Paper No. IMECE

Transcription

1 Proceedings of the ASME International Mechanical Engineering Congress & Exposition IMECE2010 November 12-18, 2010, Vancouver, British Columbia Paper No. IMECE SUBMERGED JET IMPINGEMENT COOLING OF A NANOSTRUCTURED PLATE Muhsincan Sesen Ali Kosar Ebru Demir Evrim Kurtoglu Nazli Kaplan Hadi Cagdas Erk Wisam Khudhayer University of Arkansas at Little Rock Little Rock, AR, 72204, USA Tansel Karabacak University of Arkansas at Little Rock Little Rock, AR, 72204, USA ABSTRACT In this paper, the results of a series of heat transfer experiments conducted on a compact electronics cooling device based on single phase jet impingement techniques are reported. Deionized-water is propelled into four microchannels of inner diameter 685µm which are used as nozzles and located at a nozzle to surface distance of 2.5mm. The generated jet impingement is targeted through these channels towards the surface of a nanostructured plate. This plate of size 20mmx20mm consisted of ~600 nm long copper nanorod arrays with an average nanorod diameter of ~150 nm, which were integrated on top of a silicon wafer substrate coated with a copper thin film layer (i.e. Cu-nanorod/Cu-film/Silicon-wafer). Heat removal characteristics induced through jet impingement are investigated using the nanostructured plate and compared to results obtained from a flat plate of copper thin film coated on silicon wafer surface. Enhancement in heat transfer up to 15% using the nanostructured plate has been reported in this paper. Heat generated by small scale electronic devices is simulated using a thin film heater placed on an aluminum base. Surface temperatures are recorded by a data acquisition system with the thermocouples integrated on the surface at various locations. Constant heat flux provided by the film heater is delivered to the nanostructured plate placed on top of the base. Volumetric flow rate and heat flux values were varied in order to better characterize the potential enhancement in heat transfer by nanostructured surfaces. NOMENCLATURE Symbol Description A Surface area of the nanostructured plate A c Total cross sectional area of the jets P Power input q Heat flux T s Surface temperature d i Nozzle diameter h Heat transfer coefficient Nu Nusselt number Re Reynolds number k Thermal conductivity T i Inlet fluid temperature ν Kinematic viscosity u Exit velocity Volumetric flow rate 1.0 INTRODUCTION In terms of the capability of providing high heat transfer rates, jet impingement is one of the most efficient cooling Paper No. IMECE Copyright 2010 by ASME

2 mechanisms. Jet impingement cooling not only offers high heat transfer rates but also has the benefit of removing all thermal interface resistances between the surface and the cooling fluid [1]. In a wide range of industrial applications such as annealing of metals [2], cooling of gas turbine blades [3], cooling in grinding processes [4] and cooling of photovoltaic cells [5]; jet impingement cooling becomes a preferential method for the manufacturers. For instance, in gas turbine applications, this cooling method has been used for a long time in order to assure durability during long operating intervals [2]. Moreover, impingement systems play an important role in micro scale applications such as cooling of electronic components, microprocessors and MEMS devices [1]. Recently, nanostructured surfaces have been utilized to achieve high heat transfer performance with enhanced heat transfer area and positive effect on heat transfer coefficients with diminishing length scale [6, 7]. In order to keep up with the miniaturization process, heat transfer and fluid flow at micro and nano scale have been rigorously studied in the literature to achieve higher heat removal capabilities [8]. Flow and heat transfer characteristics of multiple impinging jets can differ substantially from those of single jets depending mainly on geometrical conditions, and it is clear that the higher the number of jets in the array and the smaller the jet diameter, the higher the heat transfer rates are [1]. Multiple jet flows interact with each other, therefore employing jet arrays becomes considerably complex or even erroneous compared to single jet configurations. While heat transfer rates due to single jets can be functionally expressed by relatively simple powerfunctions of Reynolds (Re) and Prandtl number, correlations for multijet heat transfer rates require the consideration of a number of additional