- Email: [email protected]
Modification and enhancement of cryogenic quenching heat transfer by a nanoporous surface
International Journal of Heat and Mass Transfer 80 (2015) 636–643
Contents lists available at ScienceDirect
International Journal of Heat and Mass Transfer journal homepage: www.elsevier.com/locate/ijhmt
Modification and enhancement of cryogenic quenching heat transfer by a nanoporous surface Hong Hu a, Cheng Xu b, Yang Zhao b, Reid Shaeffer a, Kirk J. Ziegler b, J.N. Chung a,⇑ a b
Cryogenics Heat Transfer Laboratory, Department of Mechanical and Aerospace Engineering, University of Florida, Gainesville, FL 32611-6300, USA Nanostructured Interfaces Laboratory, Department of Chemical Engineering, University of Florida, Gainesville, FL 32611-6005, USA
a r t i c l e
i n f o
Article history: Received 27 June 2014 Received in revised form 21 September 2014 Accepted 24 September 2014
Keywords: Cryogenics Chilldown Nanoporous surfaces Leidenfrost point Critical heat flux Quenching curve
a b s t r a c t Since the average chilldown efficiency for cryogenic systems is only about 8%, significant improvements to heat transfer are needed for many applications. An experiment was performed to evaluate the modification and enhancement on the quenching heat transfer by a nanoporous heat transfer surface in this study. For comparison purposes, two sample surfaces were used. One is the mechanically polished conventional normal aluminum surface serving as the base case and the other is an aluminum surface with the anodized aluminum oxide (AAO) nanoporous finish. In this work, the effect of the nanoporous surface on the heat transfer during chilldown in a liquid nitrogen pool is investigated. The results indicated that the nanoporous surface completely modified and enhanced the phase-change heat transfer in all three quenching regimes. Comparing to the conventional surface case, the Leidenfrost temperature was increased by 32 K and the critical heat flux (CHF) was raised by 160% due to the nanoporous surface. However, the most significant modification on the boiling mechanisms by the nanoporous surface was found in the transition regime that is composed of transitional film and transitional nucleate sub-regimes with quite different quenching curve slopes. For cryogenic quenching applications, it is estimated that the nanoporous surface could save 20% in the amount of cryogen consumption by shortening the chilldown time. The modification and enhancement are mainly attributed to the superhydrophilic property and nanoscale nucleation sites offered by the nanoporous surface. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Cryogenic fluids are widely used in industrial applications, space exploration, and cryosurgery systems. Specifically, in space exploration, cryogenic fluids are used in power and propulsion, thermal management, and life-support systems of a spacecraft during space missions that involves transport, handling, and storage of these fluids under both terrestrial and microgravity conditions. For example, cryogens are usually used as liquid fuels, such as liquid hydrogen and oxygen, that are burned in liquid-fueled rocket engines. When a cryogenic fluid transport system is first started up, its walls and hardware must go through a transient chilldown period prior to reaching steady-state operation. Therefore, chilldown is the process of adjusting the system to the low temperature regime, which is usually several hundred degrees below room temperature. The chilldown or quenching process is a complicated phasechange phenomenon involving unsteady two-phase flows as well as heat and mass transfer; its characteristics have not been fully understood, resulting in very low chilldown efficiencies. ⇑ Corresponding author. Tel.: +1 352 392 9607; fax: +1 352 392 1071. E-mail address: jnchung@ufl.edu (J.N. Chung). http://dx.doi.org/10.1016/j.ijheatmasstransfer.2014.09.056 0017-9310/Ó 2014 Elsevier Ltd. All rights reserved.
Fig. 1 is a typical boiling curve where the heater surface heat flux, q00 , is plotted against the heater surface degree of superheating, TW Tsat, where TW is the surface temperature and Tsat is the saturation temperature corresponding to the system pressure. Boiling is characterized by the surface temperature and heat flux on the wall according to the boiling curve. Since the wall temperature is strictly determined by an energy balance of the wall, the heat fluxes in and out of the tube wall are important parameters. In boiling, the heat source is externally supplied and, thus, can be controlled independently, such as the constant wall heat flux condition. In this case, boiling is a heat flux (independent variable) controlled process. For the heat flux controlled condition, boiling follows the route of A ? B ? D. In order to avoid a huge temperature jump, boiling usually runs safely below B in the nucleate boiling regime. In contrast, the wall does not have an external source of heat (except a small amount of residual heating) during chilldown; therefore, heat coming out of the wall can only be supplied internally from the thermal capacity (stored energy) of the wall. The only way to get heat to be removed from the system is by lowering the inner wall surface temperature by cooling from the cryogenic flow. Accordingly, the wall inner surface temperature is the independent variable and the control parameter. In summary, quenching
H. Hu et al. / International Journal of Heat and Mass Transfer 80 (2015) 636–643
is a conjugate process involving both heat transfer and wall–fluid interaction. The quantity of heat (heat flux) that is released to the fluid can only be associated with the temperature change of the wall inner surface. As a result, the wall surface temperature is the controlling parameter that forces the quenching process to follow the route D ? C ? B ? A. In boiling applications, such as those in the cooling of a nuclear reactor, a super high flux of heat transfer in the nucleate boiling regime is standard practice. The only precaution is not to exceed the critical heat flux (CHF) and enter the film boiling regime. However, film boiling is the first mode of boiling encountered in chilldown and it cannot be avoided. Due to its low heat fluxes at high wall temperatures, the chilldown efficiency in general is extremely low. According to Shaeffer et al. [1], the average chilldown efficiency that is defined as the ratio of the amount of thermal energy removed from the wall versus the maximum cooling capability of the cryogen spent in a phase change process is about 8%, highlighting the tremendous need to improve the quenching efficiency for many applications that require a cryogen as the working fluid. Advances in nanotechnology enable modification and enhancement of heat transfer based on the phase change from liquid to vapor. This paper reports significant modification and enhancement to heat transfer throughout the entire boiling regime with the purpose of substantially improving the quenching thermal efficiency.
2. Literature review 2.1. Cryogenic chilldown experiments Because of the difficulties associated with the design and measurements of cryogenic experiments, there is less chilldown data for cryogenic fluids compared to traditional fluids. Researchers began to investigate heat and mass transfer during cryogenic liquid
637
transfer line chilldown because that this area is of extreme importance to spacecraft design. Chilldown of these transfer lines were first studied by Brennan et al. [2] under normal gravity conditions for liquid nitrogen (LN2) and liquid hydrogen (LH2). Antar and Collins found a unique flow pattern in microgravity conditions, which is known as the filament flow. Kawanami et al. [3] also observed a similar flow pattern in terrestrial conditions. They claimed that the gravity effects on heat transfer of forced convective boiling decreased with an increase in the mass velocity. Verthier et al. [4] have studied the effect of gravity during quenching using FC72. LN2 was used as the working fluid [4] to investigate the heat transfer characteristics and flow pattern during the quenching of a vertical tube under both terrestrial and ten-second microgravity conditions. It is claimed that the heat transfer and quenching front velocity under microgravity conditions increased up to 20% compared to those in traditional gravity conditions. 2.2. Heat transfer enhancement by nanoscale surface structures The majority, if not all, of the boiling heat transfer enhancement endeavors focused on the nucleate boiling regime and the CHF. There have been a number of experiments to measure the contact angle and wettability on the nanoporous surface. Luo’s group [5–7] have introduced more detail discussion both theoretically and experimentally on the wettability on nano pillar surface and derived an angle criterion determining transition from Cassie-Baxter to Wenzel states. Singh et al. [8] investigated the wetting and evaporation of sessile droplets on nanoporous anodic aluminum oxide (AAO) substrates having different pore distribution (uniform, random and linearly arranged) morphologies and pore sizes (70– 120 nm). They claimed that nano structured surface is an applicable tool to control wettability as well as the diffusive evaporation process. In addition, physical morphology and pore distribution affect wettability as well as evaporation rates. Tasaltin et al. [9]
Fig. 1. A typical boiling curve including different boiling regimes and corresponding flow patterns.
