Effect of Mixed Super-Hydrophilic and Super-Hydrophobic Surface Coatings on Droplets Freezing and Subsequent Frost Growth During Air Forced Convection Channel Flows.
Frost accretion is a common occurrence that penalizes energy efficiency of air-source heat pump systems and refrigeration systems. Studies exist in the open literature that examined the effect of surface wettability on frost formation of flat plates and fin structures (Lee et. al, 2004, Wu et. al, 2007, Kim and Lee, 2011, Huang et. al, 2012, Rahman and Jacobi, 2015, Sommers et. al, 2018, Hermes, et. al, 2018) and excellent recent review papers on this topic are the ones of Kim et. al (2017) and Song and Dang (2018). A challenge with frost accretion studies is that frost nucleation and frost growth are transient phenomena and the observations are difficult to generalize to broader range of conditions and surface types. In addition, the instantaneous values of all key variables necessary to develop and validate frost models are often not measured concurrently and distinctly during the same frosting test, or not completely documented and reported. In our previous work, we presented a new model for frost nucleation and frost growth on flat plates (Harges and Cremaschi, 2018a) and included the variability of the surface contact angle on the frost nucleation (Harges and Cremaschi, 2018b). Our model was compared with data from the literature but an in-depth validation was recommended. More recently, we developed a new methodology to measure droplets growth and freezing time as well as frost thickness during frost nucleation and subsequent frost growth processes (Cremaschi et. al, 2018). Results on a mirror-polished Aluminum surface, on a super-hydrophobic-coated, and on super-hydrophilic-coated surface were presented for two mode of runs. In the first mode, the surface temperature had small variations during frosting nucleation. This mode was referred as to "nitrogen displacement test mode" and it provided data that served for future frost models development. In the second mode of run, the surface temperature decreased from ambient to sub-freezing conditions, mimicking the actual behavior of a fin in a heat pump system. This mode was referred as to "pull-down" mode and it was helpful for comparing the frosting performance of the coatings in realistic field type operating conditions. However, each physical quantity varied significantly with time and it was difficult to use the data from this mode of operation to develop physic based frost models.
Building on our previous work, this paper presents new data of frost nucleation and subsequent frost growth for additional three biphilic surfaces. These biphilic surfaces had interlaced wettability and large distinct regions of different wettability. The top and the leading edge of the surfaces were recorded to provide the instantaneous measurements of droplets growth, droplets freezing time, as well as frost growth and frost layer thickness during frost nucleation and subsequent frost growth periods. An in-situ calibrated non-invasive infrared thermal camera measured the frost surface temperature while the test apparatus measured the instantaneous heat flux and frost mass deposited on the surfaces. Our present data of frost thickness and density resulted similar to those in the literature during the frost growth phase. However, we were not able to find data from the literature for the very first few minutes of frost accretion, that is, during the frost nucleation and crystal growth phases, and for the type of surface coatings used in the present work.
The test set up consisted of an airflow wind tunnel that controlled the air stream temperature and humidity. A second smaller airflow wind tunnel, shown in Figure 1(a), was installed inside the large wind tunnel and it accommodated the cold flat plates. Two thermoelectric coolers (TECs) and an in-house built stainless-steel heat flux meter controlled the surface temperature of the test plates during frosting. An infrared camera and a high-definition video microborescope measured temperature of the frost, droplets size during freezing, and subsequent frost thickness. Sensors of the test apparatus shown in Figure 1(a) measured the time dependent heat flux, surface temperature, air dry bulb and dew point temperatures at the inlet and outlet of the test plates, airflow rate, and air pressure. Main derivative quantities were mass of frost, frost density, and the time dependent airside convective heat transfer coefficient. Details of the test apparatus and test methodology are described in authors' previous work (Cremaschi et. al, 2018, and Cremaschi et. al, 2012). Six aluminum plates were machined to dimensions of 25mm length (i.e., depth of the plate along the airflow direction--see Figure 1(b)) by 152 mm width and 6 mm thickness (1 in x 6 in. x 0.25 in). These plates are referred to as the "test plates" throughout this paper. The test plates were exposed to convective airflow frosting conditions on their top surfaces. Air entered the test plates at 5[degrees]C (41[degrees]F) dry bulb temperature. The dew point temperature was 2[degrees]C (35.5[degrees]F), which yielded an entering relative humidity of about 80% and an absolute humidity of 0.0043 kg-water/kg-air. The airflow rate was constant at about 8.5 [m.sup.3]/h (~5 cfm) for the entire tests, that is, during both phases of frost nucleation and subsequent frost growth.
