Flexible polymeric asymmetric vapor chambers incorporating microchannels of extreme wettability for electronic device cooling

Abstract

Flexible and lightweight thermal management technologies are essential for emerging applications such as wearable electronics, soft robotics, and flexible devices, yet traditional metal-based vapor chambers remain too rigid, heavy, and electrically conductive for such systems. In this work, we present a complete microfabrication process to a flexible and ultra-thin (250 μm) asymmetric vapor chamber made from Cyclic olefin copolymer that integrates superhydrophilic microchannels designed to enhance capillary-driven liquid motion inside the device. The device’s architecture enables efficient two-phase heat transport while maintaining mechanical flexibility. The heat transfer performance is evaluated, under varying bending angles (10°, 20° and 45°) and different water filling ratios (28%, 42% and 70%) to identify the optimum configuration of vapor chamber in respect to mechanical deformation, working liquid volume, and surface wettability properties. The results show that the superhydrophilic microchannels treated for 9-min with oxygen plasma and Polyethylene Glycol coating exhibit low and stable thermal resistance across different deformation states (0–20°) which is 20% lower compared to the devices with untreated microchannels. This work provides a sustainable microfabrication approach for flexible passive cooling devices.

1 Introduction

For decades, vapor chambers have been predominantly fabricated from metals, especially copper, owing to their high bulk thermal conductivity (∼400 W/mK for copper) and mechanical durability, compatibility with established sealing methods such as brazing and welding (Yu et al., 2023), and wick fabrication techniques including sintered metal, grooved/mesh structures (Filippou et al., 2024). As a result, metallic vapor chambers have become the standard for high performance computing and industrial applications where maximum heat spreading, and long-term reliability are primary concerns. However, metallic designs have significant constraints, as they are relatively heavy, rigid, electrically conductive, prone to corrosion, and have no flexibility. These drawbacks are increasingly problematic for next-generation applications such as lightweight electronics, flexible or wearable devices, aerospace systems, and consumer products.

Polymeric vapor chambers have emerged as a promising alternative offering low weight, electrical insulation, and compatibility with scalable manufacturing processes such as molding, lamination, hot embossing and 3D printing (Chang et al., 2025). Silicone, polyimide, polyurethane and polyethylene have been used as casing materials in polymeric cooling devices as they present good flexibility and can be shaped with a variety of fabrication methods as stated already. The challenge lies in overcoming polymers’ inherently low thermal conductivity and gas permeability -issues that recent advances in composite fillers, (Slippey et al., 2014), coatings (Xu et al., 2022) and surface functionalization (Tselepi et al., 2025) are beginning to address. By adding carbon-based fillers, the thermal conductivity of polymers like Polyphenyl Sulfone (PPSU) can be increased from 0.35 W/mK to as high as 5.2 W/mK, significantly bridging the gap with metallic performance (Gulia and Sur, 2022). However, even with enhanced bulk conductivity, the thermal performance of a vapor chamber remains critically dependent on its internal wick structure, which governs capillary-driven liquid return. This represents the primary design challenge for fully polymeric devices, as conventional metallic wick structures are incompatible with the flexibility and thickness requirements of next-generation applications. Traditional wick structures in vapor chambers include sintered metal powders, mesh wicks, and microgrooves, each presenting distinct trade-offs between capillary pressure, permeability, fabrication complexity, and compatibility with flexible substrates. Metallic mesh wicks offer good capillary performance but similarly add thickness and manufacturing complexity, and their integration into polymeric devices requires hybrid fabrication approaches that compromise the fully polymeric nature of the device. To illustrate this, Shi et al. (2016), developed a horizontal position polymer-based flat heat pipe which combines a wick structure of polymer grooves that are covered with copper mesh for structural support and high capillary pressure that has been tested in different tilt angles (0°, 45°, 90°). During forced air convection, the device operated between 2.8 and 14.2 W of heat power input with a minimum thermal resistance of 1.02 K/W at 10.3 W which is lower than its copper counterpart of 4.6 K/W. While this performance is competitive, the reliance on copper mesh for structural support and capillary pressure highlights the difficulty of achieving high performance with purely polymeric wick structures. Similarly, Cho and Joshi (2019) showed that patterning copper micropillar wicks on a PCB can improve thermal performance by ∼20% over standard copper-plated alternatives. To further optimize thermal and mechanical performance, researchers have looked toward nature for architectural inspiration. Shen and Guo (2025) proposed a novel biomimetic vapor chamber design inspired by wettability phenomena and leaf vein capillary transport channels. This design utilizes biomimetic wicks with asymmetric branches and wedge-shaped grooves to balance capillary force and permeability more effectively than traditional porous media. Combined with an additive-manufactured honeycomb lattice support structure, this design significantly enhances both thermal and structural performance as it reaches a steady state sooner, shortening startup time by 11.7% at a 20% filling ratio, increasing the heat isothermal performance by up to 7.74% and incorporating composite support columns to reduce maximum shell strain by 48.26% under equivalent loads. Oshman et al. (2011), combined liquid crystal polymer films as the casing material with a wick structure made up of grooves filled with copper mesh and etched micropillars. When tested horizontally, it exhibited an effective thermal conductivity of 830 W/m K at power inputs between 3 and 11.94 W. Frequently, metallic meshes are integrated with microgrooves to optimize capillary pressure and enhance the device’s overall heat transport capacity, a configuration characterized as a composite wick structure. Even though metal sintered wicks are common in metallic and are also used for a lack of a better alternative in polymeric chambers, polymeric wicks are needed but they require different fabrication strategies.

