Experimental Study on Thermal Performance of a ...

Proceedings of the ASME 2009 2nd Micro/Nanoscale Heat & Mass Transfer International Conference MNHMT2009 December 18-21, 2009, Shanghai, China

MNHMT2009-18525 EXPERIMENTAL STUDY ON THERMAL PERFORMANCE OF A SILICON-BASED MICRO PULSATING HEAT PIPE Jian QU, Huiying WU*, Ping CHENG School of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China ACT ABSTR ABSTRA In this paper, an experimental investigation was conducted on the thermal performance of a silicon-based micro-pulsating heat pipe (SMPHP) using FC-72 and R113 as working fluids. The SMPHP, covering an area of 46×19mm2, consisted of fourteen meandering trapezoidal channels with a hydraulic diameter of 352μm. The effects of gravity, filling ratio, and working fluids on the overall thermal resistance of the SMPHP were discussed. Experimental results show that gravity had an impact on the thermal performance of the SMPHP, and selfsustained oscillation could not be achieved at the horizontal orientation. The SMPHP worked as a true pulsating device when the filling ratio varied from 30% to 65%. For FC-72 and R113, there was an optimal filling ratio of 55% and 41%, respectively, for the best thermal performance of the SMPHP. As compared to the SMPHP with 0% filling ratio (or charged with the air), the thermal resistances of the SMPHP charged with FC-72 (at 55% filling ratio) and R113 (at 41% filling ratio) were decreased maximally by 7.24°℃/W (or 56.5%) and 7.51℃/W (or 59.7%), respectively. It is also found that the R113 was favorable for the operation of the SMPHP at lower power inputs, while FC-72 was favorable at relatively higher power inputs. Keywords: micro pulsating heat pipe, silicon wafer, thermal performance, filing ratio, inclining angle 1. INTRODUCTION With the trends towards greater miniaturization and higher integration of semiconductor components, microelectronic chips are undergoing a continuously rapid rise in heating power density, and thus engendering severe thermal management problems. Since microelectronic chip cooling is usually conducted within a very limited space, it is inevitable that micro cooling devices are to be applied. Micro heat pipes were

considered to be one of the most promising options for microelectronic cooling [1, 2] due to high efficiency, simple structure and low cost. Micro heat pipes are composed of micro-sized channels with diameters ranging from tens to hundreds of microns, so they are capable of being directly fabricated as an integral part of a semiconductor chip to facilitate direct heat removal and to reduce temperature gradients. By now, several types of heat pipe, including grooved heat pipe, capillary pump loop (CPL), and loop heat pipe (LHP), have been miniaturized and fabricated in silicon wafers [3-6]. As a new member of the wickless heat pipes, the pulsating heat pipes (PHPs) with higher effective thermal conductivity than the conventional heat pipes are also suitable for the miniaturization. Borgmeyer and Ma [7] fabricated a pulsating heat pipe with the size of 76.2×76.2mm2 in a flat plate copper using a milling machine. The channels in their PHP had a square cross-section of 1.5785×1.5875mm2. Most recently, a smaller pulsating heat pipe with length, width, and internal diameter of 56mm, 50mm and 2mm respectively was manufactured using polydimethylsiloxane (PDMS) by Lin et al. [8]. The experiments showed that gravity had an impact on the performance of these two PHPs. Sugimoto et al. [9] proposed a micro counter-stream-mode oscillating flow heat pipe with a groove dimension of 575×400μm2 fabricated in a silicon wafer. Though the device worked like a PHP with oscillating flow between the heat source and heat sink, it is not the true passive PHP since the oscillation motion was assisted by two dual piezoelectric pumps, which drove the fluid by push-pull operation scheme. In this paper, a silicon-based micro pulsating heat pipe (SMPHP) in a true passive working mode was fabricated and developed with the aid of the MEMS (Micro Electro Mechanical System) technology. The micro structure, fabrication process, as well as the measured thermal

