8

th

International Conference on Multiphase Flow ICMF 2013, Jeju, Korea, May 26 - 31, 2013

High-speed visualization of droplet impingement on high temperature solid and liquid metal surfaces
Nejdet Erkan1, Tetsui Yasui1, Koji Okamoto1
1

Nuclear Professional School, The University of Tokyo, Japan

Keywords: Droplet, impingement, sub-atmospheric pressure, boiling, liquid metal, visualization

Abstract Liquid droplet impingement boiling on the hot surfaces has crucial importance for the industries employing spray cooling applications Although numerous experimental and computational efforts exist in the literature, several contradictory approaches, pertaining to the thermal-hydraulic behaviour of droplets, are encountered due to the parametric sensitivity of the phenomena. Complex and uncontrollable boundary conditions, such as surface roughness and ambient pressure, may affect the results in variety of ways that cause one to end up with irrelevant conclusions in every occasion. In order to understand the physical mechanism behind some series of experiments are performed under atmospheric and sub-atmospheric pressures with heated cupper plate and liquid metal which eliminates the effect of surface roughness. A distinct boiling behaviour is observed on the surfaces with/without surface roughness. On the polished cupper surface droplet keeps its integrity with intermittent touching to the surface and bouncing back from the surface at 313 oC, in contrast, trapped vapour layer bursting from periphery breaks the droplet suddenly into small flying out liquid fragments on liquid metal surface. In addition to the surface roughness, ambient pressure is decreased to 50 kPa at which time delay of boiling incipience follows different tendencies in both ambient pressures at similar super heat values suggesting that lesser air molar concentration in the environment cause remarkable delays in the incipience of boiling.

Introduction Hydrodynamic and thermodynamic characteristics of liquid droplet impingement onto hot-surfaces have profound significance for various industrial applications, which utilize spray cooling, such as ex-vessel cooling for in-vessel melted core retention in case of a nuclear power plant (NPP) severe accident, hot-core reflooding by the emergency cooling systems of an NPP, cooling of electronic systems, combustion inside an engine, metallurgical quenching and spacecrafts heat rejection systems etc. Spray cooling is an effective method of cooling employing liquid droplets fragmented by a spray nozzle and directed to a targeted hot surface. The droplets spread on the surface and evaporate or form a thin liquid film removing large amounts of energy due to the latent heat of evaporation. Heat transfer rates, much higher than in pool boiling, can be obtained with sprays since less resistance exists against the removal of vapor film formed in between the hot-surface and liquid (Celata et al 2009, Mohapatra et al 2012, Kim 2007 and so on). In essence, that film layer is speculated to be a mixture of vapor and ambient gas, and it degrades the heat flux in wall-liquid interface severely. Although the spray systems include multiple droplets and an individual droplet behavior cannot be extrapolated to the dynamics of large spray systems, the physics behind still demands fundamental investigations in simplified scales to develop efficient cooling systems.

Although numerous studies, dedicated to the understanding of thermal-hydraulic mechanism of the droplet impingement onto hot-surfaces, exists in the literature, the number of investigations still continues to increase due to the recent availability of more advanced and high-speed experimental and computational tools. Even though those noteworthy amounts of progresses in this branch of research available, contradictory approaches related to the boiling physics of a droplet have arisen due to the parametric sensitivity the phenomena which are originated from the complex uncontrollable boundary conditions (BC) (e.g. surface properties, ambient conditions and so on) of the problem (Moreira et al 2010). Heat transfer from the hot-surface to the droplet is dictated by several parameters which are needed to be identified for most of the cooing applications due to the variety of surface properties determining the critical parameters such as critical heat flux, Leidenfrost boiling point etc. In this study, we presented a parametric study addressing significant factors of droplet impingement boiling phenomena with the experimental results obtained from high-speed visualization tests performed under well-controlled BCs. Polished and etched cupper plate and a liquid metal, contained in a stainless steel pot, services as a hot surface under atmospheric pressure and sub-atmospheric pressure. Liquid metal (U-Alloy) as a hot-surface provides a smooth surface due to its high surface tension that discards the major effects of surface roughness. Additional to the elimination of surface roughness effect, droplet impingement on liquid metal

th

International Conference on Multiphase Flow ICMF 2013, Jeju, Jeju Korea, May 26 - 31, 2013

surface under sub-atmospheric atmospheric pressure is investigated to elucidate interference of air molecules with vapor-liquid interface which forms a heat flux resistance controlling the boiling behavior.