Recent advances demonstrate that ultra-thin and flexible two-phase elements can achieve thermal performance far beyond what their dimensions or base materials would suggest. Ultra-thin vapor chambers (0.5 mm) incorporating composite mesh wicks (Huang et al., 2020) have achieved effective conductivities of up to ∼2.5 × 10⁴ W/mK in gravity-assisted orientation, and heat-transport capacities that depend strongly on filling ratio and orientation, highlighting the importance of fluid inventory for miniaturized devices. Flat copper–water heat pipes employing fine, sintered copper powder wicks (Li and Lv, 2016) have been shown to dissipate very high heat fluxes (∼100 W/cm²) with thermal resistances on the order of 0.2 °C/W in horizontal operation, demonstrating that thin working-fluid layers and carefully engineered wick microstructure can amplify heat spreading relative to solid metal sheets. Composite ultra-thin heat pipes integrating copper foam-mesh hybrid wicks have achieved evaporation and condensation resistances of 0.29 and 0.45 K/W respectively, while highlighting the strong sensitivity of thin devices to under- and over-filling (Zhou et al., 2017).

The development of flexible and ultra-thin heat pipes and vapor chambers has gained significant attention driven by the demands of wearable and conformable electronics. Flexible aluminum flat heat pipes (Liu et al., 2019) fabricated with laser-ablated casings and gradient wetting functional copper mesh wicks maintain low thermal resistance (∼0.525 °C/W) even after repeated bending (0°–180°), although bending can disturb vapor pathways and increase interfacial resistance in the evaporator. Designs incorporating oxidized copper mesh with multi-scale micro/nanostructures and polymer connectors have achieved thermal resistances below 0.01 K/W that remain stable under bending from 30° to 120°, revealing that mechanical deformation does not degrade performance when wetting is sufficiently strong. (Yang et al., 2022); Oshman et al. (2013) fabricated fully polymer-based flexible heat pipes using laminated film casings and sintered mesh wicks that maintain thermal resistances between 1.2 and 3.0 K/W during bending up to 90°, far exceeding those of solid copper references. Demonstrating the growing potential of such systems, Hsieh and Yang (2013), on their study fabricated a flexible polymer-based heat pipe that consists of copper/silicone and has better thermal performance under bending angle of 15° (5 °C/W) than those without bending (6 °C/W). Lewis et al. (2015) using Kapton film (casing polymer) and SU-8 with an ultra-thin TiO₂ film to fabricate the wick, developed a 0.3 mm thick, flexible two-phase heat exchange heat pipe with water as the working liquid. Thermal resistance and effective thermal conductivity were measured and compared to a sample of copper at varying heat power inputs. The optimal filling ratio was 0.15 ml, which resulted in the lowest thermal resistance of 11.92 °C/W. In a previous work we have demonstrated that a fully polymeric, ultra-thin (310 μm) thermal diode can achieve high-efficiency two-phase heat transfer without a conventional wick. By employing plasma micro- and nano-texturing followed by chemical functionalization, we engineered a device comprising by a superhydrophilic evaporator and a superhydrophobic condenser, to facilitate capillary liquid spreading and dropwise condensation, respectively. This “wickless” architecture yielded a heat transfer coefficient of 3650 W/m²·K, representing a 200% performance enhancement over non-textured counterparts (Tselepi et al., 2025). These studies reveal that wick structure and multi-scale structure in particular, is essential to achieving high performance in thin and flexible devices, while filling ratio plus mechanical deformation interact strongly with capillary return and vapor transport.

Surface functionalization and wettability engineering have emerged as powerful strategies to enhance capillary performance and other functionalities in microdevices without adding thickness or metallic components (Gogolides et al., 2015). Laser texturing has been shown to create superhydrophilic wicking surfaces with complex micro-scale grooves and pores. By adjusting laser pulse energy, Gu et al. (2023) enhanced the wicking performance, allowing an ultra-thin vapor chamber (0.4 mm thickness) to achieve a maximum heat transfer capacity of 21 W at 60 °C. Notably, these laser-textured structures maintained effective liquid return even in anti-gravity orientations, which is critical for portable electronics. Rasouli et al. (2022) utilized digital light synthesis 3D printing to develop a high-effectiveness Micro-pin array polymer heat exchanger. Their design replaces traditional metallic tubes with mm-scale “water plates” containing staggered pin fin architectures to reduce form drag for airflow. Under equivalent pressure drops and volumes, the device transferred 55% more thermal energy than a standard finned-tube heat exchanger, demonstrating the potential of polymer-based additive manufacturing in high-performance cooling applications.