638
H. Hu et al. / International Journal of Heat and Mass Transfer 80 (2015) 636–643
made nanoporous AAO surfaces with a hexagonal structure of circular pores with 250 nm in diameter and 300 nm in depth and found the surfaces to be superhydrophobic. The contact angle was measured at 153.2 ± 2° for vapor-phase hexamethyldisilazane (HMDS), that is larger than the average water contact angle of 82.9 ± 3° on smooth thin film AAO surfaces. The excellent surface wettability associated with these nanoporous surfaces has significant potential to enhance heat transfer. Liter and Kaviany [10] reported a nearly three times improvement to heat transfer for a porous-layer coating over a conventional surface. Arik et al. [11] reported an experimental study finding that microporous, diamond-coated surfaces provided an average 1.6 enhancement factor on FC-72 CHF. To our knowledge, there is no literature discussing cryogenic quenching with nanoporous surfaces. Li et al. [12] reported an experimental study that investigated the CHF of a high-velocity circular water jet impingement boiling on the stagnation zone of a nano-structured heater surface with different surface wettability parameters. The wettability of the copper heater was modified with nanostructures to change the surface topography and chemistry to exhibit either hydrophilic or hydrophobic character. They found that the wettability holds a strong influence on the CHF. For example, the CHF almost doubled when the contact angle was reduced from 105° to 5°. Hsu and Chen [13] experimentally investigated the effects of surface wettability on water pool boiling heat transfer. They used coatings of nano-silica particles to change the wettability of the copper surface from superhydrophobic to superhydrophilic. They demonstrated that CHF values increased from 300 to 1483 kW/m2 when the contact angle decreased from 145° (hydrophobic) to less than 10° (hydrophilic). O’Hanley et al. [14] experimentally evaluated the separate effects of surface wettability, porosity, and roughness on the CHF of water using engineered surfaces. They reported that porous hydrophilic surfaces enhanced CHF by 50–60%, while porous hydrophobic surfaces resulted in a reduction of CHF by 97%. Phan et al. [15] also focused on the influence of surface wettability on nucleate boiling heat transfer. They used nanocoating techniques to vary the water contact angle from 20° to 110°. They reported that a greater surface wettability due to the hydrophilic nature of
the surface increases the vapor bubble departure radius and reduces the bubble emission frequency. The best heat transfer coefficients were obtained for the surfaces that had a water contact angle close to either 0° or 90°. Hens et al. [16] used molecular dynamics simulations to understand the boiling mechanisms of argon bubble formation on a platinum substrate with a particular emphasis on the effects of nanosized texture on the surface. The bubble nucleation and vapor film formation showed a dependence on the degree of superheat and solid–liquid interfacial wettability since bubbles do not form easily on non-wetting surfaces. They concluded that a wettable (hydrophilic) surface facilitates higher heat transfer. The only study found that addressed the effects of nano-textured surfaces on the Leidenfrost temperature is given by Kruse et al. [17]. The micro- and nano-structures were fabricated on a stainless steel surface by a femtosecond laser. The traditional Leidenfrost droplet experiment was performed on both machine-polished and nanotextured surfaces. They observed an extraordinary shift in the Leidenfrost temperature from 280 °C for the machine-polished surface to 450 °C for the nano-textured superhydrophilic surface. The gigantic 175 °C increase in the Leidenfrost temperature was attributed to a reduction in the contact angle that results in intermittent liquid/solid contact and capillary wicking action.
3. Experiment 3.1. Experimental apparatus and procedure A simple system was built that can explore all the fundamental physics and mechanisms of phase-change heat transfer while eliminating all unnecessary complicating factors. Instead of an internal convective flow chilldown system, a pool quenching experiment was used since the basic characteristics of the quenching curve (reversed boiling curve) are not fundamentally affected by the bulk flows. The basic schematic of the design is illustrated in Fig. 2. A transparent glass dewar with a vacuum insulation from the surroundings was used to form a liquid nitrogen pool that was
Fig. 2. Schematic of experimental apparatus (A) overview of test apparatus made of 8020 framework. (B) Test rod design.
H. Hu et al. / International Journal of Heat and Mass Transfer 80 (2015) 636–643
maintained at atmospheric pressure by an opening to atmosphere. The test section was an aluminum rod (99.99% purity) with a thermocouple bead embedded along the centerline of the rod and 0.4 inch from the bottom surface. A fast response and high sensitivity T-type 40AWG (0.0799 mm diameter) micro thermocouple was used to record the temperature history at 14 Hz data acquisition frequency. A set of background lights and a high speed camera were setup to collect the video data on the two-phase bubble dynamics. The video has been utilized to analyze the behavior of the interaction between vapor and liquid. The recording speed was set at 2000 frames per second. To study the effect of nanoporous surfaces on cryogenic quenching, chilldown experiments were conducted with two different surfaces. One is a commercially available machine polished conventional aluminum surface, which is referred as the normal surface, and an aluminum surface coated with a thin anodized aluminum oxide (AAO) nanoporous finish. The experiment started when the test rod was submerged in the dewar containing LN2. In order to demonstrate the repeatability and reduce aleatory uncertainty, 15 identical tests were conducted on each nanoporous surface and normal surface.
3.2. Preparation of anodized aluminum oxide (AAO) nanoporous surface The AAO nanoporous surface was fabricated by a two-step anodization process. A pure aluminum rod was cleaned by sonication in soapy de-ionized water, acetone, and ethanol, respectively. The aluminum rod was then put into an electropolish solution (Electro Polish System Inc.) at 65 °C and a constant voltage of 17 V for 20 min. After that, the first anodization was carried out at 15 °C and a constant voltage of 40 V in 0.3 M oxalic acid solution with stirring for 16 h. After stripping off the AAO layer in a mixture of 10 wt% phosphoric acid and 1.8 wt% chromic acid, a second anodization was performed at the same conditions. The time was varied to control the final AAO thickness to approximately 5 lm. At the end of the second anodization, the voltage was reduced at a rate of 0.5 V per 30 s to thin the barrier layer. The pore opening process was in 5 wt% phosphoric acid at room temperature and was monitored by an electrochemical setup (VersaStat 3, Princeton Applied Research) using a small voltage of 0.1 V applied against a carbon counter electrode. The current was controlled carefully so that the pore opening process stopped as soon as the current increased remarkably, which indicates that the pores were fully opened. The finished surface results in a pattern of perfectly arranged ordered nanopores formed on the aluminum rod surface, as shown in Fig. 3. The nanoporous surface has a hexagonal structure with pores approximately 50 nm in diameter and depths of 5 lm. The contact angles of the surfaces were measured with a 1 lL drop of deionized water at room temperature using a Kruss DSA100 Drop Shape Analyzer. Contact angles were measured for
639
3 drops at different locations on the surface. Fig. 4 shows the difference in contract angles for each surface. The average contact angle on the nanoporous and normal surface is 13.1 ± 1.7° and 74.6 ± 1.6°, respectively. Therefore, the nanoporous surface demonstrates superhydrophilic properties. However as discussed above, Tasaltin et al. [9] made nanoporous AAO surfaces with a hexagonal structure of circular pores with 250 nm in diameter and 300 nm in depth and found the surfaces to be superhydrophobic. Therefore the wettability depends strongly on the dimension and depth of the nano pores. The contact angles for liquid nitrogen on most conventional normal metal surfaces are relatively small (less than 7°) [18]. Also there is no report in the literature that addresses the effects of nanoporous surface on the contact angle of liquid nitrogen. Due to the low temperate nature and a lack of experimental facility, we were not able to perform the measurement. As the contact angle theory [19] has pointed out that an increasing surface free energy leads to a decreasing contact angle, nanoporous surfaces have been shown to possess larger surface energy values than normal surfaces based on the experimental results. Therefore, we expect that this is the case for our nanoporous surface. As the contact angle of liquid nitrogen on a normal metal surface is already very small, a contact angel drop of just a few degrees would mean a large percentage decrease that is certainly reflected in our experimental results. 3.3. Uncertainty analysis The dimensions of the test rod were determined by a caliper. The mass of the test rod was measured through an electronic scale. The thermocouples had the uncertainty of ±1.5% full scale according to the manufacturer. Another uncertainty source for temperature measurements comes from the DAQ whose uncertainty was ±0.1% full scale. In the current experiment, the inverse heat conduction methods [20] were used to obtain the surface heat fluxes from the measured temperatures. The uncertainty of the heat flux is calculated through a square root summation of partial derivatives. Both the uncertainties from the experimental apparatus and calculated parameters are listed in Table 1. 4. Results and discussion 4.1. Data reduction and analysis For the study of the liquid-to-vapor phase-change heat transfer, the so-called ‘‘boiling curve’’ (see Fig. 1) is often of great interest because it denotes the regimes of the different heat transfer mechanisms and is crucial for engineering applications. However, as mentioned above, the surface temperature at the bottom of the rod (point A in Fig. 2) is the independent variable during the chilldown or quenching process; therefore, the heat flux at point
Fig. 3. (A) Top- and (B) cross-view SEM images of the nanoporous surface.
640
H. Hu et al. / International Journal of Heat and Mass Transfer 80 (2015) 636–643
Fig. 4. Contact angle on (A) nanoporous surface and (B) normal surface.