Experimental Facility xperimental setup for the A schematic layout of experimental atmospheric and sub-atmospheric atmospheric pressures is shown in Figure 1. Precision flow controller drives the distilled water flow with a syringe pump. . A plastic microtube, having inner and outer diameters of 150 m and 360 m respectively, is connected to the syringe pump. Droplet dispatch and impinge onto hot-surface surface from a decided height when it reaches to maximum size which can be supported by the surface tension forces against gravity. Utilizing surface tension and gravity force balance, balance nearly uniform dispatching droplet diameter of around 2.5 mm is produced at the tip of the tube. Temperature of the hot plate is controlled according to thermocouple data which is recorded 2 mm below the liquid metal surface. surface U-alloy 70 (an alloy of Bi, Pb, Cd, and Sn, having g melting point of 70 o C, density of 9946.27 kg/m3) is used as a hot surface and assumed to be non-deformable deformable due to the droplet impact since its density and surface tension are much higher than those of water. Liquid metal surface temperature agitated to lover temperatures due to the droplet impingement; however, it recovers to 95% of settled initial temperature in 35s. As a solid hot-surface, surface, polished cupper plate (30x30 mm, 0.5 mm thick) is employed. Time resolved images are recorded with a high-speed camera mera operating at spatial and temporal resolutions of 256x256 pixels and 10 kHz respectively. 200 mm Nikon Micro Nikkor lens is adapted to the camera to have larger spatial resolutions.

tremendously varying characteristics characteristic associated with the type of material and surface chemistry. Figure 2 and Figure 3 show temporal evolution of water droplet after initial contact with the cupper plate and liquid metal respectively with inverted (a) and filtered images (b) under ambient conditions with similar Weber numbers. To amplify the surface roughness effects, relatively relativel low impact velocities are chosen in order to prevent the splashing effects, resulted from energetic impacts, from suppressing the surface roughness contribution. When the droplet impinge onto the cupper plate having a temperature of 313 oC which is well above 225 oC estimated Leidenfrost point (LFP) by Bernardin et al (1997) and regarded as lower temperature boundary of film boiling regime, wavy structures on droplet surface can be discerned from varying non-uniformity uniformity of image intensities (Figure 2a). Disturbances observed after 1.3 ms on droplet surface are characterized with unordinary random motions which are made more explicit with Sobel filtering in Figure 2b. Two modes of surface waves are distinguished during droplet collision onto the hot surface above LFP. First mode is created due to the droplet impact kinetic energy which is partially transferred to the droplet and generates uniform energetic waves encircling the droplet and propagating from the bottom of the droplet to the upward direction. Those hydrodynamic waves are visible to some extent from 1 ms and 1.3 ms after droplets contact with the surface (Figure 2a). . Parallel horizontal lines and smooth pulses on the edge of the droplet (Figure 2b) ) until the 2.3 ms also confirm those structures from another perspective. perspective The he second mode is characterized as nonuniform random fluctuations which begin to appear a on the droplet surface after 2.3 ms that indicates nonuniform non pressure transients taking place beneath the droplet due to the instantaneous vapor generation from the nucleation spots. spots Vapor layer is having intermittent pressure fluctuations which disturbs lower surface of droplet, promote unstable waves propagating whole across the liquid surface. Pressure variations are likely to be caused by non-uniform vapor generation rate attributed to the non-uniformity non of surface structures (surface roughness), since indented uneven structures lead to non-uniform uniform heat transfer to the liquid surface with intermittent direct contact conduction by rupturing the vapor film. Elevated heat transfer rates enhance vapor por production rate resulting in droplet bounce back from the surface with keeping its integrity (7.6 ms later in Figure 2 ).

Figure 1: Schematic view of experimental setup.

Results and Discussions Effect of surface roughness One of the important parameter in boiling phenomena whose effects still cannot be resolved sufficiently due to its

th

International Conference on Multiphase Flow ICMF 2013, Jeju, Jeju Korea, May 26 - 31, 2013

droplet without creating any non-uniform non strain on the water interface (from 0 ms to 2.9 ms in Figure 3). Upon vapor layer reaching sufficient pressure for breaking out its entrapment, , but not sufficient for lifting the t droplet yet, it erupt from the periphery instead of lifting the droplet off (at 2.9 and 3.1 ms in Figure 3).

