow of condensed water and heat loss due to external factors. Collected condensate is stored in a 60 l auxiliary tank below the test section, designed to accommodate condensation experiments for up to 4 h at maximum steam flow rate. Two heat exchanger tubes can be installed inside the test section, allowing the examination of the condensation heat transfer performance of two tubes with different properties under the same conditions. The test section features acrylic visualization windows on both sides and is connected to a vacuum pump capable of maintaining 0.02 bar pressure.
The steam generator has a 10 kW capacity and can produce up to 15 kg hr −1 of steam, with the flow rate adjustable via a needle valve. An auxiliary tank and heater are used to supply heated water to the steam generator to prevent cold tap water from interfering with steam production. A steam-liquid separator is installed at the steam generator's output to ensure only pure steam is supplied to the test section. The cooling system has a maximum capacity of 16.2 kW, more than sufficient to cool the steam generator's maximum heat output of 10 kW. The cooling system uses tap water as the working fluid and can maintain water temperatures between 5 and 35°C. In the experiments, the cooling water temperature is set at 15 ± 0.1°C. Measurement devices include K-type thermocouples, pressure transducers, paddlewheel flow meters, and differential pressure-based steam flow meters. Five K-type thermocouples are used in the experiment with an accuracy of ±0.15°C. Pressure  transducers have a measurement error of ±0.25% in the 0-2.1 bar range. Paddlewheel flow meters are installed to measure cooling water flow rate within a 10-110 lPM range with an error of ±1%. The steam flow meter has a measuring range of 4-20 kg hr −1 with an error of ±1% [28].
To verify the properties of the materials, we used two experimental setups, which are described below. We inspected the stainless steel and titanium surfaces using a Scanning Electron Microscope. (figure 2(b)) The properties of the contact angle were evaluated using a droplet analyzer. (figure 2(c)) For each test specimen, contact angles were measured by 5 μl droplets at five separate points on the surface. Subsequently, we calculated the mean and standard deviation for these measurements.

Experimental conditions for condensing performance evaluation
The experiment was carried out following the steps laid out in tables 1, 2. To avoid any effects from noncondensable gases during the experiment, we managed the saturation pressure by controlling the amount of steam we put into the vacuumed chamber. We also adjusted the temperature of the coolant and the Reynolds number based on what was required in the test matrix in tables 1, 2. The coolant flow rate corresponded to Reynolds numbers of 10,000 and 20,000, while the saturation pressure values were 0.2 bar, 0.4 bar, and 0.6 bar. For a Reynolds number of 10,000, the saturation pressure was increased incrementally from 0.2 bar to 0.6 bar, assigning Conditions 1 through 3 for each 0.2 bar interval. Once the saturation pressure reached 0.6 bar, the supersaturation level was elevated by adjusting the Reynolds number to 20,000 (Condition 4), followed by a decrease in saturation pressure to 0.4 bar (Condition 5) [29].

Fabrication process of the superhydrophobic stainless steel tube
While a variety of methods exist to fabricate superhydrophobic and superhydrophilic surfaces, each approach has notable limitations. The use of a femtosecond laser, for instance, is hampered by the cost and complexity of the necessary setup. Additionally, the projective nature of the process makes it challenging to apply to threedimensional surfaces, thus limiting its scalability and practicality in an industrial context. Yu et al leveraged a strong acid mixture to create surface roughness and then applied a fluorosilane coating to the surface [30]. However, their method didn't account for the reconstruction of the stainless steel's passivation layer, a crucial element in ensuring corrosion resistance. This oversight could compromise the integrity of the stainless steel surface over time. In short, although these techniques can modify stainless steel surfaces to be superhydrophobic or superhydrophilic, their limitations present significant challenges to durability, effectiveness, and scalability.
All tubes used in this study have the following characteristics. (Outer diameter: 25 mm, thickness: 2 mm, length: 500 mm) A passivation layer that prevents corrosion exists on the surface of stainless steel, which is a factor that hinders the surface modification process because it is chemically robust [31]. Therefore, we immersed in a mixed solution of 35% ferric chloride solution and hydrochloric acid 2 M for 4 h to remove the passivation layer and create microstructures at the same time. Then, it was immersed in a 20% hydrogen peroxide solution for 30 min to create a nanostructure on the microstructure. Finally, the stainless steel tube was made superhydrophobic by creating a self-assembled monolayer (SAM) on its surface, which was achieved by immersing it in a mixture of trichloro(1H,1H,2H,2H-perfluorooctyl)silane (HDFS) and hexane for a period of 2 h (figure 3).

