Microelectronics Reliability 44 (2004) 315–321 www.elsevier.com/locate/microrel
Experimental study on the thermal performance of micro-heat pipe with cross-section of polygon Seok Hwan Moon, Gunn Hwang *, Sang Choon Ko, Youn Tae Kim Microsystem Team, Human Information Department, Electronics and Telecommunications Research Institute, 161 Gajong-Dong, Yusong-Gu, Daejon 305-350, South Korea Received 12 February 2003; received in revised form 4 May 2003
Abstract The electronic industries have been tried to develop the electronic packaging technology for high-density, light weight, and miniaturization of mobile personal information terminals such as notebook PC, personal digital assistant (PDA), or handheld PC as the performance and the portability of the system have been become important more and more. The generated heat from the CPU of the electronic system should be dissipated effectively for the stability in operation and the long life-time of the system. In the present study, micro-heat pipes (MHPs) with cross-section of polygon have been manufactured and tested for operating characteristics and heat transfer limit. An additional process for installing a wick that is an inevitable component of heat pipes is not required for the MHP. For a small-sized heat pipe in the present study, a high precision technology is needed in manufacturing process. The MHPs tested in this paper have a triangular cross-section with curved sides and a rectangular cross-section with curved sides. The material of the MHP is copper and the working fluid of it is pure water. The test was performed in a vacuum chamber to minimize heat loss. The operating temperatures of the MHP were considered from 60 to 90 °C. From the test results, the MHP with triangular cross-section can dissipate up to a thermal load of 7 W. Ó 2003 Elsevier Ltd. All rights reserved.
1. Introduction The major trends in mobile information technology have been the miniaturization and the performance of mobile information device, such as handheld PC, personal digital assistant (PDA), and notebook PC. The power consumption and the heat generation per unit area of the device have been increased with the technology development. The importance of the thermal management in the mobile device will be growing more and more for reliability and life-time, especially for highly integrated electronic systems.
*
Corresponding author. Tel.: +82-42-860-5395; fax: +82-42860-6836. E-mail addresses:
[email protected] (S.H. Moon),
[email protected] (G. Hwang),
[email protected] (S.C. Ko), ytkim@ etri.re.kr (Y.T. Kim).
It is difficult to apply the traditional cooling techniques, e.g. heat sink, heat pipe, fan, and etc., to the ultra-slim mobile devices. It may be possible for the heat pipe with diameter of 3–5 mm to be pressed to 2–3 mm thickness to apply to the ultra-slim device. But there still remains a problem that the maximum pressed thickness of the heat pipe is constrained by the cross-sectional area of the pipe that is closely related with the performance of the heat pipe [1]. Considering the trend of miniaturized packaging technology for cooling components used for a very limited space, it is inevitable to introduce a microcooling device to those cases. The one of the promising techniques is a micro-heat pipe (MHP) Technology. Many papers about the MHP have been reported since Cotter [2] proposed a novel research in 1984, but a few papers have been dealt with experimental data. In this study, the MHPs with polygon cross-section and equivalent diameter of 1.5 mm were designed and manufactured. The target of this study is to analyze the thermal characteristics of the MHPs with triangular
0026-2714/$ - see front matter Ó 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0026-2714(03)00160-4
316
S.H. Moon et al. / Microelectronics Reliability 44 (2004) 315–321
cross-section and the one with rectangular cross-section. Those results can be used as the basic data to design cooling modules with MHPs applicable to the compact slim type mobile system, such as sub-notebook PC, tablet PC, mobile phone, PDA, handheld PC, and etc.