Table 1 Summary of the uncertainty. Source of uncertainty
Symbol
Uncertainty
Rod inside diameter Rod mass Outside temperature Inner temperature Heat flux
di (mm) Mr (g) To (K) Ti (K) qi (K)
±0.001 mm ±0.0001 g ±1.6% ±1.6% ±6.8%
A cannot be controlled independently. Additionally, the direct measurement of temperature for point A on the test rod is difficult. Therefore, the surface heat flux can only be inferred from the temperature history recorded by the embedded thermocouple. The heat flux and the surface temperature at point A were obtained by solving the inverse heat conduction problem (IHCP). The test section is approximated as a one-dimensional circular slab. The effect of radiation on the outer surface is neglected. An analytical method was derived [20] to obtain the surface temperature and heat flux at point A of the exposed surface based on the temperature of a thermocouple positioned inside the rod of the test section by the following equation,
To ¼ Ti þ
2 2 3 3 1 x2 dT i 1 x2 d T i 1 x2 d T i þ þ þ 2 3 2! ar dt 4! ar 6! ar dt dt
ð1Þ
where the subscript o denotes rod bottom surface (point A in Fig. 2) while i indicates the thermocouple location. The distance x is between the thermocouple and point A, t is the measurement time, and ar is the thermal conductivity of the rod. The heat flux at point A can be approximated by the first three leading terms of an infinite series given below,
q00o
" 2 3 # 2 dT i 1 x2 d T i 1 x2 d T i ¼ qr C p;r x þ þ 3 dt 3! ar dt2 5! ar dt
ð2Þ
where Cp,r is the specific heat of the rod and qr is the density of the aluminum rod. 4.2. Chilldown heat transfer characteristics The quenching (chilldown) process takes place when the rod is chilled in the pool from room temperature (300 K) down to the liquid nitrogen pool temperature at 77.35 K, which is the saturated temperature under the atmospheric pressure. The time dependent surface temperatures and heat fluxes at point A (see Fig. 2) during chilldown were estimated using the thermocouple measured temperatures and Eqs. (1) and (2), as outlined above. With the temperature and heat flux histories obtained, the chilldown heat transfer characteristics can be generated accordingly. For the purposes of illustrating the differences of chilldown heat transfer between a mechanically finished conventional surface and a nanoporous surface, respective quenching and chilldown curves are given in
Fig. 5(A) and (B). It is noted that the two plots share the same vertical axis that indicates the temperature of the rod bottom surface at point A in Fig. 2. As seen in Fig. 5(B), the total chilldown time is shorter for the nanoporous surface. Specifically, the time for the surface to reach 80 K is 28 s and 35 s for the nanoporous and normal surface, respectively, corresponding to a 20% improvement in time. This substantial savings in chilldown time was mainly achieved by the nanoporous surface in the film boiling regime. The normal surface spent 32 s in the film boiling regime while it only took 24 s for the nanoporous surface. As shown in Fig. 5(A), the nanoporous surface facilitated a higher Leidenfrost temperature (158 K) that resulted in a quicker switch from the film boiling regime to the transition boiling regime as compared to the normal surface whose Leidenfrost temperature was found at 126 K that took 8 s longer to reach. Other than the Leidenfrost point, the nanoporous surface produced a substantially modified transition boiling process as compared to the normal surface that presents an ordinary transition boiling pattern. The transition boiling on the nanoporous surface is considered to be composed of two different heat transfer mechanisms that resulted in two apparently different slopes. Mainly due to the drastically modified transition regime, both the heat flux at the CHF point and those in the nucleate boiling regime were raised substantially. 4.3. Flow visualization of two-phase quenching process Fig. 6 provides images of the solid–fluid interactions on the normal and nanoporous rod surfaces during the chilldown process in the specific boiling regimes. It is seen that the liquid–vapor hydrodynamics and the resulting flow patterns along the surfaces evolve during chilldown. We found that distinct differences were observed between the two surfaces. For the film boiling regime shown in Fig. 6(A) and (D), it can be seen that the vapor film on the normal surface is much thicker and contains more dynamic fluctuations and liquid–vapor interactions than that on the nanoporous surface. The thinner films explain the higher heat fluxes associated with the nanoporous surface. The nanoporous surface also produced smoother liquid–vapor interfaces, which can be explained by a ‘micro-damper’ theory. Millions of nanoscale porous cavities work as millions of micro-dampers that facilitate the attenuation of the high frequency perturbations. As a result, perturbations cannot be amplified on the nanoporous surface, while the surface roughness would magnify and amplify the perturbations on the normal surface to form turbulent flows and interactions. In the transition boiling regime as shown in Fig. 6(B) and (E), the normal surface still produced thicker and more chaotic two-phase patterns, similar to those observed in the film boiling regime. However, the liquid–vapor layers are thicker than those in the film boiling, which is due to the fact that more vapor is generated as the
H. Hu et al. / International Journal of Heat and Mass Transfer 80 (2015) 636–643
641
Fig. 5. (A) Quenching and (B) chilldown curves for conventional and nanoporous surfaces. The different regimes during chilldown are also shown.
Fig. 6. Comparison of the liquid–vapor flow patterns formed on conventional (top panel) and nanoporous surfaces (bottom panel) during liquid nitrogen chilldown. (A, D) Film boiling region; (B, E) transition boiling region; (C, F) nucleate boiling region.
liquid starts to contact the surface, resulting in high rates of nucleation and bubble generation. In the nucleate boiling regime, the two-phase flow pattern on the nanoporous surface shown in Fig 6(F) indicates that a large
amount of small and separated bubbles are generated in an explosive manner. However, in Fig. 6(C) the normal surface results in a different pattern where bubbles tend to merge together to form larger vapor slugs that point outwards from the surface. This
642
H. Hu et al. / International Journal of Heat and Mass Transfer 80 (2015) 636–643
Fig. 7. Comparison of the heat flux for nanoporous and conventional surfaces during chilldown.
phenomenon results in longer attachment times for the bubbles on the surface, which is the main reason why the normal surface has lower heat fluxes in the nucleate boiling regime. The proper sized and patterned nano pores on the nanoporous surface alters the wettability, allowing more nucleation sites to be created that enhance nucleation and bubble escape from the surface. 4.4. Enhancement and modification by the nanoporous surface The two quenching curves (reversed boiling curves) obtained from the normal and nanoporous surface are plotted together in Fig. 7 to summarize the effects produced by the nanoporous surface structures under identical experimental conditions (quenching fluid and its thermodynamic conditions, test rod geometry, orientation and position in the pool, and test rod material). Therefore, the differences between the two quenching curves are solely due to the nanoporous structures on the test rod surface. The large difference between the two quenching curves does not present as a surprise as Witte and Lienhard [21] has pointed out this possibility using several experimental evidences. Their main concept is that the surface properties such as the surface nanoscale cavity characteristics and wettability could modify the nucleation and vaporization mechanisms that can enhance the heat transfer rates and accordingly the quenching curve characteristics. Next, the effects from the nanoporous surface are analyzed for each boiling regime. 4.4.1. Effects on film boiling As shown in Fig. 7, the nanoporous surface shows a higher heat flux in the film boiling regime when the surface temperature is higher than 200 K. After this temperature, the heat flux from the nanoporous surface tends to be lower until the minimum heat flux point (Leidenfrost point) is reached. However, the average heat transfer in the film boiling regime is higher for the nanoporous surface as its surface chilled down faster in the film boiling regime than that of the normal surface. Another significant difference is that the nanoporous surface has produced a relatively large increase in the Leidenfrost temperature from 126 ± 6 K to 158 ± 2.6 K, a 32 K difference, corresponding to a 25% increase that shortens the film boiling regime substantially. We would attribute this change in the Leidenfrost temperature solely due to the increase in the wettability by the superhydrophilic nanoporous surface as pointed out by Kruse et al. [17] that was discussed above in the literature review section. According to Witte and Lienhard [21] the Leidenfrost point on the normal surface is lower because the surface is not hot enough to generate sufficient vapor to keep the film from collapsing. However, the liquid is able to contact the surface on the nanoporous surface due to the enhanced wettability at a higher temperature, terminating the film boiling regime.