Fabrication process of the superhydrophobic titanium tube
In the creation of the superhydrophobic titanium tube, we initially conducted chemical etching to develop a microstructure on the curved surface. The etching process involved immersing the tube in a 70% sulfuric acid solution at 40°C for a duration of 24 h [27]. Following this, we subjected the etched tube to a 30-minute ultrasonic cleaning in distilled water, then dried it in a 60°C oven for approximately 1 h. Subsequently, we carried out anodization to form nanostructures atop the microstructures, using an electrolyte composed of ethanol and ammonium fluoride at a consistent 40 V voltage for 4 h. Similar to the prior step, we carried out a 30-minute ultrasonic cleaning before drying it in a 60°C oven for 1 h, successfully generating nanoholes on the microstructure. Lastly, the tube was immersed in trichloro(1H,1H,2H,2H-perfluorooctyl)silane (HDFS) for 2 h and oven-dried at 60°C for an additional 2 h to complete the self-assembled monolayer (SAM) coating (figure 4).

Condensation experiments and additional analysis to increase the reliability of the experiment data
We conducted further testing on the superhydrophobic stainless steel and titanium tubes under long-term conditions. This was done to evaluate whether the condensation performance would degrade over time. During the long-term testing, we aimed to allocate equal periods for each structure. However, due to factors such as optimizing the experimental equipment and scheduling, there were slight deviations in the periods. The results are presented in figures 5(a), (b), and tables 3, 4. On the whole, we observed no substantial change in condensation performance over time. These findings reaffirmed that the aforementioned surfaces encountered no issues during repeated condensation tests and can withstand prolonged use. We calculated the uncertainty of the overall heat transfer coefficient using equations (2) and (4). During this process, we assumed that the level gauge of the condensate quantity had an error of ±3 s, and the thermocouple had an error of ±0.8 K. The uncertainty analysis revealed that the uncertainty of the overall heat transfer coefficient for each condition did not exceed 5%. Figures 5(c), (d) present graphs of the total thermal resistance for stainless steel and titanium, respectively, represented in linear scale according to their condition number. The conductive thermal resistance was found to be identical for both the superhydrophobic and bare tubes across all conditions, while the convective thermal resistance was observed to be determined by the Reynolds number. In the 0.2 bar condition, where the overall heat transfer coefficient showed a significant increase in the superhydrophobic tube compared to the bare tube, the proportion of the condensation thermal resistance in the total thermal resistance was observed to decrease by about 60% for both. Similarly, in the 0.4 bar condition, which showed a similar increase in the overall heat transfer coefficient, this proportion decreased to 49.9% and 54%, respectively. In the 0.6 bar condition, which also showed an increase in the overall heat transfer coefficient, the proportion of condensation thermal resistance decreased to 43% and 51%, respectively. Ultimately, the analysis of thermal resistance also observed a  similar trend in terms of improvements in the overall heat transfer coefficient and condensation performance. The uncertainty of the thermal resistance was found to be between 8.5% and 14.8% under all conditions.
The efficacy of each test specimen was assessed by comparing the overall heat transfer coefficient (U). We performed a condensation heat transfer performance analysis on superhydrophobic surface-modified stainless steel and titanium tubes. Figure 6 presents a graph that depicts the variation in the overall heat transfer coefficient value of the surface-modified stainless steel tube. We witnessed dropwise condensation under Conditions 1 and 2, that is, at low supersaturation levels, resulting in a significantly enhanced heat transfer coefficient for the superhydrophobic tube in comparison to the bare tube. (figure 6(a)) As the experiment progressed from Condition 3 to 4, the dropwise condensation gave way to flooded condensation ( figure 6(b)). Under Condition 5, no notable difference was seen in the heat transfer coefficient between the bare and superhydrophobic tubes as the surface was entirely flooded ( figure 6(c)).   Figure 7 represents a graph showcasing the changes in the overall heat transfer coefficient value for the bare and superhydrophobic titanium tubes. Under Condition 1, dropwise condensation was evident on the superhydrophobic tube ( figure 7(a)), but it transitioned to flooded condensation in Conditions 2 and 3. (figure 7(b)) In Conditions 4 and 5, the tube was fully flooded, and no significant difference was seen in the heat transfer coefficient between the superhydrophobic and the bare tubes. (figure 7(c)) Unlike the results from the stainless steel data, we noticed a relatively minor improvement in the heat transfer coefficient for the superhydrophobic tube versus the bare tube under Condition 1. Additionally, the shift from dropwise to flooded condensation occurred at a relatively low supersaturation level under Condition 2.