2. Experimental investigation The size of the MHP developed in this study is smaller than that of the miniature heat pipe with the diameter of 3–4 mm. The MHP could be manufactured using current mechanical technology of the simple manufacturing process. Therefore, this manufacturing process of the MHP has a good productivity. In fact, the micro-structures of the MHP may be manufactured with etching process [3], however that process has disadvantages in the view of productivity and cost. The container of the MHP was manufactured with drawing process and Fig. 1 shows the cross-section of it. This MHP does not have additional wick installed on the inner wall of the general heat pipe, but has sharp corners made with structural deformation of its wall which roles as wick. The condensed liquid at the condenser of the MHP returns to the evaporator by capillary force of the liquid. It is important for MHP to have sharp edge in the corners for making the working liquid return from the condenser to the evaporator. The rectangular MHP with curved sides (Fig. 1(a)) has advantages for thermal performance, such as larger inner space for vapor flowing and one more corner for liquid path than those of the triangular MHP (Fig. 1(b)). Because the edge angle at the corner of the triangular MHP is sharper than that of the rectangular MHP, the capillary force of the working fluid at the triangular MHP is larger than
Table 1 Experimental specification of micro-heat pipe Triangular MHP Total length (mm) Evaporator length (mm) Adiabatic section length (mm) Condenser length (mm) Working fluid Fill ratio of working fluid (%) Number of corner Container material Container manufacturing method
Curved rectangular MHP
50/100 10 15 25 Pure water 20 3 4 Oxygen free copper Drawing
that at the rectangular MHP. The specification of the MHP is shown in Table 1.
3. Experimental setup and procedure The testing apparatus for the thermal performance of the MHP was composed of a MHP, a vacuum chamber unit, a constant temperature bath for cooling the MHP, a data acquisition system, and a DC power supplying unit as shown in Fig. 2. The evaporator of the MHP was heated using the electric resistance heater and DC power supply unit. The wire with 0.36 mm diameter and 10 X/m resistance per meter, as an heater, was wound around the copper block with an interval of 0.5 mm for supplying the constant thermal load, which was attached on the outer wall of the evaporator. The condenser of MHP is cooled by the water jacket with circulating water. Thermal grease (0.74 W/m °C) was filled between the pipe wall and
Fig. 1. Cross-section of micro-heat pipe: (a) curved rectangular type and (b) curved triangular type.
S.H. Moon et al. / Microelectronics Reliability 44 (2004) 315–321
Fig. 2. Experimental apparatus.
the water jacket in order to minimize the thermal contact resistance. The vacuum chamber (102 –103 Torr) made of acryl was used to minimize the heat loss to the environment as shown in Fig. 2. This vacuum chamber with a cylindrical body could be set rotationally for the inclination angle tests. To measure the wall temperature of the MHP, K-type thermocouples (£ 0.08 mm) were bonded by soldering at two points at the evaporator wall, one point at the adiabatic section, and two points at the condenser. The locations of the thermocouples are shown as X in Fig. 2 for the MHP with the total length of 50 mm. In case of 100 mm length, the distance between the locations of the thermocouples would be two times longer than that for the case of 50 mm length. The measured temperatures were recorded by using the data acquisition system. The MHP is made of oxygen free copper and the wall thickness of it is between 0.27 and 0.28 mm. Pure water was used for the working fluid, which has relatively large surface tension and latent heat characteristic from 30 to 160 °C of the operating temperature. The filling ratio of the working fluid was 20% to the inner volume of the MHP. The liquid blocking region where the vapor could not be reached may be created at the condenser, the end of the MHP with 1–2 mm of the equivalent diameter. The heat transfer by phase change of the working fluid could not be accomplished in the liquid blocking region. The temperature at the end of the condenser is 5–20 °C lower than that of the area adjacent to the condenser, and the heat transfer rate of the MHP is decreased due to the liquid blocking. Therefore the inactivated amount of the working fluid due to the liquid blocking should be considered at the design step of the MHP for calculating the filling ratio of the working fluid. The generation mechanism for the liquid blocking has not been reported in detail at any papers and still remained as an undeveloped field.
317
The experimental test was performed to investigate the thermal performance of the MHP. The operating temperature of the MHP identical with the temperature at the adiabatic section of the MHP was considered for the four cases of 60, 70, 80, and 90 °C. The temperature and the amount of the coolant circulating between the water jacket and the constant temperature bath, were controlled carefully to maintain the conditions of the constant operating temperature of the MHP. The thermal load supplied to the MHP was increased stepwise by 0.5 W from 0.5 W. The wall temperature of the MHP was recorded at the steady state by each thermal load step. The measurement was stopped when the wall temperature of the evaporator of the MHP was rapidly increased due to dryout. The wall temperatures of the evaporator, the adiabatic section and the condenser of the MHP were averaged in the each zone. This test was measured in the chamber with vacuum condition of 102 –103 Torr. The results of the present study are included the errors in the measurement, i.e., the tolerance in the heat supply (0.05 V for voltage, 0.01 A for current) and that in the temperature measurement (0.1 °C).