(a)

(a)

(b) Figure 2: Droplet impingement on cupper plate a) original inverted images b) filtered with Sobel, (313 oC, We=5.2, u=0.39 m/s). On the contrary, in the case of zero surface roughness (liquid metal surface) first mode of waves continued to be monitored until droplet reaches its largest spreading diameter up to 2.9 ms, any notion of uneven disturbances on the droplet surface cannot be discerned (Figure Figure 3a-b). Soon after at 3.1 ms a liquid spurt urt come into sight from the water and liquid metal interface suggesting that droplet spread on the surface as much as surface tension forces allowed and vapor layer reached ched the maximum maximu necessary pressure for disrupting the liquid layer on it and erupted. erupted This blast of vapor wieldy destroys and fragments water in the proximity of interface resulting in flying out of small secondary droplets. Unlike to cupper surface, no major m non-structured disturbances developed on the droplet surface until a sudden blast taking place through the interface up to 3.1 ms. Based on this observation, we may speculate that a stable vapor layer is formed under the droplet in contrast to cupper cuppe surface case in which localized pressure disturbances associated with spotted vapor production are discarded by localized strains of water surface without any lumped liquid fragmentation. However, absence of surface roughness on the liquid metal surface leads to a uniform vapor generation that degraded the heat transfer rate from surface to the droplet equally for every spatial position beneath the water

(b) Figure 3: Droplet impingement on liquid metal a) original inverted images b) filtered with Sobel, (307oC, We=4.7, u=0.37 m/s). Contact angle and boiling incipience dependence on surface roughness The contact angle between a liquid and a solid surface is regarded as one of the important parameters parameter governing boiling phenomena since it characterizes wettability of a certain solid surface by a specific fluid (Tong et al 1990). During droplet impingement onto hot liquid metal surface, we observed erved a sudden stepwise decline in contact angle whereas on cupper er plate such a rigorous variation could not be detected. As mentioned and discussed partially in previous section, droplet on the hot liquid metal surface sits on a vapor/air cushion for a while then suddenly wetted with the liquid metal surface which initiate initia the boiling. g. A description of that rapid change in contact angle is demonstrated in Error! Reference source not found.,

th

International Conference on Multiphase Flow ICMF 2013, Jeju, Jeju Korea, May 26 - 31, 2013

from those time sequential images transient variation of contact angle is detected and

captured at the instants of 2.7 and 2.9 ms in Figure 3, it is clearly visible that sudden change in contact angle is followed by incipience of boiling. Based on this finding a parameter regarded as delay time and identified as the time elapsed until a drastic change of contact angle takes place is employed to investigate the droplet impingement boiling on a surface. Effect of droplet velocity on boiling incipience

Figure 4: Stepwise change of contact angle.

Temperature 307 202 106 108 (Copperplate)

150 [degree] 100 50 0

3 Time[ms]

Figure 5 Variation of contact angle after the collision of droplets to the liquid metal surface with 0.37 m/s velocity. velocity

150 [degree] 100 50 0

Temperature 309 202 105 110 (Copperplate)

Zero surface roughness changes the droplet drople boiling characteristics due to the effect of wettability of droplet with the liquid metal. Delay time for the surface wetting and incipience of boiling is explored for different impact velocities and illustrated in Error! Reference source not found.. For all velocities delay time remains nearly unchanged up to the 80K superheat (corresponding ( to a surface temperature of around 180 oC), later on a peak having different magnitudes are observed. General tendency of delay time variation, especially for the case of 0.37 m/s impact velocity, with respect to the wall super heat resembles the drop evaporation curve introduced by Bernardin and Mudawar (1999). In their study, sessile drop evaporation time me decreases in nucleate boiling regime until the CHF temperature corresponding to minimum drop life time, thereafter droplet life time start to increase in transition to film boiling regime which continues up to the LFP point temperature. From that perspective, perspe the peak in delay times for the case of 0.37 m/s corresponds to an implicit evidence of boiling mode transition point which can be speculated to be critical temperature for the transition to film boiling regime. However, onset of transition to film boiling regime occurs in stepwise manner that differs from the case of sessile droplet. Transition boiling continues up the surface temperature of 238 oC, after that temperature droplet and surface contact could not be observed that t implies the LFP point is i already exceeded.