In this work, we present a complete, scalable microfabrication process for a flexible, ultra-thin (250 μm) polymeric asymmetric vapor chamber. While prior polymer-based devices rely on the addition of metallic wick components to achieve competitive thermal performance, in the proposed device the wick structure is incorporated in the base plate by creating superhydrophilic microchannels that enable rapid capillary action. The device is characterized as “asymmetric” due to its specific geometric configuration, the evaporator and condenser sections are spatially offset on the same plane, resulting in a horizontal heat transfer path. Extreme wettability is achieved through simultaneous micro-nanotexturing and chemical functionalization via plasma etching of the microchannels. The contribution of this study lies in the fully polymeric configuration with an asymmetric planar vapor chamber geometry, enabling in-plane heat spreading and a systematic experimental evaluation of how treatment duration (0, 3 and 9 min), filling ratio (28%, 42%, 70%) and mechanical bending (10°, 20°, 45°) synergistically govern heat transfer performance in flexible two-phase polymer devices. Results show that plasma-functionalized microchannels significantly enhance thermal performance across all filling ratios and bending angles tested. The device treated for 9 min in oxygen plasma at an intermediate filling ratio of 42% exhibited stable thermal resistance across deformations up to 20°, with degradation observed only at 45°. The lowest thermal resistance value was recorded for low filling ratio (28%) and specific bending angle of 10°.

2 Materials and methods

2.1 Materials

Cyclic olefin copolymer (COC) with thickness of 100 μm, glass transition temperature 142 °C and melting temperature 220–260 °C (obtained from Microfluidic ChipShop GmbH) was used as substrate for the fabrication of the asymmetric vapor chambers. COC was chosen because it shows low shrinkage, low moisture absorption, low birefringence and it is also hydrophobic with water static contact angle ranging 85–90°. Aluminum plates (alloy 1,050) with thickness 0.5 mm, dimensions 8 cm × 3 cm and engraved microchannels by CNC micro milling were used as stamps for pattern transfer. Polyethylene Glycol 8,000 (Sigma Aldrich) has been used as a hydrophilic coating for the stabilization of the hydrophilic properties of the microchannels.

2.2 Pattern transfer with hot embossing

For the pattern transfer with the hot embossing method, two hydraulic presses (Carver, United States and GRAVEDA) were used, that both enable temperature and pressure control. Firstly, the plates of the press are set to the desired temperature (160 °C), that is above the glass transition temperature of the material (142 °C), then the substrate material and the stamp are placed on the plates and pressure is applied (2.5 MPa). After 20 min the temperature is lowered to a temperature lower than the glass transition temperature. When the temperature drops below 60 °C pressure is released, and the samples are demolded.

2.3 Surface micro-nano texturing, chemistry modification and asymmetric vapor chamber fabrication process

The polymeric COC substrates after hot embossing were cleaned prior to their plasma treatment with isopropanol, then rinsed with deionized water and dried with dry nitrogen stream. Plasma etching is performed in a custom-built Inductively Coupled Plasma (ICP) reactor, equipped with a double helical source (operating at 13.56 MHz). The reactor has an anodized grate under which the sample is placed. The grate is in contact with the electrode through bolts and acts as a sputtering resulting in the formation of a micromask on the sample, which enables the micro-nano texturing of the polymeric surface. Three different process durations were studied (3, 6 and 9 min) with plasma processing conditions: 300 W power, 250 V bias voltage, 6 mTorr oxygen gas pressure, 100 sccm gas flow rate, in order to choose the minimum time required to induce the appropriate micro-nanotexture for superhydrophilicity on the substrates. More details, including the schematic representation of this custom build reactor and how roughness control can be achieved using a specially designed electrostatic shield can be found in the work by Zeniou et al. (2021).

To extract the plasma micro-nanotexturing step duration, we have performed experiments with lower duration (3 min). For low etching durations (i.e. 3 min) the roughness created is low and therefore the lowest etching time was set to 3 min, while 9 min correspond to the upper limit in which an extremely low water contact angle is observed. In this type of reactor, these conditions enable highly anisotropic reactive ion etching and the creation of micro-nanotexturing is the combinational result of this process and the small amounts (usually less than <1%) of unetchable material coming from the reactor wall and substrate acting as nano-inhibitors. More details about the process can be found in previous works. The micro and nanotexturing method for polymers is highly reproducible and it has been used by several research groups worldwide to tune on demand the surface topography and chemistry (Palumbo et al., 2019), (Coulon et al., 2022) (Balu et al., 2008). It has been therefore implemented for the fabrication of micro-nanotextured surfaces focused on a plethora of applications (i.e., atmospheric water collection, antibacterial surfaces, functional microdevices, passive fungal proliferation control (Tselepi et al., 2024), electronic device cooling (Tselepi et al., 2025) etc.). The oxygen plasma step except from the creation of the micro- and nanotopography simultaneously modifies the surface chemistry by introducing hydrophilic groups which can be used as binding sites for efficient and homogeneous coating deposition. In order to avoid hydrophobic recovery of the microchannels a hydrophilic 1% PEG coating was deposited by spin coating for 30 s at 5000 rpm to ensure thin and uniform deposition (Nioras et al., 2024). After deposition the surfaces were UV-cured (253.7 nm) for 1 h to stabilize the coating.