1

Copyright © 2009 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 03/16/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

performance of the micro device are presented. This study provides us with an opportunity to understand the heat transfer performance of a micro pulsating heat pipe in a smaller scale and will promote its applications in electronic cooling field. 2. FABRICATION OF SILICON-BASED MICRO PULSATING HEAT PIPE As illustrated in Fig. 1, the silicon-based micro pulsating heat pipe, covering an area of 46×19mm2, was composed of a pair of silicon wafer and Pyrex 7740 glass. Firstly, fourteen meandering microchannels, which were connected by fourteen U-shape turns distributed evenly at the evaporator and condenser, were etched on the silicon wafer by the MEMS technology. Then the silicon wafer with the parallel microchannels was bonded from the top by a thin Pyrex glass plate, which enables the visualization of the two-phase flow in the pulsating heat pipe. Note that a hole was manufactured on the glass plate before its bonding with the silicon wafer, through which the SMPHP could be evacuated and charged with the working fluid.

Fig. 3 illustrates the experimental system, which was composed of the test section, data acquisition system, DC power supply (GPR-3060D, GW), water cold bath (DC-0506, BILANG), and a video recording system. The test section, with the SMPHP in it, could be rotated from the horizontal plane (with inclination angle θ = 0°) to the vertical plane (with inclination angle θ = 90°). The SMPHP, consisting of evaporation, adiabatic and condensation sections with 8, 23 and 15mm in length, respectively. The evaporator (i.e., the evaporation section) was heated from the bottom by a film heater, which was energized by the DC power supply. The voltage and current were measured by a digital multimeter (380282, EXTECH INSTRUMENTS Co.), from which the power input to the SMPHP can be determined. The condenser (i.e., the condensation section) was cooled by the cooling water pumped from the cold bath. The inlet temperature and flow rate of the cooling water were kept constant during the experiment. To measure the wall temperature, six OMEGA T-type thermocouples with a diameter of 0.1mm and an accuracy of ±0.1℃, being evenly spaced from the heating end to the cooling end, were attached at the bottom of the silicon wafer. The output signals from the thermocouples were collected through a computerized data acquisition system (34970A, Agilent). The two-phase flow pattern in the SMPHP during operation was recorded by a CCD camera (TK-C1381, JVC), which was located perpendicularly to the SMPHP.

Fig. 1 Fabrication process of a silicon-based micro pulsating heat pipe Fig. 2 gives the schematic diagram of the trapezoidal crosssection of the microchannels etched in the silicon wafer. The top width, bottom width, depth and hydraulic diameter of the parallel trapezoidal microchannels were 820μm, 294μm, and 303μm, 352μm, respectively. The distance between two neighboring microchannels was 578μm. Fig. 3 Schematic diagram of the experimental setup

Fig. 2 Schematic diagram of the cross section of the SMPHP 3. DESCRIPTION OF THE EXPERIMENT

In this study, FC-72 and R113 were chosen as the working fluids due to their relatively lower boiling point. Before the experiment, the microchannels in the SMPHP were firstly evacuated by a vacuum pump, and then they were charged with the working fluid by a special syringe. As long as a desired filling ratio (defined as the filled liquid volume divided by the total inside volume of the SMPHP) reached, the charging tube was sealed. In this paper, the filling ratio for FC-72 and R113 were in the ranges of 0% ~ 70%, and 0% ~76%, respectively.

2



The overall thermal resistance (R) is an important parameter normally used to evaluate the thermal performance of the SMPHP, which was defined by:

R=

Te − Tc Q

(1)

where Te, Tc and Q are the evaporator wall temperature, condenser wall temperature, and heating power input, respectively. Because the SMPHP was well thermally insulated at the evaporation and adiabatic sections, and the heat loss from these two sections to the ambience was negligibly small, the heat transported by the SMPHP, Q, can be calculated by:

Q = UI

(2)

where U and I are the input voltage and current at the evaporator. As shown in Fig. 3, thermocouple #1 measured the evaporator wall temperature, thermocouples #2, #3 and #4 measured the wall temperatures at the adiabatic sections, and thermocouples #5 and #6 measured the condenser wall temperature. Thus, the evaporator temperature Te and condenser temperature Tc in Eq. (1) could be substituted by T1 and (T5+T6)/2, respectively. Substituting Eq. (2), T1, and (T5+T6)/2 into Eq. (1), the thermal resistance can be written as:

R=

T1 − (T5 + T6 ) / 2 UI

(3)

4. RESULTS AND DISCUSSION 4.1 Effect of the inclination angle In a PHP, liquid slug and vapor plugs oscillate between the evaporator and condenser. The formation of liquid slugs in a channel or tube greatly depends on the Bond number (i.e., Bo = d g ( ρl − ρ v ) / σ l ), which is a measure of the

orientation, it could not persist for a long time. If the oscillation was disrupted occasionally, the fluid motion often stopped at once and could not continue. However, when the SMPHP was held at an inclination angle with respect to the horizontal plane while other conditions remain the same, oscillation motion started again. Fig. 4 shows the thermal performance of the SMPHP at different inclination angles charged with FC-72 and R113 under filling ratios of 42% and 41%, respectively. Since almost no oscillation movement was observed when the inclination angle was 30° or less, the SMPHP worked mainly based on heat conduction and had larger thermal resistance (i.e., a lower thermal performance). However, when the inclination angle increased from 45° to 90° both for FC-72 and R113, the SMPHP worked with self-sustained oscillation and had a better thermal performance. Clearly, the inclination angle had a more important impact on the thermal performance of the SMPHP charged with R113 when compared with FC-72 at relatively lower power inputs as shown in Fig. 4(b), from which it is found that the inclination angle of 45 ° is an optimal value for enhancing heat transfer performance. Besides, it is also found in Fig. 4 that the dry-out appeared within the experimental range, which can be explained as follows: the increase of the power input increases the evaporator temperature, which leads to a faster evaporation of the liquid film. Although the evaporator was intermittently flushed by the liquid through the slug oscillation motion, the flushing frequency decreased as the evaporator temperature increased and could not replenish the evaporator eventually. As a result, the dry-out occurred. Fig. 4 shows that the onset of dry-out with respect to the power input depends on the inclination angle. At a larger inclination angle, the onset of dry-out was often delayed at a higher power input.

importance of gravitation force compared to surface tension forces. Khandekar et al. [10] constructed a circular tube PHP with a Bo no larger than 2 (or a critical Bo of 2) for the formation of liquid slug and vapor plug system. For a PHP with sharp angle corner cross section, Yang et al. [11] proposed that a larger number more than 2 should be used as the critical Bo. For the SMPHP used in this experiment, the Bo is less than 0.5 when FC-72 or R113 were used as working fluids, which indicates that the surface tension completely dominates in the liquid slug and vapor plug system. Interestingly, however, the experimental results show that the influence of gravity on the thermal performance of the SMPHP could not be neglected as well and will be discussed in the following. Under some filling ratios, although oscillation motion with large amplitudes was observed in the SMPHP at horizontal

(a)

3



(b) Fig. 4 Effect of inclination angle on the thermal resistance of the SMPHP charged with (a) FC-72 (φ =42%) and (b) R113 (φ =41%)) 4.2 Effect of the filling ratio Different from a circular tube PHP, the liquid in the PHP with sharp angled corners tends to accumulate in the corners, which gives rise to some capillary action and has a fundamental

(a) φ =23%

influence on the operating characteristics [11]. Clearly, the SMPHP with the trapezoidal cross section will generate larger capillary force and allow the liquid phase to be trapped in the corners. As a result, the SMPHP behaves as an interconnected array of capillary assisted micro two-phase thermosyphons when the filling ratio is low (≤ 30%). The falling liquid film was unstable and often disrupted by the randomly formed liquid bridge due to liquid-vapor interface perturbation. For a high filling ratio (≥65%), the fluid oscillation motion in the SMPHP was restrained by the increased flow friction and the reduced movement freedom, and thus only local oscillation motion or even no oscillation was observed at high filling ratios. For the filling ratio between 30% and 65%, the SMPHP worked as a true pulsating device. Fig. 5 shows the flow behavior of the SMPHP charged with FC-72 at the filling ratios of 23%, 55% and 70%, respectively when the micro device was vertical orientated. As shown in Fig. 5(a), the evaporator of the SMPHP at φ = 23% was featured by falling film, and the liquid slugs were pushed to the condenser with small perturbation. At a filling ratio of 70%, the intermittent oscillation with small amplitude only occurred at local part of the channels (see Fig. 5(c)). The self-sustained oscillation motion with large amplitude dominated in the channels when φ = 55% (see Fig. 6(b)).

