Boiling heat transfer on modified copper surfaces with inhomogeneous wettability

05012


Introduction
The needs of microelectronics dictate the need to develop more efficient cooling systems, using, in particular, boiling heat transfer.One of the promising ways to intensify boiling heat transfer is the use of surfaces with inhomogeneous wettability or biphilic surfaces [1,2].The idea of using such surfaces is to combine the advantages of hydrophobic surfaces (highly efficient vaporization centers) and hydrophilic ones (high critical heat flux).Usually, for this goal, local hydrophobic regions are formed on the hydrophilic surface.
Biphilic surfaces are characterized by several parameters besides morphology and contact angle: sizes of hydrophobic regions, distances between them, number of hydrophobic regions, and area ratio [3,4].The area ratio is defined as the ratio of the total area of all hydrophobic regions to the total surface area.Some of these parameters are related, but in different situations, some are more convenient than others.
There are many works devoted to the study of such surfaces.Usually, the process of creating experimental samples consists of several stages, some of which are not directly related to imparting inhomogeneous wettability to the surface.These stages can make their own contribution to the enhancement of heat transfer, without additionally imparting biphilic properties to the surface, for example, by changing the roughness.There are much fewer works where the creation of a biphilic surfaces is carried out in one stage [5].
In this work, we compare the intensity of heat transfer on modified surfaces with uniform wettability and on similar surfaces, but with additionally created local hydrophobic areas.

Technologies for creating experimental samples
Below is a table with a brief description of the experimental samples used in the work.Samples #1-#5 were created according to the first technology, which involves the deposition of a fluoropolymer through a special mask on a smooth copper surface or over an array of silicon oxide micrococoons.Samples #6 and #7 were created using the second technology, when arrays of caverns are created on the surface by laser ablation.

Biphilic surfaces with fluoropolymer spots
We used two stainless steel masks with spots 50 µm and 100 µm in diameter located at the nodes of a square lattice with a pitch (distance between the centers of the spots) of 200 µm and 500 µm, respectively.These masks were tightly pressed against the surface of the heater during the deposition of the fluoropolymer coating.As a result, round hydrophobic spots 50 µm in diameter with a step of 200 µm and 100 µm in diameter with a step of 500 µm were formed on the smooth copper surface.The number of hydrophobic spots was 506 and 69, respectively, for each of the masks.A detailed description of the HWCVD fluoropolymer coating is given in [6].
Another type of surface texture for fluoropoymer coating was obtained using micrococoon arrays of silicon oxide nanowires.A micrococoon is a submicronsized tin catalyst particle coated on all sides with SiOx nanowires.Silicon oxide nanowires on tin catalyst particles were synthesized by gas jet electron beam plasma chemical vapor deposition.The process of synthesizing arrays of micrococoons from SiOx nanowires on copper heater surfaces is described in more detail in [7]. Figure 1(a) shows the surface morphology with micrococoon arrays of silicon oxide nanowires before the application of a hydrophobic layer (fluoropolymer film).This surface is #3 in table 1.
The fluoropolymer film was deposited on these heaters by the HWCVD method, as in the previous case, using the same masks and under the same conditions.As a result, round hydrophobic spots 50 µm in diameter with a 200 µm step and 100 µm in diameter with a 500 µm step with arrays of micrococoons were formed on the surface of the copper heater.The number of hydrophobic spots is 506 and 69, respectively, similar to the previous case.The elemental composition of these coatings was analyzed by the EDS method [8].Thus, it was confirmed that the composition of the spots corresponds to the fluoropolymer, and there is no fluoropolymer in the space between the spots.

Biphilic surfaces with arrays of superhydrophobic caverns
Texturing of pre-polished copper surfaces was carried out by laser ablation.After polishing, the substrates were cleaned in an ultrasonic bath in a surfactant solution, isopropanol, distilled water, and dried in nitrogen.Then arrays of caverns with an equivalent diameter of approximately 70 µm were formed on the surface by laser ablation.For texturing, we used an Argent-M laser marking complex with an infrared ytterbium fiber laser with a wavelength of 1.064 μm, a focused beam diameter of 40 μm, and a peak power of up to 0.95 mJ in the TEM00 mode.A single or multiple laser treatment of local surface areas was used with a pulse duration of 200 ns and a pulse frequency of 20 kHz.As a result, arrays of caverns of triangular shape in cross section were formed.The number of caverns is 32.In Figure 2 SEM images of individual cavities at different magnifications is shown.The parameters of the obtained arrays of caverns are presented in Table 1.
After the formation of caverns, the copper surface of sample #6 was hydrophobized.First, the surface was activated using UV radiation, then fluorinated methoxysilane was chemisorbed from vapors at a temperature of 100-110°C.As a result, a layer of twodimensional chemically cross-linked fluoroxysilane was formed, which had a hydroxyl bond with the surface at the sites of ablation [9].It is expected that the hydrophobized layer on a smooth copper surface will be quickly destroyed, and in the areas treated with a laser, it will remain for a long time.The main factor in this process should be different types of adhesion: stronger chemical in laser-treated areas and less strong physical in untreated areas.As a consequence, during the boiling, uneven degradation of the hydrophobic coating will occur, which leads to the rapid destruction of the coating between the caverns and to the relative stability of the coating inside the caverns.Thus, a biphilic surface is formed with arrays of hydrophobic caverns and non-hydrophobic areas between the cavenrs.Sample #7 was created from sample #6 by plasma cleaning.Processing conditions in oxygen plasma of direct current discharge: voltage  450 V, current 100 mA, power 45 W, pressure 1 Torr, oxygen consumption 5.5 cm 3 /min, treatment time 3 min.After this treatment caverns on sample #6 lost their superhydrophobic properties.
The contact angle for the smooth copper surface was 54±2°.After fluoropolymer deposition it was 140±2°.