For the assembly of the asymmetric vapor chambers the as treated surfaces were cut into pieces with dimensions of 85 × 35 mm. Then, double-sided tape was applied at the edges of the bottom surface, leaving an active area of 80 × 30 mm. A PEEK tube 4 cm long was cut and placed on the bottom surface, so that at least 2 cm extends beyond the surface. Finally, a flat COC sheet was placed on top of the treated surface and the hole stack with the aluminum frames is pressed for sealing, forming the vapor chamber with thickness of 250 μm.

To fill the vapor chambers with DI water, the tube extending from the chamber is connected to a 3-way valve acquired by Labsmith. Initially, air is drawn out using a vacuum syringe, followed by filling with DI water using a filling syringe. Three different filling volumes are achieved: 0.2, 0.3 and 0.5 ml that correspond to different filling ratios (28%, 42% and 70%) as has been already described in (Tselepi et al., 2025). After completing the charge with water, the part of the tube protruding from the diode is placed in the press and sealed. Sealing integrity was validated through continuous visual inspection during bending, which showed no delamination, as well as through the stability of the thermal evaluation results across multiple testing cycles, confirming the absence of fluid leakage. The complete fabrication process of the asymmetric devices is presented in Figure 1.

FIGURE 1

In order to evaluate whether the stamp pattern has been transferred successfully to the COC substrates optical profilometry has been used. The stamp microchannels exhibit height of 100 μm, width 1 mm and spacing 2 mm (Figure 1a). After the hot embossing process the design of the microchannels has been successfully transferred to COC as can be seen in Figure 1b. The height of the embossed channels is also 100 μm and width 1 mm with the spacing 2 mm as shown in the measurement from the optical profilometer (Figure 1b).

2.4 Surface morphology and wetting properties characterization

The characterization of the morphology induced on the surfaces was done by scanning electron microscopy using the JEOL JSM-7401F FEG, at 2 kV beam voltage and an Optical Profilometer (Profilm 3D, Filmetrics). The topography roughness characteristics were extracted using Profilm software and ISO 4287. 3D profiles and line roughness profiles have been also extracted to evaluate topography height, width and spacing.

For the wetting properties characterization, the water static contact angle was measured using Data Physics OCA11 Goniometer system. Our instrument has accuracy of ±1° and resolution ±0.01°. For the water static contact angle measurements, 5 μl of deionized water drops were dispensed on the surfaces, the average value as well as the standard deviation of 3 measurements is calculated using the dpiMAX software. In cases where the water is fully absorbed the static contact angle is approximately zero. Water immersion tests have been performed for at least 3 months at the fabricated polymeric plates to determine their performance.

2.5 Thermal properties evaluation set up

The heat transfer performance of the asymmetric vapor chambers is being assessed using a home-made experimental setup that consists of two main parts. The first part is the heating section, which includes an aluminum plate (dimensions 40 × 40 × 2 mm and thermal conductivity 180 W/mK) that heats the evaporation section of the vapor chamber, one 40 W heating cartridge, and a power supply unit (Axiomet AX-3005DS, accuracy 0.01 W). The second part is the temperature measurement and acquisition section, which consists of K-type thermocouples (operating temperature: −50 °C – 204 °C, accuracy: <400 °C ± 1.5 °C) connected to a microcontroller and a computer for data collection and analysis. The thermocouples are placed as seen in Figures 2a,b and the temperatures of the evaporator and condenser are monitored, with measurements recorded every second. To minimize thermal losses, the heating plate is insulated with a thermal insulating material. The vapor chambers were evaluated in heating temperatures ranging from 35 °C to 85 °C. These temperature values correspond to the evaporator temperature when in contact with the heating block. In Figure 2c the calibration curve shows the different temperature values (evaporator) for different heat source input power values is provided. Thermal resistance (Rth) is used to evaluate the thermal performance of the asymmetric vapor chambers. It is represented as the quotient of the temperature difference between two given points by the heat flow between the two points, and is given by Equation 1,

FIGURE 2

Where T_E is the evaporator temperature, T_C is the condenser temperature and Q the input power. The temperatures of the evaporator and condenser are the average values from k-type thermocouples readings in each section of the thermal diode. Thermal resistance is a metric of the material’s resistance to the heat flux and indicates how much temperature difference is required to transfer a unit of heat flux (in Watts) through the material. The lower the thermal resistance is, the better the thermal conductivity.

Every experimental measurement across all bending angles and filling ratios was performed in triplicate. The data points presented in our figures represent the average of these three independent runs, and the standard deviation was used to verify the high repeatability of the device’s performance. An independent two-sample t-test (Welch’s t-test) to extract the statistical significance of our observations was also performed. It is considered that a p-value < 0.05 typically indicates that the result is statistically significant (Thukral et al., 2023).

3 Results and discussion

3.1 Surface morphology, wetting properties and asymmetric vapor chamber working principle

In Figure 3, SEM images of the surface topography before and after the oxygen plasma micro nanotexturing step for 3 and 9 min are presented. The mechanism responsible for roughness formation is discussed in more detail in Section 2.3 and has been previously explained (Gogolides et al., 2011). In short, micro and nanotexturing is mainly formed due to the anisotropic etching conditions used and the simultaneous deposition of a really small amount of alumina molecules sputtered from the electrode of the reactor. Alumina is unetchable from oxygen plasma, and these particles are creating masking effects that enable the creation of nano-grass -which depends on the material properties and the process duration-that grows to larger micro or hierarchical micro-nano topography features.