Comparing to the Leidenfrost point of the normal surface, Witte and Lienhard [21] called the higher Leidenfrost temperature of the nanoporous surface a ‘‘jump’’ from the Leidenfrost temperature on the traditional surface. The Leidenfrost point on the normal surface is due to the phenomenon explained by hydrodynamic theory that the surface is not hot enough to generate enough vapor to keep the film from collapsing. However the reason for the ‘‘jump’’ on the nanoporous surface is owing to the enhanced wettability that enables the liquid to contact the surface at a higher temperature resulting in the ending of the film boiling. As discussed in the above, the difference in heat transfer between the two surfaces is small in the film boiling regime (above 150 K) that is mainly due to the fact that the nanoporous surface basically makes the surface more hydrophilic that enhances the wettability and the nanoporous surface also provides more nucleation sites, both of those increase the rate of nucleation process. However, in film boiling, nucleation is not a contributing heat transfer mechanism. 4.4.2. Effects on transition boiling The most significant modification to the boiling mechanisms by the nanoporous surface was found in the transition regime. If we examine the quenching curve of the nanoporous surface given in Fig. 7 for the transition regime, it is obvious that the curve is composed of two sub-regimes with different slopes connected at a surface temperature of about 110 K. Once again, Witte and Lienhard [21] has provided some insights for this phenomenon and suggested that the lower heat flux sub-regime be called the transitional film boiling while the higher heat flux one be called the transitional nucleate boiling. The transitional film boiling subregime is essentially an extension of the film boiling regime where the film boiling mechanism is still dominant but some liquid–solid contact can occur. Even the slightest liquid contact can lead to substantial nucleation and vapor generation due to the relatively higher surface temperatures. As the temperature of the surface decreases, more surface area can support the liquid contact and vapor generation that leads to a rise in heat flux. As the wall continues to cool down, at some point the surface area that allows liquid contact eventually overwhelms the dry area, moving the heat transfer process into the transitional nucleate boiling sub-regime. As the wall temperature continues to drop, more nucleation and vapor generation yield further increases in heat flux until the vapor escape path is no longer able to vent all the vapor generated, which corresponds to the CHF. Compared to the normal surface, the nanoporous surface produced much higher heat fluxes in the transition boiling regime that is mainly due to the enhanced wettability and more nucleation cavity sites offered by the nanoporous surface. The enhanced wettability raised the Leidenfrost temperature that gives rise to relatively higher temperatures (more thermal energy potential) in the transition regime. The higher temperatures combined with more nucleation capability promote higher heat fluxes in the transition boiling regime. 4.4.3. Effects on CHF and nucleate boiling As discussed above, CHF is simply the end point of the transition nucleate boiling sub-regime. The enhancement mechanism facilitated by the nanoporous surface described above results in a nearly 160% increase in CHF from 40 ± 3.1 to 105 ± 7.8 kW/m2. In the nucleate boiling regime, the surface is completely wettable and heat transfer is solely by nucleation and bubble generation. As mentioned above, the nanoporous surface is superhydrophilic and possesses extra nanoscale nucleation cavities that strongly enhance the nucleation and bubble generation rates, yielding much higher heat fluxes than the normal surface.
H. Hu et al. / International Journal of Heat and Mass Transfer 80 (2015) 636–643
4.5. Chilldown applications In the current experiment, a test rod was used to evaluate the modification and enhancement of a nanoporous surface on the quenching heat transfer. In actual cryogenic chilldown applications, the quenching process takes place inside a tube or pipe. As seen in Fig. 5(B), the chilldown time is shorter for the nanoporous surface by approximately 20%. If we assume the same mass flow rates are used to chill two pipes with identical dimensions, wall thickness and material, it is reasonable to suggest that a pipe with a nanoporous inner surface would save approximately 20% in cryogenic fluid consumption. 5. Conclusion A quenching experiment has been conducted to compare the chilldown performance between a normal aluminum surface and a nanoporous surface. The nanoporous surface has shown superior phase-change heat transfer performance than the normal surface in all three boiling regimes. Several important discoveries are listed as follows: 1. The anodized aluminum oxide (AAO) nanoporous surface has proven to survive the cryogenic temperature environment and also demonstrated its potential to be used in future space exploration applications. 2. The nanoporous surface results in a 160% improvement in the CHF and a 32 K increase in the Leidenfrost temperature, leading to a reduction in chilldown time by 20%. 3. The most significant enhancement effects produced by the nanoporous surface are the increase in the Leidenfrost temperature and much higher heat fluxes in the transition and nucleate boiling regimes. Conflict of interest None declared. Acknowledgment This research was partially supported by the Andrew H. Hines, Jr./Progress Energy Endowment Fund. References [1] R. Shaeffer, H. Hu, J.N. Chung, An experimental study on liquid nitrogen pipe chilldown and heat transfer with pulse flows, Int. J. Heat Mass Transfer 67 (2013) 955–966.
643
[2] E.G. Brentari, J.A. Brennan, R.V. Smith, W.G. Steward, An Experimental Report on Cooldown of Cryogenic Transfer Lines, National Bureau of Standards, Boulder, Colorado, 1966. [3] O. Kawanami, T. Nishida, I. Honda, Y. Kawashima, H. Ohta, Flow and heat transfer on cryogenic flow boiling during tube quenching under upward and downward flow, Microgravity Sci. Technol. 19 (3–4) (2007) 137–138. [4] B. Verthier, G.P. Celata, G. Zummo, C. Colin, J. Follet, Effect of gravity on film boiling heat transfer and rewetting temperature during quenching, Microgravity Sci. Technol. 21 (S1) (2009) 185–191. [5] C. Luo, M. Xiang, Angle inequality for judging the transition from Cassie–Baxter to Wenzel states when a water drop contacts bottoms of grooves between micropillars, Langmuir: ACS J. Surf. Colloids 28 (38) (2012) 13636–13642. [6] M. Xiang, A. Wilhelm, C. Luo, Existence and role of large micropillars on the leaf surfaces of The President lotus, Langmuir: ACS J. Surf. Colloids 29 (25) (2013) 7715–7725. [7] C. Luo, M. Xiang, X. Heng, A stable intermediate wetting state after a water drop contacts the bottom of a microchannel or is placed on a single corner, Langmuir: ACS J. Surf. Colloids 28 (25) (2012) 9554–9561. [8] S.K. Singh, S. Khandekar, D. Pratap, S.A. Ramakrishna, Wetting dynamics and evaporation of sessile droplets on nano-porous alumina surfaces, Colloids Surf. A 432 (2013) 71–81. [9] N. Tasaltin, D. Sanli, A. Jonas, A. Kiraz, C. Erkey, Preparation and characterization of superhydrophobic surfaces based on hexamethyldisilazane-modified nanoporous alumina, Nanoscale Res. Lett. 6 (2011) 487. [10] M.K. Scott, G. Liter, Pool-boiling CHF enhancement by modulated porous-layer coating- theory and experiment, Int. J. Heat Mass Transfer 44 (2001) 4277– 4310. [11] M. Arik, A. Bar-Cohen, S.M. You, Enhancement of pool boiling critical heat flux in dielectric liquids by microporous coatings, Int. J. Heat Mass Transfer 50 (5– 6) (2007) 997–1009. [12] Y.-Y. Li, Z.-H. Liu, G.-S. Wang, L. Pang, Experimental study on critical heat flux of high-velocity circular jet impingement boiling on the nano-characteristic stagnation zone, Int. J. Heat Mass Transfer 67 (2013) 560–568. [13] C.-C. Hsu, P.-H. Chen, Surface wettability effects on critical heat flux of boiling heat transfer using nanoparticle coatings, Int. J. Heat Mass Transfer 55 (13–14) (2012) 3713–3719. [14] H. O’Hanley, C. Coyle, J. Buongiorno, T. McKrell, L.-W. Hu, M. Rubner, R. Cohen, Separate effects of surface roughness, wettability, and porosity on the boiling critical heat flux, Appl. Phys. Lett. 103 (2) (2013) 024102. [15] H.T. Phan, N. Caney, P. Marty, S. Colasson, J. Gavillet, Surface wettability control by nanocoating: the effects on pool boiling heat transfer and nucleation mechanism, Int. J. Heat Mass Transfer 52 (23–24) (2009) 5459– 5471. [16] A. Hens, R. Agarwal, G. Biswas, Nanoscale study of boiling and evaporation in a liquid Ar film on a Pt heater using molecular dynamics simulation, Int. J. Heat Mass Transfer 71 (2014) 303–312. [17] C. Kruse, T. Anderson, C. Wilson, C. Zuhlke, D. Alexander, G. Gogos, S. Ndao, Extraordinary shifts of the Leidenfrost temperature from multiscale micro/ nanostructured surfaces, Langmuir: ACS J. Surf. Colloids 29 (31) (2013) 9798– 9806. [18] X. Li, W. Wei, R. Wang, Y. Shi, Numerical and experimental investigation of heat transfer on heating surface during subcooled boiling flow of liquid nitrogen, Int. J. Heat Mass Transfer 52 (2009) 1510–1516. [19] H.J.A.P. van de Boogaard, A.R. Balkenende, M. Scholten, N.P. Willard, Evaluation of different approaches to assess the surface tension of low-energy solids by means of contact angle measurements, Langmuir: ACS J. Surf. Colloids 14 (20) (1998) 5907–5912. [20] O.R. Burggraf, An exact solution of the inverse problem in heat conduction theory and applications, J. Heat Transfer 86 (3) (1962) 373–380. [21] J.H. Lienhard, L.C. Witte, On the existence of two ’transition’ boiling curves, Int. J. Heat Mass Transfer 25 (6) (1982) 771–779.