4. Results and discussion 4.1. Heat transfer characteristics A heat pipe can transport a large amount of heat with a slight temperature difference between the evaporator and the condenser. In general, one of the test procedures to check whether a non-condensable gas exists in the heat pipe or a heat pipe is operated well just after the end of manufacturing process, is to measure the temperature difference between the evaporator and the condenser. In that case, however, we should remember that each heat pipe has a temperature difference as oneÕs own. In the case of a small-sized heat pipe like the one in the present study, the high precision technologies are needed in the manufacturing process because the presence of non-condensable gases or contaminants can be detrimental to the heat pipe performance though the amount of those are small. Fig. 3 shows the temperature distribution by the axial length of 50 mm. The tested MHP has a curved triangular cross-section and a 20% filling ratio to the inner total volume. The heat was dissipated only at the condenser with the conduction heat transfer. The temperatures were averaged over 60 s after steady state to reduce minor temperature fluctuation error. As shown in Fig. 3, the wall temperatures of the MHP are increased as the thermal load is increased. This means that the thermal equilibrium, namely isothermal property of the MHP from the evaporator to the condenser is well
318
S.H. Moon et al. / Microelectronics Reliability 44 (2004) 315–321 150
102 Adiabatic section
Condenser
120 105
Q in=0.5W Q in=1W Q in=2W Q in=3W Q in=4W
90 75 60 45
Adiabatic section
Evaporator 99
Wall Temperature ( oC)
Wall Temperature ( o C)
135 Evaporator
Condenser Q in=4W Q in=5W Q in=6W Q in=7W Q in=8W Q in=9W
96 93 90 87
30 0
10
20
30
40
84
50
0
Axial Length (mm)
10
20
30
40
50
Axial Length (mm)
accomplished. The temperature differences between the evaporator and the condenser were 4.3–9.8 °C over the thermal loads of 0.5–4 W. However, the temperature difference of 9.8 °C between the evaporator and the condenser in the thermal load of 1 W is higher than that in other thermal loads. This is due that the amount of latent heat to be transported toward the condenser is small by insufficient vaporization at the evaporator, and the thermal resistance is high by a relatively thick liquid film under low thermal load near 1 W. Fig. 4 shows the temperature distribution by the axial length at the operating temperature of 90 °C that equal to the temperature at the adiabatic section. The tested MHP is the same as the one in Fig. 3. It is seen that the temperature difference between the evaporator and the condenser is increased as the thermal load is increased at the constant operating temperature of 90 °C. It can be explained that the vapor flow velocity is increased as the thermal load is increased. Therefore the friction force on vapor–liquid interface and the pressure drop in liquid flow, are increased. Because the space for the vapor flow in the MHP is narrower than that in the conventional heat pipe, the pressure drop by the friction on vapor– liquid interface may largely affect to the MHP performance [4]. Fig. 5 shows the effect of the inclination angle on the thermal performance of the triangular MHP. In Fig. 5, the negative inclination angle means a top heating mode in which the evaporator is located higher than the condenser, and conversely, the positive inclination angle means a bottom heating mode in which the evaporator is located lower than the condenser. As shown in Fig. 5, the effect of the inclination angle on the thermal performance was small. The thermal performance of the MHP was almost same for the tilting mode from the horizontal mode to the top heating mode with negative 90°. However, there was a decrease of the thermal performance of the MHP as the inclination of the MHP was
Fig. 4. Wall temperature distribution along the longitudinal axis of the MHP at Tv ¼ 90 °C.
25
Heat Transfer Limit (W)
Fig. 3. Wall temperature distribution along the longitudinal axis by thermal loads.