characteristic numbers [9]. Heat transfer in jet impingement systems are also greatly influenced by nozzle geometry. In previous studies reported in the literature, for a constant Reynolds number, decreasing the jet diameter yields higher stagnation and average heat transfer coefficients [10-12]. This can be attributed to the higher jet velocities created by the smaller nozzles [2]. In this study, a nanostructured surface with nanorods was used and compared with the results obtained from the same surface without the nanostructures. In addition to this, multiple impinging jets were used instead of a single jet since heat transfer under an impinging jet is very high in the stagnation zone, but it decreases quickly away from the jet [1]. Employed multiple jet arrays increase the number of available stagnation zones, thus they enhance the heat transfer from the impingement surface. It has been shown that nanostructured surfaces augment the heat transfer performance of jet impingement cooling systems, and this study reveals the advantages of using nanostructured surfaces and multiple impinging jets in microscale cooling. 2.0 OVERVIEW ON NANOSTRUCTURED PLATES The glancing angle deposition (GLAD) technique is a physical self-assembly growth technique that can provide a novel capability for growing 3D nanostructure arrays with interesting material properties such as high electrical/thermal conductivity and also reduced oxidation compared to the polycrystalline films [13-15]. It is a simple and single-step process that offers a cost and time efficient method to fabricate nanostructured arrays of almost any material in the periodic table as well as alloys and oxides. The GLAD technique uses the shadowing effect, which is a physical self-assembly through which some of the obliquely incident atoms may not reach certain points on the substrate due to the concurrent growth of parallel structures. Due to the statistical fluctuations in the growth and effect of initial substrate surface roughness, some rods grow faster in the vertical direction. Due to their higher height, they capture the incident particles, while the shorter rods get shadowed and cannot grow anymore. This leads to the formation of isolated nanostructures. The shadowing effect can be controlled by adjusting the deposition rate, incidence angle, substrate rotation speed, working gas pressure, substrate temperature, and the initial surface topography of the substrate. 3.0 NANOSTRUCTURE DEPOSITION The schematic of the custom-made GLAD experimental setup in the present study is shown in Fig.1. For the fabrication of vertically aligned Cu nanorod arrays, the DC magnetron sputter GLAD technique is employed. Cu nanorods were deposited on the native oxide p-si (100) substrates (2 cm 2 ) using a 99.9% pure Cu cathode (diameter about 7.6 cm). The substrates were mounted on the sample holder located at a distance of about 12 cm from the cathode. During the growth, Figure 1. A schematic of the glancing angle deposition (GLAD) technique used for the fabrication of vertical nanorod arrays is shown. the substrate was tilted so that the angle θ between the surface normal of the target and the surface normal of the substrate is 87 o. The substrate was attached to a stepper motor and rotated at a speed of 2 rpm for growing vertical nanorods. The depositions were performed under a base pressure of 5 x 10-7 Torr which was achieved by utilizing a turbo-molecular pump backed by a mechanical pump. During Cu deposition experiments, the power was 200 W with an ultrapure Ar working gas pressure of 2.5 mtorr. The deposition time of GLAD Cu nanorods was 75 min. For comparison purposes, conventional smooth Cu thin film samples (i.e. plain surface configuration) were also prepared by normal incidence deposition (θ = 0 o ) with a substrate rotation of 2 rpm. Deposition rate of the vertical nanorods was measured utilizing Paper No. IMECE Copyright 2010 by ASME