Contents lists available at ScienceDirect
International Journal of Heat and Mass Transfer journal homepage: www.elsevier.com/locate/ijhmt
Modification and enhancement of cryogenic quenching heat transfer by a nanoporous surface Hong Hu a, Cheng Xu b, Yang Zhao b, Reid Shaeffer a, Kirk J. Ziegler b, J.N. Chung a,⇑ a b
Cryogenics Heat Transfer Laboratory, Department of Mechanical and Aerospace Engineering, University of Florida, Gainesville, FL 32611-6300, USA Nanostructured Interfaces Laboratory, Department of Chemical Engineering, University of Florida, Gainesville, FL 32611-6005, USA
a r t i c l e
i n f o
Article history: Received 27 June 2014 Received in revised form 21 September 2014 Accepted 24 September 2014
Keywords: Cryogenics Chilldown Nanoporous surfaces Leidenfrost point Critical heat flux Quenching curve
a b s t r a c t Since the average chilldown efficiency for cryogenic systems is only about 8%, significant improvements to heat transfer are needed for many applications. An experiment was performed to evaluate the modification and enhancement on the quenching heat transfer by a nanoporous heat transfer surface in this study. For comparison purposes, two sample surfaces were used. One is the mechanically polished conventional normal aluminum surface serving as the base case and the other is an aluminum surface with the anodized aluminum oxide (AAO) nanoporous finish. In this work, the effect of the nanoporous surface on the heat transfer during chilldown in a liquid nitrogen pool is investigated. The results indicated that the nanoporous surface completely modified and enhanced the phase-change heat transfer in all three quenching regimes. Comparing to the conventional surface case, the Leidenfrost temperature was increased by 32 K and the critical heat flux (CHF) was raised by 160% due to the nanoporous surface. However, the most significant modification on the boiling mechanisms by the nanoporous surface was found in the transition regime that is composed of transitional film and transitional nucleate sub-regimes with quite different quenching curve slopes. For cryogenic quenching applications, it is estimated that the nanoporous surface could save 20% in the amount of cryogen consumption by shortening the chilldown time. The modification and enhancement are mainly attributed to the superhydrophilic property and nanoscale nucleation sites offered by the nanoporous surface. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Cryogenic fluids are widely used in industrial applications, space exploration, and cryosurgery systems. Specifically, in space exploration, cryogenic fluids are used in power and propulsion, thermal management, and life-support systems of a spacecraft during space missions that involves transport, handling, and storage of these fluids under both terrestrial and microgravity conditions. For example, cryogens are usually used as liquid fuels, such as liquid hydrogen and oxygen, that are burned in liquid-fueled rocket engines. When a cryogenic fluid transport system is first started up, its walls and hardware must go through a transient chilldown period prior to reaching steady-state operation. Therefore, chilldown is the process of adjusting the system to the low temperature regime, which is usually several hundred degrees below room temperature. The chilldown or quenching process is a complicated phasechange phenomenon involving unsteady two-phase flows as well as heat and mass transfer; its characteristics have not been fully understood, resulting in very low chilldown efficiencies. ⇑ Corresponding author. Tel.: +1 352 392 9607; fax: +1 352 392 1071. E-mail address: jnchung@ufl.edu (J.N. Chung). http://dx.doi.org/10.1016/j.ijheatmasstransfer.2014.09.056 0017-9310/Ó 2014 Elsevier Ltd. All rights reserved.
Fig. 1 is a typical boiling curve where the heater surface heat flux, q00 , is plotted against the heater surface degree of superheating, TW Tsat, where TW is the surface temperature and Tsat is the saturation temperature corresponding to the system pressure. Boiling is characterized by the surface temperature and heat flux on the wall according to the boiling curve. Since the wall temperature is strictly determined by an energy balance of the wall, the heat fluxes in and out of the tube wall are important parameters. In boiling, the heat source is externally supplied and, thus, can be controlled independently, such as the constant wall heat flux condition. In this case, boiling is a heat flux (independent variable) controlled process. For the heat flux controlled condition, boiling follows the route of A ? B ? D. In order to avoid a huge temperature jump, boiling usually runs safely below B in the nucleate boiling regime. In contrast, the wall does not have an external source of heat (except a small amount of residual heating) during chilldown; therefore, heat coming out of the wall can only be supplied internally from the thermal capacity (stored energy) of the wall. The only way to get heat to be removed from the system is by lowering the inner wall surface temperature by cooling from the cryogenic flow. Accordingly, the wall inner surface temperature is the independent variable and the control parameter. In summary, quenching
H. Hu et al. / International Journal of Heat and Mass Transfer 80 (2015) 636–643
is a conjugate process involving both heat transfer and wall–fluid interaction. The quantity of heat (heat flux) that is released to the fluid can only be associated with the temperature change of the wall inner surface. As a result, the wall surface temperature is the controlling parameter that forces the quenching process to follow the route D ? C ? B ? A. In boiling applications, such as those in the cooling of a nuclear reactor, a super high flux of heat transfer in the nucleate boiling regime is standard practice. The only precaution is not to exceed the critical heat flux (CHF) and enter the film boiling regime. However, film boiling is the first mode of boiling encountered in chilldown and it cannot be avoided. Due to its low heat fluxes at high wall temperatures, the chilldown efficiency in general is extremely low. According to Shaeffer et al. [1], the average chilldown efficiency that is defined as the ratio of the amount of thermal energy removed from the wall versus the maximum cooling capability of the cryogen spent in a phase change process is about 8%, highlighting the tremendous need to improve the quenching efficiency for many applications that require a cryogen as the working fluid. Advances in nanotechnology enable modification and enhancement of heat transfer based on the phase change from liquid to vapor. This paper reports significant modification and enhancement to heat transfer throughout the entire boiling regime with the purpose of substantially improving the quenching thermal efficiency.
2. Literature review 2.1. Cryogenic chilldown experiments Because of the difficulties associated with the design and measurements of cryogenic experiments, there is less chilldown data for cryogenic fluids compared to traditional fluids. Researchers began to investigate heat and mass transfer during cryogenic liquid
637
transfer line chilldown because that this area is of extreme importance to spacecraft design. Chilldown of these transfer lines were first studied by Brennan et al. [2] under normal gravity conditions for liquid nitrogen (LN2) and liquid hydrogen (LH2). Antar and Collins found a unique flow pattern in microgravity conditions, which is known as the filament flow. Kawanami et al. [3] also observed a similar flow pattern in terrestrial conditions. They claimed that the gravity effects on heat transfer of forced convective boiling decreased with an increase in the mass velocity. Verthier et al. [4] have studied the effect of gravity during quenching using FC72. LN2 was used as the working fluid [4] to investigate the heat transfer characteristics and flow pattern during the quenching of a vertical tube under both terrestrial and ten-second microgravity conditions. It is claimed that the heat transfer and quenching front velocity under microgravity conditions increased up to 20% compared to those in traditional gravity conditions. 2.2. Heat transfer enhancement by nanoscale surface structures The majority, if not all, of the boiling heat transfer enhancement endeavors focused on the nucleate boiling regime and the CHF. There have been a number of experiments to measure the contact angle and wettability on the nanoporous surface. Luo’s group [5–7] have introduced more detail discussion both theoretically and experimentally on the wettability on nano pillar surface and derived an angle criterion determining transition from Cassie-Baxter to Wenzel states. Singh et al. [8] investigated the wetting and evaporation of sessile droplets on nanoporous anodic aluminum oxide (AAO) substrates having different pore distribution (uniform, random and linearly arranged) morphologies and pore sizes (70– 120 nm). They claimed that nano structured surface is an applicable tool to control wettability as well as the diffusive evaporation process. In addition, physical morphology and pore distribution affect wettability as well as evaporation rates. Tasaltin et al. [9]
Fig. 1. A typical boiling curve including different boiling regimes and corresponding flow patterns.