3 Time[mc]

Figure 6 Variation of contact angle after the collision of droplets to the liquid metal surface with 0.41 m/s velocity. velocity plotted in Error! Reference source not found. and Error! Reference source not found. for the impact velocities of 0.37 m/s and 0.41 m/s respectively. At a first glance on these graphs, droplet is wetted with the liquid metal surface earlier when it impinge on with a higher velocity ve (Error! Reference source not found.) compared to the lower velocity case (Error! Reference source not found.). On the other hand, impingement velocity does not affect gradually the time evolution of contact angle with cupper surface which has 108 oC surface temperature. It increases in first 1 ms then reaches a maximum value and retain this value val for a relatively longer time for both impingement speeds that is likely to be due to the surface roughness of cupper plate. When the shape of two sequential images of one droplet

Figure 7: Delay time or boiling incipience time variation due to the surface temperature for several droplet velocities. Similarly, smaller maller peaks detected at 180 oC surface temperature for the higher impact velocities, however, they may not be attributed to a transition point due to the fact that delay times continues to decrease after that point

th

International Conference on Multiphase Flow ICMF 2013, Jeju, Jeju Korea, May 26 - 31, 2013

as well. That decrement in delay times turn to increment with a steep climb at around 280 oC (at around super heat he of 180K) and any wettability event cannot be observed after around 200K wall superheat. 180K superheat value is likely to be a minimal point for the transition regime to film boiling. Bernardin et al (1997) reported critical heat flux (CHF) temperature on a polished cupper surface for the impinging droplets having a 3 mm diameter and 0.7 m/s velocity as 130 oC. They hey concluded that CHF and LFP temperatures, temperatures 130 oC and 225 oC respectively, seems fairly insensitive to droplet impact velocity over tested range. On the other hand, they demonstrated that while CHF temperature is insensitive to surface roughness, LFP temperature shows higher sensitivity against surface roughness. Considering the e boiling behavior of droplets impinging on solid metallic surfaces, which has some particular level of surface roughness even though surface treatment applied, it would not be inappropriate to come into such a conclusion that under the zero surface roughness ess LFP and CHF temperatures are likely to be varying rying drastically according to the surface superheat and droplet impingement velocity as it is evidenced from the surface wettability transients. transients

beginning earlier than those of atmospheric pressure case ca and rises up to elevated levels at a superheat super temperature of 80K. . If we took into account solely the differences in latent heats of both cases and the time to reach a stable film boiling, the peak in boiling delay in sub-atmospheric sub pressure must have been developed at higher superheats and it must have not been much diverged from the atmospheric pressure. From this point of view, another factor must come into play and govern the film boiling regime by generating more stable blockage against heat transfer er and prevents droplets wetting during boiling in sub-atmospheric pressures.

Effect of pressure The noncondensible gas and ambient pressure press is another important parameter governing verning the boiling phenomena. Emmerson (1975) investigated sessile droplet boiling and found that droplet evaporation time is reduced as the pressure increases. He explained lained that tendency associating with the decrease ease in latent heat of vaporization due to increasing pressure. He also noted that whereas LFP increases with pressure, droplet evaporation time or droplet lifetime remain shorter. Buchmuller et al (2012) studied droplet impingement boiling under high pressures, up to t 16 bars, with water droplets having 2.4 mm diameter and 0.77 m/s impact velocity. They observed shifts in boiling transition points including LFP due to the increase in saturation temperature. Error! Reference source not found. presents pre the delay time variation versus wall superheat heat under atmospheric pressure and sub-atmospheric pressures for the droplets hitting to surface at the speeds of 0.41 m/s and 0.43 m/s. m/s Red line denotes the surface temperatures and superheat values at which peaks are detected for both cases. cases While no significant delay in wettability due to the boiling is observed for the case of sub-atmospheric atmospheric pressure in between 0-60 oK superheat regions, a particular particu level of delay is detected under atmospheric pressure in the same range of superheat. If we extrapolate the Emmerson (1975)s point of view (valid for the pressures higher than the atmospheric one) to the sub-atmospheric pressure; since ince the latent heat of saturated water at 50 kPa (2305 kJ/kg) higher than that of atmospheric pressure (2257 kJ/kg), apparently more heat energy is needed for the creation of vapor layer in sub-atmospheric atmospheric pressure, hence, the reason of no delay in boiling around low super-heats (below 25K) is likely to be the lack of sufficiently thick vapor layer formation. On the other hand, , we observed an intriguing sharp increase in delay time under the sub-atmospheric atmospheric pressure