3 quartz crystal microbalance (Inficon- Q-pod QCM monitor, crystal: 6 MHz gold coated standard quartz) measurements and cross-sectional scanning electron microscopy (SEM) image analysis to be about 8.6 nm/min. The SEM unit (FESEM- 6330F, JEOL Ltd, Tokyo, Japan) was used to study the morphology of the deposited nanorods. The top and side view SEM images of Cu nanorods are shown in Fig. 2 in which an isolated columnar morphology can be seen. However, for the conventional Cu film deposited at normal incidence, its surface was observed to be flat as indicated by the SEM images (not shown here). sharp tips. This property will allow reduced surface oxidation which can greatly increase the thermal conductivity, robustness, and resistance to oxidation-degradation of our nanorods in the present study. 4.0 EXPERIMENTAL SETUP AND PROCEDURE Experimental Setup The main components constituting the cooling system are an aluminum base containing the film heater, a nanostructured plate placed on top of it, four microchannels providing the impinging jets, and thermocouples (Fig. 3). The aluminum base of dimensions 25mmx20mmx5mm holds the miniature film heater which is treated with thermal grease and sealed to the base with an aluminum cap in order to enhance heat transfer rate and minimize heat losses (Fig. 4). It provides constant heat flux to the system due to the constant voltage applied to its both ends, simulating the heat generated by microchips/microprocessors. On top of the base, the nanostructured copper plate of dimensions 20mmx20mm is placed which greatly enhances heat transfer surface area (Fig. 4). The plate is also treated with thermal grease so as to improve the efficiency of the cooling process by enhancing the heat transfer rate. The whole setup is carefully sealed to prevent any leakages. Jet impingement targeted on the nanostructured plate carry the unwanted heat away from the plate effectively. The impinging jets are provided by four microchannels of inner diameter 685 µm that are connected to the experimental setup using a multiple element sealing. DI-water is driven into the channels using a micro gear pump that can be precisely tuned with a controller which allows the conduction of the experiment at different flow rates. Flow meters integrated to the system are used to measure the volumetric flow rate through the channels. To determine the pressure drop across the setup, a pressure gauge is attached to the inlet. Outside pressure is assumed to be atmospheric pressure. Thermocouples placed on the surface of the miniature film heater are used to acquire accurate steady surface temperature data. Figure 2. a) Top and b) cross-section views of glancing angle deposited (GLAD) Cu nanorods. Scale bar is 1μm. At early stages of GLAD growth, the number density of the nanorods was larger, and they had diameters as small as about 5-10 nm. As they grew longer and some of the rods stopped growing, due to the shadowing effect, their diameter grew up to about 150 nm. The average height of the individual rod was measured to be about 600 nm and the average gap among the nanorods also changed with their length from 5-10 nm up to nm at later stages. As can be seen from Fig. 2a, the top of the vertical nanorods has a pyramidal shape with four facets, which indicates that an individual nanorod has a single crystal structure. This observation was confirmed by previous studies [16-18] which reported that individual metallic nanorods fabricated by GLAD are typically single crystal. Single crystal rods do not have any interior grain boundaries and have faceted Figure 3. Experimental Setup Paper No. IMECE Copyright 2010 by ASME

4 Data is gathered through data acquisition system (NI-SCXI 1000). Data acquisition system records 100 data points per second at 100Hz sampling rate. These data points are then exported through data acquisition software LABVIEW for averaging via MS Visual Studio and MATLAB software. Experimental Procedure After the experimental setup is prepared as explained, the surface temperatures are measured as a function of the input power data gathered from the readings of the power supply. This procedure is carried out for various flow rates, which are obtained from the inlet region of the setup. In addition to the measurements of flow rates and power values, inlet temperatures, surface temperatures, pressure drops across the system and currents flowing through the film heater were also measured with the appropriate sensors. This procedure is then executed for the samples of nanostructured plate and the plate with plain surface in order to investigate the potential positive effects of the nanostructured plate on heat transfer. The effective areas of the heaters are tabulated in the manufacturer s guide, from which their values are extracted. These values are used to analytically calculate the constant heat flux input to the system. where d i is the inside diameter of each nozzle, k is the thermal conductivity of the fluid. The velocity, u, is