3.2.
Comparison of heat transfer enhancement on stainless steel, titanium superhydrophobic surface compared to bare surfaces Table 5 presents the overall heat transfer coefficient values for stainless steel and titanium under each condition. Upon comparing the heat transfer coefficient values for bare and superhydrophobic surfaces, we observed an increase in all conditions (Conditions 1-5). This finding confirms that the heat transfer performance of superhydrophobic surfaces is superior to that of bare surfaces in both stainless steel and titanium. The improvement rate analysis revealed that stainless steel exhibited the highest improvement rate of about 400% under Condition 1, while titanium showed the highest improvement rate of approximately 260% under Condition 2. Despite having the same superhydrophobic surface, stainless steel exhibited a higher rate of heat  transfer improvement compared to titanium. There are several reasons why stainless steel has a higher rate of heat transfer improvement than titanium, even with the same superhydrophobic surface.

Surface structures of superhydrophobic stainless steel and titanium surfaces
The results of micro/nanostructure observation of the stainless steel surface are shown in figure 8. Figures 8(a),

Dynamic contact angle and contact angle hysteresis
To fully understand the behavior of droplets, it is necessary to know not only the static contact angle (θ 0 ), which is the contact angle measured in a static state, but also the dynamic contact angle (θ D ), which occurs when the droplet moves under gravity due to the presence of a surface slope. Hydrophilicity and hydrophobicity are divided based on a static contact angle of 90 degrees, which is considered superhydrophobic when it is greater than 150 degrees. However, in the real world, there are many forces acting on a droplet, and the droplet will inevitably move, which is one of the reasons why the dynamic contact angle is important. When a droplet is pulled down by gravity on an inclined surface, the contact angle at the part where the droplet is driven in the direction of the force is called the advancing contact angle (θ a ). On the other hand, the contact angle on the side that is away from the direction of the droplet's movement is called the receding contact angle (θ r ). ( figure 10) The value of the advancing contact angle minus the receding contact angle is called contact angle hysteresis (CAH =θ a −θ r ), which is one of the important values that can be quantitatively evaluated for the behavior of droplets on a surface [32]. Superhydrophobic surfaces are distinguished by their θ 0 , typically above 150 degrees, and minimal CAH. The maintenance of a low degree of hysteresis is of utmost importance as it signifies a diminished capillary pinning force acting between the substrate and the droplet.