Curved triangular MHP Fill ratio of 20%
20
15
10
5
0 -90
-45
0
45
90
Inclination Angle (degree) Fig. 5. Thermal performance by inclination angle.
rotated from the bottom heating mode to the top heating mode. Namely, the thermal performance of the triangular MHP was very stable in the bottom heating mode and it is seen that the capillary force of the working fluid was enough to flow from the condenser to the evaporator. The triangular MHP has the limiting power of 4.51 W at the top heating mode of negative 90°. Fig. 6 shows the overall heat transfer coefficient according to the total length of the triangular MHP. The considered lengths of the MHP were 50 and 100 mm. In the case of the MHP with a small sized equivalent diameter smaller than 2 mm, the effect of the pipe length on the thermal performance of the MHP could be large. This is due that the pressure losses by friction at the vapor–liquid interface and the capillary limitation for returning of condensed liquid are significantly dominant as an increase of the pipe length. As shown in Fig. 6, the
S.H. Moon et al. / Microelectronics Reliability 44 (2004) 315–321
319
2o
25
L total=100mm L total=50mm
1200
Thermal Resistance ( o C/W)
Overall Heat Transfer Coefficient (W/m C)
1400
1000 800 600 400 200
Curved rectangular MHP Curved triangular MHP
20
15
10
5
0
0
0 0
1
2
3
1
4
4
5
6
7
8
Fig. 7. Performance comparison by cross-section type.
Fig. 6. Overall heat transfer coefficient by total length.
thermal performance of the triangular MHP tends to be increased according to the decrease of the pipe length. In the case of the triangular MHP, the overall heat transfer coefficient was enhanced about 92% when the total length was decreased from 100 to 50 mm for the thermal load of 3 W. In the future, more detailed experimental results for the effect of the pipe length on the thermal performance of the MHP will be studied. 4.2. Heat transfer limit When the temperature at the evaporator lowest end for the bottom heating mode is abruptly increased compared to other temperatures at the evaporator, this phenomenon is defined as dryout state. The thermal load just prior to the state in which heat transfer by phase change cannot be conducted any more due to a dryout in the evaporator, is defined as the heat transfer limit. Fig. 7 shows thermal resistances and heat transfer limits for the triangular MHP and the rectangular MHP. The thermal resistance can be calculated by (1), Te Tc Q
3
Input Power (W)
Input Power (W)
R¼
2
ð1Þ
where Te and Tc are wall temperatures at the evaporator and the condenser of the MHP respectively, and QðW Þ is a thermal load at the evaporator. The tested MHP has a fill ratio of 20% to the internal total volume of the MHP. The operating temperature is not constant but increased as the thermal load is increased. The heat dissipating at the condenser of the MHP was accomplished by circulating of 20 °C water which was controlled by a constant temperature bath. As shown in Fig. 7, the heat transfer limit of the triangular MHP is 1.6 times larger than that of the rectangular MHP. The heat transfer limits were 4.5 and 7 W
for the rectangular MHP and the triangular MHP, respectively. This result is due that the corners for the rectangular MHP case is not developed sharply compared to that for the triangular MHP, and the capillary force needed for returning condensed liquid to the evaporator cannot be obtained sufficiently. Because the rectangular MHP has one more corner than those of the triangular MHP, this property may make an advantage in thermal performance. However, because a radius of curvature at corner is not small sufficiently for retaining capillary pressure, the performance of the rectangular MHP cannot be superior to that of the triangular MHP. The performance of the MHP is largely restricted by the capillary limit [5]. The factor which affects mainly on the capillary limit is the radius of curvature at corner. The radius of curvature ðrÞ is a function of a corner aperture angle ð/Þ, a contact angle ðaÞ, and a location of the meniscus contact point in the corner ðvÞ as shown in (2) [6]. Except the two factors of a and v which is constant as a heat pipe type and operating conditions, / is the most important factor to the radius of curvature. The smaller the corner aperture angle is, the smaller the radius of curvature is. Then high capillary pumping pressure and high maximum heat transfer limit can be obtained under this condition. r ¼ f ð/; a; vÞ
ð2Þ
The corner aperture angle of 60–70° in the rectangular MHP of the present study is larger than that of 30– 40° in the analytical result of Zaghdoudi et al. [6], therefore the small radius of curvature cannot be obtained. Fig. 8 presents the experimental thermal resistance and heat transfer limit of the triangular MHP for various operating temperatures. The tested MHP has a fill ratio of 20% and was performed for the various
320
S.H. Moon et al. / Microelectronics Reliability 44 (2004) 315–321
Thermal Resistance ( o C/W)
10 Tv=60 o C Tv=70 o C Tv=80 o C Tv=90 o C
8
Dryout points
6
4
2
Fig. 10. Photograph of cooling module with two triangular MHPs.