638
H. Hu et al. / International Journal of Heat and Mass Transfer 80 (2015) 636–643
made nanoporous AAO surfaces with a hexagonal structure of circular pores with 250 nm in diameter and 300 nm in depth and found the surfaces to be superhydrophobic. The contact angle was measured at 153.2 ± 2° for vapor-phase hexamethyldisilazane (HMDS), that is larger than the average water contact angle of 82.9 ± 3° on smooth thin film AAO surfaces. The excellent surface wettability associated with these nanoporous surfaces has significant potential to enhance heat transfer. Liter and Kaviany [10] reported a nearly three times improvement to heat transfer for a porous-layer coating over a conventional surface. Arik et al. [11] reported an experimental study finding that microporous, diamond-coated surfaces provided an average 1.6 enhancement factor on FC-72 CHF. To our knowledge, there is no literature discussing cryogenic quenching with nanoporous surfaces. Li et al. [12] reported an experimental study that investigated the CHF of a high-velocity circular water jet impingement boiling on the stagnation zone of a nano-structured heater surface with different surface wettability parameters. The wettability of the copper heater was modified with nanostructures to change the surface topography and chemistry to exhibit either hydrophilic or hydrophobic character. They found that the wettability holds a strong influence on the CHF. For example, the CHF almost doubled when the contact angle was reduced from 105° to 5°. Hsu and Chen [13] experimentally investigated the effects of surface wettability on water pool boiling heat transfer. They used coatings of nano-silica particles to change the wettability of the copper surface from superhydrophobic to superhydrophilic. They demonstrated that CHF values increased from 300 to 1483 kW/m2 when the contact angle decreased from 145° (hydrophobic) to less than 10° (hydrophilic). O’Hanley et al. [14] experimentally evaluated the separate effects of surface wettability, porosity, and roughness on the CHF of water using engineered surfaces. They reported that porous hydrophilic surfaces enhanced CHF by 50–60%, while porous hydrophobic surfaces resulted in a reduction of CHF by 97%. Phan et al. [15] also focused on the influence of surface wettability on nucleate boiling heat transfer. They used nanocoating techniques to vary the water contact angle from 20° to 110°. They reported that a greater surface wettability due to the hydrophilic nature of
the surface increases the vapor bubble departure radius and reduces the bubble emission frequency. The best heat transfer coefficients were obtained for the surfaces that had a water contact angle close to either 0° or 90°. Hens et al. [16] used molecular dynamics simulations to understand the boiling mechanisms of argon bubble formation on a platinum substrate with a particular emphasis on the effects of nanosized texture on the surface. The bubble nucleation and vapor film formation showed a dependence on the degree of superheat and solid–liquid interfacial wettability since bubbles do not form easily on non-wetting surfaces. They concluded that a wettable (hydrophilic) surface facilitates higher heat transfer. The only study found that addressed the effects of nano-textured surfaces on the Leidenfrost temperature is given by Kruse et al. [17]. The micro- and nano-structures were fabricated on a stainless steel surface by a femtosecond laser. The traditional Leidenfrost droplet experiment was performed on both machine-polished and nanotextured surfaces. They observed an extraordinary shift in the Leidenfrost temperature from 280 °C for the machine-polished surface to 450 °C for the nano-textured superhydrophilic surface. The gigantic 175 °C increase in the Leidenfrost temperature was attributed to a reduction in the contact angle that results in intermittent liquid/solid contact and capillary wicking action.
3. Experiment 3.1. Experimental apparatus and procedure A simple system was built that can explore all the fundamental physics and mechanisms of phase-change heat transfer while eliminating all unnecessary complicating factors. Instead of an internal convective flow chilldown system, a pool quenching experiment was used since the basic characteristics of the quenching curve (reversed boiling curve) are not fundamentally affected by the bulk flows. The basic schematic of the design is illustrated in Fig. 2. A transparent glass dewar with a vacuum insulation from the surroundings was used to form a liquid nitrogen pool that was
Fig. 2. Schematic of experimental apparatus (A) overview of test apparatus made of 8020 framework. (B) Test rod design.
H. Hu et al. / International Journal of Heat and Mass Transfer 80 (2015) 636–643
maintained at atmospheric pressure by an opening to atmosphere. The test section was an aluminum rod (99.99% purity) with a thermocouple bead embedded along the centerline of the rod and 0.4 inch from the bottom surface. A fast response and high sensitivity T-type 40AWG (0.0799 mm diameter) micro thermocouple was used to record the temperature history at 14 Hz data acquisition frequency. A set of background lights and a high speed camera were setup to collect the video data on the two-phase bubble dynamics. The video has been utilized to analyze the behavior of the interaction between vapor and liquid. The recording speed was set at 2000 frames per second. To study the effect of nanoporous surfaces on cryogenic quenching, chilldown experiments were conducted with two different surfaces. One is a commercially available machine polished conventional aluminum surface, which is referred as the normal surface, and an aluminum surface coated with a thin anodized aluminum oxide (AAO) nanoporous finish. The experiment started when the test rod was submerged in the dewar containing LN2. In order to demonstrate the repeatability and reduce aleatory uncertainty, 15 identical tests were conducted on each nanoporous surface and normal surface.
3.2. Preparation of anodized aluminum oxide (AAO) nanoporous surface The AAO nanoporous surface was fabricated by a two-step anodization process. A pure aluminum rod was cleaned by sonication in soapy de-ionized water, acetone, and ethanol, respectively. The aluminum rod was then put into an electropolish solution (Electro Polish System Inc.) at 65 °C and a constant voltage of 17 V for 20 min. After that, the first anodization was carried out at 15 °C and a constant voltage of 40 V in 0.3 M oxalic acid solution with stirring for 16 h. After stripping off the AAO layer in a mixture of 10 wt% phosphoric acid and 1.8 wt% chromic acid, a second anodization was performed at the same conditions. The time was varied to control the final AAO thickness to approximately 5 lm. At the end of the second anodization, the voltage was reduced at a rate of 0.5 V per 30 s to thin the barrier layer. The pore opening process was in 5 wt% phosphoric acid at room temperature and was monitored by an electrochemical setup (VersaStat 3, Princeton Applied Research) using a small voltage of 0.1 V applied against a carbon counter electrode. The current was controlled carefully so that the pore opening process stopped as soon as the current increased remarkably, which indicates that the pores were fully opened. The finished surface results in a pattern of perfectly arranged ordered nanopores formed on the aluminum rod surface, as shown in Fig. 3. The nanoporous surface has a hexagonal structure with pores approximately 50 nm in diameter and depths of 5 lm. The contact angles of the surfaces were measured with a 1 lL drop of deionized water at room temperature using a Kruss DSA100 Drop Shape Analyzer. Contact angles were measured for
639
3 drops at different locations on the surface. Fig. 4 shows the difference in contract angles for each surface. The average contact angle on the nanoporous and normal surface is 13.1 ± 1.7° and 74.6 ± 1.6°, respectively. Therefore, the nanoporous surface demonstrates superhydrophilic properties. However as discussed above, Tasaltin et al. [9] made nanoporous AAO surfaces with a hexagonal structure of circular pores with 250 nm in diameter and 300 nm in depth and found the surfaces to be superhydrophobic. Therefore the wettability depends strongly on the dimension and depth of the nano pores. The contact angles for liquid nitrogen on most conventional normal metal surfaces are relatively small (less than 7°) [18]. Also there is no report in the literature that addresses the effects of nanoporous surface on the contact angle of liquid nitrogen. Due to the low temperate nature and a lack of experimental facility, we were not able to perform the measurement. As the contact angle theory [19] has pointed out that an increasing surface free energy leads to a decreasing contact angle, nanoporous surfaces have been shown to possess larger surface energy values than normal surfaces based on the experimental results. Therefore, we expect that this is the case for our nanoporous surface. As the contact angle of liquid nitrogen on a normal metal surface is already very small, a contact angel drop of just a few degrees would mean a large percentage decrease that is certainly reflected in our experimental results. 3.3. Uncertainty analysis The dimensions of the test rod were determined by a caliper. The mass of the test rod was measured through an electronic scale. The thermocouples had the uncertainty of ±1.5% full scale according to the manufacturer. Another uncertainty source for temperature measurements comes from the DAQ whose uncertainty was ±0.1% full scale. In the current experiment, the inverse heat conduction methods [20] were used to obtain the surface heat fluxes from the measured temperatures. The uncertainty of the heat flux is calculated through a square root summation of partial derivatives. Both the uncertainties from the experimental apparatus and calculated parameters are listed in Table 1. 4. Results and discussion 4.1. Data reduction and analysis For the study of the liquid-to-vapor phase-change heat transfer, the so-called ‘‘boiling curve’’ (see Fig. 1) is often of great interest because it denotes the regimes of the different heat transfer mechanisms and is crucial for engineering applications. However, as mentioned above, the surface temperature at the bottom of the rod (point A in Fig. 2) is the independent variable during the chilldown or quenching process; therefore, the heat flux at point
Fig. 3. (A) Top- and (B) cross-view SEM images of the nanoporous surface.
640
H. Hu et al. / International Journal of Heat and Mass Transfer 80 (2015) 636–643
Fig. 4. Contact angle on (A) nanoporous surface and (B) normal surface.