In nitrogen displacement test mode, a nitrogen displacement technique was used to limit the variation of the surface temperature during the first few minutes of the test. Nitrogen gas was metered through the test section before the test started. The dew point temperature reduced significantly and the thermoelectric cooling modules were energized without producing any water vapor condensation. When the test plate temperature was reached, the nitrogen gas flow was stopped, and the air was circulated on the top surface of the test plate.
The test plates were rectangular bar of about 6 mm (0.25 in) thickness. The surface exposed to the air was polished to mirror finish and grooves were machined on the other side of the bar for embedding thermocouples. Haque and Betz (2018) provided a detailed description of the coatings used on the top surfaces of the test plates. The first plate was uncoated Aluminum surface with mirror finish roughness and contact angle of about [theta] [approximately equal to] 75[degrees]. The second plate had its top surface coated with super-hydrophobic ([theta] [approximately equal to] 110-116[degrees]) coating and the third plate had super-hydrophilic ([theta] [approximately equal to] 19-29[degrees]) coating. Additional three coatings were investigated and they consisted of interlaced and mixed regions of super-hydrophilic and super-hydrophobic coatings according to the schematic of Figure 1(b). We refer to these coatings throughout this paper to as biphilic coating no 1, 2, and 3.
Frost Characteristics during Tests with Small Variations of the Surface Temperature
Figure 2 presents frost thickness at the leading edge of the test plates as a function of time for all surfaces (aluminum, super-hydrophobic, super-hydrophilic, and the three biphilic surfaces) during the nitrogen displacement tests. The frost thickness, 5f, was normalized with respect to the height of the channel, [H.sub.channel] = 4 mm (0.16 in), of the cross section of the rectangular duct available to the airflow right above the test plates. It was observed that there was a rapid increase of frost thickness during the first four to six minutes of frosting and the graph in Figure 2(b) presents zoomed-out plots during this period of the frosting tests. The thickness trends for the aluminum and hydrophilic surfaces were similar. They both had significantly lower thickness during the initial frost accretion period. Then, the frost thickness increased slowly and with almost constant slope for the remaining of the frosting period. The surface characteristics affected the time and the initial thickness of the frost after the droplets froze on the test plates. This was also the starting point for the frost growth phase. The biphilic surfaces 2 and 3 behaved similarly to the super-hydrophobic coating surface while the biphilic surface 1 led to frost thickness that was about an average between that of super-hydrophilic and super-hydrophobic coatings. This result was consistent given the parallel configuration of the two types of constituent coatings with respect to the airflow direction for biphilic surface 1. Figure 3 shows the average frost density as a function of time for all surfaces during the nitrogen displacement tests. The density for the aluminum surface was the highest and showed large fluctuations during the early stage of frost growth, as shown in the zoom out of Figure 3(b). Biphilic coatings had small range of densities and, again, biphilic 2 and 3 followed the super-hydrophobic surface trends. Interesting, the biphilic 1 had frost density that was lower than that of each individual type of constituent coating during the first six minute. Then, after 30 minutes, the biphilic 1 resulted in averaging the density data measured for the super-hydrophilic and super-hydrophobic coatings and gradually increased toward the hydrophilic density values toward the end of the test. Error bars are reported for various representative points in all the figures and they accounts for the overall experimental error of the measurements.
Figure 4(a) provides an example of the frost surface temperature as a function of time for each surface during the nitrogen displacement tests. The frost surface temperatures were very similar to each other until about forty minutes into the tests. Then, the temperature dropped for the hydrophobic surface and for the biphilic surface 1, but then began to climb again toward the end of the test. This frost temperatures drop were likely due to melting and refreezing of the very top of the frost layers. As the frost melted, the liquid water seeped into the frost layer and refroze, providing higher density and thermal conductivity of the frost layer. This resulted in higher heat conduction and lower temperature of the frost surface. Then, as frost started to grow on the top surfaces again, the temperature began to increase. This behavior was observed for the hydrophobic surface, it was repeatable, and it was observed again in the experiments with biphilic surfaces 1 and 2 but had lower magnitude. Figure 4(b) shows the airside convective heat transfer coefficient for each surface during the nitrogen displacement tests. The heat transfer coefficients were calculated based on actual instantaneous heat fluxes at the test plate surface, air inlet dry bulb temperature, and frost surface temperatures. The data were normalized with respect [h.sub.0] = 120 W/[m.sup.2]-K measured during steady-state wet tests. During frost nucleation, the heat transfer coefficients were about 60 to 70% lower than that of steady state wet-test value. They also remained quite low during the first six to ten minutes, that is, during the droplet growth and ice crystal growth phases. During frost accretion, the heat transfer coefficients increased over one because the airflow rate was constant and, as frost built up, it reduced the free flow area. Thus, the local air velocity immediately adjacent the top of the surfaces increased significantly. The heat transfer coefficients for the biphilic surfaces were similar, while that for the aluminum surface was the lowest of all of them. Biphilic surface 1 seemed to exceed the heat transfer coefficients of each individual constituent coatings (super-hydrophilic and super-hydrophobic), especially when looking at the data at about forty minutes of frost growth. However, the data converged to close ranges and within the experimental uncertainty.