Predicting condensation performance with contact angle hysteresis
As illustrated in table 6, both stainless steel and titanium are superhydrophobic, with static contact angles exceeding 150 degrees. However, measurements of contact angle hysteresis show that stainless steel is lower than titanium. In this experiment involving saturated steam, numerous very small droplets are concurrently produced on the surface. For optimal heat transfer efficiency, most of these droplets must fall off without adhering to the surface. As you can see in figure 11, stainless steel exhibits weak partial wetting, where the droplets do not stick to the surface prior to coalescing, and most of them fall off, leaving only a small portion of the structure slightly wet. Conversely, titanium exhibits strong partial wetting, where droplets form and infiltrate the structure simultaneously. Therefore, even though both materials undergo dropwise condensation, there is a noticeable difference in heat transfer performance. The primary factors contributing to the force with which a surface attracts droplets are the geometry and size of the surface structure.   Zarei et al [33] suggested that the size and shape of micro/nanostructures can influence condensation heat transfer, particularly with respect to dropwise condensation performance. They conducted a theoretical comparison and analysis of the differences in overall heat transfer according to the height, roughness, and shape of two different structures. Their findings indicated that a reduction in structure height in both structures led to an improvement in heat transfer performance by roughly 3.5 times. Given that the aspect ratio of the microstructure on titanium is larger, the absolute height of the structure is greater than that of stainless steel. This confirms that the stainless steel surface demonstrates superior heat transfer performance.
A more specific reason for the enhanced rate of heat transfer improvement in stainless steel is due to the premature loss of superhydrophobicity in titanium. Jo et al [34] sought to analyze the reasons for a superhydrophobic surface losing its properties during condensation. The cause was identified as a partially wetted condition wherein droplets penetrate the structure due to pressure condensation starting within the structure. They proposed two strategies for ensuring the smooth removal of droplets from hydrophobic micro/ nano-structured surfaces. One strategy involved maintaining the Cassie-Baxter state by promoting nucleation on the surface structure when the gap size is smaller than the critical size under supersaturation conditions. The second strategy proposed was reducing droplet coalescence, or alternatively, intentionally inducing coalescence to gather larger droplets and drop them off the surface. For these surfaces, the gap size is unavoidably larger than the critical gap size due to the etching process used to create the microstructure. Consequently, coalescence is inevitable on both stainless steel and titanium surfaces. During this process, coalescence occurs more frequently with stainless steel, which allows for rapid droplet detachment and contributes to improved heat transfer performance.
Suboptimal droplet detachment can result in droplets remaining within the surface structure, which can in turn impede heat transfer performance. In our experiment, as we moved from Conditions 1-5, there was an increase in the supersaturation level, leading to a rise in the number of droplets infiltrating the structures. The aspect ratio of the structures on stainless steel was relatively low, facilitating easier droplet detachment. However, the structures on titanium had a higher aspect ratio than those on stainless steel, making it more challenging for droplets to detach once they have penetrated the structures. These results suggest that the high aspect ratio of the microstructure on titanium surfaces results in an increase in the degree of supersaturation, which impedes the smooth desorption of droplets that have penetrated the structure, ultimately leading to a decrease in heat transfer performance.

Conclusions
Our research aimed to bridge the gap between laboratory-scale studies and real-world industrial applications. We focused on enhancing the condensation performance of two materials commonly used in industrial heat exchangers: stainless steel and titanium. Prior research methods often involve complex procedures and high costs, making them impractical for large-scale industrial applications. Our work has overcome these challenges by adopting a more practical, cost-effective approach to surface modification, enabling real-world application in industrial heat exchangers. (1) In order to evaluate the overall heat transfer coefficient and heat resistance, we carried out a vapor condensation experiment using both bare stainless steel and titanium tubes, and superhydrophobic stainless steel and titanium tubes in the same experimental setup. We monitored changes in condensation behavior under various supersaturation conditions and assessed the average heat transfer coefficient (U value) for each surface.
(3) We utilized a low-cost and scalable micro/nano fabrication method to modify the surface of stainless steel and titanium steam condenser tubes, measuring 500 mm in length. The process conditions involved cleaning/etching, oxidation, and self-assembly monolayer coating. To ensure the stability of the surface structure, we conducted SEM analysis.
(4) We conducted dynamic contact angle measurements to verify contact angle hysteresis (CAH). It was discovered that CAH provided a more accurate prediction of the droplet's interaction with the surface, and there was a correlation with condensation behavior and performance. Surface-treated stainless steel had a lower CAH value of 12 degrees compared to titanium's 21 degrees. This demonstrated that the droplet detachment phenomenon was more active on the surface-treated stainless steel with a lower CAH, leading to improved condensation heat transfer performance.