Table 1 Summary of the uncertainty. Source of uncertainty
Symbol
Uncertainty
Rod inside diameter Rod mass Outside temperature Inner temperature Heat flux
di (mm) Mr (g) To (K) Ti (K) qi (K)
±0.001 mm ±0.0001 g ±1.6% ±1.6% ±6.8%
A cannot be controlled independently. Additionally, the direct measurement of temperature for point A on the test rod is difficult. Therefore, the surface heat flux can only be inferred from the temperature history recorded by the embedded thermocouple. The heat flux and the surface temperature at point A were obtained by solving the inverse heat conduction problem (IHCP). The test section is approximated as a one-dimensional circular slab. The effect of radiation on the outer surface is neglected. An analytical method was derived [20] to obtain the surface temperature and heat flux at point A of the exposed surface based on the temperature of a thermocouple positioned inside the rod of the test section by the following equation,
To ¼ Ti þ
2 2 3 3 1 x2 dT i 1 x2 d T i 1 x2 d T i þ þ þ 2 3 2! ar dt 4! ar 6! ar dt dt
ð1Þ
where the subscript o denotes rod bottom surface (point A in Fig. 2) while i indicates the thermocouple location. The distance x is between the thermocouple and point A, t is the measurement time, and ar is the thermal conductivity of the rod. The heat flux at point A can be approximated by the first three leading terms of an infinite series given below,
q00o
" 2 3 # 2 dT i 1 x2 d T i 1 x2 d T i ¼ qr C p;r x þ þ 3 dt 3! ar dt2 5! ar dt
ð2Þ
where Cp,r is the specific heat of the rod and qr is the density of the aluminum rod. 4.2. Chilldown heat transfer characteristics The quenching (chilldown) process takes place when the rod is chilled in the pool from room temperature (300 K) down to the liquid nitrogen pool temperature at 77.35 K, which is the saturated temperature under the atmospheric pressure. The time dependent surface temperatures and heat fluxes at point A (see Fig. 2) during chilldown were estimated using the thermocouple measured temperatures and Eqs. (1) and (2), as outlined above. With the temperature and heat flux histories obtained, the chilldown heat transfer characteristics can be generated accordingly. For the purposes of illustrating the differences of chilldown heat transfer between a mechanically finished conventional surface and a nanoporous surface, respective quenching and chilldown curves are given in
Fig. 5(A) and (B). It is noted that the two plots share the same vertical axis that indicates the temperature of the rod bottom surface at point A in Fig. 2. As seen in Fig. 5(B), the total chilldown time is shorter for the nanoporous surface. Specifically, the time for the surface to reach 80 K is 28 s and 35 s for the nanoporous and normal surface, respectively, corresponding to a 20% improvement in time. This substantial savings in chilldown time was mainly achieved by the nanoporous surface in the film boiling regime. The normal surface spent 32 s in the film boiling regime while it only took 24 s for the nanoporous surface. As shown in Fig. 5(A), the nanoporous surface facilitated a higher Leidenfrost temperature (158 K) that resulted in a quicker switch from the film boiling regime to the transition boiling regime as compared to the normal surface whose Leidenfrost temperature was found at 126 K that took 8 s longer to reach. Other than the Leidenfrost point, the nanoporous surface produced a substantially modified transition boiling process as compared to the normal surface that presents an ordinary transition boiling pattern. The transition boiling on the nanoporous surface is considered to be composed of two different heat transfer mechanisms that resulted in two apparently different slopes. Mainly due to the drastically modified transition regime, both the heat flux at the CHF point and those in the nucleate boiling regime were raised substantially. 4.3. Flow visualization of two-phase quenching process Fig. 6 provides images of the solid–fluid interactions on the normal and nanoporous rod surfaces during the chilldown process in the specific boiling regimes. It is seen that the liquid–vapor hydrodynamics and the resulting flow patterns along the surfaces evolve during chilldown. We found that distinct differences were observed between the two surfaces. For the film boiling regime shown in Fig. 6(A) and (D), it can be seen that the vapor film on the normal surface is much thicker and contains more dynamic fluctuations and liquid–vapor interactions than that on the nanoporous surface. The thinner films explain the higher heat fluxes associated with the nanoporous surface. The nanoporous surface also produced smoother liquid–vapor interfaces, which can be explained by a ‘micro-damper’ theory. Millions of nanoscale porous cavities work as millions of micro-dampers that facilitate the attenuation of the high frequency perturbations. As a result, perturbations cannot be amplified on the nanoporous surface, while the surface roughness would magnify and amplify the perturbations on the normal surface to form turbulent flows and interactions. In the transition boiling regime as shown in Fig. 6(B) and (E), the normal surface still produced thicker and more chaotic two-phase patterns, similar to those observed in the film boiling regime. However, the liquid–vapor layers are thicker than those in the film boiling, which is due to the fact that more vapor is generated as the
H. Hu et al. / International Journal of Heat and Mass Transfer 80 (2015) 636–643
641
Fig. 5. (A) Quenching and (B) chilldown curves for conventional and nanoporous surfaces. The different regimes during chilldown are also shown.
Fig. 6. Comparison of the liquid–vapor flow patterns formed on conventional (top panel) and nanoporous surfaces (bottom panel) during liquid nitrogen chilldown. (A, D) Film boiling region; (B, E) transition boiling region; (C, F) nucleate boiling region.
liquid starts to contact the surface, resulting in high rates of nucleation and bubble generation. In the nucleate boiling regime, the two-phase flow pattern on the nanoporous surface shown in Fig 6(F) indicates that a large
amount of small and separated bubbles are generated in an explosive manner. However, in Fig. 6(C) the normal surface results in a different pattern where bubbles tend to merge together to form larger vapor slugs that point outwards from the surface. This
642
H. Hu et al. / International Journal of Heat and Mass Transfer 80 (2015) 636–643
Fig. 7. Comparison of the heat flux for nanoporous and conventional surfaces during chilldown.
phenomenon results in longer attachment times for the bubbles on the surface, which is the main reason why the normal surface has lower heat fluxes in the nucleate boiling regime. The proper sized and patterned nano pores on the nanoporous surface alters the wettability, allowing more nucleation sites to be created that enhance nucleation and bubble escape from the surface. 4.4. Enhancement and modification by the nanoporous surface The two quenching curves (reversed boiling curves) obtained from the normal and nanoporous surface are plotted together in Fig. 7 to summarize the effects produced by the nanoporous surface structures under identical experimental conditions (quenching fluid and its thermodynamic conditions, test rod geometry, orientation and position in the pool, and test rod material). Therefore, the differences between the two quenching curves are solely due to the nanoporous structures on the test rod surface. The large difference between the two quenching curves does not present as a surprise as Witte and Lienhard [21] has pointed out this possibility using several experimental evidences. Their main concept is that the surface properties such as the surface nanoscale cavity characteristics and wettability could modify the nucleation and vaporization mechanisms that can enhance the heat transfer rates and accordingly the quenching curve characteristics. Next, the effects from the nanoporous surface are analyzed for each boiling regime. 4.4.1. Effects on film boiling As shown in Fig. 7, the nanoporous surface shows a higher heat flux in the film boiling regime when the surface temperature is higher than 200 K. After this temperature, the heat flux from the nanoporous surface tends to be lower until the minimum heat flux point (Leidenfrost point) is reached. However, the average heat transfer in the film boiling regime is higher for the nanoporous surface as its surface chilled down faster in the film boiling regime than that of the normal surface. Another significant difference is that the nanoporous surface has produced a relatively large increase in the Leidenfrost temperature from 126 ± 6 K to 158 ± 2.6 K, a 32 K difference, corresponding to a 25% increase that shortens the film boiling regime substantially. We would attribute this change in the Leidenfrost temperature solely due to the increase in the wettability by the superhydrophilic nanoporous surface as pointed out by Kruse et al. [17] that was discussed above in the literature review section. According to Witte and Lienhard [21] the Leidenfrost point on the normal surface is lower because the surface is not hot enough to generate sufficient vapor to keep the film from collapsing. However, the liquid is able to contact the surface on the nanoporous surface due to the enhanced wettability at a higher temperature, terminating the film boiling regime.