FIGURE 3

It is clear that plasma micro-nanotexturing step induces significant morphological changes on the polymeric substrates as there is a really big difference between untreated and all treated surfaces. The 3-min plasma etched surfaces exhibit random and dense hierarchical, micro and nanostructured features that increase in height and form larger structures as the process duration increases further to 9 min. Small nanostructures are also observed on top of these bigger formations for all plasma micro-nanotexturing durations.

The micro and nanotexturing induced to the microfluidic channels cause alteration to their wetting properties. As seen in Figure 4 the results of the ageing tests of the induced wetting properties of the microchannels are presented.

FIGURE 4

The untreated COC films are intrinsically hydrophobic (WSCA = 90°) and as plasma etching is performed roughness increases and results to an extremely low contact angle of 0°. This is caused by the simultaneous increase in surface roughness and chemistry modification happening during the plasma micro-nanotexturing step. The day that the samples are undergone plasma treatment is assigned as Day zero. After 1 week of laboratory storage (ambient conditions 20 °C–30 °C, relative humidity 30%–50%) both the 3- and 9-min treated surfaces regain their hydrophobic nature with contact angles reaching 40° and after 3 months of storage they show even higher contact angles of 128° and 139° respectively.

During the oxygen plasma treatment, micro- and nanotexturing along with polar functional groups are formed on the surface, such as hydroxyl, carbonyl, and carboxyl groups. These functional groups are prone to aging, as they slowly react with the atmosphere, reducing their concentration over time. The increased treatment duration enhances the number of functional groups formed and improves their long-term stability (Nioras et al., 2024), but not for long enough. In order to improve the durability of the hydrophilic properties, a hydrophilic coating (1% PEG 8000) is added on the microchannels. The microchannels that have PEG 8000 even after 3 months do not degrade and a steady superhydrophilic state is achieved since PEG offers hydroxyl functional groups, that promote hydrogen bonding with water molecules.

On top of that, PEG durability and performance has been investigated in our two recent publication in which we observed stable condensation performance over successive operation cycles and stable wetting properties under prolonged water immersion (Tselepi et al., 2025), (Nioras et al., 2024).

In conclusion, the proposed approach results in the transformation of the inherently hydrophobic COC surface into superhydrophilic enabling capillary pumping of water inside the device. Videos of water drops spreading on 3 and 9 min treated surfaces coated with PEG and also videos showing the capillary pumping of water on 3 and 9 min treated, PEG coated open microchannels are providing as supporting information. If one compares the wettability of the two surfaces it is obvious that in 9 min treated surface coated with PEG water is rapidly adsorbed and contact angle becomes zero (superhydrophilic state), whereas in 3 min surface water spreading is slower and contact angle becomes <10°, but not 0° as in the 9 min surface. In addition, in the cross-sectional videos of the microchannels it is observed that in the 9 min treated microchannels water is pumped and no meniscus is formed, whereas in the 3 min surface water is pumped much slower and a meniscus of the water drop is still visible, indicating much slower capillary pumping as expected from the contact angle measurements. This is caused because the porous layer in the 9 min treated surface is higher compared to the 3 min treated surface. We have previously calculated the capillary velocity on plasma treated (with similar conditions), superhydrophilic microchannels made from PMMA Poly (methyl methacrylate) and PEEK (Polyether Ether Ketone) and capillary velocities were found 5.0 × 10³ up to 10⁴ μm s⁻¹ in open microchannels (Tsougeni et al., 2010).

3.2 Heat transfer performance evaluation of ultra-thin flexible asymmetric vapor chambers with and without bending

Three distinct asymmetric vapor chamber configurations have been developed, a) one comprising of a plate with untreated microchannels (untreated), b) one with microchannels of intermediate wetting properties (treated for 3 min) and c) one with superhydrophilic microchannels (treated for 9 min). In the two latter configurations, the microchannels of the thermal diode are designed to have different wetting properties deriving from the different topography created after 3- and 9- minute etching. These different configurations enable the evaluation of the effect of the different wetting properties as well as the microchannel’s roughness scale in each device. Also, different filling ratios were used to determine the optimum filling ratio in each type of device (28%, 42% and 70%). Finally, the devices performance was tested in four bending angles (0°, 10°, 20° and 45°), to also examine their ability to operate under bending. In Figure 5 the heat transfer performance of the three types of asymmetric polymeric devices is presented.