(b) φ =55%

(c) φ =70%

Fig. 5 The flow behavior of FC-72 in the SMPHP with the filling ratios of (a) 23%, (b) 55% and (c) 70% at the vertical orientation Fig. 6 shows the thermal resistance as a function of power input with various filling ratios at the vertical orientation. It is found that although the effect of filling ratio on the thermal performance was seemed to be disordered at relatively low power inputs, there existed an optimal value corresponding to the best thermal performance at higher power inputs. At relatively higher power inputs, the thermal resistances of the SMPHP charged with FC-72 (Fig. 6(a)) and R113 (Fig. 6(b)) decreased firstly with increasing the filling ratio, but increased with further increasing the filling ratio. As to R113, for instance, the thermal resistance decreased with the increase of the filling ratio from 0% to 41% when the power input was larger than

5.4W, but it increased with the further increase of the filling ratio from 41% to 76%. Thus, φ = 41% was the optimal filling ratio for the best thermal performance for R113. For the SMPHP charged with FC-72, the optimal value was about 55%. As compared with the SMPHP charged with only air (φ = 0%), the thermal resistance of the SMPHP charged with FC-72 could be maximally decreased by 7.24℃/W (or 56.5%) at the optimal filling ratio of 55%, while for the SMPHP charged with R113, the thermal resistance could be maximally decreased by 7.51℃/W (or 59.7%) at the optimal filling ratio of 41%.

4



(a)

(b) Fig. 6 Thermal performance of the SMPHP charged with (a) FC72 and (b) R113 at different filling ratios at the vertical orientation (β = 90°) 4.3 Effect of the working fluid Thermophysical properties of the working fluids also have impact on the two phase flow and thermal performance in the PHP [12]. In this paper, FC-72 and R113 were chosen as the working fluids due to their favorable properties for the SMPHP. Fig. 7 shows the effect of working fluids on the thermal performance of the SMPHP. Since the filling ratio is difficult to be precisely controlled during the charging process, there is a tiny difference (less than 2%) in the filling ratios between FC-72 and R113, however, this difference was small and would not cause a sensitive influence on the thermal performance. It is seen from Fig. 7 that though the SMPHP charged with R113 suffered the onset of dry-out earlier (especially at the inclination angle of 45°), it showed better thermal performance at most power inputs as compared with that of FC-72 at both inclination angles ( 45°

and 90°). Table 1 gives the evaporator temperatures with respect to the power inputs of 4.1, 5.4, and 6.3W at inclination angles of 45° and 90° both for FC-72 and R113. As consistent with the thermal resistance, relatively lower evaporator temperatures were obtained for the SMPHP charged with R113 as compared with that charged with FC-72 except for the inclination angle of 45° when φ was 53%. At power inputs of 4.1W and 5.4W and filling ratio of 41%, using R113 instead of FC-72 could decrease the evaporator wall temperature by 17.7℃ and 17.5℃, respectively. However, when the power inputs exceeded 7.6W, the SMPHP charged with R113 operated inefficiently since dryout occurred. This is different from the SMPHP charged with FC-72, which could endure higher power inputs. Thus, while R113 was favorable for the operation of the SMPHP at lower power inputs, FC-72 could work more efficiently at relatively higher power inputs.