0 0
2
4
6
8
12
10
Input Power (W)
Fig. 8. Thermal performance by the operating temperature.
operating temperature of 60, 70, 80, and 90 °C. As shown in Fig. 8, the heat transfer limit is a function of the operating temperature, and increased as the operating temperature is increased. The heat transfer limits were 6.18, 7.59, 8.01, and 10 W for the operating temperatures of 60, 70, 80, and 90 °C, respectively. Fig. 9 presents the experimental results of the present study and Moon et al. [7] for the heat transfer limit. The MHP tested in Moon et al. had a curved triangular cross-section and a stainless steel as container material. The pure water was used as working fluid. As shown in Fig. 9, the heat transfer limit of the present study is 1.7–2.1 times larger than those of Moon et al. over the operating temperature of 60–80 °C. This result means that the large capillary limit in the present study was obtained compared to that of Moon et al. [8]. High productivity and simple manufacturing process are considered and enhanced performance was obtained
Heat Transfer Limit (W)
12 Experimental data(Present study) Experimental data(Moon et al,1999)
10
8
6
4
2 50
60
70
80
90
Operating Temperature ( o C)
Fig. 9. Experimental results comparison between the present study and Moon et al.
compared to that of Moon et al. for the future applications. Fig. 10 shows a cooling module to dissipate heat from a CPU of sub-notebook PC with the two triangular MHPs. In the future, the triangular MHP developed through this study could be applied to the device where thin packaging space is available for cooling module.
5. Conclusions The MHPs with polygonal cross-section applicable to the electronic units with thin structure were manufactured and tested. High productivity and simple manufacturing process were considered for the future applications. The manufactured MHP was shown the good isothermal property over the total length, and the temperature differences between the evaporator and the condenser were about 4–6 °C. The effect of the inclination angle on the thermal performance was slight and the thermal characteristic was stable over from the top heating mode to the bottom heating mode. The effect of the total pipe length on the thermal performance of the triangular MHP was dominant. In the case of the triangular MHP, the overall heat transfer coefficient was enhanced about 92% when the total length was decreased from 100 to 50 mm for 3 W of the thermal load. The heat transfer limit of the triangular MHP was 1.6 times larger than that of the rectangular MHP. The heat transfer limits were 4.5 and 7 W for the rectangular MHP and the triangular MHP, respectively. The heat transfer limit was the function of the operating temperature and increased as the operating temperature was increased. The maximum heat transfer limit of the triangular MHP was 10 W for the operating temperatures of 90 °C. The heat transfer limit of the present study was 1.7–2.1 times larger than that of Moon et al. over the operating temperature of 60–80 °C. The manufactured MHP in the present study was shown a superior heat dissipation capacity therefore it will be able to be widely used in integrated electronic units as a cooling module.
S.H. Moon et al. / Microelectronics Reliability 44 (2004) 315–321
References [1] Moon SH, Hwang G, Yun HG, Choy TG, Kang YI. Improving thermal performance of miniature heat pipe for notebook PC cooling. Microelectron Reliab 2002;42:135–40. [2] Cotter TP. Principles and prospects for micro heat pipes. In: Proceedings of the 5th International Heat Pipe Conference, 1984. [3] Gerner FM. Flow limitation in micro heat pipes. AFSOR Final Report No. F49620-88-6-0053, Wright-Patterson, AFB, Dayton, OH, 1989. [4] Xie H, Aghazadeh M, Toth J. The use of heat pipes in the cooling of portables with high power packages. Thermacore Co. Technical Note.
321
[5] Hopkins R, Faghri A, Khrustalev D. Flat miniature heat pipe with micro capillary grooves. Trans ASME 1999;121: 102–9. [6] Zaghdoudi MC, Sartre V, Lallemand M. Theoretical investigation of micro heat pipes performance. In: 10th International Heat Pipe Conference, Germany, 21–25 September, F-9, 1997. [7] Moon SH et al. An experimental study on the performance limitation of a micro heat pipe with triangular cross-section. In: 11th International Heat Pipe Conference, Japan, 12–16 September, A1-4, 1999. [8] Moon SH, Hwang G, Kim YT, Seo JK, Kim SJ. Thermal analysis and testing of a miniature heat pipe with woven wire wick. ETRI J, in press.