FIGURE 5

Figure 5 presents the thermal resistance of the device as a function of constant input power (6.7 W) that corresponds to evaporation temperature 110 °C, for three different filling ratios (28%, 42%, 70%) and four bending angles (0°, 10°, 20°, 45°) for the devices with untreated and treated microchannels. It is clear that across all filling ratios devices with the treated microchannels exhibit lower thermal resistance values than those with untreated, confirming that enhanced circulation from the condenser to the evaporator area is achieved compensating dry out. This is the result of the capillary action promoted by the plasma textured microchannel surface. The lowest filling ratio, 28%, has the highest thermal resistance for 0° bending at evaporator temperature 110 °C for all three configurations (8.2 °C/W for the untreated, 7.7 °C/W for the 3-min and 7.3 °C/W for the 9-min treatment). At such a high evaporator temperature, insufficient filling with cooling liquid causes rapid evaporation and localized dry-out in the heating region. At this filling ratio, p-value is low (p ≈ 0.001), which indicates that the obtained results are highly reproducible. Increasing the filling ratio to 42% improves performance significantly, as the greater liquid volume ensures a sustained evaporation-condensation cycle with effective latent heat transfer. The condensed liquid is driven back to the evaporator through capillary action, resulting in thermal resistance values of approximately 7.5 °C/W for the untreated devices and as low as 6.3 °C/W for the treated configurations (3 and 9-min). In particular, this improvement translates to a p value of 0.0143 marking this improvement observation as statistically important, however for the 42% filling ratio, the change from 3 min to 9 min of treatment did not result in a statistically significant difference (p = 0.12), implying that 3 min might be “good enough” in these testing conditions. When the filling ratio increases to 70%, the excess liquid volume partially “floods” the microchannels, hindering vapor transport and increasing thermal resistance to 7.5 °C–8 °C/W across all configurations under no bending. Although devices with treated microchannels show some marginal improvement due to their superhydrophilic nature -which promotes water flow across the microchannels’ surfaces rather than flooding-the overall heat transfer performance does not improve drastically at this filling ratio.

Since the fabricated asymmetric devices are flexible, bending angles are introduced as an additional testing parameter to evaluate how mechanical deformation influences the two-phase heat-transfer performance. In the same Figure 5 the thermal resistance is presented for the ultra-thin devices for bending angles 10°, 20°, and 45° for the same three filling ratios. Across all bending angles, devices with treated microchannels consistently exhibit lower thermal resistance compared to untreated ones, demonstrating that enhanced surface wettability continues to positively influence heat transfer performance even under mechanical deformation. In particular, the 9-min treatment provides a statistically significant improvement in thermal resistance in 11 out of 12 configurations tested with p < 0.05. The only outlier observed is the 42% filling ratio at 20° exhibiting a p ≈ 0.208, this is the only configuration where the treatment is not providing statistically significant improvement. This is mainly driven by the large error bar (high variance) in the untreated sample at that specific point.

On the other hand, when bending is induced (10°) at the lowest filling ratio (28%) thermal resistance reaches its minimum value (5.3 °C/W) for the 9-min treated device. This small deformation promotes liquid redistribution inside the device and enhances the return of condensed liquid from the condenser to the evaporator, thereby partially compensating for the dry-out tendency associated with this low filling ratio. However, this beneficial effect is specific to this bending angle, as thermal resistance increases considerably at higher angles. Devices with filling ratio 42% demonstrate a more stable performance across the bending angles with minimum thermal resistance values of 6 °C/W for the 9-min treatment, 6.4 °C/W for the 3-min treatment and 7.5 °C/W for the untreated devices. At a bending angle of 45° the wetting properties of the microchannels no longer provide a significant advantage, as they seem to behave just like the untreated ones with thermal resistance almost 8 °C/W. This behavior is attributed to the geometric distortion induced by large bending angles, which compresses the internal cross-sectional area of the vapor chamber and creates a flow bottleneck. This effect hinders vapor transport from the evaporator to the condenser and restricts liquid return, disrupting the two-phase circulation and raising the overall thermal resistance. When the filling ratio rises to 70% for the devices under bending, treatment time of the microchannels also has limited effect in thermal resistance values. Although from the graph a general enhancement is shown (∼7 °C/W) for devices with both treatment times, it is similar to the performance of the untreated devices (up to 8 °C/W). Overall, it is concluded that the polymeric asymmetric devices that show the best performance are those with the treated microchannels with treatment time of 9 min. The 9-min treated devices appear to successfully modify the microchannels’ surface area promoting improved capillary flow. Statistically, this effect is most pronounced at the 0° (horizontal) and 10° angles, where the p-values are lowest, effectively “saving” the device from the high resistance typically seen in the untreated horizontal state. Moreover, the 9-min treated devices exhibit smaller error bars compared to the untreated samples. This suggests that the surface treatment does not just lower thermal resistance; it also makes the thermal performance more predictable and stable. In particular, across all filling ratios at bending angle (10°), the reduction in thermal resistance compared to untreated samples is statistically significant (p < 0.05).

4 Conclusion

This study presents a comprehensive investigation of the design, microfabrication process and the performance evaluation of ultra-thin (250 μm) flexible, polymeric (COC) asymmetric vapor chambers, with a focus on optimizing heat transfer performance through the control of wetting properties, surface roughness and filling ratio, including under bending conditions. These devices represent a leap forward in thermal management technology due to their unique design, material selection and wick properties which offer considerable benefits over traditional vapor chambers. Unlike traditional designs, these devices eliminate the need for separate wick layers or specially engineered materials that typically occupy internal space and add fabrication complexity and thickness, by performing plasma micro-nanotexturing directly on the embossed microchannels with height 100 μm. In order to identify the optimum filling ratio, multiple filling ratios were evaluated both before and after bending using a custom assembled experimental set-up. It is demonstrated that the devices with superhydrophilic channels after 9 min of plasma micro-nanotexturing exhibit the best performance, achieving thermal resistance as low as 6 °C/W for a filling ratio of 42% even under bending conditions for evaporator temperature 110 °C (p = 0.0143). This enhanced performance is attributed to the superior wetting properties of the microchannels (resulting from plasma micro-nanotexturing and stable surface chemistry provided by the PEG coating) which facilitate uniform and rapid liquid spreading and efficient liquid return to the evaporator. In Table 1, we provide examples of ultra-thin, passive cooling devices from the literature together with their heat transfer capabilities.