Frost Characteristics during Pull-Down Tests
Table 1 summarizes the radius of the ice beads, the freezing time, the freezing duration, and the period it took for the frost thickness to block the cross sectional area of the channel until the pressure drop across the air flow above the test plate reached 87 Pa (0.35 in H2O) during the pull-down tests. We recall in here that in this mode of operation, the surface temperature of the test plates decreased gradually from initial temperature of 5[degrees]C, which were in thermal equilibrium with the airflow above the plates, down to -15[degrees]C. During this pull down period, water vapor condensed on the surfaces and large droplets (with respect to those observed in the nitrogen displacement tests) formed before turning into ice beads. In Table 1, the freezing time was defined as the time elapsed between turning on of the thermoelectric coolers for cooling down the test plates and the midpoint of the droplet freezing process. The freezing duration was defined as the number of seconds it took for all the droplets in the viewing window to appear frozen. Right after freezing, the iced droplets on the biphilic surface 2 were the largest and those on the hydrophilic surface were the smallest. The freezing time for the biphilic surface 3 was about 18 minutes, that is, the longest time to freeze among all the surfaces. Super-hydrophilic surface froze very quickly, in only about 12 minutes. The radius of the frozen droplets was dependent on the freezing time, which was in turn function of the type of coating. Even though droplets on hydrophilic surfaces appeared flat and more "spread out", the frozen droplets were much smaller than any other droplets measured on the other surfaces. Ice nucleation on the super-hydrophilic surface occurred quickly, leaving less time for disc-like shaped water droplets to continue to grow in the radial and vertical directions before freezing. On the contrary, the freezing duration and the droplets radius were high for the biphilic surfaces because the super-hydrophobicity characteristic of the biphilic coatings dominated the freezing process. Large mass and volume of the droplets on these surfaces were observed before they turned into ice beads. Interesting, all biphilic coatings had longer freezing time and larger droplets than that measured for each individual constituent coatings.
Figure 5 shows frost thickness vs. time for all surfaces during the pull-down tests. Figure 5a gives thickness data for the entire pull down test period and the trends are clearly not linear. They followed "S" shaped like profiles and Figure 5b zooms out across the 15 minutes period during the transitions of droplets to ice beads, to crystals growth, and to initial frost accretion on the top of the ice beads. These phase changeovers are indicated in Figure 5(b) by variations of the slope of the frost thickness data when plotted vs the elapsed time: the first slope change shows the transition from droplets/iced beads to crystal growth phase and the second slope change represent the switch from crystal growth phase to frost growth phase. After this phase changeover, frost accretion during the remaining frost growth phase increased almost linearly with time. The wettability characteristics of the surfaces coatings affected the elapsed time of the phase changeovers and the thresholds of the frost thickness at the switch to the frost growth phase. Super-hydrophobic surface and biphilic surface 2 had early transitions and high thresholds of the frost thickness before switching to the frost growth phase. This result was due to the presence of large droplets on the surfaces before the droplets froze into ice beads. Biphilic surfaces 1 and 3 also had transitions and thresholds that followed the hydrophobic trends. Aluminum surface had delayed changeovers with respect to all other surfaces and it had the lowest frost thickness during the pull-down tests.
Figure 6 depicts frost density vs. time for all surfaces during the pull-down tests. The evolutions of the frost density with time were similar: they started high, decreased to a minimum value during the transitions periods from crystal growth to frost growth phases, and then they started to increase again during the frost growth phase. A zoom out of the frost densities during the phase changeover period is in Figure 6(b). The initial high density was because droplets grew for some time before they froze. These liquid droplets were more dense and compact than typical porous solid frost, leading to large masses but associated small volumes (i.e., large initial densities). As frost crystals developed on the frozen ice beads, they contributed to decrease the average frost density. This effect continued until minimum values of density were observed. Then the frost behaved as solid that growth and become less and less porous. In pull down tests, the super-hydrophobic and biphilic 2 surfaces had lowest range of densities. While the trends of the density data shown in Figure 6 are, according to authors' opinion, reasonable, caution should be used during comparison among the actual values from the different surface wettability cases. Due to transient nature of the pull-down tests and due to the repeatability of the measurements, the experimental uncertainty on the density data of Figure 6(b) could be quite high. Further investigation and additional tests are needed to confirm the repeatability of the results shown in Figure 6(b).