Experimental setup
The scheme of the experimental setup is shown in Figure 3 showing the main components.The main element of the setup is a thermostated chamber with five optical windows for monitoring the processes occurring in the course of experiments.The chamber had double walls, between which a liquid was pumped by means of a thermostat to maintain a uniform and constant temperature distribution on the chamber walls.On each side of the setup there was a window for visual observation of the course of the experiment, as well as thermocouples for monitoring the wall temperature.Two more thermocouples were installed in the working volume to control the liquid and vapor temperatures.The chamber is filled to the required level with the working fluid through the fluid supply channel.The steam subsequently entered the liquid condenser and returned to the installation.Thus, the installation had a closed loop. .At the bottom of the chamber there is a Teflon cylindrical working area with a copper core, fixed with a special clamping device.The thermocouples used to measure and calculate the heat flux from the heater are distributed over the working area.
The modified surfaces were created on the head (5 mm in diameter) of the copper core.In the core, near the working surface, two thermocouples are located at different depths to control the wall temperature and determine the heat flux.The instrument and control system consisted of an NI-9214 data acquisition system and software.The temperature measurement sensitivity of this device is 0.1 °C.The thermocouples were connected to the cold junction (without using automatic cold junction compensation) to improve accuracy.All thermocouples of the setup were calibrated individually.The calibration error was 0.1 °C.Distilled, deionized, degassed water (Milli-Q) was used as the working fluid.Atmospheric pressure.

Results
Below are the results of pool boiling heat transfer experiments on the samples described above.The Rohsenow correlation [10] is used everywhere as a reference smooth copper surface.
Comparison of the obtained data for samples #1-#3 with each other and with the correlation for smooth copper shows that the preliminary modification of the surface by creating an array of micrococoons did not have a significant effect on heat transfer.The boiling curve of sample #1 (micrococoons without fluoropolymer spots) practically does not differ from the theoretical and experimental curves for smooth copper.At the same time, the intensity of heat transfer on samples #2 and #3 (biphilic, micrococoons with fluoropolymer spots) is significantly higher than on #1.Similar results were obtained for samples #4 and #5, which were created by conformal (that is, without changing the surface morphology) deposition of fluoropolymer spots on a smooth copper surface.The intensity of heat transfer on biphilic surfaces is much higher than on a smooth surface without fluoropolymer.In this case, no preliminary or additional processing was carried out, the fluoropolymer was deposited conformally.When comparing Figures 5 and 6, one more pattern is seen: the intensity of boiling heat transfer on surfaces with fluoropolymer spots of 50 µm in size and a step of 200 µm is higher than on surfaces with spots of 100 µm in size and a step of 500 µm in both cases, regardless of the presence or absence of preliminary modification by the array micrococoons.This fact can be associated both with a larger area ratio of samples #3 and #5 than those of #2 and #4 (5.1% vs. 2.8%), and with a more optimal configuration of hydrophobic spots.
Boiling data on samples with caverns obtained by laser ablation are shown in Figure 7.As in the previous cases, the intensity of heat transfer on sample #6 (superhydrophobic caverns) is higher than on sample #7 (devoid of hydrophobic properties).The intensity of boiling heat transfer on sample #7 in the region of high heat fluxes practically does not differ from a smooth surface.At the same time, the technology for creating and morphology of these samples differs significantly from previous cases.Thus, it can be concluded that it is the biphilic properties that are decisive in the intensification of boiling heat transfer on biphilic surfaces, and not additional concomitant treatment.

Conclusion
Experiments were carried out to intensify heat transfer during boiling on biphilic surfaces of three types, performed using two technologies: deposition of a fluoropolymer through a mask and hydrophobization with fluorinated methoxysilane after preliminary laser texturing.
All described biphilic surfaces were compared with similar surfaces without hydrophobic treatment, that is, non-biphilic (smooth surface, arrays of micrococoons, arrays of caverns).
Surface with micrococoon arrays shown no difference in heat transfer from smooth copper surface.
For all biphilic surfaces, a significant intensification of heat transfer has been achieved, compared with a smooth surface.
For all biphilic surfaces, the key contribution of biphilic properties to the boiling heat transfer is shown in comparison with other associated treatments.
Figure 1(b) shows the surface morphology with arrays of micrococoons after application of a fluoropolymer film.

Fig. 3 .
Fig. 3. Scheme of the experimental setup with the designation of individual elements.

Fig. 4 .
Fig. 4. Scheme of a cylindrical copper core.The boiling surface is the upper end of the core.

Fig. 5 .
Fig. 5. Dependence of the heat flux on the temperature difference for samples with micrococoon array without fluoropolymer spots (#1) and with fluoropolymer spots (#2 and #3).

Fig. 6 .
Fig. 6.Dependence of the heat flux on the temperature difference for samples with fluoropolymer spots on smooth surfaces.

Fig. 7 .
Fig. 7. Dependence of the heat flux on the temperature difference for samples with array of superhydrophobic (fluoroxysilane) caverns (#6) and array of caverns there fluoroxysilane was removed (#7).
The work was funded by state contract with IT SB RAS #

Table 1 .
Brief description of surfaces.