Figure 8: Delay time versus super-heat sup at atmospheric (0.41 m/s) and sub-atmospheric atmospheric (0.43 m/s) m/s pressures. Air bubble entrapment beneath the impinging droplets on solid surfaces in atmospheric pressures is known from some researches in the literature. Thoroddsen et al (2005) investigated the air bubble entrapment under the impacting droplets onto cold solid surfaces. faces. They explored dynamics of air bubble for several time periods elapsed after the first contact of the droplet. They observed air bubble for various impact velocities of the water droplets having a diameter of 4-5.5 mm and showed that air entrapment occurs ccurs in wide range of Weber numbers. Recently reported by Driscoll et al (2012) that at sub-atmospheric sub pressures, a ring of highly populated microbubbles beneath a 3.1 mm diameter silicon oil droplet impacting with 3.19 m/s velocity, in contrast to the atmospheric tmospheric pressure conditions, encircling a larger bubble are entrapped and their sizes grow with decreasing pressure. Under the light of aforementioned previous findings it can be hypothesized that larger delays under sub-atmoshperic atmoshperic pressure is likely to be caused by entrapped air bubbles and accompanying microbubbles which are blocking the droplet wettability and blanketing conductive heat transfer to the droplet leading to larger delays. Even this is the case, further investigation and quantification is essential for the confirmation of this effect.

Conclusions

th

International Conference on Multiphase Flow ICMF 2013, Jeju, Korea, May 26 - 31, 2013

Effective parameters including surface roughness and ambient pressure in droplet impingement boiling phenomena are investigated utilizing polished cupper plate and liquid metal. Boiling behaviour of water droplets demonstrated completely different manner on the liquid metal surface, it is characterized by sudden vapour burst and droplet fragmentation instead of bounce back from the surface. Wetting behavior of droplet shows different tendency with respect to surface roughness. When the droplet hits the liquid metal surface, it remains nonwetted for a while then a rapid wetting is observed, whereas on the cupper surface contact angle increases monotonically and remained constant. With zero surface roughness boiling transition temperatures are likely to be varying drastically according to the surface superheat and droplet impingement velocity as it is evidenced from the surface wettability transients. Under sub-atmospheric pressure boiling behavior of the droplet indicates harsh difference such that delay time increases while the pressure decreases. The reason of this tendency is likely to be associated with the differences in air entrapment dynamics in subatmospheric pressure. References Bernardin J.D. Stebbins C.J. and Mudawar I. Effects of surface roughness on water droplet impact history and heat transfer regimes, Int. J. Heat Mass Transfer, 40-1: 73-88 (1997). Bernardin J.D. Mudawar I. The Leidenfrost point: Experimental study and assessment of existing models. Journal of Heat Transfer, 121:894-903 (1999). Buchmuller I. Roisman I, Tropea C. Influence of elevated pressure on impingment of a droplet upon a hot surface. ICLASS 2012, Heidelberg, Germany (2012). Celata G.P, Cumo M, Mariani A., Saraceno L. A comparison between spray cooling and film boiling during the rewetting of a hot surface. Heat Mass Transfer, 45:1029-1035 (2009). Driscoll M.M. Stevens C.S. and Nagel S.R. Thin film formation during splashing of viscous liquids, Ph. Rev. E, 82:036302 (2010). Emmerson G.S: The effects of pressure and surface material on the Leidenfrost point of discrete dropes of water, Int. J. Heat Mass Transfer, 18:381-386 (1975).

Tong W. Bar-cohen A. Simon T.W and You S.M. Contact angle effects on boiling incipience of highly wetting liquids. Int. J. Heat Mass Transfer, 33-1:91-103 (1990). Thoroddsen S.T. Etoh T.G. Takehara K. Ootsuka N. And Hatsuki Y. The air bubble entrapment under a drop impacting on a solid surface. J. Fluid Mech. 545:203-2012 (2005).

Kim J. Spray cooling heat transfer: The state of the art, Int. J. of Heat and Fluid Flow, 28:753-767 (2007). Mohapatra S.S. Chakraborty S. and Pal K. Experimental studies on different cooling process to achieve ultra-fast cooling rate for hot steel plate. Exp. Heat Transfer, 25:111-126 (2012). Moreira A.L.N. Moita A.S. Panao M.R. Advance and challenges in explaining fuel spray impingement: How much of single droplet impact research is useful?, Prog. in Energy and Combustion Sci., 36:554-580 (2010).