This paper presented new experimental data of frost nucleation and frost growth on cold flat plates operating in frosting conditions with air forced convective flow. The plate surfaces had different wettability from an uncoated Aluminum surface with mirror finish roughness and contact angle of [theta] [approximately equal to] 75[degrees], to super-hydrophobic ([theta] [approximately equal to] 110-116[degrees]) coating and super-hydrophilic ([theta] [approximately equal to] 19-29[degrees]) coating. Additional three biphilic coatings were investigated. Droplets size, distribution and freezing times, local temperatures, heat fluxes and rate of frost mass accretion were directly measured by using a newly developed experimental technique. Air entered to the test plate at 5[degrees]C (41[degrees]F) dry bulb temperature. The airflow rate was constant at about 8.5 [m.sup.3]/h (~5 cfm) for the phases of frost nucleation and subsequent frost growth.
The frost thicknesses for the mirror-finished Aluminum surface and super-hydrophilic coating were similar. They both had significantly lower thickness during the initial frost accretion period. Then, the frost thickness increased slowly and with almost constant slope for the remaining of the frost growth phase. The surface wettability characteristics affected the elapsed time and the initial thickness of the frost after the droplets froze on the test plates. They also affected the starting point for the phase changeover from crystal growth to frost growth. Two biphilic surfaces behaved similarly to the super-hydrophobic coating surface while a third biphilic surface, which had super-hydrophobic and super-hydrophilic coatings adjacent to each other and in parallel configuration with respect to the airflow direction, led to frost thickness that was about an average between that of the two constituent coatings. During frost nucleation and early frost growth stages, the airside convective heat transfer coefficients were about 60 to 70% lower than that of steady state wet-test value.
During pull-down tests, water vapor condensed on the surfaces and large droplets formed before turning into ice beads. Ice nucleation on the super-hydrophilic surface occurred quickly, leaving less time for disc-like shaped water droplets to continue to grow in the radial and vertical directions before freezing. On the contrary, the freezing duration and the droplets radius were high for the biphilic surfaces because the super-hydrophobicity characteristics of the biphilic coatings tended to dominate the freezing process. Frost thicknesses followed "S" shaped like profiles due to transitions of droplets to ice beads, to crystals growth, and to initial frost accretion on the top of the ice beads. These changeover were identified by variations of the slope of the frost thickness profiles when plotted vs elapsed time. The wettability characteristics of the surfaces affected the elapsed time at which these changeovers occurred and the thresholds of the frost thickness when switching to the frost growth phase. Super-hydrophobic surface and one biphilic surface had early transitions and high thresholds. Mirror-finished Aluminum surface had delayed transitions.
The authors would like to acknowledge and thank the National Science Foundation, Chemical, Bioengineering, Environmental, and Transport Systems Division for supporting the present work through the Award No. 1604084.
We would also like to acknowledge and thank Dr. Amy Betz and her research group at Kansas State University (KS, USA) for providing the coatings of the test plates used in the present study.
[[delta].sub.f] = frost thickness, m (or in)
h = air-side convective heat transfer coefficient, W/[m.sup.2]-[degrees]C (of Btu/hr-[ft.sup.2]-[degrees]F)
[H.sub.channel] = height of the channel above the plate, m (or in)
Cremaschi, L., Harges, E., Adanur, B., and Strong, A., 2018, Frost nucleation and frost growth on hydrophobic and hydrophilic surfaces for heat exchangers fin structures, Proceedings of the 17th International Refrigeration and Air Conditioning Conference at Purdue University, West Lafayette, IN (USA), July 9-12, Paper No. 2508, Pages 1-10.
Cremaschi, L., Hong, T., & Moallem, E., & Fisher, D. (2012). Measurements of Frost Growth on Louvered Folded Fins of Microchannel Heat Exchangers Part 1: Experimental Methodology (RP-1589)(CH-12-034). ASHRAE Transactions. 118.
Haque, M., R., and Betz, A., R., 2018, Frost Formation on Aluminum and Hydrophobic Surfaces, Proceedings of the ASME 2018 International Conference on Nanochannels, Microchannels, and Minichannels (ICNMM 2018), Paper No. ICNMM2018-7609, June 10-13, Dubrovnik, Croatia.