expressed as where is the flow rate of the water and is the total crossectional area of nozzles. Reynolds number, Re, is given as where is the kinematic viscosity of the working fluid. Uncertainty Analysis The uncertainties of the measured values are given in Table 1 and are derived from the manufacturer s specification sheet while the uncertainties of the derived parameters are obtained using the propagation of uncertainty method developed by Kline and McClintock [19]. Uncertainty Power Area Nozzle Diameter Temperature Volumetric Flow Rate Heat Flux Heat Transfer Coefficient Nusselt Number Velocity of Water Reynolds Number Error ±0.01W ±0.01mm 2 ±0.01mm ±0.1ºC ±1ml/min 0.04% 0.17% 1.47% 0.35% 1.5% Table 1. Uncertainty Figures in Data Figure 4. Cooling device exploded view Data Reduction Heat flux provided to the system, q, is obtained from where P is the power input and A is the heated area of the plate. The heat transfer coefficient, h, is then calculated by where T s is the surface temperature and T i is the inlet fluid temperature. Nusselt number, Nu, is extracted from 5.0 RESULTS & DISCUSSION The experimental results are obtained as explained in Section 4 above. Data points for the constant volumetric flow rate of 218 ml/min are given in Fig. 5. Heat transfer coefficient, h, which quantifies the convective heat transfer capability of a heat sink, is plotted against the heat flux provided from the miniature film heater. As can be seen from Fig. 5, heat transfer coefficient values increase with the input heat flux since the experiments are carried out only in the single phase region. Enhanced heat transfer coefficients can be observed from the figure with the application of the nanostructured plate. The nanostructured plate enhances the heat transfer area available to remove heat from the surface of the base. The nanorods also act as pin-fins in improving heat transfer rates from the surface to the liquid jets due to the minimized heat transfer resistance induced by the presence of a thin layer of the fluid on the subjected surfaces, which can be easily broken on the Paper No. IMECE Copyright 2010 by ASME

5 h (W/m²K) Nusselt Number h (W/m²K) Nusselt Number nanostructured surfaces compared to plain surfaces, thereby further contributing to heat transfer Figure 5. Heat transfer coefficient versus heat flux is plotted at a constant volumetric flow rate of 218 ml/min. The results for a higher constant volumetric flow rate 288ml/min are quite similar to the results for the 218ml/min flow rate (Fig. 6). As the flow rate is increased, further enhancement in heat transfer can be observed from the increased heat transfer coefficients. It can also be noted that the improvement in the heat transfer coefficient on nanostructured plate compared to the plain surface configuration tends to further increase for higher heat fluxes in all the results. Enhancement in heat transfer coefficient up to 15% has been observed in this experimental study Heat Transfer Coefficient vs. Heat constant Q=218 ml/min Heat Transfer Coefficient vs. Heat constant Q=288 ml/min Figure 6. Heat transfer coefficient versus heat flux is plotted at a constant volumetric flow rate of 288 ml/min. Nusselt number, Nu, profiles are displayed in Figure 7 and Figure 8 for two different flow rates. In these figures, an increase in Nusselt number with the nanostructured surface is apparent. Nusselt numbers are greater for the higher flow rate when compared to the lower flow rate at the same heat flux. This result implies Nusselt number has a strong dependence on Reynolds number, which is also extensively reported [20]. 3,25 3 2,75 2,5 2,25 2 Figure 7. Nusselt number versus heat flux is plotted at a constant volumetric flow rate of 218ml/min. 3,25 3 2,75 2,5 2,25 2 Nusselt Number vs. Heat constant Q=218ml/min Nusselt Number vs. Heat constant Q=288ml/min Figure 8. Nusselt number versus heat flux is plotted at a constant volumetric flow rate of 288ml/min. The enhanced heat transfer performance with nanostructured surfaces in jet impingement agrees with the results on nanostructured surfaces in other heat transfer modes. Pool boiling and single phase flow in a rectangular channel with nanostructured surfaces were also investigated and enhancements in heat transfer were reported [6, 7]. 6.0 CONCLUSION The results gathered from our experimental work indicate the advantageous effects of a nanostructured plate on heat transfer enhancement via a single-phase submerged jet impingement cooling device. The vertical nanorods integrated Paper No. IMECE Copyright 2010 by ASME