Fig. 7 thermal performance of the SMPHP charged with FC-72 (φ= 42% and 55%) and R113 (φ= 41% and 53%)

5. CONCLUSIONS A silicon-based micro pulsating heat pipe (SMPHP) consisting of meandering trapezoidal channels having a hydraulic diameter of 352μm was fabricated and tested. The following conclusions can be obtained: (1) Oscillation motion can be activated in a micro pulsating heat pipe at a Bond number less than 0.5. (2) Gravity had an impact on the performance of the SMPHP. A SMPHP could not work and self-sustained at the horizontal orientation. For a relatively larger inclination angle, the onset of dry-out often appeared at a higher power input. (3) The SMPHP worked as a true pulsating device when charged with FC-72 or R113 at the filling ratio ranging from 30% to 65%. There was an optimal filling ratio corresponding to the best thermal performance, which was 55% for FC-72 and 41% for R113 respectively. (4) At the optimal filling ratio (55% for FC-72 and 41% for R113), the thermal resistances of the SMPHP charged with

5



FC-72 and R113 decreased maximally by 7.24℃/W (or 56.5%) and 7.51℃/W (or 59.7%), respectively, as compared with those charged with only air.

(5) While R113 was favorable for the operation of the SMPHP at lower power inputs, FC-72 was more suitable due to its better thermal performance at relatively higher power inputs.

Table 1 Evaporator temperature at inclination angles of 45° and 90°

ACKNOWLEGEMENTS This work was supported by the National Natural Science Foundation of China through grant no. 50925624, the Program for New Century Excellent Talents in University of China through grant no. NCET-06-0406, the Shanghai Municipal Education Commission through grant nos. 08ZZ10 and 08GG05, as well as the Shanghai Municipal Project through grant no. 049.

experimental study”, Applied Thermal Engineering 23 (2003), 707-719. [11] H Yang, S Khandekar, M Groll, “Performance characteristics of pulsating heat pipes as integral thermal spreaders”, 48 (2009), 815-824. [12] S Khandekar, N Dollinger, M Groll, “Understanding operational regimes of closed loop pulsating heat pipes: an experimental study”, Applied Thermal Engineering, 23 (2003), 707-719.

REFERENCES [1] C B Sobhan, R L Rag, G P Peterson, “A review and comparative study of the investigations on micro heat pipes”, Int. J. Energy Res. 31 (2007), 664-688. [2] Jian Qu, Huiying Wu, Ping Cheng, “Effects of functional surface on performance of a micro heat pipe”, Int. Comm. Heat and Mass Transfer 35 (2008), 523-528. [3] G P Peterson, A B Duncan, M H Weichold, “Experimental investigation of micro heat pipes fabricated in silicon wafers”, J. Heat Transfer, 115 (1993), 750-756. [4] M K Berre, S Launay, V Sartre, M Lallemand, “Fabrication and experimental investigation of silicon micro heat pipes for cooling electronics”, 13 (2003), 436-441. [5] J Kirshberg, K Yerkes, D Liepmann, “Micro-cooler for chip-level temperature control”, SAE Aerospace Power Systems Conference 1999, 341. [6] Chin-Chun Hsu, Shung-Wen Kang, Tung-Fu Hou, “Performance Testing of Micro loop Heat Pipes”, Tamkang J. Sci. Eng. 8 (2005), 123-132 [7] B Borgmeyer, H B Ma, “Experimental investigation of oscillating motions in a flat plate pulsating heat pipe”, AIAA Journal of Thermophysics and heat transfer 21 (2007), 405-409. [8] Yu-Hsing Lin, Shung-Wen Kang, Tsung-Yu Wu, “Fabrication of polydimethylsiloxane (PDMS) pulsating heat pipe”, Applied Thermal Engineering 29 (2009), 573-580. [9] M Sugimoto, K Minai, M Uemura, S Nishio, O Tabata, “A novel micro counter-stream-mode oscillating-folw (COSMOS) heat-pipe”, Micro Electro Mechanical Systems, 18th IEEE International Conference on 2005, 606-609. [10] S Khandekar, N Dollinger, M Groll, , “ Understanding operational regimes of closed loop pulsating heat pipes: an

6