Table 1 shows that the thermal diodes presented in this study exhibit excellent heat transfer properties particularly remarkable given that the device is fabricated entirely from a polymeric material (COC). While a direct comparison remains complex due to varying architectures and operating conditions, the table highlights the unique ability of the present fully polymeric design to achieve competitive heat transfer performance. In addition, the absence of a separate wick layer not only reduces the total thickness of the device but also simplifies the fabrication process and reduces the cost per device. In particular, the materials used in this work are commercially available, at a low cost and are transformed into advanced interfaces using plasma micro-nanotexturing within just a few minutes. We note that plasma processing is in principle a dry method with practically no chemical waste. Thus, this approach is considered highly cost effective and environmentally friendly compared to alternatives that require specially designed materials or complex fabrication methods.

References

• 1

  BaluB.BreedveldV.HessD. W. (2008). Fabrication of ‘Roll-off’ and ‘Sticky’ superhydrophobic cellulose surfaces via plasma processing. Langmuir24 (9), 4785–4790. 10.1021/la703766c

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 2

  ChangC.FuB.LiuL.JiangQ.GuoF.SuF.et al (2025). Fabrication and performance evaluation of a polymer-based vapor chamber via 3D printing technology. Int. J. Heat. Mass Transf.246, 127095. 10.1016/J.IJHEATMASSTRANSFER.2025.127095

  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  ChoS.JoshiY. (2019). Thermal performance of microelectronic substrates with submillimeter integrated vapor chamber. J. Heat. Transf.141 (5), 051401. 10.1115/1.4042328

  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  CoulonJ.-F.DebarnotD.Poncin-EpaillardF. (2022). Plasma Texturing of Polymers, 91–119. 10.1007/978-3-030-52264-3_4

  • CrossRef
  • Google Scholar
  • View reference in article

• 5

  FilippouI.TselepiV.EllinasK. (2024). A review of microfabrication approaches for the development of thin, flattened heat pipes and vapor chambers for passive electronic cooling applications. Micro Nano Eng.22, 100235. 10.1016/j.mne.2023.100235

  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  GogolidesE.ConstantoudisV.KokkorisG.KontziampasisD.TsougeniK.BoulousisG.et al (2011). Controlling roughness: from etching to nanotexturing and plasma-directed organization on organic and inorganic materials. J. Phys. D. Appl. Phys.44 (17), 174021. 10.1088/0022-3727/44/17/174021

  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  GogolidesE.EllinasK.TserepiA. (2015). Hierarchical micro and nano structured, hydrophilic, superhydrophobic and superoleophobic surfaces incorporated in microfluidics, microarrays and lab on chip microsystems. Microelectron. Eng.132, 135–155. 10.1016/J.MEE.2014.10.002

  • CrossRef
  • Google Scholar
  • View reference in article

• 8

  GuZ.LiuH.YangK.WangQ.XuH.ZhangL. (2023). Enhancing heat transfer performance in 3D-printed integrated vapor chamber using composite structures. Appl. Therm. Eng.234 (Nov), 121292. 10.1016/j.applthermaleng.2023.121292

  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  GuliaV.SurA. (2022). A COMPREHENSIVE REVIEW ON MICROCHANNEL HEAT EXCHANGERS, HEAT SINK, AND POLYMER HEAT EXCHANGERS: CURRENT STATE OF THE ART. Front. Heat Mass Transf.18. 10.5098/hmt.18.40

  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  HsiehS.-S.YangY.-R. (2013). Design, fabrication and performance tests for a polymer-based flexible flat heat pipe. Energy Convers. Manag.70, 10–19. 10.1016/j.enconman.2013.02.011

  • CrossRef
  • Google Scholar
  • View reference in article

• 11

  HuangG.LiuW.LuoY.LiY. (2020). A novel ultra-thin vapor chamber for heat dissipation in ultra-thin portable electronic devices. Appl. Therm. Eng.167, 114726. 10.1016/J.APPLTHERMALENG.2019.114726

  • CrossRef
  • Google Scholar
  • View reference in article

• 12

  LewisR.LiewL. A.XuS.LeeY. C.YangR. (2015). Microfabricated ultra-thin all-polymer thermal ground planes. Sci. Bull. (Beijing).60 (7), 701–706. 10.1007/s11434-015-0760-9

  • CrossRef
  • Google Scholar
  • View reference in article

• 13

  LiJ.LvL. (2016). Experimental studies on a novel thin flat heat pipe heat spreader. Appl. Therm. Eng.93, 139–146. 10.1016/j.applthermaleng.2015.09.038