Comparing to the Leidenfrost point of the normal surface, Witte and Lienhard [21] called the higher Leidenfrost temperature of the nanoporous surface a ‘‘jump’’ from the Leidenfrost temperature on the traditional surface. The Leidenfrost point on the normal surface is due to the phenomenon explained by hydrodynamic theory that the surface is not hot enough to generate enough vapor to keep the film from collapsing. However the reason for the ‘‘jump’’ on the nanoporous surface is owing to the enhanced wettability that enables the liquid to contact the surface at a higher temperature resulting in the ending of the film boiling. As discussed in the above, the difference in heat transfer between the two surfaces is small in the film boiling regime (above 150 K) that is mainly due to the fact that the nanoporous surface basically makes the surface more hydrophilic that enhances the wettability and the nanoporous surface also provides more nucleation sites, both of those increase the rate of nucleation process. However, in film boiling, nucleation is not a contributing heat transfer mechanism. 4.4.2. Effects on transition boiling The most significant modification to the boiling mechanisms by the nanoporous surface was found in the transition regime. If we examine the quenching curve of the nanoporous surface given in Fig. 7 for the transition regime, it is obvious that the curve is composed of two sub-regimes with different slopes connected at a surface temperature of about 110 K. Once again, Witte and Lienhard [21] has provided some insights for this phenomenon and suggested that the lower heat flux sub-regime be called the transitional film boiling while the higher heat flux one be called the transitional nucleate boiling. The transitional film boiling subregime is essentially an extension of the film boiling regime where the film boiling mechanism is still dominant but some liquid–solid contact can occur. Even the slightest liquid contact can lead to substantial nucleation and vapor generation due to the relatively higher surface temperatures. As the temperature of the surface decreases, more surface area can support the liquid contact and vapor generation that leads to a rise in heat flux. As the wall continues to cool down, at some point the surface area that allows liquid contact eventually overwhelms the dry area, moving the heat transfer process into the transitional nucleate boiling sub-regime. As the wall temperature continues to drop, more nucleation and vapor generation yield further increases in heat flux until the vapor escape path is no longer able to vent all the vapor generated, which corresponds to the CHF. Compared to the normal surface, the nanoporous surface produced much higher heat fluxes in the transition boiling regime that is mainly due to the enhanced wettability and more nucleation cavity sites offered by the nanoporous surface. The enhanced wettability raised the Leidenfrost temperature that gives rise to relatively higher temperatures (more thermal energy potential) in the transition regime. The higher temperatures combined with more nucleation capability promote higher heat fluxes in the transition boiling regime. 4.4.3. Effects on CHF and nucleate boiling As discussed above, CHF is simply the end point of the transition nucleate boiling sub-regime. The enhancement mechanism facilitated by the nanoporous surface described above results in a nearly 160% increase in CHF from 40 ± 3.1 to 105 ± 7.8 kW/m2. In the nucleate boiling regime, the surface is completely wettable and heat transfer is solely by nucleation and bubble generation. As mentioned above, the nanoporous surface is superhydrophilic and possesses extra nanoscale nucleation cavities that strongly enhance the nucleation and bubble generation rates, yielding much higher heat fluxes than the normal surface.
H. Hu et al. / International Journal of Heat and Mass Transfer 80 (2015) 636–643
4.5. Chilldown applications In the current experiment, a test rod was used to evaluate the modification and enhancement of a nanoporous surface on the quenching heat transfer. In actual cryogenic chilldown applications, the quenching process takes place inside a tube or pipe. As seen in Fig. 5(B), the chilldown time is shorter for the nanoporous surface by approximately 20%. If we assume the same mass flow rates are used to chill two pipes with identical dimensions, wall thickness and material, it is reasonable to suggest that a pipe with a nanoporous inner surface would save approximately 20% in cryogenic fluid consumption. 5. Conclusion A quenching experiment has been conducted to compare the chilldown performance between a normal aluminum surface and a nanoporous surface. The nanoporous surface has shown superior phase-change heat transfer performance than the normal surface in all three boiling regimes. Several important discoveries are listed as follows: 1. The anodized aluminum oxide (AAO) nanoporous surface has proven to survive the cryogenic temperature environment and also demonstrated its potential to be used in future space exploration applications. 2. The nanoporous surface results in a 160% improvement in the CHF and a 32 K increase in the Leidenfrost temperature, leading to a reduction in chilldown time by 20%. 3. The most significant enhancement effects produced by the nanoporous surface are the increase in the Leidenfrost temperature and much higher heat fluxes in the transition and nucleate boiling regimes. Conflict of interest None declared. Acknowledgment This research was partially supported by the Andrew H. Hines, Jr./Progress Energy Endowment Fund. References [1] R. Shaeffer, H. Hu, J.N. Chung, An experimental study on liquid nitrogen pipe chilldown and heat transfer with pulse flows, Int. J. Heat Mass Transfer 67 (2013) 955–966.
643
[2] E.G. Brentari, J.A. Brennan, R.V. Smith, W.G. Steward, An Experimental Report on Cooldown of Cryogenic Transfer Lines, National Bureau of Standards, Boulder, Colorado, 1966. [3] O. Kawanami, T. Nishida, I. Honda, Y. Kawashima, H. Ohta, Flow and heat transfer on cryogenic flow boiling during tube quenching under upward and downward flow, Microgravity Sci. Technol. 19 (3–4) (2007) 137–138. [4] B. Verthier, G.P. Celata, G. Zummo, C. Colin, J. Follet, Effect of gravity on film boiling heat transfer and rewetting temperature during quenching, Microgravity Sci. Technol. 21 (S1) (2009) 185–191. [5] C. Luo, M. Xiang, Angle inequality for judging the transition from Cassie–Baxter to Wenzel states when a water drop contacts bottoms of grooves between micropillars, Langmuir: ACS J. Surf. Colloids 28 (38) (2012) 13636–13642. [6] M. Xiang, A. Wilhelm, C. Luo, Existence and role of large micropillars on the leaf surfaces of The President lotus, Langmuir: ACS J. Surf. Colloids 29 (25) (2013) 7715–7725. [7] C. Luo, M. Xiang, X. Heng, A stable intermediate wetting state after a water drop contacts the bottom of a microchannel or is placed on a single corner, Langmuir: ACS J. Surf. Colloids 28 (25) (2012) 9554–9561. [8] S.K. Singh, S. Khandekar, D. Pratap, S.A. Ramakrishna, Wetting dynamics and evaporation of sessile droplets on nano-porous alumina surfaces, Colloids Surf. A 432 (2013) 71–81. [9] N. Tasaltin, D. Sanli, A. Jonas, A. Kiraz, C. Erkey, Preparation and characterization of superhydrophobic surfaces based on hexamethyldisilazane-modified nanoporous alumina, Nanoscale Res. Lett. 6 (2011) 487. [10] M.K. Scott, G. Liter, Pool-boiling CHF enhancement by modulated porous-layer coating- theory and experiment, Int. J. Heat Mass Transfer 44 (2001) 4277– 4310. [11] M. Arik, A. Bar-Cohen, S.M. You, Enhancement of pool boiling critical heat flux in dielectric liquids by microporous coatings, Int. J. Heat Mass Transfer 50 (5– 6) (2007) 997–1009. [12] Y.-Y. Li, Z.-H. Liu, G.-S. Wang, L. Pang, Experimental study on critical heat flux of high-velocity circular jet impingement boiling on the nano-characteristic stagnation zone, Int. J. Heat Mass Transfer 67 (2013) 560–568. [13] C.-C. Hsu, P.-H. Chen, Surface wettability effects on critical heat flux of boiling heat transfer using nanoparticle coatings, Int. J. Heat Mass Transfer 55 (13–14) (2012) 3713–3719. [14] H. O’Hanley, C. Coyle, J. Buongiorno, T. McKrell, L.-W. Hu, M. Rubner, R. Cohen, Separate effects of surface roughness, wettability, and porosity on the boiling critical heat flux, Appl. Phys. Lett. 103 (2) (2013) 024102. [15] H.T. Phan, N. Caney, P. Marty, S. Colasson, J. Gavillet, Surface wettability control by nanocoating: the effects on pool boiling heat transfer and nucleation mechanism, Int. J. Heat Mass Transfer 52 (23–24) (2009) 5459– 5471. [16] A. Hens, R. Agarwal, G. Biswas, Nanoscale study of boiling and evaporation in a liquid Ar film on a Pt heater using molecular dynamics simulation, Int. J. Heat Mass Transfer 71 (2014) 303–312. [17] C. Kruse, T. Anderson, C. Wilson, C. Zuhlke, D. Alexander, G. Gogos, S. Ndao, Extraordinary shifts of the Leidenfrost temperature from multiscale micro/ nanostructured surfaces, Langmuir: ACS J. Surf. Colloids 29 (31) (2013) 9798– 9806. [18] X. Li, W. Wei, R. Wang, Y. Shi, Numerical and experimental investigation of heat transfer on heating surface during subcooled boiling flow of liquid nitrogen, Int. J. Heat Mass Transfer 52 (2009) 1510–1516. [19] H.J.A.P. van de Boogaard, A.R. Balkenende, M. Scholten, N.P. Willard, Evaluation of different approaches to assess the surface tension of low-energy solids by means of contact angle measurements, Langmuir: ACS J. Surf. Colloids 14 (20) (1998) 5907–5912. [20] O.R. Burggraf, An exact solution of the inverse problem in heat conduction theory and applications, J. Heat Transfer 86 (3) (1962) 373–380. [21] J.H. Lienhard, L.C. Witte, On the existence of two ’transition’ boiling curves, Int. J. Heat Mass Transfer 25 (6) (1982) 771–779.