Harges, E., and Cremaschi, L., 2018, A New Model for Frost Growth Incorporating Droplet Condensation and Crystal Growth Phases, ASHRAE Transactions, ASHRAE Conference Paper, ASHRAE Winter Conference, Chicago, IL, USA, Jan 20 - 24.
Harges, E., Cremaschi, L., 2018, Modeling of Frost Growth On Surfaces With Varying Contact Angle, Paper TFEC-2018-20908, Proceedings of the 3rd Thermal and Fluid Engineering Conference, Ft. Lauderdale, FL, Mar 4-7
Hermes, C.,J.,L., Sommers, A.,D., Gebhart, C.,W., and Nascimento, V.,S., 2018, A semi-empirical model for predicting the frost accretion on hydrophilic and hydrophobic surfaces, Int. J. Refrig., xx, pages xx-yy, doi.org/10.1016/j.ijrefrig.2017.09.022
Huang, L., Liu, Z., Liu, Y., Gou, Y., and Wang, L., 2012, Effect of contact angle on water droplet freezing process on a cold flat surface, Exp. Therm. Fluid Sci. 40, pp. 74-80
Kim, K., and Lee, K., S., 2011, Frosting and defrosting characteristics of a fin according to surface contact angle, Int. J. Heat Mass Transfer 54, pp. 2758-2764
Kim, M-H, Kim, H., Lee, K-S, and Kim, D-R, 2017, Frosting characteristics on hydrophobic and superhydrophobic surfaces: A review, Energy Conversion and Management 138, pp. 1-11
Lee, H., Shin, J., Ha, S., Choi, B., and Lee, J., 2004, Frost formation on a plate with different surface hydrophilicity, Int. J. Heat Mass Transfer 47 pp. 4881-4893.
Rahman, M., A., and Jacobi, A., M., 2015, Effects of microgroove geometry on the early stages of frost formation and frost properties, App. Therm. Eng. 56, pp. 91-100
Sommers, A., D., Gebhart, C., W., and Hermes, C., J., 2018, The role of surface wettability on natural convection frosting: Frost growth data and a new correlation for hydrophilic and hydrophobic surfaces, Int. J. Heat Mass Transfer 122, pp. 78-88
Song, M., and Dang, C., 2018, Review on the measurement and calculation of frost characteristics, Int. J. Heat Mass Transfer 124, pp. 586-614
Wu, X., Dai, W., Shan, X., Wang, W., and Tang, L., 2007, Visual and theoretical analyses of the early stage of frost formation on cold surfaces, J. Enhanced Heat Transfer, 14, pp. 257-268
Lorenzo Cremaschi, PhD
Student Member ASHRAE
Student Member ASHRAE
Student Member ASHRAE
Table 1. Frost variables during the pull-down tests for the all surfaces (air at 5 [degrees]C and 78-80 % R.H., 8.5 [m.sup.3]/h) Surface Coating Frozen Droplet Freezing Freezing Type Radius (mm) Time Duration (min) (min) Mirror-finished 0.174 17.6 1.1 Aluminum ([theta] [approximately equal to] 75[degrees]) Super-Hydrophobic 0.139 12.9 2.2 ([theta] = 110-116[degrees]) Super-Hydrophilic 0.0985 11.6 0.9 ([theta] = 19-29[degrees]) Biphilic 1 (parallel) 0.134 / 0.187 15.8 2.5 philic / phobic Biphilic 2 (series) 0.181 / 0.245 16.3 1.8 philic / phobic Biphilic 3 (alternate) 0.131 17.8 1.2 Surface Coating Time until Type [DELTA]p = 0.35 in. w.c. (min) Mirror-finished ~65 ([+ or -]0.5) Aluminum ([theta] [approximately equal to] 75[degrees]) Super-Hydrophobic ~45 ([+ or -]0.5) ([theta] = 110-116[degrees]) Super-Hydrophilic ~57 ([+ or -]0.5) ([theta] = 19-29[degrees]) Biphilic 1 (parallel) ~45 ([+ or -]0.5) Biphilic 2 (series) ~36 ([+ or -]0.5) Biphilic 3 (alternate) ~45 ([+ or -]0.5)
|Printer friendly Cite/link Email Feedback|
|Author:||Cremaschi, Lorenzo; Harges, Ellyn; Adanur, Burak; Strong, Amy|
|Date:||Jan 1, 2019|
|Previous Article:||Design of Multi Nozzles for a Portable Air Flow Meter using Numerical Simulation.|
|Next Article:||Optimization of Microchannel Shape Using Analytical Models for Condensation.|