6 to the copper thin film layer on silicon wafer introduce enhanced surface area and lower heat transfer resistance which is induced by the presence of a thin layer of fluid formed on its surface compared to the plain surface. The nanorods also act as nanoscale pin-fins which significantly contributes to heat transfer enhancement with the nanostructured surface. By using a compact setup with integrated nanostructures, the cooling system acts more efficiently. In the light of these tabulated results, more in-depth systematic studies to control the length, spacing, and diameter of nanorods are critically important of fundamental understanding of heat transfer occurring on the nanostructured surfaces as well as to clarify the potential benefits/limitations of this technology in various cooling applications of small electronic devices, such as micro reactors, micro propulsion, biotechnology, fuel cells and air conditioning. Moreover, there is a need to develop empirical correlations for calculating the heat transfer coefficients of the nanostructured surfaces which can be useful in designing such devices for cooling applications. REFERENCES [1] Agostini, B., Fabbri, M., Park, J. E., Wojtan, L., Thome J. R., Michel B., 2007, State of the Art of High Heat Flux Cooling Technologies, Heat transfer Engineering, 28(4), pp [2] Glynn, C., O Donovan, T., Murray, D. B., Feidt, M., 2005, Jet Impingement Cooling, UK Heat Transfer Conference [3] Cardenas, R., Mani, P., Narayanan, V., 2010, Saturated Mini Jet Impingement Boiling, Proceedings of ASME th International Conference on Nanochannels, Microchannels, and Minichannels, ICNMM [4] Babic, D., Murray, D. B., Torrance, A. A., 2005, Mist Jet Cooling of Grinding Processes, Int. J. Mach. Tools Manufact., 45, pp [5] Royne, A., and Dey, C., 2004, Experimental Study of a Jet Impingement Device for Cooling of Photovoltaic Cells Under High Concentration, Proceedings of Solar 2004: Life, the Universe and Renewables. [6] Sesen, M., Khudhayer, W., Karabacak, T., Kosar, A., 2010, A Compact Nanostructure Integrated Pool Boiler for Microscale Cooling Applications, Micro&Nano Letters, (in press). [7] Sesen, M., Khudhayer, W., Karabacak, T., Kosar, A. B., Ahishalioglu, B. A., Kosar, A., 2010, A Compact Nanostructure Enhanced Heat Sink with Flow in a Rectangular Channel, Proceedings of the ASME th Biennial Conference on Engineering Systems Design and Analysis, ESDA [8] Zhang, Z. M., 2007, "Nano/microscale heat transfer," McGraw-Hill. [9] Weigand, B. and Spring, S., 2009, Multiple Jet Impingement A Review, Int. Symp on Heat Transfer in Gas Turbine Systems. [10] Womac, D. J., Aharoni, G., Ramadhyani, S., Incropera, F. P., 1990, "Single Phase Liquid Jet Impingement Cooling of Small Heat Sources," Proceedings of the International Heat Transfer Conference, pp [11] Garimella, S. V. and Rice, R. A., 1995, "Confined and Submerged Liquid Jet Impingement Heat Transfer," Journal of Heat Transfer, 117, pp [12] Garimella, S. V., Nenaydykh, B., 1996, "Nozzle-Geometry Effects in Liquid Jet Impingement Heat Transfer," International Journal of Heat and Mass Transfer, 39(14), pp [13] Robbie, K., Beydaghyan, G., Brown, T., Dean, C., Adams, J. and Buzea, C., 2004, Rev. Sci. Instrum [14] Karabacak, T. and Lu, T.-M., 2005, Handbook of Theoretical and Computational Nanotechnology ed M Rieth and W Schommers, Stevenson Ranch, CA: American Scientific Publishers, 69, p [15] Karabacak, T., Wang, G. C. and Lu, T.-M., 2004, Physical self-assembly and the nucleation of 3D nanostructures by oblique angle deposition, J. Vac. Sci. Technol. A 22(4), pp [16] Karabacak, T., DeLuca, J. S., Ye, D., Wang, P.-I., Wang G.-C., and Lu T.-M., 2006, Low temperature melting of copper nanorod arrays, J. Appl. Phys. 99(6), [17] Karabacak, T., Wang, P.-I., Wang, G.-C, and Lu, T.-M., 2005, Phase transformation of single crystal β-tungsten nanorods at elevated temperatures, Thin Solid Films, 493(1-2), pp [18] Karabacak, T., Wang, P.-I., Wang, G.-C., and Lu, T.-M, 2004, Growth of Single Crystal Tungsten Nanorods by Oblique Angle Sputter Deposition, Mat. Res. Soc. Symp. Proc.,788, 75. [19] Kline, S. J., and McClintock, F. A., 1953, "Describing Uncertainties in Single-Sample Experiments," Mech. Eng., p. 3 [20] Incropera, F. P. and Dewitt, D. P., 2006, "Fundamentals of Heat and Mass Transfer 6th Edition," Wiley. Paper No. IMECE Copyright 2010 by ASME