  • CrossRef
  • Google Scholar
  • View reference in article

• 14

  LiuC.LiQ.FanD. (2019). Fabrication and performance evaluation of flexible flat heat pipes for the thermal control of deployable structure. Int. J. Heat. Mass Transf.144, 118661. 10.1016/j.ijheatmasstransfer.2019.118661

  • CrossRef
  • Google Scholar
  • View reference in article

• 15

  NiorasD.GogolidesE.EllinasK. (2024). Durable surfaces of both wettability extremes with stable dew harvesting performance during liquid-vapor-phase transitions. Ind. Eng. Chem. Res.63, 20872–20882. 10.1021/acs.iecr.4c02374

  • CrossRef
  • Google Scholar
  • View reference in article

• 16

  OshmanC.ShiB.LiC.YangR.LeeY. C.PetersonG. P.et al (2011). The development of polymer-based flat heat pipes. J. Microelectromechanical Syst.20 (2), 410–417. 10.1109/JMEMS.2011.2107885

  • CrossRef
  • Google Scholar
  • View reference in article

• 17

  OshmanC.LiQ.LiewL.-A.YangR.BrightV. M.LeeY. C. (2013). Flat flexible polymer heat pipes. J. Micromechanics Microengineering23 (1), 015001. 10.1088/0960-1317/23/1/015001

  • CrossRef
  • Google Scholar
  • View reference in article

• 18

  PalumboF.Lo PortoC.FaviaP. (2019). Plasma nano-texturing of polymers for wettability control: why, what and how. Coatings9 (10), 640. 10.3390/coatings9100640

  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  RasouliE.FrickeE.NarayananV. (2022). High efficiency 3-D printed microchannel polymer heat exchangers for air conditioning applications. Sci. Technol. Built Environ.28 (3), 289–306. 10.1080/23744731.2022.2026693

  • CrossRef
  • Google Scholar
  • View reference in article

• 20

  ShenJ.GuoC. (2025). Biomimetic design for enhanced thermal performance of vapor chambers. Energies (Basel)18 (5), 1250. 10.3390/en18051250

  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  ShiB.WangY. B.ShanY. J. (2016). An experimental investigation of thermal performance of a polymer-based flat heat pipe. Heat Transfer—Asian Res.45 (8), 746–757. 10.1002/htj.21187

  • CrossRef
  • Google Scholar
  • View reference in article

• 22

  SlippeyA.EllisM.ConwayB.YunH. C. (2014). Heat Pipe Embedded Carbon Fiber Reinforced Polymer Composite Enclosures for Avionics Thermal Management. 10.4271/2014-01-2189

  • CrossRef
  • Google Scholar
  • View reference in article

• 23

  ThukralS.KovacS.PaturuM. (2023). t-test. “t-test,” Transl. Interv. Radiol., 139–143. 10.1016/B978-0-12-823026-8.00104-8

  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  TselepiV.SarkirisP.NiorasD.TsoukoE.SarrisD.GogolidesE.et al (2024). Functional surfaces for passive fungal proliferation control: effect of surface Micro- and nanotopography, material, and wetting properties. ACS Appl. Bio Mat.7 (7), 4510–4518. 10.1021/acsabm.4c00387

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 25

  TselepiV.NiorasD.FilippouI.EllinasK. (2025). Ultra-thin, thermal diodes incorporating surfaces of extreme wettability, for phase change, heat transfer applications. Appl. Therm. Eng.272, 126419. 10.1016/j.applthermaleng.2025.126419

  • CrossRef
  • Google Scholar
  • View reference in article

• 26

  TsougeniK.PapageorgiouD.TserepiA.GogolidesE. (2010). Smart’ polymeric microfluidics fabricated by plasma processing: controlled wetting, capillary filling and hydrophobic valving. Lab. Chip10 (4), 462–469. 10.1039/B916566E

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 27

  XuF.ZhangM.CuiY.BaoD.PengJ.GaoY.et al (2022). A novel polymer composite coating with high thermal conductivity and unique anti-corrosion performance. Chem. Eng. J.439, 135660. 10.1016/J.CEJ.2022.135660

  • CrossRef
  • Google Scholar
  • View reference in article

• 28

  YangY.ChenS.XuanX. (2022). Experimental study on the thermal performances of a non-pump-driven thermo-activated building system based on flat-plate heat pipe array. Appl. Therm. Eng.215, 118934. 10.1016/j.applthermaleng.2022.118934

  • CrossRef
  • Google Scholar
  • View reference in article

• 29

  YuJ.LiY.ChenZ.XinZ.TangX.ChenH.et al (2023). Experimental investigation of the thermal characteristics of a novel pure-metal-based flexible ultrathin vapour chamber. Appl. Therm. Eng.227, 120354. 10.1016/j.applthermaleng.2023.120354

  • CrossRef
  • Google Scholar
  • View reference in article

• 30

  ZeniouA.SmyrnakisA.ConstantoudisV.AwsiukK.GogolidesE. (2021). One-step control of hierarchy and functionality of polymeric surfaces in a new plasma nanotechnology reactor. Nanotechnology32 (23), 235305–235314. 10.1088/1361-6528/abe6ca

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 31

  ZhouW.XieP.LiY.YanY.LiB. (2017). Thermal Performance of Ultra-thin Flattened Heat Pipes.

  • Google Scholar
